Re: [git pull] machine check recovery fix

[Linus Torvalds]

On Thu, May 17, 2012 at 5:14 PM, Tony Luck <tony.luck@xxxxxxxxx> wrote:
> On Thu, May 17, 2012 at 3:45 PM, Linus Torvalds
> <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>> Also, the whole "nonrestartable state flag" means - if I understood
>> things correctly - that you really cannot do the "iret" even from the
>> NMI handler.
>
> Not quite ... we can "iret" ... but not back to the instruction that was
> executing when the machine check occurred. We need to go some
> place else .... hence we send a signal that will either kill the process

Tony, I don't think you understand.

If the machine check happened in kernel space, we currently *are*
returning to the instruction that executed. With or without your
patch. That's my argument.

Your _TIF_MCE_NOTIFY games do *nothing*, because they only get tested
at return to user space - not on return to the MC faulting kernel
space instruction.

This is what I was talking about - the thing looks to work entirely
*accidentally* - and only for the user-space case.

                        Linus
--
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/