[PATCH] Reduce TLB flushing during process migration

From: Martin Hicks <mort@wildopensource.com>

Another optimization patch from Jack Steiner, intended to reduce TLB
flushes during process migration.

Most architextures should define tlb_migrate_prepare() to be flush_tlb_mm(),
but on i386, it would be a wasted flush, because i386 disconnects previous
cpus from the tlb flush automatically.

include/asm-generic/tlb.h

@@ -146,4 +146,6 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
 		__pmd_free_tlb(tlb, pmdp);			\
 	} while (0)
 
+#define tlb_migrate_prepare(mm) do { } while(0)
+
 #endif /* _ASM_GENERIC__TLB_H */

include/asm-ia64/tlb.h

@@ -211,6 +211,8 @@ __tlb_remove_tlb_entry (struct mmu_gather *tlb, pte_t *ptep, unsigned long addre
 	tlb->end_addr = address + PAGE_SIZE;
 }
 
+#define tlb_migrate_prepare(mm) flush_tlb_mm(mm)
+
 #define tlb_start_vma(tlb, vma)			do { } while (0)
 #define tlb_end_vma(tlb, vma)			do { } while (0)
 

kernel/sched.c

@@ -26,6 +26,8 @@
 #include <linux/highmem.h>
 #include <linux/smp_lock.h>
 #include <asm/mmu_context.h>
+#include <linux//pagemap.h>
+#include <asm/tlb.h>
 #include <linux/interrupt.h>
 #include <linux/completion.h>
 #include <linux/kernel_stat.h>
@@ -1127,6 +1129,14 @@ static void sched_migrate_task(task_t *p, int dest_cpu)
 		task_rq_unlock(rq, &flags);
 		wake_up_process(rq->migration_thread);
 		wait_for_completion(&req.done);
+
+		/*
+		 * we want a new context here. This eliminates TLB
+		 * flushes on the cpus where the process executed prior to
+		 * the migration.
+		 */
+		tlb_migrate_prepare(current->mm);
+
 		return;
 	}
 out: