[patch 4/4] powerpc 2.6.21-rt1: reduce scheduling latency by changing tlb flush size

Tsutomu OWA tsutomu.owa at toshiba.co.jp
Tue May 15 16:27:26 EST 2007


At Mon, 14 May 2007 16:40:02 +0200, Arnd Bergmann wrote:

> > +#if defined(CONFIG_PPC_CELLEB) && defined(CONFIG_PREEMPT_RT)
> > +/* Since tlb flush takes long time on Celleb, reduce it to 1 when Celleb && RT */
> > +#define PPC64_TLB_BATCH_NR 1

> With this code, you get silent side-effects of enabling PPC_CELLEB
> along with another platform.

> Maybe instead you should change the hpte_need_flush() to always flush
> when running on the celleb platform and PREEMPT_RT is enabled.

  OK, how about this one?

  thanks a lot!

Since flushing tlb needs expensive hypervisor call(s) on celleb,
always flush it on RT to reduce scheduling latency.

Signed-off-by: Tsutomu OWA <tsutomu.owa at toshiba.co.jp>
-- owa

--- linux-2.6.21-rt1/arch/powerpc/mm/tlb_64.c	2007-05-07 14:08:12.000000000 +0900
+++ rt/arch/powerpc/mm/tlb_64.c	2007-05-15 15:19:34.000000000 +0900
@@ -31,6 +31,7 @@
 #include <asm/tlbflush.h>
 #include <asm/tlb.h>
 #include <asm/bug.h>
+#include <asm/machdep.h>
 
 DEFINE_PER_CPU(struct ppc64_tlb_batch, ppc64_tlb_batch);
 
@@ -180,6 +181,18 @@ void hpte_update(struct mm_struct *mm, u
 	batch->vaddr[i] = (vsid << 28 ) | (addr & 0x0fffffff);
 	batch->pte[i] = __real_pte(__pte(pte), ptep);
 	batch->index = ++i;
+
+#ifdef CONFIG_PREEMPT_RT
+	/*
+	 * Since flushing tlb needs expensive hypervisor call(s) on celleb,
+	 * always flush it on RT to reduce scheduling latency.
+	 */
+	if (machine_is(celleb)) {
+		flush_tlb_pending();
+		return;
+	}
+#endif /* CONFIG_PREEMPT_RT */
+
 	if (i >= PPC64_TLB_BATCH_NR)
 		flush_tlb_pending();
 }



More information about the Linuxppc-dev mailing list