Re: [HW PROBLEM] Intel I7 MCE. Erratum or not?

From: Eric Dumazet
Date: Mon Dec 08 2008 - 05:03:49 EST


Andi Kleen a écrit :
>
>> IIRC, this error is not what happen on the time [301.7320xx] during
>> boot, but happen before the boot. Since the record says "Processor
>> context corrupt," MCE handler should call panic(or do something stop
>> the system) if the context actually corrupted during the boot.
>
> The weird thing is that 301 seconds is quite a long delay for that.
> It should happen relatively quickly at boot as the CPUs are initialized.
>

Rings a bell here

include/linux/jiffies.h:157:#define INITIAL_JIFFIES ((unsigned long)(unsigned int) (-300*HZ))

Could it be related to INITIAL_JIFFIES ?

>>
>> In other words, it seems that 1) the error was recorded at last time
>> when your machine crashed unexpectedly(by cosmic-ray etc.) and not
>> cleared
>> yet, or 2) your machine is doing something wrong in every reset/poweroff.
>
> When it happens consistently at each boot then yes it's likely something
> leaking from the BIOS initialization sequence. Perhaps try a BIOS update?
>
>

--
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/