Re: [PATCH] [4/4] x86: MCE: Fix EIPV behaviour with !PCC

[Huang Ying <ying.huang@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 1330 bytes --]

Add some description for the patch, hope that to be more clear.

Best Regards,
Huang Ying
-------------------------------------------------->
Impact: Spec compliance

Tolerant level 0 means: always panic on uncorrected errors, that is,
panic even for recoverable uncorrected errors. This is a useful option
for someone think panic is the better hardware error containment
mechanism than trying to recover.

Current implementation does not comply with the tolerant == 0 spec,
that is, it tries to recover (by killing related processes) for
recoverable uncorrected errors (errors triggered in userspace) when
tolerant == 0. This patch fixes this by going panic for that case.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>

---
 arch/x86/kernel/cpu/mcheck/mce_64.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -400,7 +400,7 @@ void do_machine_check(struct pt_regs * r
 		 * force_sig() takes an awful lot of locks and has a slight
 		 * risk of deadlocking.
 		 */
-		if (user_space) {
+		if (user_space && tolerant > 0) {
 			force_sig(SIGBUS, current);
 		} else if (panic_on_oops || tolerant < 2) {
 			mce_panic("Uncorrected machine check",


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]