mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] Avoid double flush in do_tlb_flush_all()
@ 2008-08-30 10:20 Andi Kleen
  2008-08-30 16:01 ` Arjan van de Ven
  0 siblings, 1 reply; 2+ messages in thread
From: Andi Kleen @ 2008-08-30 10:20 UTC (permalink / raw)
  To: x86, linux-kernel

Avoid double flush in do_tlb_flush_all()

leave_mm already flushes the TLBs, so don't need to flush directly before that
in do_flush_tlb_all(). Avoid the double flush.

Pointed out by Rainer Weikusat

Signed-off-by: Andi Kleen <ak@linux.intel.com>

---
 arch/x86/kernel/tlb_32.c |    5 ++---
 arch/x86/kernel/tlb_64.c |    7 +++----
 2 files changed, 5 insertions(+), 7 deletions(-)

Index: linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_32.c
===================================================================
--- linux-2.6.27-rc4-misc.orig/arch/x86/kernel/tlb_32.c
+++ linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_32.c
@@ -229,11 +229,10 @@ EXPORT_SYMBOL(flush_tlb_page);
 
 static void do_flush_tlb_all(void *info)
 {
-	unsigned long cpu = smp_processor_id();
-
-	__flush_tlb_all();
 	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY)
 		leave_mm(cpu);
+	else
+		__flush_tlb_all(smp_processor_id());
 }
 
 void flush_tlb_all(void)
Index: linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_64.c
===================================================================
--- linux-2.6.27-rc4-misc.orig/arch/x86/kernel/tlb_64.c
+++ linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_64.c
@@ -266,11 +266,10 @@ void flush_tlb_page(struct vm_area_struc
 
 static void do_flush_tlb_all(void *info)
 {
-	unsigned long cpu = smp_processor_id();
-
-	__flush_tlb_all();
 	if (read_pda(mmu_state) == TLBSTATE_LAZY)
-		leave_mm(cpu);
+		leave_mm(smp_processor_id());
+	else
+		__flush_tlb_all();
 }
 
 void flush_tlb_all(void)

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] Avoid double flush in do_tlb_flush_all()
  2008-08-30 10:20 [PATCH] Avoid double flush in do_tlb_flush_all() Andi Kleen
@ 2008-08-30 16:01 ` Arjan van de Ven
  0 siblings, 0 replies; 2+ messages in thread
From: Arjan van de Ven @ 2008-08-30 16:01 UTC (permalink / raw)
  To: Andi Kleen; +Cc: x86, linux-kernel

On Sat, 30 Aug 2008 12:20:12 +0200
Andi Kleen <andi@firstfloor.org> wrote:

> Avoid double flush in do_tlb_flush_all()
> 
> leave_mm already flushes the TLBs, so don't need to flush directly
> before that in do_flush_tlb_all(). Avoid the double flush.

but... the leave_mm won't flush the global tlbs, while
__flush_tlb_all() does, right?

sounds like we do need the behavior

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2008-08-30 16:02 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-08-30 10:20 [PATCH] Avoid double flush in do_tlb_flush_all() Andi Kleen
2008-08-30 16:01 ` Arjan van de Ven

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®