[prev in list] [next in list] [prev in thread] [next in thread] 

List:       alsa-devel
Subject:    Re: Crash in acpi_ns_validate_handle triggered by soundwire on Linux 5.10
From:       Marcin_Ślusarz <marcin.slusarz () gmail ! com>
Date:       2021-01-29 20:03:18
Message-ID: CA+GA0_u8NA90GmMSDO-Ejg-C2YEXCn8rnVZpk-_+eXcDm7XRpg () mail ! gmail ! com
[Download RAW message or body]

pt., 29 sty 2021 o 19:59 Marcin =C5=9Alusarz <marcin.slusarz@gmail.com> nap=
isa=C5=82(a):
>
> czw., 28 sty 2021 o 15:32 Marcin =C5=9Alusarz <marcin.slusarz@gmail.com> =
napisa=C5=82(a):
> >
> > czw., 28 sty 2021 o 13:39 Rafael J. Wysocki <rafael@kernel.org> napisa=
=C5=82(a):
> > > The only explanation for that I can think about (and which does not
> > > involve supernatural intervention so to speak) is a stack corruption
> > > occurring between these two calls in sdw_intel_acpi_cb().  IOW,
> > > something scribbles on the handle in the meantime, but ATM I have no
> > > idea what that can be.
> >
> > I tried KASAN but it didn't find anything and kernel actually booted
> > successfully.
>
> I investigated this and it looks like a compiler bug (or something nastie=
r),
> but I can't find where exactly registers get corrupted because if I add p=
rintks
> the corruption seems on the printk side, but if I don't add them it seems
> the value gets corrupted earlier.
(...)
> I'm using gcc 10.2.1 from Debian testing.

Someone on IRC, after hearing only that "gcc miscompiles the kernel",
suggested disabling CONFIG_STACKPROTECTOR_STRONG.
It helped indeed and it matches my observations, so it's quite likely it
is the culprit.

What do we do now?

Marcin
[prev in list] [next in list] [prev in thread] [next in thread]