aboutsummaryrefslogtreecommitdiffstats
path: root/kernel
diff options
context:
space:
mode:
Diffstat (limited to 'kernel')
-rw-r--r--kernel/acct.c6
-rw-r--r--kernel/auditsc.c5
-rw-r--r--kernel/capability.c2
-rw-r--r--kernel/compat.c35
-rw-r--r--kernel/cpuset.c106
-rw-r--r--kernel/exit.c40
-rw-r--r--kernel/fork.c15
-rw-r--r--kernel/futex.c10
-rw-r--r--kernel/hrtimer.c20
-rw-r--r--kernel/irq/chip.c6
-rw-r--r--kernel/kexec.c8
-rw-r--r--kernel/kfifo.c28
-rw-r--r--kernel/kmod.c12
-rw-r--r--kernel/lockdep.c17
-rw-r--r--kernel/module.c26
-rw-r--r--kernel/panic.c1
-rw-r--r--kernel/params.c15
-rw-r--r--kernel/posix-cpu-timers.c101
-rw-r--r--kernel/posix-timers.c21
-rw-r--r--kernel/ptrace.c1
-rw-r--r--kernel/rcutorture.c8
-rw-r--r--kernel/relay.c36
-rw-r--r--kernel/rtmutex.c51
-rw-r--r--kernel/sched.c74
-rw-r--r--kernel/signal.c6
-rw-r--r--kernel/softirq.c4
-rw-r--r--kernel/softlockup.c3
-rw-r--r--kernel/spinlock.c11
-rw-r--r--kernel/stop_machine.c3
-rw-r--r--kernel/sys.c9
-rw-r--r--kernel/sys_ni.c5
-rw-r--r--kernel/sysctl.c13
-rw-r--r--kernel/timer.c53
-rw-r--r--kernel/unwind.c4
34 files changed, 511 insertions, 244 deletions
diff --git a/kernel/acct.c b/kernel/acct.c
index 2a7c933651c7..f4330acead46 100644
--- a/kernel/acct.c
+++ b/kernel/acct.c
@@ -483,10 +483,14 @@ static void do_acct_process(struct file *file)
483 ac.ac_ppid = current->parent->tgid; 483 ac.ac_ppid = current->parent->tgid;
484#endif 484#endif
485 485
486 read_lock(&tasklist_lock); /* pin current->signal */ 486 mutex_lock(&tty_mutex);
487 /* FIXME: Whoever is responsible for current->signal locking needs
488 to use the same locking all over the kernel and document it */
489 read_lock(&tasklist_lock);
487 ac.ac_tty = current->signal->tty ? 490 ac.ac_tty = current->signal->tty ?
488 old_encode_dev(tty_devnum(current->signal->tty)) : 0; 491 old_encode_dev(tty_devnum(current->signal->tty)) : 0;
489 read_unlock(&tasklist_lock); 492 read_unlock(&tasklist_lock);
493 mutex_unlock(&tty_mutex);
490 494
491 spin_lock_irq(&current->sighand->siglock); 495 spin_lock_irq(&current->sighand->siglock);
492 ac.ac_utime = encode_comp_t(jiffies_to_AHZ(cputime_to_jiffies(pacct->ac_utime))); 496 ac.ac_utime = encode_comp_t(jiffies_to_AHZ(cputime_to_jiffies(pacct->ac_utime)));
diff --git a/kernel/auditsc.c b/kernel/auditsc.c
index fb83c5cb8c32..105147631753 100644
--- a/kernel/auditsc.c
+++ b/kernel/auditsc.c
@@ -817,6 +817,8 @@ static void audit_log_exit(struct audit_context *context, struct task_struct *ts
817 audit_log_format(ab, " success=%s exit=%ld", 817 audit_log_format(ab, " success=%s exit=%ld",
818 (context->return_valid==AUDITSC_SUCCESS)?"yes":"no", 818 (context->return_valid==AUDITSC_SUCCESS)?"yes":"no",
819 context->return_code); 819 context->return_code);
820
821 mutex_lock(&tty_mutex);
820 if (tsk->signal && tsk->signal->tty && tsk->signal->tty->name) 822 if (tsk->signal && tsk->signal->tty && tsk->signal->tty->name)
821 tty = tsk->signal->tty->name; 823 tty = tsk->signal->tty->name;
822 else 824 else
@@ -838,6 +840,9 @@ static void audit_log_exit(struct audit_context *context, struct task_struct *ts
838 context->gid, 840 context->gid,
839 context->euid, context->suid, context->fsuid, 841 context->euid, context->suid, context->fsuid,
840 context->egid, context->sgid, context->fsgid, tty); 842 context->egid, context->sgid, context->fsgid, tty);
843
844 mutex_unlock(&tty_mutex);
845
841 audit_log_task_info(ab, tsk); 846 audit_log_task_info(ab, tsk);
842 if (context->filterkey) { 847 if (context->filterkey) {
843 audit_log_format(ab, " key="); 848 audit_log_format(ab, " key=");
diff --git a/kernel/capability.c b/kernel/capability.c
index c7685ad00a97..edb845a6e84a 100644
--- a/kernel/capability.c
+++ b/kernel/capability.c
@@ -133,7 +133,7 @@ static inline int cap_set_all(kernel_cap_t *effective,
133 int found = 0; 133 int found = 0;
134 134
135 do_each_thread(g, target) { 135 do_each_thread(g, target) {
136 if (target == current || target->pid == 1) 136 if (target == current || is_init(target))
137 continue; 137 continue;
138 found = 1; 138 found = 1;
139 if (security_capset_check(target, effective, inheritable, 139 if (security_capset_check(target, effective, inheritable,
diff --git a/kernel/compat.c b/kernel/compat.c
index 126dee9530aa..b4fbd838cd77 100644
--- a/kernel/compat.c
+++ b/kernel/compat.c
@@ -22,9 +22,12 @@
22#include <linux/security.h> 22#include <linux/security.h>
23#include <linux/timex.h> 23#include <linux/timex.h>
24#include <linux/migrate.h> 24#include <linux/migrate.h>
25#include <linux/posix-timers.h>
25 26
26#include <asm/uaccess.h> 27#include <asm/uaccess.h>
27 28
29extern void sigset_from_compat(sigset_t *set, compat_sigset_t *compat);
30
28int get_compat_timespec(struct timespec *ts, const struct compat_timespec __user *cts) 31int get_compat_timespec(struct timespec *ts, const struct compat_timespec __user *cts)
29{ 32{
30 return (!access_ok(VERIFY_READ, cts, sizeof(*cts)) || 33 return (!access_ok(VERIFY_READ, cts, sizeof(*cts)) ||
@@ -601,6 +604,30 @@ long compat_sys_clock_getres(clockid_t which_clock,
601 return err; 604 return err;
602} 605}
603 606
607static long compat_clock_nanosleep_restart(struct restart_block *restart)
608{
609 long err;
610 mm_segment_t oldfs;
611 struct timespec tu;
612 struct compat_timespec *rmtp = (struct compat_timespec *)(restart->arg1);
613
614 restart->arg1 = (unsigned long) &tu;
615 oldfs = get_fs();
616 set_fs(KERNEL_DS);
617 err = clock_nanosleep_restart(restart);
618 set_fs(oldfs);
619
620 if ((err == -ERESTART_RESTARTBLOCK) && rmtp &&
621 put_compat_timespec(&tu, rmtp))
622 return -EFAULT;
623
624 if (err == -ERESTART_RESTARTBLOCK) {
625 restart->fn = compat_clock_nanosleep_restart;
626 restart->arg1 = (unsigned long) rmtp;
627 }
628 return err;
629}
630
604long compat_sys_clock_nanosleep(clockid_t which_clock, int flags, 631long compat_sys_clock_nanosleep(clockid_t which_clock, int flags,
605 struct compat_timespec __user *rqtp, 632 struct compat_timespec __user *rqtp,
606 struct compat_timespec __user *rmtp) 633 struct compat_timespec __user *rmtp)
@@ -608,6 +635,7 @@ long compat_sys_clock_nanosleep(clockid_t which_clock, int flags,
608 long err; 635 long err;
609 mm_segment_t oldfs; 636 mm_segment_t oldfs;
610 struct timespec in, out; 637 struct timespec in, out;
638 struct restart_block *restart;
611 639
612 if (get_compat_timespec(&in, rqtp)) 640 if (get_compat_timespec(&in, rqtp))
613 return -EFAULT; 641 return -EFAULT;
@@ -618,9 +646,16 @@ long compat_sys_clock_nanosleep(clockid_t which_clock, int flags,
618 (struct timespec __user *) &in, 646 (struct timespec __user *) &in,
619 (struct timespec __user *) &out); 647 (struct timespec __user *) &out);
620 set_fs(oldfs); 648 set_fs(oldfs);
649
621 if ((err == -ERESTART_RESTARTBLOCK) && rmtp && 650 if ((err == -ERESTART_RESTARTBLOCK) && rmtp &&
622 put_compat_timespec(&out, rmtp)) 651 put_compat_timespec(&out, rmtp))
623 return -EFAULT; 652 return -EFAULT;
653
654 if (err == -ERESTART_RESTARTBLOCK) {
655 restart = &current_thread_info()->restart_block;
656 restart->fn = compat_clock_nanosleep_restart;
657 restart->arg1 = (unsigned long) rmtp;
658 }
624 return err; 659 return err;
625} 660}
626 661
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 1b32c2c04c15..8c3c400cce91 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -240,7 +240,7 @@ static struct super_block *cpuset_sb;
240 * A cpuset can only be deleted if both its 'count' of using tasks 240 * A cpuset can only be deleted if both its 'count' of using tasks
241 * is zero, and its list of 'children' cpusets is empty. Since all 241 * is zero, and its list of 'children' cpusets is empty. Since all
242 * tasks in the system use _some_ cpuset, and since there is always at 242 * tasks in the system use _some_ cpuset, and since there is always at
243 * least one task in the system (init, pid == 1), therefore, top_cpuset 243 * least one task in the system (init), therefore, top_cpuset
244 * always has either children cpusets and/or using tasks. So we don't 244 * always has either children cpusets and/or using tasks. So we don't
245 * need a special hack to ensure that top_cpuset cannot be deleted. 245 * need a special hack to ensure that top_cpuset cannot be deleted.
246 * 246 *
@@ -912,6 +912,10 @@ static int update_nodemask(struct cpuset *cs, char *buf)
912 int fudge; 912 int fudge;
913 int retval; 913 int retval;
914 914
915 /* top_cpuset.mems_allowed tracks node_online_map; it's read-only */
916 if (cs == &top_cpuset)
917 return -EACCES;
918
915 trialcs = *cs; 919 trialcs = *cs;
916 retval = nodelist_parse(buf, trialcs.mems_allowed); 920 retval = nodelist_parse(buf, trialcs.mems_allowed);
917 if (retval < 0) 921 if (retval < 0)
@@ -1221,7 +1225,12 @@ static int attach_task(struct cpuset *cs, char *pidbuf, char **ppathbuf)
1221 1225
1222 task_lock(tsk); 1226 task_lock(tsk);
1223 oldcs = tsk->cpuset; 1227 oldcs = tsk->cpuset;
1224 if (!oldcs) { 1228 /*
1229 * After getting 'oldcs' cpuset ptr, be sure still not exiting.
1230 * If 'oldcs' might be the top_cpuset due to the_top_cpuset_hack
1231 * then fail this attach_task(), to avoid breaking top_cpuset.count.
1232 */
1233 if (tsk->flags & PF_EXITING) {
1225 task_unlock(tsk); 1234 task_unlock(tsk);
1226 mutex_unlock(&callback_mutex); 1235 mutex_unlock(&callback_mutex);
1227 put_task_struct(tsk); 1236 put_task_struct(tsk);
@@ -2036,33 +2045,104 @@ out:
2036 return err; 2045 return err;
2037} 2046}
2038 2047
2048#if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_MEMORY_HOTPLUG)
2039/* 2049/*
2040 * The top_cpuset tracks what CPUs and Memory Nodes are online, 2050 * If common_cpu_mem_hotplug_unplug(), below, unplugs any CPUs
2041 * period. This is necessary in order to make cpusets transparent 2051 * or memory nodes, we need to walk over the cpuset hierarchy,
2042 * (of no affect) on systems that are actively using CPU hotplug 2052 * removing that CPU or node from all cpusets. If this removes the
2043 * but making no active use of cpusets. 2053 * last CPU or node from a cpuset, then the guarantee_online_cpus()
2044 * 2054 * or guarantee_online_mems() code will use that emptied cpusets
2045 * This handles CPU hotplug (cpuhp) events. If someday Memory 2055 * parent online CPUs or nodes. Cpusets that were already empty of
2046 * Nodes can be hotplugged (dynamically changing node_online_map) 2056 * CPUs or nodes are left empty.
2047 * then we should handle that too, perhaps in a similar way. 2057 *
2058 * This routine is intentionally inefficient in a couple of regards.
2059 * It will check all cpusets in a subtree even if the top cpuset of
2060 * the subtree has no offline CPUs or nodes. It checks both CPUs and
2061 * nodes, even though the caller could have been coded to know that
2062 * only one of CPUs or nodes needed to be checked on a given call.
2063 * This was done to minimize text size rather than cpu cycles.
2064 *
2065 * Call with both manage_mutex and callback_mutex held.
2066 *
2067 * Recursive, on depth of cpuset subtree.
2048 */ 2068 */
2049 2069
2050#ifdef CONFIG_HOTPLUG_CPU 2070static void guarantee_online_cpus_mems_in_subtree(const struct cpuset *cur)
2051static int cpuset_handle_cpuhp(struct notifier_block *nb, 2071{
2052 unsigned long phase, void *cpu) 2072 struct cpuset *c;
2073
2074 /* Each of our child cpusets mems must be online */
2075 list_for_each_entry(c, &cur->children, sibling) {
2076 guarantee_online_cpus_mems_in_subtree(c);
2077 if (!cpus_empty(c->cpus_allowed))
2078 guarantee_online_cpus(c, &c->cpus_allowed);
2079 if (!nodes_empty(c->mems_allowed))
2080 guarantee_online_mems(c, &c->mems_allowed);
2081 }
2082}
2083
2084/*
2085 * The cpus_allowed and mems_allowed nodemasks in the top_cpuset track
2086 * cpu_online_map and node_online_map. Force the top cpuset to track
2087 * whats online after any CPU or memory node hotplug or unplug event.
2088 *
2089 * To ensure that we don't remove a CPU or node from the top cpuset
2090 * that is currently in use by a child cpuset (which would violate
2091 * the rule that cpusets must be subsets of their parent), we first
2092 * call the recursive routine guarantee_online_cpus_mems_in_subtree().
2093 *
2094 * Since there are two callers of this routine, one for CPU hotplug
2095 * events and one for memory node hotplug events, we could have coded
2096 * two separate routines here. We code it as a single common routine
2097 * in order to minimize text size.
2098 */
2099
2100static void common_cpu_mem_hotplug_unplug(void)
2053{ 2101{
2054 mutex_lock(&manage_mutex); 2102 mutex_lock(&manage_mutex);
2055 mutex_lock(&callback_mutex); 2103 mutex_lock(&callback_mutex);
2056 2104
2105 guarantee_online_cpus_mems_in_subtree(&top_cpuset);
2057 top_cpuset.cpus_allowed = cpu_online_map; 2106 top_cpuset.cpus_allowed = cpu_online_map;
2107 top_cpuset.mems_allowed = node_online_map;
2058 2108
2059 mutex_unlock(&callback_mutex); 2109 mutex_unlock(&callback_mutex);
2060 mutex_unlock(&manage_mutex); 2110 mutex_unlock(&manage_mutex);
2111}
2112#endif
2113
2114#ifdef CONFIG_HOTPLUG_CPU
2115/*
2116 * The top_cpuset tracks what CPUs and Memory Nodes are online,
2117 * period. This is necessary in order to make cpusets transparent
2118 * (of no affect) on systems that are actively using CPU hotplug
2119 * but making no active use of cpusets.
2120 *
2121 * This routine ensures that top_cpuset.cpus_allowed tracks
2122 * cpu_online_map on each CPU hotplug (cpuhp) event.
2123 */
2061 2124
2125static int cpuset_handle_cpuhp(struct notifier_block *nb,
2126 unsigned long phase, void *cpu)
2127{
2128 common_cpu_mem_hotplug_unplug();
2062 return 0; 2129 return 0;
2063} 2130}
2064#endif 2131#endif
2065 2132
2133#ifdef CONFIG_MEMORY_HOTPLUG
2134/*
2135 * Keep top_cpuset.mems_allowed tracking node_online_map.
2136 * Call this routine anytime after you change node_online_map.
2137 * See also the previous routine cpuset_handle_cpuhp().
2138 */
2139
2140void cpuset_track_online_nodes()
2141{
2142 common_cpu_mem_hotplug_unplug();
2143}
2144#endif
2145
2066/** 2146/**
2067 * cpuset_init_smp - initialize cpus_allowed 2147 * cpuset_init_smp - initialize cpus_allowed
2068 * 2148 *
diff --git a/kernel/exit.c b/kernel/exit.c
index d891883420f7..c189de2927ab 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -38,6 +38,7 @@
38#include <linux/pipe_fs_i.h> 38#include <linux/pipe_fs_i.h>
39#include <linux/audit.h> /* for audit_free() */ 39#include <linux/audit.h> /* for audit_free() */
40#include <linux/resource.h> 40#include <linux/resource.h>
41#include <linux/blkdev.h>
41 42
42#include <asm/uaccess.h> 43#include <asm/uaccess.h>
43#include <asm/unistd.h> 44#include <asm/unistd.h>
@@ -219,7 +220,7 @@ static int will_become_orphaned_pgrp(int pgrp, struct task_struct *ignored_task)
219 do_each_task_pid(pgrp, PIDTYPE_PGID, p) { 220 do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
220 if (p == ignored_task 221 if (p == ignored_task
221 || p->exit_state 222 || p->exit_state
222 || p->real_parent->pid == 1) 223 || is_init(p->real_parent))
223 continue; 224 continue;
224 if (process_group(p->real_parent) != pgrp 225 if (process_group(p->real_parent) != pgrp
225 && p->real_parent->signal->session == p->signal->session) { 226 && p->real_parent->signal->session == p->signal->session) {
@@ -249,17 +250,6 @@ static int has_stopped_jobs(int pgrp)
249 do_each_task_pid(pgrp, PIDTYPE_PGID, p) { 250 do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
250 if (p->state != TASK_STOPPED) 251 if (p->state != TASK_STOPPED)
251 continue; 252 continue;
252
253 /* If p is stopped by a debugger on a signal that won't
254 stop it, then don't count p as stopped. This isn't
255 perfect but it's a good approximation. */
256 if (unlikely (p->ptrace)
257 && p->exit_code != SIGSTOP
258 && p->exit_code != SIGTSTP
259 && p->exit_code != SIGTTOU
260 && p->exit_code != SIGTTIN)
261 continue;
262
263 retval = 1; 253 retval = 1;
264 break; 254 break;
265 } while_each_task_pid(pgrp, PIDTYPE_PGID, p); 255 } while_each_task_pid(pgrp, PIDTYPE_PGID, p);
@@ -292,9 +282,7 @@ static void reparent_to_init(void)
292 /* Set the exit signal to SIGCHLD so we signal init on exit */ 282 /* Set the exit signal to SIGCHLD so we signal init on exit */
293 current->exit_signal = SIGCHLD; 283 current->exit_signal = SIGCHLD;
294 284
295 if ((current->policy == SCHED_NORMAL || 285 if (!has_rt_policy(current) && (task_nice(current) < 0))
296 current->policy == SCHED_BATCH)
297 && (task_nice(current) < 0))
298 set_user_nice(current, 0); 286 set_user_nice(current, 0);
299 /* cpus_allowed? */ 287 /* cpus_allowed? */
300 /* rt_priority? */ 288 /* rt_priority? */
@@ -487,6 +475,18 @@ void fastcall put_files_struct(struct files_struct *files)
487 475
488EXPORT_SYMBOL(put_files_struct); 476EXPORT_SYMBOL(put_files_struct);
489 477
478void reset_files_struct(struct task_struct *tsk, struct files_struct *files)
479{
480 struct files_struct *old;
481
482 old = tsk->files;
483 task_lock(tsk);
484 tsk->files = files;
485 task_unlock(tsk);
486 put_files_struct(old);
487}
488EXPORT_SYMBOL(reset_files_struct);
489
490static inline void __exit_files(struct task_struct *tsk) 490static inline void __exit_files(struct task_struct *tsk)
491{ 491{
492 struct files_struct * files = tsk->files; 492 struct files_struct * files = tsk->files;
@@ -954,15 +954,15 @@ fastcall NORET_TYPE void do_exit(long code)
954 if (tsk->splice_pipe) 954 if (tsk->splice_pipe)
955 __free_pipe_info(tsk->splice_pipe); 955 __free_pipe_info(tsk->splice_pipe);
956 956
957 /* PF_DEAD causes final put_task_struct after we schedule. */
958 preempt_disable(); 957 preempt_disable();
959 BUG_ON(tsk->flags & PF_DEAD); 958 /* causes final put_task_struct in finish_task_switch(). */
960 tsk->flags |= PF_DEAD; 959 tsk->state = TASK_DEAD;
961 960
962 schedule(); 961 schedule();
963 BUG(); 962 BUG();
964 /* Avoid "noreturn function does return". */ 963 /* Avoid "noreturn function does return". */
965 for (;;) ; 964 for (;;)
965 cpu_relax(); /* For when BUG is null */
966} 966}
967 967
968EXPORT_SYMBOL_GPL(do_exit); 968EXPORT_SYMBOL_GPL(do_exit);
@@ -971,7 +971,7 @@ NORET_TYPE void complete_and_exit(struct completion *comp, long code)
971{ 971{
972 if (comp) 972 if (comp)
973 complete(comp); 973 complete(comp);
974 974
975 do_exit(code); 975 do_exit(code);
976} 976}
977 977
diff --git a/kernel/fork.c b/kernel/fork.c
index a0dad84567c9..1c999f3e0b47 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -183,7 +183,9 @@ static struct task_struct *dup_task_struct(struct task_struct *orig)
183 /* One for us, one for whoever does the "release_task()" (usually parent) */ 183 /* One for us, one for whoever does the "release_task()" (usually parent) */
184 atomic_set(&tsk->usage,2); 184 atomic_set(&tsk->usage,2);
185 atomic_set(&tsk->fs_excl, 0); 185 atomic_set(&tsk->fs_excl, 0);
186#ifdef CONFIG_BLK_DEV_IO_TRACE
186 tsk->btrace_seq = 0; 187 tsk->btrace_seq = 0;
188#endif
187 tsk->splice_pipe = NULL; 189 tsk->splice_pipe = NULL;
188 return tsk; 190 return tsk;
189} 191}
@@ -1061,7 +1063,11 @@ static struct task_struct *copy_process(unsigned long clone_flags,
1061#endif 1063#endif
1062#ifdef CONFIG_TRACE_IRQFLAGS 1064#ifdef CONFIG_TRACE_IRQFLAGS
1063 p->irq_events = 0; 1065 p->irq_events = 0;
1066#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
1067 p->hardirqs_enabled = 1;
1068#else
1064 p->hardirqs_enabled = 0; 1069 p->hardirqs_enabled = 0;
1070#endif
1065 p->hardirq_enable_ip = 0; 1071 p->hardirq_enable_ip = 0;
1066 p->hardirq_enable_event = 0; 1072 p->hardirq_enable_event = 0;
1067 p->hardirq_disable_ip = _THIS_IP_; 1073 p->hardirq_disable_ip = _THIS_IP_;
@@ -1144,7 +1150,6 @@ static struct task_struct *copy_process(unsigned long clone_flags,
1144 1150
1145 /* Our parent execution domain becomes current domain 1151 /* Our parent execution domain becomes current domain
1146 These must match for thread signalling to apply */ 1152 These must match for thread signalling to apply */
1147
1148 p->parent_exec_id = p->self_exec_id; 1153 p->parent_exec_id = p->self_exec_id;
1149 1154
1150 /* ok, now we should be set up.. */ 1155 /* ok, now we should be set up.. */
@@ -1167,6 +1172,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
1167 /* Need tasklist lock for parent etc handling! */ 1172 /* Need tasklist lock for parent etc handling! */
1168 write_lock_irq(&tasklist_lock); 1173 write_lock_irq(&tasklist_lock);
1169 1174
1175 /* for sys_ioprio_set(IOPRIO_WHO_PGRP) */
1176 p->ioprio = current->ioprio;
1177
1170 /* 1178 /*
1171 * The task hasn't been attached yet, so its cpus_allowed mask will 1179 * The task hasn't been attached yet, so its cpus_allowed mask will
1172 * not be changed, nor will its assigned CPU. 1180 * not be changed, nor will its assigned CPU.
@@ -1226,11 +1234,6 @@ static struct task_struct *copy_process(unsigned long clone_flags,
1226 } 1234 }
1227 } 1235 }
1228 1236
1229 /*
1230 * inherit ioprio
1231 */
1232 p->ioprio = current->ioprio;
1233
1234 if (likely(p->pid)) { 1237 if (likely(p->pid)) {
1235 add_parent(p); 1238 add_parent(p);
1236 if (unlikely(p->ptrace & PT_PTRACED)) 1239 if (unlikely(p->ptrace & PT_PTRACED))
diff --git a/kernel/futex.c b/kernel/futex.c
index 9d260e838cff..4b6770e9806d 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -389,7 +389,7 @@ static struct task_struct * futex_find_get_task(pid_t pid)
389{ 389{
390 struct task_struct *p; 390 struct task_struct *p;
391 391
392 read_lock(&tasklist_lock); 392 rcu_read_lock();
393 p = find_task_by_pid(pid); 393 p = find_task_by_pid(pid);
394 if (!p) 394 if (!p)
395 goto out_unlock; 395 goto out_unlock;
@@ -403,7 +403,7 @@ static struct task_struct * futex_find_get_task(pid_t pid)
403 } 403 }
404 get_task_struct(p); 404 get_task_struct(p);
405out_unlock: 405out_unlock:
406 read_unlock(&tasklist_lock); 406 rcu_read_unlock();
407 407
408 return p; 408 return p;
409} 409}
@@ -1624,7 +1624,7 @@ sys_get_robust_list(int pid, struct robust_list_head __user **head_ptr,
1624 struct task_struct *p; 1624 struct task_struct *p;
1625 1625
1626 ret = -ESRCH; 1626 ret = -ESRCH;
1627 read_lock(&tasklist_lock); 1627 rcu_read_lock();
1628 p = find_task_by_pid(pid); 1628 p = find_task_by_pid(pid);
1629 if (!p) 1629 if (!p)
1630 goto err_unlock; 1630 goto err_unlock;
@@ -1633,7 +1633,7 @@ sys_get_robust_list(int pid, struct robust_list_head __user **head_ptr,
1633 !capable(CAP_SYS_PTRACE)) 1633 !capable(CAP_SYS_PTRACE))
1634 goto err_unlock; 1634 goto err_unlock;
1635 head = p->robust_list; 1635 head = p->robust_list;
1636 read_unlock(&tasklist_lock); 1636 rcu_read_unlock();
1637 } 1637 }
1638 1638
1639 if (put_user(sizeof(*head), len_ptr)) 1639 if (put_user(sizeof(*head), len_ptr))
@@ -1641,7 +1641,7 @@ sys_get_robust_list(int pid, struct robust_list_head __user **head_ptr,
1641 return put_user(head, head_ptr); 1641 return put_user(head, head_ptr);
1642 1642
1643err_unlock: 1643err_unlock:
1644 read_unlock(&tasklist_lock); 1644 rcu_read_unlock();
1645 1645
1646 return ret; 1646 return ret;
1647} 1647}
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index 21c38a7e666b..d0ba190dfeb6 100644
--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -693,7 +693,7 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod
693 return t->task == NULL; 693 return t->task == NULL;
694} 694}
695 695
696static long __sched nanosleep_restart(struct restart_block *restart) 696long __sched hrtimer_nanosleep_restart(struct restart_block *restart)
697{ 697{
698 struct hrtimer_sleeper t; 698 struct hrtimer_sleeper t;
699 struct timespec __user *rmtp; 699 struct timespec __user *rmtp;
@@ -702,13 +702,13 @@ static long __sched nanosleep_restart(struct restart_block *restart)
702 702
703 restart->fn = do_no_restart_syscall; 703 restart->fn = do_no_restart_syscall;
704 704
705 hrtimer_init(&t.timer, restart->arg3, HRTIMER_ABS); 705 hrtimer_init(&t.timer, restart->arg0, HRTIMER_ABS);
706 t.timer.expires.tv64 = ((u64)restart->arg1 << 32) | (u64) restart->arg0; 706 t.timer.expires.tv64 = ((u64)restart->arg3 << 32) | (u64) restart->arg2;
707 707
708 if (do_nanosleep(&t, HRTIMER_ABS)) 708 if (do_nanosleep(&t, HRTIMER_ABS))
709 return 0; 709 return 0;
710 710
711 rmtp = (struct timespec __user *) restart->arg2; 711 rmtp = (struct timespec __user *) restart->arg1;
712 if (rmtp) { 712 if (rmtp) {
713 time = ktime_sub(t.timer.expires, t.timer.base->get_time()); 713 time = ktime_sub(t.timer.expires, t.timer.base->get_time());
714 if (time.tv64 <= 0) 714 if (time.tv64 <= 0)
@@ -718,7 +718,7 @@ static long __sched nanosleep_restart(struct restart_block *restart)
718 return -EFAULT; 718 return -EFAULT;
719 } 719 }
720 720
721 restart->fn = nanosleep_restart; 721 restart->fn = hrtimer_nanosleep_restart;
722 722
723 /* The other values in restart are already filled in */ 723 /* The other values in restart are already filled in */
724 return -ERESTART_RESTARTBLOCK; 724 return -ERESTART_RESTARTBLOCK;
@@ -751,11 +751,11 @@ long hrtimer_nanosleep(struct timespec *rqtp, struct timespec __user *rmtp,
751 } 751 }
752 752
753 restart = &current_thread_info()->restart_block; 753 restart = &current_thread_info()->restart_block;
754 restart->fn = nanosleep_restart; 754 restart->fn = hrtimer_nanosleep_restart;
755 restart->arg0 = t.timer.expires.tv64 & 0xFFFFFFFF; 755 restart->arg0 = (unsigned long) t.timer.base->index;
756 restart->arg1 = t.timer.expires.tv64 >> 32; 756 restart->arg1 = (unsigned long) rmtp;
757 restart->arg2 = (unsigned long) rmtp; 757 restart->arg2 = t.timer.expires.tv64 & 0xFFFFFFFF;
758 restart->arg3 = (unsigned long) t.timer.base->index; 758 restart->arg3 = t.timer.expires.tv64 >> 32;
759 759
760 return -ERESTART_RESTARTBLOCK; 760 return -ERESTART_RESTARTBLOCK;
761} 761}
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index ac1f850d4937..736cb0bd498f 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -40,10 +40,6 @@ int set_irq_chip(unsigned int irq, struct irq_chip *chip)
40 spin_lock_irqsave(&desc->lock, flags); 40 spin_lock_irqsave(&desc->lock, flags);
41 irq_chip_set_defaults(chip); 41 irq_chip_set_defaults(chip);
42 desc->chip = chip; 42 desc->chip = chip;
43 /*
44 * For compatibility only:
45 */
46 desc->chip = chip;
47 spin_unlock_irqrestore(&desc->lock, flags); 43 spin_unlock_irqrestore(&desc->lock, flags);
48 44
49 return 0; 45 return 0;
@@ -146,7 +142,7 @@ static void default_disable(unsigned int irq)
146 struct irq_desc *desc = irq_desc + irq; 142 struct irq_desc *desc = irq_desc + irq;
147 143
148 if (!(desc->status & IRQ_DELAYED_DISABLE)) 144 if (!(desc->status & IRQ_DELAYED_DISABLE))
149 irq_desc[irq].chip->mask(irq); 145 desc->chip->mask(irq);
150} 146}
151 147
152/* 148/*
diff --git a/kernel/kexec.c b/kernel/kexec.c
index 50087ecf337e..fcdd5d2bc3f4 100644
--- a/kernel/kexec.c
+++ b/kernel/kexec.c
@@ -40,7 +40,7 @@ struct resource crashk_res = {
40 40
41int kexec_should_crash(struct task_struct *p) 41int kexec_should_crash(struct task_struct *p)
42{ 42{
43 if (in_interrupt() || !p->pid || p->pid == 1 || panic_on_oops) 43 if (in_interrupt() || !p->pid || is_init(p) || panic_on_oops)
44 return 1; 44 return 1;
45 return 0; 45 return 0;
46} 46}
@@ -995,7 +995,8 @@ asmlinkage long sys_kexec_load(unsigned long entry, unsigned long nr_segments,
995 image = xchg(dest_image, image); 995 image = xchg(dest_image, image);
996 996
997out: 997out:
998 xchg(&kexec_lock, 0); /* Release the mutex */ 998 locked = xchg(&kexec_lock, 0); /* Release the mutex */
999 BUG_ON(!locked);
999 kimage_free(image); 1000 kimage_free(image);
1000 1001
1001 return result; 1002 return result;
@@ -1061,7 +1062,8 @@ void crash_kexec(struct pt_regs *regs)
1061 machine_crash_shutdown(&fixed_regs); 1062 machine_crash_shutdown(&fixed_regs);
1062 machine_kexec(kexec_crash_image); 1063 machine_kexec(kexec_crash_image);
1063 } 1064 }
1064 xchg(&kexec_lock, 0); 1065 locked = xchg(&kexec_lock, 0);
1066 BUG_ON(!locked);
1065 } 1067 }
1066} 1068}
1067 1069
diff --git a/kernel/kfifo.c b/kernel/kfifo.c
index 64ab045c3d9d..5d1d907378a2 100644
--- a/kernel/kfifo.c
+++ b/kernel/kfifo.c
@@ -122,6 +122,13 @@ unsigned int __kfifo_put(struct kfifo *fifo,
122 122
123 len = min(len, fifo->size - fifo->in + fifo->out); 123 len = min(len, fifo->size - fifo->in + fifo->out);
124 124
125 /*
126 * Ensure that we sample the fifo->out index -before- we
127 * start putting bytes into the kfifo.
128 */
129
130 smp_mb();
131
125 /* first put the data starting from fifo->in to buffer end */ 132 /* first put the data starting from fifo->in to buffer end */
126 l = min(len, fifo->size - (fifo->in & (fifo->size - 1))); 133 l = min(len, fifo->size - (fifo->in & (fifo->size - 1)));
127 memcpy(fifo->buffer + (fifo->in & (fifo->size - 1)), buffer, l); 134 memcpy(fifo->buffer + (fifo->in & (fifo->size - 1)), buffer, l);
@@ -129,6 +136,13 @@ unsigned int __kfifo_put(struct kfifo *fifo,
129 /* then put the rest (if any) at the beginning of the buffer */ 136 /* then put the rest (if any) at the beginning of the buffer */
130 memcpy(fifo->buffer, buffer + l, len - l); 137 memcpy(fifo->buffer, buffer + l, len - l);
131 138
139 /*
140 * Ensure that we add the bytes to the kfifo -before-
141 * we update the fifo->in index.
142 */
143
144 smp_wmb();
145
132 fifo->in += len; 146 fifo->in += len;
133 147
134 return len; 148 return len;
@@ -154,6 +168,13 @@ unsigned int __kfifo_get(struct kfifo *fifo,
154 168
155 len = min(len, fifo->in - fifo->out); 169 len = min(len, fifo->in - fifo->out);
156 170
171 /*
172 * Ensure that we sample the fifo->in index -before- we
173 * start removing bytes from the kfifo.
174 */
175
176 smp_rmb();
177
157 /* first get the data from fifo->out until the end of the buffer */ 178 /* first get the data from fifo->out until the end of the buffer */
158 l = min(len, fifo->size - (fifo->out & (fifo->size - 1))); 179 l = min(len, fifo->size - (fifo->out & (fifo->size - 1)));
159 memcpy(buffer, fifo->buffer + (fifo->out & (fifo->size - 1)), l); 180 memcpy(buffer, fifo->buffer + (fifo->out & (fifo->size - 1)), l);
@@ -161,6 +182,13 @@ unsigned int __kfifo_get(struct kfifo *fifo,
161 /* then get the rest (if any) from the beginning of the buffer */ 182 /* then get the rest (if any) from the beginning of the buffer */
162 memcpy(buffer + l, fifo->buffer, len - l); 183 memcpy(buffer + l, fifo->buffer, len - l);
163 184
185 /*
186 * Ensure that we remove the bytes from the kfifo -before-
187 * we update the fifo->out index.
188 */
189
190 smp_mb();
191
164 fifo->out += len; 192 fifo->out += len;
165 193
166 return len; 194 return len;
diff --git a/kernel/kmod.c b/kernel/kmod.c
index 5c470c57fb57..842f8015d7fd 100644
--- a/kernel/kmod.c
+++ b/kernel/kmod.c
@@ -176,6 +176,8 @@ static int wait_for_helper(void *data)
176 if (pid < 0) { 176 if (pid < 0) {
177 sub_info->retval = pid; 177 sub_info->retval = pid;
178 } else { 178 } else {
179 int ret;
180
179 /* 181 /*
180 * Normally it is bogus to call wait4() from in-kernel because 182 * Normally it is bogus to call wait4() from in-kernel because
181 * wait4() wants to write the exit code to a userspace address. 183 * wait4() wants to write the exit code to a userspace address.
@@ -185,7 +187,15 @@ static int wait_for_helper(void *data)
185 * 187 *
186 * Thus the __user pointer cast is valid here. 188 * Thus the __user pointer cast is valid here.
187 */ 189 */
188 sys_wait4(pid, (int __user *) &sub_info->retval, 0, NULL); 190 sys_wait4(pid, (int __user *)&ret, 0, NULL);
191
192 /*
193 * If ret is 0, either ____call_usermodehelper failed and the
194 * real error code is already in sub_info->retval or
195 * sub_info->retval is 0 anyway, so don't mess with it then.
196 */
197 if (ret)
198 sub_info->retval = ret;
189 } 199 }
190 200
191 complete(sub_info->complete); 201 complete(sub_info->complete);
diff --git a/kernel/lockdep.c b/kernel/lockdep.c
index c088e5542e84..e596525669ed 100644
--- a/kernel/lockdep.c
+++ b/kernel/lockdep.c
@@ -36,6 +36,7 @@
36#include <linux/stacktrace.h> 36#include <linux/stacktrace.h>
37#include <linux/debug_locks.h> 37#include <linux/debug_locks.h>
38#include <linux/irqflags.h> 38#include <linux/irqflags.h>
39#include <linux/utsname.h>
39 40
40#include <asm/sections.h> 41#include <asm/sections.h>
41 42
@@ -121,8 +122,8 @@ static struct list_head chainhash_table[CHAINHASH_SIZE];
121 * unique. 122 * unique.
122 */ 123 */
123#define iterate_chain_key(key1, key2) \ 124#define iterate_chain_key(key1, key2) \
124 (((key1) << MAX_LOCKDEP_KEYS_BITS/2) ^ \ 125 (((key1) << MAX_LOCKDEP_KEYS_BITS) ^ \
125 ((key1) >> (64-MAX_LOCKDEP_KEYS_BITS/2)) ^ \ 126 ((key1) >> (64-MAX_LOCKDEP_KEYS_BITS)) ^ \
126 (key2)) 127 (key2))
127 128
128void lockdep_off(void) 129void lockdep_off(void)
@@ -515,6 +516,13 @@ print_circular_bug_entry(struct lock_list *target, unsigned int depth)
515 return 0; 516 return 0;
516} 517}
517 518
519static void print_kernel_version(void)
520{
521 printk("%s %.*s\n", system_utsname.release,
522 (int)strcspn(system_utsname.version, " "),
523 system_utsname.version);
524}
525
518/* 526/*
519 * When a circular dependency is detected, print the 527 * When a circular dependency is detected, print the
520 * header first: 528 * header first:
@@ -531,6 +539,7 @@ print_circular_bug_header(struct lock_list *entry, unsigned int depth)
531 539
532 printk("\n=======================================================\n"); 540 printk("\n=======================================================\n");
533 printk( "[ INFO: possible circular locking dependency detected ]\n"); 541 printk( "[ INFO: possible circular locking dependency detected ]\n");
542 print_kernel_version();
534 printk( "-------------------------------------------------------\n"); 543 printk( "-------------------------------------------------------\n");
535 printk("%s/%d is trying to acquire lock:\n", 544 printk("%s/%d is trying to acquire lock:\n",
536 curr->comm, curr->pid); 545 curr->comm, curr->pid);
@@ -712,6 +721,7 @@ print_bad_irq_dependency(struct task_struct *curr,
712 printk("\n======================================================\n"); 721 printk("\n======================================================\n");
713 printk( "[ INFO: %s-safe -> %s-unsafe lock order detected ]\n", 722 printk( "[ INFO: %s-safe -> %s-unsafe lock order detected ]\n",
714 irqclass, irqclass); 723 irqclass, irqclass);
724 print_kernel_version();
715 printk( "------------------------------------------------------\n"); 725 printk( "------------------------------------------------------\n");
716 printk("%s/%d [HC%u[%lu]:SC%u[%lu]:HE%u:SE%u] is trying to acquire:\n", 726 printk("%s/%d [HC%u[%lu]:SC%u[%lu]:HE%u:SE%u] is trying to acquire:\n",
717 curr->comm, curr->pid, 727 curr->comm, curr->pid,
@@ -793,6 +803,7 @@ print_deadlock_bug(struct task_struct *curr, struct held_lock *prev,
793 803
794 printk("\n=============================================\n"); 804 printk("\n=============================================\n");
795 printk( "[ INFO: possible recursive locking detected ]\n"); 805 printk( "[ INFO: possible recursive locking detected ]\n");
806 print_kernel_version();
796 printk( "---------------------------------------------\n"); 807 printk( "---------------------------------------------\n");
797 printk("%s/%d is trying to acquire lock:\n", 808 printk("%s/%d is trying to acquire lock:\n",
798 curr->comm, curr->pid); 809 curr->comm, curr->pid);
@@ -1375,6 +1386,7 @@ print_irq_inversion_bug(struct task_struct *curr, struct lock_class *other,
1375 1386
1376 printk("\n=========================================================\n"); 1387 printk("\n=========================================================\n");
1377 printk( "[ INFO: possible irq lock inversion dependency detected ]\n"); 1388 printk( "[ INFO: possible irq lock inversion dependency detected ]\n");
1389 print_kernel_version();
1378 printk( "---------------------------------------------------------\n"); 1390 printk( "---------------------------------------------------------\n");
1379 printk("%s/%d just changed the state of lock:\n", 1391 printk("%s/%d just changed the state of lock:\n",
1380 curr->comm, curr->pid); 1392 curr->comm, curr->pid);
@@ -1469,6 +1481,7 @@ print_usage_bug(struct task_struct *curr, struct held_lock *this,
1469 1481
1470 printk("\n=================================\n"); 1482 printk("\n=================================\n");
1471 printk( "[ INFO: inconsistent lock state ]\n"); 1483 printk( "[ INFO: inconsistent lock state ]\n");
1484 print_kernel_version();
1472 printk( "---------------------------------\n"); 1485 printk( "---------------------------------\n");
1473 1486
1474 printk("inconsistent {%s} -> {%s} usage.\n", 1487 printk("inconsistent {%s} -> {%s} usage.\n",
diff --git a/kernel/module.c b/kernel/module.c
index b7fe6e840963..05625d5dc758 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -933,6 +933,15 @@ static ssize_t module_sect_show(struct module_attribute *mattr,
933 return sprintf(buf, "0x%lx\n", sattr->address); 933 return sprintf(buf, "0x%lx\n", sattr->address);
934} 934}
935 935
936static void free_sect_attrs(struct module_sect_attrs *sect_attrs)
937{
938 int section;
939
940 for (section = 0; section < sect_attrs->nsections; section++)
941 kfree(sect_attrs->attrs[section].name);
942 kfree(sect_attrs);
943}
944
936static void add_sect_attrs(struct module *mod, unsigned int nsect, 945static void add_sect_attrs(struct module *mod, unsigned int nsect,
937 char *secstrings, Elf_Shdr *sechdrs) 946 char *secstrings, Elf_Shdr *sechdrs)
938{ 947{
@@ -949,21 +958,26 @@ static void add_sect_attrs(struct module *mod, unsigned int nsect,
949 + nloaded * sizeof(sect_attrs->attrs[0]), 958 + nloaded * sizeof(sect_attrs->attrs[0]),
950 sizeof(sect_attrs->grp.attrs[0])); 959 sizeof(sect_attrs->grp.attrs[0]));
951 size[1] = (nloaded + 1) * sizeof(sect_attrs->grp.attrs[0]); 960 size[1] = (nloaded + 1) * sizeof(sect_attrs->grp.attrs[0]);
952 if (! (sect_attrs = kmalloc(size[0] + size[1], GFP_KERNEL))) 961 sect_attrs = kzalloc(size[0] + size[1], GFP_KERNEL);
962 if (sect_attrs == NULL)
953 return; 963 return;
954 964
955 /* Setup section attributes. */ 965 /* Setup section attributes. */
956 sect_attrs->grp.name = "sections"; 966 sect_attrs->grp.name = "sections";
957 sect_attrs->grp.attrs = (void *)sect_attrs + size[0]; 967 sect_attrs->grp.attrs = (void *)sect_attrs + size[0];
958 968
969 sect_attrs->nsections = 0;
959 sattr = &sect_attrs->attrs[0]; 970 sattr = &sect_attrs->attrs[0];
960 gattr = &sect_attrs->grp.attrs[0]; 971 gattr = &sect_attrs->grp.attrs[0];
961 for (i = 0; i < nsect; i++) { 972 for (i = 0; i < nsect; i++) {
962 if (! (sechdrs[i].sh_flags & SHF_ALLOC)) 973 if (! (sechdrs[i].sh_flags & SHF_ALLOC))
963 continue; 974 continue;
964 sattr->address = sechdrs[i].sh_addr; 975 sattr->address = sechdrs[i].sh_addr;
965 strlcpy(sattr->name, secstrings + sechdrs[i].sh_name, 976 sattr->name = kstrdup(secstrings + sechdrs[i].sh_name,
966 MODULE_SECT_NAME_LEN); 977 GFP_KERNEL);
978 if (sattr->name == NULL)
979 goto out;
980 sect_attrs->nsections++;
967 sattr->mattr.show = module_sect_show; 981 sattr->mattr.show = module_sect_show;
968 sattr->mattr.store = NULL; 982 sattr->mattr.store = NULL;
969 sattr->mattr.attr.name = sattr->name; 983 sattr->mattr.attr.name = sattr->name;
@@ -979,7 +993,7 @@ static void add_sect_attrs(struct module *mod, unsigned int nsect,
979 mod->sect_attrs = sect_attrs; 993 mod->sect_attrs = sect_attrs;
980 return; 994 return;
981 out: 995 out:
982 kfree(sect_attrs); 996 free_sect_attrs(sect_attrs);
983} 997}
984 998
985static void remove_sect_attrs(struct module *mod) 999static void remove_sect_attrs(struct module *mod)
@@ -989,13 +1003,13 @@ static void remove_sect_attrs(struct module *mod)
989 &mod->sect_attrs->grp); 1003 &mod->sect_attrs->grp);
990 /* We are positive that no one is using any sect attrs 1004 /* We are positive that no one is using any sect attrs
991 * at this point. Deallocate immediately. */ 1005 * at this point. Deallocate immediately. */
992 kfree(mod->sect_attrs); 1006 free_sect_attrs(mod->sect_attrs);
993 mod->sect_attrs = NULL; 1007 mod->sect_attrs = NULL;
994 } 1008 }
995} 1009}
996 1010
997
998#else 1011#else
1012
999static inline void add_sect_attrs(struct module *mod, unsigned int nsect, 1013static inline void add_sect_attrs(struct module *mod, unsigned int nsect,
1000 char *sectstrings, Elf_Shdr *sechdrs) 1014 char *sectstrings, Elf_Shdr *sechdrs)
1001{ 1015{
diff --git a/kernel/panic.c b/kernel/panic.c
index 6ceb664fb52a..525e365f7239 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -21,7 +21,6 @@
21#include <linux/debug_locks.h> 21#include <linux/debug_locks.h>
22 22
23int panic_on_oops; 23int panic_on_oops;
24int panic_on_unrecovered_nmi;
25int tainted; 24int tainted;
26static int pause_on_oops; 25static int pause_on_oops;
27static int pause_on_oops_flag; 26static int pause_on_oops_flag;
diff --git a/kernel/params.c b/kernel/params.c
index 91aea7aa532e..f406655d6653 100644
--- a/kernel/params.c
+++ b/kernel/params.c
@@ -547,6 +547,7 @@ static void __init kernel_param_sysfs_setup(const char *name,
547 unsigned int name_skip) 547 unsigned int name_skip)
548{ 548{
549 struct module_kobject *mk; 549 struct module_kobject *mk;
550 int ret;
550 551
551 mk = kzalloc(sizeof(struct module_kobject), GFP_KERNEL); 552 mk = kzalloc(sizeof(struct module_kobject), GFP_KERNEL);
552 BUG_ON(!mk); 553 BUG_ON(!mk);
@@ -554,7 +555,8 @@ static void __init kernel_param_sysfs_setup(const char *name,
554 mk->mod = THIS_MODULE; 555 mk->mod = THIS_MODULE;
555 kobj_set_kset_s(mk, module_subsys); 556 kobj_set_kset_s(mk, module_subsys);
556 kobject_set_name(&mk->kobj, name); 557 kobject_set_name(&mk->kobj, name);
557 kobject_register(&mk->kobj); 558 ret = kobject_register(&mk->kobj);
559 BUG_ON(ret < 0);
558 560
559 /* no need to keep the kobject if no parameter is exported */ 561 /* no need to keep the kobject if no parameter is exported */
560 if (!param_sysfs_setup(mk, kparam, num_params, name_skip)) { 562 if (!param_sysfs_setup(mk, kparam, num_params, name_skip)) {
@@ -684,13 +686,20 @@ decl_subsys(module, &module_ktype, NULL);
684 */ 686 */
685static int __init param_sysfs_init(void) 687static int __init param_sysfs_init(void)
686{ 688{
687 subsystem_register(&module_subsys); 689 int ret;
690
691 ret = subsystem_register(&module_subsys);
692 if (ret < 0) {
693 printk(KERN_WARNING "%s (%d): subsystem_register error: %d\n",
694 __FILE__, __LINE__, ret);
695 return ret;
696 }
688 697
689 param_sysfs_builtin(); 698 param_sysfs_builtin();
690 699
691 return 0; 700 return 0;
692} 701}
693__initcall(param_sysfs_init); 702subsys_initcall(param_sysfs_init);
694 703
695EXPORT_SYMBOL(param_set_byte); 704EXPORT_SYMBOL(param_set_byte);
696EXPORT_SYMBOL(param_get_byte); 705EXPORT_SYMBOL(param_get_byte);
diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c
index d38d9ec3276c..479b16b44f79 100644
--- a/kernel/posix-cpu-timers.c
+++ b/kernel/posix-cpu-timers.c
@@ -1393,25 +1393,13 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
1393 } 1393 }
1394} 1394}
1395 1395
1396static long posix_cpu_clock_nanosleep_restart(struct restart_block *); 1396static int do_cpu_nanosleep(const clockid_t which_clock, int flags,
1397 1397 struct timespec *rqtp, struct itimerspec *it)
1398int posix_cpu_nsleep(const clockid_t which_clock, int flags,
1399 struct timespec *rqtp, struct timespec __user *rmtp)
1400{ 1398{
1401 struct restart_block *restart_block =
1402 &current_thread_info()->restart_block;
1403 struct k_itimer timer; 1399 struct k_itimer timer;
1404 int error; 1400 int error;
1405 1401
1406 /* 1402 /*
1407 * Diagnose required errors first.
1408 */
1409 if (CPUCLOCK_PERTHREAD(which_clock) &&
1410 (CPUCLOCK_PID(which_clock) == 0 ||
1411 CPUCLOCK_PID(which_clock) == current->pid))
1412 return -EINVAL;
1413
1414 /*
1415 * Set up a temporary timer and then wait for it to go off. 1403 * Set up a temporary timer and then wait for it to go off.
1416 */ 1404 */
1417 memset(&timer, 0, sizeof timer); 1405 memset(&timer, 0, sizeof timer);
@@ -1422,11 +1410,12 @@ int posix_cpu_nsleep(const clockid_t which_clock, int flags,
1422 timer.it_process = current; 1410 timer.it_process = current;
1423 if (!error) { 1411 if (!error) {
1424 static struct itimerspec zero_it; 1412 static struct itimerspec zero_it;
1425 struct itimerspec it = { .it_value = *rqtp, 1413
1426 .it_interval = {} }; 1414 memset(it, 0, sizeof *it);
1415 it->it_value = *rqtp;
1427 1416
1428 spin_lock_irq(&timer.it_lock); 1417 spin_lock_irq(&timer.it_lock);
1429 error = posix_cpu_timer_set(&timer, flags, &it, NULL); 1418 error = posix_cpu_timer_set(&timer, flags, it, NULL);
1430 if (error) { 1419 if (error) {
1431 spin_unlock_irq(&timer.it_lock); 1420 spin_unlock_irq(&timer.it_lock);
1432 return error; 1421 return error;
@@ -1454,49 +1443,89 @@ int posix_cpu_nsleep(const clockid_t which_clock, int flags,
1454 * We were interrupted by a signal. 1443 * We were interrupted by a signal.
1455 */ 1444 */
1456 sample_to_timespec(which_clock, timer.it.cpu.expires, rqtp); 1445 sample_to_timespec(which_clock, timer.it.cpu.expires, rqtp);
1457 posix_cpu_timer_set(&timer, 0, &zero_it, &it); 1446 posix_cpu_timer_set(&timer, 0, &zero_it, it);
1458 spin_unlock_irq(&timer.it_lock); 1447 spin_unlock_irq(&timer.it_lock);
1459 1448
1460 if ((it.it_value.tv_sec | it.it_value.tv_nsec) == 0) { 1449 if ((it->it_value.tv_sec | it->it_value.tv_nsec) == 0) {
1461 /* 1450 /*
1462 * It actually did fire already. 1451 * It actually did fire already.
1463 */ 1452 */
1464 return 0; 1453 return 0;
1465 } 1454 }
1466 1455
1456 error = -ERESTART_RESTARTBLOCK;
1457 }
1458
1459 return error;
1460}
1461
1462int posix_cpu_nsleep(const clockid_t which_clock, int flags,
1463 struct timespec *rqtp, struct timespec __user *rmtp)
1464{
1465 struct restart_block *restart_block =
1466 &current_thread_info()->restart_block;
1467 struct itimerspec it;
1468 int error;
1469
1470 /*
1471 * Diagnose required errors first.
1472 */
1473 if (CPUCLOCK_PERTHREAD(which_clock) &&
1474 (CPUCLOCK_PID(which_clock) == 0 ||
1475 CPUCLOCK_PID(which_clock) == current->pid))
1476 return -EINVAL;
1477
1478 error = do_cpu_nanosleep(which_clock, flags, rqtp, &it);
1479
1480 if (error == -ERESTART_RESTARTBLOCK) {
1481
1482 if (flags & TIMER_ABSTIME)
1483 return -ERESTARTNOHAND;
1467 /* 1484 /*
1468 * Report back to the user the time still remaining. 1485 * Report back to the user the time still remaining.
1469 */ 1486 */
1470 if (rmtp != NULL && !(flags & TIMER_ABSTIME) && 1487 if (rmtp != NULL && copy_to_user(rmtp, &it.it_value, sizeof *rmtp))
1471 copy_to_user(rmtp, &it.it_value, sizeof *rmtp))
1472 return -EFAULT; 1488 return -EFAULT;
1473 1489
1474 restart_block->fn = posix_cpu_clock_nanosleep_restart; 1490 restart_block->fn = posix_cpu_nsleep_restart;
1475 /* Caller already set restart_block->arg1 */
1476 restart_block->arg0 = which_clock; 1491 restart_block->arg0 = which_clock;
1477 restart_block->arg1 = (unsigned long) rmtp; 1492 restart_block->arg1 = (unsigned long) rmtp;
1478 restart_block->arg2 = rqtp->tv_sec; 1493 restart_block->arg2 = rqtp->tv_sec;
1479 restart_block->arg3 = rqtp->tv_nsec; 1494 restart_block->arg3 = rqtp->tv_nsec;
1480
1481 error = -ERESTART_RESTARTBLOCK;
1482 } 1495 }
1483
1484 return error; 1496 return error;
1485} 1497}
1486 1498
1487static long 1499long posix_cpu_nsleep_restart(struct restart_block *restart_block)
1488posix_cpu_clock_nanosleep_restart(struct restart_block *restart_block)
1489{ 1500{
1490 clockid_t which_clock = restart_block->arg0; 1501 clockid_t which_clock = restart_block->arg0;
1491 struct timespec __user *rmtp; 1502 struct timespec __user *rmtp;
1492 struct timespec t; 1503 struct timespec t;
1504 struct itimerspec it;
1505 int error;
1493 1506
1494 rmtp = (struct timespec __user *) restart_block->arg1; 1507 rmtp = (struct timespec __user *) restart_block->arg1;
1495 t.tv_sec = restart_block->arg2; 1508 t.tv_sec = restart_block->arg2;
1496 t.tv_nsec = restart_block->arg3; 1509 t.tv_nsec = restart_block->arg3;
1497 1510
1498 restart_block->fn = do_no_restart_syscall; 1511 restart_block->fn = do_no_restart_syscall;
1499 return posix_cpu_nsleep(which_clock, TIMER_ABSTIME, &t, rmtp); 1512 error = do_cpu_nanosleep(which_clock, TIMER_ABSTIME, &t, &it);
1513
1514 if (error == -ERESTART_RESTARTBLOCK) {
1515 /*
1516 * Report back to the user the time still remaining.
1517 */
1518 if (rmtp != NULL && copy_to_user(rmtp, &it.it_value, sizeof *rmtp))
1519 return -EFAULT;
1520
1521 restart_block->fn = posix_cpu_nsleep_restart;
1522 restart_block->arg0 = which_clock;
1523 restart_block->arg1 = (unsigned long) rmtp;
1524 restart_block->arg2 = t.tv_sec;
1525 restart_block->arg3 = t.tv_nsec;
1526 }
1527 return error;
1528
1500} 1529}
1501 1530
1502 1531
@@ -1524,6 +1553,10 @@ static int process_cpu_nsleep(const clockid_t which_clock, int flags,
1524{ 1553{
1525 return posix_cpu_nsleep(PROCESS_CLOCK, flags, rqtp, rmtp); 1554 return posix_cpu_nsleep(PROCESS_CLOCK, flags, rqtp, rmtp);
1526} 1555}
1556static long process_cpu_nsleep_restart(struct restart_block *restart_block)
1557{
1558 return -EINVAL;
1559}
1527static int thread_cpu_clock_getres(const clockid_t which_clock, 1560static int thread_cpu_clock_getres(const clockid_t which_clock,
1528 struct timespec *tp) 1561 struct timespec *tp)
1529{ 1562{
@@ -1544,6 +1577,10 @@ static int thread_cpu_nsleep(const clockid_t which_clock, int flags,
1544{ 1577{
1545 return -EINVAL; 1578 return -EINVAL;
1546} 1579}
1580static long thread_cpu_nsleep_restart(struct restart_block *restart_block)
1581{
1582 return -EINVAL;
1583}
1547 1584
1548static __init int init_posix_cpu_timers(void) 1585static __init int init_posix_cpu_timers(void)
1549{ 1586{
@@ -1553,6 +1590,7 @@ static __init int init_posix_cpu_timers(void)
1553 .clock_set = do_posix_clock_nosettime, 1590 .clock_set = do_posix_clock_nosettime,
1554 .timer_create = process_cpu_timer_create, 1591 .timer_create = process_cpu_timer_create,
1555 .nsleep = process_cpu_nsleep, 1592 .nsleep = process_cpu_nsleep,
1593 .nsleep_restart = process_cpu_nsleep_restart,
1556 }; 1594 };
1557 struct k_clock thread = { 1595 struct k_clock thread = {
1558 .clock_getres = thread_cpu_clock_getres, 1596 .clock_getres = thread_cpu_clock_getres,
@@ -1560,6 +1598,7 @@ static __init int init_posix_cpu_timers(void)
1560 .clock_set = do_posix_clock_nosettime, 1598 .clock_set = do_posix_clock_nosettime,
1561 .timer_create = thread_cpu_timer_create, 1599 .timer_create = thread_cpu_timer_create,
1562 .nsleep = thread_cpu_nsleep, 1600 .nsleep = thread_cpu_nsleep,
1601 .nsleep_restart = thread_cpu_nsleep_restart,
1563 }; 1602 };
1564 1603
1565 register_posix_clock(CLOCK_PROCESS_CPUTIME_ID, &process); 1604 register_posix_clock(CLOCK_PROCESS_CPUTIME_ID, &process);
diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c
index ac6dc8744429..e5ebcc1ec3a0 100644
--- a/kernel/posix-timers.c
+++ b/kernel/posix-timers.c
@@ -973,3 +973,24 @@ sys_clock_nanosleep(const clockid_t which_clock, int flags,
973 return CLOCK_DISPATCH(which_clock, nsleep, 973 return CLOCK_DISPATCH(which_clock, nsleep,
974 (which_clock, flags, &t, rmtp)); 974 (which_clock, flags, &t, rmtp));
975} 975}
976
977/*
978 * nanosleep_restart for monotonic and realtime clocks
979 */
980static int common_nsleep_restart(struct restart_block *restart_block)
981{
982 return hrtimer_nanosleep_restart(restart_block);
983}
984
985/*
986 * This will restart clock_nanosleep. This is required only by
987 * compat_clock_nanosleep_restart for now.
988 */
989long
990clock_nanosleep_restart(struct restart_block *restart_block)
991{
992 clockid_t which_clock = restart_block->arg0;
993
994 return CLOCK_DISPATCH(which_clock, nsleep_restart,
995 (restart_block));
996}
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 8aad0331d82e..4d50e06fd745 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -440,6 +440,7 @@ struct task_struct *ptrace_get_task_struct(pid_t pid)
440 child = find_task_by_pid(pid); 440 child = find_task_by_pid(pid);
441 if (child) 441 if (child)
442 get_task_struct(child); 442 get_task_struct(child);
443
443 read_unlock(&tasklist_lock); 444 read_unlock(&tasklist_lock);
444 if (!child) 445 if (!child)
445 return ERR_PTR(-ESRCH); 446 return ERR_PTR(-ESRCH);
diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c
index 4d1c3d247127..4f2c4272d59c 100644
--- a/kernel/rcutorture.c
+++ b/kernel/rcutorture.c
@@ -192,13 +192,13 @@ static struct rcu_torture_ops *cur_ops = NULL;
192 * Definitions for rcu torture testing. 192 * Definitions for rcu torture testing.
193 */ 193 */
194 194
195static int rcu_torture_read_lock(void) 195static int rcu_torture_read_lock(void) __acquires(RCU)
196{ 196{
197 rcu_read_lock(); 197 rcu_read_lock();
198 return 0; 198 return 0;
199} 199}
200 200
201static void rcu_torture_read_unlock(int idx) 201static void rcu_torture_read_unlock(int idx) __releases(RCU)
202{ 202{
203 rcu_read_unlock(); 203 rcu_read_unlock();
204} 204}
@@ -250,13 +250,13 @@ static struct rcu_torture_ops rcu_ops = {
250 * Definitions for rcu_bh torture testing. 250 * Definitions for rcu_bh torture testing.
251 */ 251 */
252 252
253static int rcu_bh_torture_read_lock(void) 253static int rcu_bh_torture_read_lock(void) __acquires(RCU_BH)
254{ 254{
255 rcu_read_lock_bh(); 255 rcu_read_lock_bh();
256 return 0; 256 return 0;
257} 257}
258 258
259static void rcu_bh_torture_read_unlock(int idx) 259static void rcu_bh_torture_read_unlock(int idx) __releases(RCU_BH)
260{ 260{
261 rcu_read_unlock_bh(); 261 rcu_read_unlock_bh();
262} 262}
diff --git a/kernel/relay.c b/kernel/relay.c
index 85786ff2a4f9..1d63ecddfa70 100644
--- a/kernel/relay.c
+++ b/kernel/relay.c
@@ -95,7 +95,7 @@ int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma)
95 * @buf: the buffer struct 95 * @buf: the buffer struct
96 * @size: total size of the buffer 96 * @size: total size of the buffer
97 * 97 *
98 * Returns a pointer to the resulting buffer, NULL if unsuccessful. The 98 * Returns a pointer to the resulting buffer, %NULL if unsuccessful. The
99 * passed in size will get page aligned, if it isn't already. 99 * passed in size will get page aligned, if it isn't already.
100 */ 100 */
101static void *relay_alloc_buf(struct rchan_buf *buf, size_t *size) 101static void *relay_alloc_buf(struct rchan_buf *buf, size_t *size)
@@ -132,10 +132,9 @@ depopulate:
132 132
133/** 133/**
134 * relay_create_buf - allocate and initialize a channel buffer 134 * relay_create_buf - allocate and initialize a channel buffer
135 * @alloc_size: size of the buffer to allocate 135 * @chan: the relay channel
136 * @n_subbufs: number of sub-buffers in the channel
137 * 136 *
138 * Returns channel buffer if successful, NULL otherwise 137 * Returns channel buffer if successful, %NULL otherwise.
139 */ 138 */
140struct rchan_buf *relay_create_buf(struct rchan *chan) 139struct rchan_buf *relay_create_buf(struct rchan *chan)
141{ 140{
@@ -163,6 +162,7 @@ free_buf:
163 162
164/** 163/**
165 * relay_destroy_channel - free the channel struct 164 * relay_destroy_channel - free the channel struct
165 * @kref: target kernel reference that contains the relay channel
166 * 166 *
167 * Should only be called from kref_put(). 167 * Should only be called from kref_put().
168 */ 168 */
@@ -194,6 +194,7 @@ void relay_destroy_buf(struct rchan_buf *buf)
194 194
195/** 195/**
196 * relay_remove_buf - remove a channel buffer 196 * relay_remove_buf - remove a channel buffer
197 * @kref: target kernel reference that contains the relay buffer
197 * 198 *
198 * Removes the file from the fileystem, which also frees the 199 * Removes the file from the fileystem, which also frees the
199 * rchan_buf_struct and the channel buffer. Should only be called from 200 * rchan_buf_struct and the channel buffer. Should only be called from
@@ -374,7 +375,7 @@ void relay_reset(struct rchan *chan)
374} 375}
375EXPORT_SYMBOL_GPL(relay_reset); 376EXPORT_SYMBOL_GPL(relay_reset);
376 377
377/** 378/*
378 * relay_open_buf - create a new relay channel buffer 379 * relay_open_buf - create a new relay channel buffer
379 * 380 *
380 * Internal - used by relay_open(). 381 * Internal - used by relay_open().
@@ -448,12 +449,12 @@ static inline void setup_callbacks(struct rchan *chan,
448/** 449/**
449 * relay_open - create a new relay channel 450 * relay_open - create a new relay channel
450 * @base_filename: base name of files to create 451 * @base_filename: base name of files to create
451 * @parent: dentry of parent directory, NULL for root directory 452 * @parent: dentry of parent directory, %NULL for root directory
452 * @subbuf_size: size of sub-buffers 453 * @subbuf_size: size of sub-buffers
453 * @n_subbufs: number of sub-buffers 454 * @n_subbufs: number of sub-buffers
454 * @cb: client callback functions 455 * @cb: client callback functions
455 * 456 *
456 * Returns channel pointer if successful, NULL otherwise. 457 * Returns channel pointer if successful, %NULL otherwise.
457 * 458 *
458 * Creates a channel buffer for each cpu using the sizes and 459 * Creates a channel buffer for each cpu using the sizes and
459 * attributes specified. The created channel buffer files 460 * attributes specified. The created channel buffer files
@@ -585,7 +586,7 @@ EXPORT_SYMBOL_GPL(relay_switch_subbuf);
585 * subbufs_consumed should be the number of sub-buffers newly consumed, 586 * subbufs_consumed should be the number of sub-buffers newly consumed,
586 * not the total consumed. 587 * not the total consumed.
587 * 588 *
588 * NOTE: kernel clients don't need to call this function if the channel 589 * NOTE: Kernel clients don't need to call this function if the channel
589 * mode is 'overwrite'. 590 * mode is 'overwrite'.
590 */ 591 */
591void relay_subbufs_consumed(struct rchan *chan, 592void relay_subbufs_consumed(struct rchan *chan,
@@ -641,7 +642,7 @@ EXPORT_SYMBOL_GPL(relay_close);
641 * relay_flush - close the channel 642 * relay_flush - close the channel
642 * @chan: the channel 643 * @chan: the channel
643 * 644 *
644 * Flushes all channel buffers i.e. forces buffer switch. 645 * Flushes all channel buffers, i.e. forces buffer switch.
645 */ 646 */
646void relay_flush(struct rchan *chan) 647void relay_flush(struct rchan *chan)
647{ 648{
@@ -729,7 +730,7 @@ static int relay_file_release(struct inode *inode, struct file *filp)
729 return 0; 730 return 0;
730} 731}
731 732
732/** 733/*
733 * relay_file_read_consume - update the consumed count for the buffer 734 * relay_file_read_consume - update the consumed count for the buffer
734 */ 735 */
735static void relay_file_read_consume(struct rchan_buf *buf, 736static void relay_file_read_consume(struct rchan_buf *buf,
@@ -756,7 +757,7 @@ static void relay_file_read_consume(struct rchan_buf *buf,
756 } 757 }
757} 758}
758 759
759/** 760/*
760 * relay_file_read_avail - boolean, are there unconsumed bytes available? 761 * relay_file_read_avail - boolean, are there unconsumed bytes available?
761 */ 762 */
762static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos) 763static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos)
@@ -793,6 +794,8 @@ static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos)
793 794
794/** 795/**
795 * relay_file_read_subbuf_avail - return bytes available in sub-buffer 796 * relay_file_read_subbuf_avail - return bytes available in sub-buffer
797 * @read_pos: file read position
798 * @buf: relay channel buffer
796 */ 799 */
797static size_t relay_file_read_subbuf_avail(size_t read_pos, 800static size_t relay_file_read_subbuf_avail(size_t read_pos,
798 struct rchan_buf *buf) 801 struct rchan_buf *buf)
@@ -818,6 +821,8 @@ static size_t relay_file_read_subbuf_avail(size_t read_pos,
818 821
819/** 822/**
820 * relay_file_read_start_pos - find the first available byte to read 823 * relay_file_read_start_pos - find the first available byte to read
824 * @read_pos: file read position
825 * @buf: relay channel buffer
821 * 826 *
822 * If the read_pos is in the middle of padding, return the 827 * If the read_pos is in the middle of padding, return the
823 * position of the first actually available byte, otherwise 828 * position of the first actually available byte, otherwise
@@ -844,6 +849,9 @@ static size_t relay_file_read_start_pos(size_t read_pos,
844 849
845/** 850/**
846 * relay_file_read_end_pos - return the new read position 851 * relay_file_read_end_pos - return the new read position
852 * @read_pos: file read position
853 * @buf: relay channel buffer
854 * @count: number of bytes to be read
847 */ 855 */
848static size_t relay_file_read_end_pos(struct rchan_buf *buf, 856static size_t relay_file_read_end_pos(struct rchan_buf *buf,
849 size_t read_pos, 857 size_t read_pos,
@@ -865,7 +873,7 @@ static size_t relay_file_read_end_pos(struct rchan_buf *buf,
865 return end_pos; 873 return end_pos;
866} 874}
867 875
868/** 876/*
869 * subbuf_read_actor - read up to one subbuf's worth of data 877 * subbuf_read_actor - read up to one subbuf's worth of data
870 */ 878 */
871static int subbuf_read_actor(size_t read_start, 879static int subbuf_read_actor(size_t read_start,
@@ -890,7 +898,7 @@ static int subbuf_read_actor(size_t read_start,
890 return ret; 898 return ret;
891} 899}
892 900
893/** 901/*
894 * subbuf_send_actor - send up to one subbuf's worth of data 902 * subbuf_send_actor - send up to one subbuf's worth of data
895 */ 903 */
896static int subbuf_send_actor(size_t read_start, 904static int subbuf_send_actor(size_t read_start,
@@ -933,7 +941,7 @@ typedef int (*subbuf_actor_t) (size_t read_start,
933 read_descriptor_t *desc, 941 read_descriptor_t *desc,
934 read_actor_t actor); 942 read_actor_t actor);
935 943
936/** 944/*
937 * relay_file_read_subbufs - read count bytes, bridging subbuf boundaries 945 * relay_file_read_subbufs - read count bytes, bridging subbuf boundaries
938 */ 946 */
939static inline ssize_t relay_file_read_subbufs(struct file *filp, 947static inline ssize_t relay_file_read_subbufs(struct file *filp,
diff --git a/kernel/rtmutex.c b/kernel/rtmutex.c
index 3e13a1e5856f..4ab17da46fd8 100644
--- a/kernel/rtmutex.c
+++ b/kernel/rtmutex.c
@@ -251,6 +251,7 @@ static int rt_mutex_adjust_prio_chain(struct task_struct *task,
251 251
252 /* Grab the next task */ 252 /* Grab the next task */
253 task = rt_mutex_owner(lock); 253 task = rt_mutex_owner(lock);
254 get_task_struct(task);
254 spin_lock_irqsave(&task->pi_lock, flags); 255 spin_lock_irqsave(&task->pi_lock, flags);
255 256
256 if (waiter == rt_mutex_top_waiter(lock)) { 257 if (waiter == rt_mutex_top_waiter(lock)) {
@@ -269,7 +270,6 @@ static int rt_mutex_adjust_prio_chain(struct task_struct *task,
269 __rt_mutex_adjust_prio(task); 270 __rt_mutex_adjust_prio(task);
270 } 271 }
271 272
272 get_task_struct(task);
273 spin_unlock_irqrestore(&task->pi_lock, flags); 273 spin_unlock_irqrestore(&task->pi_lock, flags);
274 274
275 top_waiter = rt_mutex_top_waiter(lock); 275 top_waiter = rt_mutex_top_waiter(lock);
@@ -409,7 +409,7 @@ static int task_blocks_on_rt_mutex(struct rt_mutex *lock,
409 struct task_struct *owner = rt_mutex_owner(lock); 409 struct task_struct *owner = rt_mutex_owner(lock);
410 struct rt_mutex_waiter *top_waiter = waiter; 410 struct rt_mutex_waiter *top_waiter = waiter;
411 unsigned long flags; 411 unsigned long flags;
412 int boost = 0, res; 412 int chain_walk = 0, res;
413 413
414 spin_lock_irqsave(&current->pi_lock, flags); 414 spin_lock_irqsave(&current->pi_lock, flags);
415 __rt_mutex_adjust_prio(current); 415 __rt_mutex_adjust_prio(current);
@@ -433,25 +433,23 @@ static int task_blocks_on_rt_mutex(struct rt_mutex *lock,
433 plist_add(&waiter->pi_list_entry, &owner->pi_waiters); 433 plist_add(&waiter->pi_list_entry, &owner->pi_waiters);
434 434
435 __rt_mutex_adjust_prio(owner); 435 __rt_mutex_adjust_prio(owner);
436 if (owner->pi_blocked_on) { 436 if (owner->pi_blocked_on)
437 boost = 1; 437 chain_walk = 1;
438 /* gets dropped in rt_mutex_adjust_prio_chain()! */
439 get_task_struct(owner);
440 }
441 spin_unlock_irqrestore(&owner->pi_lock, flags);
442 }
443 else if (debug_rt_mutex_detect_deadlock(waiter, detect_deadlock)) {
444 spin_lock_irqsave(&owner->pi_lock, flags);
445 if (owner->pi_blocked_on) {
446 boost = 1;
447 /* gets dropped in rt_mutex_adjust_prio_chain()! */
448 get_task_struct(owner);
449 }
450 spin_unlock_irqrestore(&owner->pi_lock, flags); 438 spin_unlock_irqrestore(&owner->pi_lock, flags);
451 } 439 }
452 if (!boost) 440 else if (debug_rt_mutex_detect_deadlock(waiter, detect_deadlock))
441 chain_walk = 1;
442
443 if (!chain_walk)
453 return 0; 444 return 0;
454 445
446 /*
447 * The owner can't disappear while holding a lock,
448 * so the owner struct is protected by wait_lock.
449 * Gets dropped in rt_mutex_adjust_prio_chain()!
450 */
451 get_task_struct(owner);
452
455 spin_unlock(&lock->wait_lock); 453 spin_unlock(&lock->wait_lock);
456 454
457 res = rt_mutex_adjust_prio_chain(owner, detect_deadlock, lock, waiter, 455 res = rt_mutex_adjust_prio_chain(owner, detect_deadlock, lock, waiter,
@@ -532,7 +530,7 @@ static void remove_waiter(struct rt_mutex *lock,
532 int first = (waiter == rt_mutex_top_waiter(lock)); 530 int first = (waiter == rt_mutex_top_waiter(lock));
533 struct task_struct *owner = rt_mutex_owner(lock); 531 struct task_struct *owner = rt_mutex_owner(lock);
534 unsigned long flags; 532 unsigned long flags;
535 int boost = 0; 533 int chain_walk = 0;
536 534
537 spin_lock_irqsave(&current->pi_lock, flags); 535 spin_lock_irqsave(&current->pi_lock, flags);
538 plist_del(&waiter->list_entry, &lock->wait_list); 536 plist_del(&waiter->list_entry, &lock->wait_list);
@@ -554,19 +552,20 @@ static void remove_waiter(struct rt_mutex *lock,
554 } 552 }
555 __rt_mutex_adjust_prio(owner); 553 __rt_mutex_adjust_prio(owner);
556 554
557 if (owner->pi_blocked_on) { 555 if (owner->pi_blocked_on)
558 boost = 1; 556 chain_walk = 1;
559 /* gets dropped in rt_mutex_adjust_prio_chain()! */ 557
560 get_task_struct(owner);
561 }
562 spin_unlock_irqrestore(&owner->pi_lock, flags); 558 spin_unlock_irqrestore(&owner->pi_lock, flags);
563 } 559 }
564 560
565 WARN_ON(!plist_node_empty(&waiter->pi_list_entry)); 561 WARN_ON(!plist_node_empty(&waiter->pi_list_entry));
566 562
567 if (!boost) 563 if (!chain_walk)
568 return; 564 return;
569 565
566 /* gets dropped in rt_mutex_adjust_prio_chain()! */
567 get_task_struct(owner);
568
570 spin_unlock(&lock->wait_lock); 569 spin_unlock(&lock->wait_lock);
571 570
572 rt_mutex_adjust_prio_chain(owner, 0, lock, NULL, current); 571 rt_mutex_adjust_prio_chain(owner, 0, lock, NULL, current);
@@ -592,10 +591,10 @@ void rt_mutex_adjust_pi(struct task_struct *task)
592 return; 591 return;
593 } 592 }
594 593
595 /* gets dropped in rt_mutex_adjust_prio_chain()! */
596 get_task_struct(task);
597 spin_unlock_irqrestore(&task->pi_lock, flags); 594 spin_unlock_irqrestore(&task->pi_lock, flags);
598 595
596 /* gets dropped in rt_mutex_adjust_prio_chain()! */
597 get_task_struct(task);
599 rt_mutex_adjust_prio_chain(task, 0, NULL, NULL, task); 598 rt_mutex_adjust_prio_chain(task, 0, NULL, NULL, task);
600} 599}
601 600
diff --git a/kernel/sched.c b/kernel/sched.c
index 5c848fd4e461..74f169ac0773 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -1755,27 +1755,27 @@ static inline void finish_task_switch(struct rq *rq, struct task_struct *prev)
1755 __releases(rq->lock) 1755 __releases(rq->lock)
1756{ 1756{
1757 struct mm_struct *mm = rq->prev_mm; 1757 struct mm_struct *mm = rq->prev_mm;
1758 unsigned long prev_task_flags; 1758 long prev_state;
1759 1759
1760 rq->prev_mm = NULL; 1760 rq->prev_mm = NULL;
1761 1761
1762 /* 1762 /*
1763 * A task struct has one reference for the use as "current". 1763 * A task struct has one reference for the use as "current".
1764 * If a task dies, then it sets EXIT_ZOMBIE in tsk->exit_state and 1764 * If a task dies, then it sets TASK_DEAD in tsk->state and calls
1765 * calls schedule one last time. The schedule call will never return, 1765 * schedule one last time. The schedule call will never return, and
1766 * and the scheduled task must drop that reference. 1766 * the scheduled task must drop that reference.
1767 * The test for EXIT_ZOMBIE must occur while the runqueue locks are 1767 * The test for TASK_DEAD must occur while the runqueue locks are
1768 * still held, otherwise prev could be scheduled on another cpu, die 1768 * still held, otherwise prev could be scheduled on another cpu, die
1769 * there before we look at prev->state, and then the reference would 1769 * there before we look at prev->state, and then the reference would
1770 * be dropped twice. 1770 * be dropped twice.
1771 * Manfred Spraul <manfred@colorfullife.com> 1771 * Manfred Spraul <manfred@colorfullife.com>
1772 */ 1772 */
1773 prev_task_flags = prev->flags; 1773 prev_state = prev->state;
1774 finish_arch_switch(prev); 1774 finish_arch_switch(prev);
1775 finish_lock_switch(rq, prev); 1775 finish_lock_switch(rq, prev);
1776 if (mm) 1776 if (mm)
1777 mmdrop(mm); 1777 mmdrop(mm);
1778 if (unlikely(prev_task_flags & PF_DEAD)) { 1778 if (unlikely(prev_state == TASK_DEAD)) {
1779 /* 1779 /*
1780 * Remove function-return probe instances associated with this 1780 * Remove function-return probe instances associated with this
1781 * task and put them back on the free list. 1781 * task and put them back on the free list.
@@ -3348,9 +3348,6 @@ need_resched_nonpreemptible:
3348 3348
3349 spin_lock_irq(&rq->lock); 3349 spin_lock_irq(&rq->lock);
3350 3350
3351 if (unlikely(prev->flags & PF_DEAD))
3352 prev->state = EXIT_DEAD;
3353
3354 switch_count = &prev->nivcsw; 3351 switch_count = &prev->nivcsw;
3355 if (prev->state && !(preempt_count() & PREEMPT_ACTIVE)) { 3352 if (prev->state && !(preempt_count() & PREEMPT_ACTIVE)) {
3356 switch_count = &prev->nvcsw; 3353 switch_count = &prev->nvcsw;
@@ -4080,6 +4077,8 @@ static void __setscheduler(struct task_struct *p, int policy, int prio)
4080 * @p: the task in question. 4077 * @p: the task in question.
4081 * @policy: new policy. 4078 * @policy: new policy.
4082 * @param: structure containing the new RT priority. 4079 * @param: structure containing the new RT priority.
4080 *
4081 * NOTE: the task may be already dead
4083 */ 4082 */
4084int sched_setscheduler(struct task_struct *p, int policy, 4083int sched_setscheduler(struct task_struct *p, int policy,
4085 struct sched_param *param) 4084 struct sched_param *param)
@@ -4107,28 +4106,32 @@ recheck:
4107 (p->mm && param->sched_priority > MAX_USER_RT_PRIO-1) || 4106 (p->mm && param->sched_priority > MAX_USER_RT_PRIO-1) ||
4108 (!p->mm && param->sched_priority > MAX_RT_PRIO-1)) 4107 (!p->mm && param->sched_priority > MAX_RT_PRIO-1))
4109 return -EINVAL; 4108 return -EINVAL;
4110 if ((policy == SCHED_NORMAL || policy == SCHED_BATCH) 4109 if (is_rt_policy(policy) != (param->sched_priority != 0))
4111 != (param->sched_priority == 0))
4112 return -EINVAL; 4110 return -EINVAL;
4113 4111
4114 /* 4112 /*
4115 * Allow unprivileged RT tasks to decrease priority: 4113 * Allow unprivileged RT tasks to decrease priority:
4116 */ 4114 */
4117 if (!capable(CAP_SYS_NICE)) { 4115 if (!capable(CAP_SYS_NICE)) {
4118 /* 4116 if (is_rt_policy(policy)) {
4119 * can't change policy, except between SCHED_NORMAL 4117 unsigned long rlim_rtprio;
4120 * and SCHED_BATCH: 4118 unsigned long flags;
4121 */ 4119
4122 if (((policy != SCHED_NORMAL && p->policy != SCHED_BATCH) && 4120 if (!lock_task_sighand(p, &flags))
4123 (policy != SCHED_BATCH && p->policy != SCHED_NORMAL)) && 4121 return -ESRCH;
4124 !p->signal->rlim[RLIMIT_RTPRIO].rlim_cur) 4122 rlim_rtprio = p->signal->rlim[RLIMIT_RTPRIO].rlim_cur;
4125 return -EPERM; 4123 unlock_task_sighand(p, &flags);
4126 /* can't increase priority */ 4124
4127 if ((policy != SCHED_NORMAL && policy != SCHED_BATCH) && 4125 /* can't set/change the rt policy */
4128 param->sched_priority > p->rt_priority && 4126 if (policy != p->policy && !rlim_rtprio)
4129 param->sched_priority > 4127 return -EPERM;
4130 p->signal->rlim[RLIMIT_RTPRIO].rlim_cur) 4128
4131 return -EPERM; 4129 /* can't increase priority */
4130 if (param->sched_priority > p->rt_priority &&
4131 param->sched_priority > rlim_rtprio)
4132 return -EPERM;
4133 }
4134
4132 /* can't change other user's priorities */ 4135 /* can't change other user's priorities */
4133 if ((current->euid != p->euid) && 4136 if ((current->euid != p->euid) &&
4134 (current->euid != p->uid)) 4137 (current->euid != p->uid))
@@ -4193,14 +4196,13 @@ do_sched_setscheduler(pid_t pid, int policy, struct sched_param __user *param)
4193 return -EINVAL; 4196 return -EINVAL;
4194 if (copy_from_user(&lparam, param, sizeof(struct sched_param))) 4197 if (copy_from_user(&lparam, param, sizeof(struct sched_param)))
4195 return -EFAULT; 4198 return -EFAULT;
4196 read_lock_irq(&tasklist_lock); 4199
4200 rcu_read_lock();
4201 retval = -ESRCH;
4197 p = find_process_by_pid(pid); 4202 p = find_process_by_pid(pid);
4198 if (!p) { 4203 if (p != NULL)
4199 read_unlock_irq(&tasklist_lock); 4204 retval = sched_setscheduler(p, policy, &lparam);
4200 return -ESRCH; 4205 rcu_read_unlock();
4201 }
4202 retval = sched_setscheduler(p, policy, &lparam);
4203 read_unlock_irq(&tasklist_lock);
4204 4206
4205 return retval; 4207 return retval;
4206} 4208}
@@ -5151,7 +5153,7 @@ static void migrate_dead(unsigned int dead_cpu, struct task_struct *p)
5151 BUG_ON(p->exit_state != EXIT_ZOMBIE && p->exit_state != EXIT_DEAD); 5153 BUG_ON(p->exit_state != EXIT_ZOMBIE && p->exit_state != EXIT_DEAD);
5152 5154
5153 /* Cannot have done final schedule yet: would have vanished. */ 5155 /* Cannot have done final schedule yet: would have vanished. */
5154 BUG_ON(p->flags & PF_DEAD); 5156 BUG_ON(p->state == TASK_DEAD);
5155 5157
5156 get_task_struct(p); 5158 get_task_struct(p);
5157 5159
@@ -5272,9 +5274,11 @@ static struct notifier_block __cpuinitdata migration_notifier = {
5272int __init migration_init(void) 5274int __init migration_init(void)
5273{ 5275{
5274 void *cpu = (void *)(long)smp_processor_id(); 5276 void *cpu = (void *)(long)smp_processor_id();
5277 int err;
5275 5278
5276 /* Start one for the boot CPU: */ 5279 /* Start one for the boot CPU: */
5277 migration_call(&migration_notifier, CPU_UP_PREPARE, cpu); 5280 err = migration_call(&migration_notifier, CPU_UP_PREPARE, cpu);
5281 BUG_ON(err == NOTIFY_BAD);
5278 migration_call(&migration_notifier, CPU_ONLINE, cpu); 5282 migration_call(&migration_notifier, CPU_ONLINE, cpu);
5279 register_cpu_notifier(&migration_notifier); 5283 register_cpu_notifier(&migration_notifier);
5280 5284
diff --git a/kernel/signal.c b/kernel/signal.c
index 05853a7337e3..fb5da6d19f14 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -417,9 +417,8 @@ static int collect_signal(int sig, struct sigpending *list, siginfo_t *info)
417static int __dequeue_signal(struct sigpending *pending, sigset_t *mask, 417static int __dequeue_signal(struct sigpending *pending, sigset_t *mask,
418 siginfo_t *info) 418 siginfo_t *info)
419{ 419{
420 int sig = 0; 420 int sig = next_signal(pending, mask);
421 421
422 sig = next_signal(pending, mask);
423 if (sig) { 422 if (sig) {
424 if (current->notifier) { 423 if (current->notifier) {
425 if (sigismember(current->notifier_mask, sig)) { 424 if (sigismember(current->notifier_mask, sig)) {
@@ -432,9 +431,7 @@ static int __dequeue_signal(struct sigpending *pending, sigset_t *mask,
432 431
433 if (!collect_signal(sig, pending, info)) 432 if (!collect_signal(sig, pending, info))
434 sig = 0; 433 sig = 0;
435
436 } 434 }
437 recalc_sigpending();
438 435
439 return sig; 436 return sig;
440} 437}
@@ -451,6 +448,7 @@ int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
451 if (!signr) 448 if (!signr)
452 signr = __dequeue_signal(&tsk->signal->shared_pending, 449 signr = __dequeue_signal(&tsk->signal->shared_pending,
453 mask, info); 450 mask, info);
451 recalc_sigpending_tsk(tsk);
454 if (signr && unlikely(sig_kernel_stop(signr))) { 452 if (signr && unlikely(sig_kernel_stop(signr))) {
455 /* 453 /*
456 * Set a marker that we have dequeued a stop signal. Our 454 * Set a marker that we have dequeued a stop signal. Our
diff --git a/kernel/softirq.c b/kernel/softirq.c
index 3789ca98197c..bf25015dce16 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -612,7 +612,9 @@ static struct notifier_block __cpuinitdata cpu_nfb = {
612__init int spawn_ksoftirqd(void) 612__init int spawn_ksoftirqd(void)
613{ 613{
614 void *cpu = (void *)(long)smp_processor_id(); 614 void *cpu = (void *)(long)smp_processor_id();
615 cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); 615 int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu);
616
617 BUG_ON(err == NOTIFY_BAD);
616 cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); 618 cpu_callback(&cpu_nfb, CPU_ONLINE, cpu);
617 register_cpu_notifier(&cpu_nfb); 619 register_cpu_notifier(&cpu_nfb);
618 return 0; 620 return 0;
diff --git a/kernel/softlockup.c b/kernel/softlockup.c
index 03e6a2b0b787..50afeb813305 100644
--- a/kernel/softlockup.c
+++ b/kernel/softlockup.c
@@ -149,8 +149,9 @@ static struct notifier_block __cpuinitdata cpu_nfb = {
149__init void spawn_softlockup_task(void) 149__init void spawn_softlockup_task(void)
150{ 150{
151 void *cpu = (void *)(long)smp_processor_id(); 151 void *cpu = (void *)(long)smp_processor_id();
152 int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu);
152 153
153 cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); 154 BUG_ON(err == NOTIFY_BAD);
154 cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); 155 cpu_callback(&cpu_nfb, CPU_ONLINE, cpu);
155 register_cpu_notifier(&cpu_nfb); 156 register_cpu_notifier(&cpu_nfb);
156 157
diff --git a/kernel/spinlock.c b/kernel/spinlock.c
index 9644a41e0bef..d48143eafbfd 100644
--- a/kernel/spinlock.c
+++ b/kernel/spinlock.c
@@ -21,17 +21,6 @@
21#include <linux/debug_locks.h> 21#include <linux/debug_locks.h>
22#include <linux/module.h> 22#include <linux/module.h>
23 23
24/*
25 * Generic declaration of the raw read_trylock() function,
26 * architectures are supposed to optimize this:
27 */
28int __lockfunc generic__raw_read_trylock(raw_rwlock_t *lock)
29{
30 __raw_read_lock(lock);
31 return 1;
32}
33EXPORT_SYMBOL(generic__raw_read_trylock);
34
35int __lockfunc _spin_trylock(spinlock_t *lock) 24int __lockfunc _spin_trylock(spinlock_t *lock)
36{ 25{
37 preempt_disable(); 26 preempt_disable();
diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index 51cacd111dbd..12458040e665 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -1,3 +1,6 @@
1/* Copyright 2005 Rusty Russell rusty@rustcorp.com.au IBM Corporation.
2 * GPL v2 and any later version.
3 */
1#include <linux/stop_machine.h> 4#include <linux/stop_machine.h>
2#include <linux/kthread.h> 5#include <linux/kthread.h>
3#include <linux/sched.h> 6#include <linux/sched.h>
diff --git a/kernel/sys.c b/kernel/sys.c
index 3f894775488d..b88806c66244 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -612,7 +612,6 @@ void kernel_restart(char *cmd)
612 } else { 612 } else {
613 printk(KERN_EMERG "Restarting system with command '%s'.\n", cmd); 613 printk(KERN_EMERG "Restarting system with command '%s'.\n", cmd);
614 } 614 }
615 printk(".\n");
616 machine_restart(cmd); 615 machine_restart(cmd);
617} 616}
618EXPORT_SYMBOL_GPL(kernel_restart); 617EXPORT_SYMBOL_GPL(kernel_restart);
@@ -2084,12 +2083,12 @@ asmlinkage long sys_getcpu(unsigned __user *cpup, unsigned __user *nodep,
2084 * padding 2083 * padding
2085 */ 2084 */
2086 unsigned long t0, t1; 2085 unsigned long t0, t1;
2087 get_user(t0, &cache->t0); 2086 get_user(t0, &cache->blob[0]);
2088 get_user(t1, &cache->t1); 2087 get_user(t1, &cache->blob[1]);
2089 t0++; 2088 t0++;
2090 t1++; 2089 t1++;
2091 put_user(t0, &cache->t0); 2090 put_user(t0, &cache->blob[0]);
2092 put_user(t1, &cache->t1); 2091 put_user(t1, &cache->blob[1]);
2093 } 2092 }
2094 return err ? -EFAULT : 0; 2093 return err ? -EFAULT : 0;
2095} 2094}
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index 6991bece67e8..7a3b2e75f040 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -134,3 +134,8 @@ cond_syscall(sys_madvise);
134cond_syscall(sys_mremap); 134cond_syscall(sys_mremap);
135cond_syscall(sys_remap_file_pages); 135cond_syscall(sys_remap_file_pages);
136cond_syscall(compat_sys_move_pages); 136cond_syscall(compat_sys_move_pages);
137
138/* block-layer dependent */
139cond_syscall(sys_bdflush);
140cond_syscall(sys_ioprio_set);
141cond_syscall(sys_ioprio_get);
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 8bfa7d117c54..c57c4532e296 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -52,6 +52,10 @@
52extern int proc_nr_files(ctl_table *table, int write, struct file *filp, 52extern int proc_nr_files(ctl_table *table, int write, struct file *filp,
53 void __user *buffer, size_t *lenp, loff_t *ppos); 53 void __user *buffer, size_t *lenp, loff_t *ppos);
54 54
55#ifdef CONFIG_X86
56#include <asm/nmi.h>
57#endif
58
55#if defined(CONFIG_SYSCTL) 59#if defined(CONFIG_SYSCTL)
56 60
57/* External variables not in a header file. */ 61/* External variables not in a header file. */
@@ -74,13 +78,6 @@ extern int sysctl_drop_caches;
74extern int percpu_pagelist_fraction; 78extern int percpu_pagelist_fraction;
75extern int compat_log; 79extern int compat_log;
76 80
77#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
78int unknown_nmi_panic;
79int nmi_watchdog_enabled;
80extern int proc_nmi_enabled(struct ctl_table *, int , struct file *,
81 void __user *, size_t *, loff_t *);
82#endif
83
84/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */ 81/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */
85static int maxolduid = 65535; 82static int maxolduid = 65535;
86static int minolduid; 83static int minolduid;
@@ -1915,7 +1912,7 @@ int proc_dointvec_bset(ctl_table *table, int write, struct file *filp,
1915 return -EPERM; 1912 return -EPERM;
1916 } 1913 }
1917 1914
1918 op = (current->pid == 1) ? OP_SET : OP_AND; 1915 op = is_init(current) ? OP_SET : OP_AND;
1919 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos, 1916 return do_proc_dointvec(table,write,filp,buffer,lenp,ppos,
1920 do_proc_dointvec_bset_conv,&op); 1917 do_proc_dointvec_bset_conv,&op);
1921} 1918}
diff --git a/kernel/timer.c b/kernel/timer.c
index 1d7dd6267c2d..4f55622b0d38 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c
@@ -136,7 +136,7 @@ static void internal_add_timer(tvec_base_t *base, struct timer_list *timer)
136 list_add_tail(&timer->entry, vec); 136 list_add_tail(&timer->entry, vec);
137} 137}
138 138
139/*** 139/**
140 * init_timer - initialize a timer. 140 * init_timer - initialize a timer.
141 * @timer: the timer to be initialized 141 * @timer: the timer to be initialized
142 * 142 *
@@ -175,6 +175,7 @@ static inline void detach_timer(struct timer_list *timer,
175 */ 175 */
176static tvec_base_t *lock_timer_base(struct timer_list *timer, 176static tvec_base_t *lock_timer_base(struct timer_list *timer,
177 unsigned long *flags) 177 unsigned long *flags)
178 __acquires(timer->base->lock)
178{ 179{
179 tvec_base_t *base; 180 tvec_base_t *base;
180 181
@@ -235,7 +236,7 @@ int __mod_timer(struct timer_list *timer, unsigned long expires)
235 236
236EXPORT_SYMBOL(__mod_timer); 237EXPORT_SYMBOL(__mod_timer);
237 238
238/*** 239/**
239 * add_timer_on - start a timer on a particular CPU 240 * add_timer_on - start a timer on a particular CPU
240 * @timer: the timer to be added 241 * @timer: the timer to be added
241 * @cpu: the CPU to start it on 242 * @cpu: the CPU to start it on
@@ -255,9 +256,10 @@ void add_timer_on(struct timer_list *timer, int cpu)
255} 256}
256 257
257 258
258/*** 259/**
259 * mod_timer - modify a timer's timeout 260 * mod_timer - modify a timer's timeout
260 * @timer: the timer to be modified 261 * @timer: the timer to be modified
262 * @expires: new timeout in jiffies
261 * 263 *
262 * mod_timer is a more efficient way to update the expire field of an 264 * mod_timer is a more efficient way to update the expire field of an
263 * active timer (if the timer is inactive it will be activated) 265 * active timer (if the timer is inactive it will be activated)
@@ -291,7 +293,7 @@ int mod_timer(struct timer_list *timer, unsigned long expires)
291 293
292EXPORT_SYMBOL(mod_timer); 294EXPORT_SYMBOL(mod_timer);
293 295
294/*** 296/**
295 * del_timer - deactive a timer. 297 * del_timer - deactive a timer.
296 * @timer: the timer to be deactivated 298 * @timer: the timer to be deactivated
297 * 299 *
@@ -323,7 +325,10 @@ int del_timer(struct timer_list *timer)
323EXPORT_SYMBOL(del_timer); 325EXPORT_SYMBOL(del_timer);
324 326
325#ifdef CONFIG_SMP 327#ifdef CONFIG_SMP
326/* 328/**
329 * try_to_del_timer_sync - Try to deactivate a timer
330 * @timer: timer do del
331 *
327 * This function tries to deactivate a timer. Upon successful (ret >= 0) 332 * This function tries to deactivate a timer. Upon successful (ret >= 0)
328 * exit the timer is not queued and the handler is not running on any CPU. 333 * exit the timer is not queued and the handler is not running on any CPU.
329 * 334 *
@@ -351,7 +356,7 @@ out:
351 return ret; 356 return ret;
352} 357}
353 358
354/*** 359/**
355 * del_timer_sync - deactivate a timer and wait for the handler to finish. 360 * del_timer_sync - deactivate a timer and wait for the handler to finish.
356 * @timer: the timer to be deactivated 361 * @timer: the timer to be deactivated
357 * 362 *
@@ -401,15 +406,15 @@ static int cascade(tvec_base_t *base, tvec_t *tv, int index)
401 return index; 406 return index;
402} 407}
403 408
404/*** 409#define INDEX(N) ((base->timer_jiffies >> (TVR_BITS + (N) * TVN_BITS)) & TVN_MASK)
410
411/**
405 * __run_timers - run all expired timers (if any) on this CPU. 412 * __run_timers - run all expired timers (if any) on this CPU.
406 * @base: the timer vector to be processed. 413 * @base: the timer vector to be processed.
407 * 414 *
408 * This function cascades all vectors and executes all expired timer 415 * This function cascades all vectors and executes all expired timer
409 * vectors. 416 * vectors.
410 */ 417 */
411#define INDEX(N) ((base->timer_jiffies >> (TVR_BITS + (N) * TVN_BITS)) & TVN_MASK)
412
413static inline void __run_timers(tvec_base_t *base) 418static inline void __run_timers(tvec_base_t *base)
414{ 419{
415 struct timer_list *timer; 420 struct timer_list *timer;
@@ -970,7 +975,7 @@ void __init timekeeping_init(void)
970 975
971 976
972static int timekeeping_suspended; 977static int timekeeping_suspended;
973/* 978/**
974 * timekeeping_resume - Resumes the generic timekeeping subsystem. 979 * timekeeping_resume - Resumes the generic timekeeping subsystem.
975 * @dev: unused 980 * @dev: unused
976 * 981 *
@@ -1106,7 +1111,7 @@ static void clocksource_adjust(struct clocksource *clock, s64 offset)
1106 clock->error -= (interval - offset) << (TICK_LENGTH_SHIFT - clock->shift); 1111 clock->error -= (interval - offset) << (TICK_LENGTH_SHIFT - clock->shift);
1107} 1112}
1108 1113
1109/* 1114/**
1110 * update_wall_time - Uses the current clocksource to increment the wall time 1115 * update_wall_time - Uses the current clocksource to increment the wall time
1111 * 1116 *
1112 * Called from the timer interrupt, must hold a write on xtime_lock. 1117 * Called from the timer interrupt, must hold a write on xtime_lock.
@@ -1217,10 +1222,8 @@ static inline void calc_load(unsigned long ticks)
1217 unsigned long active_tasks; /* fixed-point */ 1222 unsigned long active_tasks; /* fixed-point */
1218 static int count = LOAD_FREQ; 1223 static int count = LOAD_FREQ;
1219 1224
1220 count -= ticks; 1225 active_tasks = count_active_tasks();
1221 if (count < 0) { 1226 for (count -= ticks; count < 0; count += LOAD_FREQ) {
1222 count += LOAD_FREQ;
1223 active_tasks = count_active_tasks();
1224 CALC_LOAD(avenrun[0], EXP_1, active_tasks); 1227 CALC_LOAD(avenrun[0], EXP_1, active_tasks);
1225 CALC_LOAD(avenrun[1], EXP_5, active_tasks); 1228 CALC_LOAD(avenrun[1], EXP_5, active_tasks);
1226 CALC_LOAD(avenrun[2], EXP_15, active_tasks); 1229 CALC_LOAD(avenrun[2], EXP_15, active_tasks);
@@ -1265,11 +1268,8 @@ void run_local_timers(void)
1265 * Called by the timer interrupt. xtime_lock must already be taken 1268 * Called by the timer interrupt. xtime_lock must already be taken
1266 * by the timer IRQ! 1269 * by the timer IRQ!
1267 */ 1270 */
1268static inline void update_times(void) 1271static inline void update_times(unsigned long ticks)
1269{ 1272{
1270 unsigned long ticks;
1271
1272 ticks = jiffies - wall_jiffies;
1273 wall_jiffies += ticks; 1273 wall_jiffies += ticks;
1274 update_wall_time(); 1274 update_wall_time();
1275 calc_load(ticks); 1275 calc_load(ticks);
@@ -1281,12 +1281,10 @@ static inline void update_times(void)
1281 * jiffies is defined in the linker script... 1281 * jiffies is defined in the linker script...
1282 */ 1282 */
1283 1283
1284void do_timer(struct pt_regs *regs) 1284void do_timer(unsigned long ticks)
1285{ 1285{
1286 jiffies_64++; 1286 jiffies_64 += ticks;
1287 /* prevent loading jiffies before storing new jiffies_64 value. */ 1287 update_times(ticks);
1288 barrier();
1289 update_times();
1290} 1288}
1291 1289
1292#ifdef __ARCH_WANT_SYS_ALARM 1290#ifdef __ARCH_WANT_SYS_ALARM
@@ -1470,8 +1468,9 @@ asmlinkage long sys_gettid(void)
1470 return current->pid; 1468 return current->pid;
1471} 1469}
1472 1470
1473/* 1471/**
1474 * sys_sysinfo - fill in sysinfo struct 1472 * sys_sysinfo - fill in sysinfo struct
1473 * @info: pointer to buffer to fill
1475 */ 1474 */
1476asmlinkage long sys_sysinfo(struct sysinfo __user *info) 1475asmlinkage long sys_sysinfo(struct sysinfo __user *info)
1477{ 1476{
@@ -1688,8 +1687,10 @@ static struct notifier_block __cpuinitdata timers_nb = {
1688 1687
1689void __init init_timers(void) 1688void __init init_timers(void)
1690{ 1689{
1691 timer_cpu_notify(&timers_nb, (unsigned long)CPU_UP_PREPARE, 1690 int err = timer_cpu_notify(&timers_nb, (unsigned long)CPU_UP_PREPARE,
1692 (void *)(long)smp_processor_id()); 1691 (void *)(long)smp_processor_id());
1692
1693 BUG_ON(err == NOTIFY_BAD);
1693 register_cpu_notifier(&timers_nb); 1694 register_cpu_notifier(&timers_nb);
1694 open_softirq(TIMER_SOFTIRQ, run_timer_softirq, NULL); 1695 open_softirq(TIMER_SOFTIRQ, run_timer_softirq, NULL);
1695} 1696}
diff --git a/kernel/unwind.c b/kernel/unwind.c
index 3430475fcd88..2e2368607aab 100644
--- a/kernel/unwind.c
+++ b/kernel/unwind.c
@@ -102,7 +102,7 @@ static struct unwind_table {
102 unsigned long size; 102 unsigned long size;
103 struct unwind_table *link; 103 struct unwind_table *link;
104 const char *name; 104 const char *name;
105} root_table, *last_table; 105} root_table;
106 106
107struct unwind_item { 107struct unwind_item {
108 enum item_location { 108 enum item_location {
@@ -174,6 +174,8 @@ void __init unwind_init(void)
174 174
175#ifdef CONFIG_MODULES 175#ifdef CONFIG_MODULES
176 176
177static struct unwind_table *last_table;
178
177/* Must be called with module_mutex held. */ 179/* Must be called with module_mutex held. */
178void *unwind_add_table(struct module *module, 180void *unwind_add_table(struct module *module,
179 const void *table_start, 181 const void *table_start,
generated by cgit v1.2.2 (git 2.25.0) at 2025-02-08 02:06:37 -0500