Re: [PATCH 5/6] x86, mce: handle "action required" errors

[Chen Gong <gong.chen@linux.intel.com>]

于 2011/12/15 5:30, Tony Luck 写道:
> On Wed, Dec 14, 2011 at 1:28 AM, Chen Gong<gong.chen@linux.intel.com>  wrote:
>>> -       if (kill_it&&    tolerant<    3)
>>>
>>> +       if (worst != MCE_AR_SEVERITY&&    kill_it&&    tolerant<    3)
>>>                 force_sig(SIGBUS, current);
>>
>>
>> I think here it should add more comments to clarify why not killing *AR*
>> case.
>> Such as: "for SRAR errors, such as DCU/IFU error, on affected logical
>> processors, it is reasonable that RIPV is 0."
>
> I'll look at this - the reason to not kill for AR is that we want to
> try to recover
> first (e.g. page could be re-read from disk into a different physical page).
> In some cases we can recover transparently to the application.

Oh, yes, these reasons are very important why not killing *AR* events. But my
point is in a *AR* supported environment, "kill_it" should not be true like
below:
         if (!(m.mcgstatus & MCG_STATUS_RIPV))
                 kill_it = 1;

the reason is what I said before. But at that time the worst severity hasn't
been determined so we have to wati until it is out.

anyway, it is an interesting coincidence, isn't it? :-)