summaryrefslogtreecommitdiffstats
path: root/arch/i386/kernel/cpu
diff options
context:
space:
mode:
authorSiddha, Suresh B <suresh.b.siddha@intel.com>2006-03-27 11:15:22 +0200
committerLinus Torvalds <torvalds@g5.osdl.org>2006-03-27 18:44:43 +0200
commit1e9f28fa1eb9773bf65bae08288c6a0a38eef4a7 (patch)
treeccfa4927ebc7a8f663f9ac9e7789a713a33253ff /arch/i386/kernel/cpu
parent[PATCH] Small schedule() optimization (diff)
downloadlinux-1e9f28fa1eb9773bf65bae08288c6a0a38eef4a7.tar.xz
linux-1e9f28fa1eb9773bf65bae08288c6a0a38eef4a7.zip
[PATCH] sched: new sched domain for representing multi-core
Add a new sched domain for representing multi-core with shared caches between cores. Consider a dual package system, each package containing two cores and with last level cache shared between cores with in a package. If there are two runnable processes, with this appended patch those two processes will be scheduled on different packages. On such systems, with this patch we have observed 8% perf improvement with specJBB(2 warehouse) benchmark and 35% improvement with CFP2000 rate(with 2 users). This new domain will come into play only on multi-core systems with shared caches. On other systems, this sched domain will be removed by domain degeneration code. This new domain can be also used for implementing power savings policy (see OLS 2005 CMP kernel scheduler paper for more details.. I will post another patch for power savings policy soon) Most of the arch/* file changes are for cpu_coregroup_map() implementation. Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch/i386/kernel/cpu')
-rw-r--r--arch/i386/kernel/cpu/common.c10
-rw-r--r--arch/i386/kernel/cpu/intel_cacheinfo.c22
2 files changed, 27 insertions, 5 deletions
diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c
index 7e3d6b6a4e96..a06a49075f10 100644
--- a/arch/i386/kernel/cpu/common.c
+++ b/arch/i386/kernel/cpu/common.c
@@ -266,7 +266,7 @@ static void __init early_cpu_detect(void)
void __cpuinit generic_identify(struct cpuinfo_x86 * c)
{
u32 tfms, xlvl;
- int junk;
+ int ebx;
if (have_cpuid_p()) {
/* Get vendor name */
@@ -282,7 +282,7 @@ void __cpuinit generic_identify(struct cpuinfo_x86 * c)
/* Intel-defined flags: level 0x00000001 */
if ( c->cpuid_level >= 0x00000001 ) {
u32 capability, excap;
- cpuid(0x00000001, &tfms, &junk, &excap, &capability);
+ cpuid(0x00000001, &tfms, &ebx, &excap, &capability);
c->x86_capability[0] = capability;
c->x86_capability[4] = excap;
c->x86 = (tfms >> 8) & 15;
@@ -292,6 +292,11 @@ void __cpuinit generic_identify(struct cpuinfo_x86 * c)
if (c->x86 >= 0x6)
c->x86_model += ((tfms >> 16) & 0xF) << 4;
c->x86_mask = tfms & 15;
+#ifdef CONFIG_SMP
+ c->apicid = phys_pkg_id((ebx >> 24) & 0xFF, 0);
+#else
+ c->apicid = (ebx >> 24) & 0xFF;
+#endif
} else {
/* Have CPUID level 0 only - unheard of */
c->x86 = 4;
@@ -474,7 +479,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
cpuid(1, &eax, &ebx, &ecx, &edx);
- c->apicid = phys_pkg_id((ebx >> 24) & 0xFF, 0);
if (!cpu_has(c, X86_FEATURE_HT) || cpu_has(c, X86_FEATURE_CMP_LEGACY))
return;
diff --git a/arch/i386/kernel/cpu/intel_cacheinfo.c b/arch/i386/kernel/cpu/intel_cacheinfo.c
index ce61921369e5..7e7fd4e67dd0 100644
--- a/arch/i386/kernel/cpu/intel_cacheinfo.c
+++ b/arch/i386/kernel/cpu/intel_cacheinfo.c
@@ -173,6 +173,10 @@ unsigned int __cpuinit init_intel_cacheinfo(struct cpuinfo_x86 *c)
unsigned int trace = 0, l1i = 0, l1d = 0, l2 = 0, l3 = 0; /* Cache sizes */
unsigned int new_l1d = 0, new_l1i = 0; /* Cache sizes from cpuid(4) */
unsigned int new_l2 = 0, new_l3 = 0, i; /* Cache sizes from cpuid(4) */
+ unsigned int l2_id = 0, l3_id = 0, num_threads_sharing, index_msb;
+#ifdef CONFIG_SMP
+ unsigned int cpu = (c == &boot_cpu_data) ? 0 : (c - cpu_data);
+#endif
if (c->cpuid_level > 3) {
static int is_initialized;
@@ -205,9 +209,15 @@ unsigned int __cpuinit init_intel_cacheinfo(struct cpuinfo_x86 *c)
break;
case 2:
new_l2 = this_leaf.size/1024;
+ num_threads_sharing = 1 + this_leaf.eax.split.num_threads_sharing;
+ index_msb = get_count_order(num_threads_sharing);
+ l2_id = c->apicid >> index_msb;
break;
case 3:
new_l3 = this_leaf.size/1024;
+ num_threads_sharing = 1 + this_leaf.eax.split.num_threads_sharing;
+ index_msb = get_count_order(num_threads_sharing);
+ l3_id = c->apicid >> index_msb;
break;
default:
break;
@@ -273,11 +283,19 @@ unsigned int __cpuinit init_intel_cacheinfo(struct cpuinfo_x86 *c)
if (new_l1i)
l1i = new_l1i;
- if (new_l2)
+ if (new_l2) {
l2 = new_l2;
+#ifdef CONFIG_SMP
+ cpu_llc_id[cpu] = l2_id;
+#endif
+ }
- if (new_l3)
+ if (new_l3) {
l3 = new_l3;
+#ifdef CONFIG_SMP
+ cpu_llc_id[cpu] = l3_id;
+#endif
+ }
if ( trace )
printk (KERN_INFO "CPU: Trace cache: %dK uops", trace);