[PATCH 0/2] x86, MCE: Catch the early worm

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@xxxxxxx>

Currently, we don't decode MCEs which are being detected during MCA
initialization and before registration of the decoders on the decoding
notifier chain. These two patches add the functionality to drain the
mcelog buffer right after we've registered a decoder.

With it, early MCEs get caught and decoded into the dmesg:

...
[    0.068003] ... generic registers:      6
[    0.068003] ... value mask:             0000ffffffffffff
[    0.068003] ... max period:             00007fffffffffff
[    0.068003] ... fixed-purpose events:   0
[    0.068003] ... event mask:             000000000000003f
[    0.068003] MCE: In-kernel MCE decoding enabled.
[    0.068003] [Hardware Error]: CPU:0  MC0_STATUS[-|UE|MiscV|-|AddrV|-|-]: 0xac80000000000833
[    0.068003] [Hardware Error]:        MC0_ADDR: 0x00000000f8080000
[    0.068003] [Hardware Error]: Data Cache Error: System Read Data Error.
[    0.068003] [Hardware Error]: cache level: L3/GEN, mem/io: MEM, mem-tx: DRD, part-proc: SRC (no timeout)
[    0.068069] Booting Node   0, Processors  #1
[    0.068276] smpboot cpu 1: start_ip = 88000
[    0.160077]  #2
[    0.160206] smpboot cpu 2: start_ip = 88000
[    0.252073]  #3 Ok.
...

Comments and suggestions are welcome, as always.

Thanks.
--
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/