perf/x86: fix broken LBR fixup code
PATCH 1/1 perf/x86: fix broken LBR fixup code
I noticed that the LBR fixups were not working anymore on programs where they used to. I tracked this down to a recent change to copy_from_user_nmi().
commit db0dc75d6403b6663c0eab4c6ccb672eb9b2ed72
Author: Arun Sharma <[email protected]> Date: Fri Apr 20 15:41:36 2012 -0700
perf/x86: Check user address explicitly in copy_from_user_nmi()
This commit added a call to __range_not_ok() to the copy_from_user_nmi() routine. The problem is that the logic of the test must be reversed. __range_not_ok() returns 0 if the range is VALID. We want to return early from copy_from_user_nmi() if the range is NOT valid.
---
arch/x86/lib/usercopy.c +1 -1
@@ -22,7 +22,7 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)void *map;int ret; -if (__range_not_ok (from, n,TASK_SIZE ) ==0 ) +if (__range_not_ok (from, n,TASK_SIZE ))return len; do {
Quoted 8 lines · Stephane Eranian
I noticed that the LBR fixups were not working anymore on programs where they used to. I tracked this down to a recent change to
copy_from_user_nmi().commit
db0dc75d6403b6663c0eab4c6ccb672eb9b2ed72Author: Arun Sharma <[email protected]> Date: Fri Apr 20 15:41:36 2012 -0700
Quoted 7 lines
perf/x86: Check user address explicitly in
copy_from_user_nmi()This commit added a call to
__range_not_ok()to thecopy_from_user_nmi()routine. The problem is that the logic of the test must be reversed.__range_not_ok()returns 0 if the range is VALID. We want to return early fromcopy_from_user_nmi()if the range is NOT valid.
D'0h.. Thanks!
Quoted 11 lines · Peter Zijlstra
Quoted 10 lines · Stephane Eranian
I noticed that the LBR fixups were not working anymore on programs where they used to. I tracked this down to a recent change to
copy_from_user_nmi().commit
db0dc75d6403b6663c0eab4c6ccb672eb9b2ed72Author: Arun Sharma<[email protected]> Date: Fri Apr 20 15:41:36 2012 -0700perf/x86: Check user address explicitly in
copy_from_user_nmi()
Quoted 7 lines
Quoted 5 lines
This commit added a call to
__range_not_ok()to thecopy_from_user_nmi()routine. The problem is that the logic of the test must be reversed.__range_not_ok()returns 0 if the range is VALID. We want to return early fromcopy_from_user_nmi()if the range is NOT valid.D'0h.. Thanks!
My bad. I was fooled by all the kernel addresses that were unwound properly. Didn't notice the broken unwinding in user space before I sent the patch. The fix looks good. Thanks.
-Arun
[tip:perf/urgent] perf/x86: Fix broken LBR fixup code
In reply to Stephane EranianCommit-ID: 25f42985825dd93f0593efe454e54c2aa13f7830 Gitweb: http://git.kernel.org/tip/25f42985825dd93f0593efe454e54c2aa13f7830 Author: Stephane Eranian <[email protected]> AuthorDate: Mon, 11 Jun 2012 15:44:26 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Wed, 13 Jun 2012 15:00:28 +0200
perf/x86: Fix broken LBR fixup code
I noticed that the LBR fixups were not working anymore on programs where they used to. I tracked this down to a recent change to copy_from_user_nmi():
db0dc75d640 ("perf/x86: Check user address explicitly in copy_from_user_nmi()")
This commit added a call to __range_not_ok() to the copy_from_user_nmi() routine. The problem is that the logic of the test must be reversed. __range_not_ok() returns 0 if the range is VALID. We want to return early from copy_from_user_nmi() if the range is NOT valid.
--- arch/x86/lib/usercopy.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/lib/usercopy.c +1 -1
@@ -22,7 +22,7 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)void *map;int ret; -if (__range_not_ok (from, n,TASK_SIZE ) ==0 ) +if (__range_not_ok (from, n,TASK_SIZE ))return len; do {