[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>

6 files changed, 23 insertions, 0 deletions

diff --git a/include/asm-i386/processor.h b/include/asm-i386/processor.h
index feca5d961e2b..af4bfd012475 100644
--- a/include/asm-i386/processor.h
+++ b/include/asm-i386/processor.h
@@ -20,6 +20,7 @@
 #include <linux/config.h>
 #include <linux/threads.h>
 #include <asm/percpu.h>
+#include <linux/cpumask.h>
 
 /* flag for disabling the tsc */
 extern int tsc_disable;
@@ -67,6 +68,9 @@ struct cpuinfo_x86 {
         char    pad0;
         int     x86_power;
         unsigned long loops_per_jiffy;
+#ifdef CONFIG_SMP
+        cpumask_t llc_shared_map;       /* cpus sharing the last level cache */
+#endif
         unsigned char x86_max_cores;    /* cpuid returned max cores value */
         unsigned char booted_cores;     /* number of cores as seen by OS */
         unsigned char apicid;
@@ -103,6 +107,7 @@ extern struct cpuinfo_x86 cpu_data[];
 
 extern  int phys_proc_id[NR_CPUS];
 extern  int cpu_core_id[NR_CPUS];
+extern  int cpu_llc_id[NR_CPUS];
 extern char ignore_fpu_irq;
 
 extern void identify_cpu(struct cpuinfo_x86 *);
diff --git a/include/asm-i386/topology.h b/include/asm-i386/topology.h
index aa958c6ee83e..b94e5eeef917 100644
--- a/include/asm-i386/topology.h
+++ b/include/asm-i386/topology.h
@@ -112,4 +112,6 @@ extern unsigned long node_remap_size[];
 
 #endif /* CONFIG_NUMA */
 
+extern cpumask_t cpu_coregroup_map(int cpu);
+
 #endif /* _ASM_I386_TOPOLOGY_H */
diff --git a/include/asm-x86_64/processor.h b/include/asm-x86_64/processor.h
index 8c8d88c036ed..1aa2cee43344 100644
--- a/include/asm-x86_64/processor.h
+++ b/include/asm-x86_64/processor.h
@@ -20,6 +20,7 @@
 #include <asm/mmsegment.h>
 #include <asm/percpu.h>
 #include <linux/personality.h>
+#include <linux/cpumask.h>
 
 #define TF_MASK         0x00000100
 #define IF_MASK         0x00000200
@@ -65,6 +66,9 @@ struct cpuinfo_x86 {
         __u32   x86_power;      
         __u32   extended_cpuid_level;   /* Max extended CPUID function supported */
         unsigned long loops_per_jiffy;
+#ifdef CONFIG_SMP
+        cpumask_t llc_shared_map;       /* cpus sharing the last level cache */
+#endif
         __u8    apicid;
         __u8    booted_cores;   /* number of cores as seen by OS */
 } ____cacheline_aligned;
diff --git a/include/asm-x86_64/smp.h b/include/asm-x86_64/smp.h
index 9ccbb2cfd5c0..a4fdaeb5c397 100644
--- a/include/asm-x86_64/smp.h
+++ b/include/asm-x86_64/smp.h
@@ -56,6 +56,7 @@ extern cpumask_t cpu_sibling_map[NR_CPUS];
 extern cpumask_t cpu_core_map[NR_CPUS];
 extern u8 phys_proc_id[NR_CPUS];
 extern u8 cpu_core_id[NR_CPUS];
+extern u8 cpu_llc_id[NR_CPUS];
 
 #define SMP_TRAMPOLINE_BASE 0x6000
 
diff --git a/include/asm-x86_64/topology.h b/include/asm-x86_64/topology.h
index c642f5d9882d..9db54e9d17bb 100644
--- a/include/asm-x86_64/topology.h
+++ b/include/asm-x86_64/topology.h
@@ -68,4 +68,6 @@ extern int __node_distance(int, int);
 
 #include <asm-generic/topology.h>
 
+extern cpumask_t cpu_coregroup_map(int cpu);
+
 #endif
diff --git a/include/linux/topology.h b/include/linux/topology.h
index e8eb0040ce3a..a305ae2e44b6 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -164,6 +164,15 @@
         .nr_balance_failed      = 0,                    \
 }
 
+#ifdef CONFIG_SCHED_MC
+#ifndef SD_MC_INIT
+/* for now its same as SD_CPU_INIT.
+ * TBD: Tune Domain parameters!
+ */
+#define SD_MC_INIT   SD_CPU_INIT
+#endif
+#endif
+
 #ifdef CONFIG_NUMA
 #ifndef SD_NODE_INIT
 #error Please define an appropriate SD_NODE_INIT in include/asm/topology.h!!!