On Tue, Jun 11, 2019 at 03:50:40PM +1000, Benjamin Herrenschmidt wrote:
> Should we fix that then instead ?
Sure.
> What are the big issues with adding some basic locking ? being called
> from NMIs ?
That is one possible issue. I know we don't call the error decoding
routines in NMI context on x86 but no clue about ARM.
> If the separate drivers operate on distinct counters I don't see a big
> problem there.
Yes, and they do.
> I'n not sure why <platform_name> ...
Well, however you'd call it. It will be some distinct name I hope :)
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
