[PATCH for 2.6.18rc2] [2/7] i386/x86-64: Add user_mode checks toprofile_pc for oprofile

From: Andi Kleen
Date: Mon Jul 24 2006 - 14:41:27 EST

Fixes a obscure user space triggerable crash during oprofiling.

Oprofile calls profile_pc from NMIs even when user_mode(regs) is not true and
the program counter is inside the kernel lock section. This opens
a race - when a user program jumps to a kernel lock address and
a NMI happens before the illegal page fault exception is raised
and the program has a unmapped esp or ebp then the kernel could
oops. NMIs have a higher priority than exceptions so that could

Add user_mode checks to i386/x86-64 profile_pc to prevent that.

Cc: John Levon <levon@xxxxxxxxxxxxxxxxx>
Signed-off-by: Andi Kleen <ak@xxxxxxx>

arch/i386/kernel/time.c | 2 +-
arch/x86_64/kernel/time.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)

Index: linux/arch/i386/kernel/time.c
--- linux.orig/arch/i386/kernel/time.c
+++ linux/arch/i386/kernel/time.c
@@ -135,7 +135,7 @@ unsigned long profile_pc(struct pt_regs
unsigned long pc = instruction_pointer(regs);

- if (in_lock_functions(pc))
+ if (!user_mode_vm(regs) && in_lock_functions(pc))
return *(unsigned long *)(regs->ebp + 4);

return pc;
Index: linux/arch/x86_64/kernel/time.c
--- linux.orig/arch/x86_64/kernel/time.c
+++ linux/arch/x86_64/kernel/time.c
@@ -195,7 +195,7 @@ unsigned long profile_pc(struct pt_regs
is just accounted to the spinlock function.
Better would be to write these functions in assembler again
and check exactly. */
- if (in_lock_functions(pc)) {
+ if (!user_mode(regs) && in_lock_functions(pc)) {
char *v = *(char **)regs->rsp;
if ((v >= _stext && v <= _etext) ||
(v >= _sinittext && v <= _einittext) ||
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/