x86: add memory clobber in switch_to()
RFC PATCH 1/1 x86: add memory clobber in switch_to()
There are more places which have segment register reloads without adding memory clobbers. Should we fix them up? Here's one of them:
Quoted 1 line
From
5f135687f7730fbdd979b6fb408fd3ca08a0a678Mon Sep 17 00:00:00 2001
From: Vegard Nossum <[email protected]>
Date: Sun, 14 Sep 2008 18:53:22 +0200
Subject: [PATCH] x86: add memory clobber in switch_to()
Segment registers are reloaded, so we should add a memory clobber. The generated assembly code is identical in my tests, but this doesn't mean it is necessarily true for all configurations/compilers.
x86_64 already has the memory clobber.
--- include/asm-x86/system.h | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-)
include/asm-x86/system.h +4 -1
@@ -64,7 +64,10 @@ do { \ \ /* regparm parameters for __switch_to(): */ \ [prev] "a" (prev), \ -[next] " d" (next)); \ +[next] " d" (next) \ +\ +: /* reloaded segment registers */ \ +" memory"); \ } while (0) /*
Signature · 1 line
1.5.5.1
Ingo Molnar
AppliedQuoted 2 lines · Vegard Nossum
There are more places which have segment register reloads without adding memory clobbers. Should we fix them up? Here's one of them:
yes, i think so.
Quoted 8 lines
Quoted 1 line
From
5f135687f7730fbdd979b6fb408fd3ca08a0a678Mon Sep 17 00:00:00 2001From: Vegard Nossum <[email protected]>
Date: Sun, 14 Sep 2008 18:53:22 +0200
Subject: [PATCH] x86: add memory clobber inswitch_to()Segment registers are reloaded, so we should add a memory clobber. The generated assembly code is identical in my tests, but this doesn't mean it is necessarily true for all configurations/compilers.
applied to tip/x86/core, thanks Vegard.
Ingo