diff options
| author | Ravikiran G Thirumalai <kiran@scalex86.org> | 2006-01-08 01:01:27 -0800 |
|---|---|---|
| committer | Linus Torvalds <torvalds@g5.osdl.org> | 2006-01-08 20:13:38 -0800 |
| commit | 22fc6eccbf4ce4eb6265e6ada7b50a7b9cc57d05 (patch) | |
| tree | 3887dc6f1eeb658d773be037971b98d6f5fb3dd7 /arch | |
| parent | 6d524aed1f50b2b1d5b4ad5a4e2fe3f38106d0a6 (diff) | |
| download | op-kernel-dev-22fc6eccbf4ce4eb6265e6ada7b50a7b9cc57d05.zip op-kernel-dev-22fc6eccbf4ce4eb6265e6ada7b50a7b9cc57d05.tar.gz | |
[PATCH] Change maxaligned_in_smp alignemnt macros to internodealigned_in_smp macros
____cacheline_maxaligned_in_smp is currently used to align critical structures
and avoid false sharing. It uses per-arch L1_CACHE_SHIFT_MAX and people find
L1_CACHE_SHIFT_MAX useless.
However, we have been using ____cacheline_maxaligned_in_smp to align
structures on the internode cacheline size. As per Andi's suggestion,
following patch kills ____cacheline_maxaligned_in_smp and introduces
INTERNODE_CACHE_SHIFT, which defaults to L1_CACHE_SHIFT for all arches.
Arches needing L3/Internode cacheline alignment can define
INTERNODE_CACHE_SHIFT in the arch asm/cache.h. Patch replaces
____cacheline_maxaligned_in_smp with ____cacheline_internodealigned_in_smp
With this patch, L1_CACHE_SHIFT_MAX can be killed
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch')
| -rw-r--r-- | arch/i386/kernel/init_task.c | 2 | ||||
| -rw-r--r-- | arch/i386/kernel/irq.c | 2 | ||||
| -rw-r--r-- | arch/x86_64/kernel/init_task.c | 2 |
3 files changed, 3 insertions, 3 deletions
diff --git a/arch/i386/kernel/init_task.c b/arch/i386/kernel/init_task.c index 9caa8e8..cff95d1 100644 --- a/arch/i386/kernel/init_task.c +++ b/arch/i386/kernel/init_task.c @@ -42,5 +42,5 @@ EXPORT_SYMBOL(init_task); * per-CPU TSS segments. Threads are completely 'soft' on Linux, * no more per-task TSS's. */ -DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_maxaligned_in_smp = INIT_TSS; +DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_internodealigned_in_smp = INIT_TSS; diff --git a/arch/i386/kernel/irq.c b/arch/i386/kernel/irq.c index 1a201a93..f3a9c78 100644 --- a/arch/i386/kernel/irq.c +++ b/arch/i386/kernel/irq.c @@ -19,7 +19,7 @@ #include <linux/cpu.h> #include <linux/delay.h> -DEFINE_PER_CPU(irq_cpustat_t, irq_stat) ____cacheline_maxaligned_in_smp; +DEFINE_PER_CPU(irq_cpustat_t, irq_stat) ____cacheline_internodealigned_in_smp; EXPORT_PER_CPU_SYMBOL(irq_stat); #ifndef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86_64/kernel/init_task.c b/arch/x86_64/kernel/init_task.c index e0ba5c1..ce31d90 100644 --- a/arch/x86_64/kernel/init_task.c +++ b/arch/x86_64/kernel/init_task.c @@ -44,6 +44,6 @@ EXPORT_SYMBOL(init_task); * section. Since TSS's are completely CPU-local, we want them * on exact cacheline boundaries, to eliminate cacheline ping-pong. */ -DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_maxaligned_in_smp = INIT_TSS; +DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_internodealigned_in_smp = INIT_TSS; #define ALIGN_TO_4K __attribute__((section(".data.init_task"))) |
