Messages in this thread Patch in this message |  | | Date | Mon, 8 Jun 2015 22:26:58 +0200 | From | Borislav Petkov <> | Subject | Re: [PATCH 4/4 Rebase] x86, MCE: Avoid potential deadlock in MCE context |
| |
On Mon, Jun 08, 2015 at 08:03:08PM +0000, Luck, Tony wrote: > @@ -156,7 +156,8 @@ void mce_log(struct mce *mce) > /* Emit the trace record: */ > trace_mce_record(mce); > > - atomic_notifier_call_chain(&x86_mce_decoder_chain, 0, mce); > + mce_genpool_add(mce); > + irq_work_queue(&mce_irq_work); > > Is it safe to call irq_work_queue() from MCE context?
Yeah, we're using it in contexts like:
do_nmi |-> default_do_nmi |-> nmi_handle |->irq_work_queue
for example. I think that is good enough for #MC. :)
> If that is OK, then I don't have any concerns.
Thanks. -- Regards/Gruss, Boris.
ECO tip #101: Trim your mails when you reply. --
|  |