2015-06-22 19:39:18

by Yann Droneaud

[permalink] [raw]
Subject: [PATCH] x86: fix copy_from_user_nmi() return if range is not ok

Commit 0a196848ca36 ("perf: Fix arch_perf_out_copy_user default"),
changes copy_from_user_nmi() to return the number of
remaining bytes so that it behave like copy_from_user().

Unfortunately, when the range is outside of the process,
memory the return value is still the number of byte
copied, eg. 0, instead of the remaining bytes.

As all users of copy_from_user_nmi() were modified as
part of commit 0a196848ca36, the function should be
fixed to return the total number of bytes if range is
not correct.

Signed-off-by: Yann Droneaud <[email protected]>
---
arch/x86/lib/usercopy.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c
index ddf9ecb53cc3..e342586db6e4 100644
--- a/arch/x86/lib/usercopy.c
+++ b/arch/x86/lib/usercopy.c
@@ -20,7 +20,7 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
unsigned long ret;

if (__range_not_ok(from, n, TASK_SIZE))
- return 0;
+ return n;

/*
* Even though this function is typically called from NMI/IRQ context
--
2.4.3


Subject: [tip:perf/urgent] perf/x86: Fix copy_from_user_nmi() return if range is not ok

Commit-ID: ebf2d2689de551d90965090bb991fc640a0c0d41
Gitweb: http://git.kernel.org/tip/ebf2d2689de551d90965090bb991fc640a0c0d41
Author: Yann Droneaud <[email protected]>
AuthorDate: Mon, 22 Jun 2015 21:38:43 +0200
Committer: Ingo Molnar <[email protected]>
CommitDate: Mon, 6 Jul 2015 14:09:27 +0200

perf/x86: Fix copy_from_user_nmi() return if range is not ok

Commit 0a196848ca36 ("perf: Fix arch_perf_out_copy_user default"),
changes copy_from_user_nmi() to return the number of
remaining bytes so that it behave like copy_from_user().

Unfortunately, when the range is outside of the process
memory, the return value is still the number of byte
copied, eg. 0, instead of the remaining bytes.

As all users of copy_from_user_nmi() were modified as
part of commit 0a196848ca36, the function should be
fixed to return the total number of bytes if range is
not correct.

Signed-off-by: Yann Droneaud <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
---
arch/x86/lib/usercopy.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c
index ddf9ecb..e342586 100644
--- a/arch/x86/lib/usercopy.c
+++ b/arch/x86/lib/usercopy.c
@@ -20,7 +20,7 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
unsigned long ret;

if (__range_not_ok(from, n, TASK_SIZE))
- return 0;
+ return n;

/*
* Even though this function is typically called from NMI/IRQ context