From aa5cb02143123289bd37c30c0ad60339f8da0bad Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Wed, 1 Mar 2006 15:07:07 +1100 Subject: [PATCH] powerpc: Expose SMT and L1 icache snoop userland features This patch makes userland aware of the icache snoop capability of the POWER5 (and possibly others in the future) and of SMT capabilities. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- include/asm-powerpc/cputable.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/asm-powerpc/cputable.h b/include/asm-powerpc/cputable.h index 90d005bb4d..5638518968 100644 --- a/include/asm-powerpc/cputable.h +++ b/include/asm-powerpc/cputable.h @@ -20,6 +20,8 @@ #define PPC_FEATURE_POWER5_PLUS 0x00020000 #define PPC_FEATURE_CELL 0x00010000 #define PPC_FEATURE_BOOKE 0x00008000 +#define PPC_FEATURE_SMT 0x00004000 +#define PPC_FEATURE_ICACHE_SNOOP 0x00002000 #ifdef __KERNEL__ #ifndef __ASSEMBLY__ -- cgit v1.2.2 From 141aa59b5347a4a021e37cfbc2258df9af9392f8 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Fri, 3 Mar 2006 16:24:06 +1100 Subject: [PATCH] powerpc: Fix incorrect pud_ERROR() message The powerpc pud_ERROR() function misleadingly prints a message indicating a pmd error. This patch fixes it. Signed-off-by: David Gibson Signed-off-by: Paul Mackerras --- include/asm-powerpc/pgtable-4k.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/asm-powerpc/pgtable-4k.h b/include/asm-powerpc/pgtable-4k.h index e9590c06ad..80a7832d27 100644 --- a/include/asm-powerpc/pgtable-4k.h +++ b/include/asm-powerpc/pgtable-4k.h @@ -88,4 +88,4 @@ (((addr) >> PUD_SHIFT) & (PTRS_PER_PUD - 1))) #define pud_ERROR(e) \ - printk("%s:%d: bad pmd %08lx.\n", __FILE__, __LINE__, pud_val(e)) + printk("%s:%d: bad pud %08lx.\n", __FILE__, __LINE__, pud_val(e)) -- cgit v1.2.2 From 4d000d5b9689734006d89fe9b7597c758b74a9fb Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 4 Mar 2006 23:23:56 -0800 Subject: [SPARC64]: Mark __ex_table section correctly. We must use the "a" (allocate) attribute every time we emit an entry into the __ex_table section. For consistency, use "a" instead of #alloc which is some Solaris compat cruft GNU as provides on Sparc. Signed-off-by: David S. Miller --- include/asm-sparc64/futex.h | 2 +- include/asm-sparc64/uaccess.h | 12 ++++++------ 2 files changed, 7 insertions(+), 7 deletions(-) (limited to 'include') diff --git a/include/asm-sparc64/futex.h b/include/asm-sparc64/futex.h index 0caf60147e..34c4b43d3f 100644 --- a/include/asm-sparc64/futex.h +++ b/include/asm-sparc64/futex.h @@ -20,7 +20,7 @@ "4: ba 3b\n" \ " mov %5, %0\n" \ " .previous\n" \ - " .section __ex_table,#alloc\n" \ + " .section __ex_table,\"a\"\n" \ " .align 4\n" \ " .word 1b, 4b\n" \ " .word 2b, 4b\n" \ diff --git a/include/asm-sparc64/uaccess.h b/include/asm-sparc64/uaccess.h index 203e8eee63..c91d1e38ea 100644 --- a/include/asm-sparc64/uaccess.h +++ b/include/asm-sparc64/uaccess.h @@ -136,7 +136,7 @@ __asm__ __volatile__( \ "b 2b\n\t" \ " mov %3, %0\n\n\t" \ ".previous\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b, 3b\n\t" \ ".previous\n\n\t" \ @@ -148,7 +148,7 @@ if (__builtin_constant_p(ret) && ret == -EFAULT) \ __asm__ __volatile__( \ "/* Put user asm ret, inline. */\n" \ "1:\t" "st"#size "a %1, [%2] %%asi\n\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b, __ret_efault\n\n\t" \ ".previous\n\n\t" \ @@ -163,7 +163,7 @@ __asm__ __volatile__( \ "ret\n\t" \ " restore %%g0, %3, %%o0\n\n\t" \ ".previous\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b, 3b\n\n\t" \ ".previous\n\n\t" \ @@ -206,7 +206,7 @@ __asm__ __volatile__( \ "b 2b\n\t" \ " mov %3, %0\n\n\t" \ ".previous\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b, 3b\n\n\t" \ ".previous\n\t" \ @@ -218,7 +218,7 @@ if (__builtin_constant_p(retval) && retval == -EFAULT) \ __asm__ __volatile__( \ "/* Get user asm ret, inline. */\n" \ "1:\t" "ld"#size "a [%1] %%asi, %0\n\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b,__ret_efault\n\n\t" \ ".previous\n\t" \ @@ -233,7 +233,7 @@ __asm__ __volatile__( \ "ret\n\t" \ " restore %%g0, %2, %%o0\n\n\t" \ ".previous\n\t" \ - ".section __ex_table,#alloc\n\t" \ + ".section __ex_table,\"a\"\n\t" \ ".align 4\n\t" \ ".word 1b, 3b\n\n\t" \ ".previous\n\t" \ -- cgit v1.2.2 From 1e4b27df55166ce3b276f55bab223fa4ae8c5525 Mon Sep 17 00:00:00 2001 From: Karsten Keil Date: Mon, 6 Mar 2006 15:42:37 -0800 Subject: [PATCH] i4l: add new PCI IDs for HFC-S PCI Add new PCI IDs for HFC-S PCI based ISDN TA 'Primux II S0' and 'Primux II S0' from Gerdes AG Signed-off-by: Martin Bachem Signed-off-by: Karsten Keil Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/pci_ids.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 82b83da25d..1709b5009d 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1752,6 +1752,8 @@ #define PCI_DEVICE_ID_CCD_B00B 0xb00b #define PCI_DEVICE_ID_CCD_B00C 0xb00c #define PCI_DEVICE_ID_CCD_B100 0xb100 +#define PCI_DEVICE_ID_CCD_B700 0xb700 +#define PCI_DEVICE_ID_CCD_B701 0xb701 #define PCI_VENDOR_ID_EXAR 0x13a8 #define PCI_DEVICE_ID_EXAR_XR17C152 0x0152 -- cgit v1.2.2 From 69239749e1ac4f3496906aa4267cb9f61ce52c9c Mon Sep 17 00:00:00 2001 From: Tony Lindgren Date: Mon, 6 Mar 2006 15:42:45 -0800 Subject: [PATCH] fix next_timer_interrupt() for hrtimer Also from Thomas Gleixner Function next_timer_interrupt() got broken with a recent patch 6ba1b91213e81aa92b5cf7539f7d2a94ff54947c as sys_nanosleep() was moved to hrtimer. This broke things as next_timer_interrupt() did not check hrtimer tree for next event. Function next_timer_interrupt() is needed with dyntick (CONFIG_NO_IDLE_HZ, VST) implementations, as the system can be in idle when next hrtimer event was supposed to happen. At least ARM and S390 currently use next_timer_interrupt(). Signed-off-by: Thomas Gleixner Cc: Martin Schwidefsky Cc: Heiko Carstens Cc: Russell King Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/hrtimer.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'include') diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h index 6361544bb6..6401c31d6a 100644 --- a/include/linux/hrtimer.h +++ b/include/linux/hrtimer.h @@ -116,6 +116,10 @@ extern int hrtimer_try_to_cancel(struct hrtimer *timer); extern ktime_t hrtimer_get_remaining(const struct hrtimer *timer); extern int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp); +#ifdef CONFIG_NO_IDLE_HZ +extern ktime_t hrtimer_get_next_event(void); +#endif + static inline int hrtimer_active(const struct hrtimer *timer) { return timer->state == HRTIMER_PENDING; -- cgit v1.2.2 From 5c8338904653365bfb92385b38915becb903d8bb Mon Sep 17 00:00:00 2001 From: Jan Blunck Date: Mon, 6 Mar 2006 15:42:46 -0800 Subject: [PATCH] s390: fix compile with VIRT_CPU_ACCOUNTING=n When CONFIG_VIRT_CPU_ACCOUNTING is not defined compiling fails with an undefined reference to account_vtime(). Signed-off-by: Jan Blunck Signed-off-by: Martin Schwidefsky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-s390/system.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/asm-s390/system.h b/include/asm-s390/system.h index b2e65e8bf8..6a89dbb03c 100644 --- a/include/asm-s390/system.h +++ b/include/asm-s390/system.h @@ -118,6 +118,8 @@ static inline void sched_cacheflush(void) extern void account_vtime(struct task_struct *); extern void account_tick_vtime(struct task_struct *); extern void account_system_vtime(struct task_struct *); +#else +#define account_vtime(x) do { /* empty */ } while (0) #endif #define finish_arch_switch(prev) do { \ -- cgit v1.2.2 From 78679302fe428f4f3dc853a51ee24f306010d874 Mon Sep 17 00:00:00 2001 From: KAMEZAWA Hiroyuki Date: Mon, 6 Mar 2006 15:42:49 -0800 Subject: [PATCH] memory-hotplug compile fix include/linux/memory_hotplug.h:53: warning: 'struct page' declared inside parameter list (akpm: I tossed in a couple more possibly-needed-sometime struct decls too) Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/memory_hotplug.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'include') diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h index 01f03bc06e..968b1aa373 100644 --- a/include/linux/memory_hotplug.h +++ b/include/linux/memory_hotplug.h @@ -6,6 +6,10 @@ #include #include +struct page; +struct zone; +struct pglist_data; + #ifdef CONFIG_MEMORY_HOTPLUG /* * pgdat resizing functions -- cgit v1.2.2 From a615fa83959896f8eac76c235953fb164cd1a9b9 Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Mon, 6 Mar 2006 15:42:50 -0800 Subject: [PATCH] Increase max kmalloc size for very large systems Systems with extemely large numbers of nodes or cpus need to kmalloc structures larger than is currently supported. This patch increases the maximum supported size for very large systems. This patch should have no effect on current systems. (akpm: why not just use alloc_pages() for sysfs_cpus?) Signed-off-by: Jack Steiner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/kmalloc_sizes.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/kmalloc_sizes.h b/include/linux/kmalloc_sizes.h index d82d4c05c1..bda23e00ed 100644 --- a/include/linux/kmalloc_sizes.h +++ b/include/linux/kmalloc_sizes.h @@ -19,8 +19,10 @@ CACHE(32768) CACHE(65536) CACHE(131072) -#ifndef CONFIG_MMU +#if (NR_CPUS > 512) || (MAX_NUMNODES > 256) || !defined(CONFIG_MMU) CACHE(262144) +#endif +#ifndef CONFIG_MMU CACHE(524288) CACHE(1048576) #ifdef CONFIG_LARGE_ALLOCS -- cgit v1.2.2 From 6a0e243069b09a323255f6e847c87d531961cd96 Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Tue, 7 Mar 2006 14:42:27 +0000 Subject: [ARM] 3352/1: DSB required for the completion of a TLB maintenance operation Patch from Catalin Marinas Chapter B2.7.3 in the latest ARM ARM (with v6 information) states that the completion of a TLB maintenance operation is only guaranteed by the execution of a DSB (Data Syncronization Barrier, formerly Data Write Barrier or Drain Write Buffer). Note that a DSB is only needed in the flush_tlb_kernel_* functions since the completion is guaranteed by a mode change (i.e. switching back to user mode) for the flush_tlb_user_* functions. Signed-off-by: Catalin Marinas Signed-off-by: Russell King --- include/asm-arm/tlbflush.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include') diff --git a/include/asm-arm/tlbflush.h b/include/asm-arm/tlbflush.h index 9387a5e1ff..0c2acc944a 100644 --- a/include/asm-arm/tlbflush.h +++ b/include/asm-arm/tlbflush.h @@ -340,6 +340,12 @@ static inline void local_flush_tlb_kernel_page(unsigned long kaddr) asm("mcr%? p15, 0, %0, c8, c6, 1" : : "r" (kaddr)); if (tlb_flag(TLB_V6_I_PAGE)) asm("mcr%? p15, 0, %0, c8, c5, 1" : : "r" (kaddr)); + + /* The ARM ARM states that the completion of a TLB maintenance + * operation is only guaranteed by a DSB instruction + */ + if (tlb_flag(TLB_V6_U_PAGE | TLB_V6_D_PAGE | TLB_V6_I_PAGE)) + asm("mcr%? p15, 0, %0, c7, c10, 4" : : "r" (zero)); } /* -- cgit v1.2.2 From 1bd79336a426c5e4f3bab142407059ceb12cadf9 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Wed, 8 Mar 2006 13:24:22 +1100 Subject: powerpc: Fix various syscall/signal/swapcontext bugs A careful reading of the recent changes to the system call entry/exit paths revealed several problems, plus some things that could be simplified and improved: * 32-bit wasn't testing the _TIF_NOERROR bit in the syscall fast exit path, so it was only doing anything with it once it saw some other bit being set. In other words, the noerror behaviour would apply to the next system call where we had to reschedule or deliver a signal, which is not necessarily the current system call. * 32-bit wasn't doing the call to ptrace_notify in the syscall exit path when the _TIF_SINGLESTEP bit was set. * _TIF_RESTOREALL was in both _TIF_USER_WORK_MASK and _TIF_PERSYSCALL_MASK, which is odd since _TIF_RESTOREALL is only set by system calls. I took it out of _TIF_USER_WORK_MASK. * On 64-bit, _TIF_RESTOREALL wasn't causing the non-volatile registers to be restored (unless perhaps a signal was delivered or the syscall was traced or single-stepped). Thus the non-volatile registers weren't restored on exit from a signal handler. We probably got away with it mostly because signal handlers written in C wouldn't alter the non-volatile registers. * On 32-bit I simplified the code and made it more like 64-bit by making the syscall exit path jump to ret_from_except to handle preemption and signal delivery. * 32-bit was calling do_signal unnecessarily when _TIF_RESTOREALL was set - but I think because of that 32-bit was actually restoring the non-volatile registers on exit from a signal handler. * I changed the order of enabling interrupts and saving the non-volatile registers before calling do_syscall_trace_leave; now we enable interrupts first. Signed-off-by: Paul Mackerras --- include/asm-powerpc/thread_info.h | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h index 237fc2b729..ffc7462d77 100644 --- a/include/asm-powerpc/thread_info.h +++ b/include/asm-powerpc/thread_info.h @@ -37,7 +37,6 @@ struct thread_info { int preempt_count; /* 0 => preemptable, <0 => BUG */ struct restart_block restart_block; - void __user *nvgprs_frame; /* low level flags - has atomic operations done on it */ unsigned long flags ____cacheline_aligned_in_smp; }; @@ -120,7 +119,6 @@ static inline struct thread_info *current_thread_info(void) #define TIF_MEMDIE 10 #define TIF_SECCOMP 11 /* secure computing */ #define TIF_RESTOREALL 12 /* Restore all regs (implies NOERROR) */ -#define TIF_SAVE_NVGPRS 13 /* Save r14-r31 in signal frame */ #define TIF_NOERROR 14 /* Force successful syscall return */ #define TIF_RESTORE_SIGMASK 15 /* Restore signal mask in do_signal */ @@ -137,15 +135,13 @@ static inline struct thread_info *current_thread_info(void) #define _TIF_SINGLESTEP (1< Date: Wed, 8 Mar 2006 14:03:09 -0800 Subject: Mark the pipe file operations static They aren't used (nor even really usable) outside of pipe.c anyway Signed-off-by: Linus Torvalds --- include/linux/fs.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'include') diff --git a/include/linux/fs.h b/include/linux/fs.h index e059da9470..0cc34b1c42 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1418,9 +1418,6 @@ extern int is_bad_inode(struct inode *); extern struct file_operations read_fifo_fops; extern struct file_operations write_fifo_fops; extern struct file_operations rdwr_fifo_fops; -extern struct file_operations read_pipe_fops; -extern struct file_operations write_pipe_fops; -extern struct file_operations rdwr_pipe_fops; extern int fs_may_remount_ro(struct super_block *); -- cgit v1.2.2 From 1c6cc5fd32978ffdc4d0acf8592d3901adefbdad Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 7 Mar 2006 21:55:20 -0800 Subject: [PATCH] powerpc: restore eeh_add_device_late() prototype stub We fixed this: arch/powerpc/platforms/pseries/eeh.c: In function `eeh_add_device_tree_late': arch/powerpc/platforms/pseries/eeh.c:901: warning: implicit declaration of function `eeh_add_device_late' arch/powerpc/platforms/pseries/eeh.c: At top level: arch/powerpc/platforms/pseries/eeh.c:918: error: conflicting types for 'eeh_add_device_late' arch/powerpc/platforms/pseries/eeh.c:901: error: previous implicit declaration of 'eeh_add_device_late' was here make[2]: *** [arch/powerpc/platforms/pseries/eeh.o] Error 1 But we forgot the !CONFIG_EEH stub. Signed-off-by: Mark Fasheh Cc: Paul Mackerras Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-powerpc/eeh.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/asm-powerpc/eeh.h b/include/asm-powerpc/eeh.h index eb392032e1..5207758a6d 100644 --- a/include/asm-powerpc/eeh.h +++ b/include/asm-powerpc/eeh.h @@ -118,6 +118,8 @@ static inline void pci_addr_cache_build(void) { } static inline void eeh_add_device_early(struct device_node *dn) { } +static inline void eeh_add_device_late(struct pci_dev *dev) { } + static inline void eeh_remove_device(struct pci_dev *dev) { } static inline void eeh_add_device_tree_early(struct device_node *dn) { } -- cgit v1.2.2 From 707ced0d718e89b52b13aa55a64653083e792cca Mon Sep 17 00:00:00 2001 From: Atsushi Nemoto Date: Tue, 7 Mar 2006 21:55:28 -0800 Subject: [PATCH] __get_unaligned() gcc-4 fix If the 'ptr' is a const, this code cause "assignment of read-only variable" error on gcc 4.x. Use __u64 instead of __typeof__(*(ptr)) for temporary variable to get rid of errors on gcc 4.x. Signed-off-by: Atsushi Nemoto Cc: Ralf Baechle Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-generic/unaligned.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/asm-generic/unaligned.h b/include/asm-generic/unaligned.h index 4dc8ddb401..09ec447fe2 100644 --- a/include/asm-generic/unaligned.h +++ b/include/asm-generic/unaligned.h @@ -78,7 +78,7 @@ static inline void __ustw(__u16 val, __u16 *addr) #define __get_unaligned(ptr, size) ({ \ const void *__gu_p = ptr; \ - __typeof__(*(ptr)) val; \ + __u64 val; \ switch (size) { \ case 1: \ val = *(const __u8 *)__gu_p; \ @@ -95,7 +95,7 @@ static inline void __ustw(__u16 val, __u16 *addr) default: \ bad_unaligned_access_length(); \ }; \ - val; \ + (__typeof__(*(ptr)))val; \ }) #define __put_unaligned(val, ptr, size) \ -- cgit v1.2.2 From e2bab3d92486fb781f4d06f56339264ed1492392 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 7 Mar 2006 21:55:31 -0800 Subject: [PATCH] percpu_counter_sum() Implement percpu_counter_sum(). This is a more accurate but slower version of percpu_counter_read_positive(). We need this for Alex's speedup-ext3_statfs patch and for the nr_file accounting fix. Otherwise these things would be too inaccurate on large CPU counts. Cc: Ravikiran G Thirumalai Cc: Alex Tomas Cc: "David S. Miller" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/percpu_counter.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include') diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h index bd6708e2c0..682525511c 100644 --- a/include/linux/percpu_counter.h +++ b/include/linux/percpu_counter.h @@ -39,6 +39,7 @@ static inline void percpu_counter_destroy(struct percpu_counter *fbc) } void percpu_counter_mod(struct percpu_counter *fbc, long amount); +long percpu_counter_sum(struct percpu_counter *fbc); static inline long percpu_counter_read(struct percpu_counter *fbc) { @@ -92,6 +93,11 @@ static inline long percpu_counter_read_positive(struct percpu_counter *fbc) return fbc->count; } +static inline long percpu_counter_sum(struct percpu_counter *fbc) +{ + return percpu_counter_read_positive(fbc); +} + #endif /* CONFIG_SMP */ static inline void percpu_counter_inc(struct percpu_counter *fbc) -- cgit v1.2.2 From 21a1ea9eb40411d4ee29448c53b9e4c0654d6ceb Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Tue, 7 Mar 2006 21:55:33 -0800 Subject: [PATCH] rcu batch tuning This patch adds new tunables for RCU queue and finished batches. There are two types of controls - number of completed RCU updates invoked in a batch (blimit) and monitoring for high rate of incoming RCUs on a cpu (qhimark, qlowmark). By default, the per-cpu batch limit is set to a small value. If the input RCU rate exceeds the high watermark, we do two things - force quiescent state on all cpus and set the batch limit of the CPU to INTMAX. Setting batch limit to INTMAX forces all finished RCUs to be processed in one shot. If we have more than INTMAX RCUs queued up, then we have bigger problems anyway. Once the incoming queued RCUs fall below the low watermark, the batch limit is set to the default. Signed-off-by: Dipankar Sarma Cc: "Paul E. McKenney" Cc: "David S. Miller" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/rcupdate.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index b87aefa082..c2ec6c7787 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -98,13 +98,17 @@ struct rcu_data { long batch; /* Batch # for current RCU batch */ struct rcu_head *nxtlist; struct rcu_head **nxttail; - long count; /* # of queued items */ + long qlen; /* # of queued callbacks */ struct rcu_head *curlist; struct rcu_head **curtail; struct rcu_head *donelist; struct rcu_head **donetail; + long blimit; /* Upper limit on a processed batch */ int cpu; struct rcu_head barrier; +#ifdef CONFIG_SMP + long last_rs_qlen; /* qlen during the last resched */ +#endif }; DECLARE_PER_CPU(struct rcu_data, rcu_data); -- cgit v1.2.2 From 529bf6be5c04f2e869d07bfdb122e9fd98ade714 Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Tue, 7 Mar 2006 21:55:35 -0800 Subject: [PATCH] fix file counting I have benchmarked this on an x86_64 NUMA system and see no significant performance difference on kernbench. Tested on both x86_64 and powerpc. The way we do file struct accounting is not very suitable for batched freeing. For scalability reasons, file accounting was constructor/destructor based. This meant that nr_files was decremented only when the object was removed from the slab cache. This is susceptible to slab fragmentation. With RCU based file structure, consequent batched freeing and a test program like Serge's, we just speed this up and end up with a very fragmented slab - llm22:~ # cat /proc/sys/fs/file-nr 587730 0 758844 At the same time, I see only a 2000+ objects in filp cache. The following patch I fixes this problem. This patch changes the file counting by removing the filp_count_lock. Instead we use a separate percpu counter, nr_files, for now and all accesses to it are through get_nr_files() api. In the sysctl handler for nr_files, we populate files_stat.nr_files before returning to user. Counting files as an when they are created and destroyed (as opposed to inside slab) allows us to correctly count open files with RCU. Signed-off-by: Dipankar Sarma Cc: "Paul E. McKenney" Cc: "David S. Miller" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/file.h | 2 -- include/linux/fs.h | 1 + 2 files changed, 1 insertion(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/file.h b/include/linux/file.h index 418b6101b5..9901b850f2 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -60,8 +60,6 @@ extern void put_filp(struct file *); extern int get_unused_fd(void); extern void FASTCALL(put_unused_fd(unsigned int fd)); struct kmem_cache; -extern void filp_ctor(void * objp, struct kmem_cache *cachep, unsigned long cflags); -extern void filp_dtor(void * objp, struct kmem_cache *cachep, unsigned long dflags); extern struct file ** alloc_fd_array(int); extern void free_fd_array(struct file **, int); diff --git a/include/linux/fs.h b/include/linux/fs.h index 0cc34b1c42..51c0c93bdf 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -35,6 +35,7 @@ struct files_stat_struct { int max_files; /* tunable */ }; extern struct files_stat_struct files_stat; +extern int get_max_files(void); struct inodes_stat_t { int nr_inodes; -- cgit v1.2.2 From 2ec5e3a867d63d04932e11c6097f63760d9be3fe Mon Sep 17 00:00:00 2001 From: Michael Matz Date: Tue, 7 Mar 2006 21:55:48 -0800 Subject: [PATCH] fix kexec asm While testing kexec and kdump we hit problems where the new kernel would freeze or instantly reboot. The easiest way to trigger it was to kexec a kernel compiled for CONFIG_M586 on an athlon cpu. Compiling for CONFIG_MK7 instead would work fine. The patch fixes a few problems with the kexec inline asm. Signed-off-by: Chris Mason Acked-by: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-powerpc/kexec.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/asm-powerpc/kexec.h b/include/asm-powerpc/kexec.h index bda2f217e6..6a2af2f685 100644 --- a/include/asm-powerpc/kexec.h +++ b/include/asm-powerpc/kexec.h @@ -93,7 +93,8 @@ static inline void crash_setup_regs(struct pt_regs *newregs, "mfxer %0\n" "std %0, 296(%2)\n" : "=&r" (tmp1), "=&r" (tmp2) - : "b" (newregs)); + : "b" (newregs) + : "memory"); } } #else -- cgit v1.2.2 From f9262c12c0084ddba445a9a42e98994018e51400 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 8 Mar 2006 17:57:25 -0800 Subject: [PATCH] i386: port ATI timer fix from x86_64 to i386 II ATI chipsets tend to generate double timer interrupts for the local APIC timer when both the 8254 and the IO-APIC timer pins are enabled. This is because they route it to both and the result is anded together and the CPU ends up processing it twice. This patch changes check_timer to disable the 8254 routing for interrupt 0. I think it would be safe on all chipsets actually (i tested it on a couple and it worked everywhere) and Windows seems to do it in a similar way, but to be conservative this patch only enables this mode on ATI (and adds options to enable/disable too) Ported over from a similar x86-64 change. I reused the ACPI earlyquirk infrastructure for the ATI bridge check, but tweaked it a bit to work even without ACPI. Inspired by a patch from Chuck Ebbert, but redone. Cc: Chuck Ebbert <76306.1226@compuserve.com> Cc: "Brown, Len" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/asm-i386/apic.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/asm-i386/apic.h b/include/asm-i386/apic.h index d30b857157..ff9ac8d19e 100644 --- a/include/asm-i386/apic.h +++ b/include/asm-i386/apic.h @@ -137,6 +137,8 @@ void switch_APIC_timer_to_ipi(void *cpumask); void switch_ipi_to_APIC_timer(void *cpumask); #define ARCH_APICTIMER_STOPS_ON_C3 1 +extern int timer_over_8254; + #else /* !CONFIG_X86_LOCAL_APIC */ static inline void lapic_shutdown(void) { } -- cgit v1.2.2 From fd2a4f1183d1e6802457d70cea067396236ed64b Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Wed, 8 Mar 2006 16:04:32 +0000 Subject: [MIPS] Undefine scr_writew and scr_readw in . This is gluing the build of cirrusfb but really the mess that would need cleaning and fixing is