From dd17c8f72993f9461e9c19250e3f155d6d99df22 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Thu, 29 Oct 2009 22:34:15 +0900 Subject: percpu: remove per_cpu__ prefix. Now that the return from alloc_percpu is compatible with the address of per-cpu vars, it makes sense to hand around the address of per-cpu variables. To make this sane, we remove the per_cpu__ prefix we used created to stop people accidentally using these vars directly. Now we have sparse, we can use that (next patch). tj: * Updated to convert stuff which were missed by or added after the original patch. * Kill per_cpu_var() macro. Signed-off-by: Rusty Russell Signed-off-by: Tejun Heo Reviewed-by: Christoph Lameter --- kernel/rcutorture.c | 8 ++++---- kernel/trace/trace.c | 6 +++--- kernel/trace/trace_functions_graph.c | 4 ++-- 3 files changed, 9 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 178967b6434e..e339ab349121 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -731,13 +731,13 @@ static void rcu_torture_timer(unsigned long unused) /* Should not happen, but... */ pipe_count = RCU_TORTURE_PIPE_LEN; } - __this_cpu_inc(per_cpu_var(rcu_torture_count)[pipe_count]); + __this_cpu_inc(rcu_torture_count[pipe_count]); completed = cur_ops->completed() - completed; if (completed > RCU_TORTURE_PIPE_LEN) { /* Should not happen, but... */ completed = RCU_TORTURE_PIPE_LEN; } - __this_cpu_inc(per_cpu_var(rcu_torture_batch)[completed]); + __this_cpu_inc(rcu_torture_batch[completed]); preempt_enable(); cur_ops->readunlock(idx); } @@ -786,13 +786,13 @@ rcu_torture_reader(void *arg) /* Should not happen, but... */ pipe_count = RCU_TORTURE_PIPE_LEN; } - __this_cpu_inc(per_cpu_var(rcu_torture_count)[pipe_count]); + __this_cpu_inc(rcu_torture_count[pipe_count]); completed = cur_ops->completed() - completed; if (completed > RCU_TORTURE_PIPE_LEN) { /* Should not happen, but... */ completed = RCU_TORTURE_PIPE_LEN; } - __this_cpu_inc(per_cpu_var(rcu_torture_batch)[completed]); + __this_cpu_inc(rcu_torture_batch[completed]); preempt_enable(); cur_ops->readunlock(idx); schedule(); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 85a5ed70b5b2..b808177af816 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -91,12 +91,12 @@ DEFINE_PER_CPU(int, ftrace_cpu_disabled); static inline void ftrace_disable_cpu(void) { preempt_disable(); - __this_cpu_inc(per_cpu_var(ftrace_cpu_disabled)); + __this_cpu_inc(ftrace_cpu_disabled); } static inline void ftrace_enable_cpu(void) { - __this_cpu_dec(per_cpu_var(ftrace_cpu_disabled)); + __this_cpu_dec(ftrace_cpu_disabled); preempt_enable(); } @@ -1085,7 +1085,7 @@ trace_function(struct trace_array *tr, struct ftrace_entry *entry; /* If we are reading the ring buffer, don't trace */ - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return; event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry), diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 90a6daa10962..8614e3241ff8 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -176,7 +176,7 @@ static int __trace_graph_entry(struct trace_array *tr, struct ring_buffer *buffer = tr->buffer; struct ftrace_graph_ent_entry *entry; - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return 0; event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_ENT, @@ -240,7 +240,7 @@ static void __trace_graph_return(struct trace_array *tr, struct ring_buffer *buffer = tr->buffer; struct ftrace_graph_ret_entry *entry; - if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled)))) + if (unlikely(__this_cpu_read(ftrace_cpu_disabled))) return; event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_RET, -- cgit v1.2.2 From f409adf5b1db55ece7e80b67a944f9c0d3fe93e9 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 1 Dec 2009 14:02:00 +0100 Subject: futex: Protect pid lookup in compat code with RCU find_task_by_vpid() in compat_sys_get_robust_list() does not require tasklist_lock. It can be protected with rcu_read_lock as done in sys_get_robust_list() already. Signed-off-by: Thomas Gleixner Cc: Darren Hart Cc: Peter Zijlstra --- kernel/futex_compat.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/futex_compat.c b/kernel/futex_compat.c index 235716556bf1..d49afb2395e5 100644 --- a/kernel/futex_compat.c +++ b/kernel/futex_compat.c @@ -146,7 +146,7 @@ compat_sys_get_robust_list(int pid, compat_uptr_t __user *head_ptr, struct task_struct *p; ret = -ESRCH; - read_lock(&tasklist_lock); + rcu_read_lock(); p = find_task_by_vpid(pid); if (!p) goto err_unlock; @@ -157,7 +157,7 @@ compat_sys_get_robust_list(int pid, compat_uptr_t __user *head_ptr, !capable(CAP_SYS_PTRACE)) goto err_unlock; head = p->compat_robust_list; - read_unlock(&tasklist_lock); + rcu_read_unlock(); } if (put_user(sizeof(*head), len_ptr)) @@ -165,7 +165,7 @@ compat_sys_get_robust_list(int pid, compat_uptr_t __user *head_ptr, return put_user(ptr_to_compat(head), head_ptr); err_unlock: - read_unlock(&tasklist_lock); + rcu_read_unlock(); return ret; } -- cgit v1.2.2 From 86fc80f16e8a2449d5827bf1a9838b7fd9f70097 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 9 Dec 2009 17:13:31 +0100 Subject: capabilities: Use RCU to protect task lookup in sys_capget cap_get_target_pid() protects the task lookup with tasklist_lock. security_capget() is called under tasklist_lock as well but tasklist_lock does not protect anything there. The capabilities are protected by RCU already. So tasklist_lock only protects the lookup and prevents the task going away, which can be done with rcu_read_lock() as well. Signed-off-by: Thomas Gleixner Signed-off-by: James Morris --- kernel/capability.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/capability.c b/kernel/capability.c index 7f876e60521f..9e4697e9b276 100644 --- a/kernel/capability.c +++ b/kernel/capability.c @@ -135,7 +135,7 @@ static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp, if (pid && (pid != task_pid_vnr(current))) { struct task_struct *target; - read_lock(&tasklist_lock); + rcu_read_lock(); target = find_task_by_vpid(pid); if (!target) @@ -143,7 +143,7 @@ static inline int cap_get_target_pid(pid_t pid, kernel_cap_t *pEp, else ret = security_capget(target, pEp, pIp, pPp); - read_unlock(&tasklist_lock); + rcu_read_unlock(); } else ret = security_capget(current, pEp, pIp, pPp); -- cgit v1.2.2 From d894837f23f491aa7ed167aae767fc07cfe6e6e6 Mon Sep 17 00:00:00 2001 From: Simon Kagstrom Date: Wed, 23 Dec 2009 11:08:18 +0100 Subject: sched: might_sleep(): Make file parameter const char * Fixes a warning when building with g++: warning: deprecated conversion from string constant to 'char*' And the file parameter use is constant, so mark it as such. Signed-off-by: Simon Kagstrom Cc: peterz@infradead.org LKML-Reference: <20091223110818.442d848e@marrow.netinsight.se> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index c535cc4f6428..64298a52eaa6 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -9694,7 +9694,7 @@ static inline int preempt_count_equals(int preempt_offset) return (nested == PREEMPT_INATOMIC_BASE + preempt_offset); } -void __might_sleep(char *file, int line, int preempt_offset) +void __might_sleep(const char *file, int line, int preempt_offset) { #ifdef in_atomic static unsigned long prev_jiffy; /* ratelimiting */ -- cgit v1.2.2 From e1783a240f491fb233f04edc042e16b18a7a79ba Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 5 Jan 2010 15:34:50 +0900 Subject: module: Use this_cpu_xx to dynamically allocate counters Use cpu ops to deal with the per cpu data instead of a local_t. Reduces memory requirements, cache footprint and decreases cycle counts. The this_cpu_xx operations are also used for !SMP mode. Otherwise we could not drop the use of __module_ref_addr() which would make per cpu data handling complicated. this_cpu_xx operations have their own fallback for !SMP. V8-V9: - Leave include asm/module.h since ringbuffer.c depends on it. Nothing else does though. Another patch will deal with that. - Remove spurious free. Signed-off-by: Christoph Lameter Acked-by: Rusty Russell Signed-off-by: Tejun Heo --- kernel/module.c | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index e96b8ed1cb6a..9bf228052ec5 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -474,9 +474,10 @@ static void module_unload_init(struct module *mod) INIT_LIST_HEAD(&mod->modules_which_use_me); for_each_possible_cpu(cpu) - local_set(__module_ref_addr(mod, cpu), 0); + per_cpu_ptr(mod->refptr, cpu)->count = 0; + /* Hold reference count during initialization. */ - local_set(__module_ref_addr(mod, raw_smp_processor_id()), 1); + __this_cpu_write(mod->refptr->count, 1); /* Backwards compatibility macros put refcount during init. */ mod->waiter = current; } @@ -619,7 +620,7 @@ unsigned int module_refcount(struct module *mod) int cpu; for_each_possible_cpu(cpu) - total += local_read(__module_ref_addr(mod, cpu)); + total += per_cpu_ptr(mod->refptr, cpu)->count; return total; } EXPORT_SYMBOL(module_refcount); @@ -796,14 +797,15 @@ static struct module_attribute refcnt = { void module_put(struct module *module) { if (module) { - unsigned int cpu = get_cpu(); - local_dec(__module_ref_addr(module, cpu)); + preempt_disable(); + __this_cpu_dec(module->refptr->count); + trace_module_put(module, _RET_IP_, - local_read(__module_ref_addr(module, cpu))); + __this_cpu_read(module->refptr->count)); /* Maybe they're waiting for us to drop reference? */ if (unlikely(!module_is_live(module))) wake_up_process(module->waiter); - put_cpu(); + preempt_enable(); } } EXPORT_SYMBOL(module_put); @@ -1394,9 +1396,9 @@ static void free_module(struct module *mod) kfree(mod->args); if (mod->percpu) percpu_modfree(mod->percpu); -#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) +#if defined(CONFIG_MODULE_UNLOAD) if (mod->refptr) - percpu_modfree(mod->refptr); + free_percpu(mod->refptr); #endif /* Free lock-classes: */ lockdep_free_key_range(mod->module_core, mod->core_size); @@ -2159,9 +2161,8 @@ static noinline struct module *load_module(void __user *umod, mod = (void *)sechdrs[modindex].sh_addr; kmemleak_load_module(mod, hdr, sechdrs, secstrings); -#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) - mod->refptr = percpu_modalloc(sizeof(local_t), __alignof__(local_t), - mod->name); +#if defined(CONFIG_MODULE_UNLOAD) + mod->refptr = alloc_percpu(struct module_ref); if (!mod->refptr) { err = -ENOMEM; goto free_init; @@ -2393,8 +2394,8 @@ static noinline struct module *load_module(void __user *umod, kobject_put(&mod->mkobj.kobj); free_unload: module_unload_free(mod); -#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) - percpu_modfree(mod->refptr); +#if defined(CONFIG_MODULE_UNLOAD) + free_percpu(mod->refptr); free_init: #endif module_free(mod, mod->module_init); -- cgit v1.2.2 From 79615760f380ec86cd58204744e774c33fab9211 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 5 Jan 2010 15:34:50 +0900 Subject: local_t: Move local.h include to ringbuffer.c and ring_buffer_benchmark.c ringbuffer*.c are the last users of local.h. Remove the include from modules.h and add it to ringbuffer files. Signed-off-by: Christoph Lameter Signed-off-by: Tejun Heo --- kernel/trace/ring_buffer.c | 1 + kernel/trace/ring_buffer_benchmark.c | 1 + 2 files changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2326b04c95c4..eb6c8988c31a 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -20,6 +20,7 @@ #include #include +#include #include "trace.h" /* diff --git a/kernel/trace/ring_buffer_benchmark.c b/kernel/trace/ring_buffer_benchmark.c index b2477caf09c2..df74c7982255 100644 --- a/kernel/trace/ring_buffer_benchmark.c +++ b/kernel/trace/ring_buffer_benchmark.c @@ -8,6 +8,7 @@ #include #include #include +#include struct rb_page { u64 ts; -- cgit v1.2.2 From 16295bec6398a3eedc9377e1af6ff4c71b98c300 Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Wed, 6 Jan 2010 19:47:10 +1100 Subject: padata: Generic parallelization/serialization interface This patch introduces an interface to process data objects in parallel. The parallelized objects return after serialization in the same order as they were before the parallelization. Signed-off-by: Steffen Klassert Signed-off-by: Herbert Xu --- kernel/Makefile | 1 + kernel/padata.c | 690 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 691 insertions(+) create mode 100644 kernel/padata.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 864ff75d65f2..6aebdeb2aa34 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -100,6 +100,7 @@ obj-$(CONFIG_SLOW_WORK_DEBUG) += slow-work-debugfs.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o +obj-$(CONFIG_PADATA) += padata.o ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra , the -fno-omit-frame-pointer is diff --git a/kernel/padata.c b/kernel/padata.c new file mode 100644 index 000000000000..6f9bcb8313d6 --- /dev/null +++ b/kernel/padata.c @@ -0,0 +1,690 @@ +/* + * padata.c - generic interface to process data streams in parallel + * + * Copyright (C) 2008, 2009 secunet Security Networks AG + * Copyright (C) 2008, 2009 Steffen Klassert + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#define MAX_SEQ_NR INT_MAX - NR_CPUS +#define MAX_OBJ_NUM 10000 * NR_CPUS + +static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index) +{ + int cpu, target_cpu; + + target_cpu = cpumask_first(pd->cpumask); + for (cpu = 0; cpu < cpu_index; cpu++) + target_cpu = cpumask_next(target_cpu, pd->cpumask); + + return target_cpu; +} + +static int padata_cpu_hash(struct padata_priv *padata) +{ + int cpu_index; + struct parallel_data *pd; + + pd = padata->pd; + + /* + * Hash the sequence numbers to the cpus by taking + * seq_nr mod. number of cpus in use. + */ + cpu_index = padata->seq_nr % cpumask_weight(pd->cpumask); + + return padata_index_to_cpu(pd, cpu_index); +} + +static void padata_parallel_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + struct padata_instance *pinst; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, pwork); + pd = queue->pd; + pinst = pd->pinst; + + spin_lock(&queue->parallel.lock); + list_replace_init(&queue->parallel.list, &local_list); + spin_unlock(&queue->parallel.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->parallel(padata); + } + + local_bh_enable(); +} + +/* + * padata_do_parallel - padata parallelization function + * + * @pinst: padata instance + * @padata: object to be parallelized + * @cb_cpu: cpu the serialization callback function will run on, + * must be in the cpumask of padata. + * + * The parallelization callback function will run with BHs off. + * Note: Every object which is parallelized by padata_do_parallel + * must be seen by padata_do_serial. + */ +int padata_do_parallel(struct padata_instance *pinst, + struct padata_priv *padata, int cb_cpu) +{ + int target_cpu, err; + struct padata_queue *queue; + struct parallel_data *pd; + + rcu_read_lock_bh(); + + pd = rcu_dereference(pinst->pd); + + err = 0; + if (!(pinst->flags & PADATA_INIT)) + goto out; + + err = -EBUSY; + if ((pinst->flags & PADATA_RESET)) + goto out; + + if (atomic_read(&pd->refcnt) >= MAX_OBJ_NUM) + goto out; + + err = -EINVAL; + if (!cpumask_test_cpu(cb_cpu, pd->cpumask)) + goto out; + + err = -EINPROGRESS; + atomic_inc(&pd->refcnt); + padata->pd = pd; + padata->cb_cpu = cb_cpu; + + if (unlikely(atomic_read(&pd->seq_nr) == pd->max_seq_nr)) + atomic_set(&pd->seq_nr, -1); + + padata->seq_nr = atomic_inc_return(&pd->seq_nr); + + target_cpu = padata_cpu_hash(padata); + queue = per_cpu_ptr(pd->queue, target_cpu); + + spin_lock(&queue->parallel.lock); + list_add_tail(&padata->list, &queue->parallel.list); + spin_unlock(&queue->parallel.lock); + + queue_work_on(target_cpu, pinst->wq, &queue->pwork); + +out: + rcu_read_unlock_bh(); + + return err; +} +EXPORT_SYMBOL(padata_do_parallel); + +static struct padata_priv *padata_get_next(struct parallel_data *pd) +{ + int cpu, num_cpus, empty, calc_seq_nr; + int seq_nr, next_nr, overrun, next_overrun; + struct padata_queue *queue, *next_queue; + struct padata_priv *padata; + struct padata_list *reorder; + + empty = 0; + next_nr = -1; + next_overrun = 0; + next_queue = NULL; + + num_cpus = cpumask_weight(pd->cpumask); + + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + reorder = &queue->reorder; + + /* + * Calculate the seq_nr of the object that should be + * next in this queue. + */ + overrun = 0; + calc_seq_nr = (atomic_read(&queue->num_obj) * num_cpus) + + queue->cpu_index; + + if (unlikely(calc_seq_nr > pd->max_seq_nr)) { + calc_seq_nr = calc_seq_nr - pd->max_seq_nr - 1; + overrun = 1; + } + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + seq_nr = padata->seq_nr; + BUG_ON(calc_seq_nr != seq_nr); + } else { + seq_nr = calc_seq_nr; + empty++; + } + + if (next_nr < 0 || seq_nr < next_nr + || (next_overrun && !overrun)) { + next_nr = seq_nr; + next_overrun = overrun; + next_queue = queue; + } + } + + padata = NULL; + + if (empty == num_cpus) + goto out; + + reorder = &next_queue->reorder; + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + if (unlikely(next_overrun)) { + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + atomic_set(&queue->num_obj, 0); + } + } + + spin_lock(&reorder->lock); + list_del_init(&padata->list); + atomic_dec(&pd->reorder_objects); + spin_unlock(&reorder->lock); + + atomic_inc(&next_queue->num_obj); + + goto out; + } + + if (next_nr % num_cpus == next_queue->cpu_index) { + padata = ERR_PTR(-ENODATA); + goto out; + } + + padata = ERR_PTR(-EINPROGRESS); +out: + return padata; +} + +static void padata_reorder(struct parallel_data *pd) +{ + struct padata_priv *padata; + struct padata_queue *queue; + struct padata_instance *pinst = pd->pinst; + +try_again: + if (!spin_trylock_bh(&pd->lock)) + goto out; + + while (1) { + padata = padata_get_next(pd); + + if (!padata || PTR_ERR(padata) == -EINPROGRESS) + break; + + if (PTR_ERR(padata) == -ENODATA) { + spin_unlock_bh(&pd->lock); + goto out; + } + + queue = per_cpu_ptr(pd->queue, padata->cb_cpu); + + spin_lock(&queue->serial.lock); + list_add_tail(&padata->list, &queue->serial.list); + spin_unlock(&queue->serial.lock); + + queue_work_on(padata->cb_cpu, pinst->wq, &queue->swork); + } + + spin_unlock_bh(&pd->lock); + + if (atomic_read(&pd->reorder_objects)) + goto try_again; + +out: + return; +} + +static void padata_serial_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, swork); + pd = queue->pd; + + spin_lock(&queue->serial.lock); + list_replace_init(&queue->serial.list, &local_list); + spin_unlock(&queue->serial.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->serial(padata); + atomic_dec(&pd->refcnt); + } + local_bh_enable(); +} + +/* + * padata_do_serial - padata serialization function + * + * @padata: object to be serialized. + * + * padata_do_serial must be called for every parallelized object. + * The serialization callback function will run with BHs off. + */ +void padata_do_serial(struct padata_priv *padata) +{ + int cpu; + struct padata_queue *queue; + struct parallel_data *pd; + + pd = padata->pd; + + cpu = get_cpu(); + queue = per_cpu_ptr(pd->queue, cpu); + + spin_lock(&queue->reorder.lock); + atomic_inc(&pd->reorder_objects); + list_add_tail(&padata->list, &queue->reorder.list); + spin_unlock(&queue->reorder.lock); + + put_cpu(); + + padata_reorder(pd); +} +EXPORT_SYMBOL(padata_do_serial); + +static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst, + const struct cpumask *cpumask) +{ + int cpu, cpu_index, num_cpus; + struct padata_queue *queue; + struct parallel_data *pd; + + cpu_index = 0; + + pd = kzalloc(sizeof(struct parallel_data), GFP_KERNEL); + if (!pd) + goto err; + + pd->queue = alloc_percpu(struct padata_queue); + if (!pd->queue) + goto err_free_pd; + + if (!alloc_cpumask_var(&pd->cpumask, GFP_KERNEL)) + goto err_free_queue; + + for_each_possible_cpu(cpu) { + queue = per_cpu_ptr(pd->queue, cpu); + + queue->pd = pd; + + if (cpumask_test_cpu(cpu, cpumask) + && cpumask_test_cpu(cpu, cpu_active_mask)) { + queue->cpu_index = cpu_index; + cpu_index++; + } else + queue->cpu_index = -1; + + INIT_LIST_HEAD(&queue->reorder.list); + INIT_LIST_HEAD(&queue->parallel.list); + INIT_LIST_HEAD(&queue->serial.list); + spin_lock_init(&queue->reorder.lock); + spin_lock_init(&queue->parallel.lock); + spin_lock_init(&queue->serial.lock); + + INIT_WORK(&queue->pwork, padata_parallel_worker); + INIT_WORK(&queue->swork, padata_serial_worker); + atomic_set(&queue->num_obj, 0); + } + + cpumask_and(pd->cpumask, cpumask, cpu_active_mask); + + num_cpus = cpumask_weight(pd->cpumask); + pd->max_seq_nr = (MAX_SEQ_NR / num_cpus) * num_cpus - 1; + + atomic_set(&pd->seq_nr, -1); + atomic_set(&pd->reorder_objects, 0); + atomic_set(&pd->refcnt, 0); + pd->pinst = pinst; + spin_lock_init(&pd->lock); + + return pd; + +err_free_queue: + free_percpu(pd->queue); +err_free_pd: + kfree(pd); +err: + return NULL; +} + +static void padata_free_pd(struct parallel_data *pd) +{ + free_cpumask_var(pd->cpumask); + free_percpu(pd->queue); + kfree(pd); +} + +static void padata_replace(struct padata_instance *pinst, + struct parallel_data *pd_new) +{ + struct parallel_data *pd_old = pinst->pd; + + pinst->flags |= PADATA_RESET; + + rcu_assign_pointer(pinst->pd, pd_new); + + synchronize_rcu(); + + while (atomic_read(&pd_old->refcnt) != 0) + yield(); + + flush_workqueue(pinst->wq); + + padata_free_pd(pd_old); + + pinst->flags &= ~PADATA_RESET; +} + +/* + * padata_set_cpumask - set the cpumask that padata should use + * + * @pinst: padata instance + * @cpumask: the cpumask to use + */ +int padata_set_cpumask(struct padata_instance *pinst, + cpumask_var_t cpumask) +{ + struct parallel_data *pd; + int err = 0; + + might_sleep(); + + mutex_lock(&pinst->lock); + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) { + err = -ENOMEM; + goto out; + } + + cpumask_copy(pinst->cpumask, cpumask); + + padata_replace(pinst, pd); + +out: + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_set_cpumask); + +static int __padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_active_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_add_cpu - add a cpu to the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to add + */ +int padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_set_cpu(cpu, pinst->cpumask); + err = __padata_add_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_add_cpu); + +static int __padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_online_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_remove_cpu - remove a cpu from the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to remove + */ +int padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_clear_cpu(cpu, pinst->cpumask); + err = __padata_remove_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_remove_cpu); + +/* + * padata_start - start the parallel processing + * + * @pinst: padata instance to start + */ +void padata_start(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags |= PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_start); + +/* + * padata_stop - stop the parallel processing + * + * @pinst: padata instance to stop + */ +void padata_stop(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags &= ~PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_stop); + +static int __cpuinit padata_cpu_callback(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + int err; + struct padata_instance *pinst; + int cpu = (unsigned long)hcpu; + + pinst = container_of(nfb, struct padata_instance, cpu_notifier); + + switch (action) { + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + + case CPU_DOWN_FAILED: + case CPU_DOWN_FAILED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + } + + return NOTIFY_OK; +} + +/* + * padata_alloc - allocate and initialize a padata instance + * + * @cpumask: cpumask that padata uses for parallelization + * @wq: workqueue to use for the allocated padata instance + */ +struct padata_instance *padata_alloc(const struct cpumask *cpumask, + struct workqueue_struct *wq) +{ + int err; + struct padata_instance *pinst; + struct parallel_data *pd; + + pinst = kzalloc(sizeof(struct padata_instance), GFP_KERNEL); + if (!pinst) + goto err; + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) + goto err_free_inst; + + rcu_assign_pointer(pinst->pd, pd); + + pinst->wq = wq; + + cpumask_copy(pinst->cpumask, cpumask); + + pinst->flags = 0; + + pinst->cpu_notifier.notifier_call = padata_cpu_callback; + pinst->cpu_notifier.priority = 0; + err = register_hotcpu_notifier(&pinst->cpu_notifier); + if (err) + goto err_free_pd; + + mutex_init(&pinst->lock); + + return pinst; + +err_free_pd: + padata_free_pd(pd); +err_free_inst: + kfree(pinst); +err: + return NULL; +} +EXPORT_SYMBOL(padata_alloc); + +/* + * padata_free - free a padata instance + * + * @ padata_inst: padata instance to free + */ +void padata_free(struct padata_instance *pinst) +{ + padata_stop(pinst); + + synchronize_rcu(); + + while (atomic_read(&pinst->pd->refcnt) != 0) + yield(); + + unregister_hotcpu_notifier(&pinst->cpu_notifier); + padata_free_pd(pinst->pd); + kfree(pinst); +} +EXPORT_SYMBOL(padata_free); -- cgit v1.2.2 From 809826a389040e0ad9d646b587bccc0e34691afd Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:34 +0800 Subject: tracing: Have __dynamic_array() define a field This is part of a patch set that removes the show_format method in the ftrace event macros. This patch set requires that all fields are added to the ftrace_event_call->fields. This patch changes __dynamic_array() to call trace_define_field() to include fields that use __dynamic_array(). Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D36.8090100@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_export.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index d4fa5dc1ee4e..9978a4f40090 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -175,7 +175,12 @@ ftrace_format_##name(struct ftrace_event_call *unused, \ return ret; #undef __dynamic_array -#define __dynamic_array(type, item) +#define __dynamic_array(type, item) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + 0, is_signed_type(type), FILTER_OTHER);\ + if (ret) \ + return ret; #undef FTRACE_ENTRY #define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -- cgit v1.2.2 From 509e760cd91c831983097ae174cb6c0b8c6c8e6b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:42 +0800 Subject: tracing: Add print_fmt field This is part of a patch set that removes the show_format method in the ftrace event macros. The print_fmt field is added to hold the string that shows the print_fmt in the event format files. This patch only adds the field but it is currently not used. Later patches will use this field to enable us to remove the show_format field and function. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D3E.2000704@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_export.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 9978a4f40090..95d14b640a66 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -203,6 +203,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) return 0; } +#undef __entry +#define __entry REC + #undef __field #define __field(type, item) @@ -218,6 +221,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) #undef __dynamic_array #define __dynamic_array(type, item) +#undef F_printk +#define F_printk(fmt, args...) #fmt ", " __stringify(args) + #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, type, tstruct, print) \ \ @@ -228,6 +234,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .id = type, \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ + .print_fmt = print, \ .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ -- cgit v1.2.2 From 50307a45f8515f6244e3b08e6b19824b9fbfe293 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:45 +0800 Subject: tracing/syscalls: Init print_fmt for syscall events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the syscall events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D41.609@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 68 ++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 67 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f372dd2..1352b0a36fac 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -191,6 +191,67 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) return trace_seq_putc(s, '\n'); } +static +int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) +{ + int i; + int pos = 0; + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s", + entry->args[i], sizeof(unsigned long), + i == entry->nb_args - 1 ? "" : ", "); + } + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, + ", ((unsigned long)(REC->%s))", entry->args[i]); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_syscall_print_fmt(struct ftrace_event_call *call) +{ + char *print_fmt; + int len; + struct syscall_metadata *entry = call->data; + + if (entry->enter_event != call) { + call->print_fmt = "\"0x%lx\", REC->ret"; + return 0; + } + + /* First: called with 0 length to calculate the needed length */ + len = __set_enter_print_fmt(entry, NULL, 0); + + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_enter_print_fmt(entry, print_fmt, len + 1); + call->print_fmt = print_fmt; + + return 0; +} + +static void free_syscall_print_fmt(struct ftrace_event_call *call) +{ + struct syscall_metadata *entry = call->data; + + if (entry->enter_event == call) + kfree(call->print_fmt); +} + int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) { int ret; @@ -386,9 +447,14 @@ int init_syscall_trace(struct ftrace_event_call *call) { int id; + if (set_syscall_print_fmt(call) < 0) + return -ENOMEM; + id = register_ftrace_event(call->event); - if (!id) + if (!id) { + free_syscall_print_fmt(call); return -ENODEV; + } call->id = id; INIT_LIST_HEAD(&call->fields); return 0; -- cgit v1.2.2 From a342a0280b981c130e32dbb94dbd3a57959c4d04 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:49 +0800 Subject: tracing/kprobes: Init print_fmt for kprobe events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the kprobe events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D45.3080100@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 64 ++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 63 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 6ea90c0e2c96..147491dccead 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1250,6 +1250,62 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, ", REC->" FIELD_STRING_RETIP); } +static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) +{ + int i; + int pos = 0; + + const char *fmt, *arg; + + if (!probe_is_return(tp)) { + fmt = "(%lx)"; + arg = "REC->" FIELD_STRING_IP; + } else { + fmt = "(%lx <- %lx)"; + arg = "REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; + } + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\"%s", fmt); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, " %s=%%lx", + tp->args[i].name); + } + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\", %s", arg); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, ", REC->%s", + tp->args[i].name); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_print_fmt(struct trace_probe *tp) +{ + int len; + char *print_fmt; + + /* First: called with 0 length to calculate the needed length */ + len = __set_print_fmt(tp, NULL, 0); + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_print_fmt(tp, print_fmt, len + 1); + tp->call.print_fmt = print_fmt; + + return 0; +} + #ifdef CONFIG_EVENT_PROFILE /* Kprobe profile handler */ @@ -1456,10 +1512,14 @@ static int register_probe_event(struct trace_probe *tp) call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } + if (set_print_fmt(tp) < 0) + return -ENOMEM; call->event = &tp->event; call->id = register_ftrace_event(&tp->event); - if (!call->id) + if (!call->id) { + kfree(call->print_fmt); return -ENODEV; + } call->enabled = 0; call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; @@ -1472,6 +1532,7 @@ static int register_probe_event(struct trace_probe *tp) ret = trace_add_event_call(call); if (ret) { pr_info("Failed to register kprobe event: %s\n", call->name); + kfree(call->print_fmt); unregister_ftrace_event(&tp->event); } return ret; @@ -1481,6 +1542,7 @@ static void unregister_probe_event(struct trace_probe *tp) { /* tp->event is unregistered in trace_remove_event_call() */ trace_remove_event_call(&tp->call); + kfree(tp->call.print_fmt); } /* Make a debugfs interface for controling probe points */ -- cgit v1.2.2 From c7ef3a9004201bca90626db246a19dadd2c29c9b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 28 Dec 2009 21:13:59 -0500 Subject: tracing: Have syscall tracing call its own init function In the clean up of having all events call one specific function, the syscall event init was changed to call this helper function. With the new print_fmt updates, the syscalls need to do special initializations. This patch converts the syscall events to call its own init function again. Cc: Lai Jiangshan Cc: Li Zefan Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 1352b0a36fac..a78e86349ecb 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -450,14 +450,14 @@ int init_syscall_trace(struct ftrace_event_call *call) if (set_syscall_print_fmt(call) < 0) return -ENOMEM; - id = register_ftrace_event(call->event); - if (!id) { + id = trace_event_raw_init(call); + + if (id < 0) { free_syscall_print_fmt(call); - return -ENODEV; + return id; } - call->id = id; - INIT_LIST_HEAD(&call->fields); - return 0; + + return id; } int __init init_ftrace_syscalls(void) -- cgit v1.2.2 From 5a65e956220efc2421e21ee56d6153fd5c533a95 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:53 +0800 Subject: tracing: Use defined fields and print_fmt to print formats The calls ftrace_format_##call() and ftrace_define_fields_##call() are almost duplicate in functionality. With the addition of the print_fmt in previous patches, these two functions can be merged into one. The trace_define_field() defines the fields and links them into the struct ftrace_event_call. The previous patches introduced the print_fmt field and this can now be used with the trace_define_field() to create the event format file fields and print_fmt field. The struct ftrace_event_call->fields are used to print the fields The struct ftrace_event_call->print_fmt is used to print the "print fmt: XXXXXXXXXXX" line. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D49.5000006@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 65 ++++++++++++++++++++++++++++----------------- 1 file changed, 41 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 189b09baf4fb..250ec865d5f5 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -528,33 +528,16 @@ extern char *__bad_type_size(void); #type, "common_" #name, offsetof(typeof(field), name), \ sizeof(field.name), is_signed_type(type) -static int trace_write_header(struct trace_seq *s) -{ - struct trace_entry field; - - /* struct trace_entry */ - return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\n", - FIELD(unsigned short, type), - FIELD(unsigned char, flags), - FIELD(unsigned char, preempt_count), - FIELD(int, pid), - FIELD(int, lock_depth)); -} - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_field *field; struct trace_seq *s; + int common_field_count = 5; char *buf; - int r; + int r = 0; if (*ppos) return 0; @@ -565,14 +548,48 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - /* If any of the first writes fail, so will the show_format. */ - trace_seq_printf(s, "name: %s\n", call->name); trace_seq_printf(s, "ID: %d\n", call->id); trace_seq_printf(s, "format:\n"); - trace_write_header(s); - r = call->show_format(call, s); + list_for_each_entry_reverse(field, &call->fields, link) { + /* + * Smartly shows the array type(except dynamic array). + * Normal: + * field:TYPE VAR + * If TYPE := TYPE[LEN], it is shown: + * field:TYPE VAR[LEN] + */ + const char *array_descriptor = strchr(field->type, '['); + + if (!strncmp(field->type, "__data_loc", 10)) + array_descriptor = NULL; + + if (!array_descriptor) { + r = trace_seq_printf(s, "\tfield:%s %s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + field->type, field->name, field->offset, + field->size, !!field->is_signed); + } else { + r = trace_seq_printf(s, "\tfield:%.*s %s%s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + (int)(array_descriptor - field->type), + field->type, field->name, + array_descriptor, field->offset, + field->size, !!field->is_signed); + } + + if (--common_field_count == 0) + r = trace_seq_printf(s, "\n"); + + if (!r) + break; + } + + if (r) + r = trace_seq_printf(s, "\nprint fmt: %s\n", + call->print_fmt); + if (!r) { /* * ug! The format output is bigger than a PAGE!! -- cgit v1.2.2 From 0fa0edaf32b9a78b9854f1da98d4511a501089b0 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:57 +0800 Subject: tracing: Remove show_format and related macros from TRACE_EVENT The previous patches added the use of print_fmt string and changes the trace_define_field() function to also create the fields and format output for the event format files. text data bss dec hex filename 5857201 1355780 9336808 16549789 fc879d vmlinux 5884589 1351684 9337896 16574169 fce6d9 vmlinux-orig The above shows the size of the vmlinux after this patch set compared to the vmlinux-orig which is before the patch set. This saves us 27k on text, 1k on bss and adds just 4k of data. The total savings of 24k in size. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D4D.40604@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 12 ------- kernel/trace/trace_export.c | 73 ---------------------------------------- kernel/trace/trace_kprobe.c | 78 ------------------------------------------- kernel/trace/trace_syscalls.c | 66 ------------------------------------ 4 files changed, 229 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 250ec865d5f5..c2a3077b7353 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -520,14 +520,6 @@ out: return ret; } -extern char *__bad_type_size(void); - -#undef FIELD -#define FIELD(type, name) \ - sizeof(type) != sizeof(field.name) ? __bad_type_size() : \ - #type, "common_" #name, offsetof(typeof(field), name), \ - sizeof(field.name), is_signed_type(type) - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -965,10 +957,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, filter); } - /* A trace may not want to export its format */ - if (!call->show_format) - return 0; - trace_create_file("format", 0444, call->dir, call, format); diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 95d14b640a66..e091f64ba6ce 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -62,78 +62,6 @@ static void __always_unused ____ftrace_check_##name(void) \ #include "trace_entries.h" - -#undef __field -#define __field(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __field_desc -#define __field_desc(type, container, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array -#define __array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array_desc -#define __array_desc(type, container, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __dynamic_array -#define __dynamic_array(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:0;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef F_printk -#define F_printk(fmt, args...) "%s, %s\n", #fmt, __stringify(args) - -#undef __entry -#define __entry REC - -#undef FTRACE_ENTRY -#define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -static int \ -ftrace_format_##name(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - struct struct_name field __attribute__((unused)); \ - int ret = 0; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#include "trace_entries.h" - #undef __field #define __field(type, item) \ ret = trace_define_field(event_call, #type, #item, \ @@ -235,7 +163,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ .print_fmt = print, \ - .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 147491dccead..c99029916c76 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1174,82 +1174,6 @@ static int kretprobe_event_define_fields(struct ftrace_event_call *event_call) return 0; } -static int __probe_event_show_format(struct trace_seq *s, - struct trace_probe *tp, const char *fmt, - const char *arg) -{ - int i; - - /* Show format */ - if (!trace_seq_printf(s, "\nprint fmt: \"%s", fmt)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, " %s=%%lx", tp->args[i].name)) - return 0; - - if (!trace_seq_printf(s, "\", %s", arg)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, ", REC->%s", tp->args[i].name)) - return 0; - - return trace_seq_puts(s, "\n"); -} - -#undef SHOW_FIELD -#define SHOW_FIELD(type, item, name) \ - do { \ - ret = trace_seq_printf(s, "\tfield:" #type " %s;\t" \ - "offset:%u;\tsize:%u;\tsigned:%d;\n", name,\ - (unsigned int)offsetof(typeof(field), item),\ - (unsigned int)sizeof(type), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; \ - } while (0) - -static int kprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, ip, FIELD_STRING_IP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx)", - "REC->" FIELD_STRING_IP); -} - -static int kretprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kretprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, func, FIELD_STRING_FUNC); - SHOW_FIELD(unsigned long, ret_ip, FIELD_STRING_RETIP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx <- %lx)", - "REC->" FIELD_STRING_FUNC - ", REC->" FIELD_STRING_RETIP); -} - static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) { int i; @@ -1504,12 +1428,10 @@ static int register_probe_event(struct trace_probe *tp) if (probe_is_return(tp)) { tp->event.trace = print_kretprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kretprobe_event_show_format; call->define_fields = kretprobe_event_define_fields; } else { tp->event.trace = print_kprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } if (set_print_fmt(tp) < 0) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index a78e86349ecb..49cea70fbf6d 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -143,54 +143,6 @@ extern char *__bad_type_size(void); #type, #name, offsetof(typeof(trace), name), \ sizeof(trace.name), is_signed_type(type) -int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int i; - int ret; - struct syscall_metadata *entry = call->data; - struct syscall_trace_enter trace; - int offset = offsetof(struct syscall_trace_enter, args); - - ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr)); - if (!ret) - return 0; - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i], - entry->args[i]); - if (!ret) - return 0; - ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;" - "\tsigned:%u;\n", offset, - sizeof(unsigned long), - is_signed_type(unsigned long)); - if (!ret) - return 0; - offset += sizeof(unsigned long); - } - - trace_seq_puts(s, "\nprint fmt: \""); - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i], - sizeof(unsigned long), - i == entry->nb_args - 1 ? "" : ", "); - if (!ret) - return 0; - } - trace_seq_putc(s, '"'); - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))", - entry->args[i]); - if (!ret) - return 0; - } - - return trace_seq_putc(s, '\n'); -} - static int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) { @@ -252,24 +204,6 @@ static void free_syscall_print_fmt(struct ftrace_event_call *call) kfree(call->print_fmt); } -int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int ret; - struct syscall_trace_exit trace; - - ret = trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr), - SYSCALL_FIELD(long, ret)); - if (!ret) - return 0; - - return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n"); -} - int syscall_enter_define_fields(struct ftrace_event_call *call) { struct syscall_trace_enter trace; -- cgit v1.2.2 From 7e53bd42d14c75192b99674c40fcc359392da59d Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 6 Jan 2010 20:08:50 +0800 Subject: tracing: Consolidate protection of reader access to the ring buffer At the beginning, access to the ring buffer was fully serialized by trace_types_lock. Patch d7350c3f4569 gives more freedom to readers, and patch b04cc6b1f6 adds code to protect trace_pipe and cpu#/trace_pipe. But actually it is not enough, ring buffer readers are not always read-only, they may consume data. This patch makes accesses to trace, trace_pipe, trace_pipe_raw cpu#/trace, cpu#/trace_pipe and cpu#/trace_pipe_raw serialized. And removes tracing_reader_cpumask which is used to protect trace_pipe. Details: Ring buffer serializes readers, but it is low level protection. The validity of the events (which returns by ring_buffer_peek() ..etc) are not protected by ring buffer. The content of events may become garbage if we allow another process to consume these events concurrently: A) the page of the consumed events may become a normal page (not reader page) in ring buffer, and this page will be rewritten by the events producer. B) The page of the consumed events may become a page for splice_read, and this page will be returned to system. This patch adds trace_access_lock() and trace_access_unlock() primitives. These primitives allow multi process access to different cpu ring buffers concurrently. These primitives don't distinguish read-only and read-consume access. Multi read-only access is also serialized. And we don't use these primitives when we open files, we only use them when we read files. Signed-off-by: Lai Jiangshan LKML-Reference: <4B447D52.1050602@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 136 ++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 97 insertions(+), 39 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0df1b0f2cb9e..abdd333a0825 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include #include @@ -102,9 +103,6 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; -/* Define which cpu buffers are currently read in trace_pipe */ -static cpumask_var_t tracing_reader_cpumask; - #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -243,12 +241,91 @@ static struct tracer *current_trace __read_mostly; /* * trace_types_lock is used to protect the trace_types list. - * This lock is also used to keep user access serialized. - * Accesses from userspace will grab this lock while userspace - * activities happen inside the kernel. */ static DEFINE_MUTEX(trace_types_lock); +/* + * serialize the access of the ring buffer + * + * ring buffer serializes readers, but it is low level protection. + * The validity of the events (which returns by ring_buffer_peek() ..etc) + * are not protected by ring buffer. + * + * The content of events may become garbage if we allow other process consumes + * these events concurrently: + * A) the page of the consumed events may become a normal page + * (not reader page) in ring buffer, and this page will be rewrited + * by events producer. + * B) The page of the consumed events may become a page for splice_read, + * and this page will be returned to system. + * + * These primitives allow multi process access to different cpu ring buffer + * concurrently. + * + * These primitives don't distinguish read-only and read-consume access. + * Multi read-only access are also serialized. + */ + +#ifdef CONFIG_SMP +static DECLARE_RWSEM(all_cpu_access_lock); +static DEFINE_PER_CPU(struct mutex, cpu_access_lock); + +static inline void trace_access_lock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + /* gain it for accessing the whole ring buffer. */ + down_write(&all_cpu_access_lock); + } else { + /* gain it for accessing a cpu ring buffer. */ + + /* Firstly block other trace_access_lock(TRACE_PIPE_ALL_CPU). */ + down_read(&all_cpu_access_lock); + + /* Secondly block other access to this @cpu ring buffer. */ + mutex_lock(&per_cpu(cpu_access_lock, cpu)); + } +} + +static inline void trace_access_unlock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + up_write(&all_cpu_access_lock); + } else { + mutex_unlock(&per_cpu(cpu_access_lock, cpu)); + up_read(&all_cpu_access_lock); + } +} + +static inline void trace_access_lock_init(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + mutex_init(&per_cpu(cpu_access_lock, cpu)); +} + +#else + +static DEFINE_MUTEX(access_lock); + +static inline void trace_access_lock(int cpu) +{ + (void)cpu; + mutex_lock(&access_lock); +} + +static inline void trace_access_unlock(int cpu) +{ + (void)cpu; + mutex_unlock(&access_lock); +} + +static inline void trace_access_lock_init(void) +{ +} + +#endif + /* trace_wait is a waitqueue for tasks blocked on trace_poll */ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); @@ -1580,12 +1657,6 @@ static void tracing_iter_reset(struct trace_iterator *iter, int cpu) } /* - * No necessary locking here. The worst thing which can - * happen is loosing events consumed at the same time - * by a trace_pipe reader. - * Other than that, we don't risk to crash the ring buffer - * because it serializes the readers. - * * The current tracer is copied to avoid a global locking * all around. */ @@ -1640,12 +1711,16 @@ static void *s_start(struct seq_file *m, loff_t *pos) } trace_event_read_lock(); + trace_access_lock(cpu_file); return p; } static void s_stop(struct seq_file *m, void *p) { + struct trace_iterator *iter = m->private; + atomic_dec(&trace_record_cmdline_disabled); + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); } @@ -2836,22 +2911,6 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); - /* We only allow one reader per cpu */ - if (cpu_file == TRACE_PIPE_ALL_CPU) { - if (!cpumask_empty(tracing_reader_cpumask)) { - ret = -EBUSY; - goto out; - } - cpumask_setall(tracing_reader_cpumask); - } else { - if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) - cpumask_set_cpu(cpu_file, tracing_reader_cpumask); - else { - ret = -EBUSY; - goto out; - } - } - /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { @@ -2907,12 +2966,6 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) mutex_lock(&trace_types_lock); - if (iter->cpu_file == TRACE_PIPE_ALL_CPU) - cpumask_clear(tracing_reader_cpumask); - else - cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); - - if (iter->trace->pipe_close) iter->trace->pipe_close(iter); @@ -3074,6 +3127,7 @@ waitagain: iter->pos = -1; trace_event_read_lock(); + trace_access_lock(iter->cpu_file); while (find_next_entry_inc(iter) != NULL) { enum print_line_t ret; int len = iter->seq.len; @@ -3090,6 +3144,7 @@ waitagain: if (iter->seq.len >= cnt) break; } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); /* Now copy what we have to the user */ @@ -3215,6 +3270,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, } trace_event_read_lock(); + trace_access_lock(iter->cpu_file); /* Fill as many pages as possible. */ for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { @@ -3238,6 +3294,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, trace_seq_init(&iter->seq); } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); mutex_unlock(&iter->mutex); @@ -3539,10 +3596,12 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, info->read = 0; + trace_access_lock(info->cpu); ret = ring_buffer_read_page(info->tr->buffer, &info->spare, count, info->cpu, 0); + trace_access_unlock(info->cpu); if (ret < 0) return 0; @@ -3670,6 +3729,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, len &= PAGE_MASK; } + trace_access_lock(info->cpu); entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) { @@ -3717,6 +3777,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); } + trace_access_unlock(info->cpu); spd.nr_pages = i; /* did we read anything? */ @@ -4153,6 +4214,8 @@ static __init int tracer_init_debugfs(void) struct dentry *d_tracer; int cpu; + trace_access_lock_init(); + d_tracer = tracing_init_dentry(); trace_create_file("tracing_enabled", 0644, d_tracer, @@ -4387,9 +4450,6 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; - if (!zalloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) - goto out_free_tracing_cpumask; - /* To save memory, keep the ring buffer size to its minimum */ if (ring_buffer_expanded) ring_buf_size = trace_buf_size; @@ -4447,8 +4507,6 @@ __init static int tracer_alloc_buffers(void) return 0; out_free_cpumask: - free_cpumask_var(tracing_reader_cpumask); -out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); -- cgit v1.2.2 From d931369b74b3d6f2044f595af6f3dd074f65d9cf Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 17:27:11 -0500 Subject: tracing: Add stack dump to trace_printk if stacktrace option is set If the ftrace stacktrace option is set, then add the stack dumps to trace_printk. Requested-by: Peter Zijlstra Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index abdd333a0825..5314c90bbc83 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1392,8 +1392,10 @@ int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) entry->fmt = fmt; memcpy(entry->buf, trace_buf, sizeof(u32) * len); - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); @@ -1466,8 +1468,10 @@ int trace_array_vprintk(struct trace_array *tr, memcpy(&entry->buf, trace_buf, len); entry->buf[len] = '\0'; - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, irq_flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); -- cgit v1.2.2 From f96e9232e04856c781d4f71923a46dd3f7b429fa Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:00 -0800 Subject: rcu: Adjust force_quiescent_state() locking, step 1 This causes rnp->lock to be held on entry to force_quiescent_state()'s switch statement. This is a first step towards prohibiting starting grace periods while force_quiescent_state() is executing, which will reduce the number and complexity of races that force_quiescent_state() is involved in. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501455-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 27 ++++++++++++++++++--------- 1 file changed, 18 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 53ae9598f798..eae331da6bee 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1204,7 +1204,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) } if (relaxed && (long)(rsp->jiffies_force_qs - jiffies) >= 0) - goto unlock_ret; /* no emergency and done recently. */ + goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); lastcomp = rsp->gpnum - 1; @@ -1213,31 +1213,32 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; spin_unlock(&rnp->lock); - goto unlock_ret; /* no GP in progress, time updated. */ + goto unlock_fqs_ret; /* no GP in progress, time updated. */ } - spin_unlock(&rnp->lock); switch (signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: + spin_unlock(&rnp->lock); break; /* grace period idle or initializing, ignore. */ case RCU_SAVE_DYNTICK: + spin_unlock(&rnp->lock); if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ if (rcu_process_dyntick(rsp, lastcomp, dyntick_save_progress_counter)) - goto unlock_ret; + goto unlock_fqs_ret; + spin_lock(&rnp->lock); /* fall into next case. */ case RCU_SAVE_COMPLETED: /* Update state, record completion counter. */ forcenow = 0; - spin_lock(&rnp->lock); if (lastcomp + 1 == rsp->gpnum && lastcomp == rsp->completed && rsp->signaled == signaled) { @@ -1245,23 +1246,31 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->completed_fqs = lastcomp; forcenow = signaled == RCU_SAVE_COMPLETED; } - spin_unlock(&rnp->lock); - if (!forcenow) + if (!forcenow) { + spin_unlock(&rnp->lock); break; + } /* fall into next case. */ case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ + spin_unlock(&rnp->lock); if (rcu_process_dyntick(rsp, rsp->completed_fqs, rcu_implicit_dynticks_qs)) - goto unlock_ret; + goto unlock_fqs_ret; /* Leave state in case more forcing is required. */ break; + + default: + + spin_unlock(&rnp->lock); + WARN_ON_ONCE(1); + break; } -unlock_ret: +unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } -- cgit v1.2.2 From 559569acf94f538b56bd6eead80b439d6a78cdff Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:01 -0800 Subject: rcu: Adjust force_quiescent_state() locking, step 2 This patch releases rnp->lock after the end of force_quiescent_state()'s switch statement. This is a second step towards prohibiting starting grace periods while force_quiescent_state() is executing, which will reduce the number and complexity of races that force_quiescent_state() is involved in. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501994-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index eae331da6bee..d42ad30c4d70 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1219,7 +1219,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_GP_IDLE: case RCU_GP_INIT: - spin_unlock(&rnp->lock); break; /* grace period idle or initializing, ignore. */ case RCU_SAVE_DYNTICK: @@ -1246,10 +1245,8 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->completed_fqs = lastcomp; forcenow = signaled == RCU_SAVE_COMPLETED; } - if (!forcenow) { - spin_unlock(&rnp->lock); + if (!forcenow) break; - } /* fall into next case. */ case RCU_FORCE_QS: @@ -1262,14 +1259,10 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Leave state in case more forcing is required. */ - break; - - default: - - spin_unlock(&rnp->lock); - WARN_ON_ONCE(1); + spin_lock(&rnp->lock); break; } + spin_unlock(&rnp->lock); unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } -- cgit v1.2.2 From 07079d5357a4d53c2b13126c4a38fb40e6e04966 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:02 -0800 Subject: rcu: Prohibit starting new grace periods while forcing quiescent states Reduce the number and variety of race conditions by prohibiting the start of a new grace period while force_quiescent_state() is active. A new fqs_active flag in the rcu_state structure is used to trace whether or not force_quiescent_state() is active, and this new flag is tested by rcu_start_gp(). If the CPU that closed out the last grace period needs another grace period, this new grace period may be delayed up to one scheduling-clock tick, but it will eventually get started. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <126264655052-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 31 +++++++++++++++++-------------- kernel/rcutree.h | 2 ++ 2 files changed, 19 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index d42ad30c4d70..41688ff60e07 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -659,7 +659,7 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) struct rcu_data *rdp = rsp->rda[smp_processor_id()]; struct rcu_node *rnp = rcu_get_root(rsp); - if (!cpu_needs_another_gp(rsp, rdp)) { + if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { if (rnp->completed == rsp->completed) { spin_unlock_irqrestore(&rnp->lock, flags); return; @@ -1195,6 +1195,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) struct rcu_node *rnp = rcu_get_root(rsp); u8 signaled; u8 forcenow; + u8 gpdone; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1206,15 +1207,16 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) (long)(rsp->jiffies_force_qs - jiffies) >= 0) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; - spin_lock(&rnp->lock); + spin_lock(&rnp->lock); /* irqs already disabled */ lastcomp = rsp->gpnum - 1; signaled = rsp->signaled; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; - spin_unlock(&rnp->lock); + spin_unlock(&rnp->lock); /* irqs remain disabled */ goto unlock_fqs_ret; /* no GP in progress, time updated. */ } + rsp->fqs_active = 1; switch (signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: @@ -1223,15 +1225,16 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_SAVE_DYNTICK: - spin_unlock(&rnp->lock); + spin_unlock(&rnp->lock); /* irqs remain disabled */ if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - if (rcu_process_dyntick(rsp, lastcomp, - dyntick_save_progress_counter)) - goto unlock_fqs_ret; - spin_lock(&rnp->lock); + gpdone = rcu_process_dyntick(rsp, lastcomp, + dyntick_save_progress_counter); + spin_lock(&rnp->lock); /* irqs already disabled */ + if (gpdone) + break; /* fall into next case. */ case RCU_SAVE_COMPLETED: @@ -1252,17 +1255,17 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ - spin_unlock(&rnp->lock); - if (rcu_process_dyntick(rsp, rsp->completed_fqs, - rcu_implicit_dynticks_qs)) - goto unlock_fqs_ret; + spin_unlock(&rnp->lock); /* irqs remain disabled */ + gpdone = rcu_process_dyntick(rsp, rsp->completed_fqs, + rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ - spin_lock(&rnp->lock); + spin_lock(&rnp->lock); /* irqs already disabled */ break; } - spin_unlock(&rnp->lock); + rsp->fqs_active = 0; + spin_unlock(&rnp->lock); /* irqs remain disabled */ unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index d2a0046f63b2..dc386a7c634f 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -277,6 +277,8 @@ struct rcu_state { u8 signaled ____cacheline_internodealigned_in_smp; /* Force QS state. */ + u8 fqs_active; /* force_quiescent_state() */ + /* is running. */ long gpnum; /* Current gp number. */ long completed; /* # of last completed gp. */ -- cgit v1.2.2 From f3a8b5c6aa543bd87764418d63632eb65b80e2f6 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:03 -0800 Subject: rcu: Eliminate local variable signaled from force_quiescent_state() Because the root rcu_node lock is held across entry to the switch statement in force_quiescent_state(), it is no longer necessary to snapshot rsp->signaled to a local variable. Eliminate both the snapshotting and the local variable. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646550602-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 41688ff60e07..1d8cfb1711fd 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1193,7 +1193,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) unsigned long flags; long lastcomp; struct rcu_node *rnp = rcu_get_root(rsp); - u8 signaled; u8 forcenow; u8 gpdone; @@ -1209,7 +1208,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->n_force_qs++; spin_lock(&rnp->lock); /* irqs already disabled */ lastcomp = rsp->gpnum - 1; - signaled = rsp->signaled; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; @@ -1217,7 +1215,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) goto unlock_fqs_ret; /* no GP in progress, time updated. */ } rsp->fqs_active = 1; - switch (signaled) { + switch (rsp->signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: @@ -1242,11 +1240,10 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Update state, record completion counter. */ forcenow = 0; if (lastcomp + 1 == rsp->gpnum && - lastcomp == rsp->completed && - rsp->signaled == signaled) { + lastcomp == rsp->completed) { + forcenow = rsp->signaled == RCU_SAVE_COMPLETED; rsp->signaled = RCU_FORCE_QS; rsp->completed_fqs = lastcomp; - forcenow = signaled == RCU_SAVE_COMPLETED; } if (!forcenow) break; -- cgit v1.2.2 From 39c0bbfc07c6e28db7346d0e11106f2d045d3035 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:04 -0800 Subject: rcu: Eliminate local variable lastcomp from force_quiescent_state() Because rsp->fqs_active is set to 1 across force_quiescent_state()'s switch statement, rcu_start_gp() will refrain from starting a new grace period during this time. Therefore, rsp->gpnum is constant, and can be propagated to all uses of lastcomp, eliminating this local variable. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465502985-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 10 +++------- kernel/rcutree.h | 2 -- 2 files changed, 3 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 1d8cfb1711fd..62b64332effb 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1191,7 +1191,6 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; - long lastcomp; struct rcu_node *rnp = rcu_get_root(rsp); u8 forcenow; u8 gpdone; @@ -1207,7 +1206,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); /* irqs already disabled */ - lastcomp = rsp->gpnum - 1; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; @@ -1228,7 +1226,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, lastcomp, + gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (gpdone) @@ -1239,11 +1237,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Update state, record completion counter. */ forcenow = 0; - if (lastcomp + 1 == rsp->gpnum && - lastcomp == rsp->completed) { + if (rsp->gpnum - 1 == rsp->completed) { forcenow = rsp->signaled == RCU_SAVE_COMPLETED; rsp->signaled = RCU_FORCE_QS; - rsp->completed_fqs = lastcomp; } if (!forcenow) break; @@ -1253,7 +1249,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rsp->completed_fqs, + gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index dc386a7c634f..534856121b06 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -296,8 +296,6 @@ struct rcu_state { long orphan_qlen; /* Number of orphaned cbs. */ spinlock_t fqslock; /* Only one task forcing */ /* quiescent states. */ - long completed_fqs; /* Value of completed @ snap. */ - /* Protected by fqslock. */ unsigned long jiffies_force_qs; /* Time at which to invoke */ /* force_quiescent_state(). */ unsigned long n_force_qs; /* Number of calls to */ -- cgit v1.2.2 From eb1ba45f1e7f6e626fefc063b340c7cbec9bd8c7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:05 -0800 Subject: rcu: Eliminate second argument of rcu_process_dyntick() At this point, the second argument to all calls to rcu_process_dyntick() is a function of the same field of the structure passed in as the first argument, namely, rsp->gpnum-1. So propagate rsp->gpnum-1 to all uses of the second argument within rcu_process_dyntick() and then eliminate the second argument. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465503786-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 62b64332effb..c7d00700fc4e 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1147,7 +1147,7 @@ void rcu_check_callbacks(int cpu, int user) * Returns 1 if the current grace period ends while scanning (possibly * because we made it end). */ -static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, +static int rcu_process_dyntick(struct rcu_state *rsp, int (*f)(struct rcu_data *)) { unsigned long bit; @@ -1159,7 +1159,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, rcu_for_each_leaf_node(rsp, rnp) { mask = 0; spin_lock_irqsave(&rnp->lock, flags); - if (rnp->completed != lastcomp) { + if (rnp->completed != rsp->gpnum - 1) { spin_unlock_irqrestore(&rnp->lock, flags); return 1; } @@ -1173,7 +1173,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rnp->completed == lastcomp) { + if (mask != 0 && rnp->completed == rsp->gpnum - 1) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1226,7 +1226,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, + gpdone = rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (gpdone) @@ -1249,8 +1249,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, - rcu_implicit_dynticks_qs); + gpdone = rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v1.2.2 From 0f10dc826646134dce3e5751512b87d30f3903e4 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:06 -0800 Subject: rcu: Eliminate rcu_process_dyntick() return value Because a new grace period cannot start while we are executing within the force_quiescent_state() function's switch statement, if any test within that switch statement or within any function called from that switch statement shows that the current grace period has ended, we can safely re-do that test any time before we leave the switch statement. This means that we no longer need a return value from rcu_process_dyntick(), as we can simply invoke rcu_gp_in_progress() to check whether the old grace period has finished -- there is no longer any need to worry about whether or not a new grace period has been started. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501857-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index c7d00700fc4e..e4971192fa9c 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1144,11 +1144,9 @@ void rcu_check_callbacks(int cpu, int user) /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. - * Returns 1 if the current grace period ends while scanning (possibly - * because we made it end). */ -static int rcu_process_dyntick(struct rcu_state *rsp, - int (*f)(struct rcu_data *)) +static void rcu_process_dyntick(struct rcu_state *rsp, + int (*f)(struct rcu_data *)) { unsigned long bit; int cpu; @@ -1161,7 +1159,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, spin_lock_irqsave(&rnp->lock, flags); if (rnp->completed != rsp->gpnum - 1) { spin_unlock_irqrestore(&rnp->lock, flags); - return 1; + return; } if (rnp->qsmask == 0) { spin_unlock_irqrestore(&rnp->lock, flags); @@ -1181,7 +1179,6 @@ static int rcu_process_dyntick(struct rcu_state *rsp, } spin_unlock_irqrestore(&rnp->lock, flags); } - return 0; } /* @@ -1193,7 +1190,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); u8 forcenow; - u8 gpdone; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1226,10 +1222,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, - dyntick_save_progress_counter); + rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ - if (gpdone) + if (!rcu_gp_in_progress(rsp)) break; /* fall into next case. */ @@ -1249,7 +1244,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); + rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v1.2.2 From ee47eb9f4da6f44af965d6d049e77ee8c8a4b822 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:07 -0800 Subject: rcu: Remove leg of force_quiescent_state() switch statement The comparisons of rsp->gpnum nad rsp->completed in rcu_process_dyntick() and force_quiescent_state() can be replaced by the much more clear rcu_gp_in_progress() predicate function. After doing this, it becomes clear that the RCU_SAVE_COMPLETED leg of the force_quiescent_state() function's switch statement is almost completely a no-op. A small change to the RCU_SAVE_DYNTICK leg renders it a complete no-op, after which it can be removed. Doing so also eliminates the forcenow local variable from force_quiescent_state(). Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501781-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 22 +++++----------------- kernel/rcutree.h | 5 ++--- 2 files changed, 7 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index e4971192fa9c..6268f37adfc4 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1144,6 +1144,7 @@ void rcu_check_callbacks(int cpu, int user) /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. + * The caller must have suppressed start of new grace periods. */ static void rcu_process_dyntick(struct rcu_state *rsp, int (*f)(struct rcu_data *)) @@ -1157,7 +1158,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, rcu_for_each_leaf_node(rsp, rnp) { mask = 0; spin_lock_irqsave(&rnp->lock, flags); - if (rnp->completed != rsp->gpnum - 1) { + if (!rcu_gp_in_progress(rsp)) { spin_unlock_irqrestore(&rnp->lock, flags); return; } @@ -1171,7 +1172,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rnp->completed == rsp->gpnum - 1) { + if (mask != 0 && rcu_gp_in_progress(rsp)) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1189,7 +1190,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); - u8 forcenow; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1224,21 +1224,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Record dyntick-idle state. */ rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ - if (!rcu_gp_in_progress(rsp)) - break; - /* fall into next case. */ - - case RCU_SAVE_COMPLETED: - - /* Update state, record completion counter. */ - forcenow = 0; - if (rsp->gpnum - 1 == rsp->completed) { - forcenow = rsp->signaled == RCU_SAVE_COMPLETED; + if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; - } - if (!forcenow) - break; - /* fall into next case. */ + break; case RCU_FORCE_QS: diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 534856121b06..edb6fae0fa94 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -237,12 +237,11 @@ struct rcu_data { #define RCU_GP_IDLE 0 /* No grace period in progress. */ #define RCU_GP_INIT 1 /* Grace period being initialized. */ #define RCU_SAVE_DYNTICK 2 /* Need to scan dyntick state. */ -#define RCU_SAVE_COMPLETED 3 /* Need to save rsp->completed. */ -#define RCU_FORCE_QS 4 /* Need to force quiescent state. */ +#define RCU_FORCE_QS 3 /* Need to force quiescent state. */ #ifdef CONFIG_NO_HZ #define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK #else /* #ifdef CONFIG_NO_HZ */ -#define RCU_SIGNAL_INIT RCU_SAVE_COMPLETED +#define RCU_SIGNAL_INIT RCU_FORCE_QS #endif /* #else #ifdef CONFIG_NO_HZ */ #define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ -- cgit v1.2.2 From 45f014c52eef022873b19d6a20eb0ec9668f2b09 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:08 -0800 Subject: rcu: Remove redundant grace-period check The rcu_process_dyntick() function checks twice for the end of the current grace period. However, it holds the current rcu_node structure's ->lock field throughout, and doesn't get to the second call to rcu_gp_in_progress() unless there is at least one CPU corresponding to this rcu_node structure that has not yet checked in for the current grace period, which would prevent the current grace period from ending. So the current grace period cannot have ended, and the second check is redundant, so remove it. Also, given that this function is used even with !CONFIG_NO_HZ, its name is quite misleading. Change from rcu_process_dyntick() to force_qs_rnp(). Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646550562-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 6268f37adfc4..d9202857d3ad 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1146,8 +1146,7 @@ void rcu_check_callbacks(int cpu, int user) * have not yet encountered a quiescent state, using the function specified. * The caller must have suppressed start of new grace periods. */ -static void rcu_process_dyntick(struct rcu_state *rsp, - int (*f)(struct rcu_data *)) +static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) { unsigned long bit; int cpu; @@ -1172,7 +1171,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rcu_gp_in_progress(rsp)) { + if (mask != 0) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1222,7 +1221,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - rcu_process_dyntick(rsp, dyntick_save_progress_counter); + force_qs_rnp(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; @@ -1232,7 +1231,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); + force_qs_rnp(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v1.2.2 From 46a1e34eda805501a8b32f26394faa435149f6d1 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:09 -0800 Subject: rcu: Make force_quiescent_state() start grace period if needed Grace periods cannot be started while force_quiescent_state() is active. This is OK in that the affected CPUs will try again later, but it does induce needless grace-period delays. This patch causes rcu_start_gp() to record a failed attempt to start a grace period. When force_quiescent_state() prepares to return, it then starts the grace period if there was such a failed attempt. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501854-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 8 ++++++++ kernel/rcutree.h | 5 +++++ 2 files changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index d9202857d3ad..55e8f6ef8195 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -660,6 +660,8 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) struct rcu_node *rnp = rcu_get_root(rsp); if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { + if (cpu_needs_another_gp(rsp, rdp)) + rsp->fqs_need_gp = 1; if (rnp->completed == rsp->completed) { spin_unlock_irqrestore(&rnp->lock, flags); return; @@ -1239,6 +1241,12 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; } rsp->fqs_active = 0; + if (rsp->fqs_need_gp) { + spin_unlock(&rsp->fqslock); /* irqs remain disabled */ + rsp->fqs_need_gp = 0; + rcu_start_gp(rsp, flags); /* releases rnp->lock */ + return; + } spin_unlock(&rnp->lock); /* irqs remain disabled */ unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); diff --git a/kernel/rcutree.h b/kernel/rcutree.h index edb6fae0fa94..bd5d78ad1c48 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -278,6 +278,11 @@ struct rcu_state { /* Force QS state. */ u8 fqs_active; /* force_quiescent_state() */ /* is running. */ + u8 fqs_need_gp; /* A CPU was prevented from */ + /* starting a new grace */ + /* period because */ + /* force_quiescent_state() */ + /* was running. */ long gpnum; /* Current gp number. */ long completed; /* # of last completed gp. */ -- cgit v1.2.2 From bf66f18e79e34c421bbd8f6511e2c556b779df2f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:10 -0800 Subject: rcu: Add force_quiescent_state() testing to rcutorture Add force_quiescent_state() testing to rcutorture, with a separate thread that repeatedly invokes force_quiescent_state() in bursts. This can greatly increase the probability of encountering certain types of race conditions. Suggested-by: Josh Triplett Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646551116-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutorture.c | 80 +++++++++++++++++++++++++++++++++++++++++++++++-- kernel/rcutree.c | 18 +++++++++++ kernel/rcutree_plugin.h | 19 ++++++++++++ 3 files changed, 115 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 9bb52177af02..adda92bfafac 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -61,6 +61,9 @@ static int test_no_idle_hz; /* Test RCU's support for tickless idle CPUs. */ static int shuffle_interval = 3; /* Interval between shuffles (in sec)*/ static int stutter = 5; /* Start/stop testing interval (in sec) */ static int irqreader = 1; /* RCU readers from irq (timers). */ +static int fqs_duration = 0; /* Duration of bursts (us), 0 to disable. */ +static int fqs_holdoff = 0; /* Hold time within burst (us). */ +static int fqs_stutter = 3; /* Wait time between bursts (s). */ static char *torture_type = "rcu"; /* What RCU implementation to torture. */ module_param(nreaders, int, 0444); @@ -79,6 +82,12 @@ module_param(stutter, int, 0444); MODULE_PARM_DESC(stutter, "Number of seconds to run/halt test"); module_param(irqreader, int, 0444); MODULE_PARM_DESC(irqreader, "Allow RCU readers from irq handlers"); +module_param(fqs_duration, int, 0444); +MODULE_PARM_DESC(fqs_duration, "Duration of fqs bursts (us)"); +module_param(fqs_holdoff, int, 0444); +MODULE_PARM_DESC(fqs_holdoff, "Holdoff time within fqs bursts (us)"); +module_param(fqs_stutter, int, 0444); +MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); module_param(torture_type, charp, 0444); MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, rcu_bh, srcu)"); @@ -99,6 +108,7 @@ static struct task_struct **reader_tasks; static struct task_struct *stats_task; static struct task_struct *shuffler_task; static struct task_struct *stutter_task; +static struct task_struct *fqs_task; #define RCU_TORTURE_PIPE_LEN 10 @@ -263,6 +273,7 @@ struct rcu_torture_ops { void (*deferred_free)(struct rcu_torture *p); void (*sync)(void); void (*cb_barrier)(void); + void (*fqs)(void); int (*stats)(char *page); int irq_capable; char *name; @@ -347,6 +358,7 @@ static struct rcu_torture_ops rcu_ops = { .deferred_free = rcu_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = rcu_barrier, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu" @@ -388,6 +400,7 @@ static struct rcu_torture_ops rcu_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_sync" @@ -403,6 +416,7 @@ static struct rcu_torture_ops rcu_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu_expedited, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_expedited" @@ -465,6 +479,7 @@ static struct rcu_torture_ops rcu_bh_ops = { .deferred_free = rcu_bh_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = rcu_barrier_bh, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh" @@ -480,6 +495,7 @@ static struct rcu_torture_ops rcu_bh_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh_sync" @@ -621,6 +637,7 @@ static struct rcu_torture_ops sched_ops = { .deferred_free = rcu_sched_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = rcu_barrier_sched, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "sched" @@ -636,6 +653,7 @@ static struct rcu_torture_ops sched_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .name = "sched_sync" }; @@ -650,11 +668,44 @@ static struct rcu_torture_ops sched_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_sched_expedited, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = rcu_expedited_torture_stats, .irq_capable = 1, .name = "sched_expedited" }; +/* + * RCU torture force-quiescent-state kthread. Repeatedly induces + * bursts of calls to force_quiescent_state(), increasing the probability + * of occurrence of some important types of race conditions. + */ +static int +rcu_torture_fqs(void *arg) +{ + unsigned long fqs_resume_time; + int fqs_burst_remaining; + + VERBOSE_PRINTK_STRING("rcu_torture_fqs task started"); + do { + fqs_resume_time = jiffies + fqs_stutter * HZ; + while (jiffies - fqs_resume_time > LONG_MAX) { + schedule_timeout_interruptible(1); + } + fqs_burst_remaining = fqs_duration; + while (fqs_burst_remaining > 0) { + cur_ops->fqs(); + udelay(fqs_holdoff); + fqs_burst_remaining -= fqs_holdoff; + } + rcu_stutter_wait("rcu_torture_fqs"); + } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); + VERBOSE_PRINTK_STRING("rcu_torture_fqs task stopping"); + rcutorture_shutdown_absorb("rcu_torture_fqs"); + while (!kthread_should_stop()) + schedule_timeout_uninterruptible(1); + return 0; +} + /* * RCU torture writer kthread. Repeatedly substitutes a new structure * for that pointed to by rcu_torture_current, freeing the old structure @@ -1030,10 +1081,11 @@ rcu_torture_print_module_parms(char *tag) printk(KERN_ALERT "%s" TORTURE_FLAG "--- %s: nreaders=%d nfakewriters=%d " "stat_interval=%d verbose=%d test_no_idle_hz=%d " - "shuffle_interval=%d stutter=%d irqreader=%d\n", + "shuffle_interval=%d stutter=%d irqreader=%d " + "fqs_duration=%d fqs_holdoff=%d fqs_stutter=%d\n", torture_type, tag, nrealreaders, nfakewriters, stat_interval, verbose, test_no_idle_hz, shuffle_interval, - stutter, irqreader); + stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter); } static struct notifier_block rcutorture_nb = { @@ -1109,6 +1161,12 @@ rcu_torture_cleanup(void) } stats_task = NULL; + if (fqs_task) { + VERBOSE_PRINTK_STRING("Stopping rcu_torture_fqs task"); + kthread_stop(fqs_task); + } + fqs_task = NULL; + /* Wait for all RCU callbacks to fire. */ if (cur_ops->cb_barrier != NULL) @@ -1154,6 +1212,11 @@ rcu_torture_init(void) mutex_unlock(&fullstop_mutex); return -EINVAL; } + if (cur_ops->fqs == NULL && fqs_duration != 0) { + printk(KERN_ALERT "rcu-torture: ->fqs NULL and non-zero " + "fqs_duration, fqs disabled.\n"); + fqs_duration = 0; + } if (cur_ops->init) cur_ops->init(); /* no "goto unwind" prior to this point!!! */ @@ -1282,6 +1345,19 @@ rcu_torture_init(void) goto unwind; } } + if (fqs_duration < 0) + fqs_duration = 0; + if (fqs_duration) { + /* Create the stutter thread */ + fqs_task = kthread_run(rcu_torture_fqs, NULL, + "rcu_torture_fqs"); + if (IS_ERR(fqs_task)) { + firsterr = PTR_ERR(fqs_task); + VERBOSE_PRINTK_ERRSTRING("Failed to create fqs"); + fqs_task = NULL; + goto unwind; + } + } register_reboot_notifier(&rcutorture_nb); mutex_unlock(&fullstop_mutex); return 0; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 55e8f6ef8195..0a4c32879398 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -156,6 +156,24 @@ long rcu_batches_completed_bh(void) } EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); +/* + * Force a quiescent state for RCU BH. + */ +void rcu_bh_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_bh_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_bh_force_quiescent_state); + +/* + * Force a quiescent state for RCU-sched. + */ +void rcu_sched_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_sched_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_sched_force_quiescent_state); + /* * Does the CPU have callbacks ready to be invoked? */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 37fbccdf41d5..f11ebd44b454 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -61,6 +61,15 @@ long rcu_batches_completed(void) } EXPORT_SYMBOL_GPL(rcu_batches_completed); +/* + * Force a quiescent state for preemptible RCU. + */ +void rcu_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_preempt_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + /* * Record a preemptable-RCU quiescent state for the specified CPU. Note * that this just means that the task currently running on the CPU is @@ -712,6 +721,16 @@ long rcu_batches_completed(void) } EXPORT_SYMBOL_GPL(rcu_batches_completed); +/* + * Force a quiescent state for RCU, which, because there is no preemptible + * RCU, becomes the same as rcu-sched. + */ +void rcu_force_quiescent_state(void) +{ + rcu_sched_force_quiescent_state(); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + /* * Because preemptable RCU does not exist, we never have to check for * CPUs being in quiescent states. -- cgit v1.2.2 From cba8244a0f1c277b6b1e48ed6504fa434119e24d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 16:04:01 -0800 Subject: rcu: Add debug check for too many rcu_read_unlock() TREE_PREEMPT_RCU maintains an rcu_read_lock_nesting counter in the task structure, which happens to be a signed int. So this patch adds a check for this counter being negative at the end of __rcu_read_unlock(). This check is under CONFIG_PROVE_LOCKING, so can be thought of as being part of lockdep. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626498423064-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree_plugin.h | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index f11ebd44b454..e77cdf321e13 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -304,6 +304,9 @@ void __rcu_read_unlock(void) if (--ACCESS_ONCE(t->rcu_read_lock_nesting) == 0 && unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) rcu_read_unlock_special(t); +#ifdef CONFIG_PROVE_LOCKING + WARN_ON_ONCE(ACCESS_ONCE(t->rcu_read_lock_nesting) < 0); +#endif /* #ifdef CONFIG_PROVE_LOCKING */ } EXPORT_SYMBOL_GPL(__rcu_read_unlock); -- cgit v1.2.2 From b6407e863934965cdc66cbc244d811ceeb6f4d77 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 16:04:02 -0800 Subject: rcu: Give different levels of the rcu_node hierarchy distinct lockdep names Previously, each level of the rcu_node hierarchy had the same rather unimaginative name: "&rcu_node_class[i]". This makes lockdep diagnostics involving these lockdep classes less helpful than would be nice. This patch fixes this by giving each level of the rcu_node hierarchy a distinct name: "rcu_node_level_0", "rcu_node_level_1", and so on. This version of the patch includes improved diagnostics suggested by Josh Triplett and Peter Zijlstra. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626498421830-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 0a4c32879398..3b13d64b010b 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1811,11 +1811,17 @@ static void __init rcu_init_levelspread(struct rcu_state *rsp) */ static void __init rcu_init_one(struct rcu_state *rsp) { + static char *buf[] = { "rcu_node_level_0", + "rcu_node_level_1", + "rcu_node_level_2", + "rcu_node_level_3" }; /* Match MAX_RCU_LVLS */ int cpustride = 1; int i; int j; struct rcu_node *rnp; + BUILD_BUG_ON(MAX_RCU_LVLS > ARRAY_SIZE(buf)); /* Fix buf[] init! */ + /* Initialize the level-tracking arrays. */ for (i = 1; i < NUM_RCU_LVLS; i++) @@ -1829,7 +1835,8 @@ static void __init rcu_init_one(struct rcu_state *rsp) rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { spin_lock_init(&rnp->lock); - lockdep_set_class(&rnp->lock, &rcu_node_class[i]); + lockdep_set_class_and_name(&rnp->lock, + &rcu_node_class[i], buf[i]); rnp->gpnum = 0; rnp->qsmask = 0; rnp->qsmaskinit = 0; -- cgit v1.2.2 From 017c426138122c8e9b9f5057fbd0567c37b35247 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 14 Jan 2010 16:10:58 -0800 Subject: rcu: Fix sparse warnings Rename local variable "i" in rcu_init() to avoid conflict with RCU_INIT_FLAVOR(), restrict the scope of RCU_TREE_NONCORE, and make __synchronize_srcu() static. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12635142581560-git-send-email-> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 6 +++--- kernel/rcutree.h | 6 ++---- kernel/srcu.c | 2 +- 3 files changed, 6 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 3b13d64b010b..099a255ede4c 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1888,7 +1888,7 @@ do { \ void __init rcu_init(void) { - int i; + int cpu; rcu_bootup_announce(); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR @@ -1908,8 +1908,8 @@ void __init rcu_init(void) * or the scheduler are operational. */ cpu_notifier(rcu_cpu_notify, 0); - for_each_online_cpu(i) - rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)i); + for_each_online_cpu(cpu) + rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu); } #include "rcutree_plugin.h" diff --git a/kernel/rcutree.h b/kernel/rcutree.h index bd5d78ad1c48..d9d032abd665 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -323,8 +323,6 @@ struct rcu_state { #define RCU_OFL_TASKS_EXP_GP 0x2 /* Tasks blocking expedited */ /* GP were moved to root. */ -#ifdef RCU_TREE_NONCORE - /* * RCU implementation internal declarations: */ @@ -339,7 +337,7 @@ extern struct rcu_state rcu_preempt_state; DECLARE_PER_CPU(struct rcu_data, rcu_preempt_data); #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ -#else /* #ifdef RCU_TREE_NONCORE */ +#ifndef RCU_TREE_NONCORE /* Forward declarations for rcutree_plugin.h */ static void rcu_bootup_announce(void); @@ -372,4 +370,4 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu); static void rcu_preempt_send_cbs_to_orphanage(void); static void __init __rcu_init_preempt(void); -#endif /* #else #ifdef RCU_TREE_NONCORE */ +#endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/srcu.c b/kernel/srcu.c index 818d7d9aa03c..31b275b9c112 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -144,7 +144,7 @@ EXPORT_SYMBOL_GPL(srcu_read_unlock); /* * Helper function for synchronize_srcu() and synchronize_srcu_expedited(). */ -void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) +static void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) { int idx; -- cgit v1.2.2 From 24a53652e361321b09df5040711e69387344ce09 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 14 Jan 2010 08:23:15 +0100 Subject: tracing: Drop the tr check from the graph tracing path Each time we save a function entry from the function graph tracer, we check if the trace array is set, which is wasteful because it is set anyway before we start the tracer. All we need is to ensure we have good read and write orderings. When we set the trace array, we just need to guarantee it to be visible before starting tracing. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Cc: Lai Jiangshan Cc: Paul E. McKenney LKML-Reference: <1263453795-7496-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index b1342c5d37cf..f2252296607c 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -212,9 +212,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) int cpu; int pc; - if (unlikely(!tr)) - return 0; - if (!ftrace_trace_task(current)) return 0; @@ -287,11 +284,20 @@ void trace_graph_return(struct ftrace_graph_ret *trace) local_irq_restore(flags); } +void set_graph_array(struct trace_array *tr) +{ + graph_array = tr; + + /* Make graph_array visible before we start tracing */ + + smp_mb(); +} + static int graph_trace_init(struct trace_array *tr) { int ret; - graph_array = tr; + set_graph_array(tr); ret = register_ftrace_graph(&trace_graph_return, &trace_graph_entry); if (ret) @@ -301,11 +307,6 @@ static int graph_trace_init(struct trace_array *tr) return 0; } -void set_graph_array(struct trace_array *tr) -{ - graph_array = tr; -} - static void graph_trace_reset(struct trace_array *tr) { tracing_stop_cmdline_record(); -- cgit v1.2.2 From 6d686f4564f3fc7c6e678852919e48ad331d276b Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Wed, 13 Jan 2010 20:21:52 -0700 Subject: sched: Don't expose local functions kernel/sched: don't expose local functions The get_rr_interval_* functions are all class methods of struct sched_class. They are not exported so make them static. Signed-off-by: H Hartley Sweeten Cc: Peter Zijlstra LKML-Reference: <201001132021.53253.hartleys@visionengravers.com> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 2 +- kernel/sched_idletask.c | 2 +- kernel/sched_rt.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 42ac3c9f66f6..71778601c103 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2076,7 +2076,7 @@ static void moved_group_fair(struct task_struct *p, int on_rq) } #endif -unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) { struct sched_entity *se = &task->se; unsigned int rr_interval = 0; diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 5f93b570d383..01332bfc61a7 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -97,7 +97,7 @@ static void prio_changed_idle(struct rq *rq, struct task_struct *p, check_preempt_curr(rq, p, 0); } -unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) { return 0; } diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index f48328ac216f..072b3fcee8d8 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1721,7 +1721,7 @@ static void set_curr_task_rt(struct rq *rq) dequeue_pushable_task(rq, p); } -unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) { /* * Time slice is 0 for SCHED_FIFO tasks -- cgit v1.2.2 From e03bcb68629c7f0728c95f1afe06ce48565c7713 Mon Sep 17 00:00:00 2001 From: Milton Miller Date: Mon, 18 Jan 2010 13:00:51 +1100 Subject: generic-ipi: Optimize accesses by using DEFINE_PER_CPU_SHARED_ALIGNED for IPI data The smp ipi data is passed around and given write access by other cpus and should be separated from per-cpu data consumed by this cpu. Looking for hot lines, I saw call_function_data shared with tick_cpu_sched. Signed-off-by: Milton Miller Acked-by: Anton Blanchard Acked-by: Jens Axboe Cc: Andrew Morton Cc: Linus Torvalds Cc: Peter Zijlstra Cc: : Nick Piggin LKML-Reference: <20100118020051.GR12666@kryten> Signed-off-by: Ingo Molnar --- kernel/smp.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index f10408422444..9867b6bfefce 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -12,8 +12,6 @@ #include #include -static DEFINE_PER_CPU(struct call_single_queue, call_single_queue); - static struct { struct list_head queue; raw_spinlock_t lock; @@ -33,12 +31,14 @@ struct call_function_data { cpumask_var_t cpumask; }; +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_function_data, cfd_data); + struct call_single_queue { struct list_head list; raw_spinlock_t lock; }; -static DEFINE_PER_CPU(struct call_function_data, cfd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_queue, call_single_queue); static int hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) @@ -256,7 +256,7 @@ void generic_smp_call_function_single_interrupt(void) } } -static DEFINE_PER_CPU(struct call_single_data, csd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_data, csd_data); /* * smp_call_function_single - Run a function on a specific CPU -- cgit v1.2.2 From 1e3c88bdeb1260edc341e45c9fb8efd182a5c511 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:00:43 +0100 Subject: sched: Move load balance code into sched_fair.c Straight fwd code movement. Since non of the load-balance abstractions are used anymore, do away with them and simplify the code some. In preparation move the code around. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched.c | 1919 +++------------------------------------------------ kernel/sched_fair.c | 1765 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 1844 insertions(+), 1840 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 64298a52eaa6..13a2acf18b2d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1805,6 +1805,51 @@ static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest) raw_spin_unlock(&busiest->lock); lock_set_subclass(&this_rq->lock.dep_map, 0, _RET_IP_); } + +/* + * double_rq_lock - safely lock two runqueues + * + * Note this does not disable interrupts like task_rq_lock, + * you need to do so manually before calling. + */ +static void double_rq_lock(struct rq *rq1, struct rq *rq2) + __acquires(rq1->lock) + __acquires(rq2->lock) +{ + BUG_ON(!irqs_disabled()); + if (rq1 == rq2) { + raw_spin_lock(&rq1->lock); + __acquire(rq2->lock); /* Fake it out ;) */ + } else { + if (rq1 < rq2) { + raw_spin_lock(&rq1->lock); + raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); + } else { + raw_spin_lock(&rq2->lock); + raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); + } + } + update_rq_clock(rq1); + update_rq_clock(rq2); +} + +/* + * double_rq_unlock - safely unlock two runqueues + * + * Note this does not restore interrupts like task_rq_unlock, + * you need to do so manually after calling. + */ +static void double_rq_unlock(struct rq *rq1, struct rq *rq2) + __releases(rq1->lock) + __releases(rq2->lock) +{ + raw_spin_unlock(&rq1->lock); + if (rq1 != rq2) + raw_spin_unlock(&rq2->lock); + else + __release(rq2->lock); +} + #endif #ifdef CONFIG_FAIR_GROUP_SCHED @@ -1834,18 +1879,14 @@ static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu) #endif } -#include "sched_stats.h" -#include "sched_idletask.c" -#include "sched_fair.c" -#include "sched_rt.c" -#ifdef CONFIG_SCHED_DEBUG -# include "sched_debug.c" -#endif +static const struct sched_class rt_sched_class; #define sched_class_highest (&rt_sched_class) #define for_each_class(class) \ for (class = sched_class_highest; class; class = class->next) +#include "sched_stats.h" + static void inc_nr_running(struct rq *rq) { rq->nr_running++; @@ -1911,6 +1952,37 @@ static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) p->se.on_rq = 0; } +/* + * activate_task - move a task to the runqueue. + */ +static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible--; + + enqueue_task(rq, p, wakeup); + inc_nr_running(rq); +} + +/* + * deactivate_task - remove a task from the runqueue. + */ +static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible++; + + dequeue_task(rq, p, sleep); + dec_nr_running(rq); +} + +#include "sched_idletask.c" +#include "sched_fair.c" +#include "sched_rt.c" +#ifdef CONFIG_SCHED_DEBUG +# include "sched_debug.c" +#endif + /* * __normal_prio - return the priority that is based on the static prio */ @@ -1957,30 +2029,6 @@ static int effective_prio(struct task_struct *p) return p->prio; } -/* - * activate_task - move a task to the runqueue. - */ -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible--; - - enqueue_task(rq, p, wakeup); - inc_nr_running(rq); -} - -/* - * deactivate_task - remove a task from the runqueue. - */ -static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible++; - - dequeue_task(rq, p, sleep); - dec_nr_running(rq); -} - /** * task_curr - is this task currently executing on a CPU? * @p: the task in question. @@ -3087,50 +3135,6 @@ static void update_cpu_load(struct rq *this_rq) #ifdef CONFIG_SMP -/* - * double_rq_lock - safely lock two runqueues - * - * Note this does not disable interrupts like task_rq_lock, - * you need to do so manually before calling. - */ -static void double_rq_lock(struct rq *rq1, struct rq *rq2) - __acquires(rq1->lock) - __acquires(rq2->lock) -{ - BUG_ON(!irqs_disabled()); - if (rq1 == rq2) { - raw_spin_lock(&rq1->lock); - __acquire(rq2->lock); /* Fake it out ;) */ - } else { - if (rq1 < rq2) { - raw_spin_lock(&rq1->lock); - raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); - } else { - raw_spin_lock(&rq2->lock); - raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); - } - } - update_rq_clock(rq1); - update_rq_clock(rq2); -} - -/* - * double_rq_unlock - safely unlock two runqueues - * - * Note this does not restore interrupts like task_rq_unlock, - * you need to do so manually after calling. - */ -static void double_rq_unlock(struct rq *rq1, struct rq *rq2) - __releases(rq1->lock) - __releases(rq2->lock) -{ - raw_spin_unlock(&rq1->lock); - if (rq1 != rq2) - raw_spin_unlock(&rq2->lock); - else - __release(rq2->lock); -} - /* * sched_exec - execve() is a valuable balancing opportunity, because at * this point the task has the smallest effective memory and cache footprint. @@ -3179,1771 +3183,6 @@ again: task_rq_unlock(rq, &flags); } -/* - * pull_task - move a task from a remote runqueue to the local runqueue. - * Both runqueues must be locked. - */ -static void pull_task(struct rq *src_rq, struct task_struct *p, - struct rq *this_rq, int this_cpu) -{ - deactivate_task(src_rq, p, 0); - set_task_cpu(p, this_cpu); - activate_task(this_rq, p, 0); - check_preempt_curr(this_rq, p, 0); -} - -/* - * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? - */ -static -int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - int tsk_cache_hot = 0; - /* - * We do not migrate tasks that are: - * 1) running (obviously), or - * 2) cannot be migrated to this CPU due to cpus_allowed, or - * 3) are cache-hot on their current CPU. - */ - if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { - schedstat_inc(p, se.nr_failed_migrations_affine); - return 0; - } - *all_pinned = 0; - - if (task_running(rq, p)) { - schedstat_inc(p, se.nr_failed_migrations_running); - return 0; - } - - /* - * Aggressive migration if: - * 1) task is cache cold, or - * 2) too many balance attempts have failed. - */ - - tsk_cache_hot = task_hot(p, rq->clock, sd); - if (!tsk_cache_hot || - sd->nr_balance_failed > sd->cache_nice_tries) { -#ifdef CONFIG_SCHEDSTATS - if (tsk_cache_hot) { - schedstat_inc(sd, lb_hot_gained[idle]); - schedstat_inc(p, se.nr_forced_migrations); - } -#endif - return 1; - } - - if (tsk_cache_hot) { - schedstat_inc(p, se.nr_failed_migrations_hot); - return 0; - } - return 1; -} - -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator) -{ - int loops = 0, pulled = 0, pinned = 0; - struct task_struct *p; - long rem_load_move = max_load_move; - - if (max_load_move == 0) - goto out; - - pinned = 1; - - /* - * Start the load-balancing iterator: - */ - p = iterator->start(iterator->arg); -next: - if (!p || loops++ > sysctl_sched_nr_migrate) - goto out; - - if ((p->se.load.weight >> 1) > rem_load_move || - !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - p = iterator->next(iterator->arg); - goto next; - } - - pull_task(busiest, p, this_rq, this_cpu); - pulled++; - rem_load_move -= p->se.load.weight; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible kernels - * will stop after the first task is pulled to minimize the critical - * section. - */ - if (idle == CPU_NEWLY_IDLE) - goto out; -#endif - - /* - * We only want to steal up to the prescribed amount of weighted load. - */ - if (rem_load_move > 0) { - if (p->prio < *this_best_prio) - *this_best_prio = p->prio; - p = iterator->next(iterator->arg); - goto next; - } -out: - /* - * Right now, this is one of only two places pull_task() is called, - * so we can safely collect pull_task() stats here rather than - * inside pull_task(). - */ - schedstat_add(sd, lb_gained[idle], pulled); - - if (all_pinned) - *all_pinned = pinned; - - return max_load_move - rem_load_move; -} - -/* - * move_tasks tries to move up to max_load_move weighted load from busiest to - * this_rq, as part of a balancing operation within domain "sd". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - const struct sched_class *class = sched_class_highest; - unsigned long total_load_moved = 0; - int this_best_prio = this_rq->curr->prio; - - do { - total_load_moved += - class->load_balance(this_rq, this_cpu, busiest, - max_load_move - total_load_moved, - sd, idle, all_pinned, &this_best_prio); - class = class->next; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible - * kernels will stop after the first task is pulled to minimize - * the critical section. - */ - if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) - break; -#endif - } while (class && max_load_move > total_load_moved); - - return total_load_moved > 0; -} - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator) -{ - struct task_struct *p = iterator->start(iterator->arg); - int pinned = 0; - - while (p) { - if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - pull_task(busiest, p, this_rq, this_cpu); - /* - * Right now, this is only the second place pull_task() - * is called, so we can safely collect pull_task() - * stats here rather than inside pull_task(). - */ - schedstat_inc(sd, lb_gained[idle]); - - return 1; - } - p = iterator->next(iterator->arg); - } - - return 0; -} - -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - const struct sched_class *class; - - for_each_class(class) { - if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) - return 1; - } - - return 0; -} -/********** Helpers for find_busiest_group ************************/ -/* - * sd_lb_stats - Structure to store the statistics of a sched_domain - * during load balancing. - */ -struct sd_lb_stats { - struct sched_group *busiest; /* Busiest group in this sd */ - struct sched_group *this; /* Local group in this sd */ - unsigned long total_load; /* Total load of all groups in sd */ - unsigned long total_pwr; /* Total power of all groups in sd */ - unsigned long avg_load; /* Average load across all groups in sd */ - - /** Statistics of this group */ - unsigned long this_load; - unsigned long this_load_per_task; - unsigned long this_nr_running; - - /* Statistics of the busiest group */ - unsigned long max_load; - unsigned long busiest_load_per_task; - unsigned long busiest_nr_running; - - int group_imb; /* Is there imbalance in this sd */ -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - int power_savings_balance; /* Is powersave balance needed for this sd */ - struct sched_group *group_min; /* Least loaded group in sd */ - struct sched_group *group_leader; /* Group which relieves group_min */ - unsigned long min_load_per_task; /* load_per_task in group_min */ - unsigned long leader_nr_running; /* Nr running of group_leader */ - unsigned long min_nr_running; /* Nr running of group_min */ -#endif -}; - -/* - * sg_lb_stats - stats of a sched_group required for load_balancing - */ -struct sg_lb_stats { - unsigned long avg_load; /*Avg load across the CPUs of the group */ - unsigned long group_load; /* Total load over the CPUs of the group */ - unsigned long sum_nr_running; /* Nr tasks running in the group */ - unsigned long sum_weighted_load; /* Weighted load of group's tasks */ - unsigned long group_capacity; - int group_imb; /* Is there an imbalance in the group ? */ -}; - -/** - * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. - * @group: The group whose first cpu is to be returned. - */ -static inline unsigned int group_first_cpu(struct sched_group *group) -{ - return cpumask_first(sched_group_cpus(group)); -} - -/** - * get_sd_load_idx - Obtain the load index for a given sched domain. - * @sd: The sched_domain whose load_idx is to be obtained. - * @idle: The Idle status of the CPU for whose sd load_icx is obtained. - */ -static inline int get_sd_load_idx(struct sched_domain *sd, - enum cpu_idle_type idle) -{ - int load_idx; - - switch (idle) { - case CPU_NOT_IDLE: - load_idx = sd->busy_idx; - break; - - case CPU_NEWLY_IDLE: - load_idx = sd->newidle_idx; - break; - default: - load_idx = sd->idle_idx; - break; - } - - return load_idx; -} - - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * init_sd_power_savings_stats - Initialize power savings statistics for - * the given sched_domain, during load balancing. - * - * @sd: Sched domain whose power-savings statistics are to be initialized. - * @sds: Variable containing the statistics for sd. - * @idle: Idle status of the CPU at which we're performing load-balancing. - */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - /* - * Busy processors will not participate in power savings - * balance. - */ - if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) - sds->power_savings_balance = 0; - else { - sds->power_savings_balance = 1; - sds->min_nr_running = ULONG_MAX; - sds->leader_nr_running = 0; - } -} - -/** - * update_sd_power_savings_stats - Update the power saving stats for a - * sched_domain while performing load balancing. - * - * @group: sched_group belonging to the sched_domain under consideration. - * @sds: Variable containing the statistics of the sched_domain - * @local_group: Does group contain the CPU for which we're performing - * load balancing ? - * @sgs: Variable containing the statistics of the group. - */ -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - - if (!sds->power_savings_balance) - return; - - /* - * If the local group is idle or completely loaded - * no need to do power savings balance at this domain - */ - if (local_group && (sds->this_nr_running >= sgs->group_capacity || - !sds->this_nr_running)) - sds->power_savings_balance = 0; - - /* - * If a group is already running at full capacity or idle, - * don't include that group in power savings calculations - */ - if (!sds->power_savings_balance || - sgs->sum_nr_running >= sgs->group_capacity || - !sgs->sum_nr_running) - return; - - /* - * Calculate the group which has the least non-idle load. - * This is the group from where we need to pick up the load - * for saving power - */ - if ((sgs->sum_nr_running < sds->min_nr_running) || - (sgs->sum_nr_running == sds->min_nr_running && - group_first_cpu(group) > group_first_cpu(sds->group_min))) { - sds->group_min = group; - sds->min_nr_running = sgs->sum_nr_running; - sds->min_load_per_task = sgs->sum_weighted_load / - sgs->sum_nr_running; - } - - /* - * Calculate the group which is almost near its - * capacity but still has some space to pick up some load - * from other group and save more power - */ - if (sgs->sum_nr_running + 1 > sgs->group_capacity) - return; - - if (sgs->sum_nr_running > sds->leader_nr_running || - (sgs->sum_nr_running == sds->leader_nr_running && - group_first_cpu(group) < group_first_cpu(sds->group_leader))) { - sds->group_leader = group; - sds->leader_nr_running = sgs->sum_nr_running; - } -} - -/** - * check_power_save_busiest_group - see if there is potential for some power-savings balance - * @sds: Variable containing the statistics of the sched_domain - * under consideration. - * @this_cpu: Cpu at which we're currently performing load-balancing. - * @imbalance: Variable to store the imbalance. - * - * Description: - * Check if we have potential to perform some power-savings balance. - * If yes, set the busiest group to be the least loaded group in the - * sched_domain, so that it's CPUs can be put to idle. - * - * Returns 1 if there is potential to perform power-savings balance. - * Else returns 0. - */ -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - if (!sds->power_savings_balance) - return 0; - - if (sds->this != sds->group_leader || - sds->group_leader == sds->group_min) - return 0; - - *imbalance = sds->min_load_per_task; - sds->busiest = sds->group_min; - - return 1; - -} -#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - return; -} - -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - return; -} - -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - return 0; -} -#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ - - -unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return SCHED_LOAD_SCALE; -} - -unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return default_scale_freq_power(sd, cpu); -} - -unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long smt_gain = sd->smt_gain; - - smt_gain /= weight; - - return smt_gain; -} - -unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) -{ - return default_scale_smt_power(sd, cpu); -} - -unsigned long scale_rt_power(int cpu) -{ - struct rq *rq = cpu_rq(cpu); - u64 total, available; - - sched_avg_update(rq); - - total = sched_avg_period() + (rq->clock - rq->age_stamp); - available = total - rq->rt_avg; - - if (unlikely((s64)total < SCHED_LOAD_SCALE)) - total = SCHED_LOAD_SCALE; - - total >>= SCHED_LOAD_SHIFT; - - return div_u64(available, total); -} - -static void update_cpu_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long power = SCHED_LOAD_SCALE; - struct sched_group *sdg = sd->groups; - - if (sched_feat(ARCH_POWER)) - power *= arch_scale_freq_power(sd, cpu); - else - power *= default_scale_freq_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - - if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { - if (sched_feat(ARCH_POWER)) - power *= arch_scale_smt_power(sd, cpu); - else - power *= default_scale_smt_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - } - - power *= scale_rt_power(cpu); - power >>= SCHED_LOAD_SHIFT; - - if (!power) - power = 1; - - sdg->cpu_power = power; -} - -static void update_group_power(struct sched_domain *sd, int cpu) -{ - struct sched_domain *child = sd->child; - struct sched_group *group, *sdg = sd->groups; - unsigned long power; - - if (!child) { - update_cpu_power(sd, cpu); - return; - } - - power = 0; - - group = child->groups; - do { - power += group->cpu_power; - group = group->next; - } while (group != child->groups); - - sdg->cpu_power = power; -} - -/** - * update_sg_lb_stats - Update sched_group's statistics for load balancing. - * @sd: The sched_domain whose statistics are to be updated. - * @group: sched_group whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @load_idx: Load index of sched_domain of this_cpu for load calc. - * @sd_idle: Idle status of the sched_domain containing group. - * @local_group: Does group contain this_cpu. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sgs: variable to hold the statistics for this group. - */ -static inline void update_sg_lb_stats(struct sched_domain *sd, - struct sched_group *group, int this_cpu, - enum cpu_idle_type idle, int load_idx, int *sd_idle, - int local_group, const struct cpumask *cpus, - int *balance, struct sg_lb_stats *sgs) -{ - unsigned long load, max_cpu_load, min_cpu_load; - int i; - unsigned int balance_cpu = -1, first_idle_cpu = 0; - unsigned long sum_avg_load_per_task; - unsigned long avg_load_per_task; - - if (local_group) { - balance_cpu = group_first_cpu(group); - if (balance_cpu == this_cpu) - update_group_power(sd, this_cpu); - } - - /* Tally up the load of all CPUs in the group */ - sum_avg_load_per_task = avg_load_per_task = 0; - max_cpu_load = 0; - min_cpu_load = ~0UL; - - for_each_cpu_and(i, sched_group_cpus(group), cpus) { - struct rq *rq = cpu_rq(i); - - if (*sd_idle && rq->nr_running) - *sd_idle = 0; - - /* Bias balancing toward cpus of our domain */ - if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { - first_idle_cpu = 1; - balance_cpu = i; - } - - load = target_load(i, load_idx); - } else { - load = source_load(i, load_idx); - if (load > max_cpu_load) - max_cpu_load = load; - if (min_cpu_load > load) - min_cpu_load = load; - } - - sgs->group_load += load; - sgs->sum_nr_running += rq->nr_running; - sgs->sum_weighted_load += weighted_cpuload(i); - - sum_avg_load_per_task += cpu_avg_load_per_task(i); - } - - /* - * First idle cpu or the first cpu(busiest) in this sched group - * is eligible for doing load balancing at this and above - * domains. In the newly idle case, we will allow all the cpu's - * to do the newly idle load balance. - */ - if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { - *balance = 0; - return; - } - - /* Adjust by relative CPU power of the group */ - sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; - - - /* - * Consider the group unbalanced when the imbalance is larger - * than the average weight of two tasks. - * - * APZ: with cgroup the avg task weight can vary wildly and - * might not be a suitable number - should we keep a - * normalized nr_running number somewhere that negates - * the hierarchy? - */ - avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / - group->cpu_power; - - if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) - sgs->group_imb = 1; - - sgs->group_capacity = - DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); -} - -/** - * update_sd_lb_stats - Update sched_group's statistics for load balancing. - * @sd: sched_domain whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @sd_idle: Idle status of the sched_domain containing group. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sds: variable to hold the statistics for this sched_domain. - */ -static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, - enum cpu_idle_type idle, int *sd_idle, - const struct cpumask *cpus, int *balance, - struct sd_lb_stats *sds) -{ - struct sched_domain *child = sd->child; - struct sched_group *group = sd->groups; - struct sg_lb_stats sgs; - int load_idx, prefer_sibling = 0; - - if (child && child->flags & SD_PREFER_SIBLING) - prefer_sibling = 1; - - init_sd_power_savings_stats(sd, sds, idle); - load_idx = get_sd_load_idx(sd, idle); - - do { - int local_group; - - local_group = cpumask_test_cpu(this_cpu, - sched_group_cpus(group)); - memset(&sgs, 0, sizeof(sgs)); - update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, - local_group, cpus, balance, &sgs); - - if (local_group && balance && !(*balance)) - return; - - sds->total_load += sgs.group_load; - sds->total_pwr += group->cpu_power; - - /* - * In case the child domain prefers tasks go to siblings - * first, lower the group capacity to one so that we'll try - * and move all the excess tasks away. - */ - if (prefer_sibling) - sgs.group_capacity = min(sgs.group_capacity, 1UL); - - if (local_group) { - sds->this_load = sgs.avg_load; - sds->this = group; - sds->this_nr_running = sgs.sum_nr_running; - sds->this_load_per_task = sgs.sum_weighted_load; - } else if (sgs.avg_load > sds->max_load && - (sgs.sum_nr_running > sgs.group_capacity || - sgs.group_imb)) { - sds->max_load = sgs.avg_load; - sds->busiest = group; - sds->busiest_nr_running = sgs.sum_nr_running; - sds->busiest_load_per_task = sgs.sum_weighted_load; - sds->group_imb = sgs.group_imb; - } - - update_sd_power_savings_stats(group, sds, local_group, &sgs); - group = group->next; - } while (group != sd->groups); -} - -/** - * fix_small_imbalance - Calculate the minor imbalance that exists - * amongst the groups of a sched_domain, during - * load balancing. - * @sds: Statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: The cpu at whose sched_domain we're performing load-balance. - * @imbalance: Variable to store the imbalance. - */ -static inline void fix_small_imbalance(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - unsigned long tmp, pwr_now = 0, pwr_move = 0; - unsigned int imbn = 2; - - if (sds->this_nr_running) { - sds->this_load_per_task /= sds->this_nr_running; - if (sds->busiest_load_per_task > - sds->this_load_per_task) - imbn = 1; - } else - sds->this_load_per_task = - cpu_avg_load_per_task(this_cpu); - - if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= - sds->busiest_load_per_task * imbn) { - *imbalance = sds->busiest_load_per_task; - return; - } - - /* - * OK, we don't have enough imbalance to justify moving tasks, - * however we may be able to increase total CPU power used by - * moving them. - */ - - pwr_now += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load); - pwr_now += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load); - pwr_now /= SCHED_LOAD_SCALE; - - /* Amount of load we'd subtract */ - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->busiest->cpu_power; - if (sds->max_load > tmp) - pwr_move += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load - tmp); - - /* Amount of load we'd add */ - if (sds->max_load * sds->busiest->cpu_power < - sds->busiest_load_per_task * SCHED_LOAD_SCALE) - tmp = (sds->max_load * sds->busiest->cpu_power) / - sds->this->cpu_power; - else - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->this->cpu_power; - pwr_move += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load + tmp); - pwr_move /= SCHED_LOAD_SCALE; - - /* Move if we gain throughput */ - if (pwr_move > pwr_now) - *imbalance = sds->busiest_load_per_task; -} - -/** - * calculate_imbalance - Calculate the amount of imbalance present within the - * groups of a given sched_domain during load balance. - * @sds: statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: Cpu for which currently load balance is being performed. - * @imbalance: The variable to store the imbalance. - */ -static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, - unsigned long *imbalance) -{ - unsigned long max_pull; - /* - * In the presence of smp nice balancing, certain scenarios can have - * max load less than avg load(as we skip the groups at or below - * its cpu_power, while calculating max_load..) - */ - if (sds->max_load < sds->avg_load) { - *imbalance = 0; - return fix_small_imbalance(sds, this_cpu, imbalance); - } - - /* Don't want to pull so many tasks that a group would go idle */ - max_pull = min(sds->max_load - sds->avg_load, - sds->max_load - sds->busiest_load_per_task); - - /* How much load to actually move to equalise the imbalance */ - *imbalance = min(max_pull * sds->busiest->cpu_power, - (sds->avg_load - sds->this_load) * sds->this->cpu_power) - / SCHED_LOAD_SCALE; - - /* - * if *imbalance is less than the average load per runnable task - * there is no gaurantee that any tasks will be moved so we'll have - * a think about bumping its value to force at least one task to be - * moved - */ - if (*imbalance < sds->busiest_load_per_task) - return fix_small_imbalance(sds, this_cpu, imbalance); - -} -/******* find_busiest_group() helpers end here *********************/ - -/** - * find_busiest_group - Returns the busiest group within the sched_domain - * if there is an imbalance. If there isn't an imbalance, and - * the user has opted for power-savings, it returns a group whose - * CPUs can be put to idle by rebalancing those tasks elsewhere, if - * such a group exists. - * - * Also calculates the amount of weighted load which should be moved - * to restore balance. - * - * @sd: The sched_domain whose busiest group is to be returned. - * @this_cpu: The cpu for which load balancing is currently being performed. - * @imbalance: Variable which stores amount of weighted load which should - * be moved to restore balance/put a group to idle. - * @idle: The idle status of this_cpu. - * @sd_idle: The idleness of sd - * @cpus: The set of CPUs under consideration for load-balancing. - * @balance: Pointer to a variable indicating if this_cpu - * is the appropriate cpu to perform load balancing at this_level. - * - * Returns: - the busiest group if imbalance exists. - * - If no imbalance and user has opted for power-savings balance, - * return the least loaded group whose CPUs can be - * put to idle by rebalancing its tasks onto our group. - */ -static struct sched_group * -find_busiest_group(struct sched_domain *sd, int this_cpu, - unsigned long *imbalance, enum cpu_idle_type idle, - int *sd_idle, const struct cpumask *cpus, int *balance) -{ - struct sd_lb_stats sds; - - memset(&sds, 0, sizeof(sds)); - - /* - * Compute the various statistics relavent for load balancing at - * this level. - */ - update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, - balance, &sds); - - /* Cases where imbalance does not exist from POV of this_cpu */ - /* 1) this_cpu is not the appropriate cpu to perform load balancing - * at this level. - * 2) There is no busy sibling group to pull from. - * 3) This group is the busiest group. - * 4) This group is more busy than the avg busieness at this - * sched_domain. - * 5) The imbalance is within the specified limit. - * 6) Any rebalance would lead to ping-pong - */ - if (balance && !(*balance)) - goto ret; - - if (!sds.busiest || sds.busiest_nr_running == 0) - goto out_balanced; - - if (sds.this_load >= sds.max_load) - goto out_balanced; - - sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; - - if (sds.this_load >= sds.avg_load) - goto out_balanced; - - if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) - goto out_balanced; - - sds.busiest_load_per_task /= sds.busiest_nr_running; - if (sds.group_imb) - sds.busiest_load_per_task = - min(sds.busiest_load_per_task, sds.avg_load); - - /* - * We're trying to get all the cpus to the average_load, so we don't - * want to push ourselves above the average load, nor do we wish to - * reduce the max loaded cpu below the average load, as either of these - * actions would just result in more rebalancing later, and ping-pong - * tasks around. Thus we look for the minimum possible imbalance. - * Negative imbalances (*we* are more loaded than anyone else) will - * be counted as no imbalance for these purposes -- we can't fix that - * by pulling tasks to us. Be careful of negative numbers as they'll - * appear as very large values with unsigned longs. - */ - if (sds.max_load <= sds.busiest_load_per_task) - goto out_balanced; - - /* Looks like there is an imbalance. Compute it */ - calculate_imbalance(&sds, this_cpu, imbalance); - return sds.busiest; - -out_balanced: - /* - * There is no obvious imbalance. But check if we can do some balancing - * to save power. - */ - if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) - return sds.busiest; -ret: - *imbalance = 0; - return NULL; -} - -/* - * find_busiest_queue - find the busiest runqueue among the cpus in group. - */ -static struct rq * -find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, - unsigned long imbalance, const struct cpumask *cpus) -{ - struct rq *busiest = NULL, *rq; - unsigned long max_load = 0; - int i; - - for_each_cpu(i, sched_group_cpus(group)) { - unsigned long power = power_of(i); - unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); - unsigned long wl; - - if (!cpumask_test_cpu(i, cpus)) - continue; - - rq = cpu_rq(i); - wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; - wl /= power; - - if (capacity && rq->nr_running == 1 && wl > imbalance) - continue; - - if (wl > max_load) { - max_load = wl; - busiest = rq; - } - } - - return busiest; -} - -/* - * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but - * so long as it is large enough. - */ -#define MAX_PINNED_INTERVAL 512 - -/* Working cpumask for load_balance and load_balance_newidle. */ -static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - */ -static int load_balance(int this_cpu, struct rq *this_rq, - struct sched_domain *sd, enum cpu_idle_type idle, - int *balance) -{ - int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; - struct sched_group *group; - unsigned long imbalance; - struct rq *busiest; - unsigned long flags; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as CPU_IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[idle]); - -redo: - update_shares(sd); - group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, - cpus, balance); - - if (*balance == 0) - goto out_balanced; - - if (!group) { - schedstat_inc(sd, lb_nobusyg[idle]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, idle, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[idle]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[idle], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* - * Attempt to move tasks. If find_busiest_group has found - * an imbalance but busiest->nr_running <= 1, the group is - * still unbalanced. ld_moved simply stays zero, so it is - * correctly treated as an imbalance. - */ - local_irq_save(flags); - double_rq_lock(this_rq, busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, idle, &all_pinned); - double_rq_unlock(this_rq, busiest); - local_irq_restore(flags); - - /* - * some other cpu did the load balance for us. - */ - if (ld_moved && this_cpu != smp_processor_id()) - resched_cpu(this_cpu); - - /* All tasks on this runqueue were pinned by CPU affinity */ - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - goto out_balanced; - } - } - - if (!ld_moved) { - schedstat_inc(sd, lb_failed[idle]); - sd->nr_balance_failed++; - - if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { - - raw_spin_lock_irqsave(&busiest->lock, flags); - - /* don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, - &busiest->curr->cpus_allowed)) { - raw_spin_unlock_irqrestore(&busiest->lock, - flags); - all_pinned = 1; - goto out_one_pinned; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - raw_spin_unlock_irqrestore(&busiest->lock, flags); - if (active_balance) - wake_up_process(busiest->migration_thread); - - /* - * We've kicked active balancing, reset the failure - * counter. - */ - sd->nr_balance_failed = sd->cache_nice_tries+1; - } - } else - sd->nr_balance_failed = 0; - - if (likely(!active_balance)) { - /* We were unbalanced, so reset the balancing interval */ - sd->balance_interval = sd->min_interval; - } else { - /* - * If we've begun active balancing, start to back off. This - * case may not be covered by the all_pinned logic if there - * is only 1 task on the busy runqueue (because we don't call - * move_tasks). - */ - if (sd->balance_interval < sd->max_interval) - sd->balance_interval *= 2; - } - - if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - - goto out; - -out_balanced: - schedstat_inc(sd, lb_balanced[idle]); - - sd->nr_balance_failed = 0; - -out_one_pinned: - /* tune up the balancing interval */ - if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || - (sd->balance_interval < sd->max_interval)) - sd->balance_interval *= 2; - - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - else - ld_moved = 0; -out: - if (ld_moved) - update_shares(sd); - return ld_moved; -} - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - * - * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). - * this_rq is locked. - */ -static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) -{ - struct sched_group *group; - struct rq *busiest = NULL; - unsigned long imbalance; - int ld_moved = 0; - int sd_idle = 0; - int all_pinned = 0; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); -redo: - update_shares_locked(this_rq, sd); - group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, - &sd_idle, cpus, NULL); - if (!group) { - schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* Attempt to move tasks */ - double_lock_balance(this_rq, busiest); - /* this_rq->clock is already updated */ - update_rq_clock(busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, CPU_NEWLY_IDLE, - &all_pinned); - double_unlock_balance(this_rq, busiest); - - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - } - } - - if (!ld_moved) { - int active_balance = 0; - - schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - - if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) - return -1; - - if (sd->nr_balance_failed++ < 2) - return -1; - - /* - * The only task running in a non-idle cpu can be moved to this - * cpu in an attempt to completely freeup the other CPU - * package. The same method used to move task in load_balance() - * have been extended for load_balance_newidle() to speedup - * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) - * - * The package power saving logic comes from - * find_busiest_group(). If there are no imbalance, then - * f_b_g() will return NULL. However when sched_mc={1,2} then - * f_b_g() will select a group from which a running task may be - * pulled to this cpu in order to make the other package idle. - * If there is no opportunity to make a package idle and if - * there are no imbalance, then f_b_g() will return NULL and no - * action will be taken in load_balance_newidle(). - * - * Under normal task pull operation due to imbalance, there - * will be more than one task in the source run queue and - * move_tasks() will succeed. ld_moved will be true and this - * active balance code will not be triggered. - */ - - /* Lock busiest in correct order while this_rq is held */ - double_lock_balance(this_rq, busiest); - - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { - double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - - } else - sd->nr_balance_failed = 0; - - update_shares_locked(this_rq, sd); - return ld_moved; - -out_balanced: - schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - sd->nr_balance_failed = 0; - - return 0; -} - -/* - * idle_balance is called by schedule() if this_cpu is about to become - * idle. Attempts to pull tasks from other CPUs. - */ -static void idle_balance(int this_cpu, struct rq *this_rq) -{ - struct sched_domain *sd; - int pulled_task = 0; - unsigned long next_balance = jiffies + HZ; - - this_rq->idle_stamp = this_rq->clock; - - if (this_rq->avg_idle < sysctl_sched_migration_cost) - return; - - for_each_domain(this_cpu, sd) { - unsigned long interval; - - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - if (sd->flags & SD_BALANCE_NEWIDLE) - /* If we've pulled tasks over stop searching: */ - pulled_task = load_balance_newidle(this_cpu, this_rq, - sd); - - interval = msecs_to_jiffies(sd->balance_interval); - if (time_after(next_balance, sd->last_balance + interval)) - next_balance = sd->last_balance + interval; - if (pulled_task) { - this_rq->idle_stamp = 0; - break; - } - } - if (pulled_task || time_after(jiffies, this_rq->next_balance)) { - /* - * We are going idle. next_balance may be set based on - * a busy processor. So reset next_balance. - */ - this_rq->next_balance = next_balance; - } -} - -/* - * active_load_balance is run by migration threads. It pushes running tasks - * off the busiest CPU onto idle CPUs. It requires at least 1 task to be - * running on each physical CPU where possible, and avoids physical / - * logical imbalances. - * - * Called with busiest_rq locked. - */ -static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) -{ - int target_cpu = busiest_rq->push_cpu; - struct sched_domain *sd; - struct rq *target_rq; - - /* Is there any task to move? */ - if (busiest_rq->nr_running <= 1) - return; - - target_rq = cpu_rq(target_cpu); - - /* - * This condition is "impossible", if it occurs - * we need to fix it. Originally reported by - * Bjorn Helgaas on a 128-cpu setup. - */ - BUG_ON(busiest_rq == target_rq); - - /* move a task from busiest_rq to target_rq */ - double_lock_balance(busiest_rq, target_rq); - update_rq_clock(busiest_rq); - update_rq_clock(target_rq); - - /* Search for an sd spanning us and the target CPU. */ - for_each_domain(target_cpu, sd) { - if ((sd->flags & SD_LOAD_BALANCE) && - cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) - break; - } - - if (likely(sd)) { - schedstat_inc(sd, alb_count); - - if (move_one_task(target_rq, target_cpu, busiest_rq, - sd, CPU_IDLE)) - schedstat_inc(sd, alb_pushed); - else - schedstat_inc(sd, alb_failed); - } - double_unlock_balance(busiest_rq, target_rq); -} - -#ifdef CONFIG_NO_HZ -static struct { - atomic_t load_balancer; - cpumask_var_t cpu_mask; - cpumask_var_t ilb_grp_nohz_mask; -} nohz ____cacheline_aligned = { - .load_balancer = ATOMIC_INIT(-1), -}; - -int get_nohz_load_balancer(void) -{ - return atomic_read(&nohz.load_balancer); -} - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * lowest_flag_domain - Return lowest sched_domain containing flag. - * @cpu: The cpu whose lowest level of sched domain is to - * be returned. - * @flag: The flag to check for the lowest sched_domain - * for the given cpu. - * - * Returns the lowest sched_domain of a cpu which contains the given flag. - */ -static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) -{ - struct sched_domain *sd; - - for_each_domain(cpu, sd) - if (sd && (sd->flags & flag)) - break; - - return sd; -} - -/** - * for_each_flag_domain - Iterates over sched_domains containing the flag. - * @cpu: The cpu whose domains we're iterating over. - * @sd: variable holding the value of the power_savings_sd - * for cpu. - * @flag: The flag to filter the sched_domains to be iterated. - * - * Iterates over all the scheduler domains for a given cpu that has the 'flag' - * set, starting from the lowest sched_domain to the highest. - */ -#define for_each_flag_domain(cpu, sd, flag) \ - for (sd = lowest_flag_domain(cpu, flag); \ - (sd && (sd->flags & flag)); sd = sd->parent) - -/** - * is_semi_idle_group - Checks if the given sched_group is semi-idle. - * @ilb_group: group to be checked for semi-idleness - * - * Returns: 1 if the group is semi-idle. 0 otherwise. - * - * We define a sched_group to be semi idle if it has atleast one idle-CPU - * and atleast one non-idle CPU. This helper function checks if the given - * sched_group is semi-idle or not. - */ -static inline int is_semi_idle_group(struct sched_group *ilb_group) -{ - cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, - sched_group_cpus(ilb_group)); - - /* - * A sched_group is semi-idle when it has atleast one busy cpu - * and atleast one idle cpu. - */ - if (cpumask_empty(nohz.ilb_grp_nohz_mask)) - return 0; - - if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) - return 0; - - return 1; -} -/** - * find_new_ilb - Finds the optimum idle load balancer for nomination. - * @cpu: The cpu which is nominating a new idle_load_balancer. - * - * Returns: Returns the id of the idle load balancer if it exists, - * Else, returns >= nr_cpu_ids. - * - * This algorithm picks the idle load balancer such that it belongs to a - * semi-idle powersavings sched_domain. The idea is to try and avoid - * completely idle packages/cores just for the purpose of idle load balancing - * when there are other idle cpu's which are better suited for that job. - */ -static int find_new_ilb(int cpu) -{ - struct sched_domain *sd; - struct sched_group *ilb_group; - - /* - * Have idle load balancer selection from semi-idle packages only - * when power-aware load balancing is enabled - */ - if (!(sched_smt_power_savings || sched_mc_power_savings)) - goto out_done; - - /* - * Optimize for the case when we have no idle CPUs or only one - * idle CPU. Don't walk the sched_domain hierarchy in such cases - */ - if (cpumask_weight(nohz.cpu_mask) < 2) - goto out_done; - - for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { - ilb_group = sd->groups; - - do { - if (is_semi_idle_group(ilb_group)) - return cpumask_first(nohz.ilb_grp_nohz_mask); - - ilb_group = ilb_group->next; - - } while (ilb_group != sd->groups); - } - -out_done: - return cpumask_first(nohz.cpu_mask); -} -#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ -static inline int find_new_ilb(int call_cpu) -{ - return cpumask_first(nohz.cpu_mask); -} -#endif - -/* - * This routine will try to nominate the ilb (idle load balancing) - * owner among the cpus whose ticks are stopped. ilb owner will do the idle - * load balancing on behalf of all those cpus. If all the cpus in the system - * go into this tickless mode, then there will be no ilb owner (as there is - * no need for one) and all the cpus will sleep till the next wakeup event - * arrives... - * - * For the ilb owner, tick is not stopped. And this tick will be used - * for idle load balancing. ilb owner will still be part of - * nohz.cpu_mask.. - * - * While stopping the tick, this cpu will become the ilb owner if there - * is no other owner. And will be the owner till that cpu becomes busy - * or if all cpus in the system stop their ticks at which point - * there is no need for ilb owner. - * - * When the ilb owner becomes busy, it nominates another owner, during the - * next busy scheduler_tick() - */ -int select_nohz_load_balancer(int stop_tick) -{ - int cpu = smp_processor_id(); - - if (stop_tick) { - cpu_rq(cpu)->in_nohz_recently = 1; - - if (!cpu_active(cpu)) { - if (atomic_read(&nohz.load_balancer) != cpu) - return 0; - - /* - * If we are going offline and still the leader, - * give up! - */ - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - - return 0; - } - - cpumask_set_cpu(cpu, nohz.cpu_mask); - - /* time for ilb owner also to sleep */ - if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { - if (atomic_read(&nohz.load_balancer) == cpu) - atomic_set(&nohz.load_balancer, -1); - return 0; - } - - if (atomic_read(&nohz.load_balancer) == -1) { - /* make me the ilb owner */ - if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) - return 1; - } else if (atomic_read(&nohz.load_balancer) == cpu) { - int new_ilb; - - if (!(sched_smt_power_savings || - sched_mc_power_savings)) - return 1; - /* - * Check to see if there is a more power-efficient - * ilb. - */ - new_ilb = find_new_ilb(cpu); - if (new_ilb < nr_cpu_ids && new_ilb != cpu) { - atomic_set(&nohz.load_balancer, -1); - resched_cpu(new_ilb); - return 0; - } - return 1; - } - } else { - if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) - return 0; - - cpumask_clear_cpu(cpu, nohz.cpu_mask); - - if (atomic_read(&nohz.load_balancer) == cpu) - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - } - return 0; -} -#endif - -static DEFINE_SPINLOCK(balancing); - -/* - * It checks each scheduling domain to see if it is due to be balanced, - * and initiates a balancing operation if so. - * - * Balancing parameters are set up in arch_init_sched_domains. - */ -static void rebalance_domains(int cpu, enum cpu_idle_type idle) -{ - int balance = 1; - struct rq *rq = cpu_rq(cpu); - unsigned long interval; - struct sched_domain *sd; - /* Earliest time when we have to do rebalance again */ - unsigned long next_balance = jiffies + 60*HZ; - int update_next_balance = 0; - int need_serialize; - - for_each_domain(cpu, sd) { - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - interval = sd->balance_interval; - if (idle != CPU_IDLE) - interval *= sd->busy_factor; - - /* scale ms to jiffies */ - interval = msecs_to_jiffies(interval); - if (unlikely(!interval)) - interval = 1; - if (interval > HZ*NR_CPUS/10) - interval = HZ*NR_CPUS/10; - - need_serialize = sd->flags & SD_SERIALIZE; - - if (need_serialize) { - if (!spin_trylock(&balancing)) - goto out; - } - - if (time_after_eq(jiffies, sd->last_balance + interval)) { - if (load_balance(cpu, rq, sd, idle, &balance)) { - /* - * We've pulled tasks over so either we're no - * longer idle, or one of our SMT siblings is - * not idle. - */ - idle = CPU_NOT_IDLE; - } - sd->last_balance = jiffies; - } - if (need_serialize) - spin_unlock(&balancing); -out: - if (time_after(next_balance, sd->last_balance + interval)) { - next_balance = sd->last_balance + interval; - update_next_balance = 1; - } - - /* - * Stop the load balance at this level. There is another - * CPU in our sched group which is doing load balancing more - * actively. - */ - if (!balance) - break; - } - - /* - * next_balance will be updated only when there is a need. - * When the cpu is attached to null domain for ex, it will not be - * updated. - */ - if (likely(update_next_balance)) - rq->next_balance = next_balance; -} - -/* - * run_rebalance_domains is triggered when needed from the scheduler tick. - * In CONFIG_NO_HZ case, the idle load balance owner will do the - * rebalancing for all the cpus for whom scheduler ticks are stopped. - */ -static void run_rebalance_domains(struct softirq_action *h) -{ - int this_cpu = smp_processor_id(); - struct rq *this_rq = cpu_rq(this_cpu); - enum cpu_idle_type idle = this_rq->idle_at_tick ? - CPU_IDLE : CPU_NOT_IDLE; - - rebalance_domains(this_cpu, idle); - -#ifdef CONFIG_NO_HZ - /* - * If this cpu is the owner for idle load balancing, then do the - * balancing on behalf of the other idle cpus whose ticks are - * stopped. - */ - if (this_rq->idle_at_tick && - atomic_read(&nohz.load_balancer) == this_cpu) { - struct rq *rq; - int balance_cpu; - - for_each_cpu(balance_cpu, nohz.cpu_mask) { - if (balance_cpu == this_cpu) - continue; - - /* - * If this cpu gets work to do, stop the load balancing - * work being done for other cpus. Next load - * balancing owner will pick it up. - */ - if (need_resched()) - break; - - rebalance_domains(balance_cpu, CPU_IDLE); - - rq = cpu_rq(balance_cpu); - if (time_after(this_rq->next_balance, rq->next_balance)) - this_rq->next_balance = rq->next_balance; - } - } -#endif -} - -static inline int on_null_domain(int cpu) -{ - return !rcu_dereference(cpu_rq(cpu)->sd); -} - -/* - * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. - * - * In case of CONFIG_NO_HZ, this is the place where we nominate a new - * idle load balancing owner or decide to stop the periodic load balancing, - * if the whole system is idle. - */ -static inline void trigger_load_balance(struct rq *rq, int cpu) -{ -#ifdef CONFIG_NO_HZ - /* - * If we were in the nohz mode recently and busy at the current - * scheduler tick, then check if we need to nominate new idle - * load balancer. - */ - if (rq->in_nohz_recently && !rq->idle_at_tick) { - rq->in_nohz_recently = 0; - - if (atomic_read(&nohz.load_balancer) == cpu) { - cpumask_clear_cpu(cpu, nohz.cpu_mask); - atomic_set(&nohz.load_balancer, -1); - } - - if (atomic_read(&nohz.load_balancer) == -1) { - int ilb = find_new_ilb(cpu); - - if (ilb < nr_cpu_ids) - resched_cpu(ilb); - } - } - - /* - * If this cpu is idle and doing idle load balancing for all the - * cpus with ticks stopped, is it time for that to stop? - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && - cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { - resched_cpu(cpu); - return; - } - - /* - * If this cpu is idle and the idle load balancing is done by - * someone else, then no need raise the SCHED_SOFTIRQ - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && - cpumask_test_cpu(cpu, nohz.cpu_mask)) - return; -#endif - /* Don't need to rebalance while attached to NULL domain */ - if (time_after_eq(jiffies, rq->next_balance) && - likely(!on_null_domain(cpu))) - raise_softirq(SCHED_SOFTIRQ); -} - -#else /* CONFIG_SMP */ - -/* - * on UP we do not need to balance between CPUs: - */ -static inline void idle_balance(int cpu, struct rq *rq) -{ -} - #endif DEFINE_PER_CPU(struct kernel_stat, kstat); diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 71778601c103..5116b81d7727 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1952,6 +1952,1762 @@ move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, return 0; } +/* + * pull_task - move a task from a remote runqueue to the local runqueue. + * Both runqueues must be locked. + */ +static void pull_task(struct rq *src_rq, struct task_struct *p, + struct rq *this_rq, int this_cpu) +{ + deactivate_task(src_rq, p, 0); + set_task_cpu(p, this_cpu); + activate_task(this_rq, p, 0); + check_preempt_curr(this_rq, p, 0); +} + +/* + * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? + */ +static +int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) +{ + int tsk_cache_hot = 0; + /* + * We do not migrate tasks that are: + * 1) running (obviously), or + * 2) cannot be migrated to this CPU due to cpus_allowed, or + * 3) are cache-hot on their current CPU. + */ + if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { + schedstat_inc(p, se.nr_failed_migrations_affine); + return 0; + } + *all_pinned = 0; + + if (task_running(rq, p)) { + schedstat_inc(p, se.nr_failed_migrations_running); + return 0; + } + + /* + * Aggressive migration if: + * 1) task is cache cold, or + * 2) too many balance attempts have failed. + */ + + tsk_cache_hot = task_hot(p, rq->clock, sd); + if (!tsk_cache_hot || + sd->nr_balance_failed > sd->cache_nice_tries) { +#ifdef CONFIG_SCHEDSTATS + if (tsk_cache_hot) { + schedstat_inc(sd, lb_hot_gained[idle]); + schedstat_inc(p, se.nr_forced_migrations); + } +#endif + return 1; + } + + if (tsk_cache_hot) { + schedstat_inc(p, se.nr_failed_migrations_hot); + return 0; + } + return 1; +} + +static unsigned long +balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, struct sched_domain *sd, + enum cpu_idle_type idle, int *all_pinned, + int *this_best_prio, struct rq_iterator *iterator) +{ + int loops = 0, pulled = 0, pinned = 0; + struct task_struct *p; + long rem_load_move = max_load_move; + + if (max_load_move == 0) + goto out; + + pinned = 1; + + /* + * Start the load-balancing iterator: + */ + p = iterator->start(iterator->arg); +next: + if (!p || loops++ > sysctl_sched_nr_migrate) + goto out; + + if ((p->se.load.weight >> 1) > rem_load_move || + !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { + p = iterator->next(iterator->arg); + goto next; + } + + pull_task(busiest, p, this_rq, this_cpu); + pulled++; + rem_load_move -= p->se.load.weight; + +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible kernels + * will stop after the first task is pulled to minimize the critical + * section. + */ + if (idle == CPU_NEWLY_IDLE) + goto out; +#endif + + /* + * We only want to steal up to the prescribed amount of weighted load. + */ + if (rem_load_move > 0) { + if (p->prio < *this_best_prio) + *this_best_prio = p->prio; + p = iterator->next(iterator->arg); + goto next; + } +out: + /* + * Right now, this is one of only two places pull_task() is called, + * so we can safely collect pull_task() stats here rather than + * inside pull_task(). + */ + schedstat_add(sd, lb_gained[idle], pulled); + + if (all_pinned) + *all_pinned = pinned; + + return max_load_move - rem_load_move; +} + +/* + * move_tasks tries to move up to max_load_move weighted load from busiest to + * this_rq, as part of a balancing operation within domain "sd". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) +{ + const struct sched_class *class = sched_class_highest; + unsigned long total_load_moved = 0; + int this_best_prio = this_rq->curr->prio; + + do { + total_load_moved += + class->load_balance(this_rq, this_cpu, busiest, + max_load_move - total_load_moved, + sd, idle, all_pinned, &this_best_prio); + class = class->next; + +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ + if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) + break; +#endif + } while (class && max_load_move > total_load_moved); + + return total_load_moved > 0; +} + +static int +iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle, + struct rq_iterator *iterator) +{ + struct task_struct *p = iterator->start(iterator->arg); + int pinned = 0; + + while (p) { + if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { + pull_task(busiest, p, this_rq, this_cpu); + /* + * Right now, this is only the second place pull_task() + * is called, so we can safely collect pull_task() + * stats here rather than inside pull_task(). + */ + schedstat_inc(sd, lb_gained[idle]); + + return 1; + } + p = iterator->next(iterator->arg); + } + + return 0; +} + +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) +{ + const struct sched_class *class; + + for_each_class(class) { + if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) + return 1; + } + + return 0; +} +/********** Helpers for find_busiest_group ************************/ +/* + * sd_lb_stats - Structure to store the statistics of a sched_domain + * during load balancing. + */ +struct sd_lb_stats { + struct sched_group *busiest; /* Busiest group in this sd */ + struct sched_group *this; /* Local group in this sd */ + unsigned long total_load; /* Total load of all groups in sd */ + unsigned long total_pwr; /* Total power of all groups in sd */ + unsigned long avg_load; /* Average load across all groups in sd */ + + /** Statistics of this group */ + unsigned long this_load; + unsigned long this_load_per_task; + unsigned long this_nr_running; + + /* Statistics of the busiest group */ + unsigned long max_load; + unsigned long busiest_load_per_task; + unsigned long busiest_nr_running; + + int group_imb; /* Is there imbalance in this sd */ +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) + int power_savings_balance; /* Is powersave balance needed for this sd */ + struct sched_group *group_min; /* Least loaded group in sd */ + struct sched_group *group_leader; /* Group which relieves group_min */ + unsigned long min_load_per_task; /* load_per_task in group_min */ + unsigned long leader_nr_running; /* Nr running of group_leader */ + unsigned long min_nr_running; /* Nr running of group_min */ +#endif +}; + +/* + * sg_lb_stats - stats of a sched_group required for load_balancing + */ +struct sg_lb_stats { + unsigned long avg_load; /*Avg load across the CPUs of the group */ + unsigned long group_load; /* Total load over the CPUs of the group */ + unsigned long sum_nr_running; /* Nr tasks running in the group */ + unsigned long sum_weighted_load; /* Weighted load of group's tasks */ + unsigned long group_capacity; + int group_imb; /* Is there an imbalance in the group ? */ +}; + +/** + * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. + * @group: The group whose first cpu is to be returned. + */ +static inline unsigned int group_first_cpu(struct sched_group *group) +{ + return cpumask_first(sched_group_cpus(group)); +} + +/** + * get_sd_load_idx - Obtain the load index for a given sched domain. + * @sd: The sched_domain whose load_idx is to be obtained. + * @idle: The Idle status of the CPU for whose sd load_icx is obtained. + */ +static inline int get_sd_load_idx(struct sched_domain *sd, + enum cpu_idle_type idle) +{ + int load_idx; + + switch (idle) { + case CPU_NOT_IDLE: + load_idx = sd->busy_idx; + break; + + case CPU_NEWLY_IDLE: + load_idx = sd->newidle_idx; + break; + default: + load_idx = sd->idle_idx; + break; + } + + return load_idx; +} + + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * init_sd_power_savings_stats - Initialize power savings statistics for + * the given sched_domain, during load balancing. + * + * @sd: Sched domain whose power-savings statistics are to be initialized. + * @sds: Variable containing the statistics for sd. + * @idle: Idle status of the CPU at which we're performing load-balancing. + */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + /* + * Busy processors will not participate in power savings + * balance. + */ + if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) + sds->power_savings_balance = 0; + else { + sds->power_savings_balance = 1; + sds->min_nr_running = ULONG_MAX; + sds->leader_nr_running = 0; + } +} + +/** + * update_sd_power_savings_stats - Update the power saving stats for a + * sched_domain while performing load balancing. + * + * @group: sched_group belonging to the sched_domain under consideration. + * @sds: Variable containing the statistics of the sched_domain + * @local_group: Does group contain the CPU for which we're performing + * load balancing ? + * @sgs: Variable containing the statistics of the group. + */ +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + + if (!sds->power_savings_balance) + return; + + /* + * If the local group is idle or completely loaded + * no need to do power savings balance at this domain + */ + if (local_group && (sds->this_nr_running >= sgs->group_capacity || + !sds->this_nr_running)) + sds->power_savings_balance = 0; + + /* + * If a group is already running at full capacity or idle, + * don't include that group in power savings calculations + */ + if (!sds->power_savings_balance || + sgs->sum_nr_running >= sgs->group_capacity || + !sgs->sum_nr_running) + return; + + /* + * Calculate the group which has the least non-idle load. + * This is the group from where we need to pick up the load + * for saving power + */ + if ((sgs->sum_nr_running < sds->min_nr_running) || + (sgs->sum_nr_running == sds->min_nr_running && + group_first_cpu(group) > group_first_cpu(sds->group_min))) { + sds->group_min = group; + sds->min_nr_running = sgs->sum_nr_running; + sds->min_load_per_task = sgs->sum_weighted_load / + sgs->sum_nr_running; + } + + /* + * Calculate the group which is almost near its + * capacity but still has some space to pick up some load + * from other group and save more power + */ + if (sgs->sum_nr_running + 1 > sgs->group_capacity) + return; + + if (sgs->sum_nr_running > sds->leader_nr_running || + (sgs->sum_nr_running == sds->leader_nr_running && + group_first_cpu(group) < group_first_cpu(sds->group_leader))) { + sds->group_leader = group; + sds->leader_nr_running = sgs->sum_nr_running; + } +} + +/** + * check_power_save_busiest_group - see if there is potential for some power-savings balance + * @sds: Variable containing the statistics of the sched_domain + * under consideration. + * @this_cpu: Cpu at which we're currently performing load-balancing. + * @imbalance: Variable to store the imbalance. + * + * Description: + * Check if we have potential to perform some power-savings balance. + * If yes, set the busiest group to be the least loaded group in the + * sched_domain, so that it's CPUs can be put to idle. + * + * Returns 1 if there is potential to perform power-savings balance. + * Else returns 0. + */ +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + if (!sds->power_savings_balance) + return 0; + + if (sds->this != sds->group_leader || + sds->group_leader == sds->group_min) + return 0; + + *imbalance = sds->min_load_per_task; + sds->busiest = sds->group_min; + + return 1; + +} +#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + return; +} + +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + return; +} + +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + return 0; +} +#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ + + +unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return SCHED_LOAD_SCALE; +} + +unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return default_scale_freq_power(sd, cpu); +} + +unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long smt_gain = sd->smt_gain; + + smt_gain /= weight; + + return smt_gain; +} + +unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) +{ + return default_scale_smt_power(sd, cpu); +} + +unsigned long scale_rt_power(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + u64 total, available; + + sched_avg_update(rq); + + total = sched_avg_period() + (rq->clock - rq->age_stamp); + available = total - rq->rt_avg; + + if (unlikely((s64)total < SCHED_LOAD_SCALE)) + total = SCHED_LOAD_SCALE; + + total >>= SCHED_LOAD_SHIFT; + + return div_u64(available, total); +} + +static void update_cpu_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long power = SCHED_LOAD_SCALE; + struct sched_group *sdg = sd->groups; + + if (sched_feat(ARCH_POWER)) + power *= arch_scale_freq_power(sd, cpu); + else + power *= default_scale_freq_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + + if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { + if (sched_feat(ARCH_POWER)) + power *= arch_scale_smt_power(sd, cpu); + else + power *= default_scale_smt_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + } + + power *= scale_rt_power(cpu); + power >>= SCHED_LOAD_SHIFT; + + if (!power) + power = 1; + + sdg->cpu_power = power; +} + +static void update_group_power(struct sched_domain *sd, int cpu) +{ + struct sched_domain *child = sd->child; + struct sched_group *group, *sdg = sd->groups; + unsigned long power; + + if (!child) { + update_cpu_power(sd, cpu); + return; + } + + power = 0; + + group = child->groups; + do { + power += group->cpu_power; + group = group->next; + } while (group != child->groups); + + sdg->cpu_power = power; +} + +/** + * update_sg_lb_stats - Update sched_group's statistics for load balancing. + * @sd: The sched_domain whose statistics are to be updated. + * @group: sched_group whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @load_idx: Load index of sched_domain of this_cpu for load calc. + * @sd_idle: Idle status of the sched_domain containing group. + * @local_group: Does group contain this_cpu. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sgs: variable to hold the statistics for this group. + */ +static inline void update_sg_lb_stats(struct sched_domain *sd, + struct sched_group *group, int this_cpu, + enum cpu_idle_type idle, int load_idx, int *sd_idle, + int local_group, const struct cpumask *cpus, + int *balance, struct sg_lb_stats *sgs) +{ + unsigned long load, max_cpu_load, min_cpu_load; + int i; + unsigned int balance_cpu = -1, first_idle_cpu = 0; + unsigned long sum_avg_load_per_task; + unsigned long avg_load_per_task; + + if (local_group) { + balance_cpu = group_first_cpu(group); + if (balance_cpu == this_cpu) + update_group_power(sd, this_cpu); + } + + /* Tally up the load of all CPUs in the group */ + sum_avg_load_per_task = avg_load_per_task = 0; + max_cpu_load = 0; + min_cpu_load = ~0UL; + + for_each_cpu_and(i, sched_group_cpus(group), cpus) { + struct rq *rq = cpu_rq(i); + + if (*sd_idle && rq->nr_running) + *sd_idle = 0; + + /* Bias balancing toward cpus of our domain */ + if (local_group) { + if (idle_cpu(i) && !first_idle_cpu) { + first_idle_cpu = 1; + balance_cpu = i; + } + + load = target_load(i, load_idx); + } else { + load = source_load(i, load_idx); + if (load > max_cpu_load) + max_cpu_load = load; + if (min_cpu_load > load) + min_cpu_load = load; + } + + sgs->group_load += load; + sgs->sum_nr_running += rq->nr_running; + sgs->sum_weighted_load += weighted_cpuload(i); + + sum_avg_load_per_task += cpu_avg_load_per_task(i); + } + + /* + * First idle cpu or the first cpu(busiest) in this sched group + * is eligible for doing load balancing at this and above + * domains. In the newly idle case, we will allow all the cpu's + * to do the newly idle load balance. + */ + if (idle != CPU_NEWLY_IDLE && local_group && + balance_cpu != this_cpu && balance) { + *balance = 0; + return; + } + + /* Adjust by relative CPU power of the group */ + sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; + + + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of two tasks. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / + group->cpu_power; + + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) + sgs->group_imb = 1; + + sgs->group_capacity = + DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); +} + +/** + * update_sd_lb_stats - Update sched_group's statistics for load balancing. + * @sd: sched_domain whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @sd_idle: Idle status of the sched_domain containing group. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sds: variable to hold the statistics for this sched_domain. + */ +static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, + enum cpu_idle_type idle, int *sd_idle, + const struct cpumask *cpus, int *balance, + struct sd_lb_stats *sds) +{ + struct sched_domain *child = sd->child; + struct sched_group *group = sd->groups; + struct sg_lb_stats sgs; + int load_idx, prefer_sibling = 0; + + if (child && child->flags & SD_PREFER_SIBLING) + prefer_sibling = 1; + + init_sd_power_savings_stats(sd, sds, idle); + load_idx = get_sd_load_idx(sd, idle); + + do { + int local_group; + + local_group = cpumask_test_cpu(this_cpu, + sched_group_cpus(group)); + memset(&sgs, 0, sizeof(sgs)); + update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, + local_group, cpus, balance, &sgs); + + if (local_group && balance && !(*balance)) + return; + + sds->total_load += sgs.group_load; + sds->total_pwr += group->cpu_power; + + /* + * In case the child domain prefers tasks go to siblings + * first, lower the group capacity to one so that we'll try + * and move all the excess tasks away. + */ + if (prefer_sibling) + sgs.group_capacity = min(sgs.group_capacity, 1UL); + + if (local_group) { + sds->this_load = sgs.avg_load; + sds->this = group; + sds->this_nr_running = sgs.sum_nr_running; + sds->this_load_per_task = sgs.sum_weighted_load; + } else if (sgs.avg_load > sds->max_load && + (sgs.sum_nr_running > sgs.group_capacity || + sgs.group_imb)) { + sds->max_load = sgs.avg_load; + sds->busiest = group; + sds->busiest_nr_running = sgs.sum_nr_running; + sds->busiest_load_per_task = sgs.sum_weighted_load; + sds->group_imb = sgs.group_imb; + } + + update_sd_power_savings_stats(group, sds, local_group, &sgs); + group = group->next; + } while (group != sd->groups); +} + +/** + * fix_small_imbalance - Calculate the minor imbalance that exists + * amongst the groups of a sched_domain, during + * load balancing. + * @sds: Statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: The cpu at whose sched_domain we're performing load-balance. + * @imbalance: Variable to store the imbalance. + */ +static inline void fix_small_imbalance(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + unsigned long tmp, pwr_now = 0, pwr_move = 0; + unsigned int imbn = 2; + + if (sds->this_nr_running) { + sds->this_load_per_task /= sds->this_nr_running; + if (sds->busiest_load_per_task > + sds->this_load_per_task) + imbn = 1; + } else + sds->this_load_per_task = + cpu_avg_load_per_task(this_cpu); + + if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= + sds->busiest_load_per_task * imbn) { + *imbalance = sds->busiest_load_per_task; + return; + } + + /* + * OK, we don't have enough imbalance to justify moving tasks, + * however we may be able to increase total CPU power used by + * moving them. + */ + + pwr_now += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load); + pwr_now += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load); + pwr_now /= SCHED_LOAD_SCALE; + + /* Amount of load we'd subtract */ + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->busiest->cpu_power; + if (sds->max_load > tmp) + pwr_move += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load - tmp); + + /* Amount of load we'd add */ + if (sds->max_load * sds->busiest->cpu_power < + sds->busiest_load_per_task * SCHED_LOAD_SCALE) + tmp = (sds->max_load * sds->busiest->cpu_power) / + sds->this->cpu_power; + else + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->this->cpu_power; + pwr_move += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load + tmp); + pwr_move /= SCHED_LOAD_SCALE; + + /* Move if we gain throughput */ + if (pwr_move > pwr_now) + *imbalance = sds->busiest_load_per_task; +} + +/** + * calculate_imbalance - Calculate the amount of imbalance present within the + * groups of a given sched_domain during load balance. + * @sds: statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: Cpu for which currently load balance is being performed. + * @imbalance: The variable to store the imbalance. + */ +static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, + unsigned long *imbalance) +{ + unsigned long max_pull; + /* + * In the presence of smp nice balancing, certain scenarios can have + * max load less than avg load(as we skip the groups at or below + * its cpu_power, while calculating max_load..) + */ + if (sds->max_load < sds->avg_load) { + *imbalance = 0; + return fix_small_imbalance(sds, this_cpu, imbalance); + } + + /* Don't want to pull so many tasks that a group would go idle */ + max_pull = min(sds->max_load - sds->avg_load, + sds->max_load - sds->busiest_load_per_task); + + /* How much load to actually move to equalise the imbalance */ + *imbalance = min(max_pull * sds->busiest->cpu_power, + (sds->avg_load - sds->this_load) * sds->this->cpu_power) + / SCHED_LOAD_SCALE; + + /* + * if *imbalance is less than the average load per runnable task + * there is no gaurantee that any tasks will be moved so we'll have + * a think about bumping its value to force at least one task to be + * moved + */ + if (*imbalance < sds->busiest_load_per_task) + return fix_small_imbalance(sds, this_cpu, imbalance); + +} +/******* find_busiest_group() helpers end here *********************/ + +/** + * find_busiest_group - Returns the busiest group within the sched_domain + * if there is an imbalance. If there isn't an imbalance, and + * the user has opted for power-savings, it returns a group whose + * CPUs can be put to idle by rebalancing those tasks elsewhere, if + * such a group exists. + * + * Also calculates the amount of weighted load which should be moved + * to restore balance. + * + * @sd: The sched_domain whose busiest group is to be returned. + * @this_cpu: The cpu for which load balancing is currently being performed. + * @imbalance: Variable which stores amount of weighted load which should + * be moved to restore balance/put a group to idle. + * @idle: The idle status of this_cpu. + * @sd_idle: The idleness of sd + * @cpus: The set of CPUs under consideration for load-balancing. + * @balance: Pointer to a variable indicating if this_cpu + * is the appropriate cpu to perform load balancing at this_level. + * + * Returns: - the busiest group if imbalance exists. + * - If no imbalance and user has opted for power-savings balance, + * return the least loaded group whose CPUs can be + * put to idle by rebalancing its tasks onto our group. + */ +static struct sched_group * +find_busiest_group(struct sched_domain *sd, int this_cpu, + unsigned long *imbalance, enum cpu_idle_type idle, + int *sd_idle, const struct cpumask *cpus, int *balance) +{ + struct sd_lb_stats sds; + + memset(&sds, 0, sizeof(sds)); + + /* + * Compute the various statistics relavent for load balancing at + * this level. + */ + update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, + balance, &sds); + + /* Cases where imbalance does not exist from POV of this_cpu */ + /* 1) this_cpu is not the appropriate cpu to perform load balancing + * at this level. + * 2) There is no busy sibling group to pull from. + * 3) This group is the busiest group. + * 4) This group is more busy than the avg busieness at this + * sched_domain. + * 5) The imbalance is within the specified limit. + * 6) Any rebalance would lead to ping-pong + */ + if (balance && !(*balance)) + goto ret; + + if (!sds.busiest || sds.busiest_nr_running == 0) + goto out_balanced; + + if (sds.this_load >= sds.max_load) + goto out_balanced; + + sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; + + if (sds.this_load >= sds.avg_load) + goto out_balanced; + + if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) + goto out_balanced; + + sds.busiest_load_per_task /= sds.busiest_nr_running; + if (sds.group_imb) + sds.busiest_load_per_task = + min(sds.busiest_load_per_task, sds.avg_load); + + /* + * We're trying to get all the cpus to the average_load, so we don't + * want to push ourselves above the average load, nor do we wish to + * reduce the max loaded cpu below the average load, as either of these + * actions would just result in more rebalancing later, and ping-pong + * tasks around. Thus we look for the minimum possible imbalance. + * Negative imbalances (*we* are more loaded than anyone else) will + * be counted as no imbalance for these purposes -- we can't fix that + * by pulling tasks to us. Be careful of negative numbers as they'll + * appear as very large values with unsigned longs. + */ + if (sds.max_load <= sds.busiest_load_per_task) + goto out_balanced; + + /* Looks like there is an imbalance. Compute it */ + calculate_imbalance(&sds, this_cpu, imbalance); + return sds.busiest; + +out_balanced: + /* + * There is no obvious imbalance. But check if we can do some balancing + * to save power. + */ + if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) + return sds.busiest; +ret: + *imbalance = 0; + return NULL; +} + +/* + * find_busiest_queue - find the busiest runqueue among the cpus in group. + */ +static struct rq * +find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, + unsigned long imbalance, const struct cpumask *cpus) +{ + struct rq *busiest = NULL, *rq; + unsigned long max_load = 0; + int i; + + for_each_cpu(i, sched_group_cpus(group)) { + unsigned long power = power_of(i); + unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); + unsigned long wl; + + if (!cpumask_test_cpu(i, cpus)) + continue; + + rq = cpu_rq(i); + wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; + wl /= power; + + if (capacity && rq->nr_running == 1 && wl > imbalance) + continue; + + if (wl > max_load) { + max_load = wl; + busiest = rq; + } + } + + return busiest; +} + +/* + * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but + * so long as it is large enough. + */ +#define MAX_PINNED_INTERVAL 512 + +/* Working cpumask for load_balance and load_balance_newidle. */ +static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); + +/* + * Check this_cpu to ensure it is balanced within domain. Attempt to move + * tasks if there is an imbalance. + */ +static int load_balance(int this_cpu, struct rq *this_rq, + struct sched_domain *sd, enum cpu_idle_type idle, + int *balance) +{ + int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; + struct sched_group *group; + unsigned long imbalance; + struct rq *busiest; + unsigned long flags; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); + + cpumask_copy(cpus, cpu_active_mask); + + /* + * When power savings policy is enabled for the parent domain, idle + * sibling can pick up load irrespective of busy siblings. In this case, + * let the state of idle sibling percolate up as CPU_IDLE, instead of + * portraying it as CPU_NOT_IDLE. + */ + if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + sd_idle = 1; + + schedstat_inc(sd, lb_count[idle]); + +redo: + update_shares(sd); + group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, + cpus, balance); + + if (*balance == 0) + goto out_balanced; + + if (!group) { + schedstat_inc(sd, lb_nobusyg[idle]); + goto out_balanced; + } + + busiest = find_busiest_queue(group, idle, imbalance, cpus); + if (!busiest) { + schedstat_inc(sd, lb_nobusyq[idle]); + goto out_balanced; + } + + BUG_ON(busiest == this_rq); + + schedstat_add(sd, lb_imbalance[idle], imbalance); + + ld_moved = 0; + if (busiest->nr_running > 1) { + /* + * Attempt to move tasks. If find_busiest_group has found + * an imbalance but busiest->nr_running <= 1, the group is + * still unbalanced. ld_moved simply stays zero, so it is + * correctly treated as an imbalance. + */ + local_irq_save(flags); + double_rq_lock(this_rq, busiest); + ld_moved = move_tasks(this_rq, this_cpu, busiest, + imbalance, sd, idle, &all_pinned); + double_rq_unlock(this_rq, busiest); + local_irq_restore(flags); + + /* + * some other cpu did the load balance for us. + */ + if (ld_moved && this_cpu != smp_processor_id()) + resched_cpu(this_cpu); + + /* All tasks on this runqueue were pinned by CPU affinity */ + if (unlikely(all_pinned)) { + cpumask_clear_cpu(cpu_of(busiest), cpus); + if (!cpumask_empty(cpus)) + goto redo; + goto out_balanced; + } + } + + if (!ld_moved) { + schedstat_inc(sd, lb_failed[idle]); + sd->nr_balance_failed++; + + if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { + + raw_spin_lock_irqsave(&busiest->lock, flags); + + /* don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, + &busiest->curr->cpus_allowed)) { + raw_spin_unlock_irqrestore(&busiest->lock, + flags); + all_pinned = 1; + goto out_one_pinned; + } + + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } + raw_spin_unlock_irqrestore(&busiest->lock, flags); + if (active_balance) + wake_up_process(busiest->migration_thread); + + /* + * We've kicked active balancing, reset the failure + * counter. + */ + sd->nr_balance_failed = sd->cache_nice_tries+1; + } + } else + sd->nr_balance_failed = 0; + + if (likely(!active_balance)) { + /* We were unbalanced, so reset the balancing interval */ + sd->balance_interval = sd->min_interval; + } else { + /* + * If we've begun active balancing, start to back off. This + * case may not be covered by the all_pinned logic if there + * is only 1 task on the busy runqueue (because we don't call + * move_tasks). + */ + if (sd->balance_interval < sd->max_interval) + sd->balance_interval *= 2; + } + + if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + + goto out; + +out_balanced: + schedstat_inc(sd, lb_balanced[idle]); + + sd->nr_balance_failed = 0; + +out_one_pinned: + /* tune up the balancing interval */ + if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || + (sd->balance_interval < sd->max_interval)) + sd->balance_interval *= 2; + + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + else + ld_moved = 0; +out: + if (ld_moved) + update_shares(sd); + return ld_moved; +} + +/* + * Check this_cpu to ensure it is balanced within domain. Attempt to move + * tasks if there is an imbalance. + * + * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). + * this_rq is locked. + */ +static int +load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) +{ + struct sched_group *group; + struct rq *busiest = NULL; + unsigned long imbalance; + int ld_moved = 0; + int sd_idle = 0; + int all_pinned = 0; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); + + cpumask_copy(cpus, cpu_active_mask); + + /* + * When power savings policy is enabled for the parent domain, idle + * sibling can pick up load irrespective of busy siblings. In this case, + * let the state of idle sibling percolate up as IDLE, instead of + * portraying it as CPU_NOT_IDLE. + */ + if (sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + sd_idle = 1; + + schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); +redo: + update_shares_locked(this_rq, sd); + group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, + &sd_idle, cpus, NULL); + if (!group) { + schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); + goto out_balanced; + } + + busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); + if (!busiest) { + schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); + goto out_balanced; + } + + BUG_ON(busiest == this_rq); + + schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); + + ld_moved = 0; + if (busiest->nr_running > 1) { + /* Attempt to move tasks */ + double_lock_balance(this_rq, busiest); + /* this_rq->clock is already updated */ + update_rq_clock(busiest); + ld_moved = move_tasks(this_rq, this_cpu, busiest, + imbalance, sd, CPU_NEWLY_IDLE, + &all_pinned); + double_unlock_balance(this_rq, busiest); + + if (unlikely(all_pinned)) { + cpumask_clear_cpu(cpu_of(busiest), cpus); + if (!cpumask_empty(cpus)) + goto redo; + } + } + + if (!ld_moved) { + int active_balance = 0; + + schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return -1; + + if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) + return -1; + + if (sd->nr_balance_failed++ < 2) + return -1; + + /* + * The only task running in a non-idle cpu can be moved to this + * cpu in an attempt to completely freeup the other CPU + * package. The same method used to move task in load_balance() + * have been extended for load_balance_newidle() to speedup + * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) + * + * The package power saving logic comes from + * find_busiest_group(). If there are no imbalance, then + * f_b_g() will return NULL. However when sched_mc={1,2} then + * f_b_g() will select a group from which a running task may be + * pulled to this cpu in order to make the other package idle. + * If there is no opportunity to make a package idle and if + * there are no imbalance, then f_b_g() will return NULL and no + * action will be taken in load_balance_newidle(). + * + * Under normal task pull operation due to imbalance, there + * will be more than one task in the source run queue and + * move_tasks() will succeed. ld_moved will be true and this + * active balance code will not be triggered. + */ + + /* Lock busiest in correct order while this_rq is held */ + double_lock_balance(this_rq, busiest); + + /* + * don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { + double_unlock_balance(this_rq, busiest); + all_pinned = 1; + return ld_moved; + } + + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } + + double_unlock_balance(this_rq, busiest); + /* + * Should not call ttwu while holding a rq->lock + */ + raw_spin_unlock(&this_rq->lock); + if (active_balance) + wake_up_process(busiest->migration_thread); + raw_spin_lock(&this_rq->lock); + + } else + sd->nr_balance_failed = 0; + + update_shares_locked(this_rq, sd); + return ld_moved; + +out_balanced: + schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return -1; + sd->nr_balance_failed = 0; + + return 0; +} + +/* + * idle_balance is called by schedule() if this_cpu is about to become + * idle. Attempts to pull tasks from other CPUs. + */ +static void idle_balance(int this_cpu, struct rq *this_rq) +{ + struct sched_domain *sd; + int pulled_task = 0; + unsigned long next_balance = jiffies + HZ; + + this_rq->idle_stamp = this_rq->clock; + + if (this_rq->avg_idle < sysctl_sched_migration_cost) + return; + + for_each_domain(this_cpu, sd) { + unsigned long interval; + + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + if (sd->flags & SD_BALANCE_NEWIDLE) + /* If we've pulled tasks over stop searching: */ + pulled_task = load_balance_newidle(this_cpu, this_rq, + sd); + + interval = msecs_to_jiffies(sd->balance_interval); + if (time_after(next_balance, sd->last_balance + interval)) + next_balance = sd->last_balance + interval; + if (pulled_task) { + this_rq->idle_stamp = 0; + break; + } + } + if (pulled_task || time_after(jiffies, this_rq->next_balance)) { + /* + * We are going idle. next_balance may be set based on + * a busy processor. So reset next_balance. + */ + this_rq->next_balance = next_balance; + } +} + +/* + * active_load_balance is run by migration threads. It pushes running tasks + * off the busiest CPU onto idle CPUs. It requires at least 1 task to be + * running on each physical CPU where possible, and avoids physical / + * logical imbalances. + * + * Called with busiest_rq locked. + */ +static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) +{ + int target_cpu = busiest_rq->push_cpu; + struct sched_domain *sd; + struct rq *target_rq; + + /* Is there any task to move? */ + if (busiest_rq->nr_running <= 1) + return; + + target_rq = cpu_rq(target_cpu); + + /* + * This condition is "impossible", if it occurs + * we need to fix it. Originally reported by + * Bjorn Helgaas on a 128-cpu setup. + */ + BUG_ON(busiest_rq == target_rq); + + /* move a task from busiest_rq to target_rq */ + double_lock_balance(busiest_rq, target_rq); + update_rq_clock(busiest_rq); + update_rq_clock(target_rq); + + /* Search for an sd spanning us and the target CPU. */ + for_each_domain(target_cpu, sd) { + if ((sd->flags & SD_LOAD_BALANCE) && + cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) + break; + } + + if (likely(sd)) { + schedstat_inc(sd, alb_count); + + if (move_one_task(target_rq, target_cpu, busiest_rq, + sd, CPU_IDLE)) + schedstat_inc(sd, alb_pushed); + else + schedstat_inc(sd, alb_failed); + } + double_unlock_balance(busiest_rq, target_rq); +} + +#ifdef CONFIG_NO_HZ +static struct { + atomic_t load_balancer; + cpumask_var_t cpu_mask; + cpumask_var_t ilb_grp_nohz_mask; +} nohz ____cacheline_aligned = { + .load_balancer = ATOMIC_INIT(-1), +}; + +int get_nohz_load_balancer(void) +{ + return atomic_read(&nohz.load_balancer); +} + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * lowest_flag_domain - Return lowest sched_domain containing flag. + * @cpu: The cpu whose lowest level of sched domain is to + * be returned. + * @flag: The flag to check for the lowest sched_domain + * for the given cpu. + * + * Returns the lowest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) + if (sd && (sd->flags & flag)) + break; + + return sd; +} + +/** + * for_each_flag_domain - Iterates over sched_domains containing the flag. + * @cpu: The cpu whose domains we're iterating over. + * @sd: variable holding the value of the power_savings_sd + * for cpu. + * @flag: The flag to filter the sched_domains to be iterated. + * + * Iterates over all the scheduler domains for a given cpu that has the 'flag' + * set, starting from the lowest sched_domain to the highest. + */ +#define for_each_flag_domain(cpu, sd, flag) \ + for (sd = lowest_flag_domain(cpu, flag); \ + (sd && (sd->flags & flag)); sd = sd->parent) + +/** + * is_semi_idle_group - Checks if the given sched_group is semi-idle. + * @ilb_group: group to be checked for semi-idleness + * + * Returns: 1 if the group is semi-idle. 0 otherwise. + * + * We define a sched_group to be semi idle if it has atleast one idle-CPU + * and atleast one non-idle CPU. This helper function checks if the given + * sched_group is semi-idle or not. + */ +static inline int is_semi_idle_group(struct sched_group *ilb_group) +{ + cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, + sched_group_cpus(ilb_group)); + + /* + * A sched_group is semi-idle when it has atleast one busy cpu + * and atleast one idle cpu. + */ + if (cpumask_empty(nohz.ilb_grp_nohz_mask)) + return 0; + + if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) + return 0; + + return 1; +} +/** + * find_new_ilb - Finds the optimum idle load balancer for nomination. + * @cpu: The cpu which is nominating a new idle_load_balancer. + * + * Returns: Returns the id of the idle load balancer if it exists, + * Else, returns >= nr_cpu_ids. + * + * This algorithm picks the idle load balancer such that it belongs to a + * semi-idle powersavings sched_domain. The idea is to try and avoid + * completely idle packages/cores just for the purpose of idle load balancing + * when there are other idle cpu's which are better suited for that job. + */ +static int find_new_ilb(int cpu) +{ + struct sched_domain *sd; + struct sched_group *ilb_group; + + /* + * Have idle load balancer selection from semi-idle packages only + * when power-aware load balancing is enabled + */ + if (!(sched_smt_power_savings || sched_mc_power_savings)) + goto out_done; + + /* + * Optimize for the case when we have no idle CPUs or only one + * idle CPU. Don't walk the sched_domain hierarchy in such cases + */ + if (cpumask_weight(nohz.cpu_mask) < 2) + goto out_done; + + for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { + ilb_group = sd->groups; + + do { + if (is_semi_idle_group(ilb_group)) + return cpumask_first(nohz.ilb_grp_nohz_mask); + + ilb_group = ilb_group->next; + + } while (ilb_group != sd->groups); + } + +out_done: + return cpumask_first(nohz.cpu_mask); +} +#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ +static inline int find_new_ilb(int call_cpu) +{ + return cpumask_first(nohz.cpu_mask); +} +#endif + +/* + * This routine will try to nominate the ilb (idle load balancing) + * owner among the cpus whose ticks are stopped. ilb owner will do the idle + * load balancing on behalf of all those cpus. If all the cpus in the system + * go into this tickless mode, then there will be no ilb owner (as there is + * no need for one) and all the cpus will sleep till the next wakeup event + * arrives... + * + * For the ilb owner, tick is not stopped. And this tick will be used + * for idle load balancing. ilb owner will still be part of + * nohz.cpu_mask.. + * + * While stopping the tick, this cpu will become the ilb owner if there + * is no other owner. And will be the owner till that cpu becomes busy + * or if all cpus in the system stop their ticks at which point + * there is no need for ilb owner. + * + * When the ilb owner becomes busy, it nominates another owner, during the + * next busy scheduler_tick() + */ +int select_nohz_load_balancer(int stop_tick) +{ + int cpu = smp_processor_id(); + + if (stop_tick) { + cpu_rq(cpu)->in_nohz_recently = 1; + + if (!cpu_active(cpu)) { + if (atomic_read(&nohz.load_balancer) != cpu) + return 0; + + /* + * If we are going offline and still the leader, + * give up! + */ + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + + return 0; + } + + cpumask_set_cpu(cpu, nohz.cpu_mask); + + /* time for ilb owner also to sleep */ + if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { + if (atomic_read(&nohz.load_balancer) == cpu) + atomic_set(&nohz.load_balancer, -1); + return 0; + } + + if (atomic_read(&nohz.load_balancer) == -1) { + /* make me the ilb owner */ + if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) + return 1; + } else if (atomic_read(&nohz.load_balancer) == cpu) { + int new_ilb; + + if (!(sched_smt_power_savings || + sched_mc_power_savings)) + return 1; + /* + * Check to see if there is a more power-efficient + * ilb. + */ + new_ilb = find_new_ilb(cpu); + if (new_ilb < nr_cpu_ids && new_ilb != cpu) { + atomic_set(&nohz.load_balancer, -1); + resched_cpu(new_ilb); + return 0; + } + return 1; + } + } else { + if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) + return 0; + + cpumask_clear_cpu(cpu, nohz.cpu_mask); + + if (atomic_read(&nohz.load_balancer) == cpu) + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + } + return 0; +} +#endif + +static DEFINE_SPINLOCK(balancing); + +/* + * It checks each scheduling domain to see if it is due to be balanced, + * and initiates a balancing operation if so. + * + * Balancing parameters are set up in arch_init_sched_domains. + */ +static void rebalance_domains(int cpu, enum cpu_idle_type idle) +{ + int balance = 1; + struct rq *rq = cpu_rq(cpu); + unsigned long interval; + struct sched_domain *sd; + /* Earliest time when we have to do rebalance again */ + unsigned long next_balance = jiffies + 60*HZ; + int update_next_balance = 0; + int need_serialize; + + for_each_domain(cpu, sd) { + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + interval = sd->balance_interval; + if (idle != CPU_IDLE) + interval *= sd->busy_factor; + + /* scale ms to jiffies */ + interval = msecs_to_jiffies(interval); + if (unlikely(!interval)) + interval = 1; + if (interval > HZ*NR_CPUS/10) + interval = HZ*NR_CPUS/10; + + need_serialize = sd->flags & SD_SERIALIZE; + + if (need_serialize) { + if (!spin_trylock(&balancing)) + goto out; + } + + if (time_after_eq(jiffies, sd->last_balance + interval)) { + if (load_balance(cpu, rq, sd, idle, &balance)) { + /* + * We've pulled tasks over so either we're no + * longer idle, or one of our SMT siblings is + * not idle. + */ + idle = CPU_NOT_IDLE; + } + sd->last_balance = jiffies; + } + if (need_serialize) + spin_unlock(&balancing); +out: + if (time_after(next_balance, sd->last_balance + interval)) { + next_balance = sd->last_balance + interval; + update_next_balance = 1; + } + + /* + * Stop the load balance at this level. There is another + * CPU in our sched group which is doing load balancing more + * actively. + */ + if (!balance) + break; + } + + /* + * next_balance will be updated only when there is a need. + * When the cpu is attached to null domain for ex, it will not be + * updated. + */ + if (likely(update_next_balance)) + rq->next_balance = next_balance; +} + +/* + * run_rebalance_domains is triggered when needed from the scheduler tick. + * In CONFIG_NO_HZ case, the idle load balance owner will do the + * rebalancing for all the cpus for whom scheduler ticks are stopped. + */ +static void run_rebalance_domains(struct softirq_action *h) +{ + int this_cpu = smp_processor_id(); + struct rq *this_rq = cpu_rq(this_cpu); + enum cpu_idle_type idle = this_rq->idle_at_tick ? + CPU_IDLE : CPU_NOT_IDLE; + + rebalance_domains(this_cpu, idle); + +#ifdef CONFIG_NO_HZ + /* + * If this cpu is the owner for idle load balancing, then do the + * balancing on behalf of the other idle cpus whose ticks are + * stopped. + */ + if (this_rq->idle_at_tick && + atomic_read(&nohz.load_balancer) == this_cpu) { + struct rq *rq; + int balance_cpu; + + for_each_cpu(balance_cpu, nohz.cpu_mask) { + if (balance_cpu == this_cpu) + continue; + + /* + * If this cpu gets work to do, stop the load balancing + * work being done for other cpus. Next load + * balancing owner will pick it up. + */ + if (need_resched()) + break; + + rebalance_domains(balance_cpu, CPU_IDLE); + + rq = cpu_rq(balance_cpu); + if (time_after(this_rq->next_balance, rq->next_balance)) + this_rq->next_balance = rq->next_balance; + } + } +#endif +} + +static inline int on_null_domain(int cpu) +{ + return !rcu_dereference(cpu_rq(cpu)->sd); +} + +/* + * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. + * + * In case of CONFIG_NO_HZ, this is the place where we nominate a new + * idle load balancing owner or decide to stop the periodic load balancing, + * if the whole system is idle. + */ +static inline void trigger_load_balance(struct rq *rq, int cpu) +{ +#ifdef CONFIG_NO_HZ + /* + * If we were in the nohz mode recently and busy at the current + * scheduler tick, then check if we need to nominate new idle + * load balancer. + */ + if (rq->in_nohz_recently && !rq->idle_at_tick) { + rq->in_nohz_recently = 0; + + if (atomic_read(&nohz.load_balancer) == cpu) { + cpumask_clear_cpu(cpu, nohz.cpu_mask); + atomic_set(&nohz.load_balancer, -1); + } + + if (atomic_read(&nohz.load_balancer) == -1) { + int ilb = find_new_ilb(cpu); + + if (ilb < nr_cpu_ids) + resched_cpu(ilb); + } + } + + /* + * If this cpu is idle and doing idle load balancing for all the + * cpus with ticks stopped, is it time for that to stop? + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && + cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { + resched_cpu(cpu); + return; + } + + /* + * If this cpu is idle and the idle load balancing is done by + * someone else, then no need raise the SCHED_SOFTIRQ + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && + cpumask_test_cpu(cpu, nohz.cpu_mask)) + return; +#endif + /* Don't need to rebalance while attached to NULL domain */ + if (time_after_eq(jiffies, rq->next_balance) && + likely(!on_null_domain(cpu))) + raise_softirq(SCHED_SOFTIRQ); +} + static void rq_online_fair(struct rq *rq) { update_sysctl(); @@ -1962,6 +3718,15 @@ static void rq_offline_fair(struct rq *rq) update_sysctl(); } +#else /* CONFIG_SMP */ + +/* + * on UP we do not need to balance between CPUs: + */ +static inline void idle_balance(int cpu, struct rq *rq) +{ +} + #endif /* CONFIG_SMP */ /* -- cgit v1.2.2 From 3d45fd804a95055ecab5b3eed81f5ab2dbb047a2 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:12:46 +0100 Subject: sched: Remove the sched_class load_balance methods Take out the sched_class methods for load-balancing. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched.c | 26 ------------------- kernel/sched_fair.c | 66 +++++++++++++++++++++++++++---------------------- kernel/sched_idletask.c | 21 ---------------- kernel/sched_rt.c | 20 --------------- 4 files changed, 37 insertions(+), 96 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 13a2acf18b2d..c0be07932a8d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1390,32 +1390,6 @@ static const u32 prio_to_wmult[40] = { /* 15 */ 119304647, 148102320, 186737708, 238609294, 286331153, }; -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup); - -/* - * runqueue iterator, to support SMP load-balancing between different - * scheduling classes, without having to expose their internal data - * structures to the load-balancing proper: - */ -struct rq_iterator { - void *arg; - struct task_struct *(*start)(void *); - struct task_struct *(*next)(void *); -}; - -#ifdef CONFIG_SMP -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator); - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator); -#endif - /* Time spent by the tasks of the cpu accounting group executing in ... */ enum cpuacct_stat_index { CPUACCT_STAT_USER, /* ... user mode */ diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 5116b81d7727..faf9a2f099ab 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1851,6 +1851,24 @@ static struct task_struct *load_balance_next_fair(void *arg) return __load_balance_iterator(cfs_rq, cfs_rq->balance_iterator); } +/* + * runqueue iterator, to support SMP load-balancing between different + * scheduling classes, without having to expose their internal data + * structures to the load-balancing proper: + */ +struct rq_iterator { + void *arg; + struct task_struct *(*start)(void *); + struct task_struct *(*next)(void *); +}; + +static unsigned long +balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, struct sched_domain *sd, + enum cpu_idle_type idle, int *all_pinned, + int *this_best_prio, struct rq_iterator *iterator); + + static unsigned long __load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -1929,8 +1947,20 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, #endif static int -move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) +iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle, + struct rq_iterator *iterator); + +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int +move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) { struct cfs_rq *busy_cfs_rq; struct rq_iterator cfs_rq_iterator; @@ -2094,16 +2124,15 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned) { - const struct sched_class *class = sched_class_highest; - unsigned long total_load_moved = 0; + unsigned long total_load_moved = 0, load_moved; int this_best_prio = this_rq->curr->prio; do { - total_load_moved += - class->load_balance(this_rq, this_cpu, busiest, + load_moved = load_balance_fair(this_rq, this_cpu, busiest, max_load_move - total_load_moved, sd, idle, all_pinned, &this_best_prio); - class = class->next; + + total_load_moved += load_moved; #ifdef CONFIG_PREEMPT /* @@ -2114,7 +2143,7 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; #endif - } while (class && max_load_move > total_load_moved); + } while (load_moved && max_load_move > total_load_moved); return total_load_moved > 0; } @@ -2145,25 +2174,6 @@ iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, return 0; } -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - const struct sched_class *class; - - for_each_class(class) { - if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) - return 1; - } - - return 0; -} /********** Helpers for find_busiest_group ************************/ /* * sd_lb_stats - Structure to store the statistics of a sched_domain @@ -3873,8 +3883,6 @@ static const struct sched_class fair_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_fair, - .load_balance = load_balance_fair, - .move_one_task = move_one_task_fair, .rq_online = rq_online_fair, .rq_offline = rq_offline_fair, diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 01332bfc61a7..a8a6d8a50947 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -44,24 +44,6 @@ static void put_prev_task_idle(struct rq *rq, struct task_struct *prev) { } -#ifdef CONFIG_SMP -static unsigned long -load_balance_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - return 0; -} - -static int -move_one_task_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - return 0; -} -#endif - static void task_tick_idle(struct rq *rq, struct task_struct *curr, int queued) { } @@ -119,9 +101,6 @@ static const struct sched_class idle_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_idle, - - .load_balance = load_balance_idle, - .move_one_task = move_one_task_idle, #endif .set_curr_task = set_curr_task_idle, diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 072b3fcee8d8..502bb614e40a 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1481,24 +1481,6 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) push_rt_tasks(rq); } -static unsigned long -load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - /* don't touch RT tasks */ - return 0; -} - -static int -move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - /* don't touch RT tasks */ - return 0; -} - static void set_cpus_allowed_rt(struct task_struct *p, const struct cpumask *new_mask) { @@ -1746,8 +1728,6 @@ static const struct sched_class rt_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_rt, - .load_balance = load_balance_rt, - .move_one_task = move_one_task_rt, .set_cpus_allowed = set_cpus_allowed_rt, .rq_online = rq_online_rt, .rq_offline = rq_offline_rt, -- cgit v1.2.2 From ee00e66ffff250fb0d3a789e5565462f71c7c9a7 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:25:20 +0100 Subject: sched: Remove rq_iterator usage from load_balance_fair Since we only ever iterate the fair class, do away with this abstraction. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 80 +++++++++++++++++++---------------------------------- 1 file changed, 29 insertions(+), 51 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index faf9a2f099ab..709deb33708a 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1866,26 +1866,9 @@ static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator); + int *this_best_prio, struct cfs_rq *busiest_cfs_rq); -static unsigned long -__load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, int *this_best_prio, - struct cfs_rq *cfs_rq) -{ - struct rq_iterator cfs_rq_iterator; - - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; - cfs_rq_iterator.arg = cfs_rq; - - return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, - this_best_prio, &cfs_rq_iterator); -} - #ifdef CONFIG_FAIR_GROUP_SCHED static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, @@ -1915,9 +1898,9 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, rem_load = (u64)rem_load_move * busiest_weight; rem_load = div_u64(rem_load, busiest_h_load + 1); - moved_load = __load_balance_fair(this_rq, this_cpu, busiest, + moved_load = balance_tasks(this_rq, this_cpu, busiest, rem_load, sd, idle, all_pinned, this_best_prio, - tg->cfs_rq[busiest_cpu]); + busiest_cfs_rq); if (!moved_load) continue; @@ -1940,7 +1923,7 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio) { - return __load_balance_fair(this_rq, this_cpu, busiest, + return balance_tasks(this_rq, this_cpu, busiest, max_load_move, sd, idle, all_pinned, this_best_prio, &busiest->cfs); } @@ -2050,53 +2033,48 @@ static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator) + int *this_best_prio, struct cfs_rq *busiest_cfs_rq) { int loops = 0, pulled = 0, pinned = 0; - struct task_struct *p; long rem_load_move = max_load_move; + struct task_struct *p, *n; if (max_load_move == 0) goto out; pinned = 1; - /* - * Start the load-balancing iterator: - */ - p = iterator->start(iterator->arg); -next: - if (!p || loops++ > sysctl_sched_nr_migrate) - goto out; + list_for_each_entry_safe(p, n, &busiest_cfs_rq->tasks, se.group_node) { + if (loops++ > sysctl_sched_nr_migrate) + break; - if ((p->se.load.weight >> 1) > rem_load_move || - !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - p = iterator->next(iterator->arg); - goto next; - } + if ((p->se.load.weight >> 1) > rem_load_move || + !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) + continue; - pull_task(busiest, p, this_rq, this_cpu); - pulled++; - rem_load_move -= p->se.load.weight; + pull_task(busiest, p, this_rq, this_cpu); + pulled++; + rem_load_move -= p->se.load.weight; #ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible kernels - * will stop after the first task is pulled to minimize the critical - * section. - */ - if (idle == CPU_NEWLY_IDLE) - goto out; + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ + if (idle == CPU_NEWLY_IDLE) + break; #endif - /* - * We only want to steal up to the prescribed amount of weighted load. - */ - if (rem_load_move > 0) { + /* + * We only want to steal up to the prescribed amount of + * weighted load. + */ + if (rem_load_move <= 0) + break; + if (p->prio < *this_best_prio) *this_best_prio = p->prio; - p = iterator->next(iterator->arg); - goto next; } out: /* -- cgit v1.2.2 From 897c395f4c94ae19302f92393a0b8304e414ee06 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:45:42 +0100 Subject: sched: Remove rq_iterator from move_one_task Again, since we only iterate the fair class, remove the abstraction. Since this is the last user of the rq_iterator, remove all that too. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 146 +++++++++++++--------------------------------------- 1 file changed, 36 insertions(+), 110 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 709deb33708a..e48e459da98d 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1814,54 +1814,6 @@ static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) * Fair scheduling class load-balancing methods: */ -/* - * Load-balancing iterator. Note: while the runqueue stays locked - * during the whole iteration, the current task might be - * dequeued so the iterator has to be dequeue-safe. Here we - * achieve that by always pre-iterating before returning - * the current task: - */ -static struct task_struct * -__load_balance_iterator(struct cfs_rq *cfs_rq, struct list_head *next) -{ - struct task_struct *p = NULL; - struct sched_entity *se; - - if (next == &cfs_rq->tasks) - return NULL; - - se = list_entry(next, struct sched_entity, group_node); - p = task_of(se); - cfs_rq->balance_iterator = next->next; - - return p; -} - -static struct task_struct *load_balance_start_fair(void *arg) -{ - struct cfs_rq *cfs_rq = arg; - - return __load_balance_iterator(cfs_rq, cfs_rq->tasks.next); -} - -static struct task_struct *load_balance_next_fair(void *arg) -{ - struct cfs_rq *cfs_rq = arg; - - return __load_balance_iterator(cfs_rq, cfs_rq->balance_iterator); -} - -/* - * runqueue iterator, to support SMP load-balancing between different - * scheduling classes, without having to expose their internal data - * structures to the load-balancing proper: - */ -struct rq_iterator { - void *arg; - struct task_struct *(*start)(void *); - struct task_struct *(*next)(void *); -}; - static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -1929,42 +1881,6 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, } #endif -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator); - -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int -move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - struct cfs_rq *busy_cfs_rq; - struct rq_iterator cfs_rq_iterator; - - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; - - for_each_leaf_cfs_rq(busiest, busy_cfs_rq) { - /* - * pass busy_cfs_rq argument into - * load_balance_[start|next]_fair iterators - */ - cfs_rq_iterator.arg = busy_cfs_rq; - if (iter_move_one_task(this_rq, this_cpu, busiest, sd, idle, - &cfs_rq_iterator)) - return 1; - } - - return 0; -} - /* * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. @@ -2029,6 +1945,42 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, return 1; } +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int +move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) +{ + struct task_struct *p, *n; + struct cfs_rq *cfs_rq; + int pinned = 0; + + for_each_leaf_cfs_rq(busiest, cfs_rq) { + list_for_each_entry_safe(p, n, &cfs_rq->tasks, se.group_node) { + + if (!can_migrate_task(p, busiest, this_cpu, + sd, idle, &pinned)) + continue; + + pull_task(busiest, p, this_rq, this_cpu); + /* + * Right now, this is only the second place pull_task() + * is called, so we can safely collect pull_task() + * stats here rather than inside pull_task(). + */ + schedstat_inc(sd, lb_gained[idle]); + return 1; + } + } + + return 0; +} + static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -2126,32 +2078,6 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, return total_load_moved > 0; } -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator) -{ - struct task_struct *p = iterator->start(iterator->arg); - int pinned = 0; - - while (p) { - if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - pull_task(busiest, p, this_rq, this_cpu); - /* - * Right now, this is only the second place pull_task() - * is called, so we can safely collect pull_task() - * stats here rather than inside pull_task(). - */ - schedstat_inc(sd, lb_gained[idle]); - - return 1; - } - p = iterator->next(iterator->arg); - } - - return 0; -} - /********** Helpers for find_busiest_group ************************/ /* * sd_lb_stats - Structure to store the statistics of a sched_domain -- cgit v1.2.2 From 230059de77a4e0f6afba98073e73bc9fd471506e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:47:12 +0100 Subject: sched: Remove from fwd decls Move code around to get rid of fwd declarations. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 127 +++++++++++++++++++++++++--------------------------- 1 file changed, 60 insertions(+), 67 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index e48e459da98d..93fccbadde23 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1814,73 +1814,6 @@ static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) * Fair scheduling class load-balancing methods: */ -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct cfs_rq *busiest_cfs_rq); - - -#ifdef CONFIG_FAIR_GROUP_SCHED -static unsigned long -load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - long rem_load_move = max_load_move; - int busiest_cpu = cpu_of(busiest); - struct task_group *tg; - - rcu_read_lock(); - update_h_load(busiest_cpu); - - list_for_each_entry_rcu(tg, &task_groups, list) { - struct cfs_rq *busiest_cfs_rq = tg->cfs_rq[busiest_cpu]; - unsigned long busiest_h_load = busiest_cfs_rq->h_load; - unsigned long busiest_weight = busiest_cfs_rq->load.weight; - u64 rem_load, moved_load; - - /* - * empty group - */ - if (!busiest_cfs_rq->task_weight) - continue; - - rem_load = (u64)rem_load_move * busiest_weight; - rem_load = div_u64(rem_load, busiest_h_load + 1); - - moved_load = balance_tasks(this_rq, this_cpu, busiest, - rem_load, sd, idle, all_pinned, this_best_prio, - busiest_cfs_rq); - - if (!moved_load) - continue; - - moved_load *= busiest_h_load; - moved_load = div_u64(moved_load, busiest_weight + 1); - - rem_load_move -= moved_load; - if (rem_load_move < 0) - break; - } - rcu_read_unlock(); - - return max_load_move - rem_load_move; -} -#else -static unsigned long -load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, - this_best_prio, &busiest->cfs); -} -#endif - /* * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. @@ -2042,6 +1975,66 @@ out: return max_load_move - rem_load_move; } +#ifdef CONFIG_FAIR_GROUP_SCHED +static unsigned long +load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned, int *this_best_prio) +{ + long rem_load_move = max_load_move; + int busiest_cpu = cpu_of(busiest); + struct task_group *tg; + + rcu_read_lock(); + update_h_load(busiest_cpu); + + list_for_each_entry_rcu(tg, &task_groups, list) { + struct cfs_rq *busiest_cfs_rq = tg->cfs_rq[busiest_cpu]; + unsigned long busiest_h_load = busiest_cfs_rq->h_load; + unsigned long busiest_weight = busiest_cfs_rq->load.weight; + u64 rem_load, moved_load; + + /* + * empty group + */ + if (!busiest_cfs_rq->task_weight) + continue; + + rem_load = (u64)rem_load_move * busiest_weight; + rem_load = div_u64(rem_load, busiest_h_load + 1); + + moved_load = balance_tasks(this_rq, this_cpu, busiest, + rem_load, sd, idle, all_pinned, this_best_prio, + busiest_cfs_rq); + + if (!moved_load) + continue; + + moved_load *= busiest_h_load; + moved_load = div_u64(moved_load, busiest_weight + 1); + + rem_load_move -= moved_load; + if (rem_load_move < 0) + break; + } + rcu_read_unlock(); + + return max_load_move - rem_load_move; +} +#else +static unsigned long +load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned, int *this_best_prio) +{ + return balance_tasks(this_rq, this_cpu, busiest, + max_load_move, sd, idle, all_pinned, + this_best_prio, &busiest->cfs); +} +#endif + /* * move_tasks tries to move up to max_load_move weighted load from busiest to * this_rq, as part of a balancing operation within domain "sd". -- cgit v1.2.2 From baa8c1102f0cd86e69c1497d61d2ee177e663663 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 18:10:09 +0100 Subject: sched: Add a lock break for PREEMPT=y Since load-balancing can hold rq->locks for quite a long while, allow breaking out early when there is lock contention. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 93fccbadde23..65d08207e925 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2065,6 +2065,10 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, */ if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; + + if (raw_spin_is_contended(&this_rq->lock) || + raw_spin_is_contended(&busiest->lock)) + break; #endif } while (load_moved && max_load_move > total_load_moved); -- cgit v1.2.2 From 1af3ed3ddf27499c3f57662c4c29871e2b95e5f9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Dec 2009 15:10:31 +0100 Subject: sched: Unify load_balance{,_newidle}() load_balance() and load_balance_newidle() look remarkably similar, one key point they differ in is the condition on when to active balance. So split out that logic into a separate function. One side effect is that previously load_balance_newidle() used to fail and return -1 under these conditions, whereas now it doesn't. I've not yet fully figured out the whole -1 return case for either load_balance{,_newidle}(). Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 115 +++++++++++++++++++++++++++------------------------- 1 file changed, 59 insertions(+), 56 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 65d08207e925..10408323794e 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2816,6 +2816,39 @@ find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, /* Working cpumask for load_balance and load_balance_newidle. */ static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); +static int need_active_balance(struct sched_domain *sd, int sd_idle, int idle) +{ + if (idle == CPU_NEWLY_IDLE) { + /* + * The only task running in a non-idle cpu can be moved to this + * cpu in an attempt to completely freeup the other CPU + * package. + * + * The package power saving logic comes from + * find_busiest_group(). If there are no imbalance, then + * f_b_g() will return NULL. However when sched_mc={1,2} then + * f_b_g() will select a group from which a running task may be + * pulled to this cpu in order to make the other package idle. + * If there is no opportunity to make a package idle and if + * there are no imbalance, then f_b_g() will return NULL and no + * action will be taken in load_balance_newidle(). + * + * Under normal task pull operation due to imbalance, there + * will be more than one task in the source run queue and + * move_tasks() will succeed. ld_moved will be true and this + * active balance code will not be triggered. + */ + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return 0; + + if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) + return 0; + } + + return unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2); +} + /* * Check this_cpu to ensure it is balanced within domain. Attempt to move * tasks if there is an imbalance. @@ -2902,8 +2935,7 @@ redo: schedstat_inc(sd, lb_failed[idle]); sd->nr_balance_failed++; - if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { - + if (need_active_balance(sd, sd_idle, idle)) { raw_spin_lock_irqsave(&busiest->lock, flags); /* don't kick the migration_thread, if the curr @@ -3049,66 +3081,37 @@ redo: int active_balance = 0; schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - - if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) - return -1; + sd->nr_balance_failed++; - if (sd->nr_balance_failed++ < 2) - return -1; + if (need_active_balance(sd, sd_idle, CPU_NEWLY_IDLE)) { + double_lock_balance(this_rq, busiest); - /* - * The only task running in a non-idle cpu can be moved to this - * cpu in an attempt to completely freeup the other CPU - * package. The same method used to move task in load_balance() - * have been extended for load_balance_newidle() to speedup - * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) - * - * The package power saving logic comes from - * find_busiest_group(). If there are no imbalance, then - * f_b_g() will return NULL. However when sched_mc={1,2} then - * f_b_g() will select a group from which a running task may be - * pulled to this cpu in order to make the other package idle. - * If there is no opportunity to make a package idle and if - * there are no imbalance, then f_b_g() will return NULL and no - * action will be taken in load_balance_newidle(). - * - * Under normal task pull operation due to imbalance, there - * will be more than one task in the source run queue and - * move_tasks() will succeed. ld_moved will be true and this - * active balance code will not be triggered. - */ + /* + * don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, + &busiest->curr->cpus_allowed)) { + double_unlock_balance(this_rq, busiest); + all_pinned = 1; + return ld_moved; + } - /* Lock busiest in correct order while this_rq is held */ - double_lock_balance(this_rq, busiest); + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; + /* + * Should not call ttwu while holding a rq->lock + */ + raw_spin_unlock(&this_rq->lock); + if (active_balance) + wake_up_process(busiest->migration_thread); + raw_spin_lock(&this_rq->lock); } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - } else sd->nr_balance_failed = 0; -- cgit v1.2.2 From f492e12ef050e02bf0185b6b57874992591b9be1 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Dec 2009 15:29:42 +0100 Subject: sched: Remove load_balance_newidle() The two functions: load_balance{,_newidle}() are very similar, with the following differences: - rq->lock usage - sb->balance_interval updates - *balance check So remove the load_balance_newidle() call with load_balance(.idle = CPU_NEWLY_IDLE), explicitly unlock the rq->lock before calling (would be done by double_lock_balance() anyway), and ignore the other differences for now. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 135 +++++----------------------------------------------- 1 file changed, 13 insertions(+), 122 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 10408323794e..de5ab1239e04 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -3009,125 +3009,6 @@ out: return ld_moved; } -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - * - * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). - * this_rq is locked. - */ -static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) -{ - struct sched_group *group; - struct rq *busiest = NULL; - unsigned long imbalance; - int ld_moved = 0; - int sd_idle = 0; - int all_pinned = 0; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); -redo: - update_shares_locked(this_rq, sd); - group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, - &sd_idle, cpus, NULL); - if (!group) { - schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* Attempt to move tasks */ - double_lock_balance(this_rq, busiest); - /* this_rq->clock is already updated */ - update_rq_clock(busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, CPU_NEWLY_IDLE, - &all_pinned); - double_unlock_balance(this_rq, busiest); - - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - } - } - - if (!ld_moved) { - int active_balance = 0; - - schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - sd->nr_balance_failed++; - - if (need_active_balance(sd, sd_idle, CPU_NEWLY_IDLE)) { - double_lock_balance(this_rq, busiest); - - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, - &busiest->curr->cpus_allowed)) { - double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - } - } else - sd->nr_balance_failed = 0; - - update_shares_locked(this_rq, sd); - return ld_moved; - -out_balanced: - schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - sd->nr_balance_failed = 0; - - return 0; -} - /* * idle_balance is called by schedule() if this_cpu is about to become * idle. Attempts to pull tasks from other CPUs. @@ -3143,16 +3024,23 @@ static void idle_balance(int this_cpu, struct rq *this_rq) if (this_rq->avg_idle < sysctl_sched_migration_cost) return; + /* + * Drop the rq->lock, but keep IRQ/preempt disabled. + */ + raw_spin_unlock(&this_rq->lock); + for_each_domain(this_cpu, sd) { unsigned long interval; + int balance = 1; if (!(sd->flags & SD_LOAD_BALANCE)) continue; - if (sd->flags & SD_BALANCE_NEWIDLE) + if (sd->flags & SD_BALANCE_NEWIDLE) { /* If we've pulled tasks over stop searching: */ - pulled_task = load_balance_newidle(this_cpu, this_rq, - sd); + pulled_task = load_balance(this_cpu, this_rq, + sd, CPU_NEWLY_IDLE, &balance); + } interval = msecs_to_jiffies(sd->balance_interval); if (time_after(next_balance, sd->last_balance + interval)) @@ -3162,6 +3050,9 @@ static void idle_balance(int this_cpu, struct rq *this_rq) break; } } + + raw_spin_lock(&this_rq->lock); + if (pulled_task || time_after(jiffies, this_rq->next_balance)) { /* * We are going idle. next_balance may be set based on -- cgit v1.2.2 From 8f190fb3f7a405682666d3723f6ec370b5afe4da Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 24 Dec 2009 14:18:21 +0100 Subject: sched: Assume *balance is valid Since all load_balance() callers will have !NULL balance parameters we can now assume so and remove a few checks. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index de5ab1239e04..0b482f5b5b3b 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2465,7 +2465,7 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, * to do the newly idle load balance. */ if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { + balance_cpu != this_cpu) { *balance = 0; return; } @@ -2528,7 +2528,7 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, local_group, cpus, balance, &sgs); - if (local_group && balance && !(*balance)) + if (local_group && !(*balance)) return; sds->total_load += sgs.group_load; @@ -2720,7 +2720,7 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, * 5) The imbalance is within the specified limit. * 6) Any rebalance would lead to ping-pong */ - if (balance && !(*balance)) + if (!(*balance)) goto ret; if (!sds.busiest || sds.busiest_nr_running == 0) -- cgit v1.2.2 From 871e35bc9733f273eaf5ceb69bbd0423b58e5285 Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Wed, 20 Jan 2010 14:02:44 -0600 Subject: sched: Fix the place where group powers are updated We want to update the sched_group_powers when balance_cpu == this_cpu. Currently the group powers are updated only if the balance_cpu is the first CPU in the local group. But balance_cpu = this_cpu could also be the first idle cpu in the group. Hence fix the place where the group powers are updated. Signed-off-by: Gautham R Shenoy Signed-off-by: Joel Schopp Signed-off-by: Peter Zijlstra LKML-Reference: <1264017764.5717.127.camel@jschopp-laptop> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 0b482f5b5b3b..22231ccb2f98 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2418,11 +2418,8 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, unsigned long sum_avg_load_per_task; unsigned long avg_load_per_task; - if (local_group) { + if (local_group) balance_cpu = group_first_cpu(group); - if (balance_cpu == this_cpu) - update_group_power(sd, this_cpu); - } /* Tally up the load of all CPUs in the group */ sum_avg_load_per_task = avg_load_per_task = 0; @@ -2470,6 +2467,8 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, return; } + update_group_power(sd, this_cpu); + /* Adjust by relative CPU power of the group */ sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; -- cgit v1.2.2 From 7c9414385ebfdd87cc542d4e7e3bb0dbb2d3ce25 Mon Sep 17 00:00:00 2001 From: Dhaval Giani Date: Wed, 20 Jan 2010 13:26:18 +0100 Subject: sched: Remove USER_SCHED Remove the USER_SCHED feature. It has been scheduled to be removed in 2.6.34 as per http://marc.info/?l=linux-kernel&m=125728479022976&w=2 Signed-off-by: Dhaval Giani Signed-off-by: Peter Zijlstra LKML-Reference: <1263990378.24844.3.camel@localhost> Signed-off-by: Ingo Molnar --- kernel/ksysfs.c | 8 -- kernel/sched.c | 114 ++------------------- kernel/sys.c | 5 - kernel/user.c | 305 -------------------------------------------------------- 4 files changed, 7 insertions(+), 425 deletions(-) (limited to 'kernel') diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c index 3feaf5a74514..6b1ccc3f0205 100644 --- a/kernel/ksysfs.c +++ b/kernel/ksysfs.c @@ -197,16 +197,8 @@ static int __init ksysfs_init(void) goto group_exit; } - /* create the /sys/kernel/uids/ directory */ - error = uids_sysfs_init(); - if (error) - goto notes_exit; - return 0; -notes_exit: - if (notes_size > 0) - sysfs_remove_bin_file(kernel_kobj, ¬es_attr); group_exit: sysfs_remove_group(kernel_kobj, &kernel_attr_group); kset_exit: diff --git a/kernel/sched.c b/kernel/sched.c index c0be07932a8d..41e76d325648 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -233,7 +233,7 @@ static void destroy_rt_bandwidth(struct rt_bandwidth *rt_b) */ static DEFINE_MUTEX(sched_domains_mutex); -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED #include @@ -243,13 +243,7 @@ static LIST_HEAD(task_groups); /* task group related information */ struct task_group { -#ifdef CONFIG_CGROUP_SCHED struct cgroup_subsys_state css; -#endif - -#ifdef CONFIG_USER_SCHED - uid_t uid; -#endif #ifdef CONFIG_FAIR_GROUP_SCHED /* schedulable entities of this group on each cpu */ @@ -274,35 +268,7 @@ struct task_group { struct list_head children; }; -#ifdef CONFIG_USER_SCHED - -/* Helper function to pass uid information to create_sched_user() */ -void set_tg_uid(struct user_struct *user) -{ - user->tg->uid = user->uid; -} - -/* - * Root task group. - * Every UID task group (including init_task_group aka UID-0) will - * be a child to this group. - */ -struct task_group root_task_group; - -#ifdef CONFIG_FAIR_GROUP_SCHED -/* Default task group's sched entity on each cpu */ -static DEFINE_PER_CPU(struct sched_entity, init_sched_entity); -/* Default task group's cfs_rq on each cpu */ -static DEFINE_PER_CPU_SHARED_ALIGNED(struct cfs_rq, init_tg_cfs_rq); -#endif /* CONFIG_FAIR_GROUP_SCHED */ - -#ifdef CONFIG_RT_GROUP_SCHED -static DEFINE_PER_CPU(struct sched_rt_entity, init_sched_rt_entity); -static DEFINE_PER_CPU_SHARED_ALIGNED(struct rt_rq, init_rt_rq_var); -#endif /* CONFIG_RT_GROUP_SCHED */ -#else /* !CONFIG_USER_SCHED */ #define root_task_group init_task_group -#endif /* CONFIG_USER_SCHED */ /* task_group_lock serializes add/remove of task groups and also changes to * a task group's cpu shares. @@ -318,11 +284,7 @@ static int root_task_group_empty(void) } #endif -#ifdef CONFIG_USER_SCHED -# define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD) -#else /* !CONFIG_USER_SCHED */ # define INIT_TASK_GROUP_LOAD NICE_0_LOAD -#endif /* CONFIG_USER_SCHED */ /* * A weight of 0 or 1 can cause arithmetics problems. @@ -348,11 +310,7 @@ static inline struct task_group *task_group(struct task_struct *p) { struct task_group *tg; -#ifdef CONFIG_USER_SCHED - rcu_read_lock(); - tg = __task_cred(p)->user->tg; - rcu_read_unlock(); -#elif defined(CONFIG_CGROUP_SCHED) +#ifdef CONFIG_CGROUP_SCHED tg = container_of(task_subsys_state(p, cpu_cgroup_subsys_id), struct task_group, css); #else @@ -383,7 +341,7 @@ static inline struct task_group *task_group(struct task_struct *p) return NULL; } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ /* CFS-related fields in a runqueue */ struct cfs_rq { @@ -7678,9 +7636,6 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED alloc_size += 2 * nr_cpu_ids * sizeof(void **); #endif -#ifdef CONFIG_USER_SCHED - alloc_size *= 2; -#endif #ifdef CONFIG_CPUMASK_OFFSTACK alloc_size += num_possible_cpus() * cpumask_size(); #endif @@ -7694,13 +7649,6 @@ void __init sched_init(void) init_task_group.cfs_rq = (struct cfs_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.se = (struct sched_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.cfs_rq = (struct cfs_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED init_task_group.rt_se = (struct sched_rt_entity **)ptr; @@ -7709,13 +7657,6 @@ void __init sched_init(void) init_task_group.rt_rq = (struct rt_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.rt_se = (struct sched_rt_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.rt_rq = (struct rt_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ #ifdef CONFIG_CPUMASK_OFFSTACK for_each_possible_cpu(i) { @@ -7735,22 +7676,13 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED init_rt_bandwidth(&init_task_group.rt_bandwidth, global_rt_period(), global_rt_runtime()); -#ifdef CONFIG_USER_SCHED - init_rt_bandwidth(&root_task_group.rt_bandwidth, - global_rt_period(), RUNTIME_INF); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED list_add(&init_task_group.list, &task_groups); INIT_LIST_HEAD(&init_task_group.children); -#ifdef CONFIG_USER_SCHED - INIT_LIST_HEAD(&root_task_group.children); - init_task_group.parent = &root_task_group; - list_add(&init_task_group.siblings, &root_task_group.children); -#endif /* CONFIG_USER_SCHED */ -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #if defined CONFIG_FAIR_GROUP_SCHED && defined CONFIG_SMP update_shares_data = __alloc_percpu(nr_cpu_ids * sizeof(unsigned long), @@ -7790,25 +7722,6 @@ void __init sched_init(void) * directly in rq->cfs (i.e init_task_group->se[] = NULL). */ init_tg_cfs_entry(&init_task_group, &rq->cfs, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - root_task_group.shares = NICE_0_LOAD; - init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, 0, NULL); - /* - * In case of task-groups formed thr' the user id of tasks, - * init_task_group represents tasks belonging to root user. - * Hence it forms a sibling of all subsequent groups formed. - * In this case, init_task_group gets only a fraction of overall - * system cpu resource, based on the weight assigned to root - * user's cpu share (INIT_TASK_GROUP_LOAD). This is accomplished - * by letting tasks of init_task_group sit in a separate cfs_rq - * (init_tg_cfs_rq) and having one entity represent this group of - * tasks in rq->cfs (i.e init_task_group->se[] != NULL). - */ - init_tg_cfs_entry(&init_task_group, - &per_cpu(init_tg_cfs_rq, i), - &per_cpu(init_sched_entity, i), i, 1, - root_task_group.se[i]); - #endif #endif /* CONFIG_FAIR_GROUP_SCHED */ @@ -7817,12 +7730,6 @@ void __init sched_init(void) INIT_LIST_HEAD(&rq->leaf_rt_rq_list); #ifdef CONFIG_CGROUP_SCHED init_tg_rt_entry(&init_task_group, &rq->rt, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - init_tg_rt_entry(&root_task_group, &rq->rt, NULL, i, 0, NULL); - init_tg_rt_entry(&init_task_group, - &per_cpu(init_rt_rq_var, i), - &per_cpu(init_sched_rt_entity, i), i, 1, - root_task_group.rt_se[i]); #endif #endif @@ -8218,7 +8125,7 @@ static inline void unregister_rt_sched_group(struct task_group *tg, int cpu) } #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED static void free_sched_group(struct task_group *tg) { free_fair_sched_group(tg); @@ -8327,7 +8234,7 @@ void sched_move_task(struct task_struct *tsk) task_rq_unlock(rq, &flags); } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #ifdef CONFIG_FAIR_GROUP_SCHED static void __set_se_shares(struct sched_entity *se, unsigned long shares) @@ -8469,13 +8376,6 @@ static int tg_schedulable(struct task_group *tg, void *data) runtime = d->rt_runtime; } -#ifdef CONFIG_USER_SCHED - if (tg == &root_task_group) { - period = global_rt_period(); - runtime = global_rt_runtime(); - } -#endif - /* * Cannot have more runtime than the period. */ diff --git a/kernel/sys.c b/kernel/sys.c index 26a6b73a6b85..f75bf0936f47 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -569,11 +569,6 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; - if (!task_can_switch_user(new_user, current)) { - free_uid(new_user); - return -EINVAL; - } - if (atomic_read(&new_user->processes) >= current->signal->rlim[RLIMIT_NPROC].rlim_cur && new_user != INIT_USER) { diff --git a/kernel/user.c b/kernel/user.c index 46d0165ca70c..766467b3bcb7 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -56,9 +56,6 @@ struct user_struct root_user = { .sigpending = ATOMIC_INIT(0), .locked_shm = 0, .user_ns = &init_user_ns, -#ifdef CONFIG_USER_SCHED - .tg = &init_task_group, -#endif }; /* @@ -75,268 +72,6 @@ static void uid_hash_remove(struct user_struct *up) put_user_ns(up->user_ns); } -#ifdef CONFIG_USER_SCHED - -static void sched_destroy_user(struct user_struct *up) -{ - sched_destroy_group(up->tg); -} - -static int sched_create_user(struct user_struct *up) -{ - int rc = 0; - - up->tg = sched_create_group(&root_task_group); - if (IS_ERR(up->tg)) - rc = -ENOMEM; - - set_tg_uid(up); - - return rc; -} - -#else /* CONFIG_USER_SCHED */ - -static void sched_destroy_user(struct user_struct *up) { } -static int sched_create_user(struct user_struct *up) { return 0; } - -#endif /* CONFIG_USER_SCHED */ - -#if defined(CONFIG_USER_SCHED) && defined(CONFIG_SYSFS) - -static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) -{ - struct user_struct *user; - struct hlist_node *h; - - hlist_for_each_entry(user, h, hashent, uidhash_node) { - if (user->uid == uid) { - /* possibly resurrect an "almost deleted" object */ - if (atomic_inc_return(&user->__count) == 1) - cancel_delayed_work(&user->work); - return user; - } - } - - return NULL; -} - -static struct kset *uids_kset; /* represents the /sys/kernel/uids/ directory */ -static DEFINE_MUTEX(uids_mutex); - -static inline void uids_mutex_lock(void) -{ - mutex_lock(&uids_mutex); -} - -static inline void uids_mutex_unlock(void) -{ - mutex_unlock(&uids_mutex); -} - -/* uid directory attributes */ -#ifdef CONFIG_FAIR_GROUP_SCHED -static ssize_t cpu_shares_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_shares(up->tg)); -} - -static ssize_t cpu_shares_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long shares; - int rc; - - sscanf(buf, "%lu", &shares); - - rc = sched_group_set_shares(up->tg, shares); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_share_attr = - __ATTR(cpu_share, 0644, cpu_shares_show, cpu_shares_store); -#endif - -#ifdef CONFIG_RT_GROUP_SCHED -static ssize_t cpu_rt_runtime_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%ld\n", sched_group_rt_runtime(up->tg)); -} - -static ssize_t cpu_rt_runtime_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_runtime; - int rc; - - sscanf(buf, "%ld", &rt_runtime); - - rc = sched_group_set_rt_runtime(up->tg, rt_runtime); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_runtime_attr = - __ATTR(cpu_rt_runtime, 0644, cpu_rt_runtime_show, cpu_rt_runtime_store); - -static ssize_t cpu_rt_period_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_rt_period(up->tg)); -} - -static ssize_t cpu_rt_period_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_period; - int rc; - - sscanf(buf, "%lu", &rt_period); - - rc = sched_group_set_rt_period(up->tg, rt_period); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_period_attr = - __ATTR(cpu_rt_period, 0644, cpu_rt_period_show, cpu_rt_period_store); -#endif - -/* default attributes per uid directory */ -static struct attribute *uids_attributes[] = { -#ifdef CONFIG_FAIR_GROUP_SCHED - &cpu_share_attr.attr, -#endif -#ifdef CONFIG_RT_GROUP_SCHED - &cpu_rt_runtime_attr.attr, - &cpu_rt_period_attr.attr, -#endif - NULL -}; - -/* the lifetime of user_struct is not managed by the core (now) */ -static void uids_release(struct kobject *kobj) -{ - return; -} - -static struct kobj_type uids_ktype = { - .sysfs_ops = &kobj_sysfs_ops, - .default_attrs = uids_attributes, - .release = uids_release, -}; - -/* - * Create /sys/kernel/uids//cpu_share file for this user - * We do not create this file for users in a user namespace (until - * sysfs tagging is implemented). - * - * See Documentation/scheduler/sched-design-CFS.txt for ramifications. - */ -static int uids_user_create(struct user_struct *up) -{ - struct kobject *kobj = &up->kobj; - int error; - - memset(kobj, 0, sizeof(struct kobject)); - if (up->user_ns != &init_user_ns) - return 0; - kobj->kset = uids_kset; - error = kobject_init_and_add(kobj, &uids_ktype, NULL, "%d", up->uid); - if (error) { - kobject_put(kobj); - goto done; - } - - kobject_uevent(kobj, KOBJ_ADD); -done: - return error; -} - -/* create these entries in sysfs: - * "/sys/kernel/uids" directory - * "/sys/kernel/uids/0" directory (for root user) - * "/sys/kernel/uids/0/cpu_share" file (for root user) - */ -int __init uids_sysfs_init(void) -{ - uids_kset = kset_create_and_add("uids", NULL, kernel_kobj); - if (!uids_kset) - return -ENOMEM; - - return uids_user_create(&root_user); -} - -/* delayed work function to remove sysfs directory for a user and free up - * corresponding structures. - */ -static void cleanup_user_struct(struct work_struct *w) -{ - struct user_struct *up = container_of(w, struct user_struct, work.work); - unsigned long flags; - int remove_user = 0; - - /* Make uid_hash_remove() + sysfs_remove_file() + kobject_del() - * atomic. - */ - uids_mutex_lock(); - - spin_lock_irqsave(&uidhash_lock, flags); - if (atomic_read(&up->__count) == 0) { - uid_hash_remove(up); - remove_user = 1; - } - spin_unlock_irqrestore(&uidhash_lock, flags); - - if (!remove_user) - goto done; - - if (up->user_ns == &init_user_ns) { - kobject_uevent(&up->kobj, KOBJ_REMOVE); - kobject_del(&up->kobj); - kobject_put(&up->kobj); - } - - sched_destroy_user(up); - key_put(up->uid_keyring); - key_put(up->session_keyring); - kmem_cache_free(uid_cachep, up); - -done: - uids_mutex_unlock(); -} - -/* IRQs are disabled and uidhash_lock is held upon function entry. - * IRQ state (as stored in flags) is restored and uidhash_lock released - * upon function exit. - */ -static void free_user(struct user_struct *up, unsigned long flags) -{ - INIT_DELAYED_WORK(&up->work, cleanup_user_struct); - schedule_delayed_work(&up->work, msecs_to_jiffies(1000)); - spin_unlock_irqrestore(&uidhash_lock, flags); -} - -#else /* CONFIG_USER_SCHED && CONFIG_SYSFS */ - static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) { struct user_struct *user; @@ -352,11 +87,6 @@ static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) return NULL; } -int uids_sysfs_init(void) { return 0; } -static inline int uids_user_create(struct user_struct *up) { return 0; } -static inline void uids_mutex_lock(void) { } -static inline void uids_mutex_unlock(void) { } - /* IRQs are disabled and uidhash_lock is held upon function entry. * IRQ state (as stored in flags) is restored and uidhash_lock released * upon function exit. @@ -365,32 +95,11 @@ static void free_user(struct user_struct *up, unsigned long flags) { uid_hash_remove(up); spin_unlock_irqrestore(&uidhash_lock, flags); - sched_destroy_user(up); key_put(up->uid_keyring); key_put(up->session_keyring); kmem_cache_free(uid_cachep, up); } -#endif - -#if defined(CONFIG_RT_GROUP_SCHED) && defined(CONFIG_USER_SCHED) -/* - * We need to check if a setuid can take place. This function should be called - * before successfully completing the setuid. - */ -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - - return sched_rt_can_attach(up->tg, tsk); - -} -#else -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - return 1; -} -#endif - /* * Locate the user_struct for the passed UID. If found, take a ref on it. The * caller must undo that ref with free_uid(). @@ -431,8 +140,6 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) /* Make uid_hash_find() + uids_user_create() + uid_hash_insert() * atomic. */ - uids_mutex_lock(); - spin_lock_irq(&uidhash_lock); up = uid_hash_find(uid, hashent); spin_unlock_irq(&uidhash_lock); @@ -445,14 +152,8 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) new->uid = uid; atomic_set(&new->__count, 1); - if (sched_create_user(new) < 0) - goto out_free_user; - new->user_ns = get_user_ns(ns); - if (uids_user_create(new)) - goto out_destoy_sched; - /* * Before adding this, check whether we raced * on adding the same user already.. @@ -475,17 +176,11 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) spin_unlock_irq(&uidhash_lock); } - uids_mutex_unlock(); - return up; -out_destoy_sched: - sched_destroy_user(new); put_user_ns(new->user_ns); -out_free_user: kmem_cache_free(uid_cachep, new); out_unlock: - uids_mutex_unlock(); return NULL; } -- cgit v1.2.2 From fabf318e5e4bda0aca2b0d617b191884fda62703 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 21 Jan 2010 21:04:57 +0100 Subject: sched: Fix fork vs hotplug vs cpuset namespaces There are a number of issues: 1) TASK_WAKING vs cgroup_clone (cpusets) copy_process(): sched_fork() child->state = TASK_WAKING; /* waiting for wake_up_new_task() */ if (current->nsproxy != p->nsproxy) ns_cgroup_clone() cgroup_clone() mutex_lock(inode->i_mutex) mutex_lock(cgroup_mutex) cgroup_attach_task() ss->can_attach() ss->attach() [ -> cpuset_attach() ] cpuset_attach_task() set_cpus_allowed_ptr(); while (child->state == TASK_WAKING) cpu_relax(); will deadlock the system. 2) cgroup_clone (cpusets) vs copy_process So even if the above would work we still have: copy_process(): if (current->nsproxy != p->nsproxy) ns_cgroup_clone() cgroup_clone() mutex_lock(inode->i_mutex) mutex_lock(cgroup_mutex) cgroup_attach_task() ss->can_attach() ss->attach() [ -> cpuset_attach() ] cpuset_attach_task() set_cpus_allowed_ptr(); ... p->cpus_allowed = current->cpus_allowed over-writing the modified cpus_allowed. 3) fork() vs hotplug if we unplug the child's cpu after the sanity check when the child gets attached to the task_list but before wake_up_new_task() shit will meet with fan. Solve all these issues by moving fork cpu selection into wake_up_new_task(). Reported-by: Serge E. Hallyn Tested-by: Serge E. Hallyn Signed-off-by: Peter Zijlstra LKML-Reference: <1264106190.4283.1314.camel@laptop> Signed-off-by: Thomas Gleixner --- kernel/fork.c | 15 --------------- kernel/sched.c | 39 +++++++++++++++++++++++++++------------ 2 files changed, 27 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 5b2959b3ffc2..f88bd984df35 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1241,21 +1241,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, /* Need tasklist lock for parent etc handling! */ write_lock_irq(&tasklist_lock); - /* - * The task hasn't been attached yet, so its cpus_allowed mask will - * not be changed, nor will its assigned CPU. - * - * The cpus_allowed mask of the parent may have changed after it was - * copied first time - so re-copy it here, then check the child's CPU - * to ensure it is on a valid CPU (and if not, just force it back to - * parent's CPU). This avoids alot of nasty races. - */ - p->cpus_allowed = current->cpus_allowed; - p->rt.nr_cpus_allowed = current->rt.nr_cpus_allowed; - if (unlikely(!cpu_isset(task_cpu(p), p->cpus_allowed) || - !cpu_online(task_cpu(p)))) - set_task_cpu(p, smp_processor_id()); - /* CLONE_PARENT re-uses the old parent */ if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) { p->real_parent = current->real_parent; diff --git a/kernel/sched.c b/kernel/sched.c index 4508fe7048be..3a8fb30a91b1 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2320,14 +2320,12 @@ static int select_fallback_rq(int cpu, struct task_struct *p) } /* - * Called from: + * Gets called from 3 sites (exec, fork, wakeup), since it is called without + * holding rq->lock we need to ensure ->cpus_allowed is stable, this is done + * by: * - * - fork, @p is stable because it isn't on the tasklist yet - * - * - exec, @p is unstable, retry loop - * - * - wake-up, we serialize ->cpus_allowed against TASK_WAKING so - * we should be good. + * exec: is unstable, retry loop + * fork & wake-up: serialize ->cpus_allowed against TASK_WAKING */ static inline int select_task_rq(struct task_struct *p, int sd_flags, int wake_flags) @@ -2620,9 +2618,6 @@ void sched_fork(struct task_struct *p, int clone_flags) if (p->sched_class->task_fork) p->sched_class->task_fork(p); -#ifdef CONFIG_SMP - cpu = select_task_rq(p, SD_BALANCE_FORK, 0); -#endif set_task_cpu(p, cpu); #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) @@ -2652,6 +2647,21 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; + int cpu = get_cpu(); + +#ifdef CONFIG_SMP + /* + * Fork balancing, do it here and not earlier because: + * - cpus_allowed can change in the fork path + * - any previously selected cpu might disappear through hotplug + * + * We still have TASK_WAKING but PF_STARTING is gone now, meaning + * ->cpus_allowed is stable, we have preemption disabled, meaning + * cpu_online_mask is stable. + */ + cpu = select_task_rq(p, SD_BALANCE_FORK, 0); + set_task_cpu(p, cpu); +#endif rq = task_rq_lock(p, &flags); BUG_ON(p->state != TASK_WAKING); @@ -2665,6 +2675,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) p->sched_class->task_woken(rq, p); #endif task_rq_unlock(rq, &flags); + put_cpu(); } #ifdef CONFIG_PREEMPT_NOTIFIERS @@ -7139,14 +7150,18 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) * the ->cpus_allowed mask from under waking tasks, which would be * possible when we change rq->lock in ttwu(), so synchronize against * TASK_WAKING to avoid that. + * + * Make an exception for freshly cloned tasks, since cpuset namespaces + * might move the task about, we have to validate the target in + * wake_up_new_task() anyway since the cpu might have gone away. */ again: - while (p->state == TASK_WAKING) + while (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) cpu_relax(); rq = task_rq_lock(p, &flags); - if (p->state == TASK_WAKING) { + if (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) { task_rq_unlock(rq, &flags); goto again; } -- cgit v1.2.2 From ea87bb7853168434f4a82426dd1ea8421f9e604d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:58:57 +0000 Subject: sched: Extend enqueue_task to allow head queueing The ability of enqueueing a task to the head of a SCHED_FIFO priority list is required to fix some violations of POSIX scheduling policy. Extend the related functions with a "head" argument. Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.734886007@linutronix.de> --- kernel/sched.c | 13 +++++++------ kernel/sched_fair.c | 3 ++- kernel/sched_rt.c | 3 ++- 3 files changed, 11 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 41e76d325648..f47560ff3346 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1856,13 +1856,14 @@ static void update_avg(u64 *avg, u64 sample) *avg += diff >> 3; } -static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task(struct rq *rq, struct task_struct *p, int wakeup, bool head) { if (wakeup) p->se.start_runtime = p->se.sum_exec_runtime; sched_info_queued(p); - p->sched_class->enqueue_task(rq, p, wakeup); + p->sched_class->enqueue_task(rq, p, wakeup, head); p->se.on_rq = 1; } @@ -1892,7 +1893,7 @@ static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) if (task_contributes_to_load(p)) rq->nr_uninterruptible--; - enqueue_task(rq, p, wakeup); + enqueue_task(rq, p, wakeup, false); inc_nr_running(rq); } @@ -4236,7 +4237,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, false); check_class_changed(rq, p, prev_class, oldprio, running); } @@ -4280,7 +4281,7 @@ void set_user_nice(struct task_struct *p, long nice) delta = p->prio - old_prio; if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, false); /* * If the task increased its priority or is running and * lowered its priority, then reschedule its CPU: @@ -8230,7 +8231,7 @@ void sched_move_task(struct task_struct *tsk) if (unlikely(running)) tsk->sched_class->set_curr_task(rq); if (on_rq) - enqueue_task(rq, tsk, 0); + enqueue_task(rq, tsk, 0, false); task_rq_unlock(rq, &flags); } diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 22231ccb2f98..0e7a7af9cf8b 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1053,7 +1053,8 @@ static inline void hrtick_update(struct rq *rq) * increased. Here we update the fair scheduling stats and * then put the task into the rbtree: */ -static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct cfs_rq *cfs_rq; struct sched_entity *se = &p->se; diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 502bb614e40a..38076dabb44a 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -878,7 +878,8 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) /* * Adding/removing a task to/from a priority array: */ -static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct sched_rt_entity *rt_se = &p->rt; -- cgit v1.2.2 From 37dad3fce97f01e5149d69de0833d8452c0e862e Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:59:01 +0000 Subject: sched: Implement head queueing for sched_rt The ability of enqueueing a task to the head of a SCHED_FIFO priority list is required to fix some violations of POSIX scheduling policy. Implement the functionality in sched_rt. Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.772169931@linutronix.de> --- kernel/sched_rt.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 38076dabb44a..ca49ceb01201 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -194,7 +194,7 @@ static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se) return rt_se->my_q; } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se); +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head); static void dequeue_rt_entity(struct sched_rt_entity *rt_se); static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) @@ -204,7 +204,7 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, false); if (rt_rq->highest_prio.curr < curr->prio) resched_task(curr); } @@ -803,7 +803,7 @@ void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) dec_rt_group(rt_se, rt_rq); } -static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { struct rt_rq *rt_rq = rt_rq_of_se(rt_se); struct rt_prio_array *array = &rt_rq->active; @@ -819,7 +819,10 @@ static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running)) return; - list_add_tail(&rt_se->run_list, queue); + if (head) + list_add(&rt_se->run_list, queue); + else + list_add_tail(&rt_se->run_list, queue); __set_bit(rt_se_prio(rt_se), array->bitmap); inc_rt_tasks(rt_se, rt_rq); @@ -856,11 +859,11 @@ static void dequeue_rt_stack(struct sched_rt_entity *rt_se) } } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { dequeue_rt_stack(rt_se); for_each_sched_rt_entity(rt_se) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, head); } static void dequeue_rt_entity(struct sched_rt_entity *rt_se) @@ -871,7 +874,7 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) struct rt_rq *rt_rq = group_rt_rq(rt_se); if (rt_rq && rt_rq->rt_nr_running) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, false); } } @@ -886,7 +889,7 @@ enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup, bool head) if (wakeup) rt_se->timeout = 0; - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, head); if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1) enqueue_pushable_task(rq, p); -- cgit v1.2.2 From 60db48cacb9b253d5607a5ff206112a59cd09e34 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:59:06 +0000 Subject: sched: Queue a deboosted task to the head of the RT prio queue rtmutex_set_prio() is used to implement priority inheritance for futexes. When a task is deboosted it gets enqueued at the tail of its RT priority list. This is violating the POSIX scheduling semantics: rt priority list X contains two runnable tasks A and B task A runs with priority X and holds mutex M task C preempts A and is blocked on mutex M -> task A is boosted to priority of task C (Y) task A unlocks the mutex M and deboosts itself -> A is dequeued from rt priority list Y -> A is enqueued to the tail of rt priority list X task C schedules away task B runs This is wrong as task A did not schedule away and therefor violates the POSIX scheduling semantics. Enqueue the task to the head of the priority list instead. Reported-by: Mathias Weber Reported-by: Carsten Emde Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.809074113@linutronix.de> --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index f47560ff3346..a56ead42d651 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4237,7 +4237,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); if (on_rq) { - enqueue_task(rq, p, 0, false); + enqueue_task(rq, p, 0, oldprio < prio); check_class_changed(rq, p, prev_class, oldprio, running); } -- cgit v1.2.2 From 74bf4076f2ed79b5510440b72a561823a8852ec0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 25 Jan 2010 15:11:53 -0500 Subject: tracing: Prevent kernel oops with corrupted buffer If the contents of the ftrace ring buffer gets corrupted and the trace file is read, it could create a kernel oops (usualy just killing the user task thread). This is caused by the checking of the pid in the buffer. If the pid is negative, it still references the cmdline cache array, which could point to an invalid address. The simple fix is to test for negative PIDs. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0df1b0f2cb9e..eac6875cb990 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -951,6 +951,11 @@ void trace_find_cmdline(int pid, char comm[]) return; } + if (WARN_ON_ONCE(pid < 0)) { + strcpy(comm, ""); + return; + } + if (pid > PID_MAX_DEFAULT) { strcpy(comm, "<...>"); return; -- cgit v1.2.2 From 7b7422a566aa0dc1e582ce263d4c7ff4a772700a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 26 Jan 2010 12:51:10 +0100 Subject: clocksource: Prevent potential kgdb dead lock commit 0f8e8ef7 (clocksource: Simplify clocksource watchdog resume logic) introduced a potential kgdb dead lock. When the kernel is stopped by kgdb inside code which holds watchdog_lock then kgdb dead locks in clocksource_resume_watchdog(). clocksource_resume_watchdog() is called from kbdg via clocksource_touch_watchdog() to avoid that the clock source watchdog marks TSC unstable after the kernel has been stopped. Solve this by replacing spin_lock with a spin_trylock and just return in case the lock is held. Not resetting the watchdog might result in TSC becoming marked unstable, but that's an acceptable penalty for using kgdb. The timekeeping is anyway easily screwed up by kgdb when the system uses either jiffies or a clock source which wraps in short intervals (e.g. pm_timer wraps about every 4.6s), so we really do not have to worry about that occasional TSC marked unstable side effect. The second caller of clocksource_resume_watchdog() is clocksource_resume(). The trylock is safe here as well because the system is UP at this point, interrupts are disabled and nothing else can hold watchdog_lock(). Reported-by: Jason Wessel LKML-Reference: <1264480000-6997-4-git-send-email-jason.wessel@windriver.com> Cc: kgdb-bugreport@lists.sourceforge.net Cc: Martin Schwidefsky Cc: John Stultz Cc: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index e85c23404d34..13700833c181 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -343,7 +343,19 @@ static void clocksource_resume_watchdog(void) { unsigned long flags; - spin_lock_irqsave(&watchdog_lock, flags); + /* + * We use trylock here to avoid a potential dead lock when + * kgdb calls this code after the kernel has been stopped with + * watchdog_lock held. When watchdog_lock is held we just + * return and accept, that the watchdog might trigger and mark + * the monitored clock source (usually TSC) unstable. + * + * This does not affect the other caller clocksource_resume() + * because at this point the kernel is UP, interrupts are + * disabled and nothing can hold watchdog_lock. + */ + if (!spin_trylock_irqsave(&watchdog_lock, flags)) + return; clocksource_reset_watchdog(); spin_unlock_irqrestore(&watchdog_lock, flags); } @@ -458,8 +470,8 @@ void clocksource_resume(void) * clocksource_touch_watchdog - Update watchdog * * Update the watchdog after exception contexts such as kgdb so as not - * to incorrectly trip the watchdog. - * + * to incorrectly trip the watchdog. This might fail when the kernel + * was stopped in code which holds watchdog_lock. */ void clocksource_touch_watchdog(void) { -- cgit v1.2.2 From 492a74f4210e15f4701422e2e1c4cd3c1e45ddae Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 25 Jan 2010 15:17:47 -0500 Subject: ring-buffer: Check if ring buffer iterator has stale data Usually reads of the ring buffer is performed by a single task. There are two types of reads from the ring buffer. One is a consuming read which will consume the entry that was read and the next read will be the entry that follows. The other is an iterator that will let the user read the contents of the ring buffer without modifying it. When an iterator is allocated, writes to the ring buffer are disabled to protect the iterator. The problem exists when consuming reads happen while an iterator is allocated. Specifically, the kind of read that swaps out an entire page (used by splice) and replaces it with a new read. If the iterator is on the page that is swapped out, then the next read may read from this swapped out page and return garbage. This patch adds a check when reading the iterator to make sure that the iterator contents are still valid. If a consuming read has taken place, the iterator is reset. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index edefe3b2801b..503b630e0bda 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -464,6 +464,8 @@ struct ring_buffer_iter { struct ring_buffer_per_cpu *cpu_buffer; unsigned long head; struct buffer_page *head_page; + struct buffer_page *cache_reader_page; + unsigned long cache_read; u64 read_stamp; }; @@ -2716,6 +2718,8 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) iter->read_stamp = cpu_buffer->read_stamp; else iter->read_stamp = iter->head_page->page->time_stamp; + iter->cache_reader_page = cpu_buffer->reader_page; + iter->cache_read = cpu_buffer->read; } /** @@ -3066,6 +3070,15 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) cpu_buffer = iter->cpu_buffer; buffer = cpu_buffer->buffer; + /* + * Check if someone performed a consuming read to + * the buffer. A consuming read invalidates the iterator + * and we need to reset the iterator in this case. + */ + if (unlikely(iter->cache_read != cpu_buffer->read || + iter->cache_reader_page != cpu_buffer->reader_page)) + rb_iter_reset(iter); + again: /* * We repeat when a timestamp is encountered. -- cgit v1.2.2 From 3c05d7482777f15e71bb4cb1ba78dee2800dfec6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 26 Jan 2010 16:14:08 -0500 Subject: ring-buffer: Check for end of page in iterator If the iterator comes to an empty page for some reason, or if the page is emptied by a consuming read. The iterator code currently does not check if the iterator is pass the contents, and may return a false entry. This patch adds a check to the ring buffer iterator to test if the current page has been completely read and sets the iterator to the next page if necessary. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 503b630e0bda..8c1b2d290718 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -3064,9 +3064,6 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event *event; int nr_loops = 0; - if (ring_buffer_iter_empty(iter)) - return NULL; - cpu_buffer = iter->cpu_buffer; buffer = cpu_buffer->buffer; @@ -3080,6 +3077,9 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) rb_iter_reset(iter); again: + if (ring_buffer_iter_empty(iter)) + return NULL; + /* * We repeat when a timestamp is encountered. * We can get multiple timestamps by nested interrupts or also @@ -3094,6 +3094,11 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) if (rb_per_cpu_empty(cpu_buffer)) return NULL; + if (iter->head >= local_read(&iter->head_page->page->commit)) { + rb_inc_iter(iter); + goto again; + } + event = rb_iter_head_event(iter); switch (event->type_len) { -- cgit v1.2.2 From 03688970347bfea32823953a7ce5886d1713205f Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Fri, 22 Jan 2010 08:12:47 -0500 Subject: tracing/documentation: Cover new frame pointer semantics Update the graph tracer examples to cover the new frame pointer semantics (in terms of passing it along). Move the HAVE_FUNCTION_GRAPH_FP_TEST docs out of the Kconfig, into the right place, and expand on the details. Signed-off-by: Mike Frysinger LKML-Reference: <1264165967-18938-1-git-send-email-vapier@gentoo.org> Signed-off-by: Steven Rostedt --- kernel/trace/Kconfig | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 6c22d8a2f289..60e2ce0181ee 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -27,9 +27,7 @@ config HAVE_FUNCTION_GRAPH_TRACER config HAVE_FUNCTION_GRAPH_FP_TEST bool help - An arch may pass in a unique value (frame pointer) to both the - entering and exiting of a function. On exit, the value is compared - and if it does not match, then it will panic the kernel. + See Documentation/trace/ftrace-design.txt config HAVE_FUNCTION_TRACE_MCOUNT_TEST bool -- cgit v1.2.2 From 48d50674179981e41f432167b2441cec782d5484 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 26 Jan 2010 19:16:41 +0100 Subject: lockdep: Fix check_usage_backwards() error message Lockdep has found the real bug, but the output doesn't look right to me: > ========================================================= > [ INFO: possible irq lock inversion dependency detected ] > 2.6.33-rc5 #77 > --------------------------------------------------------- > emacs/1609 just changed the state of lock: > (&(&tty->ctrl_lock)->rlock){+.....}, at: [] tty_fasync+0xe8/0x190 > but this lock took another, HARDIRQ-unsafe lock in the past: > (&(&sighand->siglock)->rlock){-.....} "HARDIRQ-unsafe" and "this lock took another" looks wrong, afaics. > ... key at: [] __key.46539+0x0/0x8 > ... acquired at: > [] __lock_acquire+0x1056/0x15a0 > [] lock_acquire+0x9f/0x120 > [] _raw_spin_lock_irqsave+0x52/0x90 > [] __proc_set_tty+0x3e/0x150 > [] tty_open+0x51d/0x5e0 The stack-trace shows that this lock (ctrl_lock) was taken under ->siglock (which is hopefully irq-safe). This is a clear typo in check_usage_backwards() where we tell the print a fancy routine we're forwards. Signed-off-by: Oleg Nesterov Signed-off-by: Peter Zijlstra LKML-Reference: <20100126181641.GA10460@redhat.com> Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 5feaddcdbe49..c62ec14609b9 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -2147,7 +2147,7 @@ check_usage_backwards(struct task_struct *curr, struct held_lock *this, return ret; return print_irq_inversion_bug(curr, &root, target_entry, - this, 1, irqclass); + this, 0, irqclass); } void print_irqtrace_events(struct task_struct *curr) -- cgit v1.2.2 From 11854247e2c851e7ff9ce138e501c6cffc5a4217 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 21 Jan 2010 16:34:27 +0100 Subject: sched: Fix incorrect sanity check We moved to migrate on wakeup, which means that sleeping tasks could still be present on offline cpus. Amend the check to only test running tasks. Reported-by: Heiko Carstens Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/cpu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 1c8ddd6ee940..08e54e7beaed 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -151,7 +151,7 @@ static inline void check_for_tasks(int cpu) write_lock_irq(&tasklist_lock); for_each_process(p) { - if (task_cpu(p) == cpu && + if (task_cpu(p) == cpu && p->state == TASK_RUNNING && (!cputime_eq(p->utime, cputime_zero) || !cputime_eq(p->stime, cputime_zero))) printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\ -- cgit v1.2.2 From 9d3cfc4c1d17c6d3bc1373e3b954c56b92607755 Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Mon, 25 Jan 2010 14:56:34 +0100 Subject: sched: Correct printk whitespace in warning from cpu down task check Due to an incorrect line break the output currently contains tabs. Also remove trailing space. The actual output that logcheck sent me looked like this: Task events/1 (pid = 10) is on cpu 1^I^I^I^I(state = 1, flags = 84208040) After this patch it becomes: Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040) Signed-off-by: Frans Pop Signed-off-by: Peter Zijlstra LKML-Reference: <201001251456.34996.elendil@planet.nl> Signed-off-by: Ingo Molnar --- kernel/cpu.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 08e54e7beaed..677f25376a38 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -154,10 +154,10 @@ static inline void check_for_tasks(int cpu) if (task_cpu(p) == cpu && p->state == TASK_RUNNING && (!cputime_eq(p->utime, cputime_zero) || !cputime_eq(p->stime, cputime_zero))) - printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\ - (state = %ld, flags = %x) \n", - p->comm, task_pid_nr(p), cpu, - p->state, p->flags); + printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d " + "(state = %ld, flags = %x)\n", + p->comm, task_pid_nr(p), cpu, + p->state, p->flags); } write_unlock_irq(&tasklist_lock); } -- cgit v1.2.2 From ea2c68a08fedb5053ba312d661e47df9f4d72411 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 13 Jan 2010 19:38:30 +0800 Subject: tracing: Simplify test for function_graph tracing start point In the function graph tracer, a calling function is to be traced only when it is enabled through the set_graph_function file, or when it is nested in an enabled function. Current code uses TSK_TRACE_FL_GRAPH to test whether it is nested or not. Looking at the code, we can get this: (trace->depth > 0) <==> (TSK_TRACE_FL_GRAPH is set) trace->depth is more explicit to tell that it is nested. So we use trace->depth directly and simplify the code. No functionality is changed. TSK_TRACE_FL_GRAPH is not removed yet, it is left for future usage. Signed-off-by: Lai Jiangshan Cc: Ingo Molnar Cc: Steven Rostedt LKML-Reference: <4B4DB0B6.7040607@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace.h | 2 +- kernel/trace/trace_functions_graph.c | 8 ++------ 2 files changed, 3 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4df6a77eb196..ce077fbbf552 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -504,7 +504,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count || test_tsk_trace_graph(current)) + if (!ftrace_graph_count) return 1; for (i = 0; i < ftrace_graph_count; i++) { diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index f2252296607c..616b135c9eb9 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -215,7 +215,8 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) if (!ftrace_trace_task(current)) return 0; - if (!ftrace_graph_addr(trace->func)) + /* trace it when it is-nested-in or is a function enabled. */ + if (!(trace->depth || ftrace_graph_addr(trace->func))) return 0; local_irq_save(flags); @@ -228,9 +229,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) } else { ret = 0; } - /* Only do the atomic if it is not already set */ - if (!test_tsk_trace_graph(current)) - set_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); @@ -278,8 +276,6 @@ void trace_graph_return(struct ftrace_graph_ret *trace) pc = preempt_count(); __trace_graph_return(tr, trace, flags, pc); } - if (!trace->depth) - clear_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); } -- cgit v1.2.2 From 1f5b8f8a2031ae9507eb67743cad4d424739bfff Mon Sep 17 00:00:00 2001 From: john stultz Date: Thu, 28 Jan 2010 15:02:41 -0800 Subject: ntp: Make time_esterror and time_maxerror static Make time_esterror and time_maxerror static as no one uses them outside of ntp.c Signed-off-by: John Stultz Cc: richard@rsk.demon.co.uk LKML-Reference: <1264719761.3437.47.camel@localhost.localdomain> Signed-off-by: Thomas Gleixner --- kernel/time/ntp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 4800f933910e..74b1b37b1595 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -58,10 +58,10 @@ static s64 time_offset; static long time_constant = 2; /* maximum error (usecs): */ -long time_maxerror = NTP_PHASE_LIMIT; +static long time_maxerror = NTP_PHASE_LIMIT; /* estimated error (usecs): */ -long time_esterror = NTP_PHASE_LIMIT; +static long time_esterror = NTP_PHASE_LIMIT; /* frequency offset (scaled nsecs/secs): */ static s64 time_freq; -- cgit v1.2.2 From 7e1b584774c6168ca5b27c340fbeff8f67651e4f Mon Sep 17 00:00:00 2001 From: John Stultz Date: Thu, 28 Jan 2010 20:20:44 -0800 Subject: ntp: Cleanup xtime references in ntp.c ntp.c doesn't need to access timekeeping internals directly, so change xtime references to use the get_seconds() timekeeping interface. Signed-off-by: John Stultz Cc: richard@rsk.demon.co.uk LKML-Reference: <1264738844-21935-1-git-send-email-johnstul@us.ibm.com> Signed-off-by: Thomas Gleixner --- kernel/time/ntp.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 74b1b37b1595..7c0f180d6e9d 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -142,11 +142,11 @@ static void ntp_update_offset(long offset) * Select how the frequency is to be controlled * and in which mode (PLL or FLL). */ - secs = xtime.tv_sec - time_reftime; + secs = get_seconds() - time_reftime; if (unlikely(time_status & STA_FREQHOLD)) secs = 0; - time_reftime = xtime.tv_sec; + time_reftime = get_seconds(); offset64 = offset; freq_adj = (offset64 * secs) << @@ -368,7 +368,7 @@ static inline void process_adj_status(struct timex *txc, struct timespec *ts) * reference time to current time. */ if (!(time_status & STA_PLL) && (txc->status & STA_PLL)) - time_reftime = xtime.tv_sec; + time_reftime = get_seconds(); /* only set allowed bits */ time_status &= STA_RONLY; -- cgit v1.2.2 From cc0967490c1c3824bc5b75718b6ca8a51d9f2617 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Thu, 28 Jan 2010 17:04:42 -0600 Subject: x86, hw_breakpoints, kgdb: Fix kgdb to use hw_breakpoint API In the 2.6.33 kernel, the hw_breakpoint API is now used for the performance event counters. The hw_breakpoint_handler() now consumes the hw breakpoints that were previously set by kgdb arch specific code. In order for kgdb to work in conjunction with this core API change, kgdb must use some of the low level functions of the hw_breakpoint API to install, uninstall, and deal with hw breakpoint reservations. The kgdb core required a change to call kgdb_disable_hw_debug anytime a slave cpu enters kgdb_wait() in order to keep all the hw breakpoints in sync as well as to prevent hitting a hw breakpoint while kgdb is active. During the architecture specific initialization of kgdb, it will pre-allocate 4 disabled (struct perf event **) structures. Kgdb will use these to manage the capabilities for the 4 hw breakpoint registers, per cpu. Right now the hw_breakpoint API does not have a way to ask how many breakpoints are available, on each CPU so it is possible that the install of a breakpoint might fail when kgdb restores the system to the run state. The intent of this patch is to first get the basic functionality of hw breakpoints working and leave it to the person debugging the kernel to understand what hw breakpoints are in use and what restrictions have been imposed as a result. Breakpoint constraints will be dealt with in a future patch. While atomic, the x86 specific kgdb code will call arch_uninstall_hw_breakpoint() and arch_install_hw_breakpoint() to manage the cpu specific hw breakpoints. The net result of these changes allow kgdb to use the same pool of hw_breakpoints that are used by the perf event API, but neither knows about future reservations for the available hw breakpoint slots. Signed-off-by: Jason Wessel Acked-by: Frederic Weisbecker Cc: kgdb-bugreport@lists.sourceforge.net Cc: K.Prasad Cc: Peter Zijlstra Cc: Alan Stern Cc: torvalds@linux-foundation.org LKML-Reference: <1264719883-7285-2-git-send-email-jason.wessel@windriver.com> Signed-off-by: Ingo Molnar --- kernel/kgdb.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 2eb517e23514..c7ade62e4ef0 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -583,6 +583,9 @@ static void kgdb_wait(struct pt_regs *regs) smp_wmb(); atomic_set(&cpu_in_kgdb[cpu], 1); + /* Disable any cpu specific hw breakpoints */ + kgdb_disable_hw_debug(regs); + /* Wait till primary CPU is done with debugging */ while (atomic_read(&passive_cpu_wait[cpu])) cpu_relax(); -- cgit v1.2.2 From 5352ae638e2d7d5c9b2e4d528676bbf2af6fd6f3 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Thu, 28 Jan 2010 17:04:43 -0600 Subject: perf, hw_breakpoint, kgdb: Do not take mutex for kernel debugger This patch fixes the regression in functionality where the kernel debugger and the perf API do not nicely share hw breakpoint reservations. The kernel debugger cannot use any mutex_lock() calls because it can start the kernel running from an invalid context. A mutex free version of the reservation API needed to get created for the kernel debugger to safely update hw breakpoint reservations. The possibility for a breakpoint reservation to be concurrently processed at the time that kgdb interrupts the system is improbable. Should this corner case occur the end user is warned, and the kernel debugger will prohibit updating the hardware breakpoint reservations. Any time the kernel debugger reserves a hardware breakpoint it will be a system wide reservation. Signed-off-by: Jason Wessel Acked-by: Frederic Weisbecker Cc: kgdb-bugreport@lists.sourceforge.net Cc: K.Prasad Cc: Peter Zijlstra Cc: Alan Stern Cc: torvalds@linux-foundation.org LKML-Reference: <1264719883-7285-3-git-send-email-jason.wessel@windriver.com> Signed-off-by: Ingo Molnar --- kernel/hw_breakpoint.c | 52 ++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 42 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index c030ae657f20..8a5c7d55ac9f 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -243,38 +243,70 @@ static void toggle_bp_slot(struct perf_event *bp, bool enable) * ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *)) * + max(per_cpu(nr_task_bp_pinned, *))) < HBP_NUM */ -int reserve_bp_slot(struct perf_event *bp) +static int __reserve_bp_slot(struct perf_event *bp) { struct bp_busy_slots slots = {0}; - int ret = 0; - - mutex_lock(&nr_bp_mutex); fetch_bp_busy_slots(&slots, bp); /* Flexible counters need to keep at least one slot */ - if (slots.pinned + (!!slots.flexible) == HBP_NUM) { - ret = -ENOSPC; - goto end; - } + if (slots.pinned + (!!slots.flexible) == HBP_NUM) + return -ENOSPC; toggle_bp_slot(bp, true); -end: + return 0; +} + +int reserve_bp_slot(struct perf_event *bp) +{ + int ret; + + mutex_lock(&nr_bp_mutex); + + ret = __reserve_bp_slot(bp); + mutex_unlock(&nr_bp_mutex); return ret; } +static void __release_bp_slot(struct perf_event *bp) +{ + toggle_bp_slot(bp, false); +} + void release_bp_slot(struct perf_event *bp) { mutex_lock(&nr_bp_mutex); - toggle_bp_slot(bp, false); + __release_bp_slot(bp); mutex_unlock(&nr_bp_mutex); } +/* + * Allow the kernel debugger to reserve breakpoint slots without + * taking a lock using the dbg_* variant of for the reserve and + * release breakpoint slots. + */ +int dbg_reserve_bp_slot(struct perf_event *bp) +{ + if (mutex_is_locked(&nr_bp_mutex)) + return -1; + + return __reserve_bp_slot(bp); +} + +int dbg_release_bp_slot(struct perf_event *bp) +{ + if (mutex_is_locked(&nr_bp_mutex)) + return -1; + + __release_bp_slot(bp); + + return 0; +} int register_perf_hw_breakpoint(struct perf_event *bp) { -- cgit v1.2.2 From d6ad3e286d2c075a60b9f11075a2c55aeeeca2ad Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Wed, 27 Jan 2010 16:25:22 -0600 Subject: softlockup: Add sched_clock_tick() to avoid kernel warning on kgdb resume When CONFIG_HAVE_UNSTABLE_SCHED_CLOCK is set, sched_clock() gets the time from hardware such as the TSC on x86. In this configuration kgdb will report a softlock warning message on resuming or detaching from a debug session. Sequence of events in the problem case: 1) "cpu sched clock" and "hardware time" are at 100 sec prior to a call to kgdb_handle_exception() 2) Debugger waits in kgdb_handle_exception() for 80 sec and on exit the following is called ... touch_softlockup_watchdog() --> __raw_get_cpu_var(touch_timestamp) = 0; 3) "cpu sched clock" = 100s (it was not updated, because the interrupt was disabled in kgdb) but the "hardware time" = 180 sec 4) The first timer interrupt after resuming from kgdb_handle_exception updates the watchdog from the "cpu sched clock" update_process_times() { ... run_local_timers() --> softlockup_tick() --> check (touch_timestamp == 0) (it is "YES" here, we have set "touch_timestamp = 0" at kgdb) --> __touch_softlockup_watchdog() ***(A)--> reset "touch_timestamp" to "get_timestamp()" (Here, the "touch_timestamp" will still be set to 100s.) ... scheduler_tick() ***(B)--> sched_clock_tick() (update "cpu sched clock" to "hardware time" = 180s) ... } 5) The Second timer interrupt handler appears to have a large jump and trips the softlockup warning. update_process_times() { ... run_local_timers() --> softlockup_tick() --> "cpu sched clock" - "touch_timestamp" = 180s-100s > 60s --> printk "soft lockup error messages" ... } note: ***(A) reset "touch_timestamp" to "get_timestamp(this_cpu)" Why is "touch_timestamp" 100 sec, instead of 180 sec? When CONFIG_HAVE_UNSTABLE_SCHED_CLOCK is set, the call trace of get_timestamp() is: get_timestamp(this_cpu) -->cpu_clock(this_cpu) -->sched_clock_cpu(this_cpu) -->__update_sched_clock(sched_clock_data, now) The __update_sched_clock() function uses the GTOD tick value to create a window to normalize the "now" values. So if "now" value is too big for sched_clock_data, it will be ignored. The fix is to invoke sched_clock_tick() to update "cpu sched clock" in order to recover from this state. This is done by introducing the function touch_softlockup_watchdog_sync(). This allows kgdb to request that the sched clock is updated when the watchdog thread runs the first time after a resume from kgdb. [yong.zhang0@gmail.com: Use per cpu instead of an array] Signed-off-by: Jason Wessel Signed-off-by: Dongdong Deng Cc: kgdb-bugreport@lists.sourceforge.net Cc: peterz@infradead.org LKML-Reference: <1264631124-4837-2-git-send-email-jason.wessel@windriver.com> Signed-off-by: Ingo Molnar --- kernel/kgdb.c | 6 +++--- kernel/softlockup.c | 15 +++++++++++++++ 2 files changed, 18 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 2eb517e23514..87f2cc557553 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -596,7 +596,7 @@ static void kgdb_wait(struct pt_regs *regs) /* Signal the primary CPU that we are done: */ atomic_set(&cpu_in_kgdb[cpu], 0); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sync(); clocksource_touch_watchdog(); local_irq_restore(flags); } @@ -1450,7 +1450,7 @@ acquirelock: (kgdb_info[cpu].task && kgdb_info[cpu].task->pid != kgdb_sstep_pid) && --sstep_tries) { atomic_set(&kgdb_active, -1); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sync(); clocksource_touch_watchdog(); local_irq_restore(flags); @@ -1550,7 +1550,7 @@ kgdb_restore: } /* Free kgdb_active */ atomic_set(&kgdb_active, -1); - touch_softlockup_watchdog(); + touch_softlockup_watchdog_sync(); clocksource_touch_watchdog(); local_irq_restore(flags); diff --git a/kernel/softlockup.c b/kernel/softlockup.c index d22579087e27..0d4c7898ab80 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -25,6 +25,7 @@ static DEFINE_SPINLOCK(print_lock); static DEFINE_PER_CPU(unsigned long, softlockup_touch_ts); /* touch timestamp */ static DEFINE_PER_CPU(unsigned long, softlockup_print_ts); /* print timestamp */ static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog); +static DEFINE_PER_CPU(bool, softlock_touch_sync); static int __read_mostly did_panic; int __read_mostly softlockup_thresh = 60; @@ -79,6 +80,12 @@ void touch_softlockup_watchdog(void) } EXPORT_SYMBOL(touch_softlockup_watchdog); +void touch_softlockup_watchdog_sync(void) +{ + __raw_get_cpu_var(softlock_touch_sync) = true; + __raw_get_cpu_var(softlockup_touch_ts) = 0; +} + void touch_all_softlockup_watchdogs(void) { int cpu; @@ -118,6 +125,14 @@ void softlockup_tick(void) } if (touch_ts == 0) { + if (unlikely(per_cpu(softlock_touch_sync, this_cpu))) { + /* + * If the time stamp was touched atomically + * make sure the scheduler tick is up to date. + */ + per_cpu(softlock_touch_sync, this_cpu) = false; + sched_clock_tick(); + } __touch_softlockup_watchdog(); return; } -- cgit v1.2.2 From 61ef2489dbf587258526cfd4ebf4bba3b079f401 Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Fri, 22 Jan 2010 16:16:19 +0800 Subject: resources: introduce generic page_is_ram() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It's based on walk_system_ram_range(), for archs that don't have their own page_is_ram(). The static verions in MIPS and SCORE are also made global. v4: prefer plain 1 instead of PAGE_IS_RAM (H. Peter Anvin) v3: add comment (KAMEZAWA Hiroyuki) "AFAIK, this "System RAM" information has been used for kdump to grab valid memory area and seems good for the kernel itself." v2: add PAGE_IS_RAM macro (Américo Wang) Cc: Chen Liqin Cc: Lennox Wu Cc: Américo Wang Cc: linux-mips@linux-mips.org Cc: Yinghai Lu Acked-by: Ralf Baechle Reviewed-by: KAMEZAWA Hiroyuki Signed-off-by: Wu Fengguang LKML-Reference: <20100122081619.GA6431@localhost> Cc: Andrew Morton Signed-off-by: H. Peter Anvin --- kernel/resource.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index fb11a58b9594..b4d637a55256 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -297,6 +297,19 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, #endif +static int __is_ram(unsigned long pfn, unsigned long nr_pages, void *arg) +{ + return 1; +} +/* + * This generic page_is_ram() returns true if specified address is + * registered as "System RAM" in iomem_resource list. + */ +int __attribute__((weak)) page_is_ram(unsigned long pfn) +{ + return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; +} + /* * Find empty slot in the resource tree given range and alignment. */ -- cgit v1.2.2 From e52730071567ec5b6f57e21d6693b112e01e1d0e Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 26 Jan 2010 16:31:19 -0800 Subject: Generic page_is_ram: use __weak Use __weak instead of __attribute__((weak)). Cc: Wu Fengguang Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin --- kernel/resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index b4d637a55256..e68cd7477c40 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -305,7 +305,7 @@ static int __is_ram(unsigned long pfn, unsigned long nr_pages, void *arg) * This generic page_is_ram() returns true if specified address is * registered as "System RAM" in iomem_resource list. */ -int __attribute__((weak)) page_is_ram(unsigned long pfn) +int __weak page_is_ram(unsigned long pfn) { return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; } -- cgit v1.2.2 From 90fdbdb48442a03c72cae5463e6edb64cb3a3a7d Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Sun, 31 Jan 2010 20:53:24 +0900 Subject: sched: Use for_each_bit No change in functionality. Signed-off-by: Akinobu Mita Cc: Peter Zijlstra Cc: Andrew Morton LKML-Reference: <1264938810-4173-1-git-send-email-akinobu.mita@gmail.com> Signed-off-by: Ingo Molnar --- kernel/sched_cpupri.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index 597b33099dfa..eeb3506c4834 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -47,9 +47,7 @@ static int convert_prio(int prio) } #define for_each_cpupri_active(array, idx) \ - for (idx = find_first_bit(array, CPUPRI_NR_PRIORITIES); \ - idx < CPUPRI_NR_PRIORITIES; \ - idx = find_next_bit(array, CPUPRI_NR_PRIORITIES, idx+1)) + for_each_bit(idx, array, CPUPRI_NR_PRIORITIES) /** * cpupri_find - find the best (lowest-pri) CPU in the system -- cgit v1.2.2 From 4a461c85b643258e305eb5a3aced34009db2f818 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 1 Feb 2010 11:13:39 +0100 Subject: sched: Remove unused update_shares_locked() Commit f492e12ef050e02bf0185b6b57874992591b9be1 ("sched: Remove load_balance_newidle()") removed the only user of this function, so remove it too. Reported-by: Stephen Rothwell Signed-off-by: Peter Zijlstra LKML-Reference: <1265019219.24455.128.camel@laptop> Signed-off-by: Ingo Molnar --- kernel/sched.c | 14 -------------- 1 file changed, 14 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index a56ead42d651..2386f5440002 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1633,16 +1633,6 @@ static void update_shares(struct sched_domain *sd) } } -static void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ - if (root_task_group_empty()) - return; - - raw_spin_unlock(&rq->lock); - update_shares(sd); - raw_spin_lock(&rq->lock); -} - static void update_h_load(long cpu) { if (root_task_group_empty()) @@ -1657,10 +1647,6 @@ static inline void update_shares(struct sched_domain *sd) { } -static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ -} - #endif #ifdef CONFIG_PREEMPT -- cgit v1.2.2 From 4f48f8b7fd18c44f8478174f9925cc3c059c6ce4 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 2 Feb 2010 15:32:09 +0800 Subject: tracing: Fix circular dead lock in stack trace When we cat /tracing/stack_trace, we may cause circular lock: sys_read() t_start() arch_spin_lock(&max_stack_lock); t_show() seq_printf(), vsnprintf() .... /* they are all trace-able, when they are traced, max_stack_lock may be required again. */ The following script can trigger this circular dead lock very easy: #!/bin/bash echo 1 > /proc/sys/kernel/stack_tracer_enabled mount -t debugfs xxx /mnt > /dev/null 2>&1 ( # make check_stack() zealous to require max_stack_lock for ((; ;)) { echo 1 > /mnt/tracing/stack_max_size } ) & for ((; ;)) { cat /mnt/tracing/stack_trace > /dev/null } To fix this bug, we increase the percpu trace_active before require the lock. Reported-by: Li Zefan Signed-off-by: Lai Jiangshan LKML-Reference: <4B67D4F9.9080905@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_stack.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index 678a5120ee30..f4bc9b27de5f 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -157,6 +157,7 @@ stack_max_size_write(struct file *filp, const char __user *ubuf, unsigned long val, flags; char buf[64]; int ret; + int cpu; if (count >= sizeof(buf)) return -EINVAL; @@ -171,9 +172,20 @@ stack_max_size_write(struct file *filp, const char __user *ubuf, return ret; local_irq_save(flags); + + /* + * In case we trace inside arch_spin_lock() or after (NMI), + * we will cause circular lock, so we also need to increase + * the percpu trace_active here. + */ + cpu = smp_processor_id(); + per_cpu(trace_active, cpu)++; + arch_spin_lock(&max_stack_lock); *ptr = val; arch_spin_unlock(&max_stack_lock); + + per_cpu(trace_active, cpu)--; local_irq_restore(flags); return count; @@ -206,7 +218,13 @@ t_next(struct seq_file *m, void *v, loff_t *pos) static void *t_start(struct seq_file *m, loff_t *pos) { + int cpu; + local_irq_disable(); + + cpu = smp_processor_id(); + per_cpu(trace_active, cpu)++; + arch_spin_lock(&max_stack_lock); if (*pos == 0) @@ -217,7 +235,13 @@ static void *t_start(struct seq_file *m, loff_t *pos) static void t_stop(struct seq_file *m, void *p) { + int cpu; + arch_spin_unlock(&max_stack_lock); + + cpu = smp_processor_id(); + per_cpu(trace_active, cpu)--; + local_irq_enable(); } -- cgit v1.2.2 From b8a1d37c5f981cdd2e83c9fd98198832324cd57a Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Wed, 3 Feb 2010 09:31:36 +1100 Subject: kernel/cred.c: use kmem_cache_free Free memory allocated using kmem_cache_zalloc using kmem_cache_free rather than kfree. The semantic patch that makes this change is as follows: (http://coccinelle.lip6.fr/) // @@ expression x,E,c; @@ x = \(kmem_cache_alloc\|kmem_cache_zalloc\|kmem_cache_alloc_node\)(c,...) ... when != x = E when != &x ?-kfree(x) +kmem_cache_free(c,x) // Signed-off-by: Julia Lawall Acked-by: David Howells Cc: James Morris Cc: Steve Dickson Cc: Signed-off-by: Andrew Morton Signed-off-by: James Morris --- kernel/cred.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cred.c b/kernel/cred.c index dd76cfe5f5b0..1ed8ca18790c 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -224,7 +224,7 @@ struct cred *cred_alloc_blank(void) #ifdef CONFIG_KEYS new->tgcred = kzalloc(sizeof(*new->tgcred), GFP_KERNEL); if (!new->tgcred) { - kfree(new); + kmem_cache_free(cred_jar, new); return NULL; } atomic_set(&new->tgcred->usage, 1); -- cgit v1.2.2 From bc173f7092c76a7967f135c2b3a54052ad99733b Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 2 Feb 2010 13:44:01 -0800 Subject: kfifo: fix kernel-doc notation Fix kfifo kernel-doc warnings: Warning(kernel/kfifo.c:361): No description found for parameter 'total' Warning(kernel/kfifo.c:402): bad line: @ @lenout: pointer to output variable with copied data Warning(kernel/kfifo.c:412): No description found for parameter 'lenout' Signed-off-by: Randy Dunlap Cc: Stefani Seibold Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kfifo.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kfifo.c b/kernel/kfifo.c index 32c5c15d750d..498cabba225e 100644 --- a/kernel/kfifo.c +++ b/kernel/kfifo.c @@ -349,6 +349,7 @@ EXPORT_SYMBOL(__kfifo_from_user_n); * @fifo: the fifo to be used. * @from: pointer to the data to be added. * @len: the length of the data to be added. + * @total: the actual returned data length. * * This function copies at most @len bytes from the @from into the * FIFO depending and returns -EFAULT/0. @@ -399,7 +400,7 @@ EXPORT_SYMBOL(__kfifo_to_user_n); * @fifo: the fifo to be used. * @to: where the data must be copied. * @len: the size of the destination buffer. - @ @lenout: pointer to output variable with copied data + * @lenout: pointer to output variable with copied data * * This function copies at most @len bytes from the FIFO into the * @to buffer and 0 or -EFAULT. -- cgit v1.2.2 From 4528fd0595847c2078b59f24800e751c2d6b7e41 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 2 Feb 2010 13:44:10 -0800 Subject: cgroups: fix to return errno in a failure path In cgroup_create(), if alloc_css_id() returns failure, the errno is not propagated to userspace, so mkdir will fail silently. To trigger this bug, we mount blkio (or memory subsystem), and create more then 65534 cgroups. (The number of cgroups is limited to 65535 if a subsystem has use_id == 1) # mount -t cgroup -o blkio xxx /mnt # for ((i = 0; i < 65534; i++)); do mkdir /mnt/$i; done # mkdir /mnt/65534 (should return ENOSPC) # Signed-off-by: Li Zefan Acked-by: Serge Hallyn Acked-by: Paul Menage Acked-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1fbcc748044a..aa3bee566446 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2936,14 +2936,17 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, for_each_subsys(root, ss) { struct cgroup_subsys_state *css = ss->create(ss, cgrp); + if (IS_ERR(css)) { err = PTR_ERR(css); goto err_destroy; } init_cgroup_css(css, ss, cgrp); - if (ss->use_id) - if (alloc_css_id(ss, parent, cgrp)) + if (ss->use_id) { + err = alloc_css_id(ss, parent, cgrp); + if (err) goto err_destroy; + } /* At error, ->destroy() callback has to free assigned ID. */ } -- cgit v1.2.2 From 5ecb01cfdf96c5f465192bdb2a4fd4a61a24c6cc Mon Sep 17 00:00:00 2001 From: Mikael Pettersson Date: Sat, 23 Jan 2010 22:36:29 +0100 Subject: futex_lock_pi() key refcnt fix This fixes a futex key reference count bug in futex_lock_pi(), where a key's reference count is incremented twice but decremented only once, causing the backing object to not be released. If the futex is created in a temporary file in an ext3 file system, this bug causes the file's inode to become an "undead" orphan, which causes an oops from a BUG_ON() in ext3_put_super() when the file system is unmounted. glibc's test suite is known to trigger this, see . The bug is a regression from 2.6.28-git3, namely Peter Zijlstra's 38d47c1b7075bd7ec3881141bb3629da58f88dab "[PATCH] futex: rely on get_user_pages() for shared futexes". That commit made get_futex_key() also increment the reference count of the futex key, and updated its callers to decrement the key's reference count before returning. Unfortunately the normal exit path in futex_lock_pi() wasn't corrected: the reference count is incremented by get_futex_key() and queue_lock(), but the normal exit path only decrements once, via unqueue_me_pi(). The fix is to put_futex_key() after unqueue_me_pi(), since 2.6.31 this is easily done by 'goto out_put_key' rather than 'goto out'. Signed-off-by: Mikael Pettersson Acked-by: Peter Zijlstra Acked-by: Darren Hart Signed-off-by: Thomas Gleixner Cc: --- kernel/futex.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index d9b3a2228f9d..17828033a639 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -1971,7 +1971,7 @@ retry_private: /* Unqueue and drop the lock */ unqueue_me_pi(&q); - goto out; + goto out_put_key; out_unlock_put_key: queue_unlock(&q, hb); -- cgit v1.2.2 From 51246bfd189064079c54421507236fd2723b18f3 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 2 Feb 2010 11:40:27 +0100 Subject: futex: Handle user space corruption gracefully If the owner of a PI futex dies we fix up the pi_state and set pi_state->owner to NULL. When a malicious or just sloppy programmed user space application sets the futex value to 0 e.g. by calling pthread_mutex_init(), then the futex can be acquired again. A new waiter manages to enqueue itself on the pi_state w/o damage, but on unlock the kernel dereferences pi_state->owner and oopses. Prevent this by checking pi_state->owner in the unlock path. If pi_state->owner is not current we know that user space manipulated the futex value. Ignore the mess and return -EINVAL. This catches the above case and also the case where a task hijacks the futex by setting the tid value and then tries to unlock it. Reported-by: Jermome Marchand Signed-off-by: Thomas Gleixner Acked-by: Darren Hart Acked-by: Peter Zijlstra Cc: --- kernel/futex.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index 17828033a639..06e8240d2abe 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -758,6 +758,13 @@ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this) if (!pi_state) return -EINVAL; + /* + * If current does not own the pi_state then the futex is + * inconsistent and user space fiddled with the futex value. + */ + if (pi_state->owner != current) + return -EINVAL; + raw_spin_lock(&pi_state->pi_mutex.wait_lock); new_owner = rt_mutex_next_owner(&pi_state->pi_mutex); -- cgit v1.2.2 From 59647b6ac3050dd964bc556fe6ef22f4db5b935c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 3 Feb 2010 09:33:05 +0100 Subject: futex: Handle futex value corruption gracefully The WARN_ON in lookup_pi_state which complains about a mismatch between pi_state->owner->pid and the pid which we retrieved from the user space futex is completely bogus. The code just emits the warning and then continues despite the fact that it detected an inconsistent state of the futex. A conveniant way for user space to spam the syslog. Replace the WARN_ON by a consistency check. If the values do not match return -EINVAL and let user space deal with the mess it created. This also fixes the missing task_pid_vnr() when we compare the pi_state->owner pid with the futex value. Reported-by: Jermome Marchand Signed-off-by: Thomas Gleixner Acked-by: Darren Hart Acked-by: Peter Zijlstra Cc: --- kernel/futex.c | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index 06e8240d2abe..e7a35f1039e7 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -530,8 +530,25 @@ lookup_pi_state(u32 uval, struct futex_hash_bucket *hb, return -EINVAL; WARN_ON(!atomic_read(&pi_state->refcount)); - WARN_ON(pid && pi_state->owner && - pi_state->owner->pid != pid); + + /* + * When pi_state->owner is NULL then the owner died + * and another waiter is on the fly. pi_state->owner + * is fixed up by the task which acquires + * pi_state->rt_mutex. + * + * We do not check for pid == 0 which can happen when + * the owner died and robust_list_exit() cleared the + * TID. + */ + if (pid && pi_state->owner) { + /* + * Bail out if user space manipulated the + * futex value. + */ + if (pid != task_pid_vnr(pi_state->owner)) + return -EINVAL; + } atomic_inc(&pi_state->refcount); *ps = pi_state; -- cgit v1.2.2 From b9c3032277f756e73f6c673419dc414155e04e46 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 3 Feb 2010 18:08:52 +0100 Subject: hrtimer, softirq: Fix hrtimer->softirq trampoline hrtimers callbacks are always done from hardirq context, either the jiffy tick interrupt or the hrtimer device interrupt. [ there is currently one exception that can still call a hrtimer callback from softirq, but even in that case this will still work correctly. ] Reported-by: Wei Yongjun Signed-off-by: Peter Zijlstra Cc: Yury Polyanskiy Tested-by: Wei Yongjun Acked-by: David S. Miller LKML-Reference: <1265120401.24455.306.camel@laptop> Signed-off-by: Thomas Gleixner --- kernel/softirq.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/softirq.c b/kernel/softirq.c index a09502e2ef75..7c1a67ef0274 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -500,22 +500,17 @@ EXPORT_SYMBOL(tasklet_kill); */ /* - * The trampoline is called when the hrtimer expires. If this is - * called from the hrtimer interrupt then we schedule the tasklet as - * the timer callback function expects to run in softirq context. If - * it's called in softirq context anyway (i.e. high resolution timers - * disabled) then the hrtimer callback is called right away. + * The trampoline is called when the hrtimer expires. It schedules a tasklet + * to run __tasklet_hrtimer_trampoline() which in turn will call the intended + * hrtimer callback, but from softirq context. */ static enum hrtimer_restart __hrtimer_tasklet_trampoline(struct hrtimer *timer) { struct tasklet_hrtimer *ttimer = container_of(timer, struct tasklet_hrtimer, timer); - if (hrtimer_is_hres_active(timer)) { - tasklet_hi_schedule(&ttimer->tasklet); - return HRTIMER_NORESTART; - } - return ttimer->function(timer); + tasklet_hi_schedule(&ttimer->tasklet); + return HRTIMER_NORESTART; } /* -- cgit v1.2.2 From cd757645fbdc34a8343c04bb0e74e06fccc2cb10 Mon Sep 17 00:00:00 2001 From: Mahesh Salgaonkar Date: Sat, 30 Jan 2010 10:25:18 +0530 Subject: perf: Make bp_len type to u64 generic across the arch Change 'bp_len' type to __u64 to make it work across archs as the s390 architecture watch point length can be upto 2^64. reference: http://lkml.org/lkml/2010/1/25/212 This is an ABI change that is not backward compatible with the previous hardware breakpoint info layout integrated in this development cycle, a rebuilt of perf tools is necessary for versions based on 2.6.33-rc1 - 2.6.33-rc6 to work with a kernel based on this patch. Signed-off-by: Mahesh Salgaonkar Acked-by: Peter Zijlstra Cc: Ananth N Mavinakayanahalli Cc: "K. Prasad" Cc: Maneesh Soni Cc: Heiko Carstens Cc: Martin LKML-Reference: <20100130045518.GA20776@in.ibm.com> Signed-off-by: Frederic Weisbecker --- kernel/hw_breakpoint.c | 2 +- kernel/perf_event.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 8a5c7d55ac9f..967e66143e11 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -360,8 +360,8 @@ EXPORT_SYMBOL_GPL(register_user_hw_breakpoint); int modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *attr) { u64 old_addr = bp->attr.bp_addr; + u64 old_len = bp->attr.bp_len; int old_type = bp->attr.bp_type; - int old_len = bp->attr.bp_len; int err = 0; perf_event_disable(bp); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index d27746bd3a06..2b19297742cb 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4580,7 +4580,7 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, if (attr->type >= PERF_TYPE_MAX) return -EINVAL; - if (attr->__reserved_1 || attr->__reserved_2) + if (attr->__reserved_1) return -EINVAL; if (attr->sample_type & ~(PERF_SAMPLE_MAX-1)) -- cgit v1.2.2 From 002345925e6c45861f60db6f4fc6236713fd8847 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 3 Feb 2010 15:36:43 -0800 Subject: syslog: distinguish between /proc/kmsg and syscalls This allows the LSM to distinguish between syslog functions originating from /proc/kmsg access and direct syscalls. By default, the commoncaps will now no longer require CAP_SYS_ADMIN to read an opened /proc/kmsg file descriptor. For example the kernel syslog reader can now drop privileges after opening /proc/kmsg, instead of staying privileged with CAP_SYS_ADMIN. MAC systems that implement security_syslog have unchanged behavior. Signed-off-by: Kees Cook Acked-by: Serge Hallyn Acked-by: John Johansen Signed-off-by: James Morris --- kernel/printk.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 17463ca2e229..809cf9a258a0 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -35,6 +35,7 @@ #include #include #include +#include #include @@ -273,14 +274,14 @@ static inline void boot_delay_msec(void) * 9 -- Return number of unread characters in the log buffer * 10 -- Return size of the log buffer */ -int do_syslog(int type, char __user *buf, int len) +int do_syslog(int type, char __user *buf, int len, bool from_file) { unsigned i, j, limit, count; int do_clear = 0; char c; int error = 0; - error = security_syslog(type); + error = security_syslog(type, from_file); if (error) return error; @@ -417,7 +418,7 @@ out: SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) { - return do_syslog(type, buf, len); + return do_syslog(type, buf, len, SYSLOG_FROM_CALL); } /* -- cgit v1.2.2 From d78ca3cd733d8a2c3dcd88471beb1a15d973eed8 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 3 Feb 2010 15:37:13 -0800 Subject: syslog: use defined constants instead of raw numbers Right now the syslog "type" action are just raw numbers which makes the source difficult to follow. This patch replaces the raw numbers with defined constants for some level of sanity. Signed-off-by: Kees Cook Acked-by: John Johansen Acked-by: Serge Hallyn Signed-off-by: James Morris --- kernel/printk.c | 45 +++++++++++++++++++-------------------------- 1 file changed, 19 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 809cf9a258a0..3e162d867098 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -259,21 +259,6 @@ static inline void boot_delay_msec(void) } #endif -/* - * Commands to do_syslog: - * - * 0 -- Close the log. Currently a NOP. - * 1 -- Open the log. Currently a NOP. - * 2 -- Read from the log. - * 3 -- Read all messages remaining in the ring buffer. - * 4 -- Read and clear all messages remaining in the ring buffer - * 5 -- Clear ring buffer. - * 6 -- Disable printk's to console - * 7 -- Enable printk's to console - * 8 -- Set level of messages printed to console - * 9 -- Return number of unread characters in the log buffer - * 10 -- Return size of the log buffer - */ int do_syslog(int type, char __user *buf, int len, bool from_file) { unsigned i, j, limit, count; @@ -286,11 +271,11 @@ int do_syslog(int type, char __user *buf, int len, bool from_file) return error; switch (type) { - case 0: /* Close log */ + case SYSLOG_ACTION_CLOSE: /* Close log */ break; - case 1: /* Open log */ + case SYSLOG_ACTION_OPEN: /* Open log */ break; - case 2: /* Read from log */ + case SYSLOG_ACTION_READ: /* Read from log */ error = -EINVAL; if (!buf || len < 0) goto out; @@ -321,10 +306,12 @@ int do_syslog(int type, char __user *buf, int len, bool from_file) if (!error) error = i; break; - case 4: /* Read/clear last kernel messages */ + /* Read/clear last kernel messages */ + case SYSLOG_ACTION_READ_CLEAR: do_clear = 1; /* FALL THRU */ - case 3: /* Read last kernel messages */ + /* Read last kernel messages */ + case SYSLOG_ACTION_READ_ALL: error = -EINVAL; if (!buf || len < 0) goto out; @@ -377,21 +364,25 @@ int do_syslog(int type, char __user *buf, int len, bool from_file) } } break; - case 5: /* Clear ring buffer */ + /* Clear ring buffer */ + case SYSLOG_ACTION_CLEAR: logged_chars = 0; break; - case 6: /* Disable logging to console */ + /* Disable logging to console */ + case SYSLOG_ACTION_CONSOLE_OFF: if (saved_console_loglevel == -1) saved_console_loglevel = console_loglevel; console_loglevel = minimum_console_loglevel; break; - case 7: /* Enable logging to console */ + /* Enable logging to console */ + case SYSLOG_ACTION_CONSOLE_ON: if (saved_console_loglevel != -1) { console_loglevel = saved_console_loglevel; saved_console_loglevel = -1; } break; - case 8: /* Set level of messages printed to console */ + /* Set level of messages printed to console */ + case SYSLOG_ACTION_CONSOLE_LEVEL: error = -EINVAL; if (len < 1 || len > 8) goto out; @@ -402,10 +393,12 @@ int do_syslog(int type, char __user *buf, int len, bool from_file) saved_console_loglevel = -1; error = 0; break; - case 9: /* Number of chars in the log buffer */ + /* Number of chars in the log buffer */ + case SYSLOG_ACTION_SIZE_UNREAD: error = log_end - log_start; break; - case 10: /* Size of the log buffer */ + /* Size of the log buffer */ + case SYSLOG_ACTION_SIZE_BUFFER: error = log_buf_len; break; default: -- cgit v1.2.2 From 74b7eb5885415ed41d012f432398d1b697115b5f Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Fri, 29 Jan 2010 14:57:52 +0800 Subject: sched: Change usage of rt_rq->rt_se to rt_rq->tg->rt_se[cpu] This is the first step to remove rt_rq member rt_se because it have the same meaning with tg->rt_se[cpu]. And the latter style is also used by the fair scheduling class. Signed-off-by: Yong Zhang Cc: Rusty Russell Signed-off-by: Peter Zijlstra LKML-Reference: <2674af741001282257r28c97a92o9f90cf16fe8d3d84@mail.gmail.com> Signed-off-by: Ingo Molnar --- kernel/sched_rt.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index ca49ceb01201..bf3e38fdbe6d 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -199,8 +199,11 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se); static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) { + int this_cpu = smp_processor_id(); struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr; - struct sched_rt_entity *rt_se = rt_rq->rt_se; + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) @@ -212,7 +215,10 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) static void sched_rt_rq_dequeue(struct rt_rq *rt_rq) { - struct sched_rt_entity *rt_se = rt_rq->rt_se; + int this_cpu = smp_processor_id(); + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_se && on_rt_rq(rt_se)) dequeue_rt_entity(rt_se); -- cgit v1.2.2 From 23577256953c870de9b724c3a2611ce7be6a1e4e Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Fri, 29 Jan 2010 14:58:47 +0800 Subject: sched: Remove member rt_se from struct rt_rq It's a duplicate of tg->rt_se[cpu] and the only usage is sched_rt_rq_dequeue() and sched_rt_rq_enqueue(). After the first patch to those two function. rt_se can be removed. Signed-off-by: Yong Zhang Cc: Rusty Russell Signed-off-by: Peter Zijlstra LKML-Reference: <2674af741001282258q38781619u653ca4a7dd267347@mail.gmail.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 2386f5440002..6cee227b1459 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -436,7 +436,6 @@ struct rt_rq { struct rq *rq; struct list_head leaf_rt_rq_list; struct task_group *tg; - struct sched_rt_entity *rt_se; #endif }; @@ -7592,7 +7591,6 @@ static void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq, tg->rt_rq[cpu] = rt_rq; init_rt_rq(rt_rq, rq); rt_rq->tg = tg; - rt_rq->rt_se = rt_se; rt_rq->rt_runtime = tg->rt_bandwidth.rt_runtime; if (add) list_add(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list); -- cgit v1.2.2 From c41b20e721ea4f6f20f66a66e7f0c3c97a2ca9c2 Mon Sep 17 00:00:00 2001 From: Adam Buchbinder Date: Fri, 11 Dec 2009 16:35:39 -0500 Subject: Fix misspellings of "truly" in comments. Some comments misspell "truly"; this fixes them. No code changes. Signed-off-by: Adam Buchbinder Signed-off-by: Jiri Kosina --- kernel/trace/ring_buffer.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 8c1b2d290718..9ab578f1bb65 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2541,7 +2541,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_disable); * @buffer: The ring buffer to enable writes * * Note, multiple disables will need the same number of enables - * to truely enable the writing (much like preempt_disable). + * to truly enable the writing (much like preempt_disable). */ void ring_buffer_record_enable(struct ring_buffer *buffer) { @@ -2577,7 +2577,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_disable_cpu); * @cpu: The CPU to enable. * * Note, multiple disables will need the same number of enables - * to truely enable the writing (much like preempt_disable). + * to truly enable the writing (much like preempt_disable). */ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu) { -- cgit v1.2.2 From 2a61aa401638529cd4231f6106980d307fba98fa Mon Sep 17 00:00:00 2001 From: Adam Buchbinder Date: Fri, 11 Dec 2009 16:35:40 -0500 Subject: Fix misspellings of "invocation" in comments. Some comments misspell "invocation"; this fixes them. No code changes. Signed-off-by: Adam Buchbinder Signed-off-by: Jiri Kosina --- kernel/sched_cpupri.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index 597b33099dfa..3db4b1a0e921 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -58,7 +58,7 @@ static int convert_prio(int prio) * @lowest_mask: A mask to fill in with selected CPUs (or NULL) * * Note: This function returns the recommended CPUs as calculated during the - * current invokation. By the time the call returns, the CPUs may have in + * current invocation. By the time the call returns, the CPUs may have in * fact changed priorities any number of times. While not ideal, it is not * an issue of correctness since the normal rebalancer logic will correct * any discrepancies created by racing against the uncertainty of the current -- cgit v1.2.2 From c9404c9c392d557a4687c4cbda022b03cb787ce9 Mon Sep 17 00:00:00 2001 From: Adam Buchbinder Date: Fri, 18 Dec 2009 15:40:42 -0500 Subject: Fix misspelling of "should" and "shouldn't" in comments. Some comments misspell "should" or "shouldn't"; this fixes them. No code changes. Signed-off-by: Adam Buchbinder Signed-off-by: Jiri Kosina --- kernel/audit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/audit.c b/kernel/audit.c index 5feed232be9d..78f7f86aa238 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -398,7 +398,7 @@ static void kauditd_send_skb(struct sk_buff *skb) skb_get(skb); err = netlink_unicast(audit_sock, skb, audit_nlk_pid, 0); if (err < 0) { - BUG_ON(err != -ECONNREFUSED); /* Shoudn't happen */ + BUG_ON(err != -ECONNREFUSED); /* Shouldn't happen */ printk(KERN_ERR "audit: *NO* daemon at audit_pid=%d\n", audit_pid); audit_log_lost("auditd dissapeared\n"); audit_pid = 0; -- cgit v1.2.2 From af66585270ef99aa6097faf3bd7344855077e75d Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Sun, 17 Jan 2010 19:14:26 -0200 Subject: fix comment typo boo -> boot in ksysfs.c Signed-off-by: Thadeu Lima de Souza Cascardo Signed-off-by: Jiri Kosina --- kernel/ksysfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c index 3feaf5a74514..ac08efca54c3 100644 --- a/kernel/ksysfs.c +++ b/kernel/ksysfs.c @@ -33,7 +33,7 @@ static ssize_t uevent_seqnum_show(struct kobject *kobj, } KERNEL_ATTR_RO(uevent_seqnum); -/* uevent helper program, used during early boo */ +/* uevent helper program, used during early boot */ static ssize_t uevent_helper_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { -- cgit v1.2.2 From 350f82586b7554240bee18c41cc5c842f63265ae Mon Sep 17 00:00:00 2001 From: "Edward Z. Yang" Date: Mon, 1 Feb 2010 18:26:59 -0500 Subject: Remove redundant trailing semicolons from macros Signed-off-by: Edward Z. Yang Acked-by: Rusty Russell Signed-off-by: Jiri Kosina --- kernel/params.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index cf1b69183127..2278ce244cf8 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -402,8 +402,8 @@ int param_get_string(char *buffer, struct kernel_param *kp) } /* sysfs output in /sys/modules/XYZ/parameters/ */ -#define to_module_attr(n) container_of(n, struct module_attribute, attr); -#define to_module_kobject(n) container_of(n, struct module_kobject, kobj); +#define to_module_attr(n) container_of(n, struct module_attribute, attr) +#define to_module_kobject(n) container_of(n, struct module_kobject, kobj) extern struct kernel_param __start___param[], __stop___param[]; @@ -421,7 +421,7 @@ struct module_param_attrs }; #ifdef CONFIG_SYSFS -#define to_param_attr(n) container_of(n, struct param_attribute, mattr); +#define to_param_attr(n) container_of(n, struct param_attribute, mattr) static ssize_t param_attr_show(struct module_attribute *mattr, struct module *mod, char *buf) -- cgit v1.2.2 From 9ce8e498ee58bb8a866a6c3c08fcb385ed66e9d2 Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Tue, 2 Feb 2010 08:54:51 +0200 Subject: devres: typo fix s/dev/devm/ Signed-off-by: Baruch Siach Signed-off-by: Jiri Kosina --- kernel/irq/devres.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/irq/devres.c b/kernel/irq/devres.c index d06df9c41cba..30d56bafc9c2 100644 --- a/kernel/irq/devres.c +++ b/kernel/irq/devres.c @@ -81,7 +81,7 @@ EXPORT_SYMBOL(devm_request_threaded_irq); * Except for the extra @dev argument, this function takes the * same arguments and performs the same function as free_irq(). * This function instead of free_irq() should be used to manually - * free IRQs allocated with dev_request_irq(). + * free IRQs allocated with devm_request_irq(). */ void devm_free_irq(struct device *dev, unsigned int irq, void *dev_id) { -- cgit v1.2.2 From 1537a3638cbf741d3826c1002026cce487a6bee0 Mon Sep 17 00:00:00 2001 From: Daniel Mack Date: Fri, 29 Jan 2010 15:57:49 +0800 Subject: tree-wide: fix 'lenght' typo in comments and code Some misspelled occurences of 'octet' and some comments were also fixed as I was on it. Signed-off-by: Daniel Mack Cc: Jiri Kosina Cc: Joe Perches Cc: Junio C Hamano Signed-off-by: Jiri Kosina --- kernel/trace/trace.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4df6a77eb196..e4b32c8aa85f 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -549,7 +549,7 @@ static inline int ftrace_trace_task(struct task_struct *task) * struct trace_parser - servers for reading the user input separated by spaces * @cont: set if the input is not complete - no final space char was found * @buffer: holds the parsed user input - * @idx: user input lenght + * @idx: user input length * @size: buffer size */ struct trace_parser { -- cgit v1.2.2 From 17622339af2536b32cf29699ddd4ba0fe79a61d5 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 2 Feb 2010 14:41:39 -0800 Subject: clocksource: add argument to resume callback Pass the clocksource as an argument to the clocksource resume callback. Needed so we can point out which CMT channel the sh_cmt.c driver shall resume. Signed-off-by: Magnus Damm Cc: john stultz Cc: Paul Mundt Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index e85c23404d34..08adacb2a1ed 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -449,7 +449,7 @@ void clocksource_resume(void) list_for_each_entry(cs, &clocksource_list, list) if (cs->resume) - cs->resume(); + cs->resume(cs); clocksource_resume_watchdog(); } -- cgit v1.2.2 From c54a42b19fbaae4e9f212322ecca25a6bc95c1ba Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 2 Feb 2010 14:41:41 -0800 Subject: clocksource: add suspend callback Add a clocksource suspend callback. This callback can be used by the clocksource driver to shutdown and perform any kind of late suspend activities even though the clocksource driver itself is a non-sysdev driver. One example where this is useful is to fix the sh_cmt.c platform driver that today suspends using the platform bus and shuts down the clocksource too early. With this callback in place the sh_cmt driver will suspend using the clocksource and clockevent hooks and leave the platform device pm callbacks unused. Signed-off-by: Magnus Damm Cc: Paul Mundt Cc: john stultz Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 12 ++++++++++++ kernel/time/timekeeping.c | 1 + 2 files changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 08adacb2a1ed..bd246660902c 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -440,6 +440,18 @@ static inline int clocksource_watchdog_kthread(void *data) { return 0; } #endif /* CONFIG_CLOCKSOURCE_WATCHDOG */ +/** + * clocksource_suspend - suspend the clocksource(s) + */ +void clocksource_suspend(void) +{ + struct clocksource *cs; + + list_for_each_entry_reverse(cs, &clocksource_list, list) + if (cs->suspend) + cs->suspend(cs); +} + /** * clocksource_resume - resume the clocksource(s) */ diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 7faaa32fbf4f..843d8a711b16 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -622,6 +622,7 @@ static int timekeeping_suspend(struct sys_device *dev, pm_message_t state) write_sequnlock_irqrestore(&xtime_lock, flags); clockevents_notify(CLOCK_EVT_NOTIFY_SUSPEND, NULL); + clocksource_suspend(); return 0; } -- cgit v1.2.2 From 6622e670b26fbaa9c4ae7772a4d2d0abd3414b51 Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Tue, 2 Feb 2010 14:41:42 -0800 Subject: posix-timers.c: Don't export local functions Signed-off-by: H Hartley Sweeten Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/posix-timers.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 495440779ce3..00d1fda58ab6 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -256,7 +256,7 @@ static int posix_get_monotonic_coarse(clockid_t which_clock, return 0; } -int posix_get_coarse_res(const clockid_t which_clock, struct timespec *tp) +static int posix_get_coarse_res(const clockid_t which_clock, struct timespec *tp) { *tp = ktime_to_timespec(KTIME_LOW_RES); return 0; -- cgit v1.2.2 From 50200df462023b187d80a99a52f5f2cfe3c86c26 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 2 Feb 2010 14:46:13 -0800 Subject: kernel/sched.c: Suppress unused var warning On UP: kernel/sched.c: In function 'wake_up_new_task': kernel/sched.c:2631: warning: unused variable 'cpu' Signed-off-by: Andrew Morton Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30a91b1..e3199df426e3 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2647,7 +2647,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; - int cpu = get_cpu(); + int cpu __maybe_unused = get_cpu(); #ifdef CONFIG_SMP /* -- cgit v1.2.2 From fa535a77bd3fa32b9215ba375d6a202fe73e1dd6 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 2 Feb 2010 14:46:13 -0800 Subject: sched: cpuacct: Use bigger percpu counter batch values for stats counters When CONFIG_VIRT_CPU_ACCOUNTING and CONFIG_CGROUP_CPUACCT are enabled we can call cpuacct_update_stats with values much larger than percpu_counter_batch. This means the call to percpu_counter_add will always add to the global count which is protected by a spinlock and we end up with a global spinlock in the scheduler. Based on an idea by KOSAKI Motohiro, this patch scales the batch value by cputime_one_jiffy such that we have the same batch limit as we would if CONFIG_VIRT_CPU_ACCOUNTING was disabled. His patch did this once at boot but that initialisation happened too early on PowerPC (before time_init) and it was never updated at runtime as a result of a hotplug cpu add/remove. This patch instead scales percpu_counter_batch by cputime_one_jiffy at runtime, which keeps the batch correct even after cpu hotplug operations. We cap it at INT_MAX in case of overflow. For architectures that do not support CONFIG_VIRT_CPU_ACCOUNTING, cputime_one_jiffy is the constant 1 and gcc is smart enough to optimise min(s32 percpu_counter_batch, INT_MAX) to just percpu_counter_batch at least on x86 and PowerPC. So there is no need to add an #ifdef. On a 64 thread PowerPC box with CONFIG_VIRT_CPU_ACCOUNTING and CONFIG_CGROUP_CPUACCT enabled, a context switch microbenchmark is 234x faster and almost matches a CONFIG_CGROUP_CPUACCT disabled kernel: CONFIG_CGROUP_CPUACCT disabled: 16906698 ctx switches/sec CONFIG_CGROUP_CPUACCT enabled: 61720 ctx switches/sec CONFIG_CGROUP_CPUACCT + patch: 16663217 ctx switches/sec Tested with: wget http://ozlabs.org/~anton/junkcode/context_switch.c make context_switch for i in `seq 0 63`; do taskset -c $i ./context_switch & done vmstat 1 Signed-off-by: Anton Blanchard Reviewed-by: KOSAKI Motohiro Acked-by: Balbir Singh Tested-by: Balbir Singh Cc: Peter Zijlstra Cc: Martin Schwidefsky Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/sched.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index f96be9370b75..bae6fcfe6d75 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -8997,6 +8997,23 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime) rcu_read_unlock(); } +/* + * When CONFIG_VIRT_CPU_ACCOUNTING is enabled one jiffy can be very large + * in cputime_t units. As a result, cpuacct_update_stats calls + * percpu_counter_add with values large enough to always overflow the + * per cpu batch limit causing bad SMP scalability. + * + * To fix this we scale percpu_counter_batch by cputime_one_jiffy so we + * batch the same amount of time with CONFIG_VIRT_CPU_ACCOUNTING disabled + * and enabled. We cap it at INT_MAX which is the largest allowed batch value. + */ +#ifdef CONFIG_SMP +#define CPUACCT_BATCH \ + min_t(long, percpu_counter_batch * cputime_one_jiffy, INT_MAX) +#else +#define CPUACCT_BATCH 0 +#endif + /* * Charge the system/user time to the task's accounting group. */ @@ -9004,6 +9021,7 @@ static void cpuacct_update_stats(struct task_struct *tsk, enum cpuacct_stat_index idx, cputime_t val) { struct cpuacct *ca; + int batch = CPUACCT_BATCH; if (unlikely(!cpuacct_subsys.active)) return; @@ -9012,7 +9030,7 @@ static void cpuacct_update_stats(struct task_struct *tsk, ca = task_ca(tsk); do { - percpu_counter_add(&ca->cpustat[idx], val); + __percpu_counter_add(&ca->cpustat[idx], val, batch); ca = ca->parent; } while (ca); rcu_read_unlock(); -- cgit v1.2.2 From cccc6bba3f771ef29b33e4f79e70ebc3dba245b0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 25 Dec 2009 05:07:33 -0500 Subject: Lose the first argument of audit_inode_child() it's always equal to ->d_name.name of the second argument Signed-off-by: Al Viro --- kernel/auditsc.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/auditsc.c b/kernel/auditsc.c index fc0f928167e7..f3a461c0970a 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -1988,7 +1988,6 @@ void __audit_inode(const char *name, const struct dentry *dentry) /** * audit_inode_child - collect inode info for created/removed objects - * @dname: inode's dentry name * @dentry: dentry being audited * @parent: inode of dentry parent * @@ -2000,13 +1999,14 @@ void __audit_inode(const char *name, const struct dentry *dentry) * must be hooked prior, in order to capture the target inode during * unsuccessful attempts. */ -void __audit_inode_child(const char *dname, const struct dentry *dentry, +void __audit_inode_child(const struct dentry *dentry, const struct inode *parent) { int idx; struct audit_context *context = current->audit_context; const char *found_parent = NULL, *found_child = NULL; const struct inode *inode = dentry->d_inode; + const char *dname = dentry->d_name.name; int dirlen = 0; if (!context->in_syscall) @@ -2014,9 +2014,6 @@ void __audit_inode_child(const char *dname, const struct dentry *dentry, if (inode) handle_one(inode); - /* determine matching parent */ - if (!dname) - goto add_names; /* parent is more likely, look for it first */ for (idx = 0; idx < context->name_count; idx++) { -- cgit v1.2.2 From b80109e256bc17ed66c9d559175f087b03ca2a8e Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 9 Feb 2010 15:07:40 +1100 Subject: Remove reference to kthread_create_on_cpu kthread_create_on_cpu doesn't exist so update a comment in kthread.c to reflect this. Signed-off-by: Anton Blanchard Acked-by: Rusty Russell Signed-off-by: Jiri Kosina --- kernel/kthread.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kthread.c b/kernel/kthread.c index fbb6222fe7e0..82ed0ea15194 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -101,7 +101,7 @@ static void create_kthread(struct kthread_create_info *create) * * Description: This helper function creates and names a kernel * thread. The thread will be stopped: use wake_up_process() to start - * it. See also kthread_run(), kthread_create_on_cpu(). + * it. See also kthread_run(). * * When woken, the thread will run @threadfn() with @data as its * argument. @threadfn() can either call do_exit() directly if it is a -- cgit v1.2.2 From 301ba0457f1ed853fc08e57785f8c87fe7e49c68 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 9 Feb 2010 15:07:40 +1100 Subject: kthread, sched: Remove reference to kthread_create_on_cpu kthread_create_on_cpu doesn't exist so update a comment in kthread.c to reflect this. Signed-off-by: Anton Blanchard Acked-by: Rusty Russell Cc: Peter Zijlstra LKML-Reference: <20100209040740.GB3702@kryten> Signed-off-by: Ingo Molnar --- kernel/kthread.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kthread.c b/kernel/kthread.c index fbb6222fe7e0..82ed0ea15194 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -101,7 +101,7 @@ static void create_kthread(struct kthread_create_info *create) * * Description: This helper function creates and names a kernel * thread. The thread will be stopped: use wake_up_process() to start - * it. See also kthread_run(), kthread_create_on_cpu(). + * it. See also kthread_run(). * * When woken, the thread will run @threadfn() with @data as its * argument. @threadfn() can either call do_exit() directly if it is a -- cgit v1.2.2 From c93d89f3dbf0202bf19c07960ca8602b48c2f9a0 Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Wed, 27 Jan 2010 19:13:40 +0800 Subject: Export the symbol of getboottime and mmonotonic_to_bootbased Export getboottime and monotonic_to_bootbased in order to let them could be used by following patch. Cc: stable@kernel.org Signed-off-by: Jason Wang Signed-off-by: Marcelo Tosatti --- kernel/time/timekeeping.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 7faaa32fbf4f..e2ab064c6d41 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -880,6 +880,7 @@ void getboottime(struct timespec *ts) set_normalized_timespec(ts, -boottime.tv_sec, -boottime.tv_nsec); } +EXPORT_SYMBOL_GPL(getboottime); /** * monotonic_to_bootbased - Convert the monotonic time to boot based. @@ -889,6 +890,7 @@ void monotonic_to_bootbased(struct timespec *ts) { *ts = timespec_add_safe(*ts, total_sleep_time); } +EXPORT_SYMBOL_GPL(monotonic_to_bootbased); unsigned long get_seconds(void) { -- cgit v1.2.2 From ede55c9d78101fef0d8e620940a5163f14b02f29 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 27 Jan 2010 11:25:54 -0500 Subject: tracing: Add correct/incorrect to sort keys for branch annotation output The branch annotation is a bit difficult to see the worst offenders because it only sorts by percentage: correct incorrect % Function File Line ------- --------- - -------- ---- ---- 0 163 100 qdisc_restart sch_generic.c 179 0 163 100 pfifo_fast_dequeue sch_generic.c 447 0 4 100 pskb_trim_rcsum skbuff.h 1689 0 4 100 llc_rcv llc_input.c 170 0 18 100 psmouse_interrupt psmouse-base.c 304 0 3 100 atkbd_interrupt atkbd.c 389 0 5 100 usb_alloc_dev usb.c 437 0 11 100 vsscanf vsprintf.c 1897 0 2 100 IS_ERR err.h 34 0 23 100 __rmqueue_fallback page_alloc.c 865 0 4 100 probe_wakeup_sched_switch trace_sched_wakeup.c 142 0 3 100 move_masked_irq migration.c 11 Adding the incorrect and correct values as sort keys makes this file a bit more informative: correct incorrect % Function File Line ------- --------- - -------- ---- ---- 0 366541 100 audit_syscall_entry auditsc.c 1637 0 366538 100 audit_syscall_exit auditsc.c 1685 0 115839 100 sched_info_switch sched_stats.h 269 0 74567 100 sched_info_queued sched_stats.h 222 0 66578 100 sched_info_dequeued sched_stats.h 177 0 15113 100 trace_workqueue_insertion workqueue.h 38 0 15107 100 trace_workqueue_execution workqueue.h 45 0 3622 100 syscall_trace_leave ptrace.c 1772 0 2750 100 sched_move_task sched.c 10100 0 2750 100 sched_move_task sched.c 10110 0 1815 100 pre_schedule_rt sched_rt.c 1462 0 837 100 audit_alloc auditsc.c 879 0 814 100 tcp_mss_split_point tcp_output.c 1302 Signed-off-by: Steven Rostedt --- kernel/trace/trace_branch.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 4a194f08f88c..b9bc4d470177 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -307,8 +307,23 @@ static int annotated_branch_stat_cmp(void *p1, void *p2) return -1; if (percent_a > percent_b) return 1; - else - return 0; + + if (a->incorrect < b->incorrect) + return -1; + if (a->incorrect > b->incorrect) + return 1; + + /* + * Since the above shows worse (incorrect) cases + * first, we continue that by showing best (correct) + * cases last. + */ + if (a->correct > b->correct) + return -1; + if (a->correct < b->correct) + return 1; + + return 0; } static struct tracer_stat annotated_branch_stats = { -- cgit v1.2.2 From ced5b697a76d325e7a7ac7d382dbbb632c765093 Mon Sep 17 00:00:00 2001 From: Brandon Phiilps Date: Wed, 10 Feb 2010 01:20:06 -0800 Subject: x86: Avoid race condition in pci_enable_msix() Keep chip_data in create_irq_nr and destroy_irq. When two drivers are setting up MSI-X at the same time via pci_enable_msix() there is a race. See this dmesg excerpt: [ 85.170610] ixgbe 0000:02:00.1: irq 97 for MSI/MSI-X [ 85.170611] alloc irq_desc for 99 on node -1 [ 85.170613] igb 0000:08:00.1: irq 98 for MSI/MSI-X [ 85.170614] alloc kstat_irqs on node -1 [ 85.170616] alloc irq_2_iommu on node -1 [ 85.170617] alloc irq_desc for 100 on node -1 [ 85.170619] alloc kstat_irqs on node -1 [ 85.170621] alloc irq_2_iommu on node -1 [ 85.170625] ixgbe 0000:02:00.1: irq 99 for MSI/MSI-X [ 85.170626] alloc irq_desc for 101 on node -1 [ 85.170628] igb 0000:08:00.1: irq 100 for MSI/MSI-X [ 85.170630] alloc kstat_irqs on node -1 [ 85.170631] alloc irq_2_iommu on node -1 [ 85.170635] alloc irq_desc for 102 on node -1 [ 85.170636] alloc kstat_irqs on node -1 [ 85.170639] alloc irq_2_iommu on node -1 [ 85.170646] BUG: unable to handle kernel NULL pointer dereference at 0000000000000088 As you can see igb and ixgbe are both alternating on create_irq_nr() via pci_enable_msix() in their probe function. ixgbe: While looping through irq_desc_ptrs[] via create_irq_nr() ixgbe choses irq_desc_ptrs[102] and exits the loop, drops vector_lock and calls dynamic_irq_init. Then it sets irq_desc_ptrs[102]->chip_data = NULL via dynamic_irq_init(). igb: Grabs the vector_lock now and starts looping over irq_desc_ptrs[] via create_irq_nr(). It gets to irq_desc_ptrs[102] and does this: cfg_new = irq_desc_ptrs[102]->chip_data; if (cfg_new->vector != 0) continue; This hits the NULL deref. Another possible race exists via pci_disable_msix() in a driver or in the number of error paths that call free_msi_irqs(): destroy_irq() dynamic_irq_cleanup() which sets desc->chip_data = NULL ...race window... desc->chip_data = cfg; Remove the save and restore code for cfg in create_irq_nr() and destroy_irq() and take the desc->lock when checking the irq_cfg. Reported-and-analyzed-by: Brandon Philips Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-3-git-send-email-yinghai@kernel.org> Signed-off-by: Brandon Phililps Cc: stable@kernel.org Signed-off-by: H. Peter Anvin --- kernel/irq/chip.c | 52 +++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 43 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index ecc3fa28f666..d70394f12ee9 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -18,11 +18,7 @@ #include "internals.h" -/** - * dynamic_irq_init - initialize a dynamically allocated irq - * @irq: irq number to initialize - */ -void dynamic_irq_init(unsigned int irq) +static void dynamic_irq_init_x(unsigned int irq, bool keep_chip_data) { struct irq_desc *desc; unsigned long flags; @@ -41,7 +37,8 @@ void dynamic_irq_init(unsigned int irq) desc->depth = 1; desc->msi_desc = NULL; desc->handler_data = NULL; - desc->chip_data = NULL; + if (!keep_chip_data) + desc->chip_data = NULL; desc->action = NULL; desc->irq_count = 0; desc->irqs_unhandled = 0; @@ -55,10 +52,26 @@ void dynamic_irq_init(unsigned int irq) } /** - * dynamic_irq_cleanup - cleanup a dynamically allocated irq + * dynamic_irq_init - initialize a dynamically allocated irq * @irq: irq number to initialize */ -void dynamic_irq_cleanup(unsigned int irq) +void dynamic_irq_init(unsigned int irq) +{ + dynamic_irq_init_x(irq, false); +} + +/** + * dynamic_irq_init_keep_chip_data - initialize a dynamically allocated irq + * @irq: irq number to initialize + * + * does not set irq_to_desc(irq)->chip_data to NULL + */ +void dynamic_irq_init_keep_chip_data(unsigned int irq) +{ + dynamic_irq_init_x(irq, true); +} + +static void dynamic_irq_cleanup_x(unsigned int irq, bool keep_chip_data) { struct irq_desc *desc = irq_to_desc(irq); unsigned long flags; @@ -77,7 +90,8 @@ void dynamic_irq_cleanup(unsigned int irq) } desc->msi_desc = NULL; desc->handler_data = NULL; - desc->chip_data = NULL; + if (!keep_chip_data) + desc->chip_data = NULL; desc->handle_irq = handle_bad_irq; desc->chip = &no_irq_chip; desc->name = NULL; @@ -85,6 +99,26 @@ void dynamic_irq_cleanup(unsigned int irq) raw_spin_unlock_irqrestore(&desc->lock, flags); } +/** + * dynamic_irq_cleanup - cleanup a dynamically allocated irq + * @irq: irq number to initialize + */ +void dynamic_irq_cleanup(unsigned int irq) +{ + dynamic_irq_cleanup_x(irq, false); +} + +/** + * dynamic_irq_cleanup_keep_chip_data - cleanup a dynamically allocated irq + * @irq: irq number to initialize + * + * does not set irq_to_desc(irq)->chip_data to NULL + */ +void dynamic_irq_cleanup_keep_chip_data(unsigned int irq) +{ + dynamic_irq_cleanup_x(irq, true); +} + /** * set_irq_chip - set the irq chip for an irq -- cgit v1.2.2 From 27811d8cabe56e0c3622251b049086f49face4ff Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:07 -0800 Subject: x86: Move range related operation to one file We have almost the same code for mtrr cleanup and amd_bus checkup, and this code will also be used in replacing bootmem with early_res, so try to move them together and reuse it from different parts. Also rename update_range to subtract_range as that is what the function is actually doing. -v2: update comments as Christoph requested Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-4-git-send-email-yinghai@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/Makefile | 2 +- kernel/range.c | 163 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 164 insertions(+), 1 deletion(-) create mode 100644 kernel/range.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 864ff75d65f2..ad47330ccf32 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -10,7 +10,7 @@ obj-y = sched.o fork.o exec_domain.o panic.o printk.o \ kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \ hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \ notifier.o ksysfs.o pm_qos_params.o sched_clock.o cred.o \ - async.o + async.o range.o obj-y += groups.o ifdef CONFIG_FUNCTION_TRACER diff --git a/kernel/range.c b/kernel/range.c new file mode 100644 index 000000000000..71e0021281fe --- /dev/null +++ b/kernel/range.c @@ -0,0 +1,163 @@ +/* + * Range add and subtract + */ +#include +#include +#include + +#include + +#ifndef ARRAY_SIZE +#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) +#endif + +int add_range(struct range *range, int az, int nr_range, u64 start, u64 end) +{ + if (start > end) + return nr_range; + + /* Out of slots: */ + if (nr_range >= az) + return nr_range; + + range[nr_range].start = start; + range[nr_range].end = end; + + nr_range++; + + return nr_range; +} + +int add_range_with_merge(struct range *range, int az, int nr_range, + u64 start, u64 end) +{ + int i; + + if (start > end) + return nr_range; + + /* Try to merge it with old one: */ + for (i = 0; i < nr_range; i++) { + u64 final_start, final_end; + u64 common_start, common_end; + + if (!range[i].end) + continue; + + common_start = max(range[i].start, start); + common_end = min(range[i].end, end); + if (common_start > common_end + 1) + continue; + + final_start = min(range[i].start, start); + final_end = max(range[i].end, end); + + range[i].start = final_start; + range[i].end = final_end; + return nr_range; + } + + /* Need to add it: */ + return add_range(range, az, nr_range, start, end); +} + +void subtract_range(struct range *range, int az, u64 start, u64 end) +{ + int i, j; + + if (start > end) + return; + + for (j = 0; j < az; j++) { + if (!range[j].end) + continue; + + if (start <= range[j].start && end >= range[j].end) { + range[j].start = 0; + range[j].end = 0; + continue; + } + + if (start <= range[j].start && end < range[j].end && + range[j].start < end + 1) { + range[j].start = end + 1; + continue; + } + + + if (start > range[j].start && end >= range[j].end && + range[j].end > start - 1) { + range[j].end = start - 1; + continue; + } + + if (start > range[j].start && end < range[j].end) { + /* Find the new spare: */ + for (i = 0; i < az; i++) { + if (range[i].end == 0) + break; + } + if (i < az) { + range[i].end = range[j].end; + range[i].start = end + 1; + } else { + printk(KERN_ERR "run of slot in ranges\n"); + } + range[j].end = start - 1; + continue; + } + } +} + +static int cmp_range(const void *x1, const void *x2) +{ + const struct range *r1 = x1; + const struct range *r2 = x2; + s64 start1, start2; + + start1 = r1->start; + start2 = r2->start; + + return start1 - start2; +} + +int clean_sort_range(struct range *range, int az) +{ + int i, j, k = az - 1, nr_range = 0; + + for (i = 0; i < k; i++) { + if (range[i].end) + continue; + for (j = k; j > i; j--) { + if (range[j].end) { + k = j; + break; + } + } + if (j == i) + break; + range[i].start = range[k].start; + range[i].end = range[k].end; + range[k].start = 0; + range[k].end = 0; + k--; + } + /* count it */ + for (i = 0; i < az; i++) { + if (!range[i].end) { + nr_range = i; + break; + } + } + + /* sort them */ + sort(range, nr_range, sizeof(struct range), cmp_range, NULL); + + return nr_range; +} + +void sort_range(struct range *range, int nr_range) +{ + /* sort them */ + sort(range, nr_range, sizeof(struct range), cmp_range, NULL); +} -- cgit v1.2.2 From e9a0064ad03b899938059bb576615ad9ed0f27f9 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:13 -0800 Subject: x86: Change range end to start+size So make interface more consistent with early_res. Later we can share some code with early_res. Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-10-git-send-email-yinghai@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/range.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/range.c b/kernel/range.c index 71e0021281fe..74e2e6114927 100644 --- a/kernel/range.c +++ b/kernel/range.c @@ -13,7 +13,7 @@ int add_range(struct range *range, int az, int nr_range, u64 start, u64 end) { - if (start > end) + if (start >= end) return nr_range; /* Out of slots: */ @@ -33,7 +33,7 @@ int add_range_with_merge(struct range *range, int az, int nr_range, { int i; - if (start > end) + if (start >= end) return nr_range; /* Try to merge it with old one: */ @@ -46,7 +46,7 @@ int add_range_with_merge(struct range *range, int az, int nr_range, common_start = max(range[i].start, start); common_end = min(range[i].end, end); - if (common_start > common_end + 1) + if (common_start > common_end) continue; final_start = min(range[i].start, start); @@ -65,7 +65,7 @@ void subtract_range(struct range *range, int az, u64 start, u64 end) { int i, j; - if (start > end) + if (start >= end) return; for (j = 0; j < az; j++) { @@ -79,15 +79,15 @@ void subtract_range(struct range *range, int az, u64 start, u64 end) } if (start <= range[j].start && end < range[j].end && - range[j].start < end + 1) { - range[j].start = end + 1; + range[j].start < end) { + range[j].start = end; continue; } if (start > range[j].start && end >= range[j].end && - range[j].end > start - 1) { - range[j].end = start - 1; + range[j].end > start) { + range[j].end = start; continue; } @@ -99,11 +99,11 @@ void subtract_range(struct range *range, int az, u64 start, u64 end) } if (i < az) { range[i].end = range[j].end; - range[i].start = end + 1; + range[i].start = end; } else { printk(KERN_ERR "run of slot in ranges\n"); } - range[j].end = start - 1; + range[j].end = start; continue; } } -- cgit v1.2.2 From 5c42dc7070c94622ca914b5a2e227f3744e857e7 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Thu, 11 Feb 2010 15:04:36 +0100 Subject: devres/irq: Fix devm_irq_match comment Fix the reference (in comment). Signed-off-by: Jean Delvare Signed-off-by: Jiri Kosina --- kernel/irq/devres.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/irq/devres.c b/kernel/irq/devres.c index 30d56bafc9c2..1ef4ffcdfa55 100644 --- a/kernel/irq/devres.c +++ b/kernel/irq/devres.c @@ -42,7 +42,7 @@ static int devm_irq_match(struct device *dev, void *res, void *data) * automatically freed on driver detach. * * If an IRQ allocated with this function needs to be freed - * separately, dev_free_irq() must be used. + * separately, devm_free_irq() must be used. */ int devm_request_threaded_irq(struct device *dev, unsigned int irq, irq_handler_t handler, irq_handler_t thread_fn, -- cgit v1.2.2 From c7c6b1fe9f942c1a30585ec2210a09dfff238506 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 10 Feb 2010 15:43:04 +0800 Subject: ftrace: Allow to remove a single function from function graph filter I don't see why we can only clear all functions from the filter. After patching: # echo sys_open > set_graph_function # echo sys_close >> set_graph_function # cat set_graph_function sys_open sys_close # echo '!sys_close' >> set_graph_function # cat set_graph_function sys_open Signed-off-by: Li Zefan LKML-Reference: <4B726388.2000408@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 51 +++++++++++++++++++++++++++++++-------------------- kernel/trace/trace.h | 3 ++- 2 files changed, 33 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 7968762c8167..43bec993c864 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2426,6 +2426,7 @@ static const struct file_operations ftrace_notrace_fops = { static DEFINE_MUTEX(graph_lock); int ftrace_graph_count; +int ftrace_graph_filter_enabled; unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly; static void * @@ -2448,7 +2449,7 @@ static void *g_start(struct seq_file *m, loff_t *pos) mutex_lock(&graph_lock); /* Nothing, tell g_show to print all functions are enabled */ - if (!ftrace_graph_count && !*pos) + if (!ftrace_graph_filter_enabled && !*pos) return (void *)1; return __g_next(m, pos); @@ -2494,6 +2495,7 @@ ftrace_graph_open(struct inode *inode, struct file *file) mutex_lock(&graph_lock); if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { + ftrace_graph_filter_enabled = 0; ftrace_graph_count = 0; memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs)); } @@ -2519,7 +2521,7 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) struct dyn_ftrace *rec; struct ftrace_page *pg; int search_len; - int found = 0; + int fail = 1; int type, not; char *search; bool exists; @@ -2530,37 +2532,51 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) /* decode regex */ type = filter_parse_regex(buffer, strlen(buffer), &search, ¬); - if (not) - return -EINVAL; + if (!not && *idx >= FTRACE_GRAPH_MAX_FUNCS) + return -EBUSY; search_len = strlen(search); mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { - if (*idx >= FTRACE_GRAPH_MAX_FUNCS) - break; - if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) continue; if (ftrace_match_record(rec, search, search_len, type)) { - /* ensure it is not already in the array */ + /* if it is in the array */ exists = false; - for (i = 0; i < *idx; i++) + for (i = 0; i < *idx; i++) { if (array[i] == rec->ip) { exists = true; break; } - if (!exists) - array[(*idx)++] = rec->ip; - found = 1; + } + + if (!not) { + fail = 0; + if (!exists) { + array[(*idx)++] = rec->ip; + if (*idx >= FTRACE_GRAPH_MAX_FUNCS) + goto out; + } + } else { + if (exists) { + array[i] = array[--(*idx)]; + array[*idx] = 0; + fail = 0; + } + } } } while_for_each_ftrace_rec(); - +out: mutex_unlock(&ftrace_lock); - return found ? 0 : -EINVAL; + if (fail) + return -EINVAL; + + ftrace_graph_filter_enabled = 1; + return 0; } static ssize_t @@ -2570,16 +2586,11 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, struct trace_parser parser; ssize_t read, ret; - if (!cnt || cnt < 0) + if (!cnt) return 0; mutex_lock(&graph_lock); - if (ftrace_graph_count >= FTRACE_GRAPH_MAX_FUNCS) { - ret = -EBUSY; - goto out_unlock; - } - if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX)) { ret = -ENOMEM; goto out_unlock; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index ce077fbbf552..b477fce41edf 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -497,6 +497,7 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s); #ifdef CONFIG_DYNAMIC_FTRACE /* TODO: make this variable */ #define FTRACE_GRAPH_MAX_FUNCS 32 +extern int ftrace_graph_filter_enabled; extern int ftrace_graph_count; extern unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS]; @@ -504,7 +505,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count) + if (!ftrace_graph_filter_enabled) return 1; for (i = 0; i < ftrace_graph_count; i++) { -- cgit v1.2.2 From 2225a122ae26d542bdce523d9d87a4a7ba10e07b Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Thu, 11 Feb 2010 11:51:00 -0800 Subject: ptrace: Add support for generic PTRACE_GETREGSET/PTRACE_SETREGSET Generic support for PTRACE_GETREGSET/PTRACE_SETREGSET commands which export the regsets supported by each architecture using the correponding NT_* types. These NT_* types are already part of the userland ABI, used in representing the architecture specific register sets as different NOTES in an ELF core file. 'addr' parameter for the ptrace system call encode the REGSET type (using the corresppnding NT_* type) and the 'data' parameter points to the struct iovec having the user buffer and the length of that buffer. struct iovec iov = { buf, len}; ret = ptrace(PTRACE_GETREGSET/PTRACE_SETREGSET, pid, NT_XXX_TYPE, &iov); On successful completion, iov.len will be updated by the kernel specifying how much the kernel has written/read to/from the user's iov.buf. x86 extended state registers are primarily exported using this interface. Signed-off-by: Suresh Siddha LKML-Reference: <20100211195614.886724710@sbs-t61.sc.intel.com> Acked-by: Hongjiu Lu Cc: Roland McGrath Signed-off-by: H. Peter Anvin --- kernel/ptrace.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 88 insertions(+) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 23bd09cd042e..13b4554d8fbb 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -22,6 +22,7 @@ #include #include #include +#include /* @@ -511,6 +512,47 @@ static int ptrace_resume(struct task_struct *child, long request, long data) return 0; } +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + +static const struct user_regset * +find_regset(const struct user_regset_view *view, unsigned int type) +{ + const struct user_regset *regset; + int n; + + for (n = 0; n < view->n; ++n) { + regset = view->regsets + n; + if (regset->core_note_type == type) + return regset; + } + + return NULL; +} + +static int ptrace_regset(struct task_struct *task, int req, unsigned int type, + struct iovec *kiov) +{ + const struct user_regset_view *view = task_user_regset_view(task); + const struct user_regset *regset = find_regset(view, type); + int regset_no; + + if (!regset || (kiov->iov_len % regset->size) != 0) + return -EIO; + + regset_no = regset - view->regsets; + kiov->iov_len = min(kiov->iov_len, + (__kernel_size_t) (regset->n * regset->size)); + + if (req == PTRACE_GETREGSET) + return copy_regset_to_user(task, view, regset_no, 0, + kiov->iov_len, kiov->iov_base); + else + return copy_regset_from_user(task, view, regset_no, 0, + kiov->iov_len, kiov->iov_base); +} + +#endif + int ptrace_request(struct task_struct *child, long request, long addr, long data) { @@ -573,6 +615,26 @@ int ptrace_request(struct task_struct *child, long request, return 0; return ptrace_resume(child, request, SIGKILL); +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + case PTRACE_GETREGSET: + case PTRACE_SETREGSET: + { + struct iovec kiov; + struct iovec __user *uiov = (struct iovec __user *) data; + + if (!access_ok(VERIFY_WRITE, uiov, sizeof(*uiov))) + return -EFAULT; + + if (__get_user(kiov.iov_base, &uiov->iov_base) || + __get_user(kiov.iov_len, &uiov->iov_len)) + return -EFAULT; + + ret = ptrace_regset(child, request, addr, &kiov); + if (!ret) + ret = __put_user(kiov.iov_len, &uiov->iov_len); + break; + } +#endif default: break; } @@ -711,6 +773,32 @@ int compat_ptrace_request(struct task_struct *child, compat_long_t request, else ret = ptrace_setsiginfo(child, &siginfo); break; +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + case PTRACE_GETREGSET: + case PTRACE_SETREGSET: + { + struct iovec kiov; + struct compat_iovec __user *uiov = + (struct compat_iovec __user *) datap; + compat_uptr_t ptr; + compat_size_t len; + + if (!access_ok(VERIFY_WRITE, uiov, sizeof(*uiov))) + return -EFAULT; + + if (__get_user(ptr, &uiov->iov_base) || + __get_user(len, &uiov->iov_len)) + return -EFAULT; + + kiov.iov_base = compat_ptr(ptr); + kiov.iov_len = len; + + ret = ptrace_regset(child, request, addr, &kiov); + if (!ret) + ret = __put_user(kiov.iov_len, &uiov->iov_len); + break; + } +#endif default: ret = ptrace_request(child, request, addr, data); -- cgit v1.2.2 From a9bb18f36c8056f0712fb28c52c0f85d98438dfb Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Wed, 10 Feb 2010 17:23:47 +0100 Subject: tracing/kprobes: Fix probe parsing Trying to add a probe like: echo p:myprobe 0x10000 > /sys/kernel/debug/tracing/kprobe_events will fail since the wrong pointer is passed to strict_strtoul when trying to convert the address to an unsigned long. Signed-off-by: Heiko Carstens Acked-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <20100210162346.GA6933@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_kprobe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 6ea90c0e2c96..50b1b8239806 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -689,7 +689,7 @@ static int create_trace_probe(int argc, char **argv) return -EINVAL; } /* an address specified */ - ret = strict_strtoul(&argv[0][2], 0, (unsigned long *)&addr); + ret = strict_strtoul(&argv[1][0], 0, (unsigned long *)&addr); if (ret) { pr_info("Failed to parse address.\n"); return ret; -- cgit v1.2.2 From 6f93d0a7c83772997b81c30d6f519a9a5dbab6a9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sun, 14 Feb 2010 11:12:04 +0100 Subject: perf_events: Fix FORK events Commit 22e19085 ("Honour event state for aux stream data") introduced a bug where we would drop FORK events. The thing is that we deliver FORK events to the child process' event, which at that time will be PERF_EVENT_STATE_INACTIVE because the child won't be scheduled in (we're in the middle of fork). Solve this twice, change the event state filter to exclude only disabled (STATE_OFF) or worse, and deliver FORK events to the current (parent). Signed-off-by: Peter Zijlstra Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1266142324.5273.411.camel@laptop> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 2b19297742cb..2ae7409bf38f 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3259,8 +3259,6 @@ static void perf_event_task_output(struct perf_event *event, task_event->event_id.tid = perf_event_tid(event, task); task_event->event_id.ptid = perf_event_tid(event, current); - task_event->event_id.time = perf_clock(); - perf_output_put(&handle, task_event->event_id); perf_output_end(&handle); @@ -3268,7 +3266,7 @@ static void perf_event_task_output(struct perf_event *event, static int perf_event_task_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3300,7 +3298,7 @@ static void perf_event_task_event(struct perf_task_event *task_event) cpuctx = &get_cpu_var(perf_cpu_context); perf_event_task_ctx(&cpuctx->ctx, task_event); if (!ctx) - ctx = rcu_dereference(task_event->task->perf_event_ctxp); + ctx = rcu_dereference(current->perf_event_ctxp); if (ctx) perf_event_task_ctx(ctx, task_event); put_cpu_var(perf_cpu_context); @@ -3331,6 +3329,7 @@ static void perf_event_task(struct task_struct *task, /* .ppid */ /* .tid */ /* .ptid */ + .time = perf_clock(), }, }; @@ -3380,7 +3379,7 @@ static void perf_event_comm_output(struct perf_event *event, static int perf_event_comm_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3500,7 +3499,7 @@ static void perf_event_mmap_output(struct perf_event *event, static int perf_event_mmap_match(struct perf_event *event, struct perf_mmap_event *mmap_event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) -- cgit v1.2.2 From 8b833c506c05c498d4215e2c260be44225daf6de Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 5 Feb 2010 01:24:34 -0500 Subject: kprobes: Add mcount to the kprobes blacklist Since mcount function can be called from everywhere, it should be blacklisted. Moreover, the "mcount" symbol is a special symbol name. So, it is better to put it in the generic blacklist. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Steven Rostedt LKML-Reference: <20100205062433.3745.36726.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index b7df302a0204..c4b43430d393 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -93,6 +93,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { {"native_get_debugreg",}, {"irq_entries_start",}, {"common_interrupt",}, + {"mcount",}, /* mcount can be called from everywhere */ {NULL} /* Terminator */ }; -- cgit v1.2.2 From dfff0615d28bdb3e8d213e5537dd069265912667 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Fri, 12 Feb 2010 21:58:11 +0100 Subject: tree-wide: fix typos "ass?o[sc]iac?te" -> "associate" in comments MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Uwe Kleine-König Signed-off-by: Jiri Kosina --- kernel/irq/chip.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index ecc3fa28f666..ec8a96382461 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -520,7 +520,7 @@ out: * signal. The occurence is latched into the irq controller hardware * and must be acked in order to be reenabled. After the ack another * interrupt can happen on the same source even before the first one - * is handled by the assosiacted event handler. If this happens it + * is handled by the associated event handler. If this happens it * might be necessary to disable (mask) the interrupt depending on the * controller hardware. This requires to reenable the interrupt inside * of the loop which handles the interrupts which have arrived while -- cgit v1.2.2 From 9000f05c6d1607f79c0deacf42b09693be673f4c Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Fri, 12 Feb 2010 17:14:22 -0800 Subject: sched: Fix SMT scheduler regression in find_busiest_queue() Fix a SMT scheduler performance regression that is leading to a scenario where SMT threads in one core are completely idle while both the SMT threads in another core (on the same socket) are busy. This is caused by this commit (with the problematic code highlighted) commit bdb94aa5dbd8b55e75f5a50b61312fe589e2c2d1 Author: Peter Zijlstra Date: Tue Sep 1 10:34:38 2009 +0200 sched: Try to deal with low capacity @@ -4203,15 +4223,18 @@ find_busiest_queue() ... for_each_cpu(i, sched_group_cpus(group)) { + unsigned long power = power_of(i); ... - wl = weighted_cpuload(i); + wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; + wl /= power; - if (rq->nr_running == 1 && wl > imbalance) + if (capacity && rq->nr_running == 1 && wl > imbalance) continue; On a SMT system, power of the HT logical cpu will be 589 and the scheduler load imbalance (for scenarios like the one mentioned above) can be approximately 1024 (SCHED_LOAD_SCALE). The above change of scaling the weighted load with the power will result in "wl > imbalance" and ultimately resulting in find_busiest_queue() return NULL, causing load_balance() to think that the load is well balanced. But infact one of the tasks can be moved to the idle core for optimal performance. We don't need to use the weighted load (wl) scaled by the cpu power to compare with imabalance. In that condition, we already know there is only a single task "rq->nr_running == 1" and the comparison between imbalance, wl is to make sure that we select the correct priority thread which matches imbalance. So we really need to compare the imabalnce with the original weighted load of the cpu and not the scaled load. But in other conditions where we want the most hammered(busiest) cpu, we can use scaled load to ensure that we consider the cpu power in addition to the actual load on that cpu, so that we can move the load away from the guy that is getting most hammered with respect to the actual capacity, as compared with the rest of the cpu's in that busiest group. Fix it. Reported-by: Ma Ling Initial-Analysis-by: Zhang, Yanmin Signed-off-by: Suresh Siddha Signed-off-by: Peter Zijlstra LKML-Reference: <1266023662.2808.118.camel@sbs-t61.sc.intel.com> Cc: stable@kernel.org [2.6.32.x] Signed-off-by: Thomas Gleixner --- kernel/sched.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index e3199df426e3..4d78aef4559d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4119,12 +4119,23 @@ find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, continue; rq = cpu_rq(i); - wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; - wl /= power; + wl = weighted_cpuload(i); + /* + * When comparing with imbalance, use weighted_cpuload() + * which is not scaled with the cpu power. + */ if (capacity && rq->nr_running == 1 && wl > imbalance) continue; + /* + * For the load comparisons with the other cpu's, consider + * the weighted_cpuload() scaled with the cpu power, so that + * the load can be moved away from the cpu that is potentially + * running at a lower capacity. + */ + wl = (wl * SCHED_LOAD_SCALE) / power; + if (wl > max_load) { max_load = wl; busiest = rq; -- cgit v1.2.2 From 0970d2992dfd7d5ec2c787417cf464f01eeaf42a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 15 Feb 2010 14:45:54 +0100 Subject: sched: Fix race between ttwu() and task_rq_lock() Thomas found that due to ttwu() changing a task's cpu without holding the rq->lock, task_rq_lock() might end up locking the wrong rq. Avoid this by serializing against TASK_WAKING. Reported-by: Thomas Gleixner Signed-off-by: Peter Zijlstra LKML-Reference: <1266241712.15770.420.camel@laptop> Signed-off-by: Thomas Gleixner --- kernel/sched.c | 73 ++++++++++++++++++++++++++++++++++++---------------------- 1 file changed, 46 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 4d78aef4559d..404e2017c0cf 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -940,6 +940,19 @@ static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) } #endif /* __ARCH_WANT_UNLOCKED_CTXSW */ +/* + * Check whether the task is waking, we use this to synchronize against + * ttwu() so that task_cpu() reports a stable number. + * + * We need to make an exception for PF_STARTING tasks because the fork + * path might require task_rq_lock() to work, eg. it can call + * set_cpus_allowed_ptr() from the cpuset clone_ns code. + */ +static inline int task_is_waking(struct task_struct *p) +{ + return unlikely((p->state == TASK_WAKING) && !(p->flags & PF_STARTING)); +} + /* * __task_rq_lock - lock the runqueue a given task resides on. * Must be called interrupts disabled. @@ -947,10 +960,14 @@ static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) static inline struct rq *__task_rq_lock(struct task_struct *p) __acquires(rq->lock) { + struct rq *rq; + for (;;) { - struct rq *rq = task_rq(p); + while (task_is_waking(p)) + cpu_relax(); + rq = task_rq(p); raw_spin_lock(&rq->lock); - if (likely(rq == task_rq(p))) + if (likely(rq == task_rq(p) && !task_is_waking(p))) return rq; raw_spin_unlock(&rq->lock); } @@ -967,10 +984,12 @@ static struct rq *task_rq_lock(struct task_struct *p, unsigned long *flags) struct rq *rq; for (;;) { + while (task_is_waking(p)) + cpu_relax(); local_irq_save(*flags); rq = task_rq(p); raw_spin_lock(&rq->lock); - if (likely(rq == task_rq(p))) + if (likely(rq == task_rq(p) && !task_is_waking(p))) return rq; raw_spin_unlock_irqrestore(&rq->lock, *flags); } @@ -2408,14 +2427,27 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, __task_rq_unlock(rq); cpu = select_task_rq(p, SD_BALANCE_WAKE, wake_flags); - if (cpu != orig_cpu) + if (cpu != orig_cpu) { + /* + * Since we migrate the task without holding any rq->lock, + * we need to be careful with task_rq_lock(), since that + * might end up locking an invalid rq. + */ set_task_cpu(p, cpu); + } - rq = __task_rq_lock(p); + rq = cpu_rq(cpu); + raw_spin_lock(&rq->lock); update_rq_clock(rq); + /* + * We migrated the task without holding either rq->lock, however + * since the task is not on the task list itself, nobody else + * will try and migrate the task, hence the rq should match the + * cpu we just moved it to. + */ + WARN_ON(task_cpu(p) != cpu); WARN_ON(p->state != TASK_WAKING); - cpu = task_cpu(p); #ifdef CONFIG_SCHEDSTATS schedstat_inc(rq, ttwu_count); @@ -2647,7 +2679,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; - int cpu __maybe_unused = get_cpu(); + int cpu = get_cpu(); #ifdef CONFIG_SMP /* @@ -2663,7 +2695,13 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) set_task_cpu(p, cpu); #endif - rq = task_rq_lock(p, &flags); + /* + * Since the task is not on the rq and we still have TASK_WAKING set + * nobody else will migrate this task. + */ + rq = cpu_rq(cpu); + raw_spin_lock_irqsave(&rq->lock, flags); + BUG_ON(p->state != TASK_WAKING); p->state = TASK_RUNNING; update_rq_clock(rq); @@ -7156,27 +7194,8 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) struct rq *rq; int ret = 0; - /* - * Since we rely on wake-ups to migrate sleeping tasks, don't change - * the ->cpus_allowed mask from under waking tasks, which would be - * possible when we change rq->lock in ttwu(), so synchronize against - * TASK_WAKING to avoid that. - * - * Make an exception for freshly cloned tasks, since cpuset namespaces - * might move the task about, we have to validate the target in - * wake_up_new_task() anyway since the cpu might have gone away. - */ -again: - while (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) - cpu_relax(); - rq = task_rq_lock(p, &flags); - if (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) { - task_rq_unlock(rq, &flags); - goto again; - } - if (!cpumask_intersects(new_mask, cpu_active_mask)) { ret = -EINVAL; goto out; -- cgit v1.2.2 From 1a02d59aba9b61b820517fb135086471c065b573 Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Wed, 27 Jan 2010 17:09:34 +0300 Subject: kfifo: Make kfifo_initialized work after kfifo_free After kfifo rework it's no longer possible to reliably know if kfifo is usable, since after kfifo_free(), kfifo_initialized() would still return true. The correct behaviour is needed for at least FHCI USB driver. This patch fixes the issue by resetting the kfifo to zero values (the same approach is used in kfifo_alloc() if allocation failed). Signed-off-by: Anton Vorontsov Acked-by: Stefani Seibold Signed-off-by: Greg Kroah-Hartman --- kernel/kfifo.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/kfifo.c b/kernel/kfifo.c index 498cabba225e..559fb5582b60 100644 --- a/kernel/kfifo.c +++ b/kernel/kfifo.c @@ -97,6 +97,7 @@ EXPORT_SYMBOL(kfifo_alloc); void kfifo_free(struct kfifo *fifo) { kfree(fifo->buffer); + _kfifo_init(fifo, NULL, 0); } EXPORT_SYMBOL(kfifo_free); -- cgit v1.2.2 From 5a5e0f4c7038168e38d1db6af09d1ac715ee9888 Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Wed, 27 Jan 2010 17:09:38 +0300 Subject: kfifo: Don't use integer as NULL pointer This patch fixes following sparse warnings: include/linux/kfifo.h:127:25: warning: Using plain integer as NULL pointer kernel/kfifo.c:83:21: warning: Using plain integer as NULL pointer Signed-off-by: Anton Vorontsov Acked-by: Stefani Seibold Signed-off-by: Greg Kroah-Hartman --- kernel/kfifo.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kfifo.c b/kernel/kfifo.c index 559fb5582b60..35edbe22e9a9 100644 --- a/kernel/kfifo.c +++ b/kernel/kfifo.c @@ -80,7 +80,7 @@ int kfifo_alloc(struct kfifo *fifo, unsigned int size, gfp_t gfp_mask) buffer = kmalloc(size, gfp_mask); if (!buffer) { - _kfifo_init(fifo, 0, 0); + _kfifo_init(fifo, NULL, 0); return -ENOMEM; } -- cgit v1.2.2 From 43cf38eb5cea91245502df3fcee4dbfc1c74dd1c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 2 Feb 2010 14:38:57 +0900 Subject: percpu: add __percpu sparse annotations to core kernel subsystems Add __percpu sparse annotations to core subsystems. These annotations are to make sparse consider percpu variables to be in a different address space and warn if accessed without going through percpu accessors. This patch doesn't affect normal builds. Signed-off-by: Tejun Heo Reviewed-by: Christoph Lameter Acked-by: Paul E. McKenney Cc: Jens Axboe Cc: linux-mm@kvack.org Cc: Rusty Russell Cc: Dipankar Sarma Cc: Peter Zijlstra Cc: Andrew Morton Cc: Eric Biederman --- kernel/kexec.c | 2 +- kernel/sched.c | 4 ++-- kernel/stop_machine.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index ef077fb73155..87ebe8adc474 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -41,7 +41,7 @@ #include /* Per cpu memory for storing cpu states in case of system crash. */ -note_buf_t* crash_notes; +note_buf_t __percpu *crash_notes; /* vmcoreinfo stuff */ static unsigned char vmcoreinfo_data[VMCOREINFO_BYTES]; diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30a91b1..978edfd35a96 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1566,7 +1566,7 @@ static unsigned long cpu_avg_load_per_task(int cpu) #ifdef CONFIG_FAIR_GROUP_SCHED -static __read_mostly unsigned long *update_shares_data; +static __read_mostly unsigned long __percpu *update_shares_data; static void __set_se_shares(struct sched_entity *se, unsigned long shares); @@ -10683,7 +10683,7 @@ struct cgroup_subsys cpu_cgroup_subsys = { struct cpuacct { struct cgroup_subsys_state css; /* cpuusage holds pointer to a u64-type object on every cpu */ - u64 *cpuusage; + u64 __percpu *cpuusage; struct percpu_counter cpustat[CPUACCT_STAT_NSTATS]; struct cpuacct *parent; }; diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c index 912823e2a11b..9bb9fb1bd79c 100644 --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -45,7 +45,7 @@ static int refcount; static struct workqueue_struct *stop_machine_wq; static struct stop_machine_data active, idle; static const struct cpumask *active_cpus; -static void *stop_machine_work; +static void __percpu *stop_machine_work; static void set_state(enum stopmachine_state newstate) { -- cgit v1.2.2 From 580e0ad21d6d6f932461d24b47041e3dd499c23f Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 16 Feb 2010 18:40:35 -0800 Subject: core: Move early_res from arch/x86 to kernel/ This makes the range reservation feature available to other architectures. -v2: add get_max_mapped, max_pfn_mapped only defined in x86... to fix PPC compiling -v3: according to hpa, add CONFIG_HAVE_EARLY_RES -v4: fix typo about EARLY_RES in config Signed-off-by: Yinghai Lu LKML-Reference: <4B7B5723.4070009@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/Makefile | 1 + kernel/early_res.c | 513 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 514 insertions(+) create mode 100644 kernel/early_res.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index ad47330ccf32..1292b863d667 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -11,6 +11,7 @@ obj-y = sched.o fork.o exec_domain.o panic.o printk.o \ hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \ notifier.o ksysfs.o pm_qos_params.o sched_clock.o cred.o \ async.o range.o +obj-$(CONFIG_HAVE_EARLY_RES) += early_res.o obj-y += groups.o ifdef CONFIG_FUNCTION_TRACER diff --git a/kernel/early_res.c b/kernel/early_res.c new file mode 100644 index 000000000000..aa5494ac4462 --- /dev/null +++ b/kernel/early_res.c @@ -0,0 +1,513 @@ +/* + * early_res, could be used to replace bootmem + */ +#include +#include +#include +#include +#include +#include + +/* + * Early reserved memory areas. + */ +/* + * need to make sure this one is bigger enough before + * find_fw_memmap_area could be used + */ +#define MAX_EARLY_RES_X 32 + +struct early_res { + u64 start, end; + char name[15]; + char overlap_ok; +}; +static struct early_res early_res_x[MAX_EARLY_RES_X] __initdata; + +static int max_early_res __initdata = MAX_EARLY_RES_X; +static struct early_res *early_res __initdata = &early_res_x[0]; +static int early_res_count __initdata; + +static int __init find_overlapped_early(u64 start, u64 end) +{ + int i; + struct early_res *r; + + for (i = 0; i < max_early_res && early_res[i].end; i++) { + r = &early_res[i]; + if (end > r->start && start < r->end) + break; + } + + return i; +} + +/* + * Drop the i-th range from the early reservation map, + * by copying any higher ranges down one over it, and + * clearing what had been the last slot. + */ +static void __init drop_range(int i) +{ + int j; + + for (j = i + 1; j < max_early_res && early_res[j].end; j++) + ; + + memmove(&early_res[i], &early_res[i + 1], + (j - 1 - i) * sizeof(struct early_res)); + + early_res[j - 1].end = 0; + early_res_count--; +} + +/* + * Split any existing ranges that: + * 1) are marked 'overlap_ok', and + * 2) overlap with the stated range [start, end) + * into whatever portion (if any) of the existing range is entirely + * below or entirely above the stated range. Drop the portion + * of the existing range that overlaps with the stated range, + * which will allow the caller of this routine to then add that + * stated range without conflicting with any existing range. + */ +static void __init drop_overlaps_that_are_ok(u64 start, u64 end) +{ + int i; + struct early_res *r; + u64 lower_start, lower_end; + u64 upper_start, upper_end; + char name[15]; + + for (i = 0; i < max_early_res && early_res[i].end; i++) { + r = &early_res[i]; + + /* Continue past non-overlapping ranges */ + if (end <= r->start || start >= r->end) + continue; + + /* + * Leave non-ok overlaps as is; let caller + * panic "Overlapping early reservations" + * when it hits this overlap. + */ + if (!r->overlap_ok) + return; + + /* + * We have an ok overlap. We will drop it from the early + * reservation map, and add back in any non-overlapping + * portions (lower or upper) as separate, overlap_ok, + * non-overlapping ranges. + */ + + /* 1. Note any non-overlapping (lower or upper) ranges. */ + strncpy(name, r->name, sizeof(name) - 1); + + lower_start = lower_end = 0; + upper_start = upper_end = 0; + if (r->start < start) { + lower_start = r->start; + lower_end = start; + } + if (r->end > end) { + upper_start = end; + upper_end = r->end; + } + + /* 2. Drop the original ok overlapping range */ + drop_range(i); + + i--; /* resume for-loop on copied down entry */ + + /* 3. Add back in any non-overlapping ranges. */ + if (lower_end) + reserve_early_overlap_ok(lower_start, lower_end, name); + if (upper_end) + reserve_early_overlap_ok(upper_start, upper_end, name); + } +} + +static void __init __reserve_early(u64 start, u64 end, char *name, + int overlap_ok) +{ + int i; + struct early_res *r; + + i = find_overlapped_early(start, end); + if (i >= max_early_res) + panic("Too many early reservations"); + r = &early_res[i]; + if (r->end) + panic("Overlapping early reservations " + "%llx-%llx %s to %llx-%llx %s\n", + start, end - 1, name ? name : "", r->start, + r->end - 1, r->name); + r->start = start; + r->end = end; + r->overlap_ok = overlap_ok; + if (name) + strncpy(r->name, name, sizeof(r->name) - 1); + early_res_count++; +} + +/* + * A few early reservtations come here. + * + * The 'overlap_ok' in the name of this routine does -not- mean it + * is ok for these reservations to overlap an earlier reservation. + * Rather it means that it is ok for subsequent reservations to + * overlap this one. + * + * Use this entry point to reserve early ranges when you are doing + * so out of "Paranoia", reserving perhaps more memory than you need, + * just in case, and don't mind a subsequent overlapping reservation + * that is known to be needed. + * + * The drop_overlaps_that_are_ok() call here isn't really needed. + * It would be needed if we had two colliding 'overlap_ok' + * reservations, so that the second such would not panic on the + * overlap with the first. We don't have any such as of this + * writing, but might as well tolerate such if it happens in + * the future. + */ +void __init reserve_early_overlap_ok(u64 start, u64 end, char *name) +{ + drop_overlaps_that_are_ok(start, end); + __reserve_early(start, end, name, 1); +} + +static void __init __check_and_double_early_res(u64 ex_start, u64 ex_end) +{ + u64 start, end, size, mem; + struct early_res *new; + + /* do we have enough slots left ? */ + if ((max_early_res - early_res_count) > max(max_early_res/8, 2)) + return; + + /* double it */ + mem = -1ULL; + size = sizeof(struct early_res) * max_early_res * 2; + if (early_res == early_res_x) + start = 0; + else + start = early_res[0].end; + end = ex_start; + if (start + size < end) + mem = find_fw_memmap_area(start, end, size, + sizeof(struct early_res)); + if (mem == -1ULL) { + start = ex_end; + end = get_max_mapped(); + if (start + size < end) + mem = find_fw_memmap_area(start, end, size, + sizeof(struct early_res)); + } + if (mem == -1ULL) + panic("can not find more space for early_res array"); + + new = __va(mem); + /* save the first one for own */ + new[0].start = mem; + new[0].end = mem + size; + new[0].overlap_ok = 0; + /* copy old to new */ + if (early_res == early_res_x) { + memcpy(&new[1], &early_res[0], + sizeof(struct early_res) * max_early_res); + memset(&new[max_early_res+1], 0, + sizeof(struct early_res) * (max_early_res - 1)); + early_res_count++; + } else { + memcpy(&new[1], &early_res[1], + sizeof(struct early_res) * (max_early_res - 1)); + memset(&new[max_early_res], 0, + sizeof(struct early_res) * max_early_res); + } + memset(&early_res[0], 0, sizeof(struct early_res) * max_early_res); + early_res = new; + max_early_res *= 2; + printk(KERN_DEBUG "early_res array is doubled to %d at [%llx - %llx]\n", + max_early_res, mem, mem + size - 1); +} + +/* + * Most early reservations come here. + * + * We first have drop_overlaps_that_are_ok() drop any pre-existing + * 'overlap_ok' ranges, so that we can then reserve this memory + * range without risk of panic'ing on an overlapping overlap_ok + * early reservation. + */ +void __init reserve_early(u64 start, u64 end, char *name) +{ + if (start >= end) + return; + + __check_and_double_early_res(start, end); + + drop_overlaps_that_are_ok(start, end); + __reserve_early(start, end, name, 0); +} + +void __init reserve_early_without_check(u64 start, u64 end, char *name) +{ + struct early_res *r; + + if (start >= end) + return; + + __check_and_double_early_res(start, end); + + r = &early_res[early_res_count]; + + r->start = start; + r->end = end; + r->overlap_ok = 0; + if (name) + strncpy(r->name, name, sizeof(r->name) - 1); + early_res_count++; +} + +void __init free_early(u64 start, u64 end) +{ + struct early_res *r; + int i; + + i = find_overlapped_early(start, end); + r = &early_res[i]; + if (i >= max_early_res || r->end != end || r->start != start) + panic("free_early on not reserved area: %llx-%llx!", + start, end - 1); + + drop_range(i); +} + +#ifdef CONFIG_NO_BOOTMEM +static void __init subtract_early_res(struct range *range, int az) +{ + int i, count; + u64 final_start, final_end; + int idx = 0; + + count = 0; + for (i = 0; i < max_early_res && early_res[i].end; i++) + count++; + + /* need to skip first one ?*/ + if (early_res != early_res_x) + idx = 1; + +#define DEBUG_PRINT_EARLY_RES 1 + +#if DEBUG_PRINT_EARLY_RES + printk(KERN_INFO "Subtract (%d early reservations)\n", count); +#endif + for (i = idx; i < count; i++) { + struct early_res *r = &early_res[i]; +#if DEBUG_PRINT_EARLY_RES + printk(KERN_INFO " #%d [%010llx - %010llx] %15s\n", i, + r->start, r->end, r->name); +#endif + final_start = PFN_DOWN(r->start); + final_end = PFN_UP(r->end); + if (final_start >= final_end) + continue; + subtract_range(range, az, final_start, final_end); + } + +} + +int __init get_free_all_memory_range(struct range **rangep, int nodeid) +{ + int i, count; + u64 start = 0, end; + u64 size; + u64 mem; + struct range *range; + int nr_range; + + count = 0; + for (i = 0; i < max_early_res && early_res[i].end; i++) + count++; + + count *= 2; + + size = sizeof(struct range) * count; + end = get_max_mapped(); +#ifdef MAX_DMA32_PFN + if (end > (MAX_DMA32_PFN << PAGE_SHIFT)) + start = MAX_DMA32_PFN << PAGE_SHIFT; +#endif + mem = find_fw_memmap_area(start, end, size, sizeof(struct range)); + if (mem == -1ULL) + panic("can not find more space for range free"); + + range = __va(mem); + /* use early_node_map[] and early_res to get range array at first */ + memset(range, 0, size); + nr_range = 0; + + /* need to go over early_node_map to find out good range for node */ + nr_range = add_from_early_node_map(range, count, nr_range, nodeid); +#ifdef CONFIG_X86_32 + subtract_range(range, count, max_low_pfn, -1ULL); +#endif + subtract_early_res(range, count); + nr_range = clean_sort_range(range, count); + + /* need to clear it ? */ + if (nodeid == MAX_NUMNODES) { + memset(&early_res[0], 0, + sizeof(struct early_res) * max_early_res); + early_res = NULL; + max_early_res = 0; + } + + *rangep = range; + return nr_range; +} +#else +void __init early_res_to_bootmem(u64 start, u64 end) +{ + int i, count; + u64 final_start, final_end; + int idx = 0; + + count = 0; + for (i = 0; i < max_early_res && early_res[i].end; i++) + count++; + + /* need to skip first one ?*/ + if (early_res != early_res_x) + idx = 1; + + printk(KERN_INFO "(%d/%d early reservations) ==> bootmem [%010llx - %010llx]\n", + count - idx, max_early_res, start, end); + for (i = idx; i < count; i++) { + struct early_res *r = &early_res[i]; + printk(KERN_INFO " #%d [%010llx - %010llx] %16s", i, + r->start, r->end, r->name); + final_start = max(start, r->start); + final_end = min(end, r->end); + if (final_start >= final_end) { + printk(KERN_CONT "\n"); + continue; + } + printk(KERN_CONT " ==> [%010llx - %010llx]\n", + final_start, final_end); + reserve_bootmem_generic(final_start, final_end - final_start, + BOOTMEM_DEFAULT); + } + /* clear them */ + memset(&early_res[0], 0, sizeof(struct early_res) * max_early_res); + early_res = NULL; + max_early_res = 0; + early_res_count = 0; +} +#endif + +/* Check for already reserved areas */ +static inline int __init bad_addr(u64 *addrp, u64 size, u64 align) +{ + int i; + u64 addr = *addrp; + int changed = 0; + struct early_res *r; +again: + i = find_overlapped_early(addr, addr + size); + r = &early_res[i]; + if (i < max_early_res && r->end) { + *addrp = addr = round_up(r->end, align); + changed = 1; + goto again; + } + return changed; +} + +/* Check for already reserved areas */ +static inline int __init bad_addr_size(u64 *addrp, u64 *sizep, u64 align) +{ + int i; + u64 addr = *addrp, last; + u64 size = *sizep; + int changed = 0; +again: + last = addr + size; + for (i = 0; i < max_early_res && early_res[i].end; i++) { + struct early_res *r = &early_res[i]; + if (last > r->start && addr < r->start) { + size = r->start - addr; + changed = 1; + goto again; + } + if (last > r->end && addr < r->end) { + addr = round_up(r->end, align); + size = last - addr; + changed = 1; + goto again; + } + if (last <= r->end && addr >= r->start) { + (*sizep)++; + return 0; + } + } + if (changed) { + *addrp = addr; + *sizep = size; + } + return changed; +} + +/* + * Find a free area with specified alignment in a specific range. + * only with the area.between start to end is active range from early_node_map + * so they are good as RAM + */ +u64 __init find_early_area(u64 ei_start, u64 ei_last, u64 start, u64 end, + u64 size, u64 align) +{ + u64 addr, last; + + addr = round_up(ei_start, align); + if (addr < start) + addr = round_up(start, align); + if (addr >= ei_last) + goto out; + while (bad_addr(&addr, size, align) && addr+size <= ei_last) + ; + last = addr + size; + if (last > ei_last) + goto out; + if (last > end) + goto out; + + return addr; + +out: + return -1ULL; +} + +u64 __init find_early_area_size(u64 ei_start, u64 ei_last, u64 start, + u64 *sizep, u64 align) +{ + u64 addr, last; + + addr = round_up(ei_start, align); + if (addr < start) + addr = round_up(start, align); + if (addr >= ei_last) + goto out; + *sizep = ei_last - addr; + while (bad_addr_size(&addr, sizep, align) && addr + *sizep <= ei_last) + ; + last = addr + *sizep; + if (last > ei_last) + goto out; + + return addr; + +out: + return -1ULL; +} -- cgit v1.2.2 From 83ab0aa0d5623d823444db82c3b3c34d7ec364ae Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 17 Feb 2010 09:05:48 +0100 Subject: sched: Don't use possibly stale sched_class setscheduler() saves task->sched_class outside of the rq->lock held region for a check after the setscheduler changes have become effective. That might result in checking a stale value. rtmutex_setprio() has the same problem, though it is protected by p->pi_lock against setscheduler(), but for correctness sake (and to avoid bad examples) it needs to be fixed as well. Retrieve task->sched_class inside of the rq->lock held region. Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Cc: stable@kernel.org --- kernel/sched.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index af5fa239804d..0b914fc90a55 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4249,7 +4249,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) unsigned long flags; int oldprio, on_rq, running; struct rq *rq; - const struct sched_class *prev_class = p->sched_class; + const struct sched_class *prev_class; BUG_ON(prio < 0 || prio > MAX_PRIO); @@ -4257,6 +4257,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) update_rq_clock(rq); oldprio = p->prio; + prev_class = p->sched_class; on_rq = p->se.on_rq; running = task_current(rq, p); if (on_rq) @@ -4476,7 +4477,7 @@ static int __sched_setscheduler(struct task_struct *p, int policy, { int retval, oldprio, oldpolicy = -1, on_rq, running; unsigned long flags; - const struct sched_class *prev_class = p->sched_class; + const struct sched_class *prev_class; struct rq *rq; int reset_on_fork; @@ -4590,6 +4591,7 @@ recheck: p->sched_reset_on_fork = reset_on_fork; oldprio = p->prio; + prev_class = p->sched_class; __setscheduler(rq, p, policy, param->sched_priority); if (running) -- cgit v1.2.2 From e7b8e675d9c71b868b66f62f725a948047514719 Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Tue, 26 Jan 2010 04:40:03 -0500 Subject: tracing: Unify arch_syscall_addr() implementations Most implementations of arch_syscall_addr() are the same, so create a default version in common code and move the one piece that differs (the syscall table) to asm/syscall.h. New arch ports don't have to waste time copying & pasting this simple function. The s390/sparc versions need to be different, so document why. Signed-off-by: Mike Frysinger Acked-by: David S. Miller Acked-by: Paul Mundt Acked-by: Heiko Carstens Cc: Steven Rostedt LKML-Reference: <1264498803-17278-1-git-send-email-vapier@gentoo.org> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace_syscalls.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 49cea70fbf6d..ecf00782b46c 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -394,6 +394,11 @@ int init_syscall_trace(struct ftrace_event_call *call) return id; } +unsigned long __init arch_syscall_addr(int nr) +{ + return (unsigned long)sys_call_table[nr]; +} + int __init init_ftrace_syscalls(void) { struct syscall_metadata *meta; -- cgit v1.2.2 From f850c30c8b426ba1688cb63b1a3e534eed03a138 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Wed, 10 Feb 2010 17:25:17 +0100 Subject: tracing/kprobes: Make Kconfig dependencies generic KPROBES_EVENT actually depends on the regs and stack access API (b1cf540f) and not on x86. So introduce a new config option which architectures can select if they have the API implemented and switch x86. Signed-off-by: Heiko Carstens Acked-by: Masami Hiramatsu Cc: Ingo Molnar Cc: Martin Schwidefsky LKML-Reference: <20100210162517.GB6933@osiris.boeblingen.de.ibm.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 6c22d8a2f289..40fef552f012 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -451,7 +451,7 @@ config BLK_DEV_IO_TRACE config KPROBE_EVENT depends on KPROBES - depends on X86 + depends on HAVE_REGS_AND_STACK_ACCESS_API bool "Enable kprobes-based dynamic events" select TRACING default y -- cgit v1.2.2 From febcb0c59ac19fef2081a30e371e7af3619b5e91 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:32 -0800 Subject: irq: Remove unnecessary bootmem code mem_init is moved early already. Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-29-git-send-email-yinghai@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/irq/handle.c | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 814940e7f485..0e823c0d1c9c 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -19,7 +19,6 @@ #include #include #include -#include #include #include "internals.h" @@ -87,12 +86,8 @@ void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) { void *ptr; - if (slab_is_available()) - ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), - GFP_ATOMIC, node); - else - ptr = alloc_bootmem_node(NODE_DATA(node), - nr * sizeof(*desc->kstat_irqs)); + ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), + GFP_ATOMIC, node); /* * don't overwite if can not get new one @@ -219,10 +214,7 @@ struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) if (desc) goto out_unlock; - if (slab_is_available()) - desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); - else - desc = alloc_bootmem_node(NODE_DATA(node), sizeof(*desc)); + desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); printk(KERN_DEBUG " alloc irq_desc for %d on node %d\n", irq, node); if (!desc) { -- cgit v1.2.2 From 99558f0bbe68cb09799ec38adbaa3f3b2dc7ba63 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:34 -0800 Subject: sparseirq: Change irq_desc_ptrs to static Add replace_irq_desc() instead of poking at the array directly. -v2: remove unneeded boundary check in replace_irq_desc Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-31-git-send-email-yinghai@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/irq/handle.c | 7 ++++++- kernel/irq/internals.h | 6 +----- kernel/irq/numa_migrate.c | 4 ++-- 3 files changed, 9 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 0e823c0d1c9c..266f7986aa08 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -127,7 +127,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) */ DEFINE_RAW_SPINLOCK(sparse_irq_lock); -struct irq_desc **irq_desc_ptrs __read_mostly; +static struct irq_desc **irq_desc_ptrs __read_mostly; static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { [0 ... NR_IRQS_LEGACY-1] = { @@ -192,6 +192,11 @@ struct irq_desc *irq_to_desc(unsigned int irq) return NULL; } +void replace_irq_desc(unsigned int irq, struct irq_desc *desc) +{ + irq_desc_ptrs[irq] = desc; +} + struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) { struct irq_desc *desc; diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index b2821f070a3d..c63f3bc88f0b 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -21,11 +21,7 @@ extern void clear_kstat_irqs(struct irq_desc *desc); extern raw_spinlock_t sparse_irq_lock; #ifdef CONFIG_SPARSE_IRQ -/* irq_desc_ptrs allocated at boot time */ -extern struct irq_desc **irq_desc_ptrs; -#else -/* irq_desc_ptrs is a fixed size array */ -extern struct irq_desc *irq_desc_ptrs[NR_IRQS]; +void replace_irq_desc(unsigned int irq, struct irq_desc *desc); #endif #ifdef CONFIG_PROC_FS diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index 26bac9d8f860..963559dbd858 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -70,7 +70,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, raw_spin_lock_irqsave(&sparse_irq_lock, flags); /* We have to check it to avoid races with another CPU */ - desc = irq_desc_ptrs[irq]; + desc = irq_to_desc(irq); if (desc && old_desc != desc) goto out_unlock; @@ -90,7 +90,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, goto out_unlock; } - irq_desc_ptrs[irq] = desc; + replace_irq_desc(irq, desc); raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); /* free the old one */ -- cgit v1.2.2 From b5eb78f76ddfa7caf4340cf6893b032f45d8114a Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 10 Feb 2010 01:20:35 -0800 Subject: sparseirq: Use radix_tree instead of ptrs array Use radix_tree irq_desc_tree instead of irq_desc_ptrs. -v2: according to Eric and cyrill to use radix_tree_lookup_slot and radix_tree_replace_slot Signed-off-by: Yinghai Lu LKML-Reference: <1265793639-15071-32-git-send-email-yinghai@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/irq/handle.c | 49 +++++++++++++++++++++++++------------------------ 1 file changed, 25 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 266f7986aa08..76d5a671bfe1 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -19,6 +19,7 @@ #include #include #include +#include #include #include "internals.h" @@ -127,7 +128,26 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) */ DEFINE_RAW_SPINLOCK(sparse_irq_lock); -static struct irq_desc **irq_desc_ptrs __read_mostly; +static RADIX_TREE(irq_desc_tree, GFP_ATOMIC); + +static void set_irq_desc(unsigned int irq, struct irq_desc *desc) +{ + radix_tree_insert(&irq_desc_tree, irq, desc); +} + +struct irq_desc *irq_to_desc(unsigned int irq) +{ + return radix_tree_lookup(&irq_desc_tree, irq); +} + +void replace_irq_desc(unsigned int irq, struct irq_desc *desc) +{ + void **ptr; + + ptr = radix_tree_lookup_slot(&irq_desc_tree, irq); + if (ptr) + radix_tree_replace_slot(ptr, desc); +} static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { [0 ... NR_IRQS_LEGACY-1] = { @@ -159,9 +179,6 @@ int __init early_irq_init(void) legacy_count = ARRAY_SIZE(irq_desc_legacy); node = first_online_node; - /* allocate irq_desc_ptrs array based on nr_irqs */ - irq_desc_ptrs = kcalloc(nr_irqs, sizeof(void *), GFP_NOWAIT); - /* allocate based on nr_cpu_ids */ kstat_irqs_legacy = kzalloc_node(NR_IRQS_LEGACY * nr_cpu_ids * sizeof(int), GFP_NOWAIT, node); @@ -175,28 +192,12 @@ int __init early_irq_init(void) lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); alloc_desc_masks(&desc[i], node, true); init_desc_masks(&desc[i]); - irq_desc_ptrs[i] = desc + i; + set_irq_desc(i, &desc[i]); } - for (i = legacy_count; i < nr_irqs; i++) - irq_desc_ptrs[i] = NULL; - return arch_early_irq_init(); } -struct irq_desc *irq_to_desc(unsigned int irq) -{ - if (irq_desc_ptrs && irq < nr_irqs) - return irq_desc_ptrs[irq]; - - return NULL; -} - -void replace_irq_desc(unsigned int irq, struct irq_desc *desc) -{ - irq_desc_ptrs[irq] = desc; -} - struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) { struct irq_desc *desc; @@ -208,14 +209,14 @@ struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) return NULL; } - desc = irq_desc_ptrs[irq]; + desc = irq_to_desc(irq); if (desc) return desc; raw_spin_lock_irqsave(&sparse_irq_lock, flags); /* We have to check it to avoid races with another CPU */ - desc = irq_desc_ptrs[irq]; + desc = irq_to_desc(irq); if (desc) goto out_unlock; @@ -228,7 +229,7 @@ struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) } init_one_irq_desc(irq, desc, node); - irq_desc_ptrs[irq] = desc; + set_irq_desc(irq, desc); out_unlock: raw_spin_unlock_irqrestore(&sparse_irq_lock, flags); -- cgit v1.2.2 From b54452b07a7b1b8cc1385edba3ef2ef6d4679d5a Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Thu, 18 Feb 2010 08:14:31 +0000 Subject: const: struct nla_policy Make remaining netlink policies as const. Fixup coding style where needed. Signed-off-by: Alexey Dobriyan Signed-off-by: David S. Miller --- kernel/taskstats.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/taskstats.c b/kernel/taskstats.c index ea8384d3caa7..899ca51be5e8 100644 --- a/kernel/taskstats.c +++ b/kernel/taskstats.c @@ -46,15 +46,13 @@ static struct genl_family family = { .maxattr = TASKSTATS_CMD_ATTR_MAX, }; -static struct nla_policy taskstats_cmd_get_policy[TASKSTATS_CMD_ATTR_MAX+1] -__read_mostly = { +static const struct nla_policy taskstats_cmd_get_policy[TASKSTATS_CMD_ATTR_MAX+1] = { [TASKSTATS_CMD_ATTR_PID] = { .type = NLA_U32 }, [TASKSTATS_CMD_ATTR_TGID] = { .type = NLA_U32 }, [TASKSTATS_CMD_ATTR_REGISTER_CPUMASK] = { .type = NLA_STRING }, [TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK] = { .type = NLA_STRING },}; -static struct nla_policy -cgroupstats_cmd_get_policy[CGROUPSTATS_CMD_ATTR_MAX+1] __read_mostly = { +static const struct nla_policy cgroupstats_cmd_get_policy[CGROUPSTATS_CMD_ATTR_MAX+1] = { [CGROUPSTATS_CMD_ATTR_FD] = { .type = NLA_U32 }, }; -- cgit v1.2.2 From b26b2d494b659f988b4d75eb394dfa0ddac415c9 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Fri, 1 Jan 2010 17:40:49 +0100 Subject: resource/PCI: align functions now return start of resource As suggested by Linus, align functions should return the start of a resource, not void. An update of "res->start" is no longer necessary. Cc: Bjorn Helgaas Cc: Yinghai Lu Signed-off-by: Dominik Brodowski Signed-off-by: Jesse Barnes --- kernel/resource.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index af96c1e4b54b..e697f20e2288 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -303,8 +303,10 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, static int find_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, - void (*alignf)(void *, struct resource *, - resource_size_t, resource_size_t), + resource_size_t (*alignf)(void *, + struct resource *, + resource_size_t, + resource_size_t), void *alignf_data) { struct resource *this = root->child; @@ -330,7 +332,7 @@ static int find_resource(struct resource *root, struct resource *new, tmp.end = max; tmp.start = ALIGN(tmp.start, align); if (alignf) - alignf(alignf_data, &tmp, size, align); + tmp.start = alignf(alignf_data, &tmp, size, align); if (tmp.start < tmp.end && tmp.end - tmp.start >= size - 1) { new->start = tmp.start; new->end = tmp.start + size - 1; @@ -358,8 +360,10 @@ static int find_resource(struct resource *root, struct resource *new, int allocate_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, - void (*alignf)(void *, struct resource *, - resource_size_t, resource_size_t), + resource_size_t (*alignf)(void *, + struct resource *, + resource_size_t, + resource_size_t), void *alignf_data) { int err; -- cgit v1.2.2 From 3b7a17fcdae532d29dffab9d564a28be08960988 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Fri, 1 Jan 2010 17:40:50 +0100 Subject: resource/PCI: mark struct resource as const Now that we return the new resource start position, there is no need to update "struct resource" inside the align function. Therefore, mark the struct resource as const. Cc: Bjorn Helgaas Cc: Yinghai Lu Signed-off-by: Dominik Brodowski Signed-off-by: Jesse Barnes --- kernel/resource.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index e697f20e2288..7fd123ad00aa 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -304,7 +304,7 @@ static int find_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, resource_size_t (*alignf)(void *, - struct resource *, + const struct resource *, resource_size_t, resource_size_t), void *alignf_data) @@ -361,7 +361,7 @@ int allocate_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, resource_size_t (*alignf)(void *, - struct resource *, + const struct resource *, resource_size_t, resource_size_t), void *alignf_data) -- cgit v1.2.2 From 5eeec0ec931a01e85b3701ce121b7d8a1800ec60 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 22 Dec 2009 15:02:22 -0800 Subject: resource: add release_child_resources Useful for freeing a portion of the resource tree, e.g. when trying to reallocate resources more efficiently. Signed-off-by: Yinghai Lu Acked-by: Linus Torvalds Signed-off-by: Jesse Barnes --- kernel/resource.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index 7fd123ad00aa..24e9e60c1459 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -188,6 +188,36 @@ static int __release_resource(struct resource *old) return -EINVAL; } +static void __release_child_resources(struct resource *r) +{ + struct resource *tmp, *p; + resource_size_t size; + + p = r->child; + r->child = NULL; + while (p) { + tmp = p; + p = p->sibling; + + tmp->parent = NULL; + tmp->sibling = NULL; + __release_child_resources(tmp); + + printk(KERN_DEBUG "release child resource %pR\n", tmp); + /* need to restore size, and keep flags */ + size = resource_size(tmp); + tmp->start = 0; + tmp->end = size - 1; + } +} + +void release_child_resources(struct resource *r) +{ + write_lock(&resource_lock); + __release_child_resources(r); + write_unlock(&resource_lock); +} + /** * request_resource - request and reserve an I/O or memory resource * @root: root resource descriptor -- cgit v1.2.2 From 6cbf82148ff286ec22a55be6836c3a5bffc489c1 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Wed, 17 Feb 2010 23:44:58 +0100 Subject: PCI PM: Run-time callbacks for PCI bus type Introduce run-time PM callbacks for the PCI bus type. Make the new callbacks work in analogy with the existing system sleep PM callbacks, so that the drivers already converted to struct dev_pm_ops can use their suspend and resume routines for run-time PM without modifications. Signed-off-by: Rafael J. Wysocki Signed-off-by: Jesse Barnes --- kernel/power/Kconfig | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 91e09d3b2eb2..4c9cffcf69c7 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -222,3 +222,8 @@ config PM_RUNTIME and the bus type drivers of the buses the devices are on are responsible for the actual handling of the autosuspend requests and wake-up events. + +config PM_OPS + bool + depends on PM_SLEEP || PM_RUNTIME + default y -- cgit v1.2.2 From 701188374b6f1ef9cf7e4dce4a2e69ef4c0012ac Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Mon, 22 Feb 2010 12:44:16 -0800 Subject: kernel/sys.c: fix missing rcu protection for sys_getpriority() find_task_by_vpid() is not safe without rcu_read_lock(). 2.6.33-rc7 got RCU protection for sys_setpriority() but missed it for sys_getpriority(). Signed-off-by: Tetsuo Handa Cc: Oleg Nesterov Cc: "Paul E. McKenney" Acked-by: Serge Hallyn Acked-by: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 26a6b73a6b85..18bde979f346 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -222,6 +222,7 @@ SYSCALL_DEFINE2(getpriority, int, which, int, who) if (which > PRIO_USER || which < PRIO_PROCESS) return -EINVAL; + rcu_read_lock(); read_lock(&tasklist_lock); switch (which) { case PRIO_PROCESS: @@ -267,6 +268,7 @@ SYSCALL_DEFINE2(getpriority, int, which, int, who) } out_unlock: read_unlock(&tasklist_lock); + rcu_read_unlock(); return retval; } -- cgit v1.2.2 From c6a0dd7ec6fb2d4927979ed4dc562fc5c122d826 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 22 Feb 2010 14:51:32 -0800 Subject: ptrace: Fix ptrace_regset() comments and diagnose errors specifically Return -EINVAL for the bad size and for unrecognized NT_* type in ptrace_regset() instead of -EIO. Also update the comments for this ptrace interface with more clarifications. Requested-by: Roland McGrath Requested-by: Oleg Nesterov Signed-off-by: Suresh Siddha LKML-Reference: <20100222225240.397523600@sbs-t61.sc.intel.com> Acked-by: Roland McGrath Signed-off-by: H. Peter Anvin --- kernel/ptrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 13b4554d8fbb..42ad8ae729a0 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -537,7 +537,7 @@ static int ptrace_regset(struct task_struct *task, int req, unsigned int type, int regset_no; if (!regset || (kiov->iov_len % regset->size) != 0) - return -EIO; + return -EINVAL; regset_no = regset - view->regsets; kiov->iov_len = min(kiov->iov_len, -- cgit v1.2.2 From 632ee200130899252508c478ad0e808222573fbc Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:04:45 -0800 Subject: rcu: Introduce lockdep-based checking to RCU read-side primitives Inspection is proving insufficient to catch all RCU misuses, which is understandable given that rcu_dereference() might be protected by any of four different flavors of RCU (RCU, RCU-bh, RCU-sched, and SRCU), and might also/instead be protected by any of a number of locking primitives. It is therefore time to enlist the aid of lockdep. This set of patches is inspired by earlier work by Peter Zijlstra and Thomas Gleixner, and takes the following approach: o Set up separate lockdep classes for RCU, RCU-bh, and RCU-sched. o Set up separate lockdep classes for each instance of SRCU. o Create primitives that check for being in an RCU read-side critical section. These return exact answers if lockdep is fully enabled, but if unsure, report being in an RCU read-side critical section. (We want to avoid false positives!) The primitives are: For RCU: rcu_read_lock_held(void) For RCU-bh: rcu_read_lock_bh_held(void) For RCU-sched: rcu_read_lock_sched_held(void) For SRCU: srcu_read_lock_held(struct srcu_struct *sp) o Add rcu_dereference_check(), which takes a second argument in which one places a boolean expression based on the above primitives and/or lockdep_is_held(). o A new kernel configuration parameter, CONFIG_PROVE_RCU, enables rcu_dereference_check(). This depends on CONFIG_PROVE_LOCKING, and should be quite helpful during the transition period while CONFIG_PROVE_RCU-unaware patches are in flight. The existing rcu_dereference() primitive does no checking, but upcoming patches will change that. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 10 ++++++++++ kernel/rcutorture.c | 12 ++++++++++-- kernel/srcu.c | 50 +++++++++++++++++++++++++++++++++----------------- 3 files changed, 53 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 9b7fd4723878..033cb55c26df 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -50,6 +50,16 @@ static struct lock_class_key rcu_lock_key; struct lockdep_map rcu_lock_map = STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key); EXPORT_SYMBOL_GPL(rcu_lock_map); + +static struct lock_class_key rcu_bh_lock_key; +struct lockdep_map rcu_bh_lock_map = + STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_bh", &rcu_bh_lock_key); +EXPORT_SYMBOL_GPL(rcu_bh_lock_map); + +static struct lock_class_key rcu_sched_lock_key; +struct lockdep_map rcu_sched_lock_map = + STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_sched", &rcu_sched_lock_key); +EXPORT_SYMBOL_GPL(rcu_sched_lock_map); #endif /* diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index adda92bfafac..5f43f30fcd1d 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -796,7 +796,11 @@ static void rcu_torture_timer(unsigned long unused) idx = cur_ops->readlock(); completed = cur_ops->completed(); - p = rcu_dereference(rcu_torture_current); + p = rcu_dereference_check(rcu_torture_current, + rcu_read_lock_held() || + rcu_read_lock_bh_held() || + rcu_read_lock_sched_held() || + srcu_read_lock_held(&srcu_ctl)); if (p == NULL) { /* Leave because rcu_torture_writer is not yet underway */ cur_ops->readunlock(idx); @@ -853,7 +857,11 @@ rcu_torture_reader(void *arg) } idx = cur_ops->readlock(); completed = cur_ops->completed(); - p = rcu_dereference(rcu_torture_current); + p = rcu_dereference_check(rcu_torture_current, + rcu_read_lock_held() || + rcu_read_lock_bh_held() || + rcu_read_lock_sched_held() || + srcu_read_lock_held(&srcu_ctl)); if (p == NULL) { /* Wait for rcu_torture_writer to get underway */ cur_ops->readunlock(idx); diff --git a/kernel/srcu.c b/kernel/srcu.c index 31b275b9c112..bde4295774c8 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -34,6 +34,30 @@ #include #include +static int init_srcu_struct_fields(struct srcu_struct *sp) +{ + sp->completed = 0; + mutex_init(&sp->mutex); + sp->per_cpu_ref = alloc_percpu(struct srcu_struct_array); + return sp->per_cpu_ref ? 0 : -ENOMEM; +} + +#ifdef CONFIG_DEBUG_LOCK_ALLOC + +int __init_srcu_struct(struct srcu_struct *sp, const char *name, + struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* Don't re-initialize a lock while it is held. */ + debug_check_no_locks_freed((void *)sp, sizeof(*sp)); + lockdep_init_map(&sp->dep_map, name, key, 0); +#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + return init_srcu_struct_fields(sp); +} +EXPORT_SYMBOL_GPL(__init_srcu_struct); + +#else /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + /** * init_srcu_struct - initialize a sleep-RCU structure * @sp: structure to initialize. @@ -44,13 +68,12 @@ */ int init_srcu_struct(struct srcu_struct *sp) { - sp->completed = 0; - mutex_init(&sp->mutex); - sp->per_cpu_ref = alloc_percpu(struct srcu_struct_array); - return (sp->per_cpu_ref ? 0 : -ENOMEM); + return init_srcu_struct_fields(sp); } EXPORT_SYMBOL_GPL(init_srcu_struct); +#endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + /* * srcu_readers_active_idx -- returns approximate number of readers * active on the specified rank of per-CPU counters. @@ -100,15 +123,12 @@ void cleanup_srcu_struct(struct srcu_struct *sp) } EXPORT_SYMBOL_GPL(cleanup_srcu_struct); -/** - * srcu_read_lock - register a new reader for an SRCU-protected structure. - * @sp: srcu_struct in which to register the new reader. - * +/* * Counts the new reader in the appropriate per-CPU element of the * srcu_struct. Must be called from process context. * Returns an index that must be passed to the matching srcu_read_unlock(). */ -int srcu_read_lock(struct srcu_struct *sp) +int __srcu_read_lock(struct srcu_struct *sp) { int idx; @@ -120,26 +140,22 @@ int srcu_read_lock(struct srcu_struct *sp) preempt_enable(); return idx; } -EXPORT_SYMBOL_GPL(srcu_read_lock); +EXPORT_SYMBOL_GPL(__srcu_read_lock); -/** - * srcu_read_unlock - unregister a old reader from an SRCU-protected structure. - * @sp: srcu_struct in which to unregister the old reader. - * @idx: return value from corresponding srcu_read_lock(). - * +/* * Removes the count for the old reader from the appropriate per-CPU * element of the srcu_struct. Note that this may well be a different * CPU than that which was incremented by the corresponding srcu_read_lock(). * Must be called from process context. */ -void srcu_read_unlock(struct srcu_struct *sp, int idx) +void __srcu_read_unlock(struct srcu_struct *sp, int idx) { preempt_disable(); srcu_barrier(); /* ensure compiler won't misorder critical section. */ per_cpu_ptr(sp->per_cpu_ref, smp_processor_id())->c[idx]--; preempt_enable(); } -EXPORT_SYMBOL_GPL(srcu_read_unlock); +EXPORT_SYMBOL_GPL(__srcu_read_unlock); /* * Helper function for synchronize_srcu() and synchronize_srcu_expedited(). -- cgit v1.2.2 From 0632eb3d7563d6a76d49a3860b6352d800c92854 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:04:47 -0800 Subject: rcu: Integrate rcu_dereference_check() message into lockdep Make rcu_dereference_check() print the list of held locks in addition to the stack dump to ease debugging. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-3-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index c62ec14609b9..672c436946ce 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3809,3 +3809,21 @@ void lockdep_sys_exit(void) lockdep_print_held_locks(curr); } } + +void lockdep_rcu_dereference(const char *file, const int line) +{ + struct task_struct *curr = current; + + if (!debug_locks_off()) + return; + printk("\n==============================================\n"); + printk( "[ BUG: Unsafe rcu_dereference_check() usage! ]\n"); + printk( "----------------------------------------------\n"); + printk("%s:%d invoked rcu_dereference_check() without protection!\n", + file, line); + printk("\nother info that might help us debug this:\n\n"); + lockdep_print_held_locks(curr); + printk("\nstack backtrace:\n"); + dump_stack(); +} +EXPORT_SYMBOL_GPL(lockdep_rcu_dereference); -- cgit v1.2.2 From d11c563dd20ff35da5652c3e1c989d9e10e1d6d0 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:04:50 -0800 Subject: sched: Use lockdep-based checking on rcu_dereference() Update the rcu_dereference() usages to take advantage of the new lockdep-based checking. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-6-git-send-email-paulmck@linux.vnet.ibm.com> [ -v2: fix allmodconfig missing symbol export build failure on x86 ] Signed-off-by: Ingo Molnar --- kernel/cgroup.c | 14 ++++++++++++++ kernel/exit.c | 14 +++++++++++--- kernel/fork.c | 1 + kernel/notifier.c | 6 +++--- kernel/pid.c | 2 +- kernel/sched.c | 11 ++++++++--- 6 files changed, 38 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index aa3bee566446..b1a0f5a528fe 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -166,6 +166,20 @@ static DEFINE_SPINLOCK(hierarchy_id_lock); */ static int need_forkexit_callback __read_mostly; +#ifdef CONFIG_PROVE_LOCKING +int cgroup_lock_is_held(void) +{ + return lockdep_is_held(&cgroup_mutex); +} +#else /* #ifdef CONFIG_PROVE_LOCKING */ +int cgroup_lock_is_held(void) +{ + return mutex_is_locked(&cgroup_mutex); +} +#endif /* #else #ifdef CONFIG_PROVE_LOCKING */ + +EXPORT_SYMBOL_GPL(cgroup_lock_is_held); + /* convenient tests for these bits */ inline int cgroup_is_removed(const struct cgroup *cgrp) { diff --git a/kernel/exit.c b/kernel/exit.c index 546774a31a66..45ed043b8bf5 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -85,7 +85,9 @@ static void __exit_signal(struct task_struct *tsk) BUG_ON(!sig); BUG_ON(!atomic_read(&sig->count)); - sighand = rcu_dereference(tsk->sighand); + sighand = rcu_dereference_check(tsk->sighand, + rcu_read_lock_held() || + lockdep_is_held(&tasklist_lock)); spin_lock(&sighand->siglock); posix_cpu_timers_exit(tsk); @@ -170,8 +172,10 @@ void release_task(struct task_struct * p) repeat: tracehook_prepare_release_task(p); /* don't need to get the RCU readlock here - the process is dead and - * can't be modifying its own credentials */ + * can't be modifying its own credentials. But shut RCU-lockdep up */ + rcu_read_lock(); atomic_dec(&__task_cred(p)->user->processes); + rcu_read_unlock(); proc_flush_task(p); @@ -473,9 +477,11 @@ static void close_files(struct files_struct * files) /* * It is safe to dereference the fd table without RCU or * ->file_lock because this is the last reference to the - * files structure. + * files structure. But use RCU to shut RCU-lockdep up. */ + rcu_read_lock(); fdt = files_fdtable(files); + rcu_read_unlock(); for (;;) { unsigned long set; i = j * __NFDBITS; @@ -521,10 +527,12 @@ void put_files_struct(struct files_struct *files) * at the end of the RCU grace period. Otherwise, * you can free files immediately. */ + rcu_read_lock(); fdt = files_fdtable(files); if (fdt != &files->fdtab) kmem_cache_free(files_cachep, files); free_fdtable(fdt); + rcu_read_unlock(); } } diff --git a/kernel/fork.c b/kernel/fork.c index f88bd984df35..17bbf093356d 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -86,6 +86,7 @@ int max_threads; /* tunable limit on nr_threads */ DEFINE_PER_CPU(unsigned long, process_counts) = 0; __cacheline_aligned DEFINE_RWLOCK(tasklist_lock); /* outer */ +EXPORT_SYMBOL_GPL(tasklist_lock); int nr_processes(void) { diff --git a/kernel/notifier.c b/kernel/notifier.c index acd24e7643eb..2488ba7eb568 100644 --- a/kernel/notifier.c +++ b/kernel/notifier.c @@ -78,10 +78,10 @@ static int __kprobes notifier_call_chain(struct notifier_block **nl, int ret = NOTIFY_DONE; struct notifier_block *nb, *next_nb; - nb = rcu_dereference(*nl); + nb = rcu_dereference_raw(*nl); while (nb && nr_to_call) { - next_nb = rcu_dereference(nb->next); + next_nb = rcu_dereference_raw(nb->next); #ifdef CONFIG_DEBUG_NOTIFIERS if (unlikely(!func_ptr_is_kernel_text(nb->notifier_call))) { @@ -309,7 +309,7 @@ int __blocking_notifier_call_chain(struct blocking_notifier_head *nh, * racy then it does not matter what the result of the test * is, we re-check the list after having taken the lock anyway: */ - if (rcu_dereference(nh->head)) { + if (rcu_dereference_raw(nh->head)) { down_read(&nh->rwsem); ret = notifier_call_chain(&nh->head, val, v, nr_to_call, nr_calls); diff --git a/kernel/pid.c b/kernel/pid.c index 2e17c9c92cbe..b08e697cd83f 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -367,7 +367,7 @@ struct task_struct *pid_task(struct pid *pid, enum pid_type type) struct task_struct *result = NULL; if (pid) { struct hlist_node *first; - first = rcu_dereference(pid->tasks[type].first); + first = rcu_dereference_check(pid->tasks[type].first, rcu_read_lock_held() || lockdep_is_held(&tasklist_lock)); if (first) result = hlist_entry(first, struct task_struct, pids[(type)].node); } diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30a91b1..70ae68680d4c 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -645,6 +645,11 @@ static inline int cpu_of(struct rq *rq) #endif } +#define for_each_domain_rd(p) \ + rcu_dereference_check((p), \ + rcu_read_lock_sched_held() || \ + lockdep_is_held(&sched_domains_mutex)) + /* * The domain tree (rq->sd) is protected by RCU's quiescent state transition. * See detach_destroy_domains: synchronize_sched for details. @@ -653,7 +658,7 @@ static inline int cpu_of(struct rq *rq) * preempt-disabled sections. */ #define for_each_domain(cpu, __sd) \ - for (__sd = rcu_dereference(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) + for (__sd = for_each_domain_rd(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) #define cpu_rq(cpu) (&per_cpu(runqueues, (cpu))) #define this_rq() (&__get_cpu_var(runqueues)) @@ -1531,7 +1536,7 @@ static unsigned long target_load(int cpu, int type) static struct sched_group *group_of(int cpu) { - struct sched_domain *sd = rcu_dereference(cpu_rq(cpu)->sd); + struct sched_domain *sd = rcu_dereference_sched(cpu_rq(cpu)->sd); if (!sd) return NULL; @@ -4888,7 +4893,7 @@ static void run_rebalance_domains(struct softirq_action *h) static inline int on_null_domain(int cpu) { - return !rcu_dereference(cpu_rq(cpu)->sd); + return !rcu_dereference_sched(cpu_rq(cpu)->sd); } /* -- cgit v1.2.2 From 497f0ab39cd25bed317b29482c147c967f7ecd1f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:04:51 -0800 Subject: sched: Better name for for_each_domain_rd As suggested by Peter Ziljstra, make better choice of name for for_each_domain_rd(), containing "rcu_dereference", given that it is but a wrapper for rcu_dereference_check(). The name rcu_dereference_check_sched_domain() does that and provides a separate per-subsystem name space. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-7-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 70ae68680d4c..3218f5213717 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -645,7 +645,7 @@ static inline int cpu_of(struct rq *rq) #endif } -#define for_each_domain_rd(p) \ +#define rcu_dereference_check_sched_domain(p) \ rcu_dereference_check((p), \ rcu_read_lock_sched_held() || \ lockdep_is_held(&sched_domains_mutex)) @@ -658,7 +658,7 @@ static inline int cpu_of(struct rq *rq) * preempt-disabled sections. */ #define for_each_domain(cpu, __sd) \ - for (__sd = for_each_domain_rd(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) + for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) #define cpu_rq(cpu) (&per_cpu(runqueues, (cpu))) #define this_rq() (&__get_cpu_var(runqueues)) -- cgit v1.2.2 From 8bd93a2c5d4cab2ae17d06350daa7dbf546a4634 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:04:59 -0800 Subject: rcu: Accelerate grace period if last non-dynticked CPU Currently, rcu_needs_cpu() simply checks whether the current CPU has an outstanding RCU callback, which means that the last CPU to go into dyntick-idle mode might wait a few ticks for the relevant grace periods to complete. However, if all the other CPUs are in dyntick-idle mode, and if this CPU is in a quiescent state (which it is for RCU-bh and RCU-sched any time that we are considering going into dyntick-idle mode), then the grace period is instantly complete. This patch therefore repeatedly invokes the RCU grace-period machinery in order to force any needed grace periods to complete quickly. It does so a limited number of times in order to prevent starvation by an RCU callback function that might pass itself to call_rcu(). However, if any CPU other than the current one is not in dyntick-idle mode, fall back to simply checking (with fix to bug noted by Lai Jiangshan). Also, take advantage of last grace-period forcing, the opportunity to do so noted by Steve Rostedt. And apply simplified #ifdef condition suggested by Frederic Weisbecker. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-15-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 5 ++-- kernel/rcutree_plugin.h | 69 +++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 71 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 099a255ede4c..29d88c08d875 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1550,10 +1550,9 @@ static int rcu_pending(int cpu) /* * Check to see if any future RCU-related work will need to be done * by the current CPU, even if none need be done immediately, returning - * 1 if so. This function is part of the RCU implementation; it is -not- - * an exported member of the RCU API. + * 1 if so. */ -int rcu_needs_cpu(int cpu) +static int rcu_needs_cpu_quick_check(int cpu) { /* RCU callbacks either ready or pending? */ return per_cpu(rcu_sched_data, cpu).nxtlist || diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index e77cdf321e13..a82566696b0b 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -906,3 +906,72 @@ static void __init __rcu_init_preempt(void) } #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ + +#if !defined(CONFIG_RCU_FAST_NO_HZ) + +/* + * Check to see if any future RCU-related work will need to be done + * by the current CPU, even if none need be done immediately, returning + * 1 if so. This function is part of the RCU implementation; it is -not- + * an exported member of the RCU API. + * + * Because we have preemptible RCU, just check whether this CPU needs + * any flavor of RCU. Do not chew up lots of CPU cycles with preemption + * disabled in a most-likely vain attempt to cause RCU not to need this CPU. + */ +int rcu_needs_cpu(int cpu) +{ + return rcu_needs_cpu_quick_check(cpu); +} + +#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */ + +#define RCU_NEEDS_CPU_FLUSHES 5 + +/* + * Check to see if any future RCU-related work will need to be done + * by the current CPU, even if none need be done immediately, returning + * 1 if so. This function is part of the RCU implementation; it is -not- + * an exported member of the RCU API. + * + * Because we are not supporting preemptible RCU, attempt to accelerate + * any current grace periods so that RCU no longer needs this CPU, but + * only if all other CPUs are already in dynticks-idle mode. This will + * allow the CPU cores to be powered down immediately, as opposed to after + * waiting many milliseconds for grace periods to elapse. + */ +int rcu_needs_cpu(int cpu) +{ + int c = 1; + int i; + int thatcpu; + + /* Don't bother unless we are the last non-dyntick-idle CPU. */ + for_each_cpu_not(thatcpu, nohz_cpu_mask) + if (thatcpu != cpu) + return rcu_needs_cpu_quick_check(cpu); + + /* Try to push remaining RCU-sched and RCU-bh callbacks through. */ + for (i = 0; i < RCU_NEEDS_CPU_FLUSHES && c; i++) { + c = 0; + if (per_cpu(rcu_sched_data, cpu).nxtlist) { + rcu_sched_qs(cpu); + force_quiescent_state(&rcu_sched_state, 0); + __rcu_process_callbacks(&rcu_sched_state, + &per_cpu(rcu_sched_data, cpu)); + c = !!per_cpu(rcu_sched_data, cpu).nxtlist; + } + if (per_cpu(rcu_bh_data, cpu).nxtlist) { + rcu_bh_qs(cpu); + force_quiescent_state(&rcu_bh_state, 0); + __rcu_process_callbacks(&rcu_bh_state, + &per_cpu(rcu_bh_data, cpu)); + c = !!per_cpu(rcu_bh_data, cpu).nxtlist; + } + } + + /* If RCU callbacks are still pending, RCU still needs this CPU. */ + return c; +} + +#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */ -- cgit v1.2.2 From 20133cfce7d0bbdcc0c398301030c091f5675c88 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:05:01 -0800 Subject: rcu: Stop overflowing signed integers The C standard does not specify the result of an operation that overflows a signed integer, so such operations need to be avoided. This patch changes the type of several fields from "long" to "unsigned long" and adjusts operations as needed. ULONG_CMP_GE() and ULONG_CMP_LT() macros are introduced to do the modular comparisons that are appropriate given that overflow is an expected event. Acked-by: Mathieu Desnoyers Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-17-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 11 +++++------ kernel/rcutree.h | 33 ++++++++++++++++++--------------- kernel/rcutree_trace.c | 14 +++++++------- 3 files changed, 30 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 29d88c08d875..dd0d31dffcdc 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -500,7 +500,7 @@ static void print_cpu_stall(struct rcu_state *rsp) trigger_all_cpu_backtrace(); spin_lock_irqsave(&rnp->lock, flags); - if ((long)(jiffies - rsp->jiffies_stall) >= 0) + if (ULONG_CMP_GE(jiffies, rsp->jiffies_stall)) rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; spin_unlock_irqrestore(&rnp->lock, flags); @@ -1216,8 +1216,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->n_force_qs_lh++; /* Inexact, can lose counts. Tough! */ return; /* Someone else is already on the job. */ } - if (relaxed && - (long)(rsp->jiffies_force_qs - jiffies) >= 0) + if (relaxed && ULONG_CMP_GE(rsp->jiffies_force_qs, jiffies)) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); /* irqs already disabled */ @@ -1295,7 +1294,7 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) * If an RCU GP has gone long enough, go check for dyntick * idle CPUs and, if needed, send resched IPIs. */ - if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) + if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) force_quiescent_state(rsp, 1); /* @@ -1392,7 +1391,7 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), force_quiescent_state(rsp, 0); rdp->n_force_qs_snap = rsp->n_force_qs; rdp->qlen_last_fqs_check = rdp->qlen; - } else if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) + } else if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) force_quiescent_state(rsp, 1); local_irq_restore(flags); } @@ -1525,7 +1524,7 @@ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp) /* Has an RCU GP gone long enough to send resched IPIs &c? */ if (rcu_gp_in_progress(rsp) && - ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)) { + ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) { rdp->n_rp_need_fqs++; return 1; } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index d9d032abd665..7495fed49c30 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -92,10 +92,10 @@ struct rcu_dynticks { struct rcu_node { spinlock_t lock; /* Root rcu_node's lock protects some */ /* rcu_state fields as well as following. */ - long gpnum; /* Current grace period for this node. */ + unsigned long gpnum; /* Current grace period for this node. */ /* This will either be equal to or one */ /* behind the root rcu_node's gpnum. */ - long completed; /* Last grace period completed for this node. */ + unsigned long completed; /* Last GP completed for this node. */ /* This will either be equal to or one */ /* behind the root rcu_node's gpnum. */ unsigned long qsmask; /* CPUs or groups that need to switch in */ @@ -161,11 +161,11 @@ struct rcu_node { /* Per-CPU data for read-copy update. */ struct rcu_data { /* 1) quiescent-state and grace-period handling : */ - long completed; /* Track rsp->completed gp number */ + unsigned long completed; /* Track rsp->completed gp number */ /* in order to detect GP end. */ - long gpnum; /* Highest gp number that this CPU */ + unsigned long gpnum; /* Highest gp number that this CPU */ /* is aware of having started. */ - long passed_quiesc_completed; + unsigned long passed_quiesc_completed; /* Value of completed at time of qs. */ bool passed_quiesc; /* User-mode/idle loop etc. */ bool qs_pending; /* Core waits for quiesc state. */ @@ -221,14 +221,14 @@ struct rcu_data { unsigned long resched_ipi; /* Sent a resched IPI. */ /* 5) __rcu_pending() statistics. */ - long n_rcu_pending; /* rcu_pending() calls since boot. */ - long n_rp_qs_pending; - long n_rp_cb_ready; - long n_rp_cpu_needs_gp; - long n_rp_gp_completed; - long n_rp_gp_started; - long n_rp_need_fqs; - long n_rp_need_nothing; + unsigned long n_rcu_pending; /* rcu_pending() calls since boot. */ + unsigned long n_rp_qs_pending; + unsigned long n_rp_cb_ready; + unsigned long n_rp_cpu_needs_gp; + unsigned long n_rp_gp_completed; + unsigned long n_rp_gp_started; + unsigned long n_rp_need_fqs; + unsigned long n_rp_need_nothing; int cpu; }; @@ -255,6 +255,9 @@ struct rcu_data { #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ +#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) +#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) + /* * RCU global state, including node hierarchy. This hierarchy is * represented in "heap" form in a dense array. The root (first level) @@ -283,8 +286,8 @@ struct rcu_state { /* period because */ /* force_quiescent_state() */ /* was running. */ - long gpnum; /* Current gp number. */ - long completed; /* # of last completed gp. */ + unsigned long gpnum; /* Current gp number. */ + unsigned long completed; /* # of last completed gp. */ /* End of fields guarded by root rcu_node's lock. */ diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index 9d2c88423b31..d45db2e35d27 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -50,7 +50,7 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%3d%cc=%ld g=%ld pq=%d pqc=%ld qp=%d", + seq_printf(m, "%3d%cc=%lu g=%lu pq=%d pqc=%lu qp=%d", rdp->cpu, cpu_is_offline(rdp->cpu) ? '!' : ' ', rdp->completed, rdp->gpnum, @@ -105,7 +105,7 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%d,%s,%ld,%ld,%d,%ld,%d", + seq_printf(m, "%d,%s,%lu,%lu,%d,%lu,%d", rdp->cpu, cpu_is_offline(rdp->cpu) ? "\"N\"" : "\"Y\"", rdp->completed, rdp->gpnum, @@ -155,13 +155,13 @@ static const struct file_operations rcudata_csv_fops = { static void print_one_rcu_state(struct seq_file *m, struct rcu_state *rsp) { - long gpnum; + unsigned long gpnum; int level = 0; int phase; struct rcu_node *rnp; gpnum = rsp->gpnum; - seq_printf(m, "c=%ld g=%ld s=%d jfq=%ld j=%x " + seq_printf(m, "c=%lu g=%lu s=%d jfq=%ld j=%x " "nfqs=%lu/nfqsng=%lu(%lu) fqlh=%lu oqlen=%ld\n", rsp->completed, gpnum, rsp->signaled, (long)(rsp->jiffies_force_qs - jiffies), @@ -215,12 +215,12 @@ static const struct file_operations rcuhier_fops = { static int show_rcugp(struct seq_file *m, void *unused) { #ifdef CONFIG_TREE_PREEMPT_RCU - seq_printf(m, "rcu_preempt: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_preempt: completed=%ld gpnum=%lu\n", rcu_preempt_state.completed, rcu_preempt_state.gpnum); #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ - seq_printf(m, "rcu_sched: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_sched: completed=%ld gpnum=%lu\n", rcu_sched_state.completed, rcu_sched_state.gpnum); - seq_printf(m, "rcu_bh: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_bh: completed=%ld gpnum=%lu\n", rcu_bh_state.completed, rcu_bh_state.gpnum); return 0; } -- cgit v1.2.2 From 1304afb225288a2e250d6a7495462c28e5509cbb Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:05:02 -0800 Subject: rcu: Convert to raw_spinlocks The spinlocks in rcutree need to be real spinlocks in preempt-rt. Convert them to raw_spinlocks. Signed-off-by: Thomas Gleixner Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-18-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 132 ++++++++++++++++++++++++------------------------ kernel/rcutree.h | 6 +-- kernel/rcutree_plugin.h | 46 ++++++++--------- 3 files changed, 92 insertions(+), 92 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index dd0d31dffcdc..65a807b4f58c 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -66,11 +66,11 @@ static struct lock_class_key rcu_node_class[NUM_RCU_LVLS]; .signaled = RCU_GP_IDLE, \ .gpnum = -300, \ .completed = -300, \ - .onofflock = __SPIN_LOCK_UNLOCKED(&name.onofflock), \ + .onofflock = __RAW_SPIN_LOCK_UNLOCKED(&name.onofflock), \ .orphan_cbs_list = NULL, \ .orphan_cbs_tail = &name.orphan_cbs_list, \ .orphan_qlen = 0, \ - .fqslock = __SPIN_LOCK_UNLOCKED(&name.fqslock), \ + .fqslock = __RAW_SPIN_LOCK_UNLOCKED(&name.fqslock), \ .n_force_qs = 0, \ .n_force_qs_ngp = 0, \ } @@ -457,10 +457,10 @@ static void print_other_cpu_stall(struct rcu_state *rsp) /* Only let one CPU complain about others per time interval. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); delta = jiffies - rsp->jiffies_stall; if (delta < RCU_STALL_RAT_DELAY || !rcu_gp_in_progress(rsp)) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; @@ -470,7 +470,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) * due to CPU offlining. */ rcu_print_task_stall(rnp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); /* OK, time to rat on our buddy... */ @@ -499,11 +499,11 @@ static void print_cpu_stall(struct rcu_state *rsp) smp_processor_id(), jiffies - rsp->gp_start); trigger_all_cpu_backtrace(); - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); if (ULONG_CMP_GE(jiffies, rsp->jiffies_stall)) rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); set_need_resched(); /* kick ourselves to get things going. */ } @@ -563,12 +563,12 @@ static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_save(flags); rnp = rdp->mynode; if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ - !spin_trylock(&rnp->lock)) { /* irqs already off, retry later. */ + !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } __note_new_gpnum(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -627,12 +627,12 @@ rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_save(flags); rnp = rdp->mynode; if (rdp->completed == ACCESS_ONCE(rnp->completed) || /* outside lock. */ - !spin_trylock(&rnp->lock)) { /* irqs already off, retry later. */ + !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } __rcu_process_gp_end(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -681,10 +681,10 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) if (cpu_needs_another_gp(rsp, rdp)) rsp->fqs_need_gp = 1; if (rnp->completed == rsp->completed) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* * Propagate new ->completed value to rcu_node structures @@ -692,9 +692,9 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) * of the next grace period to process their callbacks. */ rcu_for_each_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->completed = rsp->completed; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } local_irq_restore(flags); return; @@ -715,15 +715,15 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) rnp->completed = rsp->completed; rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state OK. */ rcu_start_gp_per_cpu(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } - spin_unlock(&rnp->lock); /* leave irqs disabled. */ + raw_spin_unlock(&rnp->lock); /* leave irqs disabled. */ /* Exclude any concurrent CPU-hotplug operations. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ /* * Set the quiescent-state-needed bits in all the rcu_node @@ -743,21 +743,21 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) * irqs disabled. */ rcu_for_each_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rcu_preempt_check_blocked_tasks(rnp); rnp->qsmask = rnp->qsmaskinit; rnp->gpnum = rsp->gpnum; rnp->completed = rsp->completed; if (rnp == rdp->mynode) rcu_start_gp_per_cpu(rsp, rnp, rdp); - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } rnp = rcu_get_root(rsp); - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state now OK. */ - spin_unlock(&rnp->lock); /* irqs remain disabled. */ - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } /* @@ -796,14 +796,14 @@ rcu_report_qs_rnp(unsigned long mask, struct rcu_state *rsp, if (!(rnp->qsmask & mask)) { /* Our bit has already been cleared, so done. */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } rnp->qsmask &= ~mask; if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { /* Other bits still set at this level, so done. */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } mask = rnp->grpmask; @@ -813,10 +813,10 @@ rcu_report_qs_rnp(unsigned long mask, struct rcu_state *rsp, break; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); rnp_c = rnp; rnp = rnp->parent; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); WARN_ON_ONCE(rnp_c->qsmask); } @@ -845,7 +845,7 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long las struct rcu_node *rnp; rnp = rdp->mynode; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); if (lastcomp != rnp->completed) { /* @@ -857,12 +857,12 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long las * race occurred. */ rdp->passed_quiesc = 0; /* try again later! */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } mask = rdp->grpmask; if ((rnp->qsmask & mask) == 0) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } else { rdp->qs_pending = 0; @@ -926,7 +926,7 @@ static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) if (rdp->nxtlist == NULL) return; /* irqs disabled, so comparison is stable. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ *rsp->orphan_cbs_tail = rdp->nxtlist; rsp->orphan_cbs_tail = rdp->nxttail[RCU_NEXT_TAIL]; rdp->nxtlist = NULL; @@ -934,7 +934,7 @@ static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) rdp->nxttail[i] = &rdp->nxtlist; rsp->orphan_qlen += rdp->qlen; rdp->qlen = 0; - spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ + raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ } /* @@ -945,10 +945,10 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) unsigned long flags; struct rcu_data *rdp; - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); rdp = rsp->rda[smp_processor_id()]; if (rsp->orphan_cbs_list == NULL) { - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); return; } *rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_list; @@ -957,7 +957,7 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) rsp->orphan_cbs_list = NULL; rsp->orphan_cbs_tail = &rsp->orphan_cbs_list; rsp->orphan_qlen = 0; - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } /* @@ -973,23 +973,23 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) struct rcu_node *rnp; /* Exclude any attempts to start a new grace period. */ - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); /* Remove the outgoing CPU from the masks in the rcu_node hierarchy. */ rnp = rdp->mynode; /* this is the outgoing CPU's rnp. */ mask = rdp->grpmask; /* rnp->grplo is constant. */ do { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->qsmaskinit &= ~mask; if (rnp->qsmaskinit != 0) { if (rnp != rdp->mynode) - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ break; } if (rnp == rdp->mynode) need_report = rcu_preempt_offline_tasks(rsp, rnp, rdp); else - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ mask = rnp->grpmask; rnp = rnp->parent; } while (rnp != NULL); @@ -1000,12 +1000,12 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) * because invoking rcu_report_unblock_qs_rnp() with ->onofflock * held leads to deadlock. */ - spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ + raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ rnp = rdp->mynode; if (need_report & RCU_OFL_TASKS_NORM_GP) rcu_report_unblock_qs_rnp(rnp, flags); else - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); if (need_report & RCU_OFL_TASKS_EXP_GP) rcu_report_exp_rnp(rsp, rnp); @@ -1176,13 +1176,13 @@ static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) rcu_for_each_leaf_node(rsp, rnp) { mask = 0; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); if (!rcu_gp_in_progress(rsp)) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } if (rnp->qsmask == 0) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); continue; } cpu = rnp->grplo; @@ -1197,7 +1197,7 @@ static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) rcu_report_qs_rnp(mask, rsp, rnp, flags); continue; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } } @@ -1212,18 +1212,18 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ - if (!spin_trylock_irqsave(&rsp->fqslock, flags)) { + if (!raw_spin_trylock_irqsave(&rsp->fqslock, flags)) { rsp->n_force_qs_lh++; /* Inexact, can lose counts. Tough! */ return; /* Someone else is already on the job. */ } if (relaxed && ULONG_CMP_GE(rsp->jiffies_force_qs, jiffies)) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ goto unlock_fqs_ret; /* no GP in progress, time updated. */ } rsp->fqs_active = 1; @@ -1235,13 +1235,13 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_SAVE_DYNTICK: - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ force_qs_rnp(rsp, dyntick_save_progress_counter); - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; break; @@ -1249,24 +1249,24 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ force_qs_rnp(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ break; } rsp->fqs_active = 0; if (rsp->fqs_need_gp) { - spin_unlock(&rsp->fqslock); /* irqs remain disabled */ + raw_spin_unlock(&rsp->fqslock); /* irqs remain disabled */ rsp->fqs_need_gp = 0; rcu_start_gp(rsp, flags); /* releases rnp->lock */ return; } - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ unlock_fqs_ret: - spin_unlock_irqrestore(&rsp->fqslock, flags); + raw_spin_unlock_irqrestore(&rsp->fqslock, flags); } #else /* #ifdef CONFIG_SMP */ @@ -1308,7 +1308,7 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) /* Does this CPU require a not-yet-started grace period? */ if (cpu_needs_another_gp(rsp, rdp)) { - spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags); + raw_spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags); rcu_start_gp(rsp, flags); /* releases above lock */ } @@ -1373,7 +1373,7 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), unsigned long nestflag; struct rcu_node *rnp_root = rcu_get_root(rsp); - spin_lock_irqsave(&rnp_root->lock, nestflag); + raw_spin_lock_irqsave(&rnp_root->lock, nestflag); rcu_start_gp(rsp, nestflag); /* releases rnp_root->lock. */ } @@ -1662,7 +1662,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo); rdp->nxtlist = NULL; for (i = 0; i < RCU_NEXT_SIZE; i++) @@ -1672,7 +1672,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) rdp->dynticks = &per_cpu(rcu_dynticks, cpu); #endif /* #ifdef CONFIG_NO_HZ */ rdp->cpu = cpu; - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -1690,7 +1690,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); rdp->passed_quiesc = 0; /* We could be racing with new GP, */ rdp->qs_pending = 1; /* so set up to respond to current GP. */ rdp->beenonline = 1; /* We have now been online. */ @@ -1698,7 +1698,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) rdp->qlen_last_fqs_check = 0; rdp->n_force_qs_snap = rsp->n_force_qs; rdp->blimit = blimit; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* * A new grace period might start here. If so, we won't be part @@ -1706,14 +1706,14 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) */ /* Exclude any attempts to start a new GP on large systems. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ /* Add CPU to rcu_node bitmasks. */ rnp = rdp->mynode; mask = rdp->grpmask; do { /* Exclude any attempts to start a new GP on small systems. */ - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->qsmaskinit |= mask; mask = rnp->grpmask; if (rnp == rdp->mynode) { @@ -1721,11 +1721,11 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) rdp->completed = rnp->completed; rdp->passed_quiesc_completed = rnp->completed - 1; } - spin_unlock(&rnp->lock); /* irqs already disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs already disabled. */ rnp = rnp->parent; } while (rnp != NULL && !(rnp->qsmaskinit & mask)); - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } static void __cpuinit rcu_online_cpu(int cpu) @@ -1832,7 +1832,7 @@ static void __init rcu_init_one(struct rcu_state *rsp) cpustride *= rsp->levelspread[i]; rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { - spin_lock_init(&rnp->lock); + raw_spin_lock_init(&rnp->lock); lockdep_set_class_and_name(&rnp->lock, &rcu_node_class[i], buf[i]); rnp->gpnum = 0; diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 7495fed49c30..6a82c34ce669 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -90,7 +90,7 @@ struct rcu_dynticks { * Definition for node within the RCU grace-period-detection hierarchy. */ struct rcu_node { - spinlock_t lock; /* Root rcu_node's lock protects some */ + raw_spinlock_t lock; /* Root rcu_node's lock protects some */ /* rcu_state fields as well as following. */ unsigned long gpnum; /* Current grace period for this node. */ /* This will either be equal to or one */ @@ -291,7 +291,7 @@ struct rcu_state { /* End of fields guarded by root rcu_node's lock. */ - spinlock_t onofflock; /* exclude on/offline and */ + raw_spinlock_t onofflock; /* exclude on/offline and */ /* starting new GP. Also */ /* protects the following */ /* orphan_cbs fields. */ @@ -301,7 +301,7 @@ struct rcu_state { /* going offline. */ struct rcu_head **orphan_cbs_tail; /* And tail pointer. */ long orphan_qlen; /* Number of orphaned cbs. */ - spinlock_t fqslock; /* Only one task forcing */ + raw_spinlock_t fqslock; /* Only one task forcing */ /* quiescent states. */ unsigned long jiffies_force_qs; /* Time at which to invoke */ /* force_quiescent_state(). */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index a82566696b0b..a8b2e834fd3a 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -111,7 +111,7 @@ static void rcu_preempt_note_context_switch(int cpu) /* Possibly blocking in an RCU read-side critical section. */ rdp = rcu_preempt_state.rda[cpu]; rnp = rdp->mynode; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; t->rcu_blocked_node = rnp; @@ -132,7 +132,7 @@ static void rcu_preempt_note_context_switch(int cpu) WARN_ON_ONCE(!list_empty(&t->rcu_node_entry)); phase = (rnp->gpnum + !(rnp->qsmask & rdp->grpmask)) & 0x1; list_add(&t->rcu_node_entry, &rnp->blocked_tasks[phase]); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -189,7 +189,7 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) struct rcu_node *rnp_p; if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; /* Still need more quiescent states! */ } @@ -206,8 +206,8 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) /* Report up the rest of the hierarchy. */ mask = rnp->grpmask; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ - spin_lock(&rnp_p->lock); /* irqs already disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_lock(&rnp_p->lock); /* irqs already disabled. */ rcu_report_qs_rnp(mask, &rcu_preempt_state, rnp_p, flags); } @@ -257,10 +257,10 @@ static void rcu_read_unlock_special(struct task_struct *t) */ for (;;) { rnp = t->rcu_blocked_node; - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ if (rnp == t->rcu_blocked_node) break; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } empty = !rcu_preempted_readers(rnp); empty_exp = !rcu_preempted_readers_exp(rnp); @@ -274,7 +274,7 @@ static void rcu_read_unlock_special(struct task_struct *t) * Note that rcu_report_unblock_qs_rnp() releases rnp->lock. */ if (empty) - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); else rcu_report_unblock_qs_rnp(rnp, flags); @@ -324,12 +324,12 @@ static void rcu_print_task_stall(struct rcu_node *rnp) struct task_struct *t; if (rcu_preempted_readers(rnp)) { - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); phase = rnp->gpnum & 0x1; lp = &rnp->blocked_tasks[phase]; list_for_each_entry(t, lp, rcu_node_entry) printk(" P%d", t->pid); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } } @@ -400,11 +400,11 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, lp_root = &rnp_root->blocked_tasks[i]; while (!list_empty(lp)) { tp = list_entry(lp->next, typeof(*tp), rcu_node_entry); - spin_lock(&rnp_root->lock); /* irqs already disabled */ + raw_spin_lock(&rnp_root->lock); /* irqs already disabled */ list_del(&tp->rcu_node_entry); tp->rcu_blocked_node = rnp_root; list_add(&tp->rcu_node_entry, lp_root); - spin_unlock(&rnp_root->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp_root->lock); /* irqs remain disabled */ } } return retval; @@ -528,7 +528,7 @@ static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp) unsigned long flags; unsigned long mask; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); for (;;) { if (!sync_rcu_preempt_exp_done(rnp)) break; @@ -537,12 +537,12 @@ static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp) break; } mask = rnp->grpmask; - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ rnp = rnp->parent; - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ rnp->expmask &= ~mask; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -557,11 +557,11 @@ sync_rcu_preempt_exp_init(struct rcu_state *rsp, struct rcu_node *rnp) { int must_wait; - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ list_splice_init(&rnp->blocked_tasks[0], &rnp->blocked_tasks[2]); list_splice_init(&rnp->blocked_tasks[1], &rnp->blocked_tasks[3]); must_wait = rcu_preempted_readers_exp(rnp); - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ if (!must_wait) rcu_report_exp_rnp(rsp, rnp); } @@ -606,13 +606,13 @@ void synchronize_rcu_expedited(void) /* force all RCU readers onto blocked_tasks[]. */ synchronize_sched_expedited(); - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); /* Initialize ->expmask for all non-leaf rcu_node structures. */ rcu_for_each_nonleaf_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->expmask = rnp->qsmaskinit; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } /* Snapshot current state of ->blocked_tasks[] lists. */ @@ -621,7 +621,7 @@ void synchronize_rcu_expedited(void) if (NUM_RCU_NODES > 1) sync_rcu_preempt_exp_init(rsp, rcu_get_root(rsp)); - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); /* Wait for snapshotted ->blocked_tasks[] lists to drain. */ rnp = rcu_get_root(rsp); @@ -756,7 +756,7 @@ static int rcu_preempted_readers(struct rcu_node *rnp) /* Because preemptible RCU does not exist, no quieting of tasks. */ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } #endif /* #ifdef CONFIG_HOTPLUG_CPU */ -- cgit v1.2.2 From 3acd9eb31c5f7eb97cb2009fa41472710fb4a10f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:05:03 -0800 Subject: rcu: Fix deadlock in TREE_PREEMPT_RCU CPU stall detection Under TREE_PREEMPT_RCU, print_other_cpu_stall() invokes rcu_print_task_stall() with the root rcu_node structure's ->lock held, and rcu_print_task_stall() acquires that same lock for self-deadlock. Fix this by removing the lock acquisition from rcu_print_task_stall(), and making all callers acquire the lock instead. Tested-by: John Kacur Tested-by: Thomas Gleixner Located-by: Thomas Gleixner Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-19-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 2 ++ kernel/rcutree_plugin.h | 3 --- 2 files changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 65a807b4f58c..b07be37d2aa3 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -476,7 +476,9 @@ static void print_other_cpu_stall(struct rcu_state *rsp) printk(KERN_ERR "INFO: RCU detected CPU stalls:"); rcu_for_each_leaf_node(rsp, rnp) { + raw_spin_lock_irqsave(&rnp->lock, flags); rcu_print_task_stall(rnp); + raw_spin_unlock_irqrestore(&rnp->lock, flags); if (rnp->qsmask == 0) continue; for (cpu = 0; cpu <= rnp->grphi - rnp->grplo; cpu++) diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index a8b2e834fd3a..aecfe37e0117 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -318,18 +318,15 @@ EXPORT_SYMBOL_GPL(__rcu_read_unlock); */ static void rcu_print_task_stall(struct rcu_node *rnp) { - unsigned long flags; struct list_head *lp; int phase; struct task_struct *t; if (rcu_preempted_readers(rnp)) { - raw_spin_lock_irqsave(&rnp->lock, flags); phase = rnp->gpnum & 0x1; lp = &rnp->blocked_tasks[phase]; list_for_each_entry(t, lp, rcu_node_entry) printk(" P%d", t->pid); - raw_spin_unlock_irqrestore(&rnp->lock, flags); } } -- cgit v1.2.2 From 6155fec92e85f07d99e9746234496215443ffb0d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:05:04 -0800 Subject: rcu: Fix rcutorture mod_timer argument to delay one jiffy The current "mod_timer(&t, 1)" potentially makes the timer fire immediately, change this to wait one jiffy. Signed-off-by: Dan Carpenter Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-20-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutorture.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 5f43f30fcd1d..258cdf0a91eb 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -853,7 +853,7 @@ rcu_torture_reader(void *arg) do { if (irqreader && cur_ops->irq_capable) { if (!timer_pending(&t)) - mod_timer(&t, 1); + mod_timer(&t, jiffies + 1); } idx = cur_ops->readlock(); completed = cur_ops->completed(); -- cgit v1.2.2 From 1ed509a225008c9e8c0644fbd22168e09a7383a0 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 22 Feb 2010 17:05:05 -0800 Subject: rcu: Add RCU_CPU_STALL_VERBOSE to dump detailed per-task information When RCU detects a grace-period stall, it currently just prints out the PID of any tasks doing the stalling. This patch adds RCU_CPU_STALL_VERBOSE, which enables the more-verbose reporting from sched_show_task(). Suggested-by: Thomas Gleixner Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-21-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 4 ++++ kernel/rcutree.h | 1 + kernel/rcutree_plugin.h | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 57 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index b07be37d2aa3..525d39810616 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -489,6 +489,10 @@ static void print_other_cpu_stall(struct rcu_state *rsp) smp_processor_id(), (long)(jiffies - rsp->gp_start)); trigger_all_cpu_backtrace(); + /* If so configured, complain about tasks blocking the grace period. */ + + rcu_print_detail_task_stall(rsp); + force_quiescent_state(rsp, 0); /* Kick them all. */ } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 6a82c34ce669..2ceb08388582 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -352,6 +352,7 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags); #endif /* #ifdef CONFIG_HOTPLUG_CPU */ #ifdef CONFIG_RCU_CPU_STALL_DETECTOR +static void rcu_print_detail_task_stall(struct rcu_state *rsp); static void rcu_print_task_stall(struct rcu_node *rnp); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp); diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index aecfe37e0117..3516de7091a1 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -312,6 +312,50 @@ EXPORT_SYMBOL_GPL(__rcu_read_unlock); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR +#ifdef CONFIG_RCU_CPU_STALL_VERBOSE + +/* + * Dump detailed information for all tasks blocking the current RCU + * grace period on the specified rcu_node structure. + */ +static void rcu_print_detail_task_stall_rnp(struct rcu_node *rnp) +{ + unsigned long flags; + struct list_head *lp; + int phase; + struct task_struct *t; + + if (rcu_preempted_readers(rnp)) { + raw_spin_lock_irqsave(&rnp->lock, flags); + phase = rnp->gpnum & 0x1; + lp = &rnp->blocked_tasks[phase]; + list_for_each_entry(t, lp, rcu_node_entry) + sched_show_task(t); + raw_spin_unlock_irqrestore(&rnp->lock, flags); + } +} + +/* + * Dump detailed information for all tasks blocking the current RCU + * grace period. + */ +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ + struct rcu_node *rnp = rcu_get_root(rsp); + + rcu_print_detail_task_stall_rnp(rnp); + rcu_for_each_leaf_node(rsp, rnp) + rcu_print_detail_task_stall_rnp(rnp); +} + +#else /* #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */ + +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */ + /* * Scan the current list of tasks blocked within RCU read-side critical * sections, printing out the tid of each. @@ -760,6 +804,14 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) #ifdef CONFIG_RCU_CPU_STALL_DETECTOR +/* + * Because preemptable RCU does not exist, we never have to check for + * tasks blocked within RCU read-side critical sections. + */ +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ +} + /* * Because preemptable RCU does not exist, we never have to check for * tasks blocked within RCU read-side critical sections. -- cgit v1.2.2 From c50cc75271759373bd89a036eec4d4269b291616 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 25 Feb 2010 12:02:13 +0100 Subject: sched, cgroups: Fix module export I have exported it in d11c563 - but cgroups.c did not have module.h included ... Cc: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1266887105-1528-6-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/cgroup.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b1a0f5a528fe..4fd90e129772 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -23,6 +23,7 @@ */ #include +#include #include #include #include -- cgit v1.2.2 From 86c38a31aa7f2dd6e74a262710bf8ebf7455acc5 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Wed, 24 Feb 2010 13:59:23 -0500 Subject: tracing: Fix ftrace_event_call alignment for use with gcc 4.5 GCC 4.5 introduces behavior that forces the alignment of structures to use the largest possible value. The default value is 32 bytes, so if some structures are defined with a 4-byte alignment and others aren't declared with an alignment constraint at all - it will align at 32-bytes. For things like the ftrace events, this results in a non-standard array. When initializing the ftrace subsystem, we traverse the _ftrace_events section and call the initialization callback for each event. When the structures are misaligned, we could be treating another part of the structure (or the zeroed out space between them) as a function pointer. This patch forces the alignment for all the ftrace_event_call structures to 4 bytes. Without this patch, the kernel fails to boot very early when built with gcc 4.5. It's trivial to check the alignment of the members of the array, so it might be worthwhile to add something to the build system to do that automatically. Unfortunately, that only covers this case. I've asked one of the gcc developers about adding a warning when this condition is seen. Cc: stable@kernel.org Signed-off-by: Jeff Mahoney LKML-Reference: <4B85770B.6010901@suse.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b477fce41edf..fd05bcaf91b0 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -792,7 +792,8 @@ extern const char *__stop___trace_bprintk_fmt[]; #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, id, tstruct, print) \ - extern struct ftrace_event_call event_##call; + extern struct ftrace_event_call \ + __attribute__((__aligned__(4))) event_##call; #undef FTRACE_ENTRY_DUP #define FTRACE_ENTRY_DUP(call, struct_name, id, tstruct, print) \ FTRACE_ENTRY(call, struct_name, id, PARAMS(tstruct), PARAMS(print)) -- cgit v1.2.2 From 1ab83a89411556c4d5914dcf0d5da190178ae7db Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Feb 2010 15:28:14 +0800 Subject: tracing: Remove CONFIG_TRACE_POWER from kernel config The power tracer has been converted to power trace events. Acked-by: Frederic Weisbecker Signed-off-by: Li Zefan LKML-Reference: <4B84D50E.4070806@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/Kconfig | 9 --------- 1 file changed, 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 6c22d8a2f289..ca2d3a8778b6 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -330,15 +330,6 @@ config BRANCH_TRACER Say N if unsure. -config POWER_TRACER - bool "Trace power consumption behavior" - depends on X86 - select GENERIC_TRACER - help - This tracer helps developers to analyze and optimize the kernel's - power management decisions, specifically the C-state and P-state - behavior. - config KSYM_TRACER bool "Trace read and write access on kernel memory locations" depends on HAVE_HW_BREAKPOINT -- cgit v1.2.2 From 6574658b3bc7c408581629de5efb809f125cce8c Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Wed, 24 Feb 2010 15:40:22 +0800 Subject: tracing: Fix typo in prof_sysexit_enable() Signed-off-by: Wenji Huang LKML-Reference: <1266997226-6833-1-git-send-email-wenji.huang@oracle.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 49cea70fbf6d..8cdda95da81a 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -603,7 +603,7 @@ int prof_sysexit_enable(struct ftrace_event_call *call) ret = register_trace_sys_exit(prof_syscall_exit); if (ret) { pr_info("event trace: Could not activate" - "syscall entry trace point"); + "syscall exit trace point"); } else { set_bit(num, enabled_prof_exit_syscalls); sys_prof_refcount_exit++; -- cgit v1.2.2 From a5efd925115cbc1f90195dca9a25f7b8daa10c37 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Wed, 24 Feb 2010 15:40:23 +0800 Subject: tracing: Fix typo of info text in trace_kprobe.c Signed-off-by: Wenji Huang LKML-Reference: <1266997226-6833-2-git-send-email-wenji.huang@oracle.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index c99029916c76..8d4bd16d6f70 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -651,12 +651,12 @@ static int create_trace_probe(int argc, char **argv) event = strchr(group, '/') + 1; event[-1] = '\0'; if (strlen(group) == 0) { - pr_info("Group name is not specifiled\n"); + pr_info("Group name is not specified\n"); return -EINVAL; } } if (strlen(event) == 0) { - pr_info("Event name is not specifiled\n"); + pr_info("Event name is not specified\n"); return -EINVAL; } } -- cgit v1.2.2 From c85f3a91f84d5a85f179c2504bb7a39370c82b41 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Wed, 24 Feb 2010 15:40:24 +0800 Subject: tracing: Remove unnecessary variable in print_graph_return The "cpu" variable is declared at the start of the function and also within a branch, with the exact same initialization. Remove the local variable of the same name in the branch. Signed-off-by: Wenji Huang LKML-Reference: <1266997226-6833-3-git-send-email-wenji.huang@oracle.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 616b135c9eb9..112561df2a0a 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -855,7 +855,6 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, int i; if (data) { - int cpu = iter->cpu; int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); /* -- cgit v1.2.2 From 7b60997f73865b019e595720185c85285ca3df9a Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Wed, 24 Feb 2010 15:40:26 +0800 Subject: tracing: Simplify memory recycle of trace_define_field Discard freeing field->type since it is not necessary. Reviewed-by: Li Zefan Signed-off-by: Wenji Huang LKML-Reference: <1266997226-6833-5-git-send-email-wenji.huang@oracle.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index c2a3077b7353..3f972ad98d04 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -60,10 +60,8 @@ int trace_define_field(struct ftrace_event_call *call, const char *type, return 0; err: - if (field) { + if (field) kfree(field->name); - kfree(field->type); - } kfree(field); return -ENOMEM; -- cgit v1.2.2 From 4610ee1d3638fa05ba8e87ccfa971db8e4033ae7 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Thu, 25 Feb 2010 08:33:59 -0500 Subject: kprobes: Introduce generic insn_slot framework Make insn_slot framework support various size slots. Current insn_slot just supports one-size instruction buffer slot. However, kprobes jump optimization needs larger size buffers. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Srikar Dronamraju Cc: Christoph Hellwig Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Anders Kaseorg Cc: Tim Abbott Cc: Andi Kleen Cc: Jason Baron Cc: Mathieu Desnoyers Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli LKML-Reference: <20100225133358.6725.82430.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Srikar Dronamraju Cc: Christoph Hellwig Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: H. Peter Anvin Cc: Anders Kaseorg Cc: Tim Abbott Cc: Andi Kleen Cc: Jason Baron Cc: Mathieu Desnoyers --- kernel/kprobes.c | 104 ++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 65 insertions(+), 39 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index ccec774c716d..78105623d739 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -105,57 +105,74 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { * stepping on the instruction on a vmalloced/kmalloced/data page * is a recipe for disaster */ -#define INSNS_PER_PAGE (PAGE_SIZE/(MAX_INSN_SIZE * sizeof(kprobe_opcode_t))) - struct kprobe_insn_page { struct list_head list; kprobe_opcode_t *insns; /* Page of instruction slots */ - char slot_used[INSNS_PER_PAGE]; int nused; int ngarbage; + char slot_used[]; +}; + +#define KPROBE_INSN_PAGE_SIZE(slots) \ + (offsetof(struct kprobe_insn_page, slot_used) + \ + (sizeof(char) * (slots))) + +struct kprobe_insn_cache { + struct list_head pages; /* list of kprobe_insn_page */ + size_t insn_size; /* size of instruction slot */ + int nr_garbage; }; +static int slots_per_page(struct kprobe_insn_cache *c) +{ + return PAGE_SIZE/(c->insn_size * sizeof(kprobe_opcode_t)); +} + enum kprobe_slot_state { SLOT_CLEAN = 0, SLOT_DIRTY = 1, SLOT_USED = 2, }; -static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_pages */ -static LIST_HEAD(kprobe_insn_pages); -static int kprobe_garbage_slots; -static int collect_garbage_slots(void); +static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_slots */ +static struct kprobe_insn_cache kprobe_insn_slots = { + .pages = LIST_HEAD_INIT(kprobe_insn_slots.pages), + .insn_size = MAX_INSN_SIZE, + .nr_garbage = 0, +}; +static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c); /** * __get_insn_slot() - Find a slot on an executable page for an instruction. * We allocate an executable page if there's no room on existing ones. */ -static kprobe_opcode_t __kprobes *__get_insn_slot(void) +static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) { struct kprobe_insn_page *kip; retry: - list_for_each_entry(kip, &kprobe_insn_pages, list) { - if (kip->nused < INSNS_PER_PAGE) { + list_for_each_entry(kip, &c->pages, list) { + if (kip->nused < slots_per_page(c)) { int i; - for (i = 0; i < INSNS_PER_PAGE; i++) { + for (i = 0; i < slots_per_page(c); i++) { if (kip->slot_used[i] == SLOT_CLEAN) { kip->slot_used[i] = SLOT_USED; kip->nused++; - return kip->insns + (i * MAX_INSN_SIZE); + return kip->insns + (i * c->insn_size); } } - /* Surprise! No unused slots. Fix kip->nused. */ - kip->nused = INSNS_PER_PAGE; + /* kip->nused is broken. Fix it. */ + kip->nused = slots_per_page(c); + WARN_ON(1); } } /* If there are any garbage slots, collect it and try again. */ - if (kprobe_garbage_slots && collect_garbage_slots() == 0) { + if (c->nr_garbage && collect_garbage_slots(c) == 0) goto retry; - } - /* All out of space. Need to allocate a new page. Use slot 0. */ - kip = kmalloc(sizeof(struct kprobe_insn_page), GFP_KERNEL); + + /* All out of space. Need to allocate a new page. */ + kip = kmalloc(KPROBE_INSN_PAGE_SIZE(slots_per_page(c)), GFP_KERNEL); if (!kip) return NULL; @@ -170,20 +187,23 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(void) return NULL; } INIT_LIST_HEAD(&kip->list); - list_add(&kip->list, &kprobe_insn_pages); - memset(kip->slot_used, SLOT_CLEAN, INSNS_PER_PAGE); + memset(kip->slot_used, SLOT_CLEAN, slots_per_page(c)); kip->slot_used[0] = SLOT_USED; kip->nused = 1; kip->ngarbage = 0; + list_add(&kip->list, &c->pages); return kip->insns; } + kprobe_opcode_t __kprobes *get_insn_slot(void) { - kprobe_opcode_t *ret; + kprobe_opcode_t *ret = NULL; + mutex_lock(&kprobe_insn_mutex); - ret = __get_insn_slot(); + ret = __get_insn_slot(&kprobe_insn_slots); mutex_unlock(&kprobe_insn_mutex); + return ret; } @@ -199,7 +219,7 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx) * so as not to have to set it up again the * next time somebody inserts a probe. */ - if (!list_is_singular(&kprobe_insn_pages)) { + if (!list_is_singular(&kip->list)) { list_del(&kip->list); module_free(NULL, kip->insns); kfree(kip); @@ -209,49 +229,55 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx) return 0; } -static int __kprobes collect_garbage_slots(void) +static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c) { struct kprobe_insn_page *kip, *next; /* Ensure no-one is interrupted on the garbages */ synchronize_sched(); - list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) { + list_for_each_entry_safe(kip, next, &c->pages, list) { int i; if (kip->ngarbage == 0) continue; kip->ngarbage = 0; /* we will collect all garbages */ - for (i = 0; i < INSNS_PER_PAGE; i++) { + for (i = 0; i < slots_per_page(c); i++) { if (kip->slot_used[i] == SLOT_DIRTY && collect_one_slot(kip, i)) break; } } - kprobe_garbage_slots = 0; + c->nr_garbage = 0; return 0; } -void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty) +static void __kprobes __free_insn_slot(struct kprobe_insn_cache *c, + kprobe_opcode_t *slot, int dirty) { struct kprobe_insn_page *kip; - mutex_lock(&kprobe_insn_mutex); - list_for_each_entry(kip, &kprobe_insn_pages, list) { - if (kip->insns <= slot && - slot < kip->insns + (INSNS_PER_PAGE * MAX_INSN_SIZE)) { - int i = (slot - kip->insns) / MAX_INSN_SIZE; + list_for_each_entry(kip, &c->pages, list) { + long idx = ((long)slot - (long)kip->insns) / c->insn_size; + if (idx >= 0 && idx < slots_per_page(c)) { + WARN_ON(kip->slot_used[idx] != SLOT_USED); if (dirty) { - kip->slot_used[i] = SLOT_DIRTY; + kip->slot_used[idx] = SLOT_DIRTY; kip->ngarbage++; + if (++c->nr_garbage > slots_per_page(c)) + collect_garbage_slots(c); } else - collect_one_slot(kip, i); - break; + collect_one_slot(kip, idx); + return; } } + /* Could not free this slot. */ + WARN_ON(1); +} - if (dirty && ++kprobe_garbage_slots > INSNS_PER_PAGE) - collect_garbage_slots(); - +void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty) +{ + mutex_lock(&kprobe_insn_mutex); + __free_insn_slot(&kprobe_insn_slots, slot, dirty); mutex_unlock(&kprobe_insn_mutex); } #endif -- cgit v1.2.2 From afd66255b9a48f5851326ddae50e2203fbf71dc9 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Thu, 25 Feb 2010 08:34:07 -0500 Subject: kprobes: Introduce kprobes jump optimization Introduce kprobes jump optimization arch-independent parts. Kprobes uses breakpoint instruction for interrupting execution flow, on some architectures, it can be replaced by a jump instruction and interruption emulation code. This gains kprobs' performance drastically. To enable this feature, set CONFIG_OPTPROBES=y (default y if the arch supports OPTPROBE). Changes in v9: - Fix a bug to optimize probe when enabling. - Check nearby probes can be optimize/unoptimize when disarming/arming kprobes, instead of registering/unregistering. This will help kprobe-tracer because most of probes on it are usually disabled. Changes in v6: - Cleanup coding style for readability. - Add comments around get/put_online_cpus(). Changes in v5: - Use get_online_cpus()/put_online_cpus() for avoiding text_mutex deadlock. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Srikar Dronamraju Cc: Christoph Hellwig Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Anders Kaseorg Cc: Tim Abbott Cc: Andi Kleen Cc: Jason Baron Cc: Mathieu Desnoyers Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli LKML-Reference: <20100225133407.6725.81992.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 461 +++++++++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 410 insertions(+), 51 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 78105623d739..612af2d61614 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -45,6 +45,7 @@ #include #include #include +#include #include #include @@ -280,6 +281,33 @@ void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty) __free_insn_slot(&kprobe_insn_slots, slot, dirty); mutex_unlock(&kprobe_insn_mutex); } +#ifdef CONFIG_OPTPROBES +/* For optimized_kprobe buffer */ +static DEFINE_MUTEX(kprobe_optinsn_mutex); /* Protects kprobe_optinsn_slots */ +static struct kprobe_insn_cache kprobe_optinsn_slots = { + .pages = LIST_HEAD_INIT(kprobe_optinsn_slots.pages), + /* .insn_size is initialized later */ + .nr_garbage = 0, +}; +/* Get a slot for optimized_kprobe buffer */ +kprobe_opcode_t __kprobes *get_optinsn_slot(void) +{ + kprobe_opcode_t *ret = NULL; + + mutex_lock(&kprobe_optinsn_mutex); + ret = __get_insn_slot(&kprobe_optinsn_slots); + mutex_unlock(&kprobe_optinsn_mutex); + + return ret; +} + +void __kprobes free_optinsn_slot(kprobe_opcode_t * slot, int dirty) +{ + mutex_lock(&kprobe_optinsn_mutex); + __free_insn_slot(&kprobe_optinsn_slots, slot, dirty); + mutex_unlock(&kprobe_optinsn_mutex); +} +#endif #endif /* We have preemption disabled.. so it is safe to use __ versions */ @@ -310,23 +338,324 @@ struct kprobe __kprobes *get_kprobe(void *addr) if (p->addr == addr) return p; } + return NULL; } +static int __kprobes aggr_pre_handler(struct kprobe *p, struct pt_regs *regs); + +/* Return true if the kprobe is an aggregator */ +static inline int kprobe_aggrprobe(struct kprobe *p) +{ + return p->pre_handler == aggr_pre_handler; +} + +/* + * Keep all fields in the kprobe consistent + */ +static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) +{ + memcpy(&p->opcode, &old_p->opcode, sizeof(kprobe_opcode_t)); + memcpy(&p->ainsn, &old_p->ainsn, sizeof(struct arch_specific_insn)); +} + +#ifdef CONFIG_OPTPROBES +/* + * Call all pre_handler on the list, but ignores its return value. + * This must be called from arch-dep optimized caller. + */ +void __kprobes opt_pre_handler(struct kprobe *p, struct pt_regs *regs) +{ + struct kprobe *kp; + + list_for_each_entry_rcu(kp, &p->list, list) { + if (kp->pre_handler && likely(!kprobe_disabled(kp))) { + set_kprobe_instance(kp); + kp->pre_handler(kp, regs); + } + reset_kprobe_instance(); + } +} + +/* Return true(!0) if the kprobe is ready for optimization. */ +static inline int kprobe_optready(struct kprobe *p) +{ + struct optimized_kprobe *op; + + if (kprobe_aggrprobe(p)) { + op = container_of(p, struct optimized_kprobe, kp); + return arch_prepared_optinsn(&op->optinsn); + } + + return 0; +} + +/* + * Return an optimized kprobe whose optimizing code replaces + * instructions including addr (exclude breakpoint). + */ +struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr) +{ + int i; + struct kprobe *p = NULL; + struct optimized_kprobe *op; + + /* Don't check i == 0, since that is a breakpoint case. */ + for (i = 1; !p && i < MAX_OPTIMIZED_LENGTH; i++) + p = get_kprobe((void *)(addr - i)); + + if (p && kprobe_optready(p)) { + op = container_of(p, struct optimized_kprobe, kp); + if (arch_within_optimized_kprobe(op, addr)) + return p; + } + + return NULL; +} + +/* Optimization staging list, protected by kprobe_mutex */ +static LIST_HEAD(optimizing_list); + +static void kprobe_optimizer(struct work_struct *work); +static DECLARE_DELAYED_WORK(optimizing_work, kprobe_optimizer); +#define OPTIMIZE_DELAY 5 + +/* Kprobe jump optimizer */ +static __kprobes void kprobe_optimizer(struct work_struct *work) +{ + struct optimized_kprobe *op, *tmp; + + /* Lock modules while optimizing kprobes */ + mutex_lock(&module_mutex); + mutex_lock(&kprobe_mutex); + if (kprobes_all_disarmed) + goto end; + + /* + * Wait for quiesence period to ensure all running interrupts + * are done. Because optprobe may modify multiple instructions + * there is a chance that Nth instruction is interrupted. In that + * case, running interrupt can return to 2nd-Nth byte of jump + * instruction. This wait is for avoiding it. + */ + synchronize_sched(); + + /* + * The optimization/unoptimization refers online_cpus via + * stop_machine() and cpu-hotplug modifies online_cpus. + * And same time, text_mutex will be held in cpu-hotplug and here. + * This combination can cause a deadlock (cpu-hotplug try to lock + * text_mutex but stop_machine can not be done because online_cpus + * has been changed) + * To avoid this deadlock, we need to call get_online_cpus() + * for preventing cpu-hotplug outside of text_mutex locking. + */ + get_online_cpus(); + mutex_lock(&text_mutex); + list_for_each_entry_safe(op, tmp, &optimizing_list, list) { + WARN_ON(kprobe_disabled(&op->kp)); + if (arch_optimize_kprobe(op) < 0) + op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED; + list_del_init(&op->list); + } + mutex_unlock(&text_mutex); + put_online_cpus(); +end: + mutex_unlock(&kprobe_mutex); + mutex_unlock(&module_mutex); +} + +/* Optimize kprobe if p is ready to be optimized */ +static __kprobes void optimize_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + /* Check if the kprobe is disabled or not ready for optimization. */ + if (!kprobe_optready(p) || + (kprobe_disabled(p) || kprobes_all_disarmed)) + return; + + /* Both of break_handler and post_handler are not supported. */ + if (p->break_handler || p->post_handler) + return; + + op = container_of(p, struct optimized_kprobe, kp); + + /* Check there is no other kprobes at the optimized instructions */ + if (arch_check_optimized_kprobe(op) < 0) + return; + + /* Check if it is already optimized. */ + if (op->kp.flags & KPROBE_FLAG_OPTIMIZED) + return; + + op->kp.flags |= KPROBE_FLAG_OPTIMIZED; + list_add(&op->list, &optimizing_list); + if (!delayed_work_pending(&optimizing_work)) + schedule_delayed_work(&optimizing_work, OPTIMIZE_DELAY); +} + +/* Unoptimize a kprobe if p is optimized */ +static __kprobes void unoptimize_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + if ((p->flags & KPROBE_FLAG_OPTIMIZED) && kprobe_aggrprobe(p)) { + op = container_of(p, struct optimized_kprobe, kp); + if (!list_empty(&op->list)) + /* Dequeue from the optimization queue */ + list_del_init(&op->list); + else + /* Replace jump with break */ + arch_unoptimize_kprobe(op); + op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED; + } +} + +/* Remove optimized instructions */ +static void __kprobes kill_optimized_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + op = container_of(p, struct optimized_kprobe, kp); + if (!list_empty(&op->list)) { + /* Dequeue from the optimization queue */ + list_del_init(&op->list); + op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED; + } + /* Don't unoptimize, because the target code will be freed. */ + arch_remove_optimized_kprobe(op); +} + +/* Try to prepare optimized instructions */ +static __kprobes void prepare_optimized_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + op = container_of(p, struct optimized_kprobe, kp); + arch_prepare_optimized_kprobe(op); +} + +/* Free optimized instructions and optimized_kprobe */ +static __kprobes void free_aggr_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + op = container_of(p, struct optimized_kprobe, kp); + arch_remove_optimized_kprobe(op); + kfree(op); +} + +/* Allocate new optimized_kprobe and try to prepare optimized instructions */ +static __kprobes struct kprobe *alloc_aggr_kprobe(struct kprobe *p) +{ + struct optimized_kprobe *op; + + op = kzalloc(sizeof(struct optimized_kprobe), GFP_KERNEL); + if (!op) + return NULL; + + INIT_LIST_HEAD(&op->list); + op->kp.addr = p->addr; + arch_prepare_optimized_kprobe(op); + + return &op->kp; +} + +static void __kprobes init_aggr_kprobe(struct kprobe *ap, struct kprobe *p); + +/* + * Prepare an optimized_kprobe and optimize it + * NOTE: p must be a normal registered kprobe + */ +static __kprobes void try_to_optimize_kprobe(struct kprobe *p) +{ + struct kprobe *ap; + struct optimized_kprobe *op; + + ap = alloc_aggr_kprobe(p); + if (!ap) + return; + + op = container_of(ap, struct optimized_kprobe, kp); + if (!arch_prepared_optinsn(&op->optinsn)) { + /* If failed to setup optimizing, fallback to kprobe */ + free_aggr_kprobe(ap); + return; + } + + init_aggr_kprobe(ap, p); + optimize_kprobe(ap); +} + +static void __kprobes __arm_kprobe(struct kprobe *p) +{ + struct kprobe *old_p; + + /* Check collision with other optimized kprobes */ + old_p = get_optimized_kprobe((unsigned long)p->addr); + if (unlikely(old_p)) + unoptimize_kprobe(old_p); /* Fallback to unoptimized kprobe */ + + arch_arm_kprobe(p); + optimize_kprobe(p); /* Try to optimize (add kprobe to a list) */ +} + +static void __kprobes __disarm_kprobe(struct kprobe *p) +{ + struct kprobe *old_p; + + unoptimize_kprobe(p); /* Try to unoptimize */ + arch_disarm_kprobe(p); + + /* If another kprobe was blocked, optimize it. */ + old_p = get_optimized_kprobe((unsigned long)p->addr); + if (unlikely(old_p)) + optimize_kprobe(old_p); +} + +#else /* !CONFIG_OPTPROBES */ + +#define optimize_kprobe(p) do {} while (0) +#define unoptimize_kprobe(p) do {} while (0) +#define kill_optimized_kprobe(p) do {} while (0) +#define prepare_optimized_kprobe(p) do {} while (0) +#define try_to_optimize_kprobe(p) do {} while (0) +#define __arm_kprobe(p) arch_arm_kprobe(p) +#define __disarm_kprobe(p) arch_disarm_kprobe(p) + +static __kprobes void free_aggr_kprobe(struct kprobe *p) +{ + kfree(p); +} + +static __kprobes struct kprobe *alloc_aggr_kprobe(struct kprobe *p) +{ + return kzalloc(sizeof(struct kprobe), GFP_KERNEL); +} +#endif /* CONFIG_OPTPROBES */ + /* Arm a kprobe with text_mutex */ static void __kprobes arm_kprobe(struct kprobe *kp) { + /* + * Here, since __arm_kprobe() doesn't use stop_machine(), + * this doesn't cause deadlock on text_mutex. So, we don't + * need get_online_cpus(). + */ mutex_lock(&text_mutex); - arch_arm_kprobe(kp); + __arm_kprobe(kp); mutex_unlock(&text_mutex); } /* Disarm a kprobe with text_mutex */ static void __kprobes disarm_kprobe(struct kprobe *kp) { + get_online_cpus(); /* For avoiding text_mutex deadlock */ mutex_lock(&text_mutex); - arch_disarm_kprobe(kp); + __disarm_kprobe(kp); mutex_unlock(&text_mutex); + put_online_cpus(); } /* @@ -395,7 +724,7 @@ static int __kprobes aggr_break_handler(struct kprobe *p, struct pt_regs *regs) void __kprobes kprobes_inc_nmissed_count(struct kprobe *p) { struct kprobe *kp; - if (p->pre_handler != aggr_pre_handler) { + if (!kprobe_aggrprobe(p)) { p->nmissed++; } else { list_for_each_entry_rcu(kp, &p->list, list) @@ -518,15 +847,6 @@ static void __kprobes cleanup_rp_inst(struct kretprobe *rp) free_rp_inst(rp); } -/* - * Keep all fields in the kprobe consistent - */ -static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) -{ - memcpy(&p->opcode, &old_p->opcode, sizeof(kprobe_opcode_t)); - memcpy(&p->ainsn, &old_p->ainsn, sizeof(struct arch_specific_insn)); -} - /* * Add the new probe to ap->list. Fail if this is the * second jprobe at the address - two jprobes can't coexist @@ -534,6 +854,10 @@ static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p) { BUG_ON(kprobe_gone(ap) || kprobe_gone(p)); + + if (p->break_handler || p->post_handler) + unoptimize_kprobe(ap); /* Fall back to normal kprobe */ + if (p->break_handler) { if (ap->break_handler) return -EEXIST; @@ -548,7 +872,7 @@ static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p) ap->flags &= ~KPROBE_FLAG_DISABLED; if (!kprobes_all_disarmed) /* Arm the breakpoint again. */ - arm_kprobe(ap); + __arm_kprobe(ap); } return 0; } @@ -557,12 +881,13 @@ static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p) * Fill in the required fields of the "manager kprobe". Replace the * earlier kprobe in the hlist with the manager kprobe */ -static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p) +static void __kprobes init_aggr_kprobe(struct kprobe *ap, struct kprobe *p) { + /* Copy p's insn slot to ap */ copy_kprobe(p, ap); flush_insn_slot(ap); ap->addr = p->addr; - ap->flags = p->flags; + ap->flags = p->flags & ~KPROBE_FLAG_OPTIMIZED; ap->pre_handler = aggr_pre_handler; ap->fault_handler = aggr_fault_handler; /* We don't care the kprobe which has gone. */ @@ -572,8 +897,9 @@ static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p) ap->break_handler = aggr_break_handler; INIT_LIST_HEAD(&ap->list); - list_add_rcu(&p->list, &ap->list); + INIT_HLIST_NODE(&ap->hlist); + list_add_rcu(&p->list, &ap->list); hlist_replace_rcu(&p->hlist, &ap->hlist); } @@ -587,12 +913,12 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p, int ret = 0; struct kprobe *ap = old_p; - if (old_p->pre_handler != aggr_pre_handler) { - /* If old_p is not an aggr_probe, create new aggr_kprobe. */ - ap = kzalloc(sizeof(struct kprobe), GFP_KERNEL); + if (!kprobe_aggrprobe(old_p)) { + /* If old_p is not an aggr_kprobe, create new aggr_kprobe. */ + ap = alloc_aggr_kprobe(old_p); if (!ap) return -ENOMEM; - add_aggr_kprobe(ap, old_p); + init_aggr_kprobe(ap, old_p); } if (kprobe_gone(ap)) { @@ -611,6 +937,9 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p, */ return ret; + /* Prepare optimized instructions if possible. */ + prepare_optimized_kprobe(ap); + /* * Clear gone flag to prevent allocating new slot again, and * set disabled flag because it is not armed yet. @@ -619,6 +948,7 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p, | KPROBE_FLAG_DISABLED; } + /* Copy ap's insn slot to p */ copy_kprobe(ap, p); return add_new_kprobe(ap, p); } @@ -769,27 +1099,34 @@ int __kprobes register_kprobe(struct kprobe *p) p->nmissed = 0; INIT_LIST_HEAD(&p->list); mutex_lock(&kprobe_mutex); + + get_online_cpus(); /* For avoiding text_mutex deadlock. */ + mutex_lock(&text_mutex); + old_p = get_kprobe(p->addr); if (old_p) { + /* Since this may unoptimize old_p, locking text_mutex. */ ret = register_aggr_kprobe(old_p, p); goto out; } - mutex_lock(&text_mutex); ret = arch_prepare_kprobe(p); if (ret) - goto out_unlock_text; + goto out; INIT_HLIST_NODE(&p->hlist); hlist_add_head_rcu(&p->hlist, &kprobe_table[hash_ptr(p->addr, KPROBE_HASH_BITS)]); if (!kprobes_all_disarmed && !kprobe_disabled(p)) - arch_arm_kprobe(p); + __arm_kprobe(p); + + /* Try to optimize kprobe */ + try_to_optimize_kprobe(p); -out_unlock_text: - mutex_unlock(&text_mutex); out: + mutex_unlock(&text_mutex); + put_online_cpus(); mutex_unlock(&kprobe_mutex); if (probed_mod) @@ -811,7 +1148,7 @@ static int __kprobes __unregister_kprobe_top(struct kprobe *p) return -EINVAL; if (old_p == p || - (old_p->pre_handler == aggr_pre_handler && + (kprobe_aggrprobe(old_p) && list_is_singular(&old_p->list))) { /* * Only probe on the hash list. Disarm only if kprobes are @@ -819,7 +1156,7 @@ static int __kprobes __unregister_kprobe_top(struct kprobe *p) * already have been removed. We save on flushing icache. */ if (!kprobes_all_disarmed && !kprobe_disabled(old_p)) - disarm_kprobe(p); + disarm_kprobe(old_p); hlist_del_rcu(&old_p->hlist); } else { if (p->break_handler && !kprobe_gone(p)) @@ -835,8 +1172,13 @@ noclean: list_del_rcu(&p->list); if (!kprobe_disabled(old_p)) { try_to_disable_aggr_kprobe(old_p); - if (!kprobes_all_disarmed && kprobe_disabled(old_p)) - disarm_kprobe(old_p); + if (!kprobes_all_disarmed) { + if (kprobe_disabled(old_p)) + disarm_kprobe(old_p); + else + /* Try to optimize this probe again */ + optimize_kprobe(old_p); + } } } return 0; @@ -853,7 +1195,7 @@ static void __kprobes __unregister_kprobe_bottom(struct kprobe *p) old_p = list_entry(p->list.next, struct kprobe, list); list_del(&p->list); arch_remove_kprobe(old_p); - kfree(old_p); + free_aggr_kprobe(old_p); } } @@ -1149,7 +1491,7 @@ static void __kprobes kill_kprobe(struct kprobe *p) struct kprobe *kp; p->flags |= KPROBE_FLAG_GONE; - if (p->pre_handler == aggr_pre_handler) { + if (kprobe_aggrprobe(p)) { /* * If this is an aggr_kprobe, we have to list all the * chained probes and mark them GONE. @@ -1158,6 +1500,7 @@ static void __kprobes kill_kprobe(struct kprobe *p) kp->flags |= KPROBE_FLAG_GONE; p->post_handler = NULL; p->break_handler = NULL; + kill_optimized_kprobe(p); } /* * Here, we can remove insn_slot safely, because no thread calls @@ -1267,6 +1610,11 @@ static int __init init_kprobes(void) } } +#if defined(CONFIG_OPTPROBES) && defined(__ARCH_WANT_KPROBES_INSN_SLOT) + /* Init kprobe_optinsn_slots */ + kprobe_optinsn_slots.insn_size = MAX_OPTINSN_SIZE; +#endif + /* By default, kprobes are armed */ kprobes_all_disarmed = false; @@ -1285,7 +1633,7 @@ static int __init init_kprobes(void) #ifdef CONFIG_DEBUG_FS static void __kprobes report_probe(struct seq_file *pi, struct kprobe *p, - const char *sym, int offset,char *modname) + const char *sym, int offset, char *modname, struct kprobe *pp) { char *kprobe_type; @@ -1295,19 +1643,21 @@ static void __kprobes report_probe(struct seq_file *pi, struct kprobe *p, kprobe_type = "j"; else kprobe_type = "k"; + if (sym) - seq_printf(pi, "%p %s %s+0x%x %s %s%s\n", + seq_printf(pi, "%p %s %s+0x%x %s ", p->addr, kprobe_type, sym, offset, - (modname ? modname : " "), - (kprobe_gone(p) ? "[GONE]" : ""), - ((kprobe_disabled(p) && !kprobe_gone(p)) ? - "[DISABLED]" : "")); + (modname ? modname : " ")); else - seq_printf(pi, "%p %s %p %s%s\n", - p->addr, kprobe_type, p->addr, - (kprobe_gone(p) ? "[GONE]" : ""), - ((kprobe_disabled(p) && !kprobe_gone(p)) ? - "[DISABLED]" : "")); + seq_printf(pi, "%p %s %p ", + p->addr, kprobe_type, p->addr); + + if (!pp) + pp = p; + seq_printf(pi, "%s%s%s\n", + (kprobe_gone(p) ? "[GONE]" : ""), + ((kprobe_disabled(p) && !kprobe_gone(p)) ? "[DISABLED]" : ""), + (kprobe_optimized(pp) ? "[OPTIMIZED]" : "")); } static void __kprobes *kprobe_seq_start(struct seq_file *f, loff_t *pos) @@ -1343,11 +1693,11 @@ static int __kprobes show_kprobe_addr(struct seq_file *pi, void *v) hlist_for_each_entry_rcu(p, node, head, hlist) { sym = kallsyms_lookup((unsigned long)p->addr, NULL, &offset, &modname, namebuf); - if (p->pre_handler == aggr_pre_handler) { + if (kprobe_aggrprobe(p)) { list_for_each_entry_rcu(kp, &p->list, list) - report_probe(pi, kp, sym, offset, modname); + report_probe(pi, kp, sym, offset, modname, p); } else - report_probe(pi, p, sym, offset, modname); + report_probe(pi, p, sym, offset, modname, NULL); } preempt_enable(); return 0; @@ -1425,12 +1775,13 @@ int __kprobes enable_kprobe(struct kprobe *kp) goto out; } - if (!kprobes_all_disarmed && kprobe_disabled(p)) - arm_kprobe(p); - - p->flags &= ~KPROBE_FLAG_DISABLED; if (p != kp) kp->flags &= ~KPROBE_FLAG_DISABLED; + + if (!kprobes_all_disarmed && kprobe_disabled(p)) { + p->flags &= ~KPROBE_FLAG_DISABLED; + arm_kprobe(p); + } out: mutex_unlock(&kprobe_mutex); return ret; @@ -1450,12 +1801,13 @@ static void __kprobes arm_all_kprobes(void) if (!kprobes_all_disarmed) goto already_enabled; + /* Arming kprobes doesn't optimize kprobe itself */ mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) if (!kprobe_disabled(p)) - arch_arm_kprobe(p); + __arm_kprobe(p); } mutex_unlock(&text_mutex); @@ -1482,16 +1834,23 @@ static void __kprobes disarm_all_kprobes(void) kprobes_all_disarmed = true; printk(KERN_INFO "Kprobes globally disabled\n"); + + /* + * Here we call get_online_cpus() for avoiding text_mutex deadlock, + * because disarming may also unoptimize kprobes. + */ + get_online_cpus(); mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) { if (!arch_trampoline_kprobe(p) && !kprobe_disabled(p)) - arch_disarm_kprobe(p); + __disarm_kprobe(p); } } mutex_unlock(&text_mutex); + put_online_cpus(); mutex_unlock(&kprobe_mutex); /* Allow all currently running kprobes to complete */ synchronize_sched(); -- cgit v1.2.2 From b2be84df99ebc93599c69e931a3c4a5105abfabc Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Thu, 25 Feb 2010 08:34:15 -0500 Subject: kprobes: Jump optimization sysctl interface Add /proc/sys/debug/kprobes-optimization sysctl which enables and disables kprobes jump optimization on the fly for debugging. Changes in v7: - Remove ctl_name = CTL_UNNUMBERED for upstream compatibility. Changes in v6: - Update comments and coding style. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Srikar Dronamraju Cc: Christoph Hellwig Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Anders Kaseorg Cc: Tim Abbott Cc: Andi Kleen Cc: Jason Baron Cc: Mathieu Desnoyers Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli LKML-Reference: <20100225133415.6725.8274.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 88 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-- kernel/sysctl.c | 12 ++++++++ 2 files changed, 97 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 612af2d61614..fa034d29cf73 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -42,6 +42,7 @@ #include #include #include +#include #include #include #include @@ -360,6 +361,9 @@ static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) } #ifdef CONFIG_OPTPROBES +/* NOTE: change this value only with kprobe_mutex held */ +static bool kprobes_allow_optimization; + /* * Call all pre_handler on the list, but ignores its return value. * This must be called from arch-dep optimized caller. @@ -428,7 +432,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work) /* Lock modules while optimizing kprobes */ mutex_lock(&module_mutex); mutex_lock(&kprobe_mutex); - if (kprobes_all_disarmed) + if (kprobes_all_disarmed || !kprobes_allow_optimization) goto end; /* @@ -471,7 +475,7 @@ static __kprobes void optimize_kprobe(struct kprobe *p) struct optimized_kprobe *op; /* Check if the kprobe is disabled or not ready for optimization. */ - if (!kprobe_optready(p) || + if (!kprobe_optready(p) || !kprobes_allow_optimization || (kprobe_disabled(p) || kprobes_all_disarmed)) return; @@ -588,6 +592,80 @@ static __kprobes void try_to_optimize_kprobe(struct kprobe *p) optimize_kprobe(ap); } +#ifdef CONFIG_SYSCTL +static void __kprobes optimize_all_kprobes(void) +{ + struct hlist_head *head; + struct hlist_node *node; + struct kprobe *p; + unsigned int i; + + /* If optimization is already allowed, just return */ + if (kprobes_allow_optimization) + return; + + kprobes_allow_optimization = true; + mutex_lock(&text_mutex); + for (i = 0; i < KPROBE_TABLE_SIZE; i++) { + head = &kprobe_table[i]; + hlist_for_each_entry_rcu(p, node, head, hlist) + if (!kprobe_disabled(p)) + optimize_kprobe(p); + } + mutex_unlock(&text_mutex); + printk(KERN_INFO "Kprobes globally optimized\n"); +} + +static void __kprobes unoptimize_all_kprobes(void) +{ + struct hlist_head *head; + struct hlist_node *node; + struct kprobe *p; + unsigned int i; + + /* If optimization is already prohibited, just return */ + if (!kprobes_allow_optimization) + return; + + kprobes_allow_optimization = false; + printk(KERN_INFO "Kprobes globally unoptimized\n"); + get_online_cpus(); /* For avoiding text_mutex deadlock */ + mutex_lock(&text_mutex); + for (i = 0; i < KPROBE_TABLE_SIZE; i++) { + head = &kprobe_table[i]; + hlist_for_each_entry_rcu(p, node, head, hlist) { + if (!kprobe_disabled(p)) + unoptimize_kprobe(p); + } + } + + mutex_unlock(&text_mutex); + put_online_cpus(); + /* Allow all currently running kprobes to complete */ + synchronize_sched(); +} + +int sysctl_kprobes_optimization; +int proc_kprobes_optimization_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *length, + loff_t *ppos) +{ + int ret; + + mutex_lock(&kprobe_mutex); + sysctl_kprobes_optimization = kprobes_allow_optimization ? 1 : 0; + ret = proc_dointvec_minmax(table, write, buffer, length, ppos); + + if (sysctl_kprobes_optimization) + optimize_all_kprobes(); + else + unoptimize_all_kprobes(); + mutex_unlock(&kprobe_mutex); + + return ret; +} +#endif /* CONFIG_SYSCTL */ + static void __kprobes __arm_kprobe(struct kprobe *p) { struct kprobe *old_p; @@ -1610,10 +1688,14 @@ static int __init init_kprobes(void) } } -#if defined(CONFIG_OPTPROBES) && defined(__ARCH_WANT_KPROBES_INSN_SLOT) +#if defined(CONFIG_OPTPROBES) +#if defined(__ARCH_WANT_KPROBES_INSN_SLOT) /* Init kprobe_optinsn_slots */ kprobe_optinsn_slots.insn_size = MAX_OPTINSN_SIZE; #endif + /* By default, kprobes can be optimized */ + kprobes_allow_optimization = true; +#endif /* By default, kprobes are armed */ kprobes_all_disarmed = false; diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 8a68b2448468..40d791d616b5 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -50,6 +50,7 @@ #include #include #include +#include #include #include @@ -1449,6 +1450,17 @@ static struct ctl_table debug_table[] = { .mode = 0644, .proc_handler = proc_dointvec }, +#endif +#if defined(CONFIG_OPTPROBES) + { + .procname = "kprobes-optimization", + .data = &sysctl_kprobes_optimization, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_kprobes_optimization_handler, + .extra1 = &zero, + .extra2 = &one, + }, #endif { } }; -- cgit v1.2.2 From 056ba4a9bea5f32781a36b797c562fb731e5eaa6 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 25 Feb 2010 14:06:46 -0800 Subject: rcu: Make lockdep_rcu_dereference() message less alarmist Change from "unsafe" to "suspicious", given that there will be false alarms. Suggested-by: Ingo Molnar Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267135607-7056-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 672c436946ce..0c30d0455de1 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3816,9 +3816,9 @@ void lockdep_rcu_dereference(const char *file, const int line) if (!debug_locks_off()) return; - printk("\n==============================================\n"); - printk( "[ BUG: Unsafe rcu_dereference_check() usage! ]\n"); - printk( "----------------------------------------------\n"); + printk("\n===================================================\n"); + printk( "[ INFO: suspicious rcu_dereference_check() usage. ]\n"); + printk( "---------------------------------------------------\n"); printk("%s:%d invoked rcu_dereference_check() without protection!\n", file, line); printk("\nother info that might help us debug this:\n\n"); -- cgit v1.2.2 From d9f1bb6ad7fc53c406706f47858dd5ff030b14a3 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 25 Feb 2010 14:06:47 -0800 Subject: rcu: Make rcu_read_lock_sched_held() take boot time into account Before the scheduler starts, all tasks are non-preemptible by definition. So, during that time, rcu_read_lock_sched_held() needs to always return "true". This patch makes that be so. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267135607-7056-2-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 18 ++++++++++++++++++ kernel/rcutree.c | 19 ------------------- 2 files changed, 18 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 033cb55c26df..7bfa004572b1 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -44,6 +44,7 @@ #include #include #include +#include #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; @@ -62,6 +63,23 @@ struct lockdep_map rcu_sched_lock_map = EXPORT_SYMBOL_GPL(rcu_sched_lock_map); #endif +int rcu_scheduler_active __read_mostly; + +/* + * This function is invoked towards the end of the scheduler's initialization + * process. Before this is called, the idle task might contain + * RCU read-side critical sections (during which time, this idle + * task is booting the system). After this function is called, the + * idle tasks are prohibited from containing RCU read-side critical + * sections. + */ +void rcu_scheduler_starting(void) +{ + WARN_ON(num_online_cpus() != 1); + WARN_ON(nr_context_switches() > 0); + rcu_scheduler_active = 1; +} + /* * Awaken the corresponding synchronize_rcu() instance now that a * grace period has elapsed. diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 525d39810616..335bfe4f0076 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -46,7 +46,6 @@ #include #include #include -#include #include "rcutree.h" @@ -81,9 +80,6 @@ DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state); DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); -static int rcu_scheduler_active __read_mostly; - - /* * Return true if an RCU grace period is in progress. The ACCESS_ONCE()s * permit this function to be invoked without holding the root rcu_node @@ -1565,21 +1561,6 @@ static int rcu_needs_cpu_quick_check(int cpu) rcu_preempt_needs_cpu(cpu); } -/* - * This function is invoked towards the end of the scheduler's initialization - * process. Before this is called, the idle task might contain - * RCU read-side critical sections (during which time, this idle - * task is booting the system). After this function is called, the - * idle tasks are prohibited from containing RCU read-side critical - * sections. - */ -void rcu_scheduler_starting(void) -{ - WARN_ON(num_online_cpus() != 1); - WARN_ON(nr_context_switches() > 0); - rcu_scheduler_active = 1; -} - static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL}; static atomic_t rcu_barrier_cpu_count; static DEFINE_MUTEX(rcu_barrier_mutex); -- cgit v1.2.2 From f5f654096487c6d526c47bb66308f9de81f091cf Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 25 Feb 2010 19:02:30 -0800 Subject: rcu: Export rcu_scheduler_active Kernel modules using rcu_read_lock_sched_held() must now have access to rcu_scheduler_active, so it must be exported. This should fix the fix for the boot-time RCU-lockdep splat. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <20100226030230.GA7743@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 7bfa004572b1..f1125c1a6321 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -64,6 +64,7 @@ EXPORT_SYMBOL_GPL(rcu_sched_lock_map); #endif int rcu_scheduler_active __read_mostly; +EXPORT_SYMBOL_GPL(rcu_scheduler_active); /* * This function is invoked towards the end of the scheduler's initialization -- cgit v1.2.2 From fb90ef93df654f2678933efbbf864adac0ae490e Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 24 Feb 2010 18:36:53 -0800 Subject: early_res: Add free_early_partial() To free partial areas in pcpu_setup... Reported-by: Peter Zijlstra Signed-off-by: Yinghai Lu Cc: Tejun Heo Cc: Christoph Lameter Cc: Stephen Rothwell Cc: Linus Torvalds Cc: Jesse Barnes Cc: Pekka Enberg LKML-Reference: <4B85E245.5030001@kernel.org> Signed-off-by: Ingo Molnar --- kernel/early_res.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+) (limited to 'kernel') diff --git a/kernel/early_res.c b/kernel/early_res.c index aa5494ac4462..9ab11cd84853 100644 --- a/kernel/early_res.c +++ b/kernel/early_res.c @@ -61,6 +61,40 @@ static void __init drop_range(int i) early_res_count--; } +static void __init drop_range_partial(int i, u64 start, u64 end) +{ + u64 common_start, common_end; + u64 old_start, old_end; + + old_start = early_res[i].start; + old_end = early_res[i].end; + common_start = max(old_start, start); + common_end = min(old_end, end); + + /* no overlap ? */ + if (common_start >= common_end) + return; + + if (old_start < common_start) { + /* make head segment */ + early_res[i].end = common_start; + if (old_end > common_end) { + /* add another for left over on tail */ + reserve_early_without_check(common_end, old_end, + early_res[i].name); + } + return; + } else { + if (old_end > common_end) { + /* reuse the entry for tail left */ + early_res[i].start = common_end; + return; + } + /* all covered */ + drop_range(i); + } +} + /* * Split any existing ranges that: * 1) are marked 'overlap_ok', and @@ -284,6 +318,27 @@ void __init free_early(u64 start, u64 end) drop_range(i); } +void __init free_early_partial(u64 start, u64 end) +{ + struct early_res *r; + int i; + +try_next: + i = find_overlapped_early(start, end); + if (i >= max_early_res) + return; + + r = &early_res[i]; + /* hole ? */ + if (r->end >= end && r->start <= start) { + drop_range_partial(i, start, end); + return; + } + + drop_range_partial(i, start, end); + goto try_next; +} + #ifdef CONFIG_NO_BOOTMEM static void __init subtract_early_res(struct range *range, int az) { -- cgit v1.2.2 From 3a0304e90aa5a2c0c308a05d28f7d109a48d8539 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 26 Feb 2010 10:33:41 +0100 Subject: perf_events: Report the MMAP pgoff value in bytes DaveM reported that currently perf interprets the pgoff value reported by the MMAP events as a byte range, but the kernel reports it as a page offset. Since its broken (and unusable) anyway, change the kernel behaviour (ABI) to report bytes indeed, avoiding the need for userspace to deal with PAGE_SIZE things. Reported-by: David Miller Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 087025fe3ba1..5a69abb05ac3 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3749,7 +3749,7 @@ void __perf_event_mmap(struct vm_area_struct *vma) /* .tid */ .start = vma->vm_start, .len = vma->vm_end - vma->vm_start, - .pgoff = vma->vm_pgoff, + .pgoff = (u64)vma->vm_pgoff << PAGE_SHIFT, }, }; -- cgit v1.2.2 From d76a0812ac4139ceb54daab3cc70e1bd8bd9d43a Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 8 Feb 2010 17:06:01 +0200 Subject: perf_events: Add new start/stop PMU callbacks In certain situations, the kernel may need to stop and start the same event rapidly. The current PMU callbacks do not distinguish between stop and release (i.e., stop + free the resource). Thus, a counter may be released, then it will be immediately re-acquired. Event scheduling will again take place with no guarantee to assign the same counter. On some processors, this may event yield to failure to assign the event back due to competion between cores. This patch is adding a new pair of callback to stop and restart a counter without actually release the underlying counter resource. On stop, the counter is stopped, its values saved and that's it. On start, the value is reloaded and counter is restarted (on x86, actual restart is delayed until perf_enable()). Signed-off-by: Stephane Eranian [ added fallback to ->enable/->disable for all other PMUs fixed x86_pmu_start() to call x86_pmu.enable() merged __x86_pmu_disable into x86_pmu_stop() ] Signed-off-by: Peter Zijlstra LKML-Reference: <4b703875.0a04d00a.7896.ffffb824@mx.google.com> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 5a69abb05ac3..74c60021cdbc 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1493,6 +1493,22 @@ do { \ return div64_u64(dividend, divisor); } +static void perf_event_stop(struct perf_event *event) +{ + if (!event->pmu->stop) + return event->pmu->disable(event); + + return event->pmu->stop(event); +} + +static int perf_event_start(struct perf_event *event) +{ + if (!event->pmu->start) + return event->pmu->enable(event); + + return event->pmu->start(event); +} + static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) { struct hw_perf_event *hwc = &event->hw; @@ -1513,9 +1529,9 @@ static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) if (atomic64_read(&hwc->period_left) > 8*sample_period) { perf_disable(); - event->pmu->disable(event); + perf_event_stop(event); atomic64_set(&hwc->period_left, 0); - event->pmu->enable(event); + perf_event_start(event); perf_enable(); } } -- cgit v1.2.2 From 38331f62c20456454eed9ebea2525f072c6f1d2e Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 8 Feb 2010 17:17:01 +0200 Subject: perf_events, x86: AMD event scheduling This patch adds correct AMD NorthBridge event scheduling. NB events are events measuring L3 cache, Hypertransport traffic. They are identified by an event code >= 0xe0. They measure events on the Northbride which is shared by all cores on a package. NB events are counted on a shared set of counters. When a NB event is programmed in a counter, the data actually comes from a shared counter. Thus, access to those counters needs to be synchronized. We implement the synchronization such that no two cores can be measuring NB events using the same counters. Thus, we maintain a per-NB allocation table. The available slot is propagated using the event_constraint structure. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra LKML-Reference: <4b703957.0702d00a.6bf2.7b7d@mx.google.com> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 74c60021cdbc..fb4e56eb58f4 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -98,6 +98,7 @@ void __weak hw_perf_enable(void) { barrier(); } void __weak hw_perf_event_setup(int cpu) { barrier(); } void __weak hw_perf_event_setup_online(int cpu) { barrier(); } +void __weak hw_perf_event_setup_offline(int cpu) { barrier(); } int __weak hw_perf_group_sched_in(struct perf_event *group_leader, @@ -5462,6 +5463,10 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) perf_event_exit_cpu(cpu); break; + case CPU_DEAD: + hw_perf_event_setup_offline(cpu); + break; + default: break; } -- cgit v1.2.2 From 6e37738a2fac964583debe91099bc3248554f6e5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 11 Feb 2010 13:21:58 +0100 Subject: perf_events: Simplify code by removing cpu argument to hw_perf_group_sched_in() Since the cpu argument to hw_perf_group_sched_in() is always smp_processor_id(), simplify the code a little by removing this argument and using the current cpu where needed. Signed-off-by: Peter Zijlstra Cc: David Miller Cc: Paul Mackerras Cc: Benjamin Herrenschmidt Cc: Frederic Weisbecker LKML-Reference: <1265890918.5396.3.camel@laptop> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 45 ++++++++++++++++++--------------------------- 1 file changed, 18 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index fb4e56eb58f4..05b6c6b825e3 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -103,7 +103,7 @@ void __weak hw_perf_event_setup_offline(int cpu) { barrier(); } int __weak hw_perf_group_sched_in(struct perf_event *group_leader, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, int cpu) + struct perf_event_context *ctx) { return 0; } @@ -633,14 +633,13 @@ void perf_event_disable(struct perf_event *event) static int event_sched_in(struct perf_event *event, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, - int cpu) + struct perf_event_context *ctx) { if (event->state <= PERF_EVENT_STATE_OFF) return 0; event->state = PERF_EVENT_STATE_ACTIVE; - event->oncpu = cpu; /* TODO: put 'cpu' into cpuctx->cpu */ + event->oncpu = smp_processor_id(); /* * The new state must be visible before we turn it on in the hardware: */ @@ -667,8 +666,7 @@ event_sched_in(struct perf_event *event, static int group_sched_in(struct perf_event *group_event, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, - int cpu) + struct perf_event_context *ctx) { struct perf_event *event, *partial_group; int ret; @@ -676,18 +674,18 @@ group_sched_in(struct perf_event *group_event, if (group_event->state == PERF_EVENT_STATE_OFF) return 0; - ret = hw_perf_group_sched_in(group_event, cpuctx, ctx, cpu); + ret = hw_perf_group_sched_in(group_event, cpuctx, ctx); if (ret) return ret < 0 ? ret : 0; - if (event_sched_in(group_event, cpuctx, ctx, cpu)) + if (event_sched_in(group_event, cpuctx, ctx)) return -EAGAIN; /* * Schedule in siblings as one group (if any): */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { - if (event_sched_in(event, cpuctx, ctx, cpu)) { + if (event_sched_in(event, cpuctx, ctx)) { partial_group = event; goto group_error; } @@ -761,7 +759,6 @@ static void __perf_install_in_context(void *info) struct perf_event *event = info; struct perf_event_context *ctx = event->ctx; struct perf_event *leader = event->group_leader; - int cpu = smp_processor_id(); int err; /* @@ -808,7 +805,7 @@ static void __perf_install_in_context(void *info) if (!group_can_go_on(event, cpuctx, 1)) err = -EEXIST; else - err = event_sched_in(event, cpuctx, ctx, cpu); + err = event_sched_in(event, cpuctx, ctx); if (err) { /* @@ -950,11 +947,9 @@ static void __perf_event_enable(void *info) } else { perf_disable(); if (event == leader) - err = group_sched_in(event, cpuctx, ctx, - smp_processor_id()); + err = group_sched_in(event, cpuctx, ctx); else - err = event_sched_in(event, cpuctx, ctx, - smp_processor_id()); + err = event_sched_in(event, cpuctx, ctx); perf_enable(); } @@ -1281,19 +1276,18 @@ static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, static void ctx_pinned_sched_in(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx, - int cpu) + struct perf_cpu_context *cpuctx) { struct perf_event *event; list_for_each_entry(event, &ctx->pinned_groups, group_entry) { if (event->state <= PERF_EVENT_STATE_OFF) continue; - if (event->cpu != -1 && event->cpu != cpu) + if (event->cpu != -1 && event->cpu != smp_processor_id()) continue; if (group_can_go_on(event, cpuctx, 1)) - group_sched_in(event, cpuctx, ctx, cpu); + group_sched_in(event, cpuctx, ctx); /* * If this pinned group hasn't been scheduled, @@ -1308,8 +1302,7 @@ ctx_pinned_sched_in(struct perf_event_context *ctx, static void ctx_flexible_sched_in(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx, - int cpu) + struct perf_cpu_context *cpuctx) { struct perf_event *event; int can_add_hw = 1; @@ -1322,11 +1315,11 @@ ctx_flexible_sched_in(struct perf_event_context *ctx, * Listen to the 'cpu' scheduling filter constraint * of events: */ - if (event->cpu != -1 && event->cpu != cpu) + if (event->cpu != -1 && event->cpu != smp_processor_id()) continue; if (group_can_go_on(event, cpuctx, can_add_hw)) - if (group_sched_in(event, cpuctx, ctx, cpu)) + if (group_sched_in(event, cpuctx, ctx)) can_add_hw = 0; } } @@ -1336,8 +1329,6 @@ ctx_sched_in(struct perf_event_context *ctx, struct perf_cpu_context *cpuctx, enum event_type_t event_type) { - int cpu = smp_processor_id(); - raw_spin_lock(&ctx->lock); ctx->is_active = 1; if (likely(!ctx->nr_events)) @@ -1352,11 +1343,11 @@ ctx_sched_in(struct perf_event_context *ctx, * in order to give them the best chance of going on. */ if (event_type & EVENT_PINNED) - ctx_pinned_sched_in(ctx, cpuctx, cpu); + ctx_pinned_sched_in(ctx, cpuctx); /* Then walk through the lower prio flexible groups */ if (event_type & EVENT_FLEXIBLE) - ctx_flexible_sched_in(ctx, cpuctx, cpu); + ctx_flexible_sched_in(ctx, cpuctx); perf_enable(); out: -- cgit v1.2.2 From dd5feea14a7de4edbd9f36db1a2db785de91b88d Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Tue, 23 Feb 2010 16:13:52 -0800 Subject: sched: Fix SCHED_MC regression caused by change in sched cpu_power On platforms like dual socket quad-core platform, the scheduler load balancer is not detecting the load imbalances in certain scenarios. This is leading to scenarios like where one socket is completely busy (with all the 4 cores running with 4 tasks) and leaving another socket completely idle. This causes performance issues as those 4 tasks share the memory controller, last-level cache bandwidth etc. Also we won't be taking advantage of turbo-mode as much as we would like, etc. Some of the comparisons in the scheduler load balancing code are comparing the "weighted cpu load that is scaled wrt sched_group's cpu_power" with the "weighted average load per task that is not scaled wrt sched_group's cpu_power". While this has probably been broken for a longer time (for multi socket numa nodes etc), the problem got aggrevated via this recent change: | | commit f93e65c186ab3c05ce2068733ca10e34fd00125e | Author: Peter Zijlstra | Date: Tue Sep 1 10:34:32 2009 +0200 | | sched: Restore __cpu_power to a straight sum of power | Also with this change, the sched group cpu power alone no longer reflects the group capacity that is needed to implement MC, MT performance (default) and power-savings (user-selectable) policies. We need to use the computed group capacity (sgs.group_capacity, that is computed using the SD_PREFER_SIBLING logic in update_sd_lb_stats()) to find out if the group with the max load is above its capacity and how much load to move etc. Reported-by: Ma Ling Initial-Analysis-by: Zhang, Yanmin Signed-off-by: Suresh Siddha [ -v2: build fix ] Signed-off-by: Peter Zijlstra Cc: # [2.6.32.x, 2.6.33.x] LKML-Reference: <1266970432.11588.22.camel@sbs-t61.sc.intel.com> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 76 ++++++++++++++++++++++++++++++----------------------- 1 file changed, 43 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index ff7692ccda89..3e1fd96c6cf9 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2097,6 +2097,7 @@ struct sd_lb_stats { unsigned long max_load; unsigned long busiest_load_per_task; unsigned long busiest_nr_running; + unsigned long busiest_group_capacity; int group_imb; /* Is there imbalance in this sd */ #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) @@ -2416,14 +2417,12 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, unsigned long load, max_cpu_load, min_cpu_load; int i; unsigned int balance_cpu = -1, first_idle_cpu = 0; - unsigned long sum_avg_load_per_task; - unsigned long avg_load_per_task; + unsigned long avg_load_per_task = 0; if (local_group) balance_cpu = group_first_cpu(group); /* Tally up the load of all CPUs in the group */ - sum_avg_load_per_task = avg_load_per_task = 0; max_cpu_load = 0; min_cpu_load = ~0UL; @@ -2453,7 +2452,6 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, sgs->sum_nr_running += rq->nr_running; sgs->sum_weighted_load += weighted_cpuload(i); - sum_avg_load_per_task += cpu_avg_load_per_task(i); } /* @@ -2473,7 +2471,6 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, /* Adjust by relative CPU power of the group */ sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; - /* * Consider the group unbalanced when the imbalance is larger * than the average weight of two tasks. @@ -2483,8 +2480,8 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, * normalized nr_running number somewhere that negates * the hierarchy? */ - avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / - group->cpu_power; + if (sgs->sum_nr_running) + avg_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) sgs->group_imb = 1; @@ -2553,6 +2550,7 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, sds->max_load = sgs.avg_load; sds->busiest = group; sds->busiest_nr_running = sgs.sum_nr_running; + sds->busiest_group_capacity = sgs.group_capacity; sds->busiest_load_per_task = sgs.sum_weighted_load; sds->group_imb = sgs.group_imb; } @@ -2575,6 +2573,7 @@ static inline void fix_small_imbalance(struct sd_lb_stats *sds, { unsigned long tmp, pwr_now = 0, pwr_move = 0; unsigned int imbn = 2; + unsigned long scaled_busy_load_per_task; if (sds->this_nr_running) { sds->this_load_per_task /= sds->this_nr_running; @@ -2585,8 +2584,12 @@ static inline void fix_small_imbalance(struct sd_lb_stats *sds, sds->this_load_per_task = cpu_avg_load_per_task(this_cpu); - if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= - sds->busiest_load_per_task * imbn) { + scaled_busy_load_per_task = sds->busiest_load_per_task + * SCHED_LOAD_SCALE; + scaled_busy_load_per_task /= sds->busiest->cpu_power; + + if (sds->max_load - sds->this_load + scaled_busy_load_per_task >= + (scaled_busy_load_per_task * imbn)) { *imbalance = sds->busiest_load_per_task; return; } @@ -2637,7 +2640,14 @@ static inline void fix_small_imbalance(struct sd_lb_stats *sds, static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { - unsigned long max_pull; + unsigned long max_pull, load_above_capacity = ~0UL; + + sds->busiest_load_per_task /= sds->busiest_nr_running; + if (sds->group_imb) { + sds->busiest_load_per_task = + min(sds->busiest_load_per_task, sds->avg_load); + } + /* * In the presence of smp nice balancing, certain scenarios can have * max load less than avg load(as we skip the groups at or below @@ -2648,9 +2658,29 @@ static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, return fix_small_imbalance(sds, this_cpu, imbalance); } - /* Don't want to pull so many tasks that a group would go idle */ - max_pull = min(sds->max_load - sds->avg_load, - sds->max_load - sds->busiest_load_per_task); + if (!sds->group_imb) { + /* + * Don't want to pull so many tasks that a group would go idle. + */ + load_above_capacity = (sds->busiest_nr_running - + sds->busiest_group_capacity); + + load_above_capacity *= (SCHED_LOAD_SCALE * SCHED_LOAD_SCALE); + + load_above_capacity /= sds->busiest->cpu_power; + } + + /* + * We're trying to get all the cpus to the average_load, so we don't + * want to push ourselves above the average load, nor do we wish to + * reduce the max loaded cpu below the average load. At the same time, + * we also don't want to reduce the group load below the group capacity + * (so that we can implement power-savings policies etc). Thus we look + * for the minimum possible imbalance. + * Be careful of negative numbers as they'll appear as very large values + * with unsigned longs. + */ + max_pull = min(sds->max_load - sds->avg_load, load_above_capacity); /* How much load to actually move to equalise the imbalance */ *imbalance = min(max_pull * sds->busiest->cpu_power, @@ -2718,7 +2748,6 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, * 4) This group is more busy than the avg busieness at this * sched_domain. * 5) The imbalance is within the specified limit. - * 6) Any rebalance would lead to ping-pong */ if (!(*balance)) goto ret; @@ -2737,25 +2766,6 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) goto out_balanced; - sds.busiest_load_per_task /= sds.busiest_nr_running; - if (sds.group_imb) - sds.busiest_load_per_task = - min(sds.busiest_load_per_task, sds.avg_load); - - /* - * We're trying to get all the cpus to the average_load, so we don't - * want to push ourselves above the average load, nor do we wish to - * reduce the max loaded cpu below the average load, as either of these - * actions would just result in more rebalancing later, and ping-pong - * tasks around. Thus we look for the minimum possible imbalance. - * Negative imbalances (*we* are more loaded than anyone else) will - * be counted as no imbalance for these purposes -- we can't fix that - * by pulling tasks to us. Be careful of negative numbers as they'll - * appear as very large values with unsigned longs. - */ - if (sds.max_load <= sds.busiest_load_per_task) - goto out_balanced; - /* Looks like there is an imbalance. Compute it */ calculate_imbalance(&sds, this_cpu, imbalance); return sds.busiest; -- cgit v1.2.2 From 24691ea964cc0123e386b661e03a86a481c6ee79 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 26 Feb 2010 16:36:23 +0100 Subject: perf_event: Fix preempt warning in perf_clock() A recent commit introduced a preemption warning for perf_clock(), use raw_smp_processor_id() to avoid this, it really doesn't matter which cpu we use here. Signed-off-by: Peter Zijlstra LKML-Reference: <1267198583.22519.684.camel@laptop> Cc: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 05b6c6b825e3..aa6155b5e24c 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -249,7 +249,7 @@ static void perf_unpin_context(struct perf_event_context *ctx) static inline u64 perf_clock(void) { - return cpu_clock(smp_processor_id()); + return cpu_clock(raw_smp_processor_id()); } /* -- cgit v1.2.2 From 0e06b4a891c6a108412fe24b4500f499da2cf8a1 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Sat, 23 Jan 2010 22:25:15 +0100 Subject: PM: Add a switch for disabling/enabling asynchronous suspend/resume Add sysfs attribute /sys/power/pm_async allowing the user space to disable/enable asynchronous suspend/resume of devices. Signed-off-by: Rafael J. Wysocki --- kernel/power/main.c | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/main.c b/kernel/power/main.c index 0998c7139053..b58800b21fc0 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -44,6 +44,32 @@ int pm_notifier_call_chain(unsigned long val) == NOTIFY_BAD) ? -EINVAL : 0; } +/* If set, devices may be suspended and resumed asynchronously. */ +int pm_async_enabled = 1; + +static ssize_t pm_async_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%d\n", pm_async_enabled); +} + +static ssize_t pm_async_store(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t n) +{ + unsigned long val; + + if (strict_strtoul(buf, 10, &val)) + return -EINVAL; + + if (val > 1) + return -EINVAL; + + pm_async_enabled = val; + return n; +} + +power_attr(pm_async); + #ifdef CONFIG_PM_DEBUG int pm_test_level = TEST_NONE; @@ -208,8 +234,11 @@ static struct attribute * g[] = { #ifdef CONFIG_PM_TRACE &pm_trace_attr.attr, #endif -#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_PM_DEBUG) +#ifdef CONFIG_PM_SLEEP + &pm_async_attr.attr, +#ifdef CONFIG_PM_DEBUG &pm_test_attr.attr, +#endif #endif NULL, }; -- cgit v1.2.2 From 5a2eb8585f3b38e01e30aacaa8b985a1520a993d Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Sat, 23 Jan 2010 22:25:23 +0100 Subject: PM: Add facility for advanced testing of async suspend/resume Add configuration switch CONFIG_PM_ADVANCED_DEBUG for compiling in extra PM debugging/testing code allowing one to access some PM-related attributes of devices from the user space via sysfs. If CONFIG_PM_ADVANCED_DEBUG is set, add sysfs attribute power/async for every device allowing the user space to access the device's power.async_suspend flag and modify it, if desired. Signed-off-by: Rafael J. Wysocki --- kernel/power/Kconfig | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'kernel') diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 4c9cffcf69c7..5c36ea9d55d2 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -27,6 +27,15 @@ config PM_DEBUG code. This is helpful when debugging and reporting PM bugs, like suspend support. +config PM_ADVANCED_DEBUG + bool "Extra PM attributes in sysfs for low-level debugging/testing" + depends on PM_DEBUG + default n + ---help--- + Add extra sysfs attributes allowing one to access some Power Management + fields of device objects from user space. If you are not a kernel + developer interested in debugging/testing Power Management, say "no". + config PM_VERBOSE bool "Verbose Power Management debugging" depends on PM_DEBUG @@ -85,6 +94,11 @@ config PM_SLEEP depends on SUSPEND || HIBERNATION || XEN_SAVE_RESTORE default y +config PM_SLEEP_ADVANCED_DEBUG + bool + depends on PM_ADVANCED_DEBUG + default n + config SUSPEND bool "Suspend to RAM and standby" depends on PM && ARCH_SUSPEND_POSSIBLE -- cgit v1.2.2 From b694e52ebdd439e4591d4a6651658026196174f6 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Wed, 27 Jan 2010 23:47:50 +0100 Subject: PM / Hibernate: Really deprecate deprecated user ioctls They were deprecated and removed from exported headers more than 2 years ago. Inform users about their removal in the future now. (Switch cases needed to be reorderded for an easy fall through.) And add an entry to feature-removal-schedule. Signed-off-by: Jiri Slaby Acked-by: Pavel Machek Signed-off-by: Rafael J. Wysocki --- kernel/power/user.c | 23 +++++++++++++++++++---- 1 file changed, 19 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/power/user.c b/kernel/power/user.c index bf0014d6a5f0..4d2289626a84 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -195,6 +195,15 @@ static ssize_t snapshot_write(struct file *filp, const char __user *buf, return res; } +static void snapshot_deprecated_ioctl(unsigned int cmd) +{ + if (printk_ratelimit()) + printk(KERN_NOTICE "%pf: ioctl '%.8x' is deprecated and will " + "be removed soon, update your suspend-to-disk " + "utilities\n", + __builtin_return_address(0), cmd); +} + static long snapshot_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { @@ -246,8 +255,9 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, data->frozen = 0; break; - case SNAPSHOT_CREATE_IMAGE: case SNAPSHOT_ATOMIC_SNAPSHOT: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_CREATE_IMAGE: if (data->mode != O_RDONLY || !data->frozen || data->ready) { error = -EPERM; break; @@ -275,8 +285,9 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, data->ready = 0; break; - case SNAPSHOT_PREF_IMAGE_SIZE: case SNAPSHOT_SET_IMAGE_SIZE: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_PREF_IMAGE_SIZE: image_size = arg; break; @@ -290,15 +301,17 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, error = put_user(size, (loff_t __user *)arg); break; - case SNAPSHOT_AVAIL_SWAP_SIZE: case SNAPSHOT_AVAIL_SWAP: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_AVAIL_SWAP_SIZE: size = count_swap_pages(data->swap, 1); size <<= PAGE_SHIFT; error = put_user(size, (loff_t __user *)arg); break; - case SNAPSHOT_ALLOC_SWAP_PAGE: case SNAPSHOT_GET_SWAP_PAGE: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_ALLOC_SWAP_PAGE: if (data->swap < 0 || data->swap >= MAX_SWAPFILES) { error = -ENODEV; break; @@ -321,6 +334,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, break; case SNAPSHOT_SET_SWAP_FILE: /* This ioctl is deprecated */ + snapshot_deprecated_ioctl(cmd); if (!swsusp_swap_in_use()) { /* * User space encodes device types as two-byte values, @@ -362,6 +376,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, break; case SNAPSHOT_PMOPS: /* This ioctl is deprecated */ + snapshot_deprecated_ioctl(cmd); error = -EINVAL; switch (arg) { -- cgit v1.2.2 From 09c09bc618a4ceec387c57542031b4fc35826e16 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Wed, 27 Jan 2010 23:47:56 +0100 Subject: PM / Hibernate: Swap, remove useless check from swsusp_read() It will never reach here if the sws_resume_bdev is erratic. swsusp_read() is called only from software_resume(), but after swsusp_check() which would catch the error state. Signed-off-by: Jiri Slaby Signed-off-by: Rafael J. Wysocki --- kernel/power/swap.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/power/swap.c b/kernel/power/swap.c index 09b2b0ae9e9d..1d575733d4e1 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -657,10 +657,6 @@ int swsusp_read(unsigned int *flags_p) struct swsusp_info *header; *flags_p = swsusp_header->flags; - if (IS_ERR(resume_bdev)) { - pr_debug("PM: Image device not initialised\n"); - return PTR_ERR(resume_bdev); - } memset(&snapshot, 0, sizeof(struct snapshot_handle)); error = snapshot_write_next(&snapshot, PAGE_SIZE); -- cgit v1.2.2 From 07c3bb5797d0a9a48524cca17ceb95711b48890b Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Thu, 11 Feb 2010 23:09:08 +0100 Subject: PM / Hibernate: Remove trailing space in message Remove a trailing space from a message in swsusp_save(). Signed-off-by: Frans Pop Acked-by: Pavel Machek Signed-off-by: Rafael J. Wysocki --- kernel/power/snapshot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 36cb168e4330..2eda40741aaa 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -1500,7 +1500,7 @@ asmlinkage int swsusp_save(void) { unsigned int nr_pages, nr_highmem; - printk(KERN_INFO "PM: Creating hibernation image: \n"); + printk(KERN_INFO "PM: Creating hibernation image:\n"); drain_local_pages(NULL); nr_pages = count_data_pages(); -- cgit v1.2.2 From f8bb0db8188c7ef9ef08515883dae9f9eb980984 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Sun, 21 Feb 2010 22:14:44 +0100 Subject: PM / Hibernate: Remove swsusp.c finally Its contents and entry in Makefile were already removed in 8e60c6a1348e17e68ad73589a52a03876e7059be (Shift remaining code from swsusp.c to hibernate.c) but somehow it remained in-place (rjw: which most likely was my mistake). Signed-off-by: Jiri Slaby Acked-by: Nigel Cunningham Signed-off-by: Rafael J. Wysocki --- kernel/power/swsusp.c | 58 --------------------------------------------------- 1 file changed, 58 deletions(-) delete mode 100644 kernel/power/swsusp.c (limited to 'kernel') diff --git a/kernel/power/swsusp.c b/kernel/power/swsusp.c deleted file mode 100644 index 5b3601bd1893..000000000000 --- a/kernel/power/swsusp.c +++ /dev/null @@ -1,58 +0,0 @@ -/* - * linux/kernel/power/swsusp.c - * - * This file provides code to write suspend image to swap and read it back. - * - * Copyright (C) 1998-2001 Gabor Kuti - * Copyright (C) 1998,2001-2005 Pavel Machek - * - * This file is released under the GPLv2. - * - * I'd like to thank the following people for their work: - * - * Pavel Machek : - * Modifications, defectiveness pointing, being with me at the very beginning, - * suspend to swap space, stop all tasks. Port to 2.4.18-ac and 2.5.17. - * - * Steve Doddi : - * Support the possibility of hardware state restoring. - * - * Raph : - * Support for preserving states of network devices and virtual console - * (including X and svgatextmode) - * - * Kurt Garloff : - * Straightened the critical function in order to prevent compilers from - * playing tricks with local variables. - * - * Andreas Mohr - * - * Alex Badea : - * Fixed runaway init - * - * Rafael J. Wysocki - * Reworked the freeing of memory and the handling of swap - * - * More state savers are welcome. Especially for the scsi layer... - * - * For TODOs,FIXMEs also look in Documentation/power/swsusp.txt - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "power.h" - -int in_suspend __nosavedata = 0; -- cgit v1.2.2 From a9c9b4429df437982d2fbfab1f4b46b01329e9ed Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 25 Feb 2010 22:32:37 +0100 Subject: PM / Hibernate: Fix preallocating of memory The hibernate memory preallocation code allocates memory to push some user space data out of physical RAM, so that the hibernation image is not too large. It allocates more memory than necessary for creating the image, so it has to release some pages to make room for allocations made while suspending devices and disabling nonboot CPUs, or the system will hang due to the lack of free pages to allocate from. Unfortunately, the function used for freeing these pages, free_unnecessary_pages(), contains a bug that prevents it from doing the job on all systems without highmem. Fix this problem, which is a regression from the 2.6.30 kernel, by using the right condition for the termination of the loop in free_unnecessary_pages(). Signed-off-by: Rafael J. Wysocki Reported-and-tested-by: Alan Jenkins Cc: stable@kernel.org --- kernel/power/snapshot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 2eda40741aaa..830cadecbdfc 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -1181,7 +1181,7 @@ static void free_unnecessary_pages(void) memory_bm_position_reset(©_bm); - while (to_free_normal > 0 && to_free_highmem > 0) { + while (to_free_normal > 0 || to_free_highmem > 0) { unsigned long pfn = memory_bm_next_pfn(©_bm); struct page *page = pfn_to_page(pfn); -- cgit v1.2.2 From f1c7f517a5dc23bce07efa5ed55e2c074ed9d4ba Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 26 Feb 2010 17:08:16 -0500 Subject: ftrace: Add function names to dangling } in function graph tracer The function graph tracer is currently the most invasive tracer in the ftrace family. It can easily overflow the buffer even with 10megs per CPU. This means that events can often be lost. On start up, or after events are lost, if the function return is recorded but the function enter was lost, all we get to see is the exiting '}'. Here is how a typical trace output starts: [tracing] cat trace # tracer: function_graph # # CPU DURATION FUNCTION CALLS # | | | | | | | 0) + 91.897 us | } 0) ! 567.961 us | } 0) <========== | 0) ! 579.083 us | _raw_spin_lock_irqsave(); 0) 4.694 us | _raw_spin_unlock_irqrestore(); 0) ! 594.862 us | } 0) ! 603.361 us | } 0) ! 613.574 us | } 0) ! 623.554 us | } 0) 3.653 us | fget_light(); 0) | sock_poll() { There are a series of '}' with no matching "func() {". There's no information to what functions these ending brackets belong to. This patch adds a stack on the per cpu structure used in outputting the function graph tracer to keep track of what function was outputted. Then on a function exit event, it checks the depth to see if the function exit has a matching entry event. If it does, then it only prints the '}', otherwise it adds the function name after the '}'. This allows function exit events to show what function they belong to at trace output startup, when the entry was lost due to ring buffer overflow, or even after a new task is scheduled in. Here is what the above trace will look like after this patch: [tracing] cat trace # tracer: function_graph # # CPU DURATION FUNCTION CALLS # | | | | | | | 0) + 91.897 us | } (irq_exit) 0) ! 567.961 us | } (smp_apic_timer_interrupt) 0) <========== | 0) ! 579.083 us | _raw_spin_lock_irqsave(); 0) 4.694 us | _raw_spin_unlock_irqrestore(); 0) ! 594.862 us | } (add_wait_queue) 0) ! 603.361 us | } (__pollwait) 0) ! 613.574 us | } (tcp_poll) 0) ! 623.554 us | } (sock_poll) 0) 3.653 us | fget_light(); 0) | sock_poll() { Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 52 +++++++++++++++++++++++++++++------- 1 file changed, 43 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 112561df2a0a..e998a824e9db 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -18,6 +18,7 @@ struct fgraph_cpu_data { pid_t last_pid; int depth; int ignore; + unsigned long enter_funcs[FTRACE_RETFUNC_DEPTH]; }; struct fgraph_data { @@ -670,15 +671,21 @@ print_graph_entry_leaf(struct trace_iterator *iter, duration = graph_ret->rettime - graph_ret->calltime; if (data) { + struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); + + cpu_data = per_cpu_ptr(data->cpu_data, cpu); /* * Comments display at + 1 to depth. Since * this is a leaf function, keep the comments * equal to this depth. */ - *depth = call->depth - 1; + cpu_data->depth = call->depth - 1; + + /* No need to keep this function around for this depth */ + if (call->depth < FTRACE_RETFUNC_DEPTH) + cpu_data->enter_funcs[call->depth] = 0; } /* Overhead */ @@ -718,10 +725,15 @@ print_graph_entry_nested(struct trace_iterator *iter, int i; if (data) { + struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); - *depth = call->depth; + cpu_data = per_cpu_ptr(data->cpu_data, cpu); + cpu_data->depth = call->depth; + + /* Save this function pointer to see if the exit matches */ + if (call->depth < FTRACE_RETFUNC_DEPTH) + cpu_data->enter_funcs[call->depth] = call->func; } /* No overhead */ @@ -851,18 +863,28 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, struct fgraph_data *data = iter->private; pid_t pid = ent->pid; int cpu = iter->cpu; + int func_match = 1; int ret; int i; if (data) { - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); + struct fgraph_cpu_data *cpu_data; + int cpu = iter->cpu; + + cpu_data = per_cpu_ptr(data->cpu_data, cpu); /* * Comments display at + 1 to depth. This is the * return from a function, we now want the comments * to display at the same level of the bracket. */ - *depth = trace->depth - 1; + cpu_data->depth = trace->depth - 1; + + if (trace->depth < FTRACE_RETFUNC_DEPTH) { + if (cpu_data->enter_funcs[trace->depth] != trace->func) + func_match = 0; + cpu_data->enter_funcs[trace->depth] = 0; + } } if (print_graph_prologue(iter, s, 0, 0)) @@ -887,9 +909,21 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - ret = trace_seq_printf(s, "}\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + /* + * If the return function does not have a matching entry, + * then the entry was lost. Instead of just printing + * the '}' and letting the user guess what function this + * belongs to, write out the function name. + */ + if (func_match) { + ret = trace_seq_printf(s, "}\n"); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } else { + ret = trace_seq_printf(s, "} (%ps)\n", (void *)trace->func); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } /* Overrun */ if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { -- cgit v1.2.2 From a47cd880b50e14b0b6f5e9d426ae9a2676c9c474 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 26 Feb 2010 16:38:56 -0800 Subject: rcu: Fix accelerated grace periods for last non-dynticked CPU It is invalid to invoke __rcu_process_callbacks() with irqs disabled, so do it indirectly via raise_softirq(). This requires a state-machine implementation to cycle through the grace-period machinery the required number of times. Located-by: Ingo Molnar Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267231138-27856-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.c | 3 ++ kernel/rcutree.h | 1 + kernel/rcutree_plugin.h | 73 +++++++++++++++++++++++++++++++++++-------------- 3 files changed, 57 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 335bfe4f0076..3ec8160fc75f 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1341,6 +1341,9 @@ static void rcu_process_callbacks(struct softirq_action *unused) * grace-period manipulations above. */ smp_mb(); /* See above block comment. */ + + /* If we are last CPU on way to dyntick-idle mode, accelerate it. */ + rcu_needs_cpu_flush(); } static void diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 2ceb08388582..1439eb504c22 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -373,5 +373,6 @@ static int rcu_preempt_needs_cpu(int cpu); static void __cpuinit rcu_preempt_init_percpu_data(int cpu); static void rcu_preempt_send_cbs_to_orphanage(void); static void __init __rcu_init_preempt(void); +static void rcu_needs_cpu_flush(void); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 3516de7091a1..ed241fc478f0 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -973,9 +973,19 @@ int rcu_needs_cpu(int cpu) return rcu_needs_cpu_quick_check(cpu); } +/* + * Check to see if we need to continue a callback-flush operations to + * allow the last CPU to enter dyntick-idle mode. But fast dyntick-idle + * entry is not configured, so we never do need to. + */ +static void rcu_needs_cpu_flush(void) +{ +} + #else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */ #define RCU_NEEDS_CPU_FLUSHES 5 +static DEFINE_PER_CPU(int, rcu_dyntick_drain); /* * Check to see if any future RCU-related work will need to be done @@ -988,39 +998,62 @@ int rcu_needs_cpu(int cpu) * only if all other CPUs are already in dynticks-idle mode. This will * allow the CPU cores to be powered down immediately, as opposed to after * waiting many milliseconds for grace periods to elapse. + * + * Because it is not legal to invoke rcu_process_callbacks() with irqs + * disabled, we do one pass of force_quiescent_state(), then do a + * raise_softirq() to cause rcu_process_callbacks() to be invoked later. + * The per-cpu rcu_dyntick_drain variable controls the sequencing. */ int rcu_needs_cpu(int cpu) { - int c = 1; - int i; + int c = 0; int thatcpu; /* Don't bother unless we are the last non-dyntick-idle CPU. */ for_each_cpu_not(thatcpu, nohz_cpu_mask) - if (thatcpu != cpu) + if (thatcpu != cpu) { + per_cpu(rcu_dyntick_drain, cpu) = 0; return rcu_needs_cpu_quick_check(cpu); - - /* Try to push remaining RCU-sched and RCU-bh callbacks through. */ - for (i = 0; i < RCU_NEEDS_CPU_FLUSHES && c; i++) { - c = 0; - if (per_cpu(rcu_sched_data, cpu).nxtlist) { - rcu_sched_qs(cpu); - force_quiescent_state(&rcu_sched_state, 0); - __rcu_process_callbacks(&rcu_sched_state, - &per_cpu(rcu_sched_data, cpu)); - c = !!per_cpu(rcu_sched_data, cpu).nxtlist; - } - if (per_cpu(rcu_bh_data, cpu).nxtlist) { - rcu_bh_qs(cpu); - force_quiescent_state(&rcu_bh_state, 0); - __rcu_process_callbacks(&rcu_bh_state, - &per_cpu(rcu_bh_data, cpu)); - c = !!per_cpu(rcu_bh_data, cpu).nxtlist; } + + /* Check and update the rcu_dyntick_drain sequencing. */ + if (per_cpu(rcu_dyntick_drain, cpu) <= 0) { + /* First time through, initialize the counter. */ + per_cpu(rcu_dyntick_drain, cpu) = RCU_NEEDS_CPU_FLUSHES; + } else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) { + /* We have hit the limit, so time to give up. */ + return rcu_needs_cpu_quick_check(cpu); + } + + /* Do one step pushing remaining RCU callbacks through. */ + if (per_cpu(rcu_sched_data, cpu).nxtlist) { + rcu_sched_qs(cpu); + force_quiescent_state(&rcu_sched_state, 0); + c = c || per_cpu(rcu_sched_data, cpu).nxtlist; + } + if (per_cpu(rcu_bh_data, cpu).nxtlist) { + rcu_bh_qs(cpu); + force_quiescent_state(&rcu_bh_state, 0); + c = c || per_cpu(rcu_bh_data, cpu).nxtlist; } /* If RCU callbacks are still pending, RCU still needs this CPU. */ + if (c) + raise_softirq(RCU_SOFTIRQ); return c; } +/* + * Check to see if we need to continue a callback-flush operations to + * allow the last CPU to enter dyntick-idle mode. + */ +static void rcu_needs_cpu_flush(void) +{ + int cpu = smp_processor_id(); + + if (per_cpu(rcu_dyntick_drain, cpu) <= 0) + return; + (void)rcu_needs_cpu(cpu); +} + #endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */ -- cgit v1.2.2 From 71da81324c83ef65bb196c7f874ac1c6996d8287 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 26 Feb 2010 16:38:58 -0800 Subject: rcu: Fix accelerated GPs for last non-dynticked CPU This patch disables irqs across the call to rcu_needs_cpu(). It also enforces a hold-off period so that the idle loop doesn't softirq itself to death when there are lots of RCU callbacks in flight on the last non-dynticked CPU. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267231138-27856-3-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree_plugin.h | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index ed241fc478f0..464ad2cdee00 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -986,6 +986,7 @@ static void rcu_needs_cpu_flush(void) #define RCU_NEEDS_CPU_FLUSHES 5 static DEFINE_PER_CPU(int, rcu_dyntick_drain); +static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff); /* * Check to see if any future RCU-related work will need to be done @@ -1013,6 +1014,7 @@ int rcu_needs_cpu(int cpu) for_each_cpu_not(thatcpu, nohz_cpu_mask) if (thatcpu != cpu) { per_cpu(rcu_dyntick_drain, cpu) = 0; + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1; return rcu_needs_cpu_quick_check(cpu); } @@ -1022,6 +1024,7 @@ int rcu_needs_cpu(int cpu) per_cpu(rcu_dyntick_drain, cpu) = RCU_NEEDS_CPU_FLUSHES; } else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) { /* We have hit the limit, so time to give up. */ + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; return rcu_needs_cpu_quick_check(cpu); } @@ -1038,8 +1041,10 @@ int rcu_needs_cpu(int cpu) } /* If RCU callbacks are still pending, RCU still needs this CPU. */ - if (c) + if (c) { raise_softirq(RCU_SOFTIRQ); + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; + } return c; } @@ -1050,10 +1055,13 @@ int rcu_needs_cpu(int cpu) static void rcu_needs_cpu_flush(void) { int cpu = smp_processor_id(); + unsigned long flags; if (per_cpu(rcu_dyntick_drain, cpu) <= 0) return; + local_irq_save(flags); (void)rcu_needs_cpu(cpu); + local_irq_restore(flags); } #endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */ -- cgit v1.2.2 From 44ee63587dce85593c22497140db16f4e5027860 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 17 Feb 2010 10:50:50 +0900 Subject: percpu: Add __percpu sparse annotations to hw_breakpoint Add __percpu sparse annotations to hw_breakpoint. These annotations are to make sparse consider percpu variables to be in a different address space and warn if accessed without going through percpu accessors. This patch doesn't affect normal builds. In kernel/hw_breakpoint.c, per_cpu(nr_task_bp_pinned, cpu)'s will trigger spurious noderef related warnings from sparse. Changing it to &per_cpu(nr_task_bp_pinned[0], cpu) will work around the problem but deemed to ugly by the maintainer. Leave it alone until better solution can be found. Signed-off-by: Tejun Heo Cc: Stephen Rothwell Cc: K.Prasad LKML-Reference: <4B7B4B7A.9050902@kernel.org> Signed-off-by: Frederic Weisbecker --- kernel/hw_breakpoint.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 967e66143e11..6542eacb3fa5 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -413,17 +413,17 @@ EXPORT_SYMBOL_GPL(unregister_hw_breakpoint); * * @return a set of per_cpu pointers to perf events */ -struct perf_event ** +struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered) { - struct perf_event **cpu_events, **pevent, *bp; + struct perf_event * __percpu *cpu_events, **pevent, *bp; long err; int cpu; cpu_events = alloc_percpu(typeof(*cpu_events)); if (!cpu_events) - return ERR_PTR(-ENOMEM); + return (void __percpu __force *)ERR_PTR(-ENOMEM); get_online_cpus(); for_each_online_cpu(cpu) { @@ -451,7 +451,7 @@ fail: put_online_cpus(); free_percpu(cpu_events); - return ERR_PTR(err); + return (void __percpu __force *)ERR_PTR(err); } EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); @@ -459,7 +459,7 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); * unregister_wide_hw_breakpoint - unregister a wide breakpoint in the kernel * @cpu_events: the per cpu set of events to unregister */ -void unregister_wide_hw_breakpoint(struct perf_event **cpu_events) +void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events) { int cpu; struct perf_event **pevent; -- cgit v1.2.2 From 622ea685f1fafdf84d612440535c84341f0860b8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sat, 27 Feb 2010 14:53:07 -0800 Subject: rcu: Fix holdoff for accelerated GPs for last non-dynticked CPU Make the holdoff only happen when the full number of attempts have been made. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267311188-16603-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree_plugin.h | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 464ad2cdee00..79b53bda8943 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -1010,6 +1010,10 @@ int rcu_needs_cpu(int cpu) int c = 0; int thatcpu; + /* Check for being in the holdoff period. */ + if (per_cpu(rcu_dyntick_holdoff, cpu) == jiffies) + return rcu_needs_cpu_quick_check(cpu); + /* Don't bother unless we are the last non-dyntick-idle CPU. */ for_each_cpu_not(thatcpu, nohz_cpu_mask) if (thatcpu != cpu) { @@ -1041,10 +1045,8 @@ int rcu_needs_cpu(int cpu) } /* If RCU callbacks are still pending, RCU still needs this CPU. */ - if (c) { + if (c) raise_softirq(RCU_SOFTIRQ); - per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; - } return c; } -- cgit v1.2.2 From ae1f30384baef4056438d81b305a6a5199b0d16c Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 28 Feb 2010 19:42:38 +0100 Subject: tracing: Include irqflags headers from trace clock trace_clock.c includes spinlock.h, which ends up including asm/system.h, which in turn includes linux/irqflags.h in x86. So the definition of raw_local_irq_save is luckily covered there, but this is not the case in parisc: tip/kernel/trace/trace_clock.c:86: error: implicit declaration of function 'raw_local_irq_save' tip/kernel/trace/trace_clock.c:112: error: implicit declaration of function 'raw_local_irq_restore' We need to include linux/irqflags.h directly from trace_clock.c to avoid such build error. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Robert Richter Cc: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/trace/trace_clock.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c index 84a3a7ba072a..6fbfb8f417b9 100644 --- a/kernel/trace/trace_clock.c +++ b/kernel/trace/trace_clock.c @@ -13,6 +13,7 @@ * Tracer plugins will chose a default from these clocks. */ #include +#include #include #include #include -- cgit v1.2.2 From 9a8c28c8311e30ba97499447d5a11662f5aea094 Mon Sep 17 00:00:00 2001 From: Dmitry Monakhov Date: Sat, 27 Feb 2010 20:35:13 +0300 Subject: blktrace: perform cleanup after setup error Currently even if BLKTRACESETUP ioctl has failed user must call BLKTRACETEARDOWN to be shure what all staff was cleaned, which is contr-intuitive. Let's setup ioctl make necessery cleanup by it self. Signed-off-by: Dmitry Monakhov Signed-off-by: Jens Axboe --- kernel/trace/blktrace.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index d9d6206e0b14..07f945a99430 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -540,9 +540,10 @@ int blk_trace_setup(struct request_queue *q, char *name, dev_t dev, if (ret) return ret; - if (copy_to_user(arg, &buts, sizeof(buts))) + if (copy_to_user(arg, &buts, sizeof(buts))) { + blk_trace_remove(q); return -EFAULT; - + } return 0; } EXPORT_SYMBOL_GPL(blk_trace_setup); -- cgit v1.2.2 From 1e259e0a9982078896f3404240096cbea01daca4 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 28 Feb 2010 20:51:15 +0100 Subject: hw-breakpoints: Remove stub unthrottle callback We support event unthrottling in breakpoint events. It means that if we have more than sysctl_perf_event_sample_rate/HZ, perf will throttle, ignoring subsequent events until the next tick. So if ptrace exceeds this max rate, it will omit events, which breaks the ptrace determinism that is supposed to report every triggered breakpoints. This is likely to happen if we set sysctl_perf_event_sample_rate to 1. This patch removes support for unthrottling in breakpoint events to break throttling and restore ptrace determinism. Signed-off-by: Frederic Weisbecker Cc: 2.6.33.x Cc: Peter Zijlstra Cc: K.Prasad Cc: Paul Mackerras --- kernel/hw_breakpoint.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 967e66143e11..4d99512ee149 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -489,5 +489,4 @@ struct pmu perf_ops_bp = { .enable = arch_install_hw_breakpoint, .disable = arch_uninstall_hw_breakpoint, .read = hw_breakpoint_pmu_read, - .unthrottle = hw_breakpoint_pmu_unthrottle }; -- cgit v1.2.2 From 4b17764737bb4ee3364b8bfa2059f51ebc19ccd6 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Mon, 1 Mar 2010 00:02:23 -0800 Subject: sparc: Support show_unhandled_signals. Just faults right now, will add other traps later. Signed-off-by: David S. Miller --- kernel/sysctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 8a68b2448468..33e7a38b6eb9 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1441,7 +1441,7 @@ static struct ctl_table fs_table[] = { }; static struct ctl_table debug_table[] = { -#if defined(CONFIG_X86) || defined(CONFIG_PPC) +#if defined(CONFIG_X86) || defined(CONFIG_PPC) || defined(CONFIG_SPARC) { .procname = "exception-trace", .data = &show_unhandled_signals, -- cgit v1.2.2 From 90a6501f94aedd7fb40f5556334843194fb598be Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sun, 28 Feb 2010 08:32:18 -0800 Subject: sched, rcu: Fix rcu_dereference() for RCU-lockdep Make rcu_dereference() of runqueue data structures be rcu_dereference_sched(). Located-by: Ingo Molnar Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <20100228163218.GD6846@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 3e1fd96c6cf9..5a5ea2cd924f 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -3476,7 +3476,7 @@ static void run_rebalance_domains(struct softirq_action *h) static inline int on_null_domain(int cpu) { - return !rcu_dereference(cpu_rq(cpu)->sd); + return !rcu_dereference_sched(cpu_rq(cpu)->sd); } /* -- cgit v1.2.2 From 37b99dd5372cff42f83210c280f314f10f99138e Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Mon, 1 Mar 2010 21:55:51 +0800 Subject: resource: Fix generic page_is_ram() for partial RAM pages The System RAM walk shall skip partial RAM pages and avoid calling func() on them. So that page_is_ram() return 0 for a partial RAM page. In particular, it shall not call func() with len=0. This fixes a boot time bug reported by Sachin and root caused by Thomas: > >>> WARNING: at arch/x86/mm/ioremap.c:111 __ioremap_caller+0x169/0x2f1() > >>> Hardware name: BladeCenter LS21 -[79716AA]- > >>> Modules linked in: > >>> Pid: 0, comm: swapper Not tainted 2.6.33-git6-autotest #1 > >>> Call Trace: > >>> [] ? __ioremap_caller+0x169/0x2f1 > >>> [] warn_slowpath_common+0x77/0xa4 > >>> [] warn_slowpath_null+0xf/0x11 > >>> [] __ioremap_caller+0x169/0x2f1 > >>> [] ? acpi_os_map_memory+0x12/0x1b > >>> [] ioremap_nocache+0x12/0x14 > >>> [] acpi_os_map_memory+0x12/0x1b > >>> [] acpi_tb_verify_table+0x29/0x5b > >>> [] acpi_load_tables+0x39/0x15a > >>> [] acpi_early_init+0x60/0xf5 > >>> [] start_kernel+0x397/0x3a7 > >>> [] x86_64_start_reservations+0xa5/0xa9 > >>> [] x86_64_start_kernel+0xe1/0xe8 > >>> ---[ end trace 4eaa2a86a8e2da22 ]--- > >>> ioremap reserve_memtype failed -22 The return code is -EINVAL, so it failed in the is_ram check, which is not too surprising > BIOS-provided physical RAM map: > BIOS-e820: 0000000000000000 - 000000000009c000 (usable) > BIOS-e820: 000000000009c000 - 00000000000a0000 (reserved) > BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved) > BIOS-e820: 0000000000100000 - 00000000cffa3900 (usable) > BIOS-e820: 00000000cffa3900 - 00000000cffa7400 (ACPI data) The ACPI data is not starting on a page boundary and neither does the usable RAM area end on a page boundary. Very useful ! > ACPI: DSDT 00000000cffa3900 036CE (v01 IBM SERLEWIS 00001000 INTL 20060912) ACPI is trying to map DSDT at cffa3900, which results in a check vs. cffa3000 which is the relevant page boundary. The generic is_ram check correctly identifies that as RAM because it's in the usable resource area. The old e820 based is_ram check does not take overlapping resource areas into account. That's why it works. CC: Sachin Sant CC: Thomas Gleixner CC: KAMEZAWA Hiroyuki Signed-off-by: Wu Fengguang LKML-Reference: <20100301135551.GA9998@localhost> Signed-off-by: H. Peter Anvin --- kernel/resource.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index 03c897f7935e..8f0e3d0f4bff 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -274,7 +274,7 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, void *arg, int (*func)(unsigned long, unsigned long, void *)) { struct resource res; - unsigned long pfn, len; + unsigned long pfn, end_pfn; u64 orig_end; int ret = -1; @@ -284,9 +284,10 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, orig_end = res.end; while ((res.start < res.end) && (find_next_system_ram(&res, "System RAM") >= 0)) { - pfn = (unsigned long)(res.start >> PAGE_SHIFT); - len = (unsigned long)((res.end + 1 - res.start) >> PAGE_SHIFT); - ret = (*func)(pfn, len, arg); + pfn = (res.start + PAGE_SIZE - 1) >> PAGE_SHIFT; + end_pfn = (res.end + 1) >> PAGE_SHIFT; + if (end_pfn > pfn) + ret = (*func)(pfn, end_pfn - pfn, arg); if (ret) break; res.start = res.end + 1; -- cgit v1.2.2 From dce46a04d55d6358d2d4ab44a4946a19f9425fe2 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 1 Mar 2010 18:48:52 -0800 Subject: early_res: Need to save the allocation name in drop_range_partial() During free_early_partial(), reserve_early_without_check() could end extending the early_res area from __check_and_double_early_res(); as a result, the location of the name for the current reservation could change. Therefore, we need to save a local copy of the name. [ hpa: rewrote comment and checkin description ] Signed-off-by: Yinghai Lu LKML-Reference: <4B8C7C94.7070000@kernel.org> Signed-off-by: H. Peter Anvin --- kernel/early_res.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/early_res.c b/kernel/early_res.c index 9ab11cd84853..3cb2c661bb78 100644 --- a/kernel/early_res.c +++ b/kernel/early_res.c @@ -79,9 +79,19 @@ static void __init drop_range_partial(int i, u64 start, u64 end) /* make head segment */ early_res[i].end = common_start; if (old_end > common_end) { + char name[15]; + + /* + * Save a local copy of the name, since the + * early_res array could get resized inside + * reserve_early_without_check() -> + * __check_and_double_early_res(), which would + * make the current name pointer invalid. + */ + strncpy(name, early_res[i].name, + sizeof(early_res[i].name) - 1); /* add another for left over on tail */ - reserve_early_without_check(common_end, old_end, - early_res[i].name); + reserve_early_without_check(common_end, old_end, name); } return; } else { -- cgit v1.2.2 From ad6759fbf35d104dbf573cd6f4c6784ad6823f7e Mon Sep 17 00:00:00 2001 From: john stultz Date: Mon, 1 Mar 2010 12:34:43 -0800 Subject: timekeeping: Prevent oops when GENERIC_TIME=n Aaro Koskinen reported an issue in kernel.org bugzilla #15366, where on non-GENERIC_TIME systems, accessing /sys/devices/system/clocksource/clocksource0/current_clocksource results in an oops. It seems the timekeeper/clocksource rework missed initializing the curr_clocksource value in the !GENERIC_TIME case. Thanks to Aaro for reporting and diagnosing the issue as well as testing the fix! Reported-by: Aaro Koskinen Signed-off-by: John Stultz Cc: Martin Schwidefsky Cc: stable@kernel.org LKML-Reference: <1267475683.4216.61.camel@localhost.localdomain> Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 1f663d23e85e..1f5dde637457 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -592,6 +592,10 @@ static inline void clocksource_select(void) { } */ static int __init clocksource_done_booting(void) { + mutex_lock(&clocksource_mutex); + curr_clocksource = clocksource_default_clock(); + mutex_unlock(&clocksource_mutex); + finished_booting = 1; /* -- cgit v1.2.2 From 320ebf09cbb6d01954c9a060266aa8e0d27f4638 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 2 Mar 2010 12:35:37 +0100 Subject: perf, x86: Restrict the ANY flag The ANY flag can show SMT data of another task (like 'top'), so we want to disable it when system-wide profiling is disabled. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index a661e7991865..482d5e1d3764 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -56,21 +56,6 @@ static atomic_t nr_task_events __read_mostly; */ int sysctl_perf_event_paranoid __read_mostly = 1; -static inline bool perf_paranoid_tracepoint_raw(void) -{ - return sysctl_perf_event_paranoid > -1; -} - -static inline bool perf_paranoid_cpu(void) -{ - return sysctl_perf_event_paranoid > 0; -} - -static inline bool perf_paranoid_kernel(void) -{ - return sysctl_perf_event_paranoid > 1; -} - int sysctl_perf_event_mlock __read_mostly = 512; /* 'free' kb per user */ /* -- cgit v1.2.2 From f41496607e03ab99f263b8e26689ad0fc853007f Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 2 Mar 2010 11:21:09 -0800 Subject: resource: Fix broken indentation Fix broken indentation in patch 37b99dd5372cff42f83210c280f314f10f99138e. Reported-by: Linus Torvalds Cc: Wu Fengguang LKML-Reference: <20100301135551.GA9998@localhost> Signed-off-by: H. Peter Anvin --- kernel/resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index 8f0e3d0f4bff..91f430fd467e 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -287,7 +287,7 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, pfn = (res.start + PAGE_SIZE - 1) >> PAGE_SHIFT; end_pfn = (res.end + 1) >> PAGE_SHIFT; if (end_pfn > pfn) - ret = (*func)(pfn, end_pfn - pfn, arg); + ret = (*func)(pfn, end_pfn - pfn, arg); if (ret) break; res.start = res.end + 1; -- cgit v1.2.2 From ac91d85456372a90af5b85eb6620fd2efb1e431b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 2 Mar 2010 17:54:50 +0800 Subject: tracing: Fix warning in s_next of trace file ops This warning in s_next() can be triggered by lseek(): [] ? s_next+0x77/0x80 [] warn_slowpath_common+0x81/0xa0 [] ? s_next+0x77/0x80 [] warn_slowpath_null+0x1a/0x20 [] s_next+0x77/0x80 [] traverse+0x117/0x200 [] seq_lseek+0xa3/0x120 [] ? seq_lseek+0x0/0x120 [] vfs_llseek+0x41/0x50 [] sys_llseek+0x66/0xa0 [] sysenter_do_call+0x12/0x26 The iterator "leftover" variable is zeroed in the opening of the trace file. But lseek can call s_start() which will call s_next() without reseting the "leftover" variable back to zero, which might trigger the WARN_ON_ONCE(iter->leftover) that is in s_next(). Cc: stable@kernel.org Signed-off-by: Lai Jiangshan LKML-Reference: <4B8CE06A.9090207@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 032c57ca6502..5edf410bc540 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1703,6 +1703,7 @@ static void *s_start(struct seq_file *m, loff_t *pos) ftrace_enable_cpu(); + iter->leftover = 0; for (p = iter; p && l < *pos; p = s_next(m, p, &l)) ; -- cgit v1.2.2 From 8737c9305bd5602b11f7eb4655d5695d4a42a0c6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 24 Dec 2009 06:47:55 -0500 Subject: Switch may_open() and break_lease() to passing O_... ... instead of mixing FMODE_ and O_ Signed-off-by: Al Viro --- kernel/sysctl_binary.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c index 8f5d16e0707a..8cd50d8f9bde 100644 --- a/kernel/sysctl_binary.c +++ b/kernel/sysctl_binary.c @@ -1331,7 +1331,7 @@ static ssize_t binary_sysctl(const int *name, int nlen, ssize_t result; char *pathname; int flags; - int acc_mode, fmode; + int acc_mode; pathname = sysctl_getname(name, nlen, &table); result = PTR_ERR(pathname); @@ -1342,15 +1342,12 @@ static ssize_t binary_sysctl(const int *name, int nlen, if (oldval && oldlen && newval && newlen) { flags = O_RDWR; acc_mode = MAY_READ | MAY_WRITE; - fmode = FMODE_READ | FMODE_WRITE; } else if (newval && newlen) { flags = O_WRONLY; acc_mode = MAY_WRITE; - fmode = FMODE_WRITE; } else if (oldval && oldlen) { flags = O_RDONLY; acc_mode = MAY_READ; - fmode = FMODE_READ; } else { result = 0; goto out_putname; @@ -1361,7 +1358,7 @@ static ssize_t binary_sysctl(const int *name, int nlen, if (result) goto out_putname; - result = may_open(&nd.path, acc_mode, fmode); + result = may_open(&nd.path, acc_mode, flags); if (result) goto out_putpath; -- cgit v1.2.2 From 2096f759abcb42200a81d776f597362fd9265024 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 30 Jan 2010 13:16:21 -0500 Subject: New helper: path_is_under(path1, path2) Analog of is_subdir for vfsmount,dentry pairs, moved from audit_tree.c Signed-off-by: Al Viro --- kernel/audit_tree.c | 51 ++++++++++++--------------------------------------- 1 file changed, 12 insertions(+), 39 deletions(-) (limited to 'kernel') diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index 4b05bd9479db..f09b42d9c32d 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -603,22 +603,6 @@ skip_it: mutex_unlock(&audit_filter_mutex); } -static int is_under(struct vfsmount *mnt, struct dentry *dentry, - struct path *path) -{ - if (mnt != path->mnt) { - for (;;) { - if (mnt->mnt_parent == mnt) - return 0; - if (mnt->mnt_parent == path->mnt) - break; - mnt = mnt->mnt_parent; - } - dentry = mnt->mnt_mountpoint; - } - return is_subdir(dentry, path->dentry); -} - int audit_make_tree(struct audit_krule *rule, char *pathname, u32 op) { @@ -714,29 +698,24 @@ int audit_tag_tree(char *old, char *new) { struct list_head cursor, barrier; int failed = 0; - struct path path; + struct path path1, path2; struct vfsmount *tagged; struct list_head list; - struct vfsmount *mnt; - struct dentry *dentry; int err; - err = kern_path(new, 0, &path); + err = kern_path(new, 0, &path2); if (err) return err; - tagged = collect_mounts(&path); - path_put(&path); + tagged = collect_mounts(&path2); + path_put(&path2); if (!tagged) return -ENOMEM; - err = kern_path(old, 0, &path); + err = kern_path(old, 0, &path1); if (err) { drop_collected_mounts(tagged); return err; } - mnt = mntget(path.mnt); - dentry = dget(path.dentry); - path_put(&path); list_add_tail(&list, &tagged->mnt_list); @@ -747,6 +726,7 @@ int audit_tag_tree(char *old, char *new) while (cursor.next != &tree_list) { struct audit_tree *tree; struct vfsmount *p; + int good_one = 0; tree = container_of(cursor.next, struct audit_tree, list); get_tree(tree); @@ -754,23 +734,17 @@ int audit_tag_tree(char *old, char *new) list_add(&cursor, &tree->list); mutex_unlock(&audit_filter_mutex); - err = kern_path(tree->pathname, 0, &path); - if (err) { - put_tree(tree); - mutex_lock(&audit_filter_mutex); - continue; + err = kern_path(tree->pathname, 0, &path2); + if (!err) { + good_one = path_is_under(&path1, &path2); + path_put(&path2); } - spin_lock(&vfsmount_lock); - if (!is_under(mnt, dentry, &path)) { - spin_unlock(&vfsmount_lock); - path_put(&path); + if (!good_one) { put_tree(tree); mutex_lock(&audit_filter_mutex); continue; } - spin_unlock(&vfsmount_lock); - path_put(&path); list_for_each_entry(p, &list, mnt_list) { failed = tag_chunk(p->mnt_root->d_inode, tree); @@ -820,8 +794,7 @@ int audit_tag_tree(char *old, char *new) list_del(&cursor); list_del(&list); mutex_unlock(&audit_filter_mutex); - dput(dentry); - mntput(mnt); + path_put(&path1); drop_collected_mounts(tagged); return failed; } -- cgit v1.2.2 From 1f707137b55764740981d022d29c622832a61880 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 30 Jan 2010 22:51:25 -0500 Subject: new helper: iterate_mounts() apply function to vfsmounts in set returned by collect_mounts(), stop if it returns non-zero. Signed-off-by: Al Viro --- kernel/audit_tree.c | 49 ++++++++++++++++--------------------------------- 1 file changed, 16 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index f09b42d9c32d..028e85663f27 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -548,6 +548,11 @@ int audit_remove_tree_rule(struct audit_krule *rule) return 0; } +static int compare_root(struct vfsmount *mnt, void *arg) +{ + return mnt->mnt_root->d_inode == arg; +} + void audit_trim_trees(void) { struct list_head cursor; @@ -559,7 +564,6 @@ void audit_trim_trees(void) struct path path; struct vfsmount *root_mnt; struct node *node; - struct list_head list; int err; tree = container_of(cursor.next, struct audit_tree, list); @@ -577,24 +581,16 @@ void audit_trim_trees(void) if (!root_mnt) goto skip_it; - list_add_tail(&list, &root_mnt->mnt_list); spin_lock(&hash_lock); list_for_each_entry(node, &tree->chunks, list) { - struct audit_chunk *chunk = find_chunk(node); - struct inode *inode = chunk->watch.inode; - struct vfsmount *mnt; + struct inode *inode = find_chunk(node)->watch.inode; node->index |= 1U<<31; - list_for_each_entry(mnt, &list, mnt_list) { - if (mnt->mnt_root->d_inode == inode) { - node->index &= ~(1U<<31); - break; - } - } + if (iterate_mounts(compare_root, inode, root_mnt)) + node->index &= ~(1U<<31); } spin_unlock(&hash_lock); trim_marked(tree); put_tree(tree); - list_del_init(&list); drop_collected_mounts(root_mnt); skip_it: mutex_lock(&audit_filter_mutex); @@ -622,13 +618,17 @@ void audit_put_tree(struct audit_tree *tree) put_tree(tree); } +static int tag_mount(struct vfsmount *mnt, void *arg) +{ + return tag_chunk(mnt->mnt_root->d_inode, arg); +} + /* called with audit_filter_mutex */ int audit_add_tree_rule(struct audit_krule *rule) { struct audit_tree *seed = rule->tree, *tree; struct path path; - struct vfsmount *mnt, *p; - struct list_head list; + struct vfsmount *mnt; int err; list_for_each_entry(tree, &tree_list, list) { @@ -654,16 +654,9 @@ int audit_add_tree_rule(struct audit_krule *rule) err = -ENOMEM; goto Err; } - list_add_tail(&list, &mnt->mnt_list); get_tree(tree); - list_for_each_entry(p, &list, mnt_list) { - err = tag_chunk(p->mnt_root->d_inode, tree); - if (err) - break; - } - - list_del(&list); + err = iterate_mounts(tag_mount, tree, mnt); drop_collected_mounts(mnt); if (!err) { @@ -700,7 +693,6 @@ int audit_tag_tree(char *old, char *new) int failed = 0; struct path path1, path2; struct vfsmount *tagged; - struct list_head list; int err; err = kern_path(new, 0, &path2); @@ -717,15 +709,12 @@ int audit_tag_tree(char *old, char *new) return err; } - list_add_tail(&list, &tagged->mnt_list); - mutex_lock(&audit_filter_mutex); list_add(&barrier, &tree_list); list_add(&cursor, &barrier); while (cursor.next != &tree_list) { struct audit_tree *tree; - struct vfsmount *p; int good_one = 0; tree = container_of(cursor.next, struct audit_tree, list); @@ -746,12 +735,7 @@ int audit_tag_tree(char *old, char *new) continue; } - list_for_each_entry(p, &list, mnt_list) { - failed = tag_chunk(p->mnt_root->d_inode, tree); - if (failed) - break; - } - + failed = iterate_mounts(tag_mount, tree, tagged); if (failed) { put_tree(tree); mutex_lock(&audit_filter_mutex); @@ -792,7 +776,6 @@ int audit_tag_tree(char *old, char *new) } list_del(&barrier); list_del(&cursor); - list_del(&list); mutex_unlock(&audit_filter_mutex); path_put(&path1); drop_collected_mounts(tagged); -- cgit v1.2.2 From a27341cd5fcb7cf2d2d4726e9f324009f7162c00 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 2 Mar 2010 08:36:46 -0800 Subject: Prioritize synchronous signals over 'normal' signals This makes sure that we pick the synchronous signals caused by a processor fault over any pending regular asynchronous signals sent to use by [t]kill(). This is not strictly required semantics, but it makes it _much_ easier for programs like Wine that expect to find the fault information in the signal stack. Without this, if a non-synchronous signal gets picked first, the delayed asynchronous signal will have its signal context pointing to the new signal invocation, rather than the instruction that caused the SIGSEGV or SIGBUS in the first place. This is not all that pretty, and we're discussing making the synchronous signals more explicit rather than have these kinds of implicit preferences of SIGSEGV and friends. See for example http://bugzilla.kernel.org/show_bug.cgi?id=15395 for some of the discussion. But in the meantime this is a simple and fairly straightforward work-around, and the whole if (x & Y) x &= Y; thing can be compiled into (and gcc does do it) just three instructions: movq %rdx, %rax andl $Y, %eax cmovne %rax, %rdx so it is at least a simple solution to a subtle issue. Reported-and-tested-by: Pavel Vilim Acked-by: Oleg Nesterov Signed-off-by: Linus Torvalds --- kernel/signal.c | 43 ++++++++++++++++++++++++++++++------------- 1 file changed, 30 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 934ae5e687b9..5bb9baffa4f1 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -159,6 +159,10 @@ void recalc_sigpending(void) /* Given the mask, find the first available signal that should be serviced. */ +#define SYNCHRONOUS_MASK \ + (sigmask(SIGSEGV) | sigmask(SIGBUS) | sigmask(SIGILL) | \ + sigmask(SIGTRAP) | sigmask(SIGFPE)) + int next_signal(struct sigpending *pending, sigset_t *mask) { unsigned long i, *s, *m, x; @@ -166,26 +170,39 @@ int next_signal(struct sigpending *pending, sigset_t *mask) s = pending->signal.sig; m = mask->sig; + + /* + * Handle the first word specially: it contains the + * synchronous signals that need to be dequeued first. + */ + x = *s &~ *m; + if (x) { + if (x & SYNCHRONOUS_MASK) + x &= SYNCHRONOUS_MASK; + sig = ffz(~x) + 1; + return sig; + } + switch (_NSIG_WORDS) { default: - for (i = 0; i < _NSIG_WORDS; ++i, ++s, ++m) - if ((x = *s &~ *m) != 0) { - sig = ffz(~x) + i*_NSIG_BPW + 1; - break; - } + for (i = 1; i < _NSIG_WORDS; ++i) { + x = *++s &~ *++m; + if (!x) + continue; + sig = ffz(~x) + i*_NSIG_BPW + 1; + break; + } break; - case 2: if ((x = s[0] &~ m[0]) != 0) - sig = 1; - else if ((x = s[1] &~ m[1]) != 0) - sig = _NSIG_BPW + 1; - else + case 2: + x = s[1] &~ m[1]; + if (!x) break; - sig += ffz(~x); + sig = ffz(~x) + _NSIG_BPW + 1; break; - case 1: if ((x = *s &~ *m) != 0) - sig = ffz(~x) + 1; + case 1: + /* Nothing to do */ break; } -- cgit v1.2.2 From 74781387822cd7a549123ae2b35862bf802689be Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Thu, 4 Mar 2010 13:30:22 +0800 Subject: padata: Allocate the cpumask for the padata instance The cpumask of the padata instance was used without allocated. This caused boot crashes if CONFIG_CPUMASK_OFFSTACK is enabled. This patch fixes this by doing proper allocation for this cpumask. Signed-off-by: Steffen Klassert Signed-off-by: Herbert Xu --- kernel/padata.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/padata.c b/kernel/padata.c index 6f9bcb8313d6..93caf65ff57c 100644 --- a/kernel/padata.c +++ b/kernel/padata.c @@ -642,6 +642,9 @@ struct padata_instance *padata_alloc(const struct cpumask *cpumask, if (!pd) goto err_free_inst; + if (!alloc_cpumask_var(&pinst->cpumask, GFP_KERNEL)) + goto err_free_pd; + rcu_assign_pointer(pinst->pd, pd); pinst->wq = wq; @@ -654,12 +657,14 @@ struct padata_instance *padata_alloc(const struct cpumask *cpumask, pinst->cpu_notifier.priority = 0; err = register_hotcpu_notifier(&pinst->cpu_notifier); if (err) - goto err_free_pd; + goto err_free_cpumask; mutex_init(&pinst->lock); return pinst; +err_free_cpumask: + free_cpumask_var(pinst->cpumask); err_free_pd: padata_free_pd(pd); err_free_inst: @@ -685,6 +690,7 @@ void padata_free(struct padata_instance *pinst) unregister_hotcpu_notifier(&pinst->cpu_notifier); padata_free_pd(pinst->pd); + free_cpumask_var(pinst->cpumask); kfree(pinst); } EXPORT_SYMBOL(padata_free); -- cgit v1.2.2 From db1466b3e1bd1727375cdbfcbea4bcce2f860f61 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 3 Mar 2010 07:46:56 -0800 Subject: rcu: Use wrapper function instead of exporting tasklist_lock Lockdep-RCU commit d11c563d exported tasklist_lock, which is not a good thing. This patch instead exports a function that uses lockdep to check whether tasklist_lock is held. Suggested-by: Christoph Hellwig Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com Cc: Christoph Hellwig LKML-Reference: <1267631219-8713-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/exit.c | 2 +- kernel/fork.c | 9 ++++++++- kernel/pid.c | 4 +++- 3 files changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 45ed043b8bf5..fed3a4db6f04 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -87,7 +87,7 @@ static void __exit_signal(struct task_struct *tsk) sighand = rcu_dereference_check(tsk->sighand, rcu_read_lock_held() || - lockdep_is_held(&tasklist_lock)); + lockdep_tasklist_lock_is_held()); spin_lock(&sighand->siglock); posix_cpu_timers_exit(tsk); diff --git a/kernel/fork.c b/kernel/fork.c index 17bbf093356d..8691c540a470 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -86,7 +86,14 @@ int max_threads; /* tunable limit on nr_threads */ DEFINE_PER_CPU(unsigned long, process_counts) = 0; __cacheline_aligned DEFINE_RWLOCK(tasklist_lock); /* outer */ -EXPORT_SYMBOL_GPL(tasklist_lock); + +#ifdef CONFIG_PROVE_RCU +int lockdep_tasklist_lock_is_held(void) +{ + return lockdep_is_held(&tasklist_lock); +} +EXPORT_SYMBOL_GPL(lockdep_tasklist_lock_is_held); +#endif /* #ifdef CONFIG_PROVE_RCU */ int nr_processes(void) { diff --git a/kernel/pid.c b/kernel/pid.c index b08e697cd83f..b6064405f367 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -367,7 +367,9 @@ struct task_struct *pid_task(struct pid *pid, enum pid_type type) struct task_struct *result = NULL; if (pid) { struct hlist_node *first; - first = rcu_dereference_check(pid->tasks[type].first, rcu_read_lock_held() || lockdep_is_held(&tasklist_lock)); + first = rcu_dereference_check(pid->tasks[type].first, + rcu_read_lock_held() || + lockdep_tasklist_lock_is_held()); if (first) result = hlist_entry(first, struct task_struct, pids[(type)].node); } -- cgit v1.2.2 From cc5b83a9f884fe8722a275069a5a6fde39988455 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 3 Mar 2010 07:46:59 -0800 Subject: rcu: Add control variables to lockdep_rcu_dereference() diagnostics Add the values of rcu_scheduler_active() and debug_locks() to the lockdep_rcu_dereference() output to help diagnose RCU lockdep splats that occur shortly after the scheduler starts. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267631219-8713-4-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 0c30d0455de1..681bc2e1e187 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3822,6 +3822,7 @@ void lockdep_rcu_dereference(const char *file, const int line) printk("%s:%d invoked rcu_dereference_check() without protection!\n", file, line); printk("\nother info that might help us debug this:\n\n"); + printk("\nrcu_scheduler_active = %d, debug_locks = %d\n", rcu_scheduler_active, debug_locks); lockdep_print_held_locks(curr); printk("\nstack backtrace:\n"); dump_stack(); -- cgit v1.2.2 From 8d53dd546f36073e0d29b0cfc24c665db301e3e7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 3 Mar 2010 17:50:18 -0800 Subject: rcu, ftrace: Fix RCU lockdep splat in ftrace_perf_buf_prepare() Change the pair of rcu_dereference() calls in ftrace_perf_buf_prepare() to rcu_dereference_sched(). Signed-off-by: Paul E. McKenney Acked-by: Frederic Weisbecker Cc: Steven Rostedt Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com Cc: Frederic Weisbecker LKML-Reference: <1267667418-32233-3-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_event_profile.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index f0d693005075..c1cc3ab633de 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -138,9 +138,9 @@ __kprobes void *ftrace_perf_buf_prepare(int size, unsigned short type, cpu = smp_processor_id(); if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); + trace_buf = rcu_dereference_sched(perf_trace_buf_nmi); else - trace_buf = rcu_dereference(perf_trace_buf); + trace_buf = rcu_dereference_sched(perf_trace_buf); if (!trace_buf) goto err; -- cgit v1.2.2 From 801c29fd1fdeb84f60241beb445ff5db154450ae Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 5 Mar 2010 20:02:19 -0500 Subject: function-graph: Fix unused reference to ftrace_set_func() The declaration of ftrace_set_func() is at the start of the ftrace.c file and wrapped with a #ifdef CONFIG_FUNCTION_GRAPH condition. If function graph tracing is enabled but CONFIG_DYNAMIC_FTRACE is not, a warning about that function being declared static and unused is given. This really should have been placed within the CONFIG_FUNCTION_GRAPH condition that uses ftrace_set_func(). Moving the declaration down fixes the warning and makes the code cleaner. Reported-by: Peter Zijlstra Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index d996353473fd..d0407c9f368c 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -85,10 +85,6 @@ ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub; ftrace_func_t __ftrace_trace_function __read_mostly = ftrace_stub; ftrace_func_t ftrace_pid_function __read_mostly = ftrace_stub; -#ifdef CONFIG_FUNCTION_GRAPH_TRACER -static int ftrace_set_func(unsigned long *array, int *idx, char *buffer); -#endif - static void ftrace_list_func(unsigned long ip, unsigned long parent_ip) { struct ftrace_ops *op = ftrace_list; @@ -2300,6 +2296,8 @@ __setup("ftrace_filter=", set_ftrace_filter); #ifdef CONFIG_FUNCTION_GRAPH_TRACER static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata; +static int ftrace_set_func(unsigned long *array, int *idx, char *buffer); + static int __init set_graph_function(char *str) { strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE); -- cgit v1.2.2 From a094fe04c751698a18c3a0d376a3bdb117f1e0d8 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 5 Mar 2010 20:08:58 -0500 Subject: function-graph: Use comment notation for func names of dangling '}' When a '}' does not have a matching function start, the name is printed within parenthesis. But this makes it confusing between ending '}' and function starts. This patch makes the function name appear in C comment notation. Old view: 3) 1.281 us | } (might_fault) 3) 3.620 us | } (filldir) 3) 5.251 us | } (call_filldir) 3) | call_filldir() { 3) | filldir() { New view: 3) 1.281 us | } /* might_fault */ 3) 3.620 us | } /* filldir */ 3) 5.251 us | } /* call_filldir */ 3) | call_filldir() { 3) | filldir() { Requested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index e998a824e9db..7b1f24618d97 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -920,7 +920,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, if (!ret) return TRACE_TYPE_PARTIAL_LINE; } else { - ret = trace_seq_printf(s, "} (%ps)\n", (void *)trace->func); + ret = trace_seq_printf(s, "} /* %ps */\n", (void *)trace->func); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } -- cgit v1.2.2 From 1acaa1b2d9b5904c9cce06122990a2d71046ce16 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 5 Mar 2010 18:23:50 -0300 Subject: tracing: Update the comm field in the right variable in update_max_tr The latency output showed: # | task: -3 (uid:0 nice:0 policy:1 rt_prio:99) The comm is missing in the "task:" and it looks like a minus 3 is the output. The correct display should be: # | task: migration/0-3 (uid:0 nice:0 policy:1 rt_prio:99) The problem is that the comm is being stored in the wrong data structure. The max_tr.data[cpu] is what stores the comm, not the tr->data[cpu]. Before this patch the max_tr.data[cpu]->comm was zeroed and the /debug/trace ended up showing just the '-' sign followed by the pid. Also remove a needless initialization of max_data. Signed-off-by: Arnaldo Carvalho de Melo LKML-Reference: <1267824230-23861-1-git-send-email-acme@infradead.org> Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 032c57ca6502..6efd5cb3c252 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -592,7 +592,7 @@ static void __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) { struct trace_array_cpu *data = tr->data[cpu]; - struct trace_array_cpu *max_data = tr->data[cpu]; + struct trace_array_cpu *max_data; max_tr.cpu = cpu; max_tr.time_start = data->preempt_timestamp; @@ -602,7 +602,7 @@ __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) max_data->critical_start = data->critical_start; max_data->critical_end = data->critical_end; - memcpy(data->comm, tsk->comm, TASK_COMM_LEN); + memcpy(max_data->comm, tsk->comm, TASK_COMM_LEN); max_data->pid = tsk->pid; max_data->uid = task_uid(tsk); max_data->nice = tsk->static_prio - 20 - MAX_RT_PRIO; -- cgit v1.2.2 From 0e95017355dcf43031da6d0e360a748717e56df1 Mon Sep 17 00:00:00 2001 From: Tim Bird Date: Thu, 25 Feb 2010 15:36:43 -0800 Subject: function-graph: Add tracing_thresh support to function_graph tracer Add support for tracing_thresh to the function_graph tracer. This version of this feature isolates the checks into new entry and return functions, to avoid adding more conditional code into the main function_graph paths. When the tracing_thresh is set and the function graph tracer is enabled, only the functions that took longer than the time in microseconds that was set in tracing_thresh are recorded. To do this efficiently, only the function exits are recorded: [tracing]# echo 100 > tracing_thresh [tracing]# echo function_graph > current_tracer [tracing]# cat trace # tracer: function_graph # # CPU DURATION FUNCTION CALLS # | | | | | | | 1) ! 119.214 us | } /* smp_apic_timer_interrupt */ 1) <========== | 0) ! 101.527 us | } /* __rcu_process_callbacks */ 0) ! 126.461 us | } /* rcu_process_callbacks */ 0) ! 145.111 us | } /* __do_softirq */ 0) ! 149.667 us | } /* do_softirq */ 0) ! 168.817 us | } /* irq_exit */ 0) ! 248.254 us | } /* smp_apic_timer_interrupt */ Also, add support for specifying tracing_thresh on the kernel command line. When used like so: "tracing_thresh=200 ftrace=function_graph" this can be used to analyse system startup. It is important to disable tracing soon after boot, in order to avoid losing the trace data. Acked-by: Frederic Weisbecker Signed-off-by: Tim Bird LKML-Reference: <4B87098B.4040308@am.sony.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 20 ++++++++++++++++++-- kernel/trace/trace.h | 3 ++- kernel/trace/trace_functions_graph.c | 25 +++++++++++++++++++++++-- 3 files changed, 43 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6efd5cb3c252..ababedb4e87f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -374,6 +374,21 @@ static int __init set_buf_size(char *str) } __setup("trace_buf_size=", set_buf_size); +static int __init set_tracing_thresh(char *str) +{ + unsigned long threshhold; + int ret; + + if (!str) + return 0; + ret = strict_strtoul(str, 0, &threshhold); + if (ret < 0) + return 0; + tracing_thresh = threshhold * 1000; + return 1; +} +__setup("tracing_thresh=", set_tracing_thresh); + unsigned long nsecs_to_usecs(unsigned long nsecs) { return nsecs / 1000; @@ -579,9 +594,10 @@ static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) static arch_spinlock_t ftrace_max_lock = (arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED; +unsigned long __read_mostly tracing_thresh; + #ifdef CONFIG_TRACER_MAX_TRACE unsigned long __read_mostly tracing_max_latency; -unsigned long __read_mostly tracing_thresh; /* * Copy the new maximum trace into the separate maximum-trace @@ -4248,10 +4264,10 @@ static __init int tracer_init_debugfs(void) #ifdef CONFIG_TRACER_MAX_TRACE trace_create_file("tracing_max_latency", 0644, d_tracer, &tracing_max_latency, &tracing_max_lat_fops); +#endif trace_create_file("tracing_thresh", 0644, d_tracer, &tracing_thresh, &tracing_max_lat_fops); -#endif trace_create_file("README", 0444, d_tracer, NULL, &tracing_readme_fops); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index fd05bcaf91b0..1bc8cd1431d7 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -396,9 +396,10 @@ extern int process_new_ksym_entry(char *ksymname, int op, unsigned long addr); extern unsigned long nsecs_to_usecs(unsigned long nsecs); +extern unsigned long tracing_thresh; + #ifdef CONFIG_TRACER_MAX_TRACE extern unsigned long tracing_max_latency; -extern unsigned long tracing_thresh; void update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu); void update_max_tr_single(struct trace_array *tr, diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 7b1f24618d97..e9df04b60267 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -237,6 +237,14 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) return ret; } +int trace_graph_thresh_entry(struct ftrace_graph_ent *trace) +{ + if (tracing_thresh) + return 1; + else + return trace_graph_entry(trace); +} + static void __trace_graph_return(struct trace_array *tr, struct ftrace_graph_ret *trace, unsigned long flags, @@ -290,13 +298,26 @@ void set_graph_array(struct trace_array *tr) smp_mb(); } +void trace_graph_thresh_return(struct ftrace_graph_ret *trace) +{ + if (tracing_thresh && + (trace->rettime - trace->calltime < tracing_thresh)) + return; + else + trace_graph_return(trace); +} + static int graph_trace_init(struct trace_array *tr) { int ret; set_graph_array(tr); - ret = register_ftrace_graph(&trace_graph_return, - &trace_graph_entry); + if (tracing_thresh) + ret = register_ftrace_graph(&trace_graph_thresh_return, + &trace_graph_thresh_entry); + else + ret = register_ftrace_graph(&trace_graph_return, + &trace_graph_entry); if (ret) return ret; tracing_start_cmdline_record(); -- cgit v1.2.2 From 984b3f5746ed2cde3d184651dabf26980f2b66e5 Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Fri, 5 Mar 2010 13:41:37 -0800 Subject: bitops: rename for_each_bit() to for_each_set_bit() Rename for_each_bit to for_each_set_bit in the kernel source tree. To permit for_each_clear_bit(), should that ever be added. The patch includes a macro to map the old for_each_bit() onto the new for_each_set_bit(). This is a (very) temporary thing to ease the migration. [akpm@linux-foundation.org: add temporary for_each_bit()] Suggested-by: Alexey Dobriyan Suggested-by: Andrew Morton Signed-off-by: Akinobu Mita Cc: "David S. Miller" Cc: Russell King Cc: David Woodhouse Cc: Artem Bityutskiy Cc: Stephen Rothwell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sched_cpupri.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index eeb3506c4834..82095bf2099f 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -47,7 +47,7 @@ static int convert_prio(int prio) } #define for_each_cpupri_active(array, idx) \ - for_each_bit(idx, array, CPUPRI_NR_PRIORITIES) + for_each_set_bit(idx, array, CPUPRI_NR_PRIORITIES) /** * cpupri_find - find the best (lowest-pri) CPU in the system -- cgit v1.2.2 From d559db086ff5be9bcc259e5aa50bf3d881eaf1d1 Mon Sep 17 00:00:00 2001 From: KAMEZAWA Hiroyuki Date: Fri, 5 Mar 2010 13:41:39 -0800 Subject: mm: clean up mm_counter Presently, per-mm statistics counter is defined by macro in sched.h This patch modifies it to - defined in mm.h as inlinf functions - use array instead of macro's name creation. This patch is for reducing patch size in future patch to modify implementation of per-mm counter. Signed-off-by: KAMEZAWA Hiroyuki Reviewed-by: Minchan Kim Cc: Christoph Lameter Cc: Lee Schermerhorn Cc: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 3 +-- kernel/tsacct.c | 1 + 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 17bbf093356d..7616bcf107b9 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -455,8 +455,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) (current->mm->flags & MMF_INIT_MASK) : default_dump_filter; mm->core_state = NULL; mm->nr_ptes = 0; - set_mm_counter(mm, file_rss, 0); - set_mm_counter(mm, anon_rss, 0); + memset(&mm->rss_stat, 0, sizeof(mm->rss_stat)); spin_lock_init(&mm->page_table_lock); mm->free_area_cache = TASK_UNMAPPED_BASE; mm->cached_hole_size = ~0UL; diff --git a/kernel/tsacct.c b/kernel/tsacct.c index 00d59d048edf..0a67e041edf8 100644 --- a/kernel/tsacct.c +++ b/kernel/tsacct.c @@ -21,6 +21,7 @@ #include #include #include +#include /* * fill in basic accounting fields -- cgit v1.2.2 From 34e55232e59f7b19050267a05ff1226e5cd122a5 Mon Sep 17 00:00:00 2001 From: KAMEZAWA Hiroyuki Date: Fri, 5 Mar 2010 13:41:40 -0800 Subject: mm: avoid false sharing of mm_counter Considering the nature of per mm stats, it's the shared object among threads and can be a cache-miss point in the page fault path. This patch adds per-thread cache for mm_counter. RSS value will be counted into a struct in task_struct and synchronized with mm's one at events. Now, in this patch, the event is the number of calls to handle_mm_fault. Per-thread value is added to mm at each 64 calls. rough estimation with small benchmark on parallel thread (2threads) shows [before] 4.5 cache-miss/faults [after] 4.0 cache-miss/faults Anyway, the most contended object is mmap_sem if the number of threads grows. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: KAMEZAWA Hiroyuki Cc: Minchan Kim Cc: Christoph Lameter Cc: Lee Schermerhorn Cc: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 45ed043b8bf5..10d3c5d5ae44 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -952,7 +952,8 @@ NORET_TYPE void do_exit(long code) preempt_count()); acct_update_integrals(tsk); - + /* sync mm's RSS info before statistics gathering */ + sync_mm_rss(tsk, tsk->mm); group_dead = atomic_dec_and_test(&tsk->signal->live); if (group_dead) { hrtimer_cancel(&tsk->signal->real_timer); -- cgit v1.2.2 From 5beb49305251e5669852ed541e8e2f2f7696c53e Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Fri, 5 Mar 2010 13:42:07 -0800 Subject: mm: change anon_vma linking to fix multi-process server scalability issue The old anon_vma code can lead to scalability issues with heavily forking workloads. Specifically, each anon_vma will be shared between the parent process and all its child processes. In a workload with 1000 child processes and a VMA with 1000 anonymous pages per process that get COWed, this leads to a system with a million anonymous pages in the same anon_vma, each of which is mapped in just one of the 1000 processes. However, the current rmap code needs to walk them all, leading to O(N) scanning complexity for each page. This can result in systems where one CPU is walking the page tables of 1000 processes in page_referenced_one, while all other CPUs are stuck on the anon_vma lock. This leads to catastrophic failure for a benchmark like AIM7, where the total number of processes can reach in the tens of thousands. Real workloads are still a factor 10 less process intensive than AIM7, but they are catching up. This patch changes the way anon_vmas and VMAs are linked, which allows us to associate multiple anon_vmas with a VMA. At fork time, each child process gets its own anon_vmas, in which its COWed pages will be instantiated. The parents' anon_vma is also linked to the VMA, because non-COWed pages could be present in any of the children. This reduces rmap scanning complexity to O(1) for the pages of the 1000 child processes, with O(N) complexity for at most 1/N pages in the system. This reduces the average scanning cost in heavily forking workloads from O(N) to 2. The only real complexity in this patch stems from the fact that linking a VMA to anon_vmas now involves memory allocations. This means vma_adjust can fail, if it needs to attach a VMA to anon_vma structures. This in turn means error handling needs to be added to the calling functions. A second source of complexity is that, because there can be multiple anon_vmas, the anon_vma linking in vma_adjust can no longer be done under "the" anon_vma lock. To prevent the rmap code from walking up an incomplete VMA, this patch introduces the VM_LOCK_RMAP VMA flag. This bit flag uses the same slot as the NOMMU VM_MAPPED_COPY, with an ifdef in mm.h to make sure it is impossible to compile a kernel that needs both symbolic values for the same bitflag. Some test results: Without the anon_vma changes, when AIM7 hits around 9.7k users (on a test box with 16GB RAM and not quite enough IO), the system ends up running >99% in system time, with every CPU on the same anon_vma lock in the pageout code. With these changes, AIM7 hits the cross-over point around 29.7k users. This happens with ~99% IO wait time, there never seems to be any spike in system time. The anon_vma lock contention appears to be resolved. [akpm@linux-foundation.org: cleanups] Signed-off-by: Rik van Riel Cc: KOSAKI Motohiro Cc: Larry Woodman Cc: Lee Schermerhorn Cc: Minchan Kim Cc: Andrea Arcangeli Cc: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 7616bcf107b9..bab7b254ad39 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -329,15 +329,17 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) if (!tmp) goto fail_nomem; *tmp = *mpnt; + INIT_LIST_HEAD(&tmp->anon_vma_chain); pol = mpol_dup(vma_policy(mpnt)); retval = PTR_ERR(pol); if (IS_ERR(pol)) goto fail_nomem_policy; vma_set_policy(tmp, pol); + if (anon_vma_fork(tmp, mpnt)) + goto fail_nomem_anon_vma_fork; tmp->vm_flags &= ~VM_LOCKED; tmp->vm_mm = mm; tmp->vm_next = NULL; - anon_vma_link(tmp); file = tmp->vm_file; if (file) { struct inode *inode = file->f_path.dentry->d_inode; @@ -392,6 +394,8 @@ out: flush_tlb_mm(oldmm); up_write(&oldmm->mmap_sem); return retval; +fail_nomem_anon_vma_fork: + mpol_put(pol); fail_nomem_policy: kmem_cache_free(vm_area_cachep, tmp); fail_nomem: -- cgit v1.2.2 From 452aa6999e6703ffbddd7f6ea124d3968915f3e3 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 5 Mar 2010 13:42:13 -0800 Subject: mm/pm: force GFP_NOIO during suspend/hibernation and resume There are quite a few GFP_KERNEL memory allocations made during suspend/hibernation and resume that may cause the system to hang, because the I/O operations they depend on cannot be completed due to the underlying devices being suspended. Avoid this problem by clearing the __GFP_IO and __GFP_FS bits in gfp_allowed_mask before suspend/hibernation and restoring the original values of these bits in gfp_allowed_mask durig the subsequent resume. [akpm@linux-foundation.org: fix CONFIG_PM=n linkage] Signed-off-by: Rafael J. Wysocki Reported-by: Maxim Levitsky Cc: Sebastian Ott Cc: Benjamin Herrenschmidt Cc: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/hibernate.c | 9 +++++++++ kernel/power/suspend.c | 3 +++ 2 files changed, 12 insertions(+) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index bbfe472d7524..da5288ec2392 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -323,6 +323,7 @@ static int create_image(int platform_mode) int hibernation_snapshot(int platform_mode) { int error; + gfp_t saved_mask; error = platform_begin(platform_mode); if (error) @@ -334,6 +335,7 @@ int hibernation_snapshot(int platform_mode) goto Close; suspend_console(); + saved_mask = clear_gfp_allowed_mask(GFP_IOFS); error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Recover_platform; @@ -351,6 +353,7 @@ int hibernation_snapshot(int platform_mode) dpm_resume_end(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); + set_gfp_allowed_mask(saved_mask); resume_console(); Close: platform_end(platform_mode); @@ -445,14 +448,17 @@ static int resume_target_kernel(bool platform_mode) int hibernation_restore(int platform_mode) { int error; + gfp_t saved_mask; pm_prepare_console(); suspend_console(); + saved_mask = clear_gfp_allowed_mask(GFP_IOFS); error = dpm_suspend_start(PMSG_QUIESCE); if (!error) { error = resume_target_kernel(platform_mode); dpm_resume_end(PMSG_RECOVER); } + set_gfp_allowed_mask(saved_mask); resume_console(); pm_restore_console(); return error; @@ -466,6 +472,7 @@ int hibernation_restore(int platform_mode) int hibernation_platform_enter(void) { int error; + gfp_t saved_mask; if (!hibernation_ops) return -ENOSYS; @@ -481,6 +488,7 @@ int hibernation_platform_enter(void) entering_platform_hibernation = true; suspend_console(); + saved_mask = clear_gfp_allowed_mask(GFP_IOFS); error = dpm_suspend_start(PMSG_HIBERNATE); if (error) { if (hibernation_ops->recover) @@ -518,6 +526,7 @@ int hibernation_platform_enter(void) Resume_devices: entering_platform_hibernation = false; dpm_resume_end(PMSG_RESTORE); + set_gfp_allowed_mask(saved_mask); resume_console(); Close: diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index 6f10dfc2d3e9..44cce10b582d 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -189,6 +189,7 @@ static int suspend_enter(suspend_state_t state) int suspend_devices_and_enter(suspend_state_t state) { int error; + gfp_t saved_mask; if (!suspend_ops) return -ENOSYS; @@ -199,6 +200,7 @@ int suspend_devices_and_enter(suspend_state_t state) goto Close; } suspend_console(); + saved_mask = clear_gfp_allowed_mask(GFP_IOFS); suspend_test_start(); error = dpm_suspend_start(PMSG_SUSPEND); if (error) { @@ -215,6 +217,7 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); dpm_resume_end(PMSG_RESUME); suspend_test_finish("resume devices"); + set_gfp_allowed_mask(saved_mask); resume_console(); Close: if (suspend_ops->end) -- cgit v1.2.2 From 87d5e0236d9d688fb575e9e12232764ac617617c Mon Sep 17 00:00:00 2001 From: Chen Gong Date: Fri, 5 Mar 2010 13:42:38 -0800 Subject: kernel/cpu.c: delete deprecated definition in cpu_up() Additional_cpus is only supported for IA64 now. X86_64 should not be included. Signed-off-by: Chen Gong Cc: Ingo Molnar Cc: Rusty Russell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 677f25376a38..f8cced2692b3 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -338,7 +338,7 @@ int __cpuinit cpu_up(unsigned int cpu) if (!cpu_possible(cpu)) { printk(KERN_ERR "can't online cpu %d because it is not " "configured as may-hotadd at boot time\n", cpu); -#if defined(CONFIG_IA64) || defined(CONFIG_X86_64) +#if defined(CONFIG_IA64) printk(KERN_ERR "please check additional_cpus= boot " "parameter\n"); #endif -- cgit v1.2.2 From 5f1664f92b2247111b7d37e454a050b76ac61b7f Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 5 Mar 2010 13:42:51 -0800 Subject: splice: comparing unsigned int < 0 "ret" needs to be signed or the error handling for splice_to_pipe() won't work correctly. Signed-off-by: Dan Carpenter Cc: Tom Zanussi Cc: Jens Axboe Cc: Lai Jiangshan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/relay.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/relay.c b/kernel/relay.c index c705a41b4ba3..3d97f2821611 100644 --- a/kernel/relay.c +++ b/kernel/relay.c @@ -1215,14 +1215,14 @@ static void relay_page_release(struct splice_pipe_desc *spd, unsigned int i) /* * subbuf_splice_actor - splice up to one subbuf's worth of data */ -static int subbuf_splice_actor(struct file *in, +static ssize_t subbuf_splice_actor(struct file *in, loff_t *ppos, struct pipe_inode_info *pipe, size_t len, unsigned int flags, int *nonpad_ret) { - unsigned int pidx, poff, total_len, subbuf_pages, nr_pages, ret; + unsigned int pidx, poff, total_len, subbuf_pages, nr_pages; struct rchan_buf *rbuf = in->private_data; unsigned int subbuf_size = rbuf->chan->subbuf_size; uint64_t pos = (uint64_t) *ppos; @@ -1241,6 +1241,7 @@ static int subbuf_splice_actor(struct file *in, .ops = &relay_pipe_buf_ops, .spd_release = relay_page_release, }; + ssize_t ret; if (rbuf->subbufs_produced == rbuf->subbufs_consumed) return 0; -- cgit v1.2.2 From 9c03c383563f147907f1a90cf16f1e190e2f4aae Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Fri, 5 Mar 2010 13:42:52 -0800 Subject: includecheck fix for kernel/params.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix the following 'make includecheck' warning: kernel/params.c: linux/string.h is included more than once. Signed-off-by: Jaswinder Singh Rajput Cc: André Goddard Rosa Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/params.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index cf1b69183127..8d95f5451b22 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -24,7 +24,6 @@ #include #include #include -#include #if 0 #define DEBUGP printk -- cgit v1.2.2 From f3abd4f9531becb71626bd206955d47d5ea54f06 Mon Sep 17 00:00:00 2001 From: Thiago Farina Date: Fri, 5 Mar 2010 13:42:52 -0800 Subject: kernel/exit.c: fix shadows sparse warning kernel/exit.c:1183:26: warning: symbol 'status' shadows an earlier one kernel/exit.c:1173:21: originally declared here Signed-off-by: Thiago Farina Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 10d3c5d5ae44..ce1e48c2d93d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -1189,7 +1189,7 @@ static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p) if (unlikely(wo->wo_flags & WNOWAIT)) { int exit_code = p->exit_code; - int why, status; + int why; get_task_struct(p); read_unlock(&tasklist_lock); -- cgit v1.2.2 From d4bb527438b4181cd3c564ae04dd344c381283a1 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Fri, 5 Mar 2010 13:42:53 -0800 Subject: posix-cpu-timers: cleanup rlimits usage Fetch rlimit (both hard and soft) values only once and work on them. It removes many accesses through sig structure and makes the code cleaner. Mostly a preparation for writable resource limits support. Signed-off-by: Jiri Slaby Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: john stultz Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/posix-cpu-timers.c | 32 +++++++++++++++++--------------- 1 file changed, 17 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 438ff4523513..dbb16bf15c45 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -982,6 +982,7 @@ static void check_thread_timers(struct task_struct *tsk, int maxfire; struct list_head *timers = tsk->cpu_timers; struct signal_struct *const sig = tsk->signal; + unsigned long soft; maxfire = 20; tsk->cputime_expires.prof_exp = cputime_zero; @@ -1030,9 +1031,9 @@ static void check_thread_timers(struct task_struct *tsk, /* * Check for the special case thread timers. */ - if (sig->rlim[RLIMIT_RTTIME].rlim_cur != RLIM_INFINITY) { + soft = sig->rlim[RLIMIT_RTTIME].rlim_cur; + if (soft != RLIM_INFINITY) { unsigned long hard = sig->rlim[RLIMIT_RTTIME].rlim_max; - unsigned long *soft = &sig->rlim[RLIMIT_RTTIME].rlim_cur; if (hard != RLIM_INFINITY && tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) { @@ -1043,14 +1044,13 @@ static void check_thread_timers(struct task_struct *tsk, __group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk); return; } - if (tsk->rt.timeout > DIV_ROUND_UP(*soft, USEC_PER_SEC/HZ)) { + if (tsk->rt.timeout > DIV_ROUND_UP(soft, USEC_PER_SEC/HZ)) { /* * At the soft limit, send a SIGXCPU every second. */ - if (sig->rlim[RLIMIT_RTTIME].rlim_cur - < sig->rlim[RLIMIT_RTTIME].rlim_max) { - sig->rlim[RLIMIT_RTTIME].rlim_cur += - USEC_PER_SEC; + if (soft < hard) { + soft += USEC_PER_SEC; + sig->rlim[RLIMIT_RTTIME].rlim_cur = soft; } printk(KERN_INFO "RT Watchdog Timeout: %s[%d]\n", @@ -1121,6 +1121,7 @@ static void check_process_timers(struct task_struct *tsk, unsigned long long sum_sched_runtime, sched_expires; struct list_head *timers = sig->cpu_timers; struct task_cputime cputime; + unsigned long soft; /* * Don't sample the current process CPU clocks if there are no timers. @@ -1193,11 +1194,12 @@ static void check_process_timers(struct task_struct *tsk, SIGPROF); check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_expires, utime, SIGVTALRM); - - if (sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY) { + soft = sig->rlim[RLIMIT_CPU].rlim_cur; + if (soft != RLIM_INFINITY) { unsigned long psecs = cputime_to_secs(ptime); + unsigned long hard = sig->rlim[RLIMIT_CPU].rlim_max; cputime_t x; - if (psecs >= sig->rlim[RLIMIT_CPU].rlim_max) { + if (psecs >= hard) { /* * At the hard limit, we just die. * No need to calculate anything else now. @@ -1205,17 +1207,17 @@ static void check_process_timers(struct task_struct *tsk, __group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk); return; } - if (psecs >= sig->rlim[RLIMIT_CPU].rlim_cur) { + if (psecs >= soft) { /* * At the soft limit, send a SIGXCPU every second. */ __group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk); - if (sig->rlim[RLIMIT_CPU].rlim_cur - < sig->rlim[RLIMIT_CPU].rlim_max) { - sig->rlim[RLIMIT_CPU].rlim_cur++; + if (soft < hard) { + soft++; + sig->rlim[RLIMIT_CPU].rlim_cur = soft; } } - x = secs_to_cputime(sig->rlim[RLIMIT_CPU].rlim_cur); + x = secs_to_cputime(soft); if (cputime_eq(prof_expires, cputime_zero) || cputime_lt(x, prof_expires)) { prof_expires = x; -- cgit v1.2.2 From 78d7d407b62a021e6d2e8dc24c0b90e390ab58a1 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Fri, 5 Mar 2010 13:42:54 -0800 Subject: kernel core: use helpers for rlimits Make sure compiler won't do weird things with limits. E.g. fetching them twice may return 2 different values after writable limits are implemented. I.e. either use rlimit helpers added in commit 3e10e716abf3 ("resource: add helpers for fetching rlimits") or ACCESS_ONCE if not applicable. Signed-off-by: Jiri Slaby Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: john stultz Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 10 ++++++---- kernel/perf_event.c | 2 +- kernel/posix-cpu-timers.c | 10 ++++++---- kernel/sched.c | 4 ++-- kernel/sched_rt.c | 5 +++-- kernel/signal.c | 2 +- kernel/sys.c | 3 +-- 7 files changed, 20 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index bab7b254ad39..b0ec34abc0bb 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -828,6 +828,8 @@ void __cleanup_sighand(struct sighand_struct *sighand) */ static void posix_cpu_timers_init_group(struct signal_struct *sig) { + unsigned long cpu_limit; + /* Thread group counters. */ thread_group_cputime_init(sig); @@ -842,9 +844,9 @@ static void posix_cpu_timers_init_group(struct signal_struct *sig) sig->cputime_expires.virt_exp = cputime_zero; sig->cputime_expires.sched_exp = 0; - if (sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY) { - sig->cputime_expires.prof_exp = - secs_to_cputime(sig->rlim[RLIMIT_CPU].rlim_cur); + cpu_limit = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur); + if (cpu_limit != RLIM_INFINITY) { + sig->cputime_expires.prof_exp = secs_to_cputime(cpu_limit); sig->cputimer.running = 1; } @@ -1037,7 +1039,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, #endif retval = -EAGAIN; if (atomic_read(&p->real_cred->user->processes) >= - p->signal->rlim[RLIMIT_NPROC].rlim_cur) { + task_rlimit(p, RLIMIT_NPROC)) { if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && p->real_cred->user != INIT_USER) goto bad_fork_free; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index a661e7991865..8e352c756ba7 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2610,7 +2610,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) if (user_locked > user_lock_limit) extra = user_locked - user_lock_limit; - lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; + lock_limit = rlimit(RLIMIT_MEMLOCK); lock_limit >>= PAGE_SHIFT; locked = vma->vm_mm->locked_vm + extra; diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index dbb16bf15c45..1a22dfd42df9 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -1031,9 +1031,10 @@ static void check_thread_timers(struct task_struct *tsk, /* * Check for the special case thread timers. */ - soft = sig->rlim[RLIMIT_RTTIME].rlim_cur; + soft = ACCESS_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_cur); if (soft != RLIM_INFINITY) { - unsigned long hard = sig->rlim[RLIMIT_RTTIME].rlim_max; + unsigned long hard = + ACCESS_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_max); if (hard != RLIM_INFINITY && tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) { @@ -1194,10 +1195,11 @@ static void check_process_timers(struct task_struct *tsk, SIGPROF); check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_expires, utime, SIGVTALRM); - soft = sig->rlim[RLIMIT_CPU].rlim_cur; + soft = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur); if (soft != RLIM_INFINITY) { unsigned long psecs = cputime_to_secs(ptime); - unsigned long hard = sig->rlim[RLIMIT_CPU].rlim_max; + unsigned long hard = + ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_max); cputime_t x; if (psecs >= hard) { /* diff --git a/kernel/sched.c b/kernel/sched.c index abb36b16b93b..b47ceeec1a91 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4353,7 +4353,7 @@ int can_nice(const struct task_struct *p, const int nice) /* convert nice value [19,-20] to rlimit style value [1,40] */ int nice_rlim = 20 - nice; - return (nice_rlim <= p->signal->rlim[RLIMIT_NICE].rlim_cur || + return (nice_rlim <= task_rlimit(p, RLIMIT_NICE) || capable(CAP_SYS_NICE)); } @@ -4530,7 +4530,7 @@ recheck: if (!lock_task_sighand(p, &flags)) return -ESRCH; - rlim_rtprio = p->signal->rlim[RLIMIT_RTPRIO].rlim_cur; + rlim_rtprio = task_rlimit(p, RLIMIT_RTPRIO); unlock_task_sighand(p, &flags); /* can't set/change the rt policy */ diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index bf3e38fdbe6d..5a6ed1f0990a 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1662,8 +1662,9 @@ static void watchdog(struct rq *rq, struct task_struct *p) if (!p->signal) return; - soft = p->signal->rlim[RLIMIT_RTTIME].rlim_cur; - hard = p->signal->rlim[RLIMIT_RTTIME].rlim_max; + /* max may change after cur was read, this will be fixed next tick */ + soft = task_rlimit(p, RLIMIT_RTTIME); + hard = task_rlimit_max(p, RLIMIT_RTTIME); if (soft != RLIM_INFINITY) { unsigned long next; diff --git a/kernel/signal.c b/kernel/signal.c index 5bb9baffa4f1..dbd7fe073c55 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -245,7 +245,7 @@ __sigqueue_alloc(int sig, struct task_struct *t, gfp_t flags, int override_rlimi if (override_rlimit || atomic_read(&user->sigpending) <= - t->signal->rlim[RLIMIT_SIGPENDING].rlim_cur) { + task_rlimit(t, RLIMIT_SIGPENDING)) { q = kmem_cache_alloc(sigqueue_cachep, flags); } else { print_dropped_signal(sig); diff --git a/kernel/sys.c b/kernel/sys.c index 877fe4f8e05e..9814e43fb23b 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -571,8 +571,7 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; - if (atomic_read(&new_user->processes) >= - current->signal->rlim[RLIMIT_NPROC].rlim_cur && + if (atomic_read(&new_user->processes) >= rlimit(RLIMIT_NPROC) && new_user != INIT_USER) { free_uid(new_user); return -EAGAIN; -- cgit v1.2.2 From 8aeee85a29e27e043db582bf2ae8e5f42767934f Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Fri, 5 Mar 2010 13:42:55 -0800 Subject: panic: fix panic_timeout accuracy when running on a hypervisor I've had some complaints about panic_timeout being wildly innacurate on shared processor PowerPC partitions (a 3 minute panic_timeout taking 30 minutes). The problem is we loop on mdelay(1) and with a 1ms in 10ms hypervisor timeslice each of these will take 10ms (ie 10x) longer. I expect other platforms with shared processor hypervisors will see the same issue. This patch keeps the old behaviour if we have a panic_blink (only keyboard LEDs right now) and does 1 second mdelays if we don't. Signed-off-by: Anton Blanchard Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/panic.c | 46 ++++++++++++++++++++++++++++++---------------- 1 file changed, 30 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index c787333282b8..13d966b4c14a 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -36,15 +36,36 @@ ATOMIC_NOTIFIER_HEAD(panic_notifier_list); EXPORT_SYMBOL(panic_notifier_list); -static long no_blink(long time) -{ - return 0; -} - /* Returns how long it waited in ms */ long (*panic_blink)(long time); EXPORT_SYMBOL(panic_blink); +static void panic_blink_one_second(void) +{ + static long i = 0, end; + + if (panic_blink) { + end = i + MSEC_PER_SEC; + + while (i < end) { + i += panic_blink(i); + mdelay(1); + i++; + } + } else { + /* + * When running under a hypervisor a small mdelay may get + * rounded up to the hypervisor timeslice. For example, with + * a 1ms in 10ms hypervisor timeslice we might inflate a + * mdelay(1) loop by 10x. + * + * If we have nothing to blink, spin on 1 second calls to + * mdelay to avoid this. + */ + mdelay(MSEC_PER_SEC); + } +} + /** * panic - halt the system * @fmt: The text string to print @@ -95,9 +116,6 @@ NORET_TYPE void panic(const char * fmt, ...) bust_spinlocks(0); - if (!panic_blink) - panic_blink = no_blink; - if (panic_timeout > 0) { /* * Delay timeout seconds before rebooting the machine. @@ -105,11 +123,9 @@ NORET_TYPE void panic(const char * fmt, ...) */ printk(KERN_EMERG "Rebooting in %d seconds..", panic_timeout); - for (i = 0; i < panic_timeout*1000; ) { + for (i = 0; i < panic_timeout; i++) { touch_nmi_watchdog(); - i += panic_blink(i); - mdelay(1); - i++; + panic_blink_one_second(); } /* * This will not be a clean reboot, with everything @@ -135,11 +151,9 @@ NORET_TYPE void panic(const char * fmt, ...) } #endif local_irq_enable(); - for (i = 0; ; ) { + while (1) { touch_softlockup_watchdog(); - i += panic_blink(i); - mdelay(1); - i++; + panic_blink_one_second(); } } -- cgit v1.2.2 From 9728e5d6e6c432ee8487c63ce6e479e2474d9945 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Fri, 5 Mar 2010 13:42:56 -0800 Subject: kernel/pid.c: update comment on find_task_by_pid_ns tasklist_lock does protect the task and its pid, it can't go away. The problem is that find_pid_ns() itself is unsafe without rcu lock, it can race with copy_process()->free_pid(any_pid). Protecting copy_process()->free_pid(any_pid) with tasklist_lock would make it possible to call find_task_by_pid_ns() under tasklist safely, but we don't do so because we are trying to get rid of the read_lock sites of tasklist_lock. Signed-off-by: Tetsuo Handa Cc: Oleg Nesterov Cc: "Paul E. McKenney" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index b08e697cd83f..86b296943e5f 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -376,7 +376,7 @@ struct task_struct *pid_task(struct pid *pid, enum pid_type type) EXPORT_SYMBOL(pid_task); /* - * Must be called under rcu_read_lock() or with tasklist_lock read-held. + * Must be called under rcu_read_lock(). */ struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns) { -- cgit v1.2.2 From cea83886dde49fd7524e9f4a246dd5dff4ad236a Mon Sep 17 00:00:00 2001 From: "Gustavo F. Padovan" Date: Fri, 5 Mar 2010 13:42:58 -0800 Subject: printk: avoid warning when CONFIG_PRINTK is disabled kernel/printk.c:72: warning: `saved_console_loglevel' defined but not used Signed-off-by: Gustavo F. Padovan Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 40674122ecf2..75077ad0b537 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -70,8 +70,6 @@ int console_printk[4] = { DEFAULT_CONSOLE_LOGLEVEL, /* default_console_loglevel */ }; -static int saved_console_loglevel = -1; - /* * Low level drivers may need that to know if they can schedule in * their unblank() callback or not. So let's export it. @@ -146,6 +144,7 @@ static char __log_buf[__LOG_BUF_LEN]; static char *log_buf = __log_buf; static int log_buf_len = __LOG_BUF_LEN; static unsigned logged_chars; /* Number of chars produced since last read+clear operation */ +static int saved_console_loglevel = -1; #ifdef CONFIG_KEXEC /* -- cgit v1.2.2 From 1fcccbac89f5bbc5e41aa72086960059fce372da Mon Sep 17 00:00:00 2001 From: Daisuke HATAYAMA Date: Fri, 5 Mar 2010 13:44:07 -0800 Subject: elf coredump: replace ELF_CORE_EXTRA_* macros by functions elf_core_dump() and elf_fdpic_core_dump() use #ifdef and the corresponding macro for hiding _multiline_ logics in functions. This patch removes #ifdef and replaces ELF_CORE_EXTRA_* by corresponding functions. For architectures not implemeonting ELF_CORE_EXTRA_*, we use weak functions in order to reduce a range of modification. This cleanup is for my next patches, but I think this cleanup itself is worth doing regardless of my firnal purpose. Signed-off-by: Daisuke HATAYAMA Cc: "Luck, Tony" Cc: Jeff Dike Cc: David Howells Cc: Greg Ungerer Cc: Roland McGrath Cc: Oleg Nesterov Cc: Ingo Molnar Cc: Alexander Viro Cc: Andi Kleen Cc: Alan Cox Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/Makefile | 3 +++ kernel/elfcore.c | 23 +++++++++++++++++++++++ 2 files changed, 26 insertions(+) create mode 100644 kernel/elfcore.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 7b974699f8c2..a987aa1676b5 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -91,6 +91,9 @@ obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o obj-$(CONFIG_TASKSTATS) += taskstats.o tsacct.o obj-$(CONFIG_TRACEPOINTS) += tracepoint.o obj-$(CONFIG_LATENCYTOP) += latencytop.o +obj-$(CONFIG_BINFMT_ELF) += elfcore.o +obj-$(CONFIG_COMPAT_BINFMT_ELF) += elfcore.o +obj-$(CONFIG_BINFMT_ELF_FDPIC) += elfcore.o obj-$(CONFIG_FUNCTION_TRACER) += trace/ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_X86_DS) += trace/ diff --git a/kernel/elfcore.c b/kernel/elfcore.c new file mode 100644 index 000000000000..5445741f4b4c --- /dev/null +++ b/kernel/elfcore.c @@ -0,0 +1,23 @@ +#include +#include +#include + +#include + + +Elf_Half __weak elf_core_extra_phdrs(void) +{ + return 0; +} + +int __weak elf_core_write_extra_phdrs(struct file *file, loff_t offset, size_t *size, + unsigned long limit) +{ + return 1; +} + +int __weak elf_core_write_extra_data(struct file *file, size_t *size, + unsigned long limit) +{ + return 1; +} -- cgit v1.2.2 From 8d9032bbe4671dc481261ccd4e161cd96e54b118 Mon Sep 17 00:00:00 2001 From: Daisuke HATAYAMA Date: Fri, 5 Mar 2010 13:44:10 -0800 Subject: elf coredump: add extended numbering support The current ELF dumper implementation can produce broken corefiles if program headers exceed 65535. This number is determined by the number of vmas which the process have. In particular, some extreme programs may use more than 65535 vmas. (If you google max_map_count, you can find some users facing this problem.) This kind of program never be able to generate correct coredumps. This patch implements ``extended numbering'' that uses sh_info field of the first section header instead of e_phnum field in order to represent upto 4294967295 vmas. This is supported by AMD64-ABI(http://www.x86-64.org/documentation.html) and Solaris(http://docs.sun.com/app/docs/doc/817-1984/). Of course, we are preparing patches for gdb and binutils. Signed-off-by: Daisuke HATAYAMA Cc: "Luck, Tony" Cc: Jeff Dike Cc: David Howells Cc: Greg Ungerer Cc: Roland McGrath Cc: Oleg Nesterov Cc: Ingo Molnar Cc: Alexander Viro Cc: Andi Kleen Cc: Alan Cox Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/elfcore.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/elfcore.c b/kernel/elfcore.c index 5445741f4b4c..ff915efef66d 100644 --- a/kernel/elfcore.c +++ b/kernel/elfcore.c @@ -21,3 +21,8 @@ int __weak elf_core_write_extra_data(struct file *file, size_t *size, { return 1; } + +size_t __weak elf_core_extra_data_size(void) +{ + return 0; +} -- cgit v1.2.2 From c9be0a36f9bf392a7984473124a67a12964df11f Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Tue, 5 Jan 2010 12:47:58 +0100 Subject: sysdev: Pass attribute in sysdev_class attributes show/store Passing the attribute to the low level IO functions allows all kinds of cleanups, by sharing low level IO code without requiring an own function for every piece of data. Also drivers can extend the attributes with own data fields and use that in the low level function. Similar to sysdev_attributes and normal attributes. This is a tree-wide sweep, converting everything in one go. No functional changes in this patch other than passing the new argument everywhere. Tested on x86, the non x86 parts are uncompiled. Signed-off-by: Andi Kleen Signed-off-by: Greg Kroah-Hartman --- kernel/perf_event.c | 13 ++++++++++--- kernel/sched.c | 4 ++++ 2 files changed, 14 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 8e352c756ba7..f40560b86544 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5481,13 +5481,16 @@ void __init perf_event_init(void) register_cpu_notifier(&perf_cpu_nb); } -static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, char *buf) +static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%d\n", perf_reserved_percpu); } static ssize_t perf_set_reserve_percpu(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { @@ -5516,13 +5519,17 @@ perf_set_reserve_percpu(struct sysdev_class *class, return count; } -static ssize_t perf_show_overcommit(struct sysdev_class *class, char *buf) +static ssize_t perf_show_overcommit(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + char *buf) { return sprintf(buf, "%d\n", perf_overcommit); } static ssize_t -perf_set_overcommit(struct sysdev_class *class, const char *buf, size_t count) +perf_set_overcommit(struct sysdev_class *class, + struct sysdev_class_attribute *attr, + const char *buf, size_t count) { unsigned long val; int err; diff --git a/kernel/sched.c b/kernel/sched.c index b47ceeec1a91..150b6988de49 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7406,11 +7406,13 @@ static ssize_t sched_power_savings_store(const char *buf, size_t count, int smt) #ifdef CONFIG_SCHED_MC static ssize_t sched_mc_power_savings_show(struct sysdev_class *class, + struct sysdev_class_attribute *attr, char *page) { return sprintf(page, "%u\n", sched_mc_power_savings); } static ssize_t sched_mc_power_savings_store(struct sysdev_class *class, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { return sched_power_savings_store(buf, count, 0); @@ -7422,11 +7424,13 @@ static SYSDEV_CLASS_ATTR(sched_mc_power_savings, 0644, #ifdef CONFIG_SCHED_SMT static ssize_t sched_smt_power_savings_show(struct sysdev_class *dev, + struct sysdev_class_attribute *attr, char *page) { return sprintf(page, "%u\n", sched_smt_power_savings); } static ssize_t sched_smt_power_savings_store(struct sysdev_class *dev, + struct sysdev_class_attribute *attr, const char *buf, size_t count) { return sched_power_savings_store(buf, count, 1); -- cgit v1.2.2 From 9cd43611ccfb46632bfa7d19f688924ea93f1613 Mon Sep 17 00:00:00 2001 From: Emese Revfy Date: Thu, 31 Dec 2009 14:52:51 +0100 Subject: kobject: Constify struct kset_uevent_ops Constify struct kset_uevent_ops. This is part of the ops structure constification effort started by Arjan van de Ven et al. Benefits of this constification: * prevents modification of data that is shared (referenced) by many other structure instances at runtime * detects/prevents accidental (but not intentional) modification attempts on archs that enforce read-only kernel data at runtime * potentially better optimized code as the compiler can assume that the const data cannot be changed * the compiler/linker move const data into .rodata and therefore exclude them from false sharing Signed-off-by: Emese Revfy Signed-off-by: Greg Kroah-Hartman --- kernel/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 8d95f5451b22..48370be3c0a1 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -736,7 +736,7 @@ static int uevent_filter(struct kset *kset, struct kobject *kobj) return 0; } -static struct kset_uevent_ops module_uevent_ops = { +static const struct kset_uevent_ops module_uevent_ops = { .filter = uevent_filter, }; -- cgit v1.2.2 From 52cf25d0ab7f78eeecc59ac652ed5090f69b619e Mon Sep 17 00:00:00 2001 From: Emese Revfy Date: Tue, 19 Jan 2010 02:58:23 +0100 Subject: Driver core: Constify struct sysfs_ops in struct kobj_type Constify struct sysfs_ops. This is part of the ops structure constification effort started by Arjan van de Ven et al. Benefits of this constification: * prevents modification of data that is shared (referenced) by many other structure instances at runtime * detects/prevents accidental (but not intentional) modification attempts on archs that enforce read-only kernel data at runtime * potentially better optimized code as the compiler can assume that the const data cannot be changed * the compiler/linker move const data into .rodata and therefore exclude them from false sharing Signed-off-by: Emese Revfy Acked-by: David Teigland Acked-by: Matt Domsch Acked-by: Maciej Sosnowski Acked-by: Hans J. Koch Acked-by: Pekka Enberg Acked-by: Jens Axboe Acked-by: Stephen Hemminger Signed-off-by: Greg Kroah-Hartman --- kernel/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 48370be3c0a1..68396d73c838 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -722,7 +722,7 @@ static ssize_t module_attr_store(struct kobject *kobj, return ret; } -static struct sysfs_ops module_sysfs_ops = { +static const struct sysfs_ops module_sysfs_ops = { .show = module_attr_show, .store = module_attr_store, }; -- cgit v1.2.2 From a07e4156a2ee6359d31a44946d7ee7f85dbf6bca Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 11 Feb 2010 15:23:05 -0800 Subject: sysfs: Use sysfs_attr_init and sysfs_bin_attr_init on dynamic attributes These are the non-static sysfs attributes that exist on my test machine. Fix them to use sysfs_attr_init or sysfs_bin_attr_init as appropriate. It simply requires making a sysfs attribute present to see this. So this is a little bit tedious but otherwise not too bad. Signed-off-by: Eric W. Biederman Acked-by: WANG Cong Cc: Tejun Heo Signed-off-by: Greg Kroah-Hartman --- kernel/params.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 68396d73c838..d55a53ec9234 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -516,6 +516,7 @@ static __modinit int add_sysfs_param(struct module_kobject *mk, new->grp.attrs = attrs; /* Tack new one on the end. */ + sysfs_attr_init(&new->attrs[num].mattr.attr); new->attrs[num].param = kp; new->attrs[num].mattr.show = param_attr_show; new->attrs[num].mattr.store = param_attr_store; -- cgit v1.2.2 From 361795b1eb7c08e9e65a2ebb4a4e536294d378a2 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 12 Feb 2010 13:41:56 -0800 Subject: sysfs: Use sysfs_attr_init and sysfs_bin_attr_init on module dynamic attributes A little more whack-a-mole annotating the dynamic sysfs attributes. I had everything built into my earlier test kernel, and so I missed these. Signed-off-by: Eric W. Biederman Signed-off-by: Greg Kroah-Hartman --- kernel/module.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index e5538d5f00ad..c968d3606dca 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -1085,6 +1085,7 @@ static void add_sect_attrs(struct module *mod, unsigned int nsect, if (sattr->name == NULL) goto out; sect_attrs->nsections++; + sysfs_attr_init(&sattr->mattr.attr); sattr->mattr.show = module_sect_show; sattr->mattr.store = NULL; sattr->mattr.attr.name = sattr->name; @@ -1180,6 +1181,7 @@ static void add_notes_attrs(struct module *mod, unsigned int nsect, if (sect_empty(&sechdrs[i])) continue; if (sechdrs[i].sh_type == SHT_NOTE) { + sysfs_bin_attr_init(nattr); nattr->attr.name = mod->sect_attrs->attrs[loaded].name; nattr->attr.mode = S_IRUGO; nattr->size = sechdrs[i].sh_size; @@ -1252,6 +1254,7 @@ int module_add_modinfo_attrs(struct module *mod) if (!attr->test || (attr->test && attr->test(mod))) { memcpy(temp_attr, attr, sizeof(*temp_attr)); + sysfs_attr_init(&temp_attr->attr); error = sysfs_create_file(&mod->mkobj.kobj,&temp_attr->attr); ++temp_attr; } -- cgit v1.2.2 From dc1d628a67a8f042e711ea5accc0beedc3ef0092 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 3 Mar 2010 15:55:04 +0100 Subject: perf: Provide generic perf_sample_data initialization This makes it easier to extend perf_sample_data and fixes a bug on arm and sparc, which failed to set ->raw to NULL, which can cause crashes when combined with PERF_SAMPLE_RAW. It also optimizes PowerPC and tracepoint, because the struct initialization is forced to zero out the whole structure. Signed-off-by: Peter Zijlstra Acked-by: Jean Pihet Reviewed-by: Frederic Weisbecker Acked-by: David S. Miller Cc: Jamie Iles Cc: Paul Mackerras Cc: Stephane Eranian Cc: stable@kernel.org LKML-Reference: <20100304140100.315416040@chello.nl> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 21 ++++++++------------- 1 file changed, 8 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index e68745053013..4393b9e73740 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4108,8 +4108,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi, if (rctx < 0) return; - data.addr = addr; - data.raw = NULL; + perf_sample_data_init(&data, addr); do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs); @@ -4154,11 +4153,10 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) struct perf_event *event; u64 period; - event = container_of(hrtimer, struct perf_event, hw.hrtimer); + event = container_of(hrtimer, struct perf_event, hw.hrtimer); event->pmu->read(event); - data.addr = 0; - data.raw = NULL; + perf_sample_data_init(&data, 0); data.period = event->hw.last_period; regs = get_irq_regs(); /* @@ -4322,17 +4320,15 @@ static const struct pmu perf_ops_task_clock = { void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size) { + struct pt_regs *regs = get_irq_regs(); + struct perf_sample_data data; struct perf_raw_record raw = { .size = entry_size, .data = record, }; - struct perf_sample_data data = { - .addr = addr, - .raw = &raw, - }; - - struct pt_regs *regs = get_irq_regs(); + perf_sample_data_init(&data, addr); + data.raw = &raw; if (!regs) regs = task_pt_regs(current); @@ -4448,8 +4444,7 @@ void perf_bp_event(struct perf_event *bp, void *data) struct perf_sample_data sample; struct pt_regs *regs = data; - sample.raw = NULL; - sample.addr = bp->attr.bp_addr; + perf_sample_data_init(&sample, bp->attr.bp_addr); if (!perf_exclude_event(bp, regs)) perf_swevent_add(bp, 1, 1, &sample, regs); -- cgit v1.2.2 From 3f6da3905398826d85731247e7fbcf53400c18bd Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 5 Mar 2010 13:01:18 +0100 Subject: perf: Rework and fix the arch CPU-hotplug hooks Remove the hw_perf_event_*() hotplug hooks in favour of per PMU hotplug notifiers. This has the advantage of reducing the static weak interface as well as exposing all hotplug actions to the PMU. Use this to fix x86 hotplug usage where we did things in ONLINE which should have been done in UP_PREPARE or STARTING. Signed-off-by: Peter Zijlstra Cc: Paul Mundt Cc: paulus@samba.org Cc: eranian@google.com Cc: robert.richter@amd.com Cc: fweisbec@gmail.com Cc: Arnaldo Carvalho de Melo LKML-Reference: <20100305154128.736225361@chello.nl> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 4393b9e73740..73329dedb5ad 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -81,10 +81,6 @@ extern __weak const struct pmu *hw_perf_event_init(struct perf_event *event) void __weak hw_perf_disable(void) { barrier(); } void __weak hw_perf_enable(void) { barrier(); } -void __weak hw_perf_event_setup(int cpu) { barrier(); } -void __weak hw_perf_event_setup_online(int cpu) { barrier(); } -void __weak hw_perf_event_setup_offline(int cpu) { barrier(); } - int __weak hw_perf_group_sched_in(struct perf_event *group_leader, struct perf_cpu_context *cpuctx, @@ -5382,8 +5378,6 @@ static void __cpuinit perf_event_init_cpu(int cpu) spin_lock(&perf_resource_lock); cpuctx->max_pertask = perf_max_events - perf_reserved_percpu; spin_unlock(&perf_resource_lock); - - hw_perf_event_setup(cpu); } #ifdef CONFIG_HOTPLUG_CPU @@ -5423,20 +5417,11 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) perf_event_init_cpu(cpu); break; - case CPU_ONLINE: - case CPU_ONLINE_FROZEN: - hw_perf_event_setup_online(cpu); - break; - case CPU_DOWN_PREPARE: case CPU_DOWN_PREPARE_FROZEN: perf_event_exit_cpu(cpu); break; - case CPU_DEAD: - hw_perf_event_setup_offline(cpu); - break; - default: break; } -- cgit v1.2.2 From 32975a4f114be52286f9a5bf6c230dbb8c0e1903 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sat, 6 Mar 2010 19:49:19 +0100 Subject: perf: Optimize perf_disable Currently we always call hw_perf_disable(), even if its already disabled, this seems superflous, esp. since it cannot be made NMI safe (see further patches). Signed-off-by: Peter Zijlstra Cc: paulus@samba.org Cc: eranian@google.com Cc: robert.richter@amd.com Cc: fweisbec@gmail.com Cc: Arnaldo Carvalho de Melo LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 73329dedb5ad..d8108465397d 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -93,25 +93,15 @@ void __weak perf_event_print_debug(void) { } static DEFINE_PER_CPU(int, perf_disable_count); -void __perf_disable(void) -{ - __get_cpu_var(perf_disable_count)++; -} - -bool __perf_enable(void) -{ - return !--__get_cpu_var(perf_disable_count); -} - void perf_disable(void) { - __perf_disable(); - hw_perf_disable(); + if (!__get_cpu_var(perf_disable_count)++) + hw_perf_disable(); } void perf_enable(void) { - if (__perf_enable()) + if (!--__get_cpu_var(perf_disable_count)) hw_perf_enable(); } -- cgit v1.2.2 From d4944a06666054707d23e11888e480af239e5abf Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 8 Mar 2010 13:51:20 +0100 Subject: perf: Provide better condition for event rotation Try to avoid useless rotation and PMU disables. [ Could be improved by keeping a nr_runnable count to better account for the < PERF_STAT_INACTIVE counters ] Signed-off-by: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: paulus@samba.org Cc: eranian@google.com Cc: robert.richter@amd.com Cc: fweisbec@gmail.com LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 21 +++++++++++++++------ 1 file changed, 15 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index d8108465397d..52c69a34d697 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1524,12 +1524,15 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) */ if (interrupts == MAX_INTERRUPTS) { perf_log_throttle(event, 1); + perf_disable(); event->pmu->unthrottle(event); + perf_enable(); } if (!event->attr.freq || !event->attr.sample_freq) continue; + perf_disable(); event->pmu->read(event); now = atomic64_read(&event->count); delta = now - hwc->freq_count_stamp; @@ -1537,6 +1540,7 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) if (delta > 0) perf_adjust_period(event, TICK_NSEC, delta); + perf_enable(); } raw_spin_unlock(&ctx->lock); } @@ -1546,9 +1550,6 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) */ static void rotate_ctx(struct perf_event_context *ctx) { - if (!ctx->nr_events) - return; - raw_spin_lock(&ctx->lock); /* Rotate the first entry last of non-pinned groups */ @@ -1561,19 +1562,28 @@ void perf_event_task_tick(struct task_struct *curr) { struct perf_cpu_context *cpuctx; struct perf_event_context *ctx; + int rotate = 0; if (!atomic_read(&nr_events)) return; cpuctx = &__get_cpu_var(perf_cpu_context); - ctx = curr->perf_event_ctxp; + if (cpuctx->ctx.nr_events && + cpuctx->ctx.nr_events != cpuctx->ctx.nr_active) + rotate = 1; - perf_disable(); + ctx = curr->perf_event_ctxp; + if (ctx && ctx->nr_events && ctx->nr_events != ctx->nr_active) + rotate = 1; perf_ctx_adjust_freq(&cpuctx->ctx); if (ctx) perf_ctx_adjust_freq(ctx); + if (!rotate) + return; + + perf_disable(); cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); if (ctx) task_ctx_sched_out(ctx, EVENT_FLEXIBLE); @@ -1585,7 +1595,6 @@ void perf_event_task_tick(struct task_struct *curr) cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); if (ctx) task_ctx_sched_in(curr, EVENT_FLEXIBLE); - perf_enable(); } -- cgit v1.2.2 From db2c4c7791cd04512093d05afc693c3511a65fd7 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 2 Feb 2010 23:34:40 +0100 Subject: lockdep: Move lock events under lockdep recursion protection There are rcu locked read side areas in the path where we submit a trace event. And these rcu_read_(un)lock() trigger lock events, which create recursive events. One pair in do_perf_sw_event: __lock_acquire | |--96.11%-- lock_acquire | | | |--27.21%-- do_perf_sw_event | | perf_tp_event | | | | | |--49.62%-- ftrace_profile_lock_release | | | lock_release | | | | | | | |--33.85%-- _raw_spin_unlock Another pair in perf_output_begin/end: __lock_acquire |--23.40%-- perf_output_begin | | __perf_event_overflow | | perf_swevent_overflow | | perf_swevent_add | | perf_swevent_ctx_event | | do_perf_sw_event | | perf_tp_event | | | | | |--55.37%-- ftrace_profile_lock_acquire | | | lock_acquire | | | | | | | |--37.31%-- _raw_spin_lock The problem is not that much the trace recursion itself, as we have a recursion protection already (though it's always wasteful to recurse). But the trace events are outside the lockdep recursion protection, then each lockdep event triggers a lock trace, which will trigger two other lockdep events. Here the recursive lock trace event won't be taken because of the trace recursion, so the recursion stops there but lockdep will still analyse these new events: To sum up, for each lockdep events we have: lock_*() | trace lock_acquire | ----- rcu_read_lock() | | | lock_acquire() | | | trace_lock_acquire() (stopped) | | | lockdep analyze | ----- rcu_read_unlock() | lock_release | trace_lock_release() (stopped) | lockdep analyze And you can repeat the above two times as we have two rcu read side sections when we submit an event. This is fixed in this patch by moving the lock trace event under the lockdep recursion protection. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Paul Mackerras Cc: Hitoshi Mitake Cc: Li Zefan Cc: Lai Jiangshan Cc: Masami Hiramatsu Cc: Jens Axboe --- kernel/lockdep.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 0c30d0455de1..65b5f5b7c298 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3211,8 +3211,6 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, { unsigned long flags; - trace_lock_acquire(lock, subclass, trylock, read, check, nest_lock, ip); - if (unlikely(current->lockdep_recursion)) return; @@ -3220,6 +3218,7 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, check_flags(flags); current->lockdep_recursion = 1; + trace_lock_acquire(lock, subclass, trylock, read, check, nest_lock, ip); __lock_acquire(lock, subclass, trylock, read, check, irqs_disabled_flags(flags), nest_lock, ip, 0); current->lockdep_recursion = 0; @@ -3232,14 +3231,13 @@ void lock_release(struct lockdep_map *lock, int nested, { unsigned long flags; - trace_lock_release(lock, nested, ip); - if (unlikely(current->lockdep_recursion)) return; raw_local_irq_save(flags); check_flags(flags); current->lockdep_recursion = 1; + trace_lock_release(lock, nested, ip); __lock_release(lock, nested, ip); current->lockdep_recursion = 0; raw_local_irq_restore(flags); @@ -3413,8 +3411,6 @@ void lock_contended(struct lockdep_map *lock, unsigned long ip) { unsigned long flags; - trace_lock_contended(lock, ip); - if (unlikely(!lock_stat)) return; @@ -3424,6 +3420,7 @@ void lock_contended(struct lockdep_map *lock, unsigned long ip) raw_local_irq_save(flags); check_flags(flags); current->lockdep_recursion = 1; + trace_lock_contended(lock, ip); __lock_contended(lock, ip); current->lockdep_recursion = 0; raw_local_irq_restore(flags); -- cgit v1.2.2 From 5331d7b84613b8325362dde53dc2bff2fb87d351 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 4 Mar 2010 21:15:56 +0100 Subject: perf: Introduce new perf_fetch_caller_regs() for hot regs snapshot Events that trigger overflows by interrupting a context can use get_irq_regs() or task_pt_regs() to retrieve the state when the event triggered. But this is not the case for some other class of events like trace events as tracepoints are executed in the same context than the code that triggered the event. It means we need a different api to capture the regs there, namely we need a hot snapshot to get the most important informations for perf: the instruction pointer to get the event origin, the frame pointer for the callchain, the code segment for user_mode() tests (we always use __KERNEL_CS as trace events always occur from the kernel) and the eflags for further purposes. v2: rename perf_save_regs to perf_fetch_caller_regs as per Masami's suggestion. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Masami Hiramatsu Cc: Jason Baron Cc: Archs --- kernel/perf_event.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 52c69a34d697..359d7f690c2b 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2786,6 +2786,11 @@ __weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) return NULL; } +__weak +void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip) +{ +} + /* * Output */ -- cgit v1.2.2 From c530665c31c0140b74ca7689e7f836177796e5bd Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 3 Mar 2010 07:16:16 +0100 Subject: perf: Take a hot regs snapshot for trace events We are taking a wrong regs snapshot when a trace event triggers. Either we use get_irq_regs(), which gives us the interrupted registers if we are in an interrupt, or we use task_pt_regs() which gives us the state before we entered the kernel, assuming we are lucky enough to be no kernel thread, in which case task_pt_regs() returns the initial set of regs when the kernel thread was started. What we want is different. We need a hot snapshot of the regs, so that we can get the instruction pointer to record in the sample, the frame pointer for the callchain, and some other things. Let's use the new perf_fetch_caller_regs() for that. Comparison with perf record -e lock: -R -a -f -g Before: perf [kernel] [k] __do_softirq | --- __do_softirq | |--55.16%-- __open | --44.84%-- __write_nocancel After: perf [kernel] [k] perf_tp_event | --- perf_tp_event | |--41.07%-- lock_acquire | | | |--39.36%-- _raw_spin_lock | | | | | |--7.81%-- hrtimer_interrupt | | | smp_apic_timer_interrupt | | | apic_timer_interrupt The old case was producing unreliable callchains. Now having right frame and instruction pointers, we have the trace we want. Also syscalls and kprobe events already have the right regs, let's use them instead of wasting a retrieval. v2: Follow the rename perf_save_regs() -> perf_fetch_caller_regs() Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Masami Hiramatsu Cc: Jason Baron Cc: Archs --- kernel/perf_event.c | 8 ++------ kernel/trace/trace_event_profile.c | 3 ++- kernel/trace/trace_kprobe.c | 5 +++-- kernel/trace/trace_syscalls.c | 4 ++-- 4 files changed, 9 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 359d7f690c2b..45b4b6e55891 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4318,9 +4318,8 @@ static const struct pmu perf_ops_task_clock = { #ifdef CONFIG_EVENT_TRACING void perf_tp_event(int event_id, u64 addr, u64 count, void *record, - int entry_size) + int entry_size, struct pt_regs *regs) { - struct pt_regs *regs = get_irq_regs(); struct perf_sample_data data; struct perf_raw_record raw = { .size = entry_size, @@ -4330,12 +4329,9 @@ void perf_tp_event(int event_id, u64 addr, u64 count, void *record, perf_sample_data_init(&data, addr); data.raw = &raw; - if (!regs) - regs = task_pt_regs(current); - /* Trace events already protected against recursion */ do_perf_sw_event(PERF_TYPE_TRACEPOINT, event_id, count, 1, - &data, regs); + &data, regs); } EXPORT_SYMBOL_GPL(perf_tp_event); diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index f0d693005075..e66d21e15a0f 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -2,13 +2,14 @@ * trace event based perf counter profiling * * Copyright (C) 2009 Red Hat Inc, Peter Zijlstra - * + * Copyright (C) 2009-2010 Frederic Weisbecker */ #include #include #include "trace.h" +DEFINE_PER_CPU(struct pt_regs, perf_trace_regs); static char *perf_trace_buf; static char *perf_trace_buf_nmi; diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 505c92273b1a..f7a20a8bfb31 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1240,7 +1240,7 @@ static __kprobes void kprobe_profile_func(struct kprobe *kp, for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags, regs); } /* Kretprobe profile handler */ @@ -1271,7 +1271,8 @@ static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, + irq_flags, regs); } static int probe_profile_enable(struct ftrace_event_call *call) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index cba47d7935cc..7e6e84fb7b6c 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -467,7 +467,7 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags, regs); } int prof_sysenter_enable(struct ftrace_event_call *call) @@ -542,7 +542,7 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags, regs); } int prof_sysexit_enable(struct ftrace_event_call *call) -- cgit v1.2.2 From 97d5a22005f38057b4bc0d95f81cd26510268794 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 5 Mar 2010 05:35:37 +0100 Subject: perf: Drop the obsolete profile naming for trace events Drop the obsolete "profile" naming used by perf for trace events. Perf can now do more than simple events counting, so generalize the API naming. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt Cc: Masami Hiramatsu Cc: Jason Baron --- kernel/perf_event.c | 4 +- kernel/trace/Makefile | 2 +- kernel/trace/trace_event_perf.c | 165 +++++++++++++++++++++++++++++++++++++ kernel/trace/trace_event_profile.c | 165 ------------------------------------- kernel/trace/trace_events.c | 2 +- kernel/trace/trace_kprobe.c | 28 +++---- kernel/trace/trace_syscalls.c | 72 ++++++++-------- 7 files changed, 219 insertions(+), 219 deletions(-) create mode 100644 kernel/trace/trace_event_perf.c delete mode 100644 kernel/trace/trace_event_profile.c (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 45b4b6e55891..c502b18594cc 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4347,7 +4347,7 @@ static int perf_tp_event_match(struct perf_event *event, static void tp_perf_event_destroy(struct perf_event *event) { - ftrace_profile_disable(event->attr.config); + perf_trace_disable(event->attr.config); } static const struct pmu *tp_perf_event_init(struct perf_event *event) @@ -4361,7 +4361,7 @@ static const struct pmu *tp_perf_event_init(struct perf_event *event) !capable(CAP_SYS_ADMIN)) return ERR_PTR(-EPERM); - if (ftrace_profile_enable(event->attr.config)) + if (perf_trace_enable(event->attr.config)) return NULL; event->destroy = tp_perf_event_destroy; diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index d00c6fe23f54..78edc6490038 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -52,7 +52,7 @@ obj-$(CONFIG_EVENT_TRACING) += trace_events.o obj-$(CONFIG_EVENT_TRACING) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o ifeq ($(CONFIG_PERF_EVENTS),y) -obj-$(CONFIG_EVENT_TRACING) += trace_event_profile.o +obj-$(CONFIG_EVENT_TRACING) += trace_event_perf.o endif obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c new file mode 100644 index 000000000000..f315b12a41d8 --- /dev/null +++ b/kernel/trace/trace_event_perf.c @@ -0,0 +1,165 @@ +/* + * trace event based perf event profiling/tracing + * + * Copyright (C) 2009 Red Hat Inc, Peter Zijlstra + * Copyright (C) 2009-2010 Frederic Weisbecker + */ + +#include +#include +#include "trace.h" + +DEFINE_PER_CPU(struct pt_regs, perf_trace_regs); + +static char *perf_trace_buf; +static char *perf_trace_buf_nmi; + +typedef typeof(char [PERF_MAX_TRACE_SIZE]) perf_trace_t ; + +/* Count the events in use (per event id, not per instance) */ +static int total_ref_count; + +static int perf_trace_event_enable(struct ftrace_event_call *event) +{ + char *buf; + int ret = -ENOMEM; + + if (event->perf_refcount++ > 0) + return 0; + + if (!total_ref_count) { + buf = (char *)alloc_percpu(perf_trace_t); + if (!buf) + goto fail_buf; + + rcu_assign_pointer(perf_trace_buf, buf); + + buf = (char *)alloc_percpu(perf_trace_t); + if (!buf) + goto fail_buf_nmi; + + rcu_assign_pointer(perf_trace_buf_nmi, buf); + } + + ret = event->perf_event_enable(event); + if (!ret) { + total_ref_count++; + return 0; + } + +fail_buf_nmi: + if (!total_ref_count) { + free_percpu(perf_trace_buf_nmi); + free_percpu(perf_trace_buf); + perf_trace_buf_nmi = NULL; + perf_trace_buf = NULL; + } +fail_buf: + event->perf_refcount--; + + return ret; +} + +int perf_trace_enable(int event_id) +{ + struct ftrace_event_call *event; + int ret = -EINVAL; + + mutex_lock(&event_mutex); + list_for_each_entry(event, &ftrace_events, list) { + if (event->id == event_id && event->perf_event_enable && + try_module_get(event->mod)) { + ret = perf_trace_event_enable(event); + break; + } + } + mutex_unlock(&event_mutex); + + return ret; +} + +static void perf_trace_event_disable(struct ftrace_event_call *event) +{ + char *buf, *nmi_buf; + + if (--event->perf_refcount > 0) + return; + + event->perf_event_disable(event); + + if (!--total_ref_count) { + buf = perf_trace_buf; + rcu_assign_pointer(perf_trace_buf, NULL); + + nmi_buf = perf_trace_buf_nmi; + rcu_assign_pointer(perf_trace_buf_nmi, NULL); + + /* + * Ensure every events in profiling have finished before + * releasing the buffers + */ + synchronize_sched(); + + free_percpu(buf); + free_percpu(nmi_buf); + } +} + +void perf_trace_disable(int event_id) +{ + struct ftrace_event_call *event; + + mutex_lock(&event_mutex); + list_for_each_entry(event, &ftrace_events, list) { + if (event->id == event_id) { + perf_trace_event_disable(event); + module_put(event->mod); + break; + } + } + mutex_unlock(&event_mutex); +} + +__kprobes void *perf_trace_buf_prepare(int size, unsigned short type, + int *rctxp, unsigned long *irq_flags) +{ + struct trace_entry *entry; + char *trace_buf, *raw_data; + int pc, cpu; + + pc = preempt_count(); + + /* Protect the per cpu buffer, begin the rcu read side */ + local_irq_save(*irq_flags); + + *rctxp = perf_swevent_get_recursion_context(); + if (*rctxp < 0) + goto err_recursion; + + cpu = smp_processor_id(); + + if (in_nmi()) + trace_buf = rcu_dereference(perf_trace_buf_nmi); + else + trace_buf = rcu_dereference(perf_trace_buf); + + if (!trace_buf) + goto err; + + raw_data = per_cpu_ptr(trace_buf, cpu); + + /* zero the dead bytes from align to not leak stack to user */ + *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + + entry = (struct trace_entry *)raw_data; + tracing_generic_entry_update(entry, *irq_flags, pc); + entry->type = type; + + return raw_data; +err: + perf_swevent_put_recursion_context(*rctxp); +err_recursion: + local_irq_restore(*irq_flags); + return NULL; +} +EXPORT_SYMBOL_GPL(perf_trace_buf_prepare); diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c deleted file mode 100644 index e66d21e15a0f..000000000000 --- a/kernel/trace/trace_event_profile.c +++ /dev/null @@ -1,165 +0,0 @@ -/* - * trace event based perf counter profiling - * - * Copyright (C) 2009 Red Hat Inc, Peter Zijlstra - * Copyright (C) 2009-2010 Frederic Weisbecker - */ - -#include -#include -#include "trace.h" - -DEFINE_PER_CPU(struct pt_regs, perf_trace_regs); - -static char *perf_trace_buf; -static char *perf_trace_buf_nmi; - -typedef typeof(char [FTRACE_MAX_PROFILE_SIZE]) perf_trace_t ; - -/* Count the events in use (per event id, not per instance) */ -static int total_profile_count; - -static int ftrace_profile_enable_event(struct ftrace_event_call *event) -{ - char *buf; - int ret = -ENOMEM; - - if (event->profile_count++ > 0) - return 0; - - if (!total_profile_count) { - buf = (char *)alloc_percpu(perf_trace_t); - if (!buf) - goto fail_buf; - - rcu_assign_pointer(perf_trace_buf, buf); - - buf = (char *)alloc_percpu(perf_trace_t); - if (!buf) - goto fail_buf_nmi; - - rcu_assign_pointer(perf_trace_buf_nmi, buf); - } - - ret = event->profile_enable(event); - if (!ret) { - total_profile_count++; - return 0; - } - -fail_buf_nmi: - if (!total_profile_count) { - free_percpu(perf_trace_buf_nmi); - free_percpu(perf_trace_buf); - perf_trace_buf_nmi = NULL; - perf_trace_buf = NULL; - } -fail_buf: - event->profile_count--; - - return ret; -} - -int ftrace_profile_enable(int event_id) -{ - struct ftrace_event_call *event; - int ret = -EINVAL; - - mutex_lock(&event_mutex); - list_for_each_entry(event, &ftrace_events, list) { - if (event->id == event_id && event->profile_enable && - try_module_get(event->mod)) { - ret = ftrace_profile_enable_event(event); - break; - } - } - mutex_unlock(&event_mutex); - - return ret; -} - -static void ftrace_profile_disable_event(struct ftrace_event_call *event) -{ - char *buf, *nmi_buf; - - if (--event->profile_count > 0) - return; - - event->profile_disable(event); - - if (!--total_profile_count) { - buf = perf_trace_buf; - rcu_assign_pointer(perf_trace_buf, NULL); - - nmi_buf = perf_trace_buf_nmi; - rcu_assign_pointer(perf_trace_buf_nmi, NULL); - - /* - * Ensure every events in profiling have finished before - * releasing the buffers - */ - synchronize_sched(); - - free_percpu(buf); - free_percpu(nmi_buf); - } -} - -void ftrace_profile_disable(int event_id) -{ - struct ftrace_event_call *event; - - mutex_lock(&event_mutex); - list_for_each_entry(event, &ftrace_events, list) { - if (event->id == event_id) { - ftrace_profile_disable_event(event); - module_put(event->mod); - break; - } - } - mutex_unlock(&event_mutex); -} - -__kprobes void *ftrace_perf_buf_prepare(int size, unsigned short type, - int *rctxp, unsigned long *irq_flags) -{ - struct trace_entry *entry; - char *trace_buf, *raw_data; - int pc, cpu; - - pc = preempt_count(); - - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(*irq_flags); - - *rctxp = perf_swevent_get_recursion_context(); - if (*rctxp < 0) - goto err_recursion; - - cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto err; - - raw_data = per_cpu_ptr(trace_buf, cpu); - - /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - - entry = (struct trace_entry *)raw_data; - tracing_generic_entry_update(entry, *irq_flags, pc); - entry->type = type; - - return raw_data; -err: - perf_swevent_put_recursion_context(*rctxp); -err_recursion: - local_irq_restore(*irq_flags); - return NULL; -} -EXPORT_SYMBOL_GPL(ftrace_perf_buf_prepare); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 3f972ad98d04..beab8bf2f310 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -938,7 +938,7 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, trace_create_file("enable", 0644, call->dir, call, enable); - if (call->id && call->profile_enable) + if (call->id && call->perf_event_enable) trace_create_file("id", 0444, call->dir, call, id); diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index f7a20a8bfb31..1251e367bae9 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1214,7 +1214,7 @@ static int set_print_fmt(struct trace_probe *tp) #ifdef CONFIG_PERF_EVENTS /* Kprobe profile handler */ -static __kprobes void kprobe_profile_func(struct kprobe *kp, +static __kprobes void kprobe_perf_func(struct kprobe *kp, struct pt_regs *regs) { struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); @@ -1227,11 +1227,11 @@ static __kprobes void kprobe_profile_func(struct kprobe *kp, __size = SIZEOF_KPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, + if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, "profile buffer not large enough")) return; - entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + entry = perf_trace_buf_prepare(size, call->id, &rctx, &irq_flags); if (!entry) return; @@ -1240,11 +1240,11 @@ static __kprobes void kprobe_profile_func(struct kprobe *kp, for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags, regs); + perf_trace_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags, regs); } /* Kretprobe profile handler */ -static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, +static __kprobes void kretprobe_perf_func(struct kretprobe_instance *ri, struct pt_regs *regs) { struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); @@ -1257,11 +1257,11 @@ static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, __size = SIZEOF_KRETPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, + if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, "profile buffer not large enough")) return; - entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + entry = perf_trace_buf_prepare(size, call->id, &rctx, &irq_flags); if (!entry) return; @@ -1271,11 +1271,11 @@ static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, + perf_trace_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags, regs); } -static int probe_profile_enable(struct ftrace_event_call *call) +static int probe_perf_enable(struct ftrace_event_call *call) { struct trace_probe *tp = (struct trace_probe *)call->data; @@ -1287,7 +1287,7 @@ static int probe_profile_enable(struct ftrace_event_call *call) return enable_kprobe(&tp->rp.kp); } -static void probe_profile_disable(struct ftrace_event_call *call) +static void probe_perf_disable(struct ftrace_event_call *call) { struct trace_probe *tp = (struct trace_probe *)call->data; @@ -1312,7 +1312,7 @@ int kprobe_dispatcher(struct kprobe *kp, struct pt_regs *regs) kprobe_trace_func(kp, regs); #ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) - kprobe_profile_func(kp, regs); + kprobe_perf_func(kp, regs); #endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1326,7 +1326,7 @@ int kretprobe_dispatcher(struct kretprobe_instance *ri, struct pt_regs *regs) kretprobe_trace_func(ri, regs); #ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) - kretprobe_profile_func(ri, regs); + kretprobe_perf_func(ri, regs); #endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1359,8 +1359,8 @@ static int register_probe_event(struct trace_probe *tp) call->unregfunc = probe_event_disable; #ifdef CONFIG_PERF_EVENTS - call->profile_enable = probe_profile_enable; - call->profile_disable = probe_profile_disable; + call->perf_event_enable = probe_perf_enable; + call->perf_event_disable = probe_perf_disable; #endif call->data = tp; ret = trace_add_event_call(call); diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 7e6e84fb7b6c..33c2a5b769dc 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -428,12 +428,12 @@ core_initcall(init_ftrace_syscalls); #ifdef CONFIG_PERF_EVENTS -static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls); -static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls); -static int sys_prof_refcount_enter; -static int sys_prof_refcount_exit; +static DECLARE_BITMAP(enabled_perf_enter_syscalls, NR_syscalls); +static DECLARE_BITMAP(enabled_perf_exit_syscalls, NR_syscalls); +static int sys_perf_refcount_enter; +static int sys_perf_refcount_exit; -static void prof_syscall_enter(struct pt_regs *regs, long id) +static void perf_syscall_enter(struct pt_regs *regs, long id) { struct syscall_metadata *sys_data; struct syscall_trace_enter *rec; @@ -443,7 +443,7 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) int size; syscall_nr = syscall_get_nr(current, regs); - if (!test_bit(syscall_nr, enabled_prof_enter_syscalls)) + if (!test_bit(syscall_nr, enabled_perf_enter_syscalls)) return; sys_data = syscall_nr_to_meta(syscall_nr); @@ -455,11 +455,11 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) size = ALIGN(size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, - "profile buffer not large enough")) + if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, + "perf buffer not large enough")) return; - rec = (struct syscall_trace_enter *)ftrace_perf_buf_prepare(size, + rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size, sys_data->enter_event->id, &rctx, &flags); if (!rec) return; @@ -467,10 +467,10 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags, regs); + perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs); } -int prof_sysenter_enable(struct ftrace_event_call *call) +int perf_sysenter_enable(struct ftrace_event_call *call) { int ret = 0; int num; @@ -478,34 +478,34 @@ int prof_sysenter_enable(struct ftrace_event_call *call) num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); - if (!sys_prof_refcount_enter) - ret = register_trace_sys_enter(prof_syscall_enter); + if (!sys_perf_refcount_enter) + ret = register_trace_sys_enter(perf_syscall_enter); if (ret) { pr_info("event trace: Could not activate" "syscall entry trace point"); } else { - set_bit(num, enabled_prof_enter_syscalls); - sys_prof_refcount_enter++; + set_bit(num, enabled_perf_enter_syscalls); + sys_perf_refcount_enter++; } mutex_unlock(&syscall_trace_lock); return ret; } -void prof_sysenter_disable(struct ftrace_event_call *call) +void perf_sysenter_disable(struct ftrace_event_call *call) { int num; num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); - sys_prof_refcount_enter--; - clear_bit(num, enabled_prof_enter_syscalls); - if (!sys_prof_refcount_enter) - unregister_trace_sys_enter(prof_syscall_enter); + sys_perf_refcount_enter--; + clear_bit(num, enabled_perf_enter_syscalls); + if (!sys_perf_refcount_enter) + unregister_trace_sys_enter(perf_syscall_enter); mutex_unlock(&syscall_trace_lock); } -static void prof_syscall_exit(struct pt_regs *regs, long ret) +static void perf_syscall_exit(struct pt_regs *regs, long ret) { struct syscall_metadata *sys_data; struct syscall_trace_exit *rec; @@ -515,7 +515,7 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) int size; syscall_nr = syscall_get_nr(current, regs); - if (!test_bit(syscall_nr, enabled_prof_exit_syscalls)) + if (!test_bit(syscall_nr, enabled_perf_exit_syscalls)) return; sys_data = syscall_nr_to_meta(syscall_nr); @@ -530,11 +530,11 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) * Impossible, but be paranoid with the future * How to put this check outside runtime? */ - if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, - "exit event has grown above profile buffer size")) + if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, + "exit event has grown above perf buffer size")) return; - rec = (struct syscall_trace_exit *)ftrace_perf_buf_prepare(size, + rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size, sys_data->exit_event->id, &rctx, &flags); if (!rec) return; @@ -542,10 +542,10 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags, regs); + perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs); } -int prof_sysexit_enable(struct ftrace_event_call *call) +int perf_sysexit_enable(struct ftrace_event_call *call) { int ret = 0; int num; @@ -553,30 +553,30 @@ int prof_sysexit_enable(struct ftrace_event_call *call) num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); - if (!sys_prof_refcount_exit) - ret = register_trace_sys_exit(prof_syscall_exit); + if (!sys_perf_refcount_exit) + ret = register_trace_sys_exit(perf_syscall_exit); if (ret) { pr_info("event trace: Could not activate" "syscall exit trace point"); } else { - set_bit(num, enabled_prof_exit_syscalls); - sys_prof_refcount_exit++; + set_bit(num, enabled_perf_exit_syscalls); + sys_perf_refcount_exit++; } mutex_unlock(&syscall_trace_lock); return ret; } -void prof_sysexit_disable(struct ftrace_event_call *call) +void perf_sysexit_disable(struct ftrace_event_call *call) { int num; num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); - sys_prof_refcount_exit--; - clear_bit(num, enabled_prof_exit_syscalls); - if (!sys_prof_refcount_exit) - unregister_trace_sys_exit(prof_syscall_exit); + sys_perf_refcount_exit--; + clear_bit(num, enabled_perf_exit_syscalls); + if (!sys_perf_refcount_exit) + unregister_trace_sys_exit(perf_syscall_exit); mutex_unlock(&syscall_trace_lock); } -- cgit v1.2.2 From 0b1adaa031a55e44f5dd942f234bf09d28e8a0d6 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 9 Mar 2010 19:45:54 +0100 Subject: genirq: Prevent oneshot irq thread race Lars-Peter pointed out that the oneshot threaded interrupt handler code has the following race: CPU0 CPU1 hande_level_irq(irq X) mask_ack_irq(irq X) handle_IRQ_event(irq X) wake_up(thread_handler) thread handler(irq X) runs finalize_oneshot(irq X) does not unmask due to !(desc->status & IRQ_MASKED) return from irq does not unmask due to (desc->status & IRQ_ONESHOT) This leaves the interrupt line masked forever. The reason for this is the inconsistent handling of the IRQ_MASKED flag. Instead of setting it in the mask function the oneshot support sets the flag after waking up the irq thread. The solution for this is to set/clear the IRQ_MASKED status whenever we mask/unmask an interrupt line. That's the easy part, but that cleanup opens another race: CPU0 CPU1 hande_level_irq(irq) mask_ack_irq(irq) handle_IRQ_event(irq) wake_up(thread_handler) thread handler(irq) runs finalize_oneshot_irq(irq) unmask(irq) irq triggers again handle_level_irq(irq) mask_ack_irq(irq) return from irq due to IRQ_INPROGRESS return from irq does not unmask due to (desc->status & IRQ_ONESHOT) This requires that we synchronize finalize_oneshot_irq() with the primary handler. If IRQ_INPROGESS is set we wait until the primary handler on the other CPU has returned before unmasking the interrupt line again. We probably have never seen that problem because it does not happen on UP and on SMP the irqbalancer protects us by pinning the primary handler and the thread to the same CPU. Reported-by: Lars-Peter Clausen Signed-off-by: Thomas Gleixner Cc: stable@kernel.org --- kernel/irq/chip.c | 31 ++++++++++++++++++++++--------- kernel/irq/manage.c | 18 ++++++++++++++++++ 2 files changed, 40 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index d70394f12ee9..71eba24a39a2 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -359,6 +359,23 @@ static inline void mask_ack_irq(struct irq_desc *desc, int irq) if (desc->chip->ack) desc->chip->ack(irq); } + desc->status |= IRQ_MASKED; +} + +static inline void mask_irq(struct irq_desc *desc, int irq) +{ + if (desc->chip->mask) { + desc->chip->mask(irq); + desc->status |= IRQ_MASKED; + } +} + +static inline void unmask_irq(struct irq_desc *desc, int irq) +{ + if (desc->chip->unmask) { + desc->chip->unmask(irq); + desc->status &= ~IRQ_MASKED; + } } /* @@ -484,10 +501,8 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) raw_spin_lock(&desc->lock); desc->status &= ~IRQ_INPROGRESS; - if (unlikely(desc->status & IRQ_ONESHOT)) - desc->status |= IRQ_MASKED; - else if (!(desc->status & IRQ_DISABLED) && desc->chip->unmask) - desc->chip->unmask(irq); + if (!(desc->status & (IRQ_DISABLED | IRQ_ONESHOT))) + unmask_irq(desc, irq); out_unlock: raw_spin_unlock(&desc->lock); } @@ -524,8 +539,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) action = desc->action; if (unlikely(!action || (desc->status & IRQ_DISABLED))) { desc->status |= IRQ_PENDING; - if (desc->chip->mask) - desc->chip->mask(irq); + mask_irq(desc, irq); goto out; } @@ -593,7 +607,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) irqreturn_t action_ret; if (unlikely(!action)) { - desc->chip->mask(irq); + mask_irq(desc, irq); goto out_unlock; } @@ -605,8 +619,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) if (unlikely((desc->status & (IRQ_PENDING | IRQ_MASKED | IRQ_DISABLED)) == (IRQ_PENDING | IRQ_MASKED))) { - desc->chip->unmask(irq); - desc->status &= ~IRQ_MASKED; + unmask_irq(desc, irq); } desc->status &= ~IRQ_PENDING; diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index eb6078ca60c7..69a3d7b9414c 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -483,8 +483,26 @@ static int irq_wait_for_interrupt(struct irqaction *action) */ static void irq_finalize_oneshot(unsigned int irq, struct irq_desc *desc) { +again: chip_bus_lock(irq, desc); raw_spin_lock_irq(&desc->lock); + + /* + * Implausible though it may be we need to protect us against + * the following scenario: + * + * The thread is faster done than the hard interrupt handler + * on the other CPU. If we unmask the irq line then the + * interrupt can come in again and masks the line, leaves due + * to IRQ_INPROGRESS and the irq line is masked forever. + */ + if (unlikely(desc->status & IRQ_INPROGRESS)) { + raw_spin_unlock_irq(&desc->lock); + chip_bus_sync_unlock(irq, desc); + cpu_relax(); + goto again; + } + if (!(desc->status & IRQ_DISABLED) && (desc->status & IRQ_MASKED)) { desc->status &= ~IRQ_MASKED; desc->chip->unmask(irq); -- cgit v1.2.2 From 220b140b52ab6cc133f674a7ffec8fa792054f25 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 10 Mar 2010 20:45:52 +1100 Subject: perf_event: Fix oops triggered by cpu offline/online Anton Blanchard found that he could reliably make the kernel hit a BUG_ON in the slab allocator by taking a cpu offline and then online while a system-wide perf record session was running. The reason is that when the cpu comes up, we completely reinitialize the ctx field of the struct perf_cpu_context for the cpu. If there is a system-wide perf record session running, then there will be a struct perf_event that has a reference to the context, so its refcount will be 2. (The perf_event has been removed from the context's group_entry and event_entry lists by perf_event_exit_cpu(), but that doesn't remove the perf_event's reference to the context and doesn't decrement the context's refcount.) When the cpu comes up, perf_event_init_cpu() gets called, and it calls __perf_event_init_context() on the cpu's context. That resets the refcount to 1. Then when the perf record session finishes and the perf_event is closed, the refcount gets decremented to 0 and the context gets kfreed after an RCU grace period. Since the context wasn't kmalloced -- it's part of a per-cpu variable -- bad things happen. In fact we don't need to completely reinitialize the context when the cpu comes up. It's sufficient to initialize the context once at boot, but we need to do it for all possible cpus. This moves the context initialization to happen at boot time. With this, we don't trash the refcount and the context never gets kfreed, and we don't hit the BUG_ON. Reported-by: Anton Blanchard Signed-off-by: Paul Mackerras Tested-by: Anton Blanchard Acked-by: Peter Zijlstra Cc: Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index c502b18594cc..fb3031cf9f17 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5368,12 +5368,22 @@ int perf_event_init_task(struct task_struct *child) return ret; } +static void __init perf_event_init_all_cpus(void) +{ + int cpu; + struct perf_cpu_context *cpuctx; + + for_each_possible_cpu(cpu) { + cpuctx = &per_cpu(perf_cpu_context, cpu); + __perf_event_init_context(&cpuctx->ctx, NULL); + } +} + static void __cpuinit perf_event_init_cpu(int cpu) { struct perf_cpu_context *cpuctx; cpuctx = &per_cpu(perf_cpu_context, cpu); - __perf_event_init_context(&cpuctx->ctx, NULL); spin_lock(&perf_resource_lock); cpuctx->max_pertask = perf_max_events - perf_reserved_percpu; @@ -5439,6 +5449,7 @@ static struct notifier_block __cpuinitdata perf_cpu_nb = { void __init perf_event_init(void) { + perf_event_init_all_cpus(); perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_UP_PREPARE, (void *)(long)smp_processor_id()); perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_ONLINE, -- cgit v1.2.2 From 3f379b03fbfddd20536389a85c6456f8233d1f8d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 5 Mar 2010 15:03:25 -0800 Subject: ftrace: Replace read_barrier_depends() with rcu_dereference_raw() Replace the calls to read_barrier_depends() in ftrace_list_func() with rcu_dereference_raw() to improve readability. The reason that we use rcu_dereference_raw() here is that removed entries are never freed, instead they are simply leaked. This is one of a very few cases where use of rcu_dereference_raw() is the long-term right answer. And I don't yet know of any others. ;-) Signed-off-by: Paul E. McKenney Acked-by: Steven Rostedt Cc: Frederic Weisbecker Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267830207-9474-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 83783579378f..8c5adc0e5db3 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -27,6 +27,7 @@ #include #include #include +#include #include @@ -88,18 +89,22 @@ ftrace_func_t ftrace_pid_function __read_mostly = ftrace_stub; static int ftrace_set_func(unsigned long *array, int *idx, char *buffer); #endif +/* + * Traverse the ftrace_list, invoking all entries. The reason that we + * can use rcu_dereference_raw() is that elements removed from this list + * are simply leaked, so there is no need to interact with a grace-period + * mechanism. The rcu_dereference_raw() calls are needed to handle + * concurrent insertions into the ftrace_list. + * + * Silly Alpha and silly pointer-speculation compiler optimizations! + */ static void ftrace_list_func(unsigned long ip, unsigned long parent_ip) { - struct ftrace_ops *op = ftrace_list; - - /* in case someone actually ports this to alpha! */ - read_barrier_depends(); + struct ftrace_ops *op = rcu_dereference_raw(ftrace_list); /*see above*/ while (op != &ftrace_list_end) { - /* silly alpha */ - read_barrier_depends(); op->func(ip, parent_ip); - op = op->next; + op = rcu_dereference_raw(op->next); /*see above*/ }; } @@ -154,8 +159,7 @@ static int __register_ftrace_function(struct ftrace_ops *ops) * the ops->next pointer is valid before another CPU sees * the ops pointer included into the ftrace_list. */ - smp_wmb(); - ftrace_list = ops; + rcu_assign_pointer(ftrace_list, ops); if (ftrace_enabled) { ftrace_func_t func; -- cgit v1.2.2 From 007b09243b099811124f69d492adeebe9e439f96 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 5 Mar 2010 15:03:26 -0800 Subject: rcu: Increase RCU CPU stall timeouts if PROVE_RCU CONFIG_PROVE_RCU imposes additional overhead on the kernel, so increase the RCU CPU stall timeouts in an attempt to allow for this effect. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1267830207-9474-2-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcutree.h | 21 +++++++++++++++------ 1 file changed, 15 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 1439eb504c22..4a525a30e08e 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -246,12 +246,21 @@ struct rcu_data { #define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ #ifdef CONFIG_RCU_CPU_STALL_DETECTOR -#define RCU_SECONDS_TILL_STALL_CHECK (10 * HZ) /* for rsp->jiffies_stall */ -#define RCU_SECONDS_TILL_STALL_RECHECK (30 * HZ) /* for rsp->jiffies_stall */ -#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ - /* to take at least one */ - /* scheduling clock irq */ - /* before ratting on them. */ + +#ifdef CONFIG_PROVE_RCU +#define RCU_STALL_DELAY_DELTA (5 * HZ) +#else +#define RCU_STALL_DELAY_DELTA 0 +#endif + +#define RCU_SECONDS_TILL_STALL_CHECK (10 * HZ + RCU_STALL_DELAY_DELTA) + /* for rsp->jiffies_stall */ +#define RCU_SECONDS_TILL_STALL_RECHECK (30 * HZ + RCU_STALL_DELAY_DELTA) + /* for rsp->jiffies_stall */ +#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ + /* to take at least one */ + /* scheduling clock irq */ + /* before ratting on them. */ #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ -- cgit v1.2.2 From ab3b3aa5dd01b3aaa6b15caee113b21b1b6520c4 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Sat, 6 Mar 2010 14:17:52 +0300 Subject: sched: Cleanup: remove unused variable in try_to_wake_up() We haven't used the "orig_rq" variable since 055a00865d "Fix/add missing update_rq_clock() calls" Signed-off-by: Dan Carpenter Cc: Peter Zijlstra Cc: Andreas Herrmann Cc: Gautham R Shenoy Cc: efault@gmx.de LKML-Reference: <20100306111752.GL4958@bicker> Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 6a212c97f523..2c1db81f80eb 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2359,7 +2359,7 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, { int cpu, orig_cpu, this_cpu, success = 0; unsigned long flags; - struct rq *rq, *orig_rq; + struct rq *rq; if (!sched_feat(SYNC_WAKEUPS)) wake_flags &= ~WF_SYNC; @@ -2367,7 +2367,7 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, this_cpu = get_cpu(); smp_wmb(); - rq = orig_rq = task_rq_lock(p, &flags); + rq = task_rq_lock(p, &flags); update_rq_clock(rq); if (!(p->state & state)) goto out; -- cgit v1.2.2 From 83ff56f46a8532488ee364bb93a9cb2a59490d33 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 9 Mar 2010 10:22:19 -0500 Subject: kprobes: Calculate the index correctly when freeing the out-of-line execution slot From : Ananth N Mavinakayanahalli When freeing the instruction slot, the arithmetic to calculate the index of the slot in the page needs to account for the total size of the instruction on the various architectures. Calculate the index correctly when freeing the out-of-line execution slot. Reported-by: Sachin Sant Reported-by: Heiko Carstens Signed-off-by: Ananth N Mavinakayanahalli Signed-off-by: Masami Hiramatsu LKML-Reference: <4B9667AB.9050507@redhat.com> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index fa034d29cf73..0ed46f3e51e9 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -259,7 +259,8 @@ static void __kprobes __free_insn_slot(struct kprobe_insn_cache *c, struct kprobe_insn_page *kip; list_for_each_entry(kip, &c->pages, list) { - long idx = ((long)slot - (long)kip->insns) / c->insn_size; + long idx = ((long)slot - (long)kip->insns) / + (c->insn_size * sizeof(kprobe_opcode_t)); if (idx >= 0 && idx < slots_per_page(c)) { WARN_ON(kip->slot_used[idx] != SLOT_USED); if (dirty) { -- cgit v1.2.2 From 639fe4b12f92b54c9c3b38c82cdafaa38cfd3e63 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Thu, 11 Mar 2010 15:30:35 +0800 Subject: perf: export perf_trace_regs and perf_arch_fetch_caller_regs Export perf_trace_regs and perf_arch_fetch_caller_regs since module will use these. Signed-off-by: Xiao Guangrong [ use EXPORT_PER_CPU_SYMBOL_GPL() ] Signed-off-by: Peter Zijlstra LKML-Reference: <4B989C1B.2090407@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_event_perf.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index f315b12a41d8..0709e4f75114 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -10,6 +10,7 @@ #include "trace.h" DEFINE_PER_CPU(struct pt_regs, perf_trace_regs); +EXPORT_PER_CPU_SYMBOL_GPL(perf_trace_regs); static char *perf_trace_buf; static char *perf_trace_buf_nmi; -- cgit v1.2.2 From 3d07467b7aa91623b31d7b5888a123a2c8c8e9cc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 10 Mar 2010 17:07:24 +0100 Subject: sched: Fix pick_next_highest_task_rt() for cgroups Since pick_next_highest_task_rt() already iterates all the cgroups and is really only interested in tasks, skip over the !task entries. Reported-by: Dhaval Giani Signed-off-by: Peter Zijlstra Tested-by: Dhaval Giani LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched_rt.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index bf3e38fdbe6d..c4fb42a66cab 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1146,7 +1146,12 @@ static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) if (next && next->prio < idx) continue; list_for_each_entry(rt_se, array->queue + idx, run_list) { - struct task_struct *p = rt_task_of(rt_se); + struct task_struct *p; + + if (!rt_entity_is_task(rt_se)) + continue; + + p = rt_task_of(rt_se); if (pick_rt_task(rq, p, cpu)) { next = p; break; -- cgit v1.2.2 From 80a05b9ffa7dc13f6693902dd8999a2b61a3a0d7 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 12 Mar 2010 17:34:14 +0100 Subject: clockevents: Sanitize min_delta_ns adjustment and prevent overflows The current logic which handles clock events programming failures can increase min_delta_ns unlimited and even can cause overflows. Sanitize it by: - prevent zero increase when min_delta_ns == 1 - limiting min_delta_ns to a jiffie - bail out if the jiffie limit is hit - add retries stats for /proc/timer_list so we can gather data Reported-by: Uwe Kleine-Koenig Signed-off-by: Thomas Gleixner --- kernel/time/tick-oneshot.c | 52 +++++++++++++++++++++++++++++++++++----------- kernel/time/timer_list.c | 3 ++- 2 files changed, 42 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-oneshot.c b/kernel/time/tick-oneshot.c index 0a8a213016f0..aada0e52680a 100644 --- a/kernel/time/tick-oneshot.c +++ b/kernel/time/tick-oneshot.c @@ -22,6 +22,29 @@ #include "tick-internal.h" +/* Limit min_delta to a jiffie */ +#define MIN_DELTA_LIMIT (NSEC_PER_SEC / HZ) + +static int tick_increase_min_delta(struct clock_event_device *dev) +{ + /* Nothing to do if we already reached the limit */ + if (dev->min_delta_ns >= MIN_DELTA_LIMIT) + return -ETIME; + + if (dev->min_delta_ns < 5000) + dev->min_delta_ns = 5000; + else + dev->min_delta_ns += dev->min_delta_ns >> 1; + + if (dev->min_delta_ns > MIN_DELTA_LIMIT) + dev->min_delta_ns = MIN_DELTA_LIMIT; + + printk(KERN_WARNING "CE: %s increased min_delta_ns to %llu nsec\n", + dev->name ? dev->name : "?", + (unsigned long long) dev->min_delta_ns); + return 0; +} + /** * tick_program_event internal worker function */ @@ -37,23 +60,28 @@ int tick_dev_program_event(struct clock_event_device *dev, ktime_t expires, if (!ret || !force) return ret; + dev->retries++; /* - * We tried 2 times to program the device with the given - * min_delta_ns. If that's not working then we double it + * We tried 3 times to program the device with the given + * min_delta_ns. If that's not working then we increase it * and emit a warning. */ if (++i > 2) { /* Increase the min. delta and try again */ - if (!dev->min_delta_ns) - dev->min_delta_ns = 5000; - else - dev->min_delta_ns += dev->min_delta_ns >> 1; - - printk(KERN_WARNING - "CE: %s increasing min_delta_ns to %llu nsec\n", - dev->name ? dev->name : "?", - (unsigned long long) dev->min_delta_ns << 1); - + if (tick_increase_min_delta(dev)) { + /* + * Get out of the loop if min_delta_ns + * hit the limit already. That's + * better than staying here forever. + * + * We clear next_event so we have a + * chance that the box survives. + */ + printk(KERN_WARNING + "CE: Reprogramming failure. Giving up\n"); + dev->next_event.tv64 = KTIME_MAX; + return -ETIME; + } i = 0; } diff --git a/kernel/time/timer_list.c b/kernel/time/timer_list.c index bdfb8dd1050c..1a4a7dd78777 100644 --- a/kernel/time/timer_list.c +++ b/kernel/time/timer_list.c @@ -228,6 +228,7 @@ print_tickdevice(struct seq_file *m, struct tick_device *td, int cpu) SEQ_printf(m, " event_handler: "); print_name_offset(m, dev->event_handler); SEQ_printf(m, "\n"); + SEQ_printf(m, " retries: %lu\n", dev->retries); } static void timer_list_show_tickdevices(struct seq_file *m) @@ -257,7 +258,7 @@ static int timer_list_show(struct seq_file *m, void *v) u64 now = ktime_to_ns(ktime_get()); int cpu; - SEQ_printf(m, "Timer List Version: v0.5\n"); + SEQ_printf(m, "Timer List Version: v0.6\n"); SEQ_printf(m, "HRTIMER_MAX_CLOCK_BASES: %d\n", HRTIMER_MAX_CLOCK_BASES); SEQ_printf(m, "now at %Ld nsecs\n", (unsigned long long)now); -- cgit v1.2.2 From 829b6c1ef488856c6a46a2f705f5068062d5f34c Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 11 Mar 2010 14:04:30 -0800 Subject: timer stats: Fix del_timer_sync() and try_to_del_timer_sync() These functions forgot to run timer_stats_timer_clear_start_info(). It's unobvious what effect this has and whether it matters much - we won't be printing it out anyway if the timer's detached. Untested, just an Ingo trollpatch. [ Nevertheless correct - tglx ] Signed-off-by: Andrew Morton Cc: Ingo Molnar Cc: johnstul@us.ibm.com Signed-off-by: Thomas Gleixner --- kernel/timer.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/timer.c b/kernel/timer.c index c61a7949387f..fc965eae0e87 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -880,6 +880,7 @@ int try_to_del_timer_sync(struct timer_list *timer) if (base->running_timer == timer) goto out; + timer_stats_timer_clear_start_info(timer); ret = 0; if (timer_pending(timer)) { detach_timer(timer, 1); -- cgit v1.2.2 From 15365c108ea27598e265f8c13e7051d99ca5b0b9 Mon Sep 17 00:00:00 2001 From: Stanislaw Gruszka Date: Thu, 11 Mar 2010 14:04:31 -0800 Subject: posix-cpu-timers: Reset expire cache when no timer is running When a process deletes cpu timer or a timer expires we do not clear the expiration cache sig->cputimer_expires. As a result the fastpath_timer_check() which prevents us to loop over all threads in case no timer is active is not working and we run the slow path needlessly on every tick. Zero sig->cputimer_expires in stop_process_timers(). Signed-off-by: Stanislaw Gruszka Cc: Ingo Molnar Cc: Oleg Nesterov Cc: Peter Zijlstra Cc: Hidetoshi Seto Cc: Spencer Candland Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/posix-cpu-timers.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 438ff4523513..edec25a68e9e 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -1060,9 +1060,9 @@ static void check_thread_timers(struct task_struct *tsk, } } -static void stop_process_timers(struct task_struct *tsk) +static void stop_process_timers(struct signal_struct *sig) { - struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; + struct thread_group_cputimer *cputimer = &sig->cputimer; unsigned long flags; if (!cputimer->running) @@ -1071,6 +1071,10 @@ static void stop_process_timers(struct task_struct *tsk) spin_lock_irqsave(&cputimer->lock, flags); cputimer->running = 0; spin_unlock_irqrestore(&cputimer->lock, flags); + + sig->cputime_expires.prof_exp = cputime_zero; + sig->cputime_expires.virt_exp = cputime_zero; + sig->cputime_expires.sched_exp = 0; } static u32 onecputick; @@ -1131,7 +1135,7 @@ static void check_process_timers(struct task_struct *tsk, list_empty(&timers[CPUCLOCK_VIRT]) && cputime_eq(sig->it[CPUCLOCK_VIRT].expires, cputime_zero) && list_empty(&timers[CPUCLOCK_SCHED])) { - stop_process_timers(tsk); + stop_process_timers(sig); return; } -- cgit v1.2.2 From baed7fc9b580bd3fb8252ff1d9b36eaf1f86b670 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 10 Mar 2010 15:21:18 -0800 Subject: Add generic sys_ipc wrapper Add a generic implementation of the ipc demultiplexer syscall. Except for s390 and sparc64 all implementations of the sys_ipc are nearly identical. There are slight differences in the types of the parameters, where mips and powerpc as the only 64-bit architectures with sys_ipc use unsigned long for the "third" argument as it gets casted to a pointer later, while it traditionally is an "int" like most other paramters. frv goes even further and uses unsigned long for all parameters execept for "ptr" which is a pointer type everywhere. The change from int to unsigned long for "third" and back to "int" for the others on frv should be fine due to the in-register calling conventions for syscalls (we already had a similar issue with the generic sys_ptrace), but I'd prefer to have the arch maintainers looks over this in details. Except for that h8300, m68k and m68knommu lack an impplementation of the semtimedop sub call which this patch adds, and various architectures have gets used - at least on i386 it seems superflous as the compat code on x86-64 and ia64 doesn't even bother to implement it. [akpm@linux-foundation.org: add sys_ipc to sys_ni.c] Signed-off-by: Christoph Hellwig Cc: Ralf Baechle Cc: Benjamin Herrenschmidt Cc: Paul Mundt Cc: Jeff Dike Cc: Hirokazu Takata Cc: Thomas Gleixner Cc: Ingo Molnar Reviewed-by: H. Peter Anvin Cc: Al Viro Cc: Arnd Bergmann Cc: Heiko Carstens Cc: Martin Schwidefsky Cc: "Luck, Tony" Cc: James Morris Cc: Andreas Schwab Acked-by: Jesper Nilsson Acked-by: Russell King Acked-by: David Howells Acked-by: Kyle McMartin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys_ni.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 695384f12a7d..70f2ea758ffe 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -126,6 +126,7 @@ cond_syscall(sys_setreuid16); cond_syscall(sys_setuid16); cond_syscall(sys_vm86old); cond_syscall(sys_vm86); +cond_syscall(sys_ipc); cond_syscall(compat_sys_ipc); cond_syscall(compat_sys_sysctl); cond_syscall(sys_flock); -- cgit v1.2.2 From e28cbf22933d0c0ccaf3c4c27a1a263b41f73859 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 10 Mar 2010 15:21:19 -0800 Subject: improve sys_newuname() for compat architectures On an architecture that supports 32-bit compat we need to override the reported machine in uname with the 32-bit value. Instead of doing this separately in every architecture introduce a COMPAT_UTS_MACHINE define in and apply it directly in sys_newuname(). Signed-off-by: Christoph Hellwig Cc: Ralf Baechle Cc: Benjamin Herrenschmidt Cc: Paul Mundt Cc: Jeff Dike Cc: Hirokazu Takata Cc: Thomas Gleixner Cc: Ingo Molnar Cc: H. Peter Anvin Cc: Al Viro Cc: Arnd Bergmann Cc: Heiko Carstens Cc: Martin Schwidefsky Cc: "Luck, Tony" Cc: James Morris Cc: Andreas Schwab Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 9814e43fb23b..e483eb5530e4 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -1114,6 +1115,15 @@ out: DECLARE_RWSEM(uts_sem); +#ifdef COMPAT_UTS_MACHINE +#define override_architecture(name) \ + (current->personality == PER_LINUX32 && \ + copy_to_user(name->machine, COMPAT_UTS_MACHINE, \ + sizeof(COMPAT_UTS_MACHINE))) +#else +#define override_architecture(name) 0 +#endif + SYSCALL_DEFINE1(newuname, struct new_utsname __user *, name) { int errno = 0; @@ -1122,6 +1132,9 @@ SYSCALL_DEFINE1(newuname, struct new_utsname __user *, name) if (copy_to_user(name, utsname(), sizeof *name)) errno = -EFAULT; up_read(&uts_sem); + + if (!errno && override_architecture(name)) + errno = -EFAULT; return errno; } -- cgit v1.2.2 From 5cacdb4add1b1e50fe75edc50ebbb7bddd9cf5e7 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 10 Mar 2010 15:21:21 -0800 Subject: Add generic sys_olduname() Add generic implementations of the old and really old uname system calls. Note that sh only implements sys_olduname but not sys_oldolduname, but I'm not going to bother with another ifdef for that special case. m32r implemented an old uname but never wired it up, so kill it, too. Signed-off-by: Christoph Hellwig Cc: Ralf Baechle Cc: Benjamin Herrenschmidt Cc: Paul Mundt Cc: Jeff Dike Cc: Hirokazu Takata Cc: Thomas Gleixner Cc: Ingo Molnar Cc: H. Peter Anvin Cc: Al Viro Cc: Arnd Bergmann Cc: Heiko Carstens Cc: Martin Schwidefsky Cc: "Luck, Tony" Cc: James Morris Cc: Andreas Schwab Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index e483eb5530e4..8298878f4f71 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1138,6 +1138,60 @@ SYSCALL_DEFINE1(newuname, struct new_utsname __user *, name) return errno; } +#ifdef __ARCH_WANT_SYS_OLD_UNAME +/* + * Old cruft + */ +SYSCALL_DEFINE1(uname, struct old_utsname __user *, name) +{ + int error = 0; + + if (!name) + return -EFAULT; + + down_read(&uts_sem); + if (copy_to_user(name, utsname(), sizeof(*name))) + error = -EFAULT; + up_read(&uts_sem); + + if (!error && override_architecture(name)) + error = -EFAULT; + return error; +} + +SYSCALL_DEFINE1(olduname, struct oldold_utsname __user *, name) +{ + int error; + + if (!name) + return -EFAULT; + if (!access_ok(VERIFY_WRITE, name, sizeof(struct oldold_utsname))) + return -EFAULT; + + down_read(&uts_sem); + error = __copy_to_user(&name->sysname, &utsname()->sysname, + __OLD_UTS_LEN); + error |= __put_user(0, name->sysname + __OLD_UTS_LEN); + error |= __copy_to_user(&name->nodename, &utsname()->nodename, + __OLD_UTS_LEN); + error |= __put_user(0, name->nodename + __OLD_UTS_LEN); + error |= __copy_to_user(&name->release, &utsname()->release, + __OLD_UTS_LEN); + error |= __put_user(0, name->release + __OLD_UTS_LEN); + error |= __copy_to_user(&name->version, &utsname()->version, + __OLD_UTS_LEN); + error |= __put_user(0, name->version + __OLD_UTS_LEN); + error |= __copy_to_user(&name->machine, &utsname()->machine, + __OLD_UTS_LEN); + error |= __put_user(0, name->machine + __OLD_UTS_LEN); + up_read(&uts_sem); + + if (!error && override_architecture(name)) + error = -EFAULT; + return error ? -EFAULT : 0; +} +#endif + SYSCALL_DEFINE2(sethostname, char __user *, name, int, len) { int errno; -- cgit v1.2.2 From 2468c7234b366eeb799ee0648cb58f9cba394a54 Mon Sep 17 00:00:00 2001 From: Daisuke Nishimura Date: Wed, 10 Mar 2010 15:22:03 -0800 Subject: cgroup: introduce cancel_attach() Add cancel_attach() operation to struct cgroup_subsys. cancel_attach() can be used when can_attach() operation prepares something for the subsys, but we should rollback what can_attach() operation has prepared if attach task fails after we've succeeded in can_attach(). Signed-off-by: Daisuke Nishimura Acked-by: Li Zefan Reviewed-by: Paul Menage Cc: Balbir Singh Acked-by: KAMEZAWA Hiroyuki Cc: Daisuke Nishimura Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 40 +++++++++++++++++++++++++++++++++------- 1 file changed, 33 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 4fd90e129772..be45d2f6008a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1554,7 +1554,7 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) { int retval = 0; - struct cgroup_subsys *ss; + struct cgroup_subsys *ss, *failed_ss = NULL; struct cgroup *oldcgrp; struct css_set *cg; struct css_set *newcg; @@ -1568,8 +1568,16 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) for_each_subsys(root, ss) { if (ss->can_attach) { retval = ss->can_attach(ss, cgrp, tsk, false); - if (retval) - return retval; + if (retval) { + /* + * Remember on which subsystem the can_attach() + * failed, so that we only call cancel_attach() + * against the subsystems whose can_attach() + * succeeded. (See below) + */ + failed_ss = ss; + goto out; + } } } @@ -1583,14 +1591,17 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) */ newcg = find_css_set(cg, cgrp); put_css_set(cg); - if (!newcg) - return -ENOMEM; + if (!newcg) { + retval = -ENOMEM; + goto out; + } task_lock(tsk); if (tsk->flags & PF_EXITING) { task_unlock(tsk); put_css_set(newcg); - return -ESRCH; + retval = -ESRCH; + goto out; } rcu_assign_pointer(tsk->cgroups, newcg); task_unlock(tsk); @@ -1616,7 +1627,22 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) * is no longer empty. */ cgroup_wakeup_rmdir_waiter(cgrp); - return 0; +out: + if (retval) { + for_each_subsys(root, ss) { + if (ss == failed_ss) + /* + * This subsystem was the one that failed the + * can_attach() check earlier, so we don't need + * to call cancel_attach() against it or any + * remaining subsystems. + */ + break; + if (ss->cancel_attach) + ss->cancel_attach(ss, cgrp, tsk, false); + } + } + return retval; } /* -- cgit v1.2.2 From d7b9fff711d5e8db8c844161c684017e556c38a0 Mon Sep 17 00:00:00 2001 From: Daisuke Nishimura Date: Wed, 10 Mar 2010 15:22:05 -0800 Subject: cgroup: introduce coalesce css_get() and css_put() Current css_get() and css_put() increment/decrement css->refcnt one by one. This patch add a new function __css_get(), which takes "count" as a arg and increment the css->refcnt by "count". And this patch also add a new arg("count") to __css_put() and change the function to decrement the css->refcnt by "count". These coalesce version of __css_get()/__css_put() will be used to improve performance of memcg's moving charge feature later, where instead of calling css_get()/css_put() repeatedly, these new functions will be used. No change is needed for current users of css_get()/css_put(). Signed-off-by: Daisuke Nishimura Acked-by: Paul Menage Cc: Balbir Singh Acked-by: KAMEZAWA Hiroyuki Cc: Li Zefan Cc: Daisuke Nishimura Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index be45d2f6008a..cace83ddbcdc 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3746,12 +3746,13 @@ static void check_for_release(struct cgroup *cgrp) } } -void __css_put(struct cgroup_subsys_state *css) +/* Caller must verify that the css is not for root cgroup */ +void __css_put(struct cgroup_subsys_state *css, int count) { struct cgroup *cgrp = css->cgroup; int val; rcu_read_lock(); - val = atomic_dec_return(&css->refcnt); + val = atomic_sub_return(count, &css->refcnt); if (val == 1) { if (notify_on_release(cgrp)) { set_bit(CGRP_RELEASABLE, &cgrp->flags); -- cgit v1.2.2 From aae8aab40367036931608fdaf9e2dc568b516f19 Mon Sep 17 00:00:00 2001 From: Ben Blum Date: Wed, 10 Mar 2010 15:22:07 -0800 Subject: cgroups: revamp subsys array This patch series provides the ability for cgroup subsystems to be compiled as modules both within and outside the kernel tree. This is mainly useful for classifiers and subsystems that hook into components that are already modules. cls_cgroup and blkio-cgroup serve as the example use cases for this feature. It provides an interface cgroup_load_subsys() and cgroup_unload_subsys() which modular subsystems can use to register and depart during runtime. The net_cls classifier subsystem serves as the example for a subsystem which can be converted into a module using these changes. Patch #1 sets up the subsys[] array so its contents can be dynamic as modules appear and (eventually) disappear. Iterations over the array are modified to handle when subsystems are absent, and the dynamic section of the array is protected by cgroup_mutex. Patch #2 implements an interface for modules to load subsystems, called cgroup_load_subsys, similar to cgroup_init_subsys, and adds a module pointer in struct cgroup_subsys. Patch #3 adds a mechanism for unloading modular subsystems, which includes a more advanced rework of the rudimentary reference counting introduced in patch 2. Patch #4 modifies the net_cls subsystem, which already had some module declarations, to be configurable as a module, which also serves as a simple proof-of-concept. Part of implementing patches 2 and 4 involved updating css pointers in each css_set when the module appears or leaves. In doing this, it was discovered that css_sets always remain linked to the dummy cgroup, regardless of whether or not any subsystems are actually bound to it (i.e., not mounted on an actual hierarchy). The subsystem loading and unloading code therefore should keep in mind the special cases where the added subsystem is the only one in the dummy cgroup (and therefore all css_sets need to be linked back into it) and where the removed subsys was the only one in the dummy cgroup (and therefore all css_sets should be unlinked from it) - however, as all css_sets always stay attached to the dummy cgroup anyway, these cases are ignored. Any fix that addresses this issue should also make sure these cases are addressed in the subsystem loading and unloading code. This patch: Make subsys[] able to be dynamically populated to support modular subsystems This patch reworks the way the subsys[] array is used so that subsystems can register themselves after boot time, and enables the internals of cgroups to be able to handle when subsystems are not present or may appear/disappear. Signed-off-by: Ben Blum Acked-by: Li Zefan Cc: Paul Menage Cc: "David S. Miller" Cc: KAMEZAWA Hiroyuki Cc: Lai Jiangshan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 96 +++++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 80 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index cace83ddbcdc..c92fb9549358 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -57,10 +57,14 @@ static DEFINE_MUTEX(cgroup_mutex); -/* Generate an array of cgroup subsystem pointers */ +/* + * Generate an array of cgroup subsystem pointers. At boot time, this is + * populated up to CGROUP_BUILTIN_SUBSYS_COUNT, and modular subsystems are + * registered after that. The mutable section of this array is protected by + * cgroup_mutex. + */ #define SUBSYS(_x) &_x ## _subsys, - -static struct cgroup_subsys *subsys[] = { +static struct cgroup_subsys *subsys[CGROUP_SUBSYS_COUNT] = { #include }; @@ -448,8 +452,11 @@ static struct css_set *find_existing_css_set( struct hlist_node *node; struct css_set *cg; - /* Built the set of subsystem state objects that we want to - * see in the new css_set */ + /* + * Build the set of subsystem state objects that we want to see in the + * new css_set. while subsystems can change globally, the entries here + * won't change, so no need for locking. + */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { if (root->subsys_bits & (1UL << i)) { /* Subsystem is in this hierarchy. So we want @@ -884,7 +891,9 @@ void cgroup_release_and_wakeup_rmdir(struct cgroup_subsys_state *css) css_put(css); } - +/* + * Call with cgroup_mutex held. + */ static int rebind_subsystems(struct cgroupfs_root *root, unsigned long final_bits) { @@ -892,6 +901,8 @@ static int rebind_subsystems(struct cgroupfs_root *root, struct cgroup *cgrp = &root->top_cgroup; int i; + BUG_ON(!mutex_is_locked(&cgroup_mutex)); + removed_bits = root->actual_subsys_bits & ~final_bits; added_bits = final_bits & ~root->actual_subsys_bits; /* Check that any added subsystems are currently free */ @@ -900,6 +911,12 @@ static int rebind_subsystems(struct cgroupfs_root *root, struct cgroup_subsys *ss = subsys[i]; if (!(bit & added_bits)) continue; + /* + * Nobody should tell us to do a subsys that doesn't exist: + * parse_cgroupfs_options should catch that case and refcounts + * ensure that subsystems won't disappear once selected. + */ + BUG_ON(ss == NULL); if (ss->root != &rootnode) { /* Subsystem isn't free */ return -EBUSY; @@ -919,6 +936,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, unsigned long bit = 1UL << i; if (bit & added_bits) { /* We're binding this subsystem to this hierarchy */ + BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i]); BUG_ON(!dummytop->subsys[i]); BUG_ON(dummytop->subsys[i]->cgroup != dummytop); @@ -932,6 +950,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, mutex_unlock(&ss->hierarchy_mutex); } else if (bit & removed_bits) { /* We're removing this subsystem */ + BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); mutex_lock(&ss->hierarchy_mutex); @@ -944,6 +963,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, mutex_unlock(&ss->hierarchy_mutex); } else if (bit & final_bits) { /* Subsystem state should already exist */ + BUG_ON(ss == NULL); BUG_ON(!cgrp->subsys[i]); } else { /* Subsystem state shouldn't exist */ @@ -986,14 +1006,18 @@ struct cgroup_sb_opts { }; -/* Convert a hierarchy specifier into a bitmask of subsystems and - * flags. */ +/* + * Convert a hierarchy specifier into a bitmask of subsystems and flags. Call + * with cgroup_mutex held to protect the subsys[] array. + */ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) { char *token, *o = data ?: "all"; unsigned long mask = (unsigned long)-1; + BUG_ON(!mutex_is_locked(&cgroup_mutex)); + #ifdef CONFIG_CPUSETS mask = ~(1UL << cpuset_subsys_id); #endif @@ -1009,6 +1033,8 @@ static int parse_cgroupfs_options(char *data, opts->subsys_bits = 0; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; + if (ss == NULL) + continue; if (!ss->disabled) opts->subsys_bits |= 1ul << i; } @@ -1053,6 +1079,8 @@ static int parse_cgroupfs_options(char *data, int i; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { ss = subsys[i]; + if (ss == NULL) + continue; if (!strcmp(token, ss->name)) { if (!ss->disabled) set_bit(i, &opts->subsys_bits); @@ -1306,7 +1334,9 @@ static int cgroup_get_sb(struct file_system_type *fs_type, struct cgroupfs_root *new_root; /* First find the desired set of subsystems */ + mutex_lock(&cgroup_mutex); ret = parse_cgroupfs_options(data, &opts); + mutex_unlock(&cgroup_mutex); if (ret) goto out_err; @@ -2918,8 +2948,14 @@ static void cgroup_lock_hierarchy(struct cgroupfs_root *root) /* We need to take each hierarchy_mutex in a consistent order */ int i; + /* + * No worry about a race with rebind_subsystems that might mess up the + * locking order, since both parties are under cgroup_mutex. + */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; + if (ss == NULL) + continue; if (ss->root == root) mutex_lock(&ss->hierarchy_mutex); } @@ -2931,6 +2967,8 @@ static void cgroup_unlock_hierarchy(struct cgroupfs_root *root) for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; + if (ss == NULL) + continue; if (ss->root == root) mutex_unlock(&ss->hierarchy_mutex); } @@ -3054,11 +3092,16 @@ static int cgroup_has_css_refs(struct cgroup *cgrp) * synchronization other than RCU, and the subsystem linked * list isn't RCU-safe */ int i; + /* + * We won't need to lock the subsys array, because the subsystems + * we're concerned about aren't going anywhere since our cgroup root + * has a reference on them. + */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; struct cgroup_subsys_state *css; - /* Skip subsystems not in this hierarchy */ - if (ss->root != cgrp->root) + /* Skip subsystems not present or not in this hierarchy */ + if (ss == NULL || ss->root != cgrp->root) continue; css = cgrp->subsys[ss->subsys_id]; /* When called from check_for_release() it's possible @@ -3279,7 +3322,8 @@ int __init cgroup_init_early(void) for (i = 0; i < CSS_SET_TABLE_SIZE; i++) INIT_HLIST_HEAD(&css_set_table[i]); - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + /* at bootup time, we don't worry about modular subsystems */ + for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; BUG_ON(!ss->name); @@ -3314,7 +3358,8 @@ int __init cgroup_init(void) if (err) return err; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + /* at bootup time, we don't worry about modular subsystems */ + for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; if (!ss->early_init) cgroup_init_subsys(ss); @@ -3423,9 +3468,16 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) int i; seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\tenabled\n"); + /* + * ideally we don't want subsystems moving around while we do this. + * cgroup_mutex is also necessary to guarantee an atomic snapshot of + * subsys/hierarchy state. + */ mutex_lock(&cgroup_mutex); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; + if (ss == NULL) + continue; seq_printf(m, "%s\t%d\t%d\t%d\n", ss->name, ss->root->hierarchy_id, ss->root->number_of_cgroups, !ss->disabled); @@ -3483,7 +3535,12 @@ void cgroup_fork_callbacks(struct task_struct *child) { if (need_forkexit_callback) { int i; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + /* + * forkexit callbacks are only supported for builtin + * subsystems, and the builtin section of the subsys array is + * immutable, so we don't need to lock the subsys array here. + */ + for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; if (ss->fork) ss->fork(ss, child); @@ -3552,7 +3609,11 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) struct css_set *cg; if (run_callbacks && need_forkexit_callback) { - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + /* + * modular subsystems can't use callbacks, so no need to lock + * the subsys array + */ + for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; if (ss->exit) ss->exit(ss, tsk); @@ -3844,8 +3905,11 @@ static int __init cgroup_disable(char *str) while ((token = strsep(&str, ",")) != NULL) { if (!*token) continue; - - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + /* + * cgroup_disable, being at boot time, can't know about module + * subsystems, so we don't worry about them. + */ + for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; if (!strcmp(token, ss->name)) { -- cgit v1.2.2 From e6a1105ba08b265023dd71a4174fb4a29ebc7083 Mon Sep 17 00:00:00 2001 From: Ben Blum Date: Wed, 10 Mar 2010 15:22:09 -0800 Subject: cgroups: subsystem module loading interface Add interface between cgroups subsystem management and module loading This patch implements rudimentary module-loading support for cgroups - namely, a cgroup_load_subsys (similar to cgroup_init_subsys) for use as a module initcall, and a struct module pointer in struct cgroup_subsys. Several functions that might be wanted by modules have had EXPORT_SYMBOL added to them, but it's unclear exactly which functions want it and which won't. Signed-off-by: Ben Blum Acked-by: Li Zefan Cc: Paul Menage Cc: "David S. Miller" Cc: KAMEZAWA Hiroyuki Cc: Lai Jiangshan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 150 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 145 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c92fb9549358..2cae38e64c59 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -44,6 +44,7 @@ #include #include #include +#include #include #include #include @@ -254,7 +255,8 @@ struct cg_cgroup_link { static struct css_set init_css_set; static struct cg_cgroup_link init_css_set_link; -static int cgroup_subsys_init_idr(struct cgroup_subsys *ss); +static int cgroup_init_idr(struct cgroup_subsys *ss, + struct cgroup_subsys_state *css); /* css_set_lock protects the list of css_set objects, and the * chain of tasks off each css_set. Nests outside task->alloc_lock @@ -2125,6 +2127,7 @@ int cgroup_add_file(struct cgroup *cgrp, error = PTR_ERR(dentry); return error; } +EXPORT_SYMBOL_GPL(cgroup_add_file); int cgroup_add_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, @@ -2139,6 +2142,7 @@ int cgroup_add_files(struct cgroup *cgrp, } return 0; } +EXPORT_SYMBOL_GPL(cgroup_add_files); /** * cgroup_task_count - count the number of tasks in a cgroup. @@ -3292,7 +3296,144 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) mutex_init(&ss->hierarchy_mutex); lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key); ss->active = 1; + + /* this function shouldn't be used with modular subsystems, since they + * need to register a subsys_id, among other things */ + BUG_ON(ss->module); +} + +/** + * cgroup_load_subsys: load and register a modular subsystem at runtime + * @ss: the subsystem to load + * + * This function should be called in a modular subsystem's initcall. If the + * subsytem is built as a module, it will be assigned a new subsys_id and set + * up for use. If the subsystem is built-in anyway, work is delegated to the + * simpler cgroup_init_subsys. + */ +int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) +{ + int i; + struct cgroup_subsys_state *css; + + /* check name and function validity */ + if (ss->name == NULL || strlen(ss->name) > MAX_CGROUP_TYPE_NAMELEN || + ss->create == NULL || ss->destroy == NULL) + return -EINVAL; + + /* + * we don't support callbacks in modular subsystems. this check is + * before the ss->module check for consistency; a subsystem that could + * be a module should still have no callbacks even if the user isn't + * compiling it as one. + */ + if (ss->fork || ss->exit) + return -EINVAL; + + /* + * an optionally modular subsystem is built-in: we want to do nothing, + * since cgroup_init_subsys will have already taken care of it. + */ + if (ss->module == NULL) { + /* a few sanity checks */ + BUG_ON(ss->subsys_id >= CGROUP_BUILTIN_SUBSYS_COUNT); + BUG_ON(subsys[ss->subsys_id] != ss); + return 0; + } + + /* + * need to register a subsys id before anything else - for example, + * init_cgroup_css needs it. + */ + mutex_lock(&cgroup_mutex); + /* find the first empty slot in the array */ + for (i = CGROUP_BUILTIN_SUBSYS_COUNT; i < CGROUP_SUBSYS_COUNT; i++) { + if (subsys[i] == NULL) + break; + } + if (i == CGROUP_SUBSYS_COUNT) { + /* maximum number of subsystems already registered! */ + mutex_unlock(&cgroup_mutex); + return -EBUSY; + } + /* assign ourselves the subsys_id */ + ss->subsys_id = i; + subsys[i] = ss; + + /* + * no ss->create seems to need anything important in the ss struct, so + * this can happen first (i.e. before the rootnode attachment). + */ + css = ss->create(ss, dummytop); + if (IS_ERR(css)) { + /* failure case - need to deassign the subsys[] slot. */ + subsys[i] = NULL; + mutex_unlock(&cgroup_mutex); + return PTR_ERR(css); + } + + list_add(&ss->sibling, &rootnode.subsys_list); + ss->root = &rootnode; + + /* our new subsystem will be attached to the dummy hierarchy. */ + init_cgroup_css(css, ss, dummytop); + /* init_idr must be after init_cgroup_css because it sets css->id. */ + if (ss->use_id) { + int ret = cgroup_init_idr(ss, css); + if (ret) { + dummytop->subsys[ss->subsys_id] = NULL; + ss->destroy(ss, dummytop); + subsys[i] = NULL; + mutex_unlock(&cgroup_mutex); + return ret; + } + } + + /* + * Now we need to entangle the css into the existing css_sets. unlike + * in cgroup_init_subsys, there are now multiple css_sets, so each one + * will need a new pointer to it; done by iterating the css_set_table. + * furthermore, modifying the existing css_sets will corrupt the hash + * table state, so each changed css_set will need its hash recomputed. + * this is all done under the css_set_lock. + */ + write_lock(&css_set_lock); + for (i = 0; i < CSS_SET_TABLE_SIZE; i++) { + struct css_set *cg; + struct hlist_node *node, *tmp; + struct hlist_head *bucket = &css_set_table[i], *new_bucket; + + hlist_for_each_entry_safe(cg, node, tmp, bucket, hlist) { + /* skip entries that we already rehashed */ + if (cg->subsys[ss->subsys_id]) + continue; + /* remove existing entry */ + hlist_del(&cg->hlist); + /* set new value */ + cg->subsys[ss->subsys_id] = css; + /* recompute hash and restore entry */ + new_bucket = css_set_hash(cg->subsys); + hlist_add_head(&cg->hlist, new_bucket); + } + } + write_unlock(&css_set_lock); + + mutex_init(&ss->hierarchy_mutex); + lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key); + ss->active = 1; + + /* + * pin the subsystem's module so it doesn't go away. this shouldn't + * fail, since the module's initcall calls us. + * TODO: with module unloading, move this elsewhere + */ + BUG_ON(!try_module_get(ss->module)); + + /* success! */ + mutex_unlock(&cgroup_mutex); + return 0; } +EXPORT_SYMBOL_GPL(cgroup_load_subsys); /** * cgroup_init_early - cgroup initialization at system boot @@ -3364,7 +3505,7 @@ int __init cgroup_init(void) if (!ss->early_init) cgroup_init_subsys(ss); if (ss->use_id) - cgroup_subsys_init_idr(ss); + cgroup_init_idr(ss, init_css_set.subsys[ss->subsys_id]); } /* Add init_css_set to the hash table */ @@ -4033,15 +4174,14 @@ err_out: } -static int __init cgroup_subsys_init_idr(struct cgroup_subsys *ss) +static int __init_or_module cgroup_init_idr(struct cgroup_subsys *ss, + struct cgroup_subsys_state *rootcss) { struct css_id *newid; - struct cgroup_subsys_state *rootcss; spin_lock_init(&ss->id_lock); idr_init(&ss->idr); - rootcss = init_css_set.subsys[ss->subsys_id]; newid = get_new_cssid(ss, 0); if (IS_ERR(newid)) return PTR_ERR(newid); -- cgit v1.2.2 From cf5d5941fda647fe3d2f2d00cf9e0245236a5f08 Mon Sep 17 00:00:00 2001 From: Ben Blum Date: Wed, 10 Mar 2010 15:22:09 -0800 Subject: cgroups: subsystem module unloading Provides support for unloading modular subsystems. This patch adds a new function cgroup_unload_subsys which is to be used for removing a loaded subsystem during module deletion. Reference counting of the subsystems' modules is moved from once (at load time) to once per attached hierarchy (in parse_cgroupfs_options and rebind_subsystems) (i.e., 0 or 1). Signed-off-by: Ben Blum Acked-by: Li Zefan Cc: Paul Menage Cc: "David S. Miller" Cc: KAMEZAWA Hiroyuki Cc: Lai Jiangshan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 167 +++++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 142 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2cae38e64c59..aa889c96cc74 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -894,7 +894,9 @@ void cgroup_release_and_wakeup_rmdir(struct cgroup_subsys_state *css) } /* - * Call with cgroup_mutex held. + * Call with cgroup_mutex held. Drops reference counts on modules, including + * any duplicate ones that parse_cgroupfs_options took. If this function + * returns an error, no reference counts are touched. */ static int rebind_subsystems(struct cgroupfs_root *root, unsigned long final_bits) @@ -950,6 +952,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, if (ss->bind) ss->bind(ss, cgrp); mutex_unlock(&ss->hierarchy_mutex); + /* refcount was already taken, and we're keeping it */ } else if (bit & removed_bits) { /* We're removing this subsystem */ BUG_ON(ss == NULL); @@ -963,10 +966,20 @@ static int rebind_subsystems(struct cgroupfs_root *root, subsys[i]->root = &rootnode; list_move(&ss->sibling, &rootnode.subsys_list); mutex_unlock(&ss->hierarchy_mutex); + /* subsystem is now free - drop reference on module */ + module_put(ss->module); } else if (bit & final_bits) { /* Subsystem state should already exist */ BUG_ON(ss == NULL); BUG_ON(!cgrp->subsys[i]); + /* + * a refcount was taken, but we already had one, so + * drop the extra reference. + */ + module_put(ss->module); +#ifdef CONFIG_MODULE_UNLOAD + BUG_ON(ss->module && !module_refcount(ss->module)); +#endif } else { /* Subsystem state shouldn't exist */ BUG_ON(cgrp->subsys[i]); @@ -1010,13 +1023,16 @@ struct cgroup_sb_opts { /* * Convert a hierarchy specifier into a bitmask of subsystems and flags. Call - * with cgroup_mutex held to protect the subsys[] array. + * with cgroup_mutex held to protect the subsys[] array. This function takes + * refcounts on subsystems to be used, unless it returns error, in which case + * no refcounts are taken. */ -static int parse_cgroupfs_options(char *data, - struct cgroup_sb_opts *opts) +static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) { char *token, *o = data ?: "all"; unsigned long mask = (unsigned long)-1; + int i; + bool module_pin_failed = false; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -1031,7 +1047,6 @@ static int parse_cgroupfs_options(char *data, return -EINVAL; if (!strcmp(token, "all")) { /* Add all non-disabled subsystems */ - int i; opts->subsys_bits = 0; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; @@ -1054,7 +1069,6 @@ static int parse_cgroupfs_options(char *data, if (!opts->release_agent) return -ENOMEM; } else if (!strncmp(token, "name=", 5)) { - int i; const char *name = token + 5; /* Can't specify an empty name */ if (!strlen(name)) @@ -1078,7 +1092,6 @@ static int parse_cgroupfs_options(char *data, return -ENOMEM; } else { struct cgroup_subsys *ss; - int i; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { ss = subsys[i]; if (ss == NULL) @@ -1117,9 +1130,54 @@ static int parse_cgroupfs_options(char *data, if (!opts->subsys_bits && !opts->name) return -EINVAL; + /* + * Grab references on all the modules we'll need, so the subsystems + * don't dance around before rebind_subsystems attaches them. This may + * take duplicate reference counts on a subsystem that's already used, + * but rebind_subsystems handles this case. + */ + for (i = CGROUP_BUILTIN_SUBSYS_COUNT; i < CGROUP_SUBSYS_COUNT; i++) { + unsigned long bit = 1UL << i; + + if (!(bit & opts->subsys_bits)) + continue; + if (!try_module_get(subsys[i]->module)) { + module_pin_failed = true; + break; + } + } + if (module_pin_failed) { + /* + * oops, one of the modules was going away. this means that we + * raced with a module_delete call, and to the user this is + * essentially a "subsystem doesn't exist" case. + */ + for (i--; i >= CGROUP_BUILTIN_SUBSYS_COUNT; i--) { + /* drop refcounts only on the ones we took */ + unsigned long bit = 1UL << i; + + if (!(bit & opts->subsys_bits)) + continue; + module_put(subsys[i]->module); + } + return -ENOENT; + } + return 0; } +static void drop_parsed_module_refcounts(unsigned long subsys_bits) +{ + int i; + for (i = CGROUP_BUILTIN_SUBSYS_COUNT; i < CGROUP_SUBSYS_COUNT; i++) { + unsigned long bit = 1UL << i; + + if (!(bit & subsys_bits)) + continue; + module_put(subsys[i]->module); + } +} + static int cgroup_remount(struct super_block *sb, int *flags, char *data) { int ret = 0; @@ -1136,21 +1194,19 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) if (ret) goto out_unlock; - /* Don't allow flags to change at remount */ - if (opts.flags != root->flags) { - ret = -EINVAL; - goto out_unlock; - } - - /* Don't allow name to change at remount */ - if (opts.name && strcmp(opts.name, root->name)) { + /* Don't allow flags or name to change at remount */ + if (opts.flags != root->flags || + (opts.name && strcmp(opts.name, root->name))) { ret = -EINVAL; + drop_parsed_module_refcounts(opts.subsys_bits); goto out_unlock; } ret = rebind_subsystems(root, opts.subsys_bits); - if (ret) + if (ret) { + drop_parsed_module_refcounts(opts.subsys_bits); goto out_unlock; + } /* (re)populate subsystem files */ cgroup_populate_dir(cgrp); @@ -1349,7 +1405,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, new_root = cgroup_root_from_opts(&opts); if (IS_ERR(new_root)) { ret = PTR_ERR(new_root); - goto out_err; + goto drop_modules; } opts.new_root = new_root; @@ -1358,7 +1414,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, if (IS_ERR(sb)) { ret = PTR_ERR(sb); cgroup_drop_root(opts.new_root); - goto out_err; + goto drop_modules; } root = sb->s_fs_info; @@ -1414,6 +1470,11 @@ static int cgroup_get_sb(struct file_system_type *fs_type, free_cg_links(&tmp_cg_links); goto drop_new_super; } + /* + * There must be no failure case after here, since rebinding + * takes care of subsystems' refcounts, which are explicitly + * dropped in the failure exit path. + */ /* EBUSY should be the only error here */ BUG_ON(ret); @@ -1452,6 +1513,8 @@ static int cgroup_get_sb(struct file_system_type *fs_type, * any) is not needed */ cgroup_drop_root(opts.new_root); + /* no subsys rebinding, so refcounts don't change */ + drop_parsed_module_refcounts(opts.subsys_bits); } simple_set_mnt(mnt, sb); @@ -1461,6 +1524,8 @@ static int cgroup_get_sb(struct file_system_type *fs_type, drop_new_super: deactivate_locked_super(sb); + drop_modules: + drop_parsed_module_refcounts(opts.subsys_bits); out_err: kfree(opts.release_agent); kfree(opts.name); @@ -3422,19 +3487,71 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key); ss->active = 1; - /* - * pin the subsystem's module so it doesn't go away. this shouldn't - * fail, since the module's initcall calls us. - * TODO: with module unloading, move this elsewhere - */ - BUG_ON(!try_module_get(ss->module)); - /* success! */ mutex_unlock(&cgroup_mutex); return 0; } EXPORT_SYMBOL_GPL(cgroup_load_subsys); +/** + * cgroup_unload_subsys: unload a modular subsystem + * @ss: the subsystem to unload + * + * This function should be called in a modular subsystem's exitcall. When this + * function is invoked, the refcount on the subsystem's module will be 0, so + * the subsystem will not be attached to any hierarchy. + */ +void cgroup_unload_subsys(struct cgroup_subsys *ss) +{ + struct cg_cgroup_link *link; + struct hlist_head *hhead; + + BUG_ON(ss->module == NULL); + + /* + * we shouldn't be called if the subsystem is in use, and the use of + * try_module_get in parse_cgroupfs_options should ensure that it + * doesn't start being used while we're killing it off. + */ + BUG_ON(ss->root != &rootnode); + + mutex_lock(&cgroup_mutex); + /* deassign the subsys_id */ + BUG_ON(ss->subsys_id < CGROUP_BUILTIN_SUBSYS_COUNT); + subsys[ss->subsys_id] = NULL; + + /* remove subsystem from rootnode's list of subsystems */ + list_del(&ss->sibling); + + /* + * disentangle the css from all css_sets attached to the dummytop. as + * in loading, we need to pay our respects to the hashtable gods. + */ + write_lock(&css_set_lock); + list_for_each_entry(link, &dummytop->css_sets, cgrp_link_list) { + struct css_set *cg = link->cg; + + hlist_del(&cg->hlist); + BUG_ON(!cg->subsys[ss->subsys_id]); + cg->subsys[ss->subsys_id] = NULL; + hhead = css_set_hash(cg->subsys); + hlist_add_head(&cg->hlist, hhead); + } + write_unlock(&css_set_lock); + + /* + * remove subsystem's css from the dummytop and free it - need to free + * before marking as null because ss->destroy needs the cgrp->subsys + * pointer to find their state. note that this also takes care of + * freeing the css_id. + */ + ss->destroy(ss, dummytop); + dummytop->subsys[ss->subsys_id] = NULL; + + mutex_unlock(&cgroup_mutex); +} +EXPORT_SYMBOL_GPL(cgroup_unload_subsys); + /** * cgroup_init_early - cgroup initialization at system boot * -- cgit v1.2.2 From 67523c48aa74d5637848edeccf285af1c60bf14a Mon Sep 17 00:00:00 2001 From: Ben Blum Date: Wed, 10 Mar 2010 15:22:11 -0800 Subject: cgroups: blkio subsystem as module Modify the Block I/O cgroup subsystem to be able to be built as a module. As the CFQ disk scheduler optionally depends on blk-cgroup, config options in block/Kconfig, block/Kconfig.iosched, and block/blk-cgroup.h are enhanced to support the new module dependency. Signed-off-by: Ben Blum Cc: Li Zefan Cc: Paul Menage Cc: "David S. Miller" Cc: KAMEZAWA Hiroyuki Cc: Lai Jiangshan Cc: Vivek Goyal Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index aa889c96cc74..521591dbab2f 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -705,6 +705,7 @@ void cgroup_lock(void) { mutex_lock(&cgroup_mutex); } +EXPORT_SYMBOL_GPL(cgroup_lock); /** * cgroup_unlock - release lock on cgroup changes @@ -715,6 +716,7 @@ void cgroup_unlock(void) { mutex_unlock(&cgroup_mutex); } +EXPORT_SYMBOL_GPL(cgroup_unlock); /* * A couple of forward declarations required, due to cyclic reference loop: @@ -1639,6 +1641,7 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) memmove(buf, start, buf + buflen - start); return 0; } +EXPORT_SYMBOL_GPL(cgroup_path); /** * cgroup_attach_task - attach task 'tsk' to cgroup 'cgrp' @@ -1805,6 +1808,7 @@ bool cgroup_lock_live_group(struct cgroup *cgrp) } return true; } +EXPORT_SYMBOL_GPL(cgroup_lock_live_group); static int cgroup_release_agent_write(struct cgroup *cgrp, struct cftype *cft, const char *buffer) @@ -4082,6 +4086,7 @@ void __css_put(struct cgroup_subsys_state *css, int count) rcu_read_unlock(); WARN_ON_ONCE(val < 1); } +EXPORT_SYMBOL_GPL(__css_put); /* * Notify userspace when a cgroup is released, by running the @@ -4197,6 +4202,7 @@ unsigned short css_id(struct cgroup_subsys_state *css) return cssid->id; return 0; } +EXPORT_SYMBOL_GPL(css_id); unsigned short css_depth(struct cgroup_subsys_state *css) { @@ -4206,6 +4212,7 @@ unsigned short css_depth(struct cgroup_subsys_state *css) return cssid->depth; return 0; } +EXPORT_SYMBOL_GPL(css_depth); bool css_is_ancestor(struct cgroup_subsys_state *child, const struct cgroup_subsys_state *root) @@ -4242,6 +4249,7 @@ void free_css_id(struct cgroup_subsys *ss, struct cgroup_subsys_state *css) spin_unlock(&ss->id_lock); call_rcu(&id->rcu_head, __free_css_id_cb); } +EXPORT_SYMBOL_GPL(free_css_id); /* * This is called by init or create(). Then, calls to this function are @@ -4358,6 +4366,7 @@ struct cgroup_subsys_state *css_lookup(struct cgroup_subsys *ss, int id) return rcu_dereference(cssid->css); } +EXPORT_SYMBOL_GPL(css_lookup); /** * css_get_next - lookup next cgroup under specified hierarchy. -- cgit v1.2.2 From b70cc5fdb445a6929a01e9c406593265b136c99d Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 10 Mar 2010 15:22:12 -0800 Subject: cgroups: clean up cgroup_pidlist_find() a bit Don't call get_pid_ns() before we locate/alloc the ns. Signed-off-by: Li Zefan Cc: Serge Hallyn Acked-by: Paul Menage Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 521591dbab2f..1bf4d6db54ab 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2597,7 +2597,8 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, { struct cgroup_pidlist *l; /* don't need task_nsproxy() if we're looking at ourself */ - struct pid_namespace *ns = get_pid_ns(current->nsproxy->pid_ns); + struct pid_namespace *ns = current->nsproxy->pid_ns; + /* * We can't drop the pidlist_mutex before taking the l->mutex in case * the last ref-holder is trying to remove l from the list at the same @@ -2607,8 +2608,6 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, mutex_lock(&cgrp->pidlist_mutex); list_for_each_entry(l, &cgrp->pidlists, links) { if (l->key.type == type && l->key.ns == ns) { - /* found a matching list - drop the extra refcount */ - put_pid_ns(ns); /* make sure l doesn't vanish out from under us */ down_write(&l->mutex); mutex_unlock(&cgrp->pidlist_mutex); @@ -2619,13 +2618,12 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, l = kmalloc(sizeof(struct cgroup_pidlist), GFP_KERNEL); if (!l) { mutex_unlock(&cgrp->pidlist_mutex); - put_pid_ns(ns); return l; } init_rwsem(&l->mutex); down_write(&l->mutex); l->key.type = type; - l->key.ns = ns; + l->key.ns = get_pid_ns(ns); l->use_count = 0; /* don't increment here */ l->list = NULL; l->owner = cgrp; -- cgit v1.2.2 From 0dea116876eefc9c7ca9c5d74fe665481e499fa3 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Wed, 10 Mar 2010 15:22:20 -0800 Subject: cgroup: implement eventfd-based generic API for notifications This patchset introduces eventfd-based API for notifications in cgroups and implements memory notifications on top of it. It uses statistics in memory controler to track memory usage. Output of time(1) on building kernel on tmpfs: Root cgroup before changes: make -j2 506.37 user 60.93s system 193% cpu 4:52.77 total Non-root cgroup before changes: make -j2 507.14 user 62.66s system 193% cpu 4:54.74 total Root cgroup after changes (0 thresholds): make -j2 507.13 user 62.20s system 193% cpu 4:53.55 total Non-root cgroup after changes (0 thresholds): make -j2 507.70 user 64.20s system 193% cpu 4:55.70 total Root cgroup after changes (1 thresholds, never crossed): make -j2 506.97 user 62.20s system 193% cpu 4:53.90 total Non-root cgroup after changes (1 thresholds, never crossed): make -j2 507.55 user 64.08s system 193% cpu 4:55.63 total This patch: Introduce the write-only file "cgroup.event_control" in every cgroup. To register new notification handler you need: - create an eventfd; - open a control file to be monitored. Callbacks register_event() and unregister_event() must be defined for the control file; - write " " to cgroup.event_control. Interpretation of args is defined by control file implementation; eventfd will be woken up by control file implementation or when the cgroup is removed. To unregister notification handler just close eventfd. If you need notification functionality for a control file you have to implement callbacks register_event() and unregister_event() in the struct cftype. [kamezawa.hiroyu@jp.fujitsu.com: Kconfig fix] Signed-off-by: Kirill A. Shutemov Reviewed-by: KAMEZAWA Hiroyuki Paul Menage Cc: Li Zefan Cc: Balbir Singh Cc: Pavel Emelyanov Cc: Dan Malek Cc: Vladislav Buzov Cc: Daisuke Nishimura Cc: Alexander Shishkin Cc: Davide Libenzi Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 228 +++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 227 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1bf4d6db54ab..ea94984a3895 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4,6 +4,10 @@ * Based originally on the cpuset system, extracted by Paul Menage * Copyright (C) 2006 Google, Inc * + * Notifications support + * Copyright (C) 2009 Nokia Corporation + * Author: Kirill A. Shutemov + * * Copyright notices from the original cpuset code: * -------------------------------------------------- * Copyright (C) 2003 BULL SA. @@ -53,6 +57,8 @@ #include #include #include /* TODO: replace with more sophisticated array */ +#include +#include #include @@ -152,6 +158,35 @@ struct css_id { unsigned short stack[0]; /* Array of Length (depth+1) */ }; +/* + * cgroup_event represents events which userspace want to recieve. + */ +struct cgroup_event { + /* + * Cgroup which the event belongs to. + */ + struct cgroup *cgrp; + /* + * Control file which the event associated. + */ + struct cftype *cft; + /* + * eventfd to signal userspace about the event. + */ + struct eventfd_ctx *eventfd; + /* + * Each of these stored in a list by the cgroup. + */ + struct list_head list; + /* + * All fields below needed to unregister event when + * userspace closes eventfd. + */ + poll_table pt; + wait_queue_head_t *wqh; + wait_queue_t wait; + struct work_struct remove; +}; /* The list of hierarchy roots */ @@ -760,14 +795,28 @@ static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) static int cgroup_call_pre_destroy(struct cgroup *cgrp) { struct cgroup_subsys *ss; + struct cgroup_event *event, *tmp; int ret = 0; for_each_subsys(cgrp->root, ss) if (ss->pre_destroy) { ret = ss->pre_destroy(ss, cgrp); if (ret) - break; + goto out; } + + /* + * Unregister events and notify userspace. + */ + spin_lock(&cgrp->event_list_lock); + list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) { + list_del(&event->list); + eventfd_signal(event->eventfd, 1); + schedule_work(&event->remove); + } + spin_unlock(&cgrp->event_list_lock); + +out: return ret; } @@ -1239,6 +1288,8 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); mutex_init(&cgrp->pidlist_mutex); + INIT_LIST_HEAD(&cgrp->event_list); + spin_lock_init(&cgrp->event_list_lock); } static void init_cgroup_root(struct cgroupfs_root *root) @@ -2077,6 +2128,16 @@ static const struct inode_operations cgroup_dir_inode_operations = { .rename = cgroup_rename, }; +/* + * Check if a file is a control file + */ +static inline struct cftype *__file_cft(struct file *file) +{ + if (file->f_dentry->d_inode->i_fop != &cgroup_file_operations) + return ERR_PTR(-EINVAL); + return __d_cft(file->f_dentry); +} + static int cgroup_create_file(struct dentry *dentry, mode_t mode, struct super_block *sb) { @@ -2930,6 +2991,166 @@ static int cgroup_write_notify_on_release(struct cgroup *cgrp, return 0; } +/* + * Unregister event and free resources. + * + * Gets called from workqueue. + */ +static void cgroup_event_remove(struct work_struct *work) +{ + struct cgroup_event *event = container_of(work, struct cgroup_event, + remove); + struct cgroup *cgrp = event->cgrp; + + /* TODO: check return code */ + event->cft->unregister_event(cgrp, event->cft, event->eventfd); + + eventfd_ctx_put(event->eventfd); + remove_wait_queue(event->wqh, &event->wait); + kfree(event); +} + +/* + * Gets called on POLLHUP on eventfd when user closes it. + * + * Called with wqh->lock held and interrupts disabled. + */ +static int cgroup_event_wake(wait_queue_t *wait, unsigned mode, + int sync, void *key) +{ + struct cgroup_event *event = container_of(wait, + struct cgroup_event, wait); + struct cgroup *cgrp = event->cgrp; + unsigned long flags = (unsigned long)key; + + if (flags & POLLHUP) { + spin_lock(&cgrp->event_list_lock); + list_del(&event->list); + spin_unlock(&cgrp->event_list_lock); + /* + * We are in atomic context, but cgroup_event_remove() may + * sleep, so we have to call it in workqueue. + */ + schedule_work(&event->remove); + } + + return 0; +} + +static void cgroup_event_ptable_queue_proc(struct file *file, + wait_queue_head_t *wqh, poll_table *pt) +{ + struct cgroup_event *event = container_of(pt, + struct cgroup_event, pt); + + event->wqh = wqh; + add_wait_queue(wqh, &event->wait); +} + +/* + * Parse input and register new cgroup event handler. + * + * Input must be in format ' '. + * Interpretation of args is defined by control file implementation. + */ +static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, + const char *buffer) +{ + struct cgroup_event *event = NULL; + unsigned int efd, cfd; + struct file *efile = NULL; + struct file *cfile = NULL; + char *endp; + int ret; + + efd = simple_strtoul(buffer, &endp, 10); + if (*endp != ' ') + return -EINVAL; + buffer = endp + 1; + + cfd = simple_strtoul(buffer, &endp, 10); + if ((*endp != ' ') && (*endp != '\0')) + return -EINVAL; + buffer = endp + 1; + + event = kzalloc(sizeof(*event), GFP_KERNEL); + if (!event) + return -ENOMEM; + event->cgrp = cgrp; + INIT_LIST_HEAD(&event->list); + init_poll_funcptr(&event->pt, cgroup_event_ptable_queue_proc); + init_waitqueue_func_entry(&event->wait, cgroup_event_wake); + INIT_WORK(&event->remove, cgroup_event_remove); + + efile = eventfd_fget(efd); + if (IS_ERR(efile)) { + ret = PTR_ERR(efile); + goto fail; + } + + event->eventfd = eventfd_ctx_fileget(efile); + if (IS_ERR(event->eventfd)) { + ret = PTR_ERR(event->eventfd); + goto fail; + } + + cfile = fget(cfd); + if (!cfile) { + ret = -EBADF; + goto fail; + } + + /* the process need read permission on control file */ + ret = file_permission(cfile, MAY_READ); + if (ret < 0) + goto fail; + + event->cft = __file_cft(cfile); + if (IS_ERR(event->cft)) { + ret = PTR_ERR(event->cft); + goto fail; + } + + if (!event->cft->register_event || !event->cft->unregister_event) { + ret = -EINVAL; + goto fail; + } + + ret = event->cft->register_event(cgrp, event->cft, + event->eventfd, buffer); + if (ret) + goto fail; + + if (efile->f_op->poll(efile, &event->pt) & POLLHUP) { + event->cft->unregister_event(cgrp, event->cft, event->eventfd); + ret = 0; + goto fail; + } + + spin_lock(&cgrp->event_list_lock); + list_add(&event->list, &cgrp->event_list); + spin_unlock(&cgrp->event_list_lock); + + fput(cfile); + fput(efile); + + return 0; + +fail: + if (cfile) + fput(cfile); + + if (event && event->eventfd && !IS_ERR(event->eventfd)) + eventfd_ctx_put(event->eventfd); + + if (!IS_ERR_OR_NULL(efile)) + fput(efile); + + kfree(event); + + return ret; +} + /* * for the common functions, 'private' gives the type of file */ @@ -2955,6 +3176,11 @@ static struct cftype files[] = { .read_u64 = cgroup_read_notify_on_release, .write_u64 = cgroup_write_notify_on_release, }, + { + .name = CGROUP_FILE_GENERIC_PREFIX "event_control", + .write_string = cgroup_write_event_control, + .mode = S_IWUGO, + }, }; static struct cftype cft_release_agent = { -- cgit v1.2.2 From 4ab78683c17d739c2a2077141dcf81a02b7fb57e Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Wed, 10 Mar 2010 15:22:34 -0800 Subject: cgroups: fix race between userspace and kernelspace Notify userspace about cgroup removing only after rmdir of cgroup directory to avoid race between userspace and kernelspace. eventfd are used to notify about two types of event: - control file-specific, like crossing memory threshold; - cgroup removing. To understand what really happen, userspace can check if the cgroup still exists. To avoid race beetween userspace and kernelspace we have to notify userspace about cgroup removing only after rmdir of cgroup directory. Signed-off-by: Kirill A. Shutemov Reviewed-by: KAMEZAWA Hiroyuki Cc: Paul Menage Acked-by: Li Zefan Cc: Balbir Singh Cc: Pavel Emelyanov Cc: Dan Malek Cc: Daisuke Nishimura Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 32 +++++++++++++++++--------------- 1 file changed, 17 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ea94984a3895..87441fc75663 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -795,28 +795,15 @@ static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) static int cgroup_call_pre_destroy(struct cgroup *cgrp) { struct cgroup_subsys *ss; - struct cgroup_event *event, *tmp; int ret = 0; for_each_subsys(cgrp->root, ss) if (ss->pre_destroy) { ret = ss->pre_destroy(ss, cgrp); if (ret) - goto out; + break; } - /* - * Unregister events and notify userspace. - */ - spin_lock(&cgrp->event_list_lock); - list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) { - list_del(&event->list); - eventfd_signal(event->eventfd, 1); - schedule_work(&event->remove); - } - spin_unlock(&cgrp->event_list_lock); - -out: return ret; } @@ -3006,7 +2993,6 @@ static void cgroup_event_remove(struct work_struct *work) event->cft->unregister_event(cgrp, event->cft, event->eventfd); eventfd_ctx_put(event->eventfd); - remove_wait_queue(event->wqh, &event->wait); kfree(event); } @@ -3024,6 +3010,7 @@ static int cgroup_event_wake(wait_queue_t *wait, unsigned mode, unsigned long flags = (unsigned long)key; if (flags & POLLHUP) { + remove_wait_queue_locked(event->wqh, &event->wait); spin_lock(&cgrp->event_list_lock); list_del(&event->list); spin_unlock(&cgrp->event_list_lock); @@ -3472,6 +3459,7 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) struct dentry *d; struct cgroup *parent; DEFINE_WAIT(wait); + struct cgroup_event *event, *tmp; int ret; /* the vfs holds both inode->i_mutex already */ @@ -3555,6 +3543,20 @@ again: set_bit(CGRP_RELEASABLE, &parent->flags); check_for_release(parent); + /* + * Unregister events and notify userspace. + * Notify userspace about cgroup removing only after rmdir of cgroup + * directory to avoid race between userspace and kernelspace + */ + spin_lock(&cgrp->event_list_lock); + list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) { + list_del(&event->list); + remove_wait_queue(event->wqh, &event->wait); + eventfd_signal(event->eventfd, 1); + schedule_work(&event->remove); + } + spin_unlock(&cgrp->event_list_lock); + mutex_unlock(&cgroup_mutex); return 0; } -- cgit v1.2.2 From a0a4db548edcce067c1201ef25cf2bc29f32dca4 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Wed, 10 Mar 2010 15:22:34 -0800 Subject: cgroups: remove events before destroying subsystem state objects Events should be removed after rmdir of cgroup directory, but before destroying subsystem state objects. Let's take reference to cgroup directory dentry to do that. Signed-off-by: Kirill A. Shutemov Acked-by: KAMEZAWA Hiroyuki Cc: Paul Menage Acked-by: Li Zefan Cc: Balbir Singh Cc: Pavel Emelyanov Cc: Dan Malek Cc: Daisuke Nishimura Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 87441fc75663..ef909a329750 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2994,6 +2994,7 @@ static void cgroup_event_remove(struct work_struct *work) eventfd_ctx_put(event->eventfd); kfree(event); + dput(cgrp->dentry); } /* @@ -3114,6 +3115,13 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, goto fail; } + /* + * Events should be removed after rmdir of cgroup directory, but before + * destroying subsystem state objects. Let's take reference to cgroup + * directory dentry to do that. + */ + dget(cgrp->dentry); + spin_lock(&cgrp->event_list_lock); list_add(&event->list, &cgrp->event_list); spin_unlock(&cgrp->event_list_lock); -- cgit v1.2.2 From a56704ef6b0c5796c9ff38cc78aa232dfb9644d7 Mon Sep 17 00:00:00 2001 From: Veaceslav Falico Date: Wed, 10 Mar 2010 15:23:01 -0800 Subject: copy_signal() cleanup: use zalloc and remove initializations Use kmem_cache_zalloc() on signal creation and remove unneeded initialization lines in copy_signal(). Signed-off-by: Veaceslav Falico Acked-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 27 +-------------------------- 1 file changed, 1 insertion(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index b0ec34abc0bb..ce2666f84d85 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -863,7 +863,7 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) if (clone_flags & CLONE_THREAD) return 0; - sig = kmem_cache_alloc(signal_cachep, GFP_KERNEL); + sig = kmem_cache_zalloc(signal_cachep, GFP_KERNEL); tsk->signal = sig; if (!sig) return -ENOMEM; @@ -871,46 +871,21 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) atomic_set(&sig->count, 1); atomic_set(&sig->live, 1); init_waitqueue_head(&sig->wait_chldexit); - sig->flags = 0; if (clone_flags & CLONE_NEWPID) sig->flags |= SIGNAL_UNKILLABLE; - sig->group_exit_code = 0; - sig->group_exit_task = NULL; - sig->group_stop_count = 0; sig->curr_target = tsk; init_sigpending(&sig->shared_pending); INIT_LIST_HEAD(&sig->posix_timers); hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); - sig->it_real_incr.tv64 = 0; sig->real_timer.function = it_real_fn; - sig->leader = 0; /* session leadership doesn't inherit */ - sig->tty_old_pgrp = NULL; - sig->tty = NULL; - - sig->utime = sig->stime = sig->cutime = sig->cstime = cputime_zero; - sig->gtime = cputime_zero; - sig->cgtime = cputime_zero; -#ifndef CONFIG_VIRT_CPU_ACCOUNTING - sig->prev_utime = sig->prev_stime = cputime_zero; -#endif - sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0; - sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0; - sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0; - sig->maxrss = sig->cmaxrss = 0; - task_io_accounting_init(&sig->ioac); - sig->sum_sched_runtime = 0; - taskstats_tgid_init(sig); - task_lock(current->group_leader); memcpy(sig->rlim, current->signal->rlim, sizeof sig->rlim); task_unlock(current->group_leader); posix_cpu_timers_init_group(sig); - acct_init_pacct(&sig->pacct); - tty_audit_fork(sig); sig->oom_adj = current->signal->oom_adj; -- cgit v1.2.2 From 4dd66e69d472f0ba5355a2529364d0db9a18a02b Mon Sep 17 00:00:00 2001 From: Veaceslav Falico Date: Wed, 10 Mar 2010 15:23:02 -0800 Subject: copy_signal() cleanup: kill taskstats_tgid_init() and acct_init_pacct() Kill unused functions taskstats_tgid_init() and acct_init_pacct() because we don't use them anywhere after using kmem_cache_zalloc() in copy_signal(). Signed-off-by: Veaceslav Falico Cc: Roland McGrath Cc: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/acct.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'kernel') diff --git a/kernel/acct.c b/kernel/acct.c index a6605ca921b6..24f8c81fc48d 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -587,16 +587,6 @@ out: revert_creds(orig_cred); } -/** - * acct_init_pacct - initialize a new pacct_struct - * @pacct: per-process accounting info struct to initialize - */ -void acct_init_pacct(struct pacct_struct *pacct) -{ - memset(pacct, 0, sizeof(struct pacct_struct)); - pacct->ac_utime = pacct->ac_stime = cputime_zero; -} - /** * acct_collect - collect accounting information into pacct_struct * @exitcode: task exit code -- cgit v1.2.2 From 93c59907c6f247d09239135caecf294a106a2ae0 Mon Sep 17 00:00:00 2001 From: Veaceslav Falico Date: Wed, 10 Mar 2010 15:23:03 -0800 Subject: copy_signal() cleanup: clean thread_group_cputime_init() Remove unneeded initializations in thread_group_cputime_init() and in posix_cpu_timers_init_group(). They are useless after kmem_cache_zalloc() was used in copy_signal(). Signed-off-by: Veaceslav Falico Acked-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 11 ----------- 1 file changed, 11 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index ce2666f84d85..1beb6c303c41 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -833,17 +833,6 @@ static void posix_cpu_timers_init_group(struct signal_struct *sig) /* Thread group counters. */ thread_group_cputime_init(sig); - /* Expiration times and increments. */ - sig->it[CPUCLOCK_PROF].expires = cputime_zero; - sig->it[CPUCLOCK_PROF].incr = cputime_zero; - sig->it[CPUCLOCK_VIRT].expires = cputime_zero; - sig->it[CPUCLOCK_VIRT].incr = cputime_zero; - - /* Cached expiration times. */ - sig->cputime_expires.prof_exp = cputime_zero; - sig->cputime_expires.virt_exp = cputime_zero; - sig->cputime_expires.sched_exp = 0; - cpu_limit = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur); if (cpu_limit != RLIM_INFINITY) { sig->cputime_expires.prof_exp = secs_to_cputime(cpu_limit); -- cgit v1.2.2 From 13aa9a6b0f2371d2ce0de57c2ede62ab7a787157 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 10 Mar 2010 15:23:09 -0800 Subject: pid_ns: zap_pid_ns_processes: use SEND_SIG_NOINFO instead of force_sig() zap_pid_ns_processes() uses force_sig(SIGKILL) to ensure SIGKILL will be delivered to sub-namespace inits as well. This is correct, but we are going to change force_sig_info() semantics. See http://bugzilla.kernel.org/show_bug.cgi?id=15395#c31 We can use send_sig_info(SEND_SIG_NOINFO) instead, since 614c517d7c00af1b26ded20646b329397d6f51a1 ("signals: SEND_SIG_NOINFO should be considered as SI_FROMUSER()") SEND_SIG_NOINFO means "from user" and therefore send_signal() will get the correct from_ancestor_ns = T flag. Signed-off-by: Oleg Nesterov Acked-by: Serge Hallyn Acked-by: Linus Torvalds Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid_namespace.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 86b3796b0436..79aac93acf99 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -161,13 +161,12 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) rcu_read_lock(); /* - * Use force_sig() since it clears SIGNAL_UNKILLABLE ensuring - * any nested-container's init processes don't ignore the - * signal + * Any nested-container's init processes won't ignore the + * SEND_SIG_NOINFO signal, see send_signal()->si_fromuser(). */ task = pid_task(find_vpid(nr), PIDTYPE_PID); if (task) - force_sig(SIGKILL, task); + send_sig_info(SIGKILL, SEND_SIG_NOINFO, task); rcu_read_unlock(); -- cgit v1.2.2 From 8467005da3ef6104b89a4cc5e9c9d9445b75565f Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Wed, 10 Mar 2010 15:23:10 -0800 Subject: nsproxy: remove INIT_NSPROXY() Remove INIT_NSPROXY(), use C99 initializer. Remove INIT_IPC_NS(), INIT_NET_NS() while I'm at it. Note: headers trim will be done later, now it's quite pointless because results will be invalidated by merge window. Signed-off-by: Alexey Dobriyan Acked-by: Serge Hallyn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/nsproxy.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 09b4ff9711b2..2ab67233ee8f 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -24,7 +24,18 @@ static struct kmem_cache *nsproxy_cachep; -struct nsproxy init_nsproxy = INIT_NSPROXY(init_nsproxy); +struct nsproxy init_nsproxy = { + .count = ATOMIC_INIT(1), + .uts_ns = &init_uts_ns, +#if defined(CONFIG_POSIX_MQUEUE) || defined(CONFIG_SYSVIPC) + .ipc_ns = &init_ipc_ns, +#endif + .mnt_ns = NULL, + .pid_ns = &init_pid_ns, +#ifdef CONFIG_NET + .net_ns = &init_net, +#endif +}; static inline struct nsproxy *create_nsproxy(void) { -- cgit v1.2.2 From eb5572fed55f4c2b7dbc42582bc82dcb47632380 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:23:59 -0800 Subject: sysctl extern cleanup: C_A_D Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move C_A_D extern variable declaration to linux/reboot.h Signed-off-by: Dave Young Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 0ef19c614f6d..72c3b1e80d7b 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -65,7 +65,6 @@ #if defined(CONFIG_SYSCTL) /* External variables not in a header file. */ -extern int C_A_D; extern int print_fatal_signals; extern int sysctl_overcommit_memory; extern int sysctl_overcommit_ratio; -- cgit v1.2.2 From d33ed52d57e794eba55cea3f5eab3c8f80b6cb5a Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:23:59 -0800 Subject: sysctl extern cleanup: signal Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move print_fatal_signals extern declaration to linux/signal.h Signed-off-by: Dave Young Cc: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 72c3b1e80d7b..a8fd10a9a501 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include #include @@ -65,7 +66,6 @@ #if defined(CONFIG_SYSCTL) /* External variables not in a header file. */ -extern int print_fatal_signals; extern int sysctl_overcommit_memory; extern int sysctl_overcommit_ratio; extern int sysctl_panic_on_oom; -- cgit v1.2.2 From e5ab67726f33b50f40db0ccf271ceb3c658554d5 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:05 -0800 Subject: sysctl extern cleanup: rcu Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move rcutorture_runnable extern declaration to linux/rcupdate.h Signed-off-by: Dave Young Acked-by: Josh Triplett Reviewed-by: "Paul E. McKenney" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index a8fd10a9a501..f18aaa7b0d65 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -87,9 +87,6 @@ extern int sysctl_nr_open_min, sysctl_nr_open_max; #ifndef CONFIG_MMU extern int sysctl_nr_trim_pages; #endif -#ifdef CONFIG_RCU_TORTURE_TEST -extern int rcutorture_runnable; -#endif /* #ifdef CONFIG_RCU_TORTURE_TEST */ #ifdef CONFIG_BLOCK extern int blk_iopoll_enabled; #endif -- cgit v1.2.2 From 5ed109103d73b0bafc92e860cead56725231384d Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:06 -0800 Subject: sysctl extern cleanup: module Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move modprobe_path extern declaration to linux/kmod.h Move modules_disabled extern declaration to linux/module.h Signed-off-by: Dave Young Cc: Rusty Russell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index f18aaa7b0d65..44e9492368fd 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -116,10 +116,6 @@ static int min_percpu_pagelist_fract = 8; static int ngroups_max = NGROUPS_MAX; -#ifdef CONFIG_MODULES -extern char modprobe_path[]; -extern int modules_disabled; -#endif #ifdef CONFIG_CHR_DEV_SG extern int sg_big_buff; #endif -- cgit v1.2.2 From 15485a4682d1d3bfee2aa78b4b1a5d36f5746b64 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:07 -0800 Subject: sysctl extern cleanup: sg Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move sg_big_buff extern declaration to scsi/sg.h Signed-off-by: Dave Young Acked-by: Doug Gilbert Cc: James Bottomley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 44e9492368fd..5290c437f151 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -61,6 +61,9 @@ #include #include #endif +#ifdef CONFIG_CHR_DEV_SG +#include +#endif #if defined(CONFIG_SYSCTL) @@ -116,10 +119,6 @@ static int min_percpu_pagelist_fract = 8; static int ngroups_max = NGROUPS_MAX; -#ifdef CONFIG_CHR_DEV_SG -extern int sg_big_buff; -#endif - #ifdef CONFIG_SPARC #include #endif -- cgit v1.2.2 From c55b7c3e82d0ad58f35a0785faaaf2f70b9b6cd3 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:08 -0800 Subject: sysctl extern cleanup: acct Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move acct_parm extern declaration to linux/acct.h Signed-off-by: Dave Young Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 5290c437f151..7635bb15f5af 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -61,6 +61,9 @@ #include #include #endif +#ifdef CONFIG_BSD_PROCESS_ACCT +#include +#endif #ifdef CONFIG_CHR_DEV_SG #include #endif @@ -140,10 +143,6 @@ extern int sysctl_userprocess_debug; extern int spin_retry; #endif -#ifdef CONFIG_BSD_PROCESS_ACCT -extern int acct_parm[]; -#endif - #ifdef CONFIG_IA64 extern int no_unaligned_warning; extern int unaligned_dump_stack; -- cgit v1.2.2 From 4f0e056fdebc15d3f4724ebc7bbf323158add1d7 Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:09 -0800 Subject: sysctl extern cleanup: rtmutex Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move max_lock_depth extern declaration to linux/rtmutex.h Signed-off-by: Dave Young Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 7635bb15f5af..622029ba5103 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -64,6 +64,9 @@ #ifdef CONFIG_BSD_PROCESS_ACCT #include #endif +#ifdef CONFIG_RT_MUTEXES +#include +#endif #ifdef CONFIG_CHR_DEV_SG #include #endif @@ -150,10 +153,6 @@ extern int unaligned_dump_stack; extern struct ratelimit_state printk_ratelimit_state; -#ifdef CONFIG_RT_MUTEXES -extern int max_lock_depth; -#endif - #ifdef CONFIG_PROC_SYSCTL static int proc_do_cad_pid(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos); -- cgit v1.2.2 From 2edf5e49800846a2b2b6461d99cdae18067c440f Mon Sep 17 00:00:00 2001 From: Dave Young Date: Wed, 10 Mar 2010 15:24:10 -0800 Subject: sysctl extern cleanup: lockdep Extern declarations in sysctl.c should be moved to their own header file, and then include them in relavant .c files. Move lockdep extern declarations to linux/lockdep.h Signed-off-by: Dave Young Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 622029ba5103..8686b0f5fc12 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -67,6 +67,9 @@ #ifdef CONFIG_RT_MUTEXES #include #endif +#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT) +#include +#endif #ifdef CONFIG_CHR_DEV_SG #include #endif @@ -191,9 +194,6 @@ extern struct ctl_table epoll_table[]; int sysctl_legacy_va_layout; #endif -extern int prove_locking; -extern int lock_stat; - /* The default sysctl tables: */ static struct ctl_table root_table[] = { -- cgit v1.2.2 From 52fbe9cde7fdb5c6fac196d7ebd2d92d05ef3cd4 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 8 Mar 2010 14:50:43 +0800 Subject: ring-buffer: Move disabled check into preempt disable section The ring buffer resizing and resetting relies on a schedule RCU action. The buffers are disabled, a synchronize_sched() is called and then the resize or reset takes place. But this only works if the disabling of the buffers are within the preempt disabled section, otherwise a window exists that the buffers can be written to while a reset or resize takes place. Cc: stable@kernel.org Reported-by: Li Zefan Signed-off-by: Lai Jiangshan LKML-Reference: <4B949E43.2010906@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 8c1b2d290718..54191d6ed195 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2232,12 +2232,12 @@ ring_buffer_lock_reserve(struct ring_buffer *buffer, unsigned long length) if (ring_buffer_flags != RB_BUFFERS_ON) return NULL; - if (atomic_read(&buffer->record_disabled)) - return NULL; - /* If we are tracing schedule, we don't want to recurse */ resched = ftrace_preempt_disable(); + if (atomic_read(&buffer->record_disabled)) + goto out_nocheck; + if (trace_recursive_lock()) goto out_nocheck; @@ -2469,11 +2469,11 @@ int ring_buffer_write(struct ring_buffer *buffer, if (ring_buffer_flags != RB_BUFFERS_ON) return -EBUSY; - if (atomic_read(&buffer->record_disabled)) - return -EBUSY; - resched = ftrace_preempt_disable(); + if (atomic_read(&buffer->record_disabled)) + goto out; + cpu = raw_smp_processor_id(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) -- cgit v1.2.2 From ea14eb714041d40fcc5180b5a586034503650149 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 12 Mar 2010 19:41:23 -0500 Subject: function-graph: Init curr_ret_stack with ret_stack If the graph tracer is active, and a task is forked but the allocating of the processes graph stack fails, it can cause crash later on. This is due to the temporary stack being NULL, but the curr_ret_stack variable is copied from the parent. If it is not -1, then in ftrace_graph_probe_sched_switch() the following: for (index = next->curr_ret_stack; index >= 0; index--) next->ret_stack[index].calltime += timestamp; Will cause a kernel OOPS. Found with Li Zefan's ftrace_stress_test. Cc: stable@kernel.org Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index d4d1238b096b..bb53edbb5c8c 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -3349,6 +3349,7 @@ void ftrace_graph_init_task(struct task_struct *t) { /* Make sure we do not use the parent ret_stack */ t->ret_stack = NULL; + t->curr_ret_stack = -1; if (ftrace_graph_active) { struct ftrace_ret_stack *ret_stack; @@ -3358,7 +3359,6 @@ void ftrace_graph_init_task(struct task_struct *t) GFP_KERNEL); if (!ret_stack) return; - t->curr_ret_stack = -1; atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); t->ftrace_timestamp = 0; -- cgit v1.2.2 From 283740c619d211e34572cc93c8cdba92ccbdb9cc Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 12 Mar 2010 19:48:41 -0500 Subject: tracing: Use same local variable when resetting the ring buffer In the ftrace code that resets the ring buffer it references the buffer with a local variable, but then uses the tr->buffer as the parameter to reset. If the wakeup tracer is running, which can switch the tr->buffer with the max saved buffer, this can break the requirement of disabling the buffer before the reset. buffer = tr->buffer; ring_buffer_record_disable(buffer); synchronize_sched(); __tracing_reset(tr->buffer, cpu); If the tr->buffer is swapped, then the reset is not happening to the buffer that was disabled. This will cause the ring buffer to fail. Found with Li Zefan's ftrace_stress_test. Cc: stable@kernel.org Reported-by: Lai Jiangshan Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6af8d7bc953b..60de37bd0f75 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -840,10 +840,10 @@ out: mutex_unlock(&trace_types_lock); } -static void __tracing_reset(struct trace_array *tr, int cpu) +static void __tracing_reset(struct ring_buffer *buffer, int cpu) { ftrace_disable_cpu(); - ring_buffer_reset_cpu(tr->buffer, cpu); + ring_buffer_reset_cpu(buffer, cpu); ftrace_enable_cpu(); } @@ -855,7 +855,7 @@ void tracing_reset(struct trace_array *tr, int cpu) /* Make sure all commits have finished */ synchronize_sched(); - __tracing_reset(tr, cpu); + __tracing_reset(buffer, cpu); ring_buffer_record_enable(buffer); } @@ -873,7 +873,7 @@ void tracing_reset_online_cpus(struct trace_array *tr) tr->time_start = ftrace_now(tr->cpu); for_each_online_cpu(cpu) - __tracing_reset(tr, cpu); + __tracing_reset(buffer, cpu); ring_buffer_record_enable(buffer); } -- cgit v1.2.2 From a2f8071428ed9a0f06865f417c962421c9a6b488 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 12 Mar 2010 19:56:00 -0500 Subject: tracing: Disable buffer switching when starting or stopping trace When the trace iterator is read, tracing_start() and tracing_stop() is called to stop tracing while the iterator is processing the trace output. These functions disable both the standard buffer and the max latency buffer. But if the wakeup tracer is running, it can switch these buffers between the two disables: buffer = global_trace.buffer; if (buffer) ring_buffer_record_disable(buffer); <<<--------- swap happens here buffer = max_tr.buffer; if (buffer) ring_buffer_record_disable(buffer); What happens is that we disabled the same buffer twice. On tracing_start() we can enable the same buffer twice. All ring_buffer_record_disable() must be matched with a ring_buffer_record_enable() or the buffer can be disable permanently, or enable prematurely, and cause a bug where a reset happens while a trace is commiting. This patch protects these two by taking the ftrace_max_lock to prevent a switch from occurring. Found with Li Zefan's ftrace_stress_test. Cc: stable@kernel.org Reported-by: Lai Jiangshan Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 60de37bd0f75..484337d33959 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -950,6 +950,8 @@ void tracing_start(void) goto out; } + /* Prevent the buffers from switching */ + arch_spin_lock(&ftrace_max_lock); buffer = global_trace.buffer; if (buffer) @@ -959,6 +961,8 @@ void tracing_start(void) if (buffer) ring_buffer_record_enable(buffer); + arch_spin_unlock(&ftrace_max_lock); + ftrace_start(); out: spin_unlock_irqrestore(&tracing_start_lock, flags); @@ -980,6 +984,9 @@ void tracing_stop(void) if (trace_stop_count++) goto out; + /* Prevent the buffers from switching */ + arch_spin_lock(&ftrace_max_lock); + buffer = global_trace.buffer; if (buffer) ring_buffer_record_disable(buffer); @@ -988,6 +995,8 @@ void tracing_stop(void) if (buffer) ring_buffer_record_disable(buffer); + arch_spin_unlock(&ftrace_max_lock); + out: spin_unlock_irqrestore(&tracing_start_lock, flags); } -- cgit v1.2.2 From b6345879ccbd9b92864fbd7eb8ac48acdb4d6b15 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 12 Mar 2010 20:03:30 -0500 Subject: tracing: Do not record user stack trace from NMI context A bug was found with Li Zefan's ftrace_stress_test that caused applications to segfault during the test. Placing a tracing_off() in the segfault code, and examining several traces, I found that the following was always the case. The lock tracer was enabled (lockdep being required) and userstack was enabled. Testing this out, I just enabled the two, but that was not good enough. I needed to run something else that could trigger it. Running a load like hackbench did not work, but executing a new program would. The following would trigger the segfault within seconds: # echo 1 > /debug/tracing/options/userstacktrace # echo 1 > /debug/tracing/events/lock/enable # while :; do ls > /dev/null ; done Enabling the function graph tracer and looking at what was happening I finally noticed that all cashes happened just after an NMI. 1) | copy_user_handle_tail() { 1) | bad_area_nosemaphore() { 1) | __bad_area_nosemaphore() { 1) | no_context() { 1) | fixup_exception() { 1) 0.319 us | search_exception_tables(); 1) 0.873 us | } [...] 1) 0.314 us | __rcu_read_unlock(); 1) 0.325 us | native_apic_mem_write(); 1) 0.943 us | } 1) 0.304 us | rcu_nmi_exit(); [...] 1) 0.479 us | find_vma(); 1) | bad_area() { 1) | __bad_area() { After capturing several traces of failures, all of them happened after an NMI. Curious about this, I added a trace_printk() to the NMI handler to read the regs->ip to see where the NMI happened. In which I found out it was here: ffffffff8135b660 : ffffffff8135b660: 48 83 ec 78 sub $0x78,%rsp ffffffff8135b664: e8 97 01 00 00 callq ffffffff8135b800 What was happening is that the NMI would happen at the place that a page fault occurred. It would call rcu_read_lock() which was traced by the lock events, and the user_stack_trace would run. This would trigger a page fault inside the NMI. I do not see where the CR2 register is saved or restored in NMI handling. This means that it would corrupt the page fault handling that the NMI interrupted. The reason the while loop of ls helped trigger the bug, was that each execution of ls would cause lots of pages to be faulted in, and increase the chances of the race happening. The simple solution is to not allow user stack traces in NMI context. After this patch, I ran the above "ls" test for a couple of hours without any issues. Without this patch, the bug would trigger in less than a minute. Cc: stable@kernel.org Reported-by: Li Zefan Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 484337d33959..e52683f7c3b2 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1284,6 +1284,13 @@ ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, int pc) if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) return; + /* + * NMIs can not handle page faults, even with fix ups. + * The save user stack can (and often does) fault. + */ + if (unlikely(in_nmi())) + return; + event = trace_buffer_lock_reserve(buffer, TRACE_USER_STACK, sizeof(*entry), flags, pc); if (!event) -- cgit v1.2.2 From cd3d8031eb4311e516329aee03c79a08333141f1 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 12 Mar 2010 16:15:36 +0900 Subject: sched: sched_getaffinity(): Allow less than NR_CPUS length [ Note, this commit changes the syscall ABI for > 1024 CPUs systems. ] Recently, some distro decided to use NR_CPUS=4096 for mysterious reasons. Unfortunately, glibc sched interface has the following definition: # define __CPU_SETSIZE 1024 # define __NCPUBITS (8 * sizeof (__cpu_mask)) typedef unsigned long int __cpu_mask; typedef struct { __cpu_mask __bits[__CPU_SETSIZE / __NCPUBITS]; } cpu_set_t; It mean, if NR_CPUS is bigger than 1024, cpu_set_t makes an ABI issue ... More recently, Sharyathi Nagesh reported following test program makes misterious syscall failure: ----------------------------------------------------------------------- #define _GNU_SOURCE #include #include #include int main() { cpu_set_t set; if (sched_getaffinity(0, sizeof(cpu_set_t), &set) < 0) printf("\n Call is failing with:%d", errno); } ----------------------------------------------------------------------- Because the kernel assumes len argument of sched_getaffinity() is bigger than NR_CPUS. But now it is not correct. Now we are faced with the following annoying dilemma, due to the limitations of the glibc interface built in years ago: (1) if we change glibc's __CPU_SETSIZE definition, we lost binary compatibility of _all_ application. (2) if we don't change it, we also lost binary compatibility of Sharyathi's use case. Then, I would propse to change the rule of the len argument of sched_getaffinity(). Old: len should be bigger than NR_CPUS New: len should be bigger than maximum possible cpu id This creates the following behavior: (A) In the real 4096 cpus machine, the above test program still return -EINVAL. (B) NR_CPUS=4096 but the machine have less than 1024 cpus (almost all machines in the world), the above can run successfully. Fortunatelly, BIG SGI machine is mainly used for HPC use case. It means they can rebuild their programs. IOW we hope they are not annoyed by this issue ... Reported-by: Sharyathi Nagesh Signed-off-by: KOSAKI Motohiro Acked-by: Ulrich Drepper Acked-by: Peter Zijlstra Cc: Linus Torvalds Cc: Andrew Morton Cc: Jack Steiner Cc: Russ Anderson Cc: Mike Travis LKML-Reference: <20100312161316.9520.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 9ab3cd7858d3..6eaef3df2d19 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4902,7 +4902,9 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, int ret; cpumask_var_t mask; - if (len < cpumask_size()) + if (len < nr_cpu_ids) + return -EINVAL; + if (len & (sizeof(unsigned long)-1)) return -EINVAL; if (!alloc_cpumask_var(&mask, GFP_KERNEL)) @@ -4910,10 +4912,12 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, ret = sched_getaffinity(pid, mask); if (ret == 0) { - if (copy_to_user(user_mask_ptr, mask, cpumask_size())) + int retlen = min(len, cpumask_size()); + + if (copy_to_user(user_mask_ptr, mask, retlen)) ret = -EFAULT; else - ret = cpumask_size(); + ret = retlen; } free_cpumask_var(mask); -- cgit v1.2.2 From e3818b8dce2a934cd1521dbc4827e5238d8f45d8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 15 Mar 2010 17:03:43 -0700 Subject: rcu: Make rcu_read_lock_bh_held() allow for disabled BH Disabling BH can stand in for rcu_read_lock_bh(), and this patch updates rcu_read_lock_bh_held() to allow for this. In order to avoid include-file hell, this function is moved out of line to kernel/rcupdate.c. This fixes a false positive RCU warning. Reported-by: Arnd Bergmann Reported-by: Eric Dumazet Signed-off-by: Paul E. McKenney Acked-by: Lai Jiangshan Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <20100316000343.GA25857@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index f1125c1a6321..63fe25433980 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -45,6 +45,7 @@ #include #include #include +#include #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; @@ -66,6 +67,28 @@ EXPORT_SYMBOL_GPL(rcu_sched_lock_map); int rcu_scheduler_active __read_mostly; EXPORT_SYMBOL_GPL(rcu_scheduler_active); +#ifdef CONFIG_DEBUG_LOCK_ALLOC + +/** + * rcu_read_lock_bh_held - might we be in RCU-bh read-side critical section? + * + * Check for bottom half being disabled, which covers both the + * CONFIG_PROVE_RCU and not cases. Note that if someone uses + * rcu_read_lock_bh(), but then later enables BH, lockdep (if enabled) + * will show the situation. + * + * Check debug_lockdep_rcu_enabled() to prevent false positives during boot. + */ +int rcu_read_lock_bh_held(void) +{ + if (!debug_lockdep_rcu_enabled()) + return 1; + return in_softirq(); +} +EXPORT_SYMBOL_GPL(rcu_read_lock_bh_held); + +#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + /* * This function is invoked towards the end of the scheduler's initialization * process. Before this is called, the idle task might contain -- cgit v1.2.2 From c890692bf37671b5b78a1870d55d6d87e1c8a509 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 11 Mar 2010 14:08:43 -0800 Subject: kernel/sched.c: Suppress unused var warning On UP: kernel/sched.c: In function 'wake_up_new_task': kernel/sched.c:2631: warning: unused variable 'cpu' Signed-off-by: Andrew Morton Cc: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 6eaef3df2d19..82975b5b42f7 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2650,7 +2650,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; - int cpu = get_cpu(); + int cpu __maybe_unused = get_cpu(); #ifdef CONFIG_SMP /* -- cgit v1.2.2 From 8bc037fb89bb3104b9ae290d18c877624cd7d9cc Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Wed, 17 Mar 2010 09:36:58 +0900 Subject: sched: Use proper type in sched_getaffinity() Using the proper type fixes the following compiler warning: kernel/sched.c:4850: warning: comparison of distinct pointer types lacks a cast Signed-off-by: KOSAKI Motohiro Cc: torvalds@linux-foundation.org Cc: travis@sgi.com Cc: peterz@infradead.org Cc: drepper@redhat.com Cc: rja@sgi.com Cc: sharyath@in.ibm.com Cc: steiner@sgi.com LKML-Reference: <20100317090046.4C79.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 82975b5b42f7..49d2fa7b687a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4912,7 +4912,7 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, ret = sched_getaffinity(pid, mask); if (ret == 0) { - int retlen = min(len, cpumask_size()); + size_t retlen = min_t(size_t, len, cpumask_size()); if (copy_to_user(user_mask_ptr, mask, retlen)) ret = -EFAULT; -- cgit v1.2.2 From dcd5c1662db59a6b82942f47fb6ac9dd63f6d3dd Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 16 Mar 2010 01:05:02 +0100 Subject: perf: Fix unexported generic perf_arch_fetch_caller_regs perf_arch_fetch_caller_regs() is exported for the overriden x86 version, but not for the generic weak version. As a general rule, weak functions should not have their symbol exported in the same file they are defined. So let's export it on trace_event_perf.c as it is used by trace events only. This fixes: ERROR: ".perf_arch_fetch_caller_regs" [fs/xfs/xfs.ko] undefined! ERROR: ".perf_arch_fetch_caller_regs" [arch/powerpc/platforms/cell/spufs/spufs.ko] undefined! -v2: And also only build it if trace events are enabled. -v3: Fix changelog mistake Reported-by: Stephen Rothwell Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Xiao Guangrong Cc: Paul Mackerras LKML-Reference: <1268697902-9518-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 2 ++ kernel/trace/trace_event_perf.c | 2 ++ 2 files changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index fb3031cf9f17..574ee58a3046 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2786,10 +2786,12 @@ __weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) return NULL; } +#ifdef CONFIG_EVENT_TRACING __weak void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip) { } +#endif /* * Output diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 0709e4f75114..7d79a10c3cde 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -12,6 +12,8 @@ DEFINE_PER_CPU(struct pt_regs, perf_trace_regs); EXPORT_PER_CPU_SYMBOL_GPL(perf_trace_regs); +EXPORT_SYMBOL_GPL(perf_arch_fetch_caller_regs); + static char *perf_trace_buf; static char *perf_trace_buf_nmi; -- cgit v1.2.2 From 2271048d1b3b0aabf83d25b29c20646dcabedc05 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 18 Mar 2010 17:54:19 -0400 Subject: ring-buffer: Do 8 byte alignment for 64 bit that can not handle 4 byte align The ring buffer uses 4 byte alignment while recording events into the buffer, even on 64bit machines. This saves space when there are lots of events being recorded at 4 byte boundaries. The ring buffer has a zero copy method to write into the buffer, with the reserving of space and then committing it. This may cause problems when writing an 8 byte word into a 4 byte alignment (not 8). For x86 and PPC this is not an issue, but on some architectures this would cause an out-of-alignment exception. This patch uses CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS to determine if it is OK to use 4 byte alignments on 64 bit machines. If it is not, it forces the ring buffer event header to be 8 bytes and not 4, and will align the length of the data to be 8 byte aligned. This keeps the data payload at 8 byte alignments and will allow these machines to run without issue. The trick to this is that the header can be either 4 bytes or 8 bytes depending on the length of the data payload. The 4 byte header has a length field that supports up to 112 bytes. If the length of the data is more than 112, the length field is set to zero, and the actual length is stored in the next 4 bytes after the header. When CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS is not set, the code forces zero in the 4 byte header forcing the length to be stored in the 4 byte array, even with a small data load. It also forces the length of the data load to be 8 byte aligned. The combination of these two guarantee that the data is always at 8 byte alignment. Tested-by: Frederic Weisbecker (on sparc64) Reported-by: Frederic Weisbecker Acked-by: David S. Miller Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 05a9f83b8819..d1187ef20caf 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -207,6 +207,14 @@ EXPORT_SYMBOL_GPL(tracing_is_on); #define RB_MAX_SMALL_DATA (RB_ALIGNMENT * RINGBUF_TYPE_DATA_TYPE_LEN_MAX) #define RB_EVNT_MIN_SIZE 8U /* two 32bit words */ +#if !defined(CONFIG_64BIT) || defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) +# define RB_FORCE_8BYTE_ALIGNMENT 0 +# define RB_ARCH_ALIGNMENT RB_ALIGNMENT +#else +# define RB_FORCE_8BYTE_ALIGNMENT 1 +# define RB_ARCH_ALIGNMENT 8U +#endif + /* define RINGBUF_TYPE_DATA for 'case RINGBUF_TYPE_DATA:' */ #define RINGBUF_TYPE_DATA 0 ... RINGBUF_TYPE_DATA_TYPE_LEN_MAX @@ -1547,7 +1555,7 @@ rb_update_event(struct ring_buffer_event *event, case 0: length -= RB_EVNT_HDR_SIZE; - if (length > RB_MAX_SMALL_DATA) + if (length > RB_MAX_SMALL_DATA || RB_FORCE_8BYTE_ALIGNMENT) event->array[0] = length; else event->type_len = DIV_ROUND_UP(length, RB_ALIGNMENT); @@ -1722,11 +1730,11 @@ static unsigned rb_calculate_event_length(unsigned length) if (!length) length = 1; - if (length > RB_MAX_SMALL_DATA) + if (length > RB_MAX_SMALL_DATA || RB_FORCE_8BYTE_ALIGNMENT) length += sizeof(event.array[0]); length += RB_EVNT_HDR_SIZE; - length = ALIGN(length, RB_ALIGNMENT); + length = ALIGN(length, RB_ARCH_ALIGNMENT); return length; } -- cgit v1.2.2 From 8c2eb4805d422bdbf60ba00ff233c794d23c3c00 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Fri, 19 Mar 2010 10:28:02 +0000 Subject: softlockup: Stop spurious softlockup messages due to overflow Ensure additions on touch_ts do not overflow. This can occur when the top 32 bits of the TSC reach 0xffffffff causing additions to touch_ts to overflow and this in turn generates spurious softlockup warnings. Signed-off-by: Colin Ian King Cc: Peter Zijlstra Cc: Eric Dumazet Cc: LKML-Reference: <1268994482.1798.6.camel@lenovo> Signed-off-by: Ingo Molnar --- kernel/softlockup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/softlockup.c b/kernel/softlockup.c index 0d4c7898ab80..4b493f67dcb5 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -155,11 +155,11 @@ void softlockup_tick(void) * Wake up the high-prio watchdog task twice per * threshold timespan. */ - if (now > touch_ts + softlockup_thresh/2) + if (time_after(now - softlockup_thresh/2, touch_ts)) wake_up_process(per_cpu(softlockup_watchdog, this_cpu)); /* Warn about unreasonable delays: */ - if (now <= (touch_ts + softlockup_thresh)) + if (time_before_eq(now - softlockup_thresh, touch_ts)) return; per_cpu(softlockup_print_ts, this_cpu) = touch_ts; -- cgit v1.2.2 From 830ec0458c390f29c6c99e1ff7feab9e36368d12 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Thu, 18 Mar 2010 14:47:30 -0700 Subject: time: Fix accumulation bug triggered by long delay. The logarithmic accumulation done in the timekeeping has some overflow protection that limits the max shift value. That means it will take more then shift loops to accumulate all of the cycles. This causes the shift decrement to underflow, which causes the loop to never exit. The simplest fix would be simply to do a: if (shift) shift--; However that is not optimal, as we know the cycle offset is larger then the interval << shift, the above would make shift drop to zero, then we would be spinning for quite awhile accumulating at interval chunks at a time. Instead, this patch only decreases shift if the offset is smaller then cycle_interval << shift. This makes sure we accumulate using the largest chunks possible without overflowing tick_length, and limits the number of iterations through the loop. This issue was found and reported by Sonic Zhang, who also tested the fix. Many thanks your explanation and testing! Reported-by: Sonic Zhang Signed-off-by: John Stultz Tested-by: Sonic Zhang LKML-Reference: <1268948850-5225-1-git-send-email-johnstul@us.ibm.com> Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 16736379a9ca..39f6177fafac 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -818,7 +818,8 @@ void update_wall_time(void) shift = min(shift, maxshift); while (offset >= timekeeper.cycle_interval) { offset = logarithmic_accumulation(offset, shift); - shift--; + if(offset < timekeeper.cycle_interval< Date: Thu, 11 Mar 2010 17:01:09 -0700 Subject: resources: add interfaces that return conflict information request_resource() and insert_resource() only return success or failure, which no information about what existing resource conflicted with the proposed new reservation. This patch adds request_resource_conflict() and insert_resource_conflict(), which return the conflicting resource. Callers may use this for better error messages or to adjust the new resource and retry the request. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- kernel/resource.c | 44 +++++++++++++++++++++++++++++++++++++------- 1 file changed, 37 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index 2d5be5d9bf5f..9c358e263534 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -219,19 +219,34 @@ void release_child_resources(struct resource *r) } /** - * request_resource - request and reserve an I/O or memory resource + * request_resource_conflict - request and reserve an I/O or memory resource * @root: root resource descriptor * @new: resource descriptor desired by caller * - * Returns 0 for success, negative error code on error. + * Returns 0 for success, conflict resource on error. */ -int request_resource(struct resource *root, struct resource *new) +struct resource *request_resource_conflict(struct resource *root, struct resource *new) { struct resource *conflict; write_lock(&resource_lock); conflict = __request_resource(root, new); write_unlock(&resource_lock); + return conflict; +} + +/** + * request_resource - request and reserve an I/O or memory resource + * @root: root resource descriptor + * @new: resource descriptor desired by caller + * + * Returns 0 for success, negative error code on error. + */ +int request_resource(struct resource *root, struct resource *new) +{ + struct resource *conflict; + + conflict = request_resource_conflict(root, new); return conflict ? -EBUSY : 0; } @@ -474,25 +489,40 @@ static struct resource * __insert_resource(struct resource *parent, struct resou } /** - * insert_resource - Inserts a resource in the resource tree + * insert_resource_conflict - Inserts resource in the resource tree * @parent: parent of the new resource * @new: new resource to insert * - * Returns 0 on success, -EBUSY if the resource can't be inserted. + * Returns 0 on success, conflict resource if the resource can't be inserted. * - * This function is equivalent to request_resource when no conflict + * This function is equivalent to request_resource_conflict when no conflict * happens. If a conflict happens, and the conflicting resources * entirely fit within the range of the new resource, then the new * resource is inserted and the conflicting resources become children of * the new resource. */ -int insert_resource(struct resource *parent, struct resource *new) +struct resource *insert_resource_conflict(struct resource *parent, struct resource *new) { struct resource *conflict; write_lock(&resource_lock); conflict = __insert_resource(parent, new); write_unlock(&resource_lock); + return conflict; +} + +/** + * insert_resource - Inserts a resource in the resource tree + * @parent: parent of the new resource + * @new: new resource to insert + * + * Returns 0 on success, -EBUSY if the resource can't be inserted. + */ +int insert_resource(struct resource *parent, struct resource *new) +{ + struct resource *conflict; + + conflict = insert_resource_conflict(parent, new); return conflict ? -EBUSY : 0; } -- cgit v1.2.2 From cc8c3b78433222e5dbc1fdfcfdde29e1743f181a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 23 Mar 2010 22:40:53 +0100 Subject: genirq: Protect access to irq_desc->action in can_request_irq() can_request_irq() accesses and dereferences irq_desc->action w/o holding irq_desc->lock. So action can be freed on another CPU before it's dereferenced. Unlikely, but ... Protect it with desc->lock. Signed-off-by: Thomas Gleixner --- kernel/irq/manage.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 69a3d7b9414c..398fda155f6e 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -382,6 +382,7 @@ int can_request_irq(unsigned int irq, unsigned long irqflags) { struct irq_desc *desc = irq_to_desc(irq); struct irqaction *action; + unsigned long flags; if (!desc) return 0; @@ -389,11 +390,14 @@ int can_request_irq(unsigned int irq, unsigned long irqflags) if (desc->status & IRQ_NOREQUEST) return 0; + raw_spin_lock_irqsave(&desc->lock, flags); action = desc->action; if (action) if (irqflags & action->flags & IRQF_SHARED) action = NULL; + raw_spin_unlock_irqrestore(&desc->lock, flags); + return !action; } -- cgit v1.2.2 From 860652bfb890bd861c999ec39fcffabe5b712f85 Mon Sep 17 00:00:00 2001 From: Henrik Kretzschmar Date: Wed, 24 Mar 2010 12:59:20 +0100 Subject: genirq: Move two IRQ functions from .init.text to .text Both functions should not be marked as __init, since they be called from modules after the init section is freed. Signed-off-by: Henrik Kretzschmar Cc: Yinghai Lu Cc: Peter Zijlstra Cc: Jiri Kosina LKML-Reference: <1269431961-5731-1-git-send-email-henne@nachtwindheim.de> Signed-off-by: Thomas Gleixner --- kernel/irq/chip.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 71eba24a39a2..3c2d6e7737f0 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -729,7 +729,7 @@ set_irq_chip_and_handler_name(unsigned int irq, struct irq_chip *chip, __set_irq_handler(irq, handle, 0, name); } -void __init set_irq_noprobe(unsigned int irq) +void set_irq_noprobe(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); unsigned long flags; @@ -744,7 +744,7 @@ void __init set_irq_noprobe(unsigned int irq) raw_spin_unlock_irqrestore(&desc->lock, flags); } -void __init set_irq_probe(unsigned int irq) +void set_irq_probe(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); unsigned long flags; -- cgit v1.2.2 From 9d34706f42f9b8c15185423d9af98d37ba21d011 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 23 Mar 2010 13:35:12 -0700 Subject: cgroups: remove duplicate include commit e6a1105b ("cgroups: subsystem module loading interface") and commit c50cc752 ("sched, cgroups: Fix module export") result in duplicate including of module.h Signed-off-by: Li Zefan Acked-by: Paul Menage Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cgroup.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ef909a329750..e2769e13980c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -27,7 +27,6 @@ */ #include -#include #include #include #include -- cgit v1.2.2 From 5ab116c9349ef52d6fbd2e2917a53f13194b048e Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Tue, 23 Mar 2010 13:35:34 -0700 Subject: cpuset: fix the problem that cpuset_mem_spread_node() returns an offline node cpuset_mem_spread_node() returns an offline node, and causes an oops. This patch fixes it by initializing task->mems_allowed to node_states[N_HIGH_MEMORY], and updating task->mems_allowed when doing memory hotplug. Signed-off-by: Miao Xie Acked-by: David Rientjes Reported-by: Nick Piggin Tested-by: Nick Piggin Cc: Paul Menage Cc: Li Zefan Cc: Ingo Molnar Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 20 ++++++++++++-------- kernel/kthread.c | 2 +- 2 files changed, 13 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index ba401fab459f..5d38bd74483c 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -920,9 +920,6 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, * call to guarantee_online_mems(), as we know no one is changing * our task's cpuset. * - * Hold callback_mutex around the two modifications of our tasks - * mems_allowed to synchronize with cpuset_mems_allowed(). - * * While the mm_struct we are migrating is typically from some * other task, the task_struct mems_allowed that we are hacking * is for our current task, which must allocate new pages for that @@ -1391,11 +1388,10 @@ static void cpuset_attach(struct cgroup_subsys *ss, struct cgroup *cont, if (cs == &top_cpuset) { cpumask_copy(cpus_attach, cpu_possible_mask); - to = node_possible_map; } else { guarantee_online_cpus(cs, cpus_attach); - guarantee_online_mems(cs, &to); } + guarantee_online_mems(cs, &to); /* do per-task migration stuff possibly for each in the threadgroup */ cpuset_attach_task(tsk, &to, cs); @@ -2090,15 +2086,23 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb, static int cpuset_track_online_nodes(struct notifier_block *self, unsigned long action, void *arg) { + nodemask_t oldmems; + cgroup_lock(); switch (action) { case MEM_ONLINE: - case MEM_OFFLINE: + oldmems = top_cpuset.mems_allowed; mutex_lock(&callback_mutex); top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; mutex_unlock(&callback_mutex); - if (action == MEM_OFFLINE) - scan_for_empty_cpusets(&top_cpuset); + update_tasks_nodemask(&top_cpuset, &oldmems, NULL); + break; + case MEM_OFFLINE: + /* + * needn't update top_cpuset.mems_allowed explicitly because + * scan_for_empty_cpusets() will update it. + */ + scan_for_empty_cpusets(&top_cpuset); break; default: break; diff --git a/kernel/kthread.c b/kernel/kthread.c index 82ed0ea15194..83911c780175 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -219,7 +219,7 @@ int kthreadd(void *unused) set_task_comm(tsk, "kthreadd"); ignore_signals(tsk); set_cpus_allowed_ptr(tsk, cpu_all_mask); - set_mems_allowed(node_possible_map); + set_mems_allowed(node_states[N_HIGH_MEMORY]); current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG; -- cgit v1.2.2 From 53feb29767c29c877f9d47dcfe14211b5b0f7ebd Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Tue, 23 Mar 2010 13:35:35 -0700 Subject: cpuset: alloc nodemask_t on the heap rather than the stack Signed-off-by: Miao Xie Acked-by: David Rientjes Cc: Nick Piggin Cc: Paul Menage Cc: Li Zefan Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpuset.c | 94 ++++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 66 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 5d38bd74483c..d10946748ec2 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -970,15 +970,20 @@ static void cpuset_change_nodemask(struct task_struct *p, struct cpuset *cs; int migrate; const nodemask_t *oldmem = scan->data; - nodemask_t newmems; + NODEMASK_ALLOC(nodemask_t, newmems, GFP_KERNEL); + + if (!newmems) + return; cs = cgroup_cs(scan->cg); - guarantee_online_mems(cs, &newmems); + guarantee_online_mems(cs, newmems); task_lock(p); - cpuset_change_task_nodemask(p, &newmems); + cpuset_change_task_nodemask(p, newmems); task_unlock(p); + NODEMASK_FREE(newmems); + mm = get_task_mm(p); if (!mm) return; @@ -1048,16 +1053,21 @@ static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, const char *buf) { - nodemask_t oldmem; + NODEMASK_ALLOC(nodemask_t, oldmem, GFP_KERNEL); int retval; struct ptr_heap heap; + if (!oldmem) + return -ENOMEM; + /* * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY]; * it's read-only */ - if (cs == &top_cpuset) - return -EACCES; + if (cs == &top_cpuset) { + retval = -EACCES; + goto done; + } /* * An empty mems_allowed is ok iff there are no tasks in the cpuset. @@ -1073,11 +1083,13 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, goto done; if (!nodes_subset(trialcs->mems_allowed, - node_states[N_HIGH_MEMORY])) - return -EINVAL; + node_states[N_HIGH_MEMORY])) { + retval = -EINVAL; + goto done; + } } - oldmem = cs->mems_allowed; - if (nodes_equal(oldmem, trialcs->mems_allowed)) { + *oldmem = cs->mems_allowed; + if (nodes_equal(*oldmem, trialcs->mems_allowed)) { retval = 0; /* Too easy - nothing to do */ goto done; } @@ -1093,10 +1105,11 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, cs->mems_allowed = trialcs->mems_allowed; mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, &oldmem, &heap); + update_tasks_nodemask(cs, oldmem, &heap); heap_free(&heap); done: + NODEMASK_FREE(oldmem); return retval; } @@ -1381,39 +1394,47 @@ static void cpuset_attach(struct cgroup_subsys *ss, struct cgroup *cont, struct cgroup *oldcont, struct task_struct *tsk, bool threadgroup) { - nodemask_t from, to; struct mm_struct *mm; struct cpuset *cs = cgroup_cs(cont); struct cpuset *oldcs = cgroup_cs(oldcont); + NODEMASK_ALLOC(nodemask_t, from, GFP_KERNEL); + NODEMASK_ALLOC(nodemask_t, to, GFP_KERNEL); + + if (from == NULL || to == NULL) + goto alloc_fail; if (cs == &top_cpuset) { cpumask_copy(cpus_attach, cpu_possible_mask); } else { guarantee_online_cpus(cs, cpus_attach); } - guarantee_online_mems(cs, &to); + guarantee_online_mems(cs, to); /* do per-task migration stuff possibly for each in the threadgroup */ - cpuset_attach_task(tsk, &to, cs); + cpuset_attach_task(tsk, to, cs); if (threadgroup) { struct task_struct *c; rcu_read_lock(); list_for_each_entry_rcu(c, &tsk->thread_group, thread_group) { - cpuset_attach_task(c, &to, cs); + cpuset_attach_task(c, to, cs); } rcu_read_unlock(); } /* change mm; only needs to be done once even if threadgroup */ - from = oldcs->mems_allowed; - to = cs->mems_allowed; + *from = oldcs->mems_allowed; + *to = cs->mems_allowed; mm = get_task_mm(tsk); if (mm) { - mpol_rebind_mm(mm, &to); + mpol_rebind_mm(mm, to); if (is_memory_migrate(cs)) - cpuset_migrate_mm(mm, &from, &to); + cpuset_migrate_mm(mm, from, to); mmput(mm); } + +alloc_fail: + NODEMASK_FREE(from); + NODEMASK_FREE(to); } /* The various types of files and directories in a cpuset file system */ @@ -1558,13 +1579,21 @@ static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs) static int cpuset_sprintf_memlist(char *page, struct cpuset *cs) { - nodemask_t mask; + NODEMASK_ALLOC(nodemask_t, mask, GFP_KERNEL); + int retval; + + if (mask == NULL) + return -ENOMEM; mutex_lock(&callback_mutex); - mask = cs->mems_allowed; + *mask = cs->mems_allowed; mutex_unlock(&callback_mutex); - return nodelist_scnprintf(page, PAGE_SIZE, mask); + retval = nodelist_scnprintf(page, PAGE_SIZE, *mask); + + NODEMASK_FREE(mask); + + return retval; } static ssize_t cpuset_common_file_read(struct cgroup *cont, @@ -1993,7 +2022,10 @@ static void scan_for_empty_cpusets(struct cpuset *root) struct cpuset *cp; /* scans cpusets being updated */ struct cpuset *child; /* scans child cpusets of cp */ struct cgroup *cont; - nodemask_t oldmems; + NODEMASK_ALLOC(nodemask_t, oldmems, GFP_KERNEL); + + if (oldmems == NULL) + return; list_add_tail((struct list_head *)&root->stack_list, &queue); @@ -2010,7 +2042,7 @@ static void scan_for_empty_cpusets(struct cpuset *root) nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY])) continue; - oldmems = cp->mems_allowed; + *oldmems = cp->mems_allowed; /* Remove offline cpus and mems from this cpuset. */ mutex_lock(&callback_mutex); @@ -2026,9 +2058,10 @@ static void scan_for_empty_cpusets(struct cpuset *root) remove_tasks_in_empty_cpuset(cp); else { update_tasks_cpumask(cp, NULL); - update_tasks_nodemask(cp, &oldmems, NULL); + update_tasks_nodemask(cp, oldmems, NULL); } } + NODEMASK_FREE(oldmems); } /* @@ -2086,16 +2119,19 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb, static int cpuset_track_online_nodes(struct notifier_block *self, unsigned long action, void *arg) { - nodemask_t oldmems; + NODEMASK_ALLOC(nodemask_t, oldmems, GFP_KERNEL); + + if (oldmems == NULL) + return NOTIFY_DONE; cgroup_lock(); switch (action) { case MEM_ONLINE: - oldmems = top_cpuset.mems_allowed; + *oldmems = top_cpuset.mems_allowed; mutex_lock(&callback_mutex); top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; mutex_unlock(&callback_mutex); - update_tasks_nodemask(&top_cpuset, &oldmems, NULL); + update_tasks_nodemask(&top_cpuset, oldmems, NULL); break; case MEM_OFFLINE: /* @@ -2108,6 +2144,8 @@ static int cpuset_track_online_nodes(struct notifier_block *self, break; } cgroup_unlock(); + + NODEMASK_FREE(oldmems); return NOTIFY_OK; } #endif -- cgit v1.2.2 From 4f598458ea4450f53e8ed929ee4e66b3404a7286 Mon Sep 17 00:00:00 2001 From: Xiaotian Feng Date: Wed, 10 Mar 2010 22:59:13 +0100 Subject: Freezer: Only show the state of tasks refusing to freeze show_state will dump all tasks state, so if freezer failed to freeze any task, kernel will dump all tasks state and flood the dmesg log. This patch makes freezer only show state of tasks refusing to freeze. Signed-off-by: Xiaotian Feng Acked-by: Pavel Machek Acked-by: David Rientjes Signed-off-by: Rafael J. Wysocki --- kernel/power/process.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/power/process.c b/kernel/power/process.c index 5ade1bdcf366..a0480cd4daaf 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -88,12 +88,11 @@ static int try_to_freeze_tasks(bool sig_only) printk(KERN_ERR "Freezing of tasks failed after %d.%02d seconds " "(%d tasks refusing to freeze):\n", elapsed_csecs / 100, elapsed_csecs % 100, todo); - show_state(); read_lock(&tasklist_lock); do_each_thread(g, p) { task_lock(p); if (freezing(p) && !freezer_should_skip(p)) - printk(KERN_ERR " %s\n", p->comm); + sched_show_task(p); cancel_freezing(p); task_unlock(p); } while_each_thread(g, p); -- cgit v1.2.2 From 5a7aadfe2fcb0f69e2acc1fbefe22a096e792fc9 Mon Sep 17 00:00:00 2001 From: Matt Helsley Date: Fri, 26 Mar 2010 23:51:44 +0100 Subject: Freezer: Fix buggy resume test for tasks frozen with cgroup freezer When the cgroup freezer is used to freeze tasks we do not want to thaw those tasks during resume. Currently we test the cgroup freezer state of the resuming tasks to see if the cgroup is FROZEN. If so then we don't thaw the task. However, the FREEZING state also indicates that the task should remain frozen. This also avoids a problem pointed out by Oren Ladaan: the freezer state transition from FREEZING to FROZEN is updated lazily when userspace reads or writes the freezer.state file in the cgroup filesystem. This means that resume will thaw tasks in cgroups which should be in the FROZEN state if there is no read/write of the freezer.state file to trigger this transition before suspend. NOTE: Another "simple" solution would be to always update the cgroup freezer state during resume. However it's a bad choice for several reasons: Updating the cgroup freezer state is somewhat expensive because it requires walking all the tasks in the cgroup and checking if they are each frozen. Worse, this could easily make resume run in N^2 time where N is the number of tasks in the cgroup. Finally, updating the freezer state from this code path requires trickier locking because of the way locks must be ordered. Instead of updating the freezer state we rely on the fact that lazy updates only manage the transition from FREEZING to FROZEN. We know that a cgroup with the FREEZING state may actually be FROZEN so test for that state too. This makes sense in the resume path even for partially-frozen cgroups -- those that really are FREEZING but not FROZEN. Reported-by: Oren Ladaan Signed-off-by: Matt Helsley Cc: stable@kernel.org Signed-off-by: Rafael J. Wysocki --- kernel/cgroup_freezer.c | 9 ++++++--- kernel/power/process.c | 2 +- 2 files changed, 7 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 59e9ef6aab40..eb3f34d57419 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -47,17 +47,20 @@ static inline struct freezer *task_freezer(struct task_struct *task) struct freezer, css); } -int cgroup_frozen(struct task_struct *task) +int cgroup_freezing_or_frozen(struct task_struct *task) { struct freezer *freezer; enum freezer_state state; task_lock(task); freezer = task_freezer(task); - state = freezer->state; + if (!freezer->css.cgroup->parent) + state = CGROUP_THAWED; /* root cgroup can't be frozen */ + else + state = freezer->state; task_unlock(task); - return state == CGROUP_FROZEN; + return (state == CGROUP_FREEZING) || (state == CGROUP_FROZEN); } /* diff --git a/kernel/power/process.c b/kernel/power/process.c index a0480cd4daaf..71ae29052ab6 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -144,7 +144,7 @@ static void thaw_tasks(bool nosig_only) if (nosig_only && should_send_signal(p)) continue; - if (cgroup_frozen(p)) + if (cgroup_freezing_or_frozen(p)) continue; thaw_process(p); -- cgit v1.2.2 From 259354deaaf03d49a02dbb9975d6ec2a54675672 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 10 Mar 2010 18:56:10 +0900 Subject: module: encapsulate percpu handling better and record percpu_size Better encapsulate module static percpu area handling so that code outsidef of CONFIG_SMP ifdef doesn't deal with mod->percpu directly and add mod->percpu_size and record percpu_size in it. Both percpu fields are compiled out on UP. While at it, mark mod->percpu w/ __percpu. This is to prepare for is_module_percpu_address(). Signed-off-by: Tejun Heo Acked-by: Rusty Russell --- kernel/module.c | 66 ++++++++++++++++++++++++++++++--------------------------- 1 file changed, 35 insertions(+), 31 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index c968d3606dca..e7a6e53fc73e 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -370,27 +370,33 @@ EXPORT_SYMBOL_GPL(find_module); #ifdef CONFIG_SMP -static void *percpu_modalloc(unsigned long size, unsigned long align, - const char *name) +static inline void __percpu *mod_percpu(struct module *mod) { - void *ptr; + return mod->percpu; +} +static int percpu_modalloc(struct module *mod, + unsigned long size, unsigned long align) +{ if (align > PAGE_SIZE) { printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n", - name, align, PAGE_SIZE); + mod->name, align, PAGE_SIZE); align = PAGE_SIZE; } - ptr = __alloc_reserved_percpu(size, align); - if (!ptr) + mod->percpu = __alloc_reserved_percpu(size, align); + if (!mod->percpu) { printk(KERN_WARNING "Could not allocate %lu bytes percpu data\n", size); - return ptr; + return -ENOMEM; + } + mod->percpu_size = size; + return 0; } -static void percpu_modfree(void *freeme) +static void percpu_modfree(struct module *mod) { - free_percpu(freeme); + free_percpu(mod->percpu); } static unsigned int find_pcpusec(Elf_Ehdr *hdr, @@ -400,24 +406,28 @@ static unsigned int find_pcpusec(Elf_Ehdr *hdr, return find_sec(hdr, sechdrs, secstrings, ".data.percpu"); } -static void percpu_modcopy(void *pcpudest, const void *from, unsigned long size) +static void percpu_modcopy(struct module *mod, + const void *from, unsigned long size) { int cpu; for_each_possible_cpu(cpu) - memcpy(pcpudest + per_cpu_offset(cpu), from, size); + memcpy(per_cpu_ptr(mod->percpu, cpu), from, size); } #else /* ... !CONFIG_SMP */ -static inline void *percpu_modalloc(unsigned long size, unsigned long align, - const char *name) +static inline void __percpu *mod_percpu(struct module *mod) { return NULL; } -static inline void percpu_modfree(void *pcpuptr) +static inline int percpu_modalloc(struct module *mod, + unsigned long size, unsigned long align) +{ + return -ENOMEM; +} +static inline void percpu_modfree(struct module *mod) { - BUG(); } static inline unsigned int find_pcpusec(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, @@ -425,8 +435,8 @@ static inline unsigned int find_pcpusec(Elf_Ehdr *hdr, { return 0; } -static inline void percpu_modcopy(void *pcpudst, const void *src, - unsigned long size) +static inline void percpu_modcopy(struct module *mod, + const void *from, unsigned long size) { /* pcpusec should be 0, and size of that section should be 0. */ BUG_ON(size != 0); @@ -1400,8 +1410,7 @@ static void free_module(struct module *mod) /* This may be NULL, but that's OK */ module_free(mod, mod->module_init); kfree(mod->args); - if (mod->percpu) - percpu_modfree(mod->percpu); + percpu_modfree(mod); #if defined(CONFIG_MODULE_UNLOAD) if (mod->refptr) free_percpu(mod->refptr); @@ -1520,7 +1529,7 @@ static int simplify_symbols(Elf_Shdr *sechdrs, default: /* Divert to percpu allocation if a percpu var. */ if (sym[i].st_shndx == pcpuindex) - secbase = (unsigned long)mod->percpu; + secbase = (unsigned long)mod_percpu(mod); else secbase = sechdrs[sym[i].st_shndx].sh_addr; sym[i].st_value += secbase; @@ -1954,7 +1963,7 @@ static noinline struct module *load_module(void __user *umod, unsigned int modindex, versindex, infoindex, pcpuindex; struct module *mod; long err = 0; - void *percpu = NULL, *ptr = NULL; /* Stops spurious gcc warning */ + void *ptr = NULL; /* Stops spurious gcc warning */ unsigned long symoffs, stroffs, *strmap; mm_segment_t old_fs; @@ -2094,15 +2103,11 @@ static noinline struct module *load_module(void __user *umod, if (pcpuindex) { /* We have a special allocation for this section. */ - percpu = percpu_modalloc(sechdrs[pcpuindex].sh_size, - sechdrs[pcpuindex].sh_addralign, - mod->name); - if (!percpu) { - err = -ENOMEM; + err = percpu_modalloc(mod, sechdrs[pcpuindex].sh_size, + sechdrs[pcpuindex].sh_addralign); + if (err) goto free_mod; - } sechdrs[pcpuindex].sh_flags &= ~(unsigned long)SHF_ALLOC; - mod->percpu = percpu; } /* Determine total sizes, and put offsets in sh_entsize. For now @@ -2317,7 +2322,7 @@ static noinline struct module *load_module(void __user *umod, sort_extable(mod->extable, mod->extable + mod->num_exentries); /* Finally, copy percpu area over. */ - percpu_modcopy(mod->percpu, (void *)sechdrs[pcpuindex].sh_addr, + percpu_modcopy(mod, (void *)sechdrs[pcpuindex].sh_addr, sechdrs[pcpuindex].sh_size); add_kallsyms(mod, sechdrs, hdr->e_shnum, symindex, strindex, @@ -2409,8 +2414,7 @@ static noinline struct module *load_module(void __user *umod, module_free(mod, mod->module_core); /* mod will be freed with core. Don't access it beyond this line! */ free_percpu: - if (percpu) - percpu_modfree(percpu); + percpu_modfree(mod); free_mod: kfree(args); kfree(strmap); -- cgit v1.2.2 From 10fad5e46f6c7bdfb01b1a012380a38e3c6ab346 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 10 Mar 2010 18:57:54 +0900 Subject: percpu, module: implement and use is_kernel/module_percpu_address() lockdep has custom code to check whether a pointer belongs to static percpu area which is somewhat broken. Implement proper is_kernel/module_percpu_address() and replace the custom code. On UP, percpu variables are regular static variables and can't be distinguished from them. Always return %false on UP. Signed-off-by: Tejun Heo Acked-by: Peter Zijlstra Cc: Rusty Russell Cc: Ingo Molnar --- kernel/lockdep.c | 21 +++++---------------- kernel/module.c | 38 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 43 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index c927a549db2c..9bbb9c841e48 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -582,9 +582,6 @@ static int static_obj(void *obj) unsigned long start = (unsigned long) &_stext, end = (unsigned long) &_end, addr = (unsigned long) obj; -#ifdef CONFIG_SMP - int i; -#endif /* * static variable? @@ -595,24 +592,16 @@ static int static_obj(void *obj) if (arch_is_kernel_data(addr)) return 1; -#ifdef CONFIG_SMP /* - * percpu var? + * in-kernel percpu var? */ - for_each_possible_cpu(i) { - start = (unsigned long) &__per_cpu_start + per_cpu_offset(i); - end = (unsigned long) &__per_cpu_start + PERCPU_ENOUGH_ROOM - + per_cpu_offset(i); - - if ((addr >= start) && (addr < end)) - return 1; - } -#endif + if (is_kernel_percpu_address(addr)) + return 1; /* - * module var? + * module static or percpu var? */ - return is_module_address(addr); + return is_module_address(addr) || is_module_percpu_address(addr); } /* diff --git a/kernel/module.c b/kernel/module.c index e7a6e53fc73e..9f8d23d8b3a8 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -415,6 +415,40 @@ static void percpu_modcopy(struct module *mod, memcpy(per_cpu_ptr(mod->percpu, cpu), from, size); } +/** + * is_module_percpu_address - test whether address is from module static percpu + * @addr: address to test + * + * Test whether @addr belongs to module static percpu area. + * + * RETURNS: + * %true if @addr is from module static percpu area + */ +bool is_module_percpu_address(unsigned long addr) +{ + struct module *mod; + unsigned int cpu; + + preempt_disable(); + + list_for_each_entry_rcu(mod, &modules, list) { + if (!mod->percpu_size) + continue; + for_each_possible_cpu(cpu) { + void *start = per_cpu_ptr(mod->percpu, cpu); + + if ((void *)addr >= start && + (void *)addr < start + mod->percpu_size) { + preempt_enable(); + return true; + } + } + } + + preempt_enable(); + return false; +} + #else /* ... !CONFIG_SMP */ static inline void __percpu *mod_percpu(struct module *mod) @@ -441,6 +475,10 @@ static inline void percpu_modcopy(struct module *mod, /* pcpusec should be 0, and size of that section should be 0. */ BUG_ON(size != 0); } +bool is_module_percpu_address(unsigned long addr) +{ + return false; +} #endif /* CONFIG_SMP */ -- cgit v1.2.2 From 88be12c440cfa2fa3f5be83507360aac9ea1c54e Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Mon, 29 Mar 2010 12:01:50 +0100 Subject: slow-work: use get_ref wrapper instead of directly calling get_ref Otherwise we can get an oops if the user has no get_ref/put_ref requirement. Signed-off-by: Dave Airlie Signed-off-by: David Howells Signed-off-by: Linus Torvalds --- kernel/slow-work.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/slow-work.c b/kernel/slow-work.c index 7494bbf5a270..7d3f4fa9ef4f 100644 --- a/kernel/slow-work.c +++ b/kernel/slow-work.c @@ -637,7 +637,7 @@ int delayed_slow_work_enqueue(struct delayed_slow_work *dwork, goto cancelled; /* the timer holds a reference whilst it is pending */ - ret = work->ops->get_ref(work); + ret = slow_work_get_ref(work); if (ret < 0) goto cant_get_ref; -- cgit v1.2.2 From a53f4f9efaeb1d87cfae066346979d4d70e1abe9 Mon Sep 17 00:00:00 2001 From: David Howells Date: Mon, 29 Mar 2010 13:08:52 +0100 Subject: SLOW_WORK: CONFIG_SLOW_WORK_PROC should be CONFIG_SLOW_WORK_DEBUG CONFIG_SLOW_WORK_PROC was changed to CONFIG_SLOW_WORK_DEBUG, but not in all instances. Change the remaining instances. This makes the debugfs file display the time mark and the owner's description again. Signed-off-by: David Howells Signed-off-by: Linus Torvalds --- kernel/slow-work.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/slow-work.h b/kernel/slow-work.h index 321f3c59d732..a29ebd1ef41d 100644 --- a/kernel/slow-work.h +++ b/kernel/slow-work.h @@ -43,28 +43,28 @@ extern void slow_work_new_thread_desc(struct slow_work *, struct seq_file *); */ static inline void slow_work_set_thread_pid(int id, pid_t pid) { -#ifdef CONFIG_SLOW_WORK_PROC +#ifdef CONFIG_SLOW_WORK_DEBUG slow_work_pids[id] = pid; #endif } static inline void slow_work_mark_time(struct slow_work *work) { -#ifdef CONFIG_SLOW_WORK_PROC +#ifdef CONFIG_SLOW_WORK_DEBUG work->mark = CURRENT_TIME; #endif } static inline void slow_work_begin_exec(int id, struct slow_work *work) { -#ifdef CONFIG_SLOW_WORK_PROC +#ifdef CONFIG_SLOW_WORK_DEBUG slow_work_execs[id] = work; #endif } static inline void slow_work_end_exec(int id, struct slow_work *work) { -#ifdef CONFIG_SLOW_WORK_PROC +#ifdef CONFIG_SLOW_WORK_DEBUG write_lock(&slow_work_execs_lock); slow_work_execs[id] = NULL; write_unlock(&slow_work_execs_lock); -- cgit v1.2.2 From eed63519e3e74d515d2007ecd895338d0ba2a85c Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Sun, 28 Mar 2010 19:42:56 -0700 Subject: x86: Do not free zero sized per cpu areas This avoids an infinite loop in free_early_partial(). Add a warning to free_early_partial() to catch future problems. -v5: put back start > end back into WARN_ONCE() -v6: use one line for warning, suggested by Linus -v7: more tests -v8: remove the function name as suggested by Johannes WARN_ONCE() will print out that function name. Signed-off-by: Ian Campbell Signed-off-by: Yinghai Lu Tested-by: Konrad Rzeszutek Wilk Tested-by: Joel Becker Tested-by: Stanislaw Gruszka Acked-by: Johannes Weiner Cc: Peter Zijlstra Cc: David Miller Cc: Benjamin Herrenschmidt Cc: Linus Torvalds LKML-Reference: <1269830604-26214-4-git-send-email-yinghai@kernel.org> Signed-off-by: Ingo Molnar --- kernel/early_res.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/early_res.c b/kernel/early_res.c index 3cb2c661bb78..31aa9332ef3f 100644 --- a/kernel/early_res.c +++ b/kernel/early_res.c @@ -333,6 +333,12 @@ void __init free_early_partial(u64 start, u64 end) struct early_res *r; int i; + if (start == end) + return; + + if (WARN_ONCE(start > end, " wrong range [%#llx, %#llx]\n", start, end)) + return; + try_next: i = find_overlapped_early(start, end); if (i >= max_early_res) -- cgit v1.2.2 From e36673ec5126f15a8cddf6049aede7bdcf484c26 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Mar 2010 10:57:37 +0800 Subject: tracing: Fix lockdep warning in global_clock() # echo 1 > events/enable # echo global > trace_clock ------------[ cut here ]------------ WARNING: at kernel/lockdep.c:3162 check_flags+0xb2/0x190() ... ---[ end trace 3f86734a89416623 ]--- possible reason: unannotated irqs-on. ... There's no reason to use the raw_local_irq_save() in trace_clock_global. The local_irq_save() version is fine, and does not cause the bug in lockdep. Acked-by: Peter Zijlstra Signed-off-by: Li Zefan LKML-Reference: <4BA97FA1.7030606@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c index 6fbfb8f417b9..9d589d8dcd1a 100644 --- a/kernel/trace/trace_clock.c +++ b/kernel/trace/trace_clock.c @@ -84,7 +84,7 @@ u64 notrace trace_clock_global(void) int this_cpu; u64 now; - raw_local_irq_save(flags); + local_irq_save(flags); this_cpu = raw_smp_processor_id(); now = cpu_clock(this_cpu); @@ -110,7 +110,7 @@ u64 notrace trace_clock_global(void) arch_spin_unlock(&trace_clock_struct.lock); out: - raw_local_irq_restore(flags); + local_irq_restore(flags); return now; } -- cgit v1.2.2 From 292f60c0c4ab44aa2d589ba03c12e64a3b3c5e38 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Mon, 29 Mar 2010 17:37:02 +0200 Subject: ring-buffer: Add missing unlock In some error handling cases the lock is not unlocked. The return is converted to a goto, to share the unlock at the end of the function. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // @r exists@ expression E1; identifier f; @@ f (...) { <+... * spin_lock_irq (E1,...); ... when != E1 * return ...; ...+> } // Signed-off-by: Julia Lawall LKML-Reference: Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d1187ef20caf..9a0f9bf6a37b 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1209,18 +1209,19 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages) for (i = 0; i < nr_pages; i++) { if (RB_WARN_ON(cpu_buffer, list_empty(cpu_buffer->pages))) - return; + goto out; p = cpu_buffer->pages->next; bpage = list_entry(p, struct buffer_page, list); list_del_init(&bpage->list); free_buffer_page(bpage); } if (RB_WARN_ON(cpu_buffer, list_empty(cpu_buffer->pages))) - return; + goto out; rb_reset_cpu(cpu_buffer); rb_check_pages(cpu_buffer); +out: spin_unlock_irq(&cpu_buffer->reader_lock); } @@ -1237,7 +1238,7 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer, for (i = 0; i < nr_pages; i++) { if (RB_WARN_ON(cpu_buffer, list_empty(pages))) - return; + goto out; p = pages->next; bpage = list_entry(p, struct buffer_page, list); list_del_init(&bpage->list); @@ -1246,6 +1247,7 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer, rb_reset_cpu(cpu_buffer); rb_check_pages(cpu_buffer); +out: spin_unlock_irq(&cpu_buffer->reader_lock); } -- cgit v1.2.2 From 570b8fb505896e007fd3bb07573ba6640e51851d Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Tue, 30 Mar 2010 00:04:00 +0100 Subject: CRED: Fix memory leak in error handling Fix a memory leak on an OOM condition in prepare_usermodehelper_creds(). Signed-off-by: Mathieu Desnoyers Signed-off-by: David Howells Signed-off-by: James Morris --- kernel/cred.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cred.c b/kernel/cred.c index 1ed8ca18790c..1b1129d0cce8 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -364,7 +364,7 @@ struct cred *prepare_usermodehelper_creds(void) new = kmem_cache_alloc(cred_jar, GFP_ATOMIC); if (!new) - return NULL; + goto free_tgcred; kdebug("prepare_usermodehelper_creds() alloc %p", new); @@ -397,6 +397,10 @@ struct cred *prepare_usermodehelper_creds(void) error: put_cred(new); +free_tgcred: +#ifdef CONFIG_KEYS + kfree(tgcred); +#endif return NULL; } -- cgit v1.2.2 From 5a0e3ad6af8660be21ca98a971cd00f331318c05 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 24 Mar 2010 17:04:11 +0900 Subject: include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo Guess-its-ok-by: Christoph Lameter Cc: Ingo Molnar Cc: Lee Schermerhorn --- kernel/async.c | 1 + kernel/audit.c | 1 + kernel/audit_tree.c | 1 + kernel/audit_watch.c | 1 + kernel/auditfilter.c | 1 + kernel/auditsc.c | 1 + kernel/cgroup_freezer.c | 1 + kernel/compat.c | 1 + kernel/cpu.c | 1 + kernel/cred.c | 1 + kernel/irq/numa_migrate.c | 1 + kernel/irq/proc.c | 1 + kernel/kallsyms.c | 1 + kernel/latencytop.c | 1 - kernel/lockdep.c | 1 + kernel/nsproxy.c | 1 + kernel/padata.c | 1 + kernel/perf_event.c | 1 + kernel/pid_namespace.c | 1 + kernel/power/hibernate.c | 1 + kernel/power/hibernate_nvs.c | 1 + kernel/power/snapshot.c | 1 + kernel/power/suspend.c | 1 + kernel/power/swap.c | 1 + kernel/res_counter.c | 1 - kernel/sched.c | 1 + kernel/sched_cpupri.c | 1 + kernel/smp.c | 1 + kernel/srcu.c | 1 - kernel/sys.c | 1 + kernel/sysctl_binary.c | 1 + kernel/taskstats.c | 1 + kernel/time.c | 1 - kernel/time/timecompare.c | 1 + kernel/timer.c | 1 + kernel/trace/blktrace.c | 1 + kernel/trace/ftrace.c | 1 + kernel/trace/power-traces.c | 1 - kernel/trace/ring_buffer.c | 1 + kernel/trace/trace.c | 2 +- kernel/trace/trace_events.c | 1 + kernel/trace/trace_events_filter.c | 1 + kernel/trace/trace_functions_graph.c | 1 + kernel/trace/trace_ksym.c | 1 + kernel/trace/trace_mmiotrace.c | 1 + kernel/trace/trace_selftest.c | 1 + kernel/trace/trace_stat.c | 1 + kernel/trace/trace_syscalls.c | 1 + kernel/trace/trace_workqueue.c | 1 + 49 files changed, 44 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/async.c b/kernel/async.c index 27235f5de198..15319d6c18fe 100644 --- a/kernel/async.c +++ b/kernel/async.c @@ -56,6 +56,7 @@ asynchronous and synchronous parts of the kernel. #include #include #include +#include #include static async_cookie_t next_cookie = 1; diff --git a/kernel/audit.c b/kernel/audit.c index 78f7f86aa238..c71bd26631a2 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -46,6 +46,7 @@ #include #include #include +#include #include #include diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index 028e85663f27..46a57b57a335 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -3,6 +3,7 @@ #include #include #include +#include struct audit_tree; struct audit_chunk; diff --git a/kernel/audit_watch.c b/kernel/audit_watch.c index cc7e87936cbc..8df43696f4ba 100644 --- a/kernel/audit_watch.c +++ b/kernel/audit_watch.c @@ -27,6 +27,7 @@ #include #include #include +#include #include #include #include "audit.h" diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index a70604047f3c..ce08041f578d 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -27,6 +27,7 @@ #include #include #include +#include #include #include "audit.h" diff --git a/kernel/auditsc.c b/kernel/auditsc.c index f3a461c0970a..97a3cef81b9c 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -49,6 +49,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 59e9ef6aab40..d2ccd2798d7a 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -15,6 +15,7 @@ */ #include +#include #include #include #include diff --git a/kernel/compat.c b/kernel/compat.c index f6c204f07ea6..7f40e9275fd9 100644 --- a/kernel/compat.c +++ b/kernel/compat.c @@ -25,6 +25,7 @@ #include #include #include +#include #include diff --git a/kernel/cpu.c b/kernel/cpu.c index f8cced2692b3..25bba73b1be3 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -14,6 +14,7 @@ #include #include #include +#include #ifdef CONFIG_SMP /* Serializes the updates to cpu_online_mask, cpu_present_mask */ diff --git a/kernel/cred.c b/kernel/cred.c index 1ed8ca18790c..d84bdef17c9f 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -10,6 +10,7 @@ */ #include #include +#include #include #include #include diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index 963559dbd858..65d3845665ac 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -6,6 +6,7 @@ */ #include +#include #include #include #include diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index 6f50eccc79c0..7a6eb04ef6b5 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -7,6 +7,7 @@ */ #include +#include #include #include #include diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index 8e5288a8a355..13aff293f4de 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -21,6 +21,7 @@ #include /* for cond_resched */ #include #include +#include #include diff --git a/kernel/latencytop.c b/kernel/latencytop.c index ca07c5c0c914..877fb306d415 100644 --- a/kernel/latencytop.c +++ b/kernel/latencytop.c @@ -56,7 +56,6 @@ #include #include #include -#include #include static DEFINE_SPINLOCK(latency_lock); diff --git a/kernel/lockdep.c b/kernel/lockdep.c index c927a549db2c..367f72452ff4 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -43,6 +43,7 @@ #include #include #include +#include #include diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 2ab67233ee8f..f74e6c00e26d 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -13,6 +13,7 @@ * Pavel Emelianov */ +#include #include #include #include diff --git a/kernel/padata.c b/kernel/padata.c index 93caf65ff57c..fd03513c7327 100644 --- a/kernel/padata.c +++ b/kernel/padata.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #define MAX_SEQ_NR INT_MAX - NR_CPUS diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 574ee58a3046..a77266e3e3e1 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 79aac93acf99..a5aff94e1f0b 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -13,6 +13,7 @@ #include #include #include +#include #define BITS_PER_PAGE (PAGE_SIZE*8) diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index da5288ec2392..aa9e916da4d5 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include diff --git a/kernel/power/hibernate_nvs.c b/kernel/power/hibernate_nvs.c index 39ac698ef836..fdcad9ed5a7b 100644 --- a/kernel/power/hibernate_nvs.c +++ b/kernel/power/hibernate_nvs.c @@ -10,6 +10,7 @@ #include #include #include +#include #include /* diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 830cadecbdfc..be861c26dda7 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -26,6 +26,7 @@ #include #include #include +#include #include #include diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index 44cce10b582d..56e7dbb8b996 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -15,6 +15,7 @@ #include #include #include +#include #include "power.h" diff --git a/kernel/power/swap.c b/kernel/power/swap.c index 1d575733d4e1..66824d71983a 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -23,6 +23,7 @@ #include #include #include +#include #include "power.h" diff --git a/kernel/res_counter.c b/kernel/res_counter.c index bcdabf37c40b..c7eaa37a768b 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -10,7 +10,6 @@ #include #include #include -#include #include #include #include diff --git a/kernel/sched.c b/kernel/sched.c index 49d2fa7b687a..86c7cc1d7c9d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -71,6 +71,7 @@ #include #include #include +#include #include #include diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index fccf9fbb0d7b..e6871cb3fc83 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -27,6 +27,7 @@ * of the License. */ +#include #include "sched_cpupri.h" /* Convert between a 140 based task->prio, and our 102 based cpupri */ diff --git a/kernel/smp.c b/kernel/smp.c index 9867b6bfefce..3fc697336183 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -9,6 +9,7 @@ #include #include #include +#include #include #include diff --git a/kernel/srcu.c b/kernel/srcu.c index bde4295774c8..2980da3fd509 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -30,7 +30,6 @@ #include #include #include -#include #include #include diff --git a/kernel/sys.c b/kernel/sys.c index 8298878f4f71..6d1a7e0f9d5b 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -36,6 +36,7 @@ #include #include #include +#include #include #include diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c index 8cd50d8f9bde..59030570f5ca 100644 --- a/kernel/sysctl_binary.c +++ b/kernel/sysctl_binary.c @@ -13,6 +13,7 @@ #include #include #include +#include #ifdef CONFIG_SYSCTL_SYSCALL diff --git a/kernel/taskstats.c b/kernel/taskstats.c index 899ca51be5e8..11281d5792bd 100644 --- a/kernel/taskstats.c +++ b/kernel/taskstats.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/time.c b/kernel/time.c index 804798005d19..656dccfe1cbb 100644 --- a/kernel/time.c +++ b/kernel/time.c @@ -35,7 +35,6 @@ #include #include #include -#include #include #include diff --git a/kernel/time/timecompare.c b/kernel/time/timecompare.c index 12f5c55090be..ac38fbb176cc 100644 --- a/kernel/time/timecompare.c +++ b/kernel/time/timecompare.c @@ -19,6 +19,7 @@ #include #include +#include #include /* diff --git a/kernel/timer.c b/kernel/timer.c index fc965eae0e87..aeb6a54f2771 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -39,6 +39,7 @@ #include #include #include +#include #include #include diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 07f945a99430..b3bc91a3f510 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index d9062f5cc0c0..2404b59b3097 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/trace/power-traces.c b/kernel/trace/power-traces.c index 9f4f565b01e6..a22582a06161 100644 --- a/kernel/trace/power-traces.c +++ b/kernel/trace/power-traces.c @@ -9,7 +9,6 @@ #include #include #include -#include #define CREATE_TRACE_POINTS #include diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d1187ef20caf..2c839ca5e5ce 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3ec2ee6f6560..44f916a04065 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -33,10 +33,10 @@ #include #include #include +#include #include #include #include -#include #include #include "trace.h" diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index beab8bf2f310..c697c7043349 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 4615f62a04f1..88c0b6dbd7fe 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -22,6 +22,7 @@ #include #include #include +#include #include "trace.h" #include "trace_output.h" diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index e6989d9b44da..9aed1a5cf553 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -9,6 +9,7 @@ #include #include #include +#include #include #include "trace.h" diff --git a/kernel/trace/trace_ksym.c b/kernel/trace/trace_ksym.c index 94103cdcf9d8..d59cd6879477 100644 --- a/kernel/trace/trace_ksym.c +++ b/kernel/trace/trace_ksym.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include "trace_output.h" diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index 0acd834659ed..017fa376505d 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -9,6 +9,7 @@ #include #include #include +#include #include #include diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 280fea470d67..81003b4d617f 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -3,6 +3,7 @@ #include #include #include +#include static inline int trace_valid_entry(struct trace_entry *entry) { diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index a4bb239eb987..96cffb269e73 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -10,6 +10,7 @@ #include +#include #include #include #include "trace_stat.h" diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 33c2a5b769dc..4d6d711717f2 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -1,5 +1,6 @@ #include #include +#include #include #include #include diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index 40cafb07dffd..cc2d2faa7d9e 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -9,6 +9,7 @@ #include #include #include +#include #include #include "trace_stat.h" #include "trace.h" -- cgit v1.2.2 From 753649dbc49345a73a2454c770a3f2d54d11aec6 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 31 Mar 2010 13:30:19 +0200 Subject: genirq: Force MSI irq handlers to run with interrupts disabled Network folks reported that directing all MSI-X vectors of their multi queue NICs to a single core can cause interrupt stack overflows when enough interrupts fire at the same time. This is caused by the fact that we run interrupt handlers by default with interrupts enabled unless the driver reuqests the interrupt with the IRQF_DISABLED set. The NIC handlers do not set this flag, so simultaneous interrupts can nest unlimited and cause the stack overflow. The only safe counter measure is to run the interrupt handlers with interrupts disabled. We can't switch to this mode in general right now, but it is safe to do so for MSI interrupts. Force IRQF_DISABLED for MSI interrupt handlers. Signed-off-by: Thomas Gleixner Cc: Andi Kleen Cc: Linus Torvalds Cc: Andrew Morton Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Alan Cox Cc: David Miller Cc: Greg Kroah-Hartman Cc: Arnaldo Carvalho de Melo Cc: stable@kernel.org --- kernel/irq/manage.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 398fda155f6e..704e488730a5 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -757,6 +757,16 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) if (new->flags & IRQF_ONESHOT) desc->status |= IRQ_ONESHOT; + /* + * Force MSI interrupts to run with interrupts + * disabled. The multi vector cards can cause stack + * overflows due to nested interrupts when enough of + * them are directed to a core and fire at the same + * time. + */ + if (desc->msi_desc) + new->flags |= IRQF_DISABLED; + if (!(desc->status & IRQ_NOAUTOEN)) { desc->depth = 0; desc->status &= ~IRQ_DISABLED; -- cgit v1.2.2 From eb1e79611cc9bfe21978230e3521e77ea2d7874a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Mar 2010 00:08:59 +0100 Subject: perf: Correctly align perf event tracing buffer The trace event buffer used by perf to record raw sample events is typed as an array of char and may then not be aligned to 8 by alloc_percpu(). But we need it to be aligned to 8 in sparc64 because we cast this buffer into a random structure type built by the TRACE_EVENT() macro to store the traces. So if a random 64 bits field is accessed inside, it may be not under an expected good alignment. Use an array of long instead to force the appropriate alignment, and perform a compile time check to ensure the size in byte of the buffer is a multiple of sizeof(long) so that its actual size doesn't get shrinked under us. This fixes unaligned accesses reported while using perf lock in sparc 64. Suggested-by: David Miller Suggested-by: Tejun Heo Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras Cc: Ingo Molnar Cc: David Miller Cc: Steven Rostedt --- kernel/trace/trace_event_perf.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 81f691eb3a30..0565bb42566f 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -17,7 +17,12 @@ EXPORT_SYMBOL_GPL(perf_arch_fetch_caller_regs); static char *perf_trace_buf; static char *perf_trace_buf_nmi; -typedef typeof(char [PERF_MAX_TRACE_SIZE]) perf_trace_t ; +/* + * Force it to be aligned to unsigned long to avoid misaligned accesses + * suprises + */ +typedef typeof(unsigned long [PERF_MAX_TRACE_SIZE / sizeof(unsigned long)]) + perf_trace_t; /* Count the events in use (per event id, not per instance) */ static int total_ref_count; @@ -130,6 +135,8 @@ __kprobes void *perf_trace_buf_prepare(int size, unsigned short type, char *trace_buf, *raw_data; int pc, cpu; + BUILD_BUG_ON(PERF_MAX_TRACE_SIZE % sizeof(unsigned long)); + pc = preempt_count(); /* Protect the per cpu buffer, begin the rcu read side */ @@ -152,7 +159,7 @@ __kprobes void *perf_trace_buf_prepare(int size, unsigned short type, raw_data = per_cpu_ptr(trace_buf, cpu); /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + memset(&raw_data[size - sizeof(u64)], 0, sizeof(u64)); entry = (struct trace_entry *)raw_data; tracing_generic_entry_update(entry, *irq_flags, pc); -- cgit v1.2.2 From e49a5bd38159dfb1928fd25b173bc9de4bbadb21 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 22 Mar 2010 19:40:03 +0100 Subject: perf: Use hot regs with software sched switch/migrate events Scheduler's task migration events don't work because they always pass NULL regs perf_sw_event(). The event hence gets filtered in perf_swevent_add(). Scheduler's context switches events use task_pt_regs() to get the context when the event occured which is a wrong thing to do as this won't give us the place in the kernel where we went to sleep but the place where we left userspace. The result is even more wrong if we switch from a kernel thread. Use the hot regs snapshot for both events as they belong to the non-interrupt/exception based events family. Unlike page faults or so that provide the regs matching the exact origin of the event, we need to save the current context. This makes the task migration event working and fix the context switch callchains and origin ip. Example: perf record -a -e cs Before: 10.91% ksoftirqd/0 0 [k] 0000000000000000 | --- (nil) perf_callchain perf_prepare_sample __perf_event_overflow perf_swevent_overflow perf_swevent_add perf_swevent_ctx_event do_perf_sw_event __perf_sw_event perf_event_task_sched_out schedule run_ksoftirqd kthread kernel_thread_helper After: 23.77% hald-addon-stor [kernel.kallsyms] [k] schedule | --- schedule | |--60.00%-- schedule_timeout | wait_for_common | wait_for_completion | blk_execute_rq | scsi_execute | scsi_execute_req | sr_test_unit_ready | | | |--66.67%-- sr_media_change | | media_changed | | cdrom_media_changed | | sr_block_media_changed | | check_disk_change | | cdrom_open v2: Always build perf_arch_fetch_caller_regs() now that software events need that too. They don't need it from modules, unlike trace events, so we keep the EXPORT_SYMBOL in trace_event_perf.c Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras Cc: Ingo Molnar Cc: David Miller --- kernel/perf_event.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 574ee58a3046..b0feb4795af3 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1164,11 +1164,9 @@ void perf_event_task_sched_out(struct task_struct *task, struct perf_event_context *ctx = task->perf_event_ctxp; struct perf_event_context *next_ctx; struct perf_event_context *parent; - struct pt_regs *regs; int do_switch = 1; - regs = task_pt_regs(task); - perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, regs, 0); + perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, NULL, 0); if (likely(!ctx || !cpuctx->task_ctx)) return; -- cgit v1.2.2 From 8bb39f9aa068262732fe44b965d7a6eb5a5a7d67 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Fri, 26 Mar 2010 11:11:33 +0100 Subject: perf: Fix 'perf sched record' deadlock perf sched record can deadlock a box should the holder of handle->data->lock take an interrupt, and then attempt to acquire an rq lock held by a CPU trying to acquire the same lock. Disable interrupts. CPU0 CPU1 sched event with rq->lock held grab handle->data->lock spin on handle->data->lock interrupt try to grab rq->lock Reported-by: Li Zefan Signed-off-by: Mike Galbraith Tested-by: Li Zefan Signed-off-by: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <1269598293.6174.8.camel@marge.simson.net> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index b0feb4795af3..96aae13c7960 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3376,15 +3376,23 @@ static void perf_event_task_output(struct perf_event *event, struct perf_task_event *task_event) { struct perf_output_handle handle; - int size; struct task_struct *task = task_event->task; - int ret; + unsigned long flags; + int size, ret; + + /* + * If this CPU attempts to acquire an rq lock held by a CPU spinning + * in perf_output_lock() from interrupt context, it's game over. + */ + local_irq_save(flags); size = task_event->event_id.header.size; ret = perf_output_begin(&handle, event, size, 0, 0); - if (ret) + if (ret) { + local_irq_restore(flags); return; + } task_event->event_id.pid = perf_event_pid(event, task); task_event->event_id.ppid = perf_event_pid(event, current); @@ -3395,6 +3403,7 @@ static void perf_event_task_output(struct perf_event *event, perf_output_put(&handle, task_event->event_id); perf_output_end(&handle); + local_irq_restore(flags); } static int perf_event_task_match(struct perf_event *event) -- cgit v1.2.2 From 269484a492d9177072ee11ec8c9bff71d256837a Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Tue, 30 Mar 2010 11:09:53 +0200 Subject: sched: Fix proc_sched_set_task() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Latencytop clearing sum_exec_runtime via proc_sched_set_task() breaks task_times(). Other places in kernel use nvcsw and nivcsw, which are being cleared as well, Clear task statistics only. Reported-by: Török Edwin Signed-off-by: Mike Galbraith Cc: Hidetoshi Seto Cc: Arjan van de Ven Signed-off-by: Peter Zijlstra LKML-Reference: <1269940193.19286.14.camel@marge.simson.net> Signed-off-by: Ingo Molnar --- kernel/sched_debug.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index 67f95aada4b9..9b49db144037 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -518,8 +518,4 @@ void proc_sched_set_task(struct task_struct *p) p->se.nr_wakeups_idle = 0; p->sched_info.bkl_count = 0; #endif - p->se.sum_exec_runtime = 0; - p->se.prev_sum_exec_runtime = 0; - p->nvcsw = 0; - p->nivcsw = 0; } -- cgit v1.2.2 From 47a70985e5c093ae03d8ccf633c70a93761d86f2 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 30 Mar 2010 18:58:29 +0200 Subject: sched: set_cpus_allowed_ptr(): Don't use rq->migration_thread after unlock Trivial typo fix. rq->migration_thread can be NULL after task_rq_unlock(), this is why we have "mt" which should be used instead. Signed-off-by: Oleg Nesterov Signed-off-by: Peter Zijlstra LKML-Reference: <20100330165829.GA18284@redhat.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 49d2fa7b687a..528a10592c16 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5387,7 +5387,7 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) get_task_struct(mt); task_rq_unlock(rq, &flags); - wake_up_process(rq->migration_thread); + wake_up_process(mt); put_task_struct(mt); wait_for_completion(&req.done); tlb_migrate_finish(p->mm); -- cgit v1.2.2 From a0279bd58060ccedbd414edf97d50cfa3778c370 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Fri, 2 Apr 2010 11:33:29 -0500 Subject: kgdb: have ebin2mem call probe_kernel_write once Rather than call probe_kernel_write() one byte at a time, process the whole buffer locally and pass the entire result in one go. This way, architectures that need to do special handling based on the length can do so, or we only end up calling memcpy() once. [sonic.zhang@analog.com: Reported original problem and preliminary patch] Signed-off-by: Jason Wessel Signed-off-by: Sonic Zhang Signed-off-by: Mike Frysinger --- kernel/kgdb.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 761fdd2b3034..42fd128127a6 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -391,27 +391,22 @@ int kgdb_mem2hex(char *mem, char *buf, int count) /* * Copy the binary array pointed to by buf into mem. Fix $, #, and - * 0x7d escaped with 0x7d. Return a pointer to the character after - * the last byte written. + * 0x7d escaped with 0x7d. Return -EFAULT on failure or 0 on success. + * The input buf is overwitten with the result to write to mem. */ static int kgdb_ebin2mem(char *buf, char *mem, int count) { - int err = 0; - char c; + int size = 0; + char *c = buf; while (count-- > 0) { - c = *buf++; - if (c == 0x7d) - c = *buf++ ^ 0x20; - - err = probe_kernel_write(mem, &c, 1); - if (err) - break; - - mem++; + c[size] = *buf++; + if (c[size] == 0x7d) + c[size] = *buf++ ^ 0x20; + size++; } - return err; + return probe_kernel_write(mem, c, size); } /* -- cgit v1.2.2 From 62fae312197a8fbcd3727261d59f5a6bd0dbf158 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Fri, 2 Apr 2010 11:47:02 -0500 Subject: kgdb: eliminate kgdb_wait(), all cpus enter the same way This is a kgdb architectural change to have all the cpus (master or slave) enter the same function. A cpu that hits an exception (wants to be the master cpu) will call kgdb_handle_exception() from the trap handler and then invoke a kgdb_roundup_cpu() to synchronize the other cpus and bring them into the kgdb_handle_exception() as well. A slave cpu will enter kgdb_handle_exception() from the kgdb_nmicallback() and set the exception state to note that the processor is a slave. Previously the salve cpu would have called kgdb_wait(). This change allows the debug core to change cpus without resuming the system in order to inspect arch specific cpu information. Signed-off-by: Jason Wessel --- kernel/kgdb.c | 165 +++++++++++++++++++++++++++++----------------------------- 1 file changed, 82 insertions(+), 83 deletions(-) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 42fd128127a6..6882c047452d 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -69,9 +69,16 @@ struct kgdb_state { struct pt_regs *linux_regs; }; +/* Exception state values */ +#define DCPU_WANT_MASTER 0x1 /* Waiting to become a master kgdb cpu */ +#define DCPU_NEXT_MASTER 0x2 /* Transition from one master cpu to another */ +#define DCPU_IS_SLAVE 0x4 /* Slave cpu enter exception */ +#define DCPU_SSTEP 0x8 /* CPU is single stepping */ + static struct debuggerinfo_struct { void *debuggerinfo; struct task_struct *task; + int exception_state; } kgdb_info[NR_CPUS]; /** @@ -557,49 +564,6 @@ static struct task_struct *getthread(struct pt_regs *regs, int tid) return find_task_by_pid_ns(tid, &init_pid_ns); } -/* - * CPU debug state control: - */ - -#ifdef CONFIG_SMP -static void kgdb_wait(struct pt_regs *regs) -{ - unsigned long flags; - int cpu; - - local_irq_save(flags); - cpu = raw_smp_processor_id(); - kgdb_info[cpu].debuggerinfo = regs; - kgdb_info[cpu].task = current; - /* - * Make sure the above info reaches the primary CPU before - * our cpu_in_kgdb[] flag setting does: - */ - smp_wmb(); - atomic_set(&cpu_in_kgdb[cpu], 1); - - /* Disable any cpu specific hw breakpoints */ - kgdb_disable_hw_debug(regs); - - /* Wait till primary CPU is done with debugging */ - while (atomic_read(&passive_cpu_wait[cpu])) - cpu_relax(); - - kgdb_info[cpu].debuggerinfo = NULL; - kgdb_info[cpu].task = NULL; - - /* fix up hardware debug registers on local cpu */ - if (arch_kgdb_ops.correct_hw_break) - arch_kgdb_ops.correct_hw_break(); - - /* Signal the primary CPU that we are done: */ - atomic_set(&cpu_in_kgdb[cpu], 0); - touch_softlockup_watchdog_sync(); - clocksource_touch_watchdog(); - local_irq_restore(flags); -} -#endif - /* * Some architectures need cache flushes when we set/clear a * breakpoint: @@ -1395,34 +1359,12 @@ static int kgdb_reenter_check(struct kgdb_state *ks) return 1; } -/* - * kgdb_handle_exception() - main entry point from a kernel exception - * - * Locking hierarchy: - * interface locks, if any (begin_session) - * kgdb lock (kgdb_active) - */ -int -kgdb_handle_exception(int evector, int signo, int ecode, struct pt_regs *regs) +static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs) { - struct kgdb_state kgdb_var; - struct kgdb_state *ks = &kgdb_var; unsigned long flags; int sstep_tries = 100; int error = 0; int i, cpu; - - ks->cpu = raw_smp_processor_id(); - ks->ex_vector = evector; - ks->signo = signo; - ks->ex_vector = evector; - ks->err_code = ecode; - ks->kgdb_usethreadid = 0; - ks->linux_regs = regs; - - if (kgdb_reenter_check(ks)) - return 0; /* Ouch, double exception ! */ - acquirelock: /* * Interrupts will be restored by the 'trap return' code, except when @@ -1430,13 +1372,42 @@ acquirelock: */ local_irq_save(flags); - cpu = raw_smp_processor_id(); + cpu = ks->cpu; + kgdb_info[cpu].debuggerinfo = regs; + kgdb_info[cpu].task = current; + /* + * Make sure the above info reaches the primary CPU before + * our cpu_in_kgdb[] flag setting does: + */ + smp_wmb(); + atomic_set(&cpu_in_kgdb[cpu], 1); /* - * Acquire the kgdb_active lock: + * CPU will loop if it is a slave or request to become a kgdb + * master cpu and acquire the kgdb_active lock: */ - while (atomic_cmpxchg(&kgdb_active, -1, cpu) != -1) + while (1) { + if (kgdb_info[cpu].exception_state & DCPU_WANT_MASTER) { + if (atomic_cmpxchg(&kgdb_active, -1, cpu) == cpu) + break; + } else if (kgdb_info[cpu].exception_state & DCPU_IS_SLAVE) { + if (!atomic_read(&passive_cpu_wait[cpu])) + goto return_normal; + } else { +return_normal: + /* Return to normal operation by executing any + * hw breakpoint fixup. + */ + if (arch_kgdb_ops.correct_hw_break) + arch_kgdb_ops.correct_hw_break(); + atomic_set(&cpu_in_kgdb[cpu], 0); + touch_softlockup_watchdog_sync(); + clocksource_touch_watchdog(); + local_irq_restore(flags); + return 0; + } cpu_relax(); + } /* * For single stepping, try to only enter on the processor @@ -1470,9 +1441,6 @@ acquirelock: if (kgdb_io_ops->pre_exception) kgdb_io_ops->pre_exception(); - kgdb_info[ks->cpu].debuggerinfo = ks->linux_regs; - kgdb_info[ks->cpu].task = current; - kgdb_disable_hw_debug(ks->linux_regs); /* @@ -1484,12 +1452,6 @@ acquirelock: atomic_set(&passive_cpu_wait[i], 1); } - /* - * spin_lock code is good enough as a barrier so we don't - * need one here: - */ - atomic_set(&cpu_in_kgdb[ks->cpu], 1); - #ifdef CONFIG_SMP /* Signal the other CPUs to enter kgdb_wait() */ if ((!kgdb_single_step) && kgdb_do_roundup) @@ -1521,8 +1483,6 @@ acquirelock: if (kgdb_io_ops->post_exception) kgdb_io_ops->post_exception(); - kgdb_info[ks->cpu].debuggerinfo = NULL; - kgdb_info[ks->cpu].task = NULL; atomic_set(&cpu_in_kgdb[ks->cpu], 0); if (!kgdb_single_step) { @@ -1555,13 +1515,52 @@ kgdb_restore: return error; } +/* + * kgdb_handle_exception() - main entry point from a kernel exception + * + * Locking hierarchy: + * interface locks, if any (begin_session) + * kgdb lock (kgdb_active) + */ +int +kgdb_handle_exception(int evector, int signo, int ecode, struct pt_regs *regs) +{ + struct kgdb_state kgdb_var; + struct kgdb_state *ks = &kgdb_var; + int ret; + + ks->cpu = raw_smp_processor_id(); + ks->ex_vector = evector; + ks->signo = signo; + ks->ex_vector = evector; + ks->err_code = ecode; + ks->kgdb_usethreadid = 0; + ks->linux_regs = regs; + + if (kgdb_reenter_check(ks)) + return 0; /* Ouch, double exception ! */ + kgdb_info[ks->cpu].exception_state |= DCPU_WANT_MASTER; + ret = kgdb_cpu_enter(ks, regs); + kgdb_info[ks->cpu].exception_state &= ~DCPU_WANT_MASTER; + return ret; +} + int kgdb_nmicallback(int cpu, void *regs) { #ifdef CONFIG_SMP + struct kgdb_state kgdb_var; + struct kgdb_state *ks = &kgdb_var; + + memset(ks, 0, sizeof(struct kgdb_state)); + ks->cpu = cpu; + ks->linux_regs = regs; + if (!atomic_read(&cpu_in_kgdb[cpu]) && - atomic_read(&kgdb_active) != cpu && - atomic_read(&cpu_in_kgdb[atomic_read(&kgdb_active)])) { - kgdb_wait((struct pt_regs *)regs); + atomic_read(&kgdb_active) != -1 && + atomic_read(&kgdb_active) != cpu) { + kgdb_info[cpu].exception_state |= DCPU_IS_SLAVE; + kgdb_cpu_enter(ks, regs); + kgdb_info[cpu].exception_state &= ~DCPU_IS_SLAVE; return 0; } #endif -- cgit v1.2.2 From ae6bf53e0255c8ab04b6fe31806e318432570e3c Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Fri, 2 Apr 2010 14:58:18 -0500 Subject: kgdb: use atomic_inc and atomic_dec instead of atomic_set Memory barriers should be used for the kgdb cpu synchronization. The atomic_set() does not imply a memory barrier. Reported-by: Will Deacon Signed-off-by: Jason Wessel --- kernel/kgdb.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 6882c047452d..2f7f454605c2 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -1379,8 +1379,7 @@ acquirelock: * Make sure the above info reaches the primary CPU before * our cpu_in_kgdb[] flag setting does: */ - smp_wmb(); - atomic_set(&cpu_in_kgdb[cpu], 1); + atomic_inc(&cpu_in_kgdb[cpu]); /* * CPU will loop if it is a slave or request to become a kgdb @@ -1400,7 +1399,7 @@ return_normal: */ if (arch_kgdb_ops.correct_hw_break) arch_kgdb_ops.correct_hw_break(); - atomic_set(&cpu_in_kgdb[cpu], 0); + atomic_dec(&cpu_in_kgdb[cpu]); touch_softlockup_watchdog_sync(); clocksource_touch_watchdog(); local_irq_restore(flags); @@ -1449,7 +1448,7 @@ return_normal: */ if (!kgdb_single_step) { for (i = 0; i < NR_CPUS; i++) - atomic_set(&passive_cpu_wait[i], 1); + atomic_inc(&passive_cpu_wait[i]); } #ifdef CONFIG_SMP @@ -1483,11 +1482,11 @@ return_normal: if (kgdb_io_ops->post_exception) kgdb_io_ops->post_exception(); - atomic_set(&cpu_in_kgdb[ks->cpu], 0); + atomic_dec(&cpu_in_kgdb[ks->cpu]); if (!kgdb_single_step) { for (i = NR_CPUS-1; i >= 0; i--) - atomic_set(&passive_cpu_wait[i], 0); + atomic_dec(&passive_cpu_wait[i]); /* * Wait till all the CPUs have quit * from the debugger. @@ -1736,11 +1735,11 @@ EXPORT_SYMBOL_GPL(kgdb_unregister_io_module); */ void kgdb_breakpoint(void) { - atomic_set(&kgdb_setting_breakpoint, 1); + atomic_inc(&kgdb_setting_breakpoint); wmb(); /* Sync point before breakpoint */ arch_kgdb_breakpoint(); wmb(); /* Sync point after breakpoint */ - atomic_set(&kgdb_setting_breakpoint, 0); + atomic_dec(&kgdb_setting_breakpoint); } EXPORT_SYMBOL_GPL(kgdb_breakpoint); -- cgit v1.2.2 From 4da75b9ceac6939cd76830ec9581bef5bb398ad3 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Fri, 2 Apr 2010 11:57:18 -0500 Subject: kgdb: Turn off tracing while in the debugger The kernel debugger should turn off kernel tracing any time the debugger is active and restore it on resume. Signed-off-by: Jason Wessel Reviewed-by: Steven Rostedt --- kernel/kgdb.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'kernel') diff --git a/kernel/kgdb.c b/kernel/kgdb.c index 2f7f454605c2..11f3515ca83f 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -1365,6 +1365,7 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs) int sstep_tries = 100; int error = 0; int i, cpu; + int trace_on = 0; acquirelock: /* * Interrupts will be restored by the 'trap return' code, except when @@ -1399,6 +1400,8 @@ return_normal: */ if (arch_kgdb_ops.correct_hw_break) arch_kgdb_ops.correct_hw_break(); + if (trace_on) + tracing_on(); atomic_dec(&cpu_in_kgdb[cpu]); touch_softlockup_watchdog_sync(); clocksource_touch_watchdog(); @@ -1474,6 +1477,9 @@ return_normal: kgdb_single_step = 0; kgdb_contthread = current; exception_level = 0; + trace_on = tracing_is_on(); + if (trace_on) + tracing_off(); /* Talk to debugger with gdbserial protocol */ error = gdb_serial_stub(ks); @@ -1505,6 +1511,8 @@ kgdb_restore: else kgdb_sstep_pid = 0; } + if (trace_on) + tracing_on(); /* Free kgdb_active */ atomic_set(&kgdb_active, -1); touch_softlockup_watchdog_sync(); -- cgit v1.2.2 From 26d80aa782e708c380a47601779d42d30bf016d6 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 3 Apr 2010 12:22:05 +0200 Subject: perf: Always build the stub perf_arch_fetch_caller_regs version Now that software events use perf_arch_fetch_caller_regs() too, we need the stub version to be always built in for archs that don't implement it. Fixes the following build error in PARISC: kernel/built-in.o: In function `perf_event_task_sched_out': (.text.perf_event_task_sched_out+0x54): undefined reference to `perf_arch_fetch_caller_regs' Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras --- kernel/perf_event.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 96aae13c7960..681af806d76b 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -2784,12 +2784,11 @@ __weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) return NULL; } -#ifdef CONFIG_EVENT_TRACING __weak void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip) { } -#endif + /* * Output -- cgit v1.2.2 From 449cedf099b23a250e7d61982e35555ccb871182 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Mon, 5 Apr 2010 16:16:26 -0400 Subject: audit: preface audit printk with audit There have been a number of reports of people seeing the message: "name_count maxed, losing inode data: dev=00:05, inode=3185" in dmesg. These usually lead to people reporting problems to the filesystem group who are in turn clueless what they mean. Eventually someone finds me and I explain what is going on and that these come from the audit system. The basics of the problem is that the audit subsystem never expects a single syscall to 'interact' (for some wish washy meaning of interact) with more than 20 inodes. But in fact some operations like loading kernel modules can cause changes to lots of inodes in debugfs. There are a couple real fixes being bandied about including removing the fixed compile time limit of 20 or not auditing changes in debugfs (or both) but neither are small and obvious so I am not sending them for immediate inclusion (I hope Al forwards a real solution next devel window). In the meantime this patch simply adds 'audit' to the beginning of the crap message so if a user sees it, they come blame me first and we can talk about what it means and make sure we understand all of the reasons it can happen and make sure this gets solved correctly in the long run. Signed-off-by: Eric Paris Signed-off-by: Linus Torvalds --- kernel/auditsc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 97a3cef81b9c..3828ad5fb8f1 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -1894,7 +1894,7 @@ static int audit_inc_name_count(struct audit_context *context, { if (context->name_count >= AUDIT_NAMES) { if (inode) - printk(KERN_DEBUG "name_count maxed, losing inode data: " + printk(KERN_DEBUG "audit: name_count maxed, losing inode data: " "dev=%02x:%02x, inode=%lu\n", MAJOR(inode->i_sb->s_dev), MINOR(inode->i_sb->s_dev), -- cgit v1.2.2 From 5fbfb18d7a5b846946d52c4a10e3aaa213ec31b6 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Thu, 1 Apr 2010 19:09:40 +1100 Subject: Fix up possibly racy module refcounting Module refcounting is implemented with a per-cpu counter for speed. However there is a race when tallying the counter where a reference may be taken by one CPU and released by another. Reference count summation may then see the decrement without having seen the previous increment, leading to lower than expected count. A module which never has its actual reference drop below 1 may return a reference count of 0 due to this race. Module removal generally runs under stop_machine, which prevents this race causing bugs due to removal of in-use modules. However there are other real bugs in module.c code and driver code (module_refcount is exported) where the callers do not run under stop_machine. Fix this by maintaining running per-cpu counters for the number of module refcount increments and the number of refcount decrements. The increments are tallied after the decrements, so any decrement seen will always have its corresponding increment counted. The final refcount is the difference of the total increments and decrements, preventing a low-refcount from being returned. Signed-off-by: Nick Piggin Acked-by: Rusty Russell Signed-off-by: Linus Torvalds --- kernel/module.c | 35 +++++++++++++++++++++++++++-------- 1 file changed, 27 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 9f8d23d8b3a8..1016b75b026a 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -521,11 +521,13 @@ static void module_unload_init(struct module *mod) int cpu; INIT_LIST_HEAD(&mod->modules_which_use_me); - for_each_possible_cpu(cpu) - per_cpu_ptr(mod->refptr, cpu)->count = 0; + for_each_possible_cpu(cpu) { + per_cpu_ptr(mod->refptr, cpu)->incs = 0; + per_cpu_ptr(mod->refptr, cpu)->decs = 0; + } /* Hold reference count during initialization. */ - __this_cpu_write(mod->refptr->count, 1); + __this_cpu_write(mod->refptr->incs, 1); /* Backwards compatibility macros put refcount during init. */ mod->waiter = current; } @@ -664,12 +666,28 @@ static int try_stop_module(struct module *mod, int flags, int *forced) unsigned int module_refcount(struct module *mod) { - unsigned int total = 0; + unsigned int incs = 0, decs = 0; int cpu; for_each_possible_cpu(cpu) - total += per_cpu_ptr(mod->refptr, cpu)->count; - return total; + decs += per_cpu_ptr(mod->refptr, cpu)->decs; + /* + * ensure the incs are added up after the decs. + * module_put ensures incs are visible before decs with smp_wmb. + * + * This 2-count scheme avoids the situation where the refcount + * for CPU0 is read, then CPU0 increments the module refcount, + * then CPU1 drops that refcount, then the refcount for CPU1 is + * read. We would record a decrement but not its corresponding + * increment so we would see a low count (disaster). + * + * Rare situation? But module_refcount can be preempted, and we + * might be tallying up 4096+ CPUs. So it is not impossible. + */ + smp_rmb(); + for_each_possible_cpu(cpu) + incs += per_cpu_ptr(mod->refptr, cpu)->incs; + return incs - decs; } EXPORT_SYMBOL(module_refcount); @@ -846,10 +864,11 @@ void module_put(struct module *module) { if (module) { preempt_disable(); - __this_cpu_dec(module->refptr->count); + smp_wmb(); /* see comment in module_refcount */ + __this_cpu_inc(module->refptr->decs); trace_module_put(module, _RET_IP_, - __this_cpu_read(module->refptr->count)); + __this_cpu_read(module->refptr->decs)); /* Maybe they're waiting for us to drop reference? */ if (unlikely(!module_is_live(module))) wake_up_process(module->waiter); -- cgit v1.2.2 From 84fba5ec91f11c0efb27d0ed6098f7447491f0df Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 6 Apr 2010 17:02:19 +1000 Subject: sched: Fix sched_getaffinity() taskset on 2.6.34-rc3 fails on one of my ppc64 test boxes with the following error: sched_getaffinity(0, 16, 0x10029650030) = -1 EINVAL (Invalid argument) This box has 128 threads and 16 bytes is enough to cover it. Commit cd3d8031eb4311e516329aee03c79a08333141f1 (sched: sched_getaffinity(): Allow less than NR_CPUS length) is comparing this 16 bytes agains nr_cpu_ids. Fix it by comparing nr_cpu_ids to the number of bits in the cpumask we pass in. Signed-off-by: Anton Blanchard Reviewed-by: KOSAKI Motohiro Cc: Sharyathi Nagesh Cc: Ulrich Drepper Cc: Peter Zijlstra Cc: Linus Torvalds Cc: Jack Steiner Cc: Russ Anderson Cc: Mike Travis LKML-Reference: <20100406070218.GM5594@kryten> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 528a10592c16..eaf5c7375dfd 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4902,7 +4902,7 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, int ret; cpumask_var_t mask; - if (len < nr_cpu_ids) + if ((len * BITS_PER_BYTE) < nr_cpu_ids) return -EINVAL; if (len & (sizeof(unsigned long)-1)) return -EINVAL; -- cgit v1.2.2 From a3a2e76c77fa22b114e421ac11dec0c56c3503fb Mon Sep 17 00:00:00 2001 From: KAMEZAWA Hiroyuki Date: Tue, 6 Apr 2010 14:34:42 -0700 Subject: mm: avoid null-pointer deref in sync_mm_rss() - We weren't zeroing p->rss_stat[] at fork() - Consequently sync_mm_rss() was dereferencing tsk->mm for kernel threads and was oopsing. - Make __sync_task_rss_stat() static, too. Addresses https://bugzilla.kernel.org/show_bug.cgi?id=15648 [akpm@linux-foundation.org: remove the BUG_ON(!mm->rss)] Reported-by: Troels Liebe Bentsen Signed-off-by: KAMEZAWA Hiroyuki "Michael S. Tsirkin" Cc: Andrea Arcangeli Cc: Rik van Riel Cc: Minchan Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 3 ++- kernel/fork.c | 3 +++ 2 files changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index cce59cb5ee6a..7f2683a10ac4 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -953,7 +953,8 @@ NORET_TYPE void do_exit(long code) acct_update_integrals(tsk); /* sync mm's RSS info before statistics gathering */ - sync_mm_rss(tsk, tsk->mm); + if (tsk->mm) + sync_mm_rss(tsk, tsk->mm); group_dead = atomic_dec_and_test(&tsk->signal->live); if (group_dead) { hrtimer_cancel(&tsk->signal->real_timer); diff --git a/kernel/fork.c b/kernel/fork.c index 4799c5f0e6d0..44b0791b0a2e 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1052,6 +1052,9 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->prev_utime = cputime_zero; p->prev_stime = cputime_zero; #endif +#if defined(SPLIT_RSS_COUNTING) + memset(&p->rss_stat, 0, sizeof(p->rss_stat)); +#endif p->default_timer_slack_ns = current->timer_slack_ns; -- cgit v1.2.2 From d88d4050dcaf09e417aaa9a5024dd9449ef71b2e Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Sat, 10 Apr 2010 22:28:56 +0200 Subject: PM / Hibernate: user.c, fix SNAPSHOT_SET_SWAP_AREA handling When CONFIG_DEBUG_BLOCK_EXT_DEVT is set we decode the device improperly by old_decode_dev and it results in an error while hibernating with s2disk. All users already pass the new device number, so switch to new_decode_dev(). Signed-off-by: Jiri Slaby Reported-and-tested-by: Jiri Kosina Signed-off-by: "Rafael J. Wysocki" --- kernel/power/user.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/user.c b/kernel/power/user.c index 4d2289626a84..a8c96212bc1b 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -420,7 +420,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, * User space encodes device types as two-byte values, * so we need to recode them */ - swdev = old_decode_dev(swap_area.dev); + swdev = new_decode_dev(swap_area.dev); if (swdev) { offset = swap_area.offset; data->swap = swap_type_of(swdev, offset, NULL); -- cgit v1.2.2 From bc293d62b26ec590afc90a9e0a31c45d355b7bd8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 15 Apr 2010 12:50:39 -0700 Subject: rcu: Make RCU lockdep check the lockdep_recursion variable The lockdep facility temporarily disables lockdep checking by incrementing the current->lockdep_recursion variable. Such disabling happens in NMIs and in other situations where lockdep might expect to recurse on itself. This patch therefore checks current->lockdep_recursion, disabling RCU lockdep splats when this variable is non-zero. In addition, this patch removes the "likely()", as suggested by Lai Jiangshan. Reported-by: Frederic Weisbecker Reported-by: David Miller Tested-by: Frederic Weisbecker Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com Cc: eric.dumazet@gmail.com LKML-Reference: <20100415195039.GA22623@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 63fe25433980..03a7ea1579f6 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -69,6 +69,13 @@ EXPORT_SYMBOL_GPL(rcu_scheduler_active); #ifdef CONFIG_DEBUG_LOCK_ALLOC +int debug_lockdep_rcu_enabled(void) +{ + return rcu_scheduler_active && debug_locks && + current->lockdep_recursion == 0; +} +EXPORT_SYMBOL_GPL(debug_lockdep_rcu_enabled); + /** * rcu_read_lock_bh_held - might we be in RCU-bh read-side critical section? * -- cgit v1.2.2 From eff30363c0b8b057f773108589bfd8881659fe74 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 20 Apr 2010 22:41:18 +0100 Subject: CRED: Fix double free in prepare_usermodehelper_creds() error handling Patch 570b8fb505896e007fd3bb07573ba6640e51851d: Author: Mathieu Desnoyers Date: Tue Mar 30 00:04:00 2010 +0100 Subject: CRED: Fix memory leak in error handling attempts to fix a memory leak in the error handling by making the offending return statement into a jump down to the bottom of the function where a kfree(tgcred) is inserted. This is, however, incorrect, as it does a kfree() after doing put_cred() if security_prepare_creds() fails. That will result in a double free if 'error' is jumped to as put_cred() will also attempt to free the new tgcred record by virtue of it being pointed to by the new cred record. Signed-off-by: David Howells Signed-off-by: James Morris --- kernel/cred.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/cred.c b/kernel/cred.c index e1dbe9eef800..ce1a52b9e8a3 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -398,6 +398,8 @@ struct cred *prepare_usermodehelper_creds(void) error: put_cred(new); + return NULL; + free_tgcred: #ifdef CONFIG_KEYS kfree(tgcred); -- cgit v1.2.2 From e134d200d57d43b171dcb0b55c178a1a0c7db14a Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 21 Apr 2010 10:28:25 +0100 Subject: CRED: Fix a race in creds_are_invalid() in credentials debugging creds_are_invalid() reads both cred->usage and cred->subscribers and then compares them to make sure the number of processes subscribed to a cred struct never exceeds the refcount of that cred struct. The problem is that this can cause a race with both copy_creds() and exit_creds() as the two counters, whilst they are of atomic_t type, are only atomic with respect to themselves, and not atomic with respect to each other. This means that if creds_are_invalid() can read the values on one CPU whilst they're being modified on another CPU, and so can observe an evolving state in which the subscribers count now is greater than the usage count a moment before. Switching the order in which the counts are read cannot help, so the thing to do is to remove that particular check. I had considered rechecking the values to see if they're in flux if the test fails, but I can't guarantee they won't appear the same, even if they've changed several times in the meantime. Note that this can only happen if CONFIG_DEBUG_CREDENTIALS is enabled. The problem is only likely to occur with multithreaded programs, and can be tested by the tst-eintr1 program from glibc's "make check". The symptoms look like: CRED: Invalid credentials CRED: At include/linux/cred.h:240 CRED: Specified credentials: ffff88003dda5878 [real][eff] CRED: ->magic=43736564, put_addr=(null) CRED: ->usage=766, subscr=766 CRED: ->*uid = { 0,0,0,0 } CRED: ->*gid = { 0,0,0,0 } CRED: ->security is ffff88003d72f538 CRED: ->security {359, 359} ------------[ cut here ]------------ kernel BUG at kernel/cred.c:850! ... RIP: 0010:[] [] __invalid_creds+0x4e/0x52 ... Call Trace: [] copy_creds+0x6b/0x23f Note the ->usage=766 and subscr=766. The values appear the same because they've been re-read since the check was made. Reported-by: Roland McGrath Signed-off-by: David Howells Signed-off-by: James Morris --- kernel/cred.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cred.c b/kernel/cred.c index ce1a52b9e8a3..62af1816c235 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -793,8 +793,6 @@ bool creds_are_invalid(const struct cred *cred) { if (cred->magic != CRED_MAGIC) return true; - if (atomic_read(&cred->usage) < atomic_read(&cred->subscribers)) - return true; #ifdef CONFIG_SECURITY_SELINUX if (selinux_is_enabled()) { if ((unsigned long) cred->security < PAGE_SIZE) -- cgit v1.2.2 From 4b402210486c6414fe5fbfd85934a0a22da56b04 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Fri, 16 Apr 2010 23:20:00 +0200 Subject: mutex: Don't spin when the owner CPU is offline or other weird cases Due to recent load-balancer changes that delay the task migration to the next wakeup, the adaptive mutex spinning ends up in a live lock when the owner's CPU gets offlined because the cpu_online() check lives before the owner running check. This patch changes mutex_spin_on_owner() to return 0 (don't spin) in any case where we aren't sure about the owner struct validity or CPU number, and if the said CPU is offline. There is no point going back & re-evaluate spinning in corner cases like that, let's just go to sleep. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Peter Zijlstra LKML-Reference: <1271212509.13059.135.camel@pasglop> Signed-off-by: Ingo Molnar --- kernel/sched.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index 6af210a7de70..de0bd26e520a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3780,7 +3780,7 @@ int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner) * the mutex owner just released it and exited. */ if (probe_kernel_address(&owner->cpu, cpu)) - goto out; + return 0; #else cpu = owner->cpu; #endif @@ -3790,14 +3790,14 @@ int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner) * the cpu field may no longer be valid. */ if (cpu >= nr_cpumask_bits) - goto out; + return 0; /* * We need to validate that we can do a * get_cpu() and that we have the percpu area. */ if (!cpu_online(cpu)) - goto out; + return 0; rq = cpu_rq(cpu); @@ -3816,7 +3816,7 @@ int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner) cpu_relax(); } -out: + return 1; } #endif -- cgit v1.2.2 From 46da27664887fb95cedba53eafcf876de812c8c1 Mon Sep 17 00:00:00 2001 From: Andreas Schwab Date: Fri, 23 Apr 2010 13:17:44 -0400 Subject: kernel/sys.c: fix compat uname machine On ppc64 you get this error: $ setarch ppc -R true setarch: ppc: Unrecognized architecture because uname still reports ppc64 as the machine. So mask off the personality flags when checking for PER_LINUX32. Signed-off-by: Andreas Schwab Reviewed-by: Christoph Hellwig Acked-by: David S. Miller Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 6d1a7e0f9d5b..7cb426a58965 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1118,7 +1118,7 @@ DECLARE_RWSEM(uts_sem); #ifdef COMPAT_UTS_MACHINE #define override_architecture(name) \ - (current->personality == PER_LINUX32 && \ + (personality(current->personality) == PER_LINUX32 && \ copy_to_user(name->machine, COMPAT_UTS_MACHINE, \ sizeof(COMPAT_UTS_MACHINE))) #else -- cgit v1.2.2 From 47dd5be2d6a82b8153e059a1d09eb3879d485bfd Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 30 Apr 2010 07:23:51 +0200 Subject: workqueue: flush_delayed_work: keep the original workqueue for re-queueing flush_delayed_work() always uses keventd_wq for re-queueing, but it should use the workqueue this dwork was queued on. Signed-off-by: Oleg Nesterov Signed-off-by: Tejun Heo --- kernel/workqueue.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index dee48658805c..5bfb213984b2 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -774,7 +774,7 @@ void flush_delayed_work(struct delayed_work *dwork) { if (del_timer_sync(&dwork->timer)) { struct cpu_workqueue_struct *cwq; - cwq = wq_per_cpu(keventd_wq, get_cpu()); + cwq = wq_per_cpu(get_wq_data(&dwork->work)->wq, get_cpu()); __queue_work(cwq, &dwork->work); put_cpu(); } -- cgit v1.2.2 From 8b08ca52f5942c21564bbb90ccfb61053f2c26a1 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 21 Apr 2010 13:02:07 -0700 Subject: rcu: Fix RCU lockdep splat in set_task_cpu on fork path Add an RCU read-side critical section to suppress this false positive. Located-by: Eric Paris Signed-off-by: Peter Zijlstra Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com Cc: eric.dumazet@gmail.com LKML-Reference: <1271880131-3951-1-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index de0bd26e520a..3c2a54f70ffe 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -323,6 +323,15 @@ static inline struct task_group *task_group(struct task_struct *p) /* Change a task's cfs_rq and parent entity if it moves across CPUs/groups */ static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { + /* + * Strictly speaking this rcu_read_lock() is not needed since the + * task_group is tied to the cgroup, which in turn can never go away + * as long as there are tasks attached to it. + * + * However since task_group() uses task_subsys_state() which is an + * rcu_dereference() user, this quiets CONFIG_PROVE_RCU. + */ + rcu_read_lock(); #ifdef CONFIG_FAIR_GROUP_SCHED p->se.cfs_rq = task_group(p)->cfs_rq[cpu]; p->se.parent = task_group(p)->se[cpu]; @@ -332,6 +341,7 @@ static inline void set_task_rq(struct task_struct *p, unsigned int cpu) p->rt.rt_rq = task_group(p)->rt_rq[cpu]; p->rt.parent = task_group(p)->rt_se[cpu]; #endif + rcu_read_unlock(); } #else -- cgit v1.2.2 From 8b46f880841aac821af8efa6581bb0e46b8b9845 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 21 Apr 2010 13:02:08 -0700 Subject: rcu: Fix RCU lockdep splat on freezer_fork path Add an RCU read-side critical section to suppress this false positive. Located-by: Eric Paris Signed-off-by: Paul E. McKenney Acked-by: Li Zefan Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com Cc: eric.dumazet@gmail.com LKML-Reference: <1271880131-3951-2-git-send-email-paulmck@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar --- kernel/cgroup_freezer.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index da5e13975531..e5c0244962b0 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -205,9 +205,12 @@ static void freezer_fork(struct cgroup_subsys *ss, struct task_struct *task) * No lock is needed, since the task isn't on tasklist yet, * so it can't be moved to another cgroup, which means the * freezer won't be removed and will be valid during this - * function call. + * function call. Nevertheless, apply RCU read-side critical + * section to suppress RCU lockdep false positives. */ + rcu_read_lock(); freezer = task_freezer(task); + rcu_read_unlock(); /* * The root cgroup is non-freezable, so we can skip the -- cgit v1.2.2 From 048c852051d2bd5da54a4488bc1f16b0fc74c695 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sat, 1 May 2010 10:11:35 +0200 Subject: perf: Fix resource leak in failure path of perf_event_open() perf_event_open() kfrees event after init failure which doesn't release all resources allocated by perf_event_alloc(). Use free_event() instead. Signed-off-by: Tejun Heo Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: LKML-Reference: <4BDBE237.1040809@kernel.org> Signed-off-by: Ingo Molnar --- kernel/perf_event.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 2f3fbf84215a..3d1552d3c12b 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4897,7 +4897,7 @@ err_fput_free_put_context: err_free_put_context: if (err < 0) - kfree(event); + free_event(event); err_put_context: if (err < 0) -- cgit v1.2.2 From 9a9686b634acc5cb6b7c601c171ae64af0318a24 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 22 Apr 2010 17:29:24 +0800 Subject: cgroup: Fix an RCU warning in cgroup_path() with CONFIG_PROVE_RCU=y, a warning can be triggered: # mount -t cgroup -o debug xxx /mnt # cat /proc/$$/cgroup ... kernel/cgroup.c:1649 invoked rcu_dereference_check() without protection! ... This is a false-positive, because cgroup_path() can be called with either rcu_read_lock() held or cgroup_mutex held. Signed-off-by: Li Zefan Signed-off-by: Paul E. McKenney --- kernel/cgroup.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e2769e13980c..4ca928db890c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1646,7 +1646,9 @@ static inline struct cftype *__d_cft(struct dentry *dentry) int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) { char *start; - struct dentry *dentry = rcu_dereference(cgrp->dentry); + struct dentry *dentry = rcu_dereference_check(cgrp->dentry, + rcu_read_lock_held() || + cgroup_lock_is_held()); if (!dentry || cgrp == dummytop) { /* @@ -1662,13 +1664,17 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) *--start = '\0'; for (;;) { int len = dentry->d_name.len; + if ((start -= len) < buf) return -ENAMETOOLONG; - memcpy(start, cgrp->dentry->d_name.name, len); + memcpy(start, dentry->d_name.name, len); cgrp = cgrp->parent; if (!cgrp) break; - dentry = rcu_dereference(cgrp->dentry); + + dentry = rcu_dereference_check(cgrp->dentry, + rcu_read_lock_held() || + cgroup_lock_is_held()); if (!cgrp->parent) continue; if (--start < buf) -- cgit v1.2.2 From fae9c791703606636c1220e47f6690660042ce7f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 22 Apr 2010 17:30:00 +0800 Subject: cgroup: Fix an RCU warning in alloc_css_id() With CONFIG_PROVE_RCU=y, a warning can be triggered: # mount -t cgroup -o memory xxx /mnt # mkdir /mnt/0 ... kernel/cgroup.c:4442 invoked rcu_dereference_check() without protection! ... This is a false-positive. It's safe to directly access parent_css->id. Signed-off-by: Li Zefan Signed-off-by: Paul E. McKenney --- kernel/cgroup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 4ca928db890c..3a53c771e503 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4561,13 +4561,13 @@ static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, { int subsys_id, i, depth = 0; struct cgroup_subsys_state *parent_css, *child_css; - struct css_id *child_id, *parent_id = NULL; + struct css_id *child_id, *parent_id; subsys_id = ss->subsys_id; parent_css = parent->subsys[subsys_id]; child_css = child->subsys[subsys_id]; - depth = css_depth(parent_css) + 1; parent_id = parent_css->id; + depth = parent_id->depth; child_id = get_new_cssid(ss, depth); if (IS_ERR(child_id)) -- cgit v1.2.2 From b629317e66fb1c6066c550dded45ab85a936163c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 22 Apr 2010 17:30:40 +0800 Subject: sched: Fix an RCU warning in print_task() With CONFIG_PROVE_RCU=y, a warning can be triggered: $ cat /proc/sched_debug ... kernel/cgroup.c:1649 invoked rcu_dereference_check() without protection! ... Both cgroup_path() and task_group() should be called with either rcu_read_lock or cgroup_mutex held. The rcu_dereference_check() does include cgroup_lock_is_held(), so we know that this lock is not held. Therefore, in a CONFIG_PREEMPT kernel, to say nothing of a CONFIG_PREEMPT_RT kernel, the original code could have ended up copying a string out of the freelist. This patch inserts RCU read-side primitives needed to prevent this scenario. Signed-off-by: Li Zefan Signed-off-by: Paul E. McKenney --- kernel/sched_debug.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index 9b49db144037..19be00ba6123 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -114,7 +114,9 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) { char path[64]; + rcu_read_lock(); cgroup_path(task_group(p)->css.cgroup, path, sizeof(path)); + rcu_read_unlock(); SEQ_printf(m, " %s", path); } #endif -- cgit v1.2.2 From ee84b8243b07c33a5c8aed42b4b2da60cb16d1d2 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 6 May 2010 09:28:41 -0700 Subject: rcu: create rcu_my_thread_group_empty() wrapper Some RCU-lockdep splat repairs need to know whether they are running in a single-threaded process. Unfortunately, the thread_group_empty() primitive is defined in sched.h, and can induce #include hell. This commit therefore introduces a rcu_my_thread_group_empty() wrapper that is defined in rcupdate.c, thus avoiding the need to include sched.h everywhere. Signed-off-by: "Paul E. McKenney" --- kernel/rcupdate.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 03a7ea1579f6..49d808e833b0 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -122,3 +122,14 @@ void wakeme_after_rcu(struct rcu_head *head) rcu = container_of(head, struct rcu_synchronize, head); complete(&rcu->completion); } + +#ifdef CONFIG_PROVE_RCU +/* + * wrapper function to avoid #include problems. + */ +int rcu_my_thread_group_empty(void) +{ + return thread_group_empty(current); +} +EXPORT_SYMBOL_GPL(rcu_my_thread_group_empty); +#endif /* #ifdef CONFIG_PROVE_RCU */ -- cgit v1.2.2