[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] Xen/MCE: adjust for future new vMCE model

Luck, Tony wrote:
>> Yes, it's vMCE specific issue, and only under the case when migrate
>> from platform A w/ bigger bank number to platform B w/ smaller bank
>> number. Under this case, it would block migration from 3.x -> 4.0 ->
>> 4.1 -> 4.2. 
> Linux (and probably every other OS) will only look to see how many
> banks there 
> are once at boot time. From that point on it will scan as many banks
> as it was told existed for the rest of eternity. So it really doesn't
> matter what value you present in MCG_CAP after a  migration - the OS
> will never look at it any way. The important thing is what you do
>       when the OS does: for (i = 0; i < banks; i++) {
>               status = mce_rdmsrl(MSR_IA32_MCx_STATUS(i));
>               ...
>       }
> because you'd better give it something reasonable for each bank that
> it was 
> told existed.

Hmm, this bug is not worry about MCG_CAP bank field. It in fact caused from 
current dirty xen vmce code and its data structure ... and because of that, it 
block migration when migrate from big bank to small bank platform :)

> Also if you are going to present a machine check to the guest, then
> you need to fill in the data for some bank that it knows about (i.e.
> i < banks ... and possibly i != 0 for certain CPUIDs).
> -Tony

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.