Messages in this thread | | | From | Andy Lutomirski <> | Date | Tue, 11 Nov 2014 09:15:18 -0800 | Subject | Re: [PATCH v2 4/5] x86/mce: Simplify flow when handling recoverable memory errors |
| |
On Tue, Nov 11, 2014 at 8:30 AM, Borislav Petkov <bp@alien8.de> wrote: > On Tue, Nov 11, 2014 at 08:22:45AM -0800, Andy Lutomirski wrote: >> I think it's okay-ish, but only if it's necessary, and I still don't >> see why it's necessary. >> >> Can't you just remove TIF_MCE_NOTIFY entirely and just do all the >> mce_notify_process work directly in do_machine_check? IOW, why do you >> need to store any state per-task when it's already on the stack >> anyway. > > I wish but memory_failure() can't run in #MC context as it noodles > quite a lot and grabs all kinds of locks and does a bunch of other > atomit-context-unsafe things.
Oh -- does it need to sleep?
I find myself wondering whether a much cleaner solution might be to sync regs and switch stacks before invoking do_machine_check rather than afterwards. Then do_machine_check would really be completely non-atomic. It would add a few lines of asm, though.
--Andy
> > And it needs to run *before* the process is killed as it looks at its > pages. > > -- > Regards/Gruss, > Boris. > > Sent from a fat crate under my desk. Formatting is fine. > --
-- Andy Lutomirski AMA Capital Management, LLC
| |