EDAC, MCE, AMD: Print CPU number when reporting the error
authorBorislav Petkov <borislav.petkov@amd.com>
Thu, 4 Aug 2011 17:25:24 +0000 (19:25 +0200)
committerBorislav Petkov <borislav.petkov@amd.com>
Thu, 6 Oct 2011 10:34:02 +0000 (12:34 +0200)
Currently, correctable ECCs go through mcelog and do not print the scary
MCE banner. In that case, however, reporting the core where the CECC
happened is important information so dump it along with the decoded
string albeit at risk of having a minor redundancy.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
drivers/edac/mce_amd.c

index 795cfbc0bf50ff39b3f3911b6596d3353f90d4ef..5bfe6997d9ff73207d6497f3bc15ffc15f9f5795 100644 (file)
@@ -769,8 +769,8 @@ int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data)
        if (amd_filter_mce(m))
                return NOTIFY_STOP;
 
-       pr_emerg(HW_ERR "MC%d_STATUS[%s|%s|%s|%s|%s",
-               m->bank,
+       pr_emerg(HW_ERR "CPU:%d MC%d_STATUS[%s|%s|%s|%s|%s",
+               m->extcpu, m->bank,
                ((m->status & MCI_STATUS_OVER)  ? "Over"  : "-"),
                ((m->status & MCI_STATUS_UC)    ? "UE"    : "CE"),
                ((m->status & MCI_STATUS_MISCV) ? "MiscV" : "-"),
This page took 0.025364 seconds and 5 git commands to generate.