Re: CVE request: kernel: timekeeping: Prevent oops when GENERIC_TIME=n

[Josh Bressers <bressers () redhat com>]

Please use CVE-2010-2243 for this.

Thanks.

----- "Eugene Teo" <eugeneteo () kernel sg> wrote:

> "Aaro Koskinen reported an issue in kernel.org bugzilla #15366, where
> on 
> non-GENERIC_TIME systems, accessing 
> /sys/devices/system/clocksource/clocksource0/current_clocksource
> results 
> in an oops.
>
> It seems the timekeeper/clocksource rework missed initializing the 
> curr_clocksource value in the !GENERIC_TIME case."
>
> Upstream commit:
> http://git.kernel.org/linus/ad6759fbf35d104dbf573cd6f4c6784ad6823f7e
>
> I believe distros are using GENERIC_TIME=y, at least this is the case
> in 
> our supported kernels.
>
> Thanks, Eugene
> -- 
> main(i) { putchar(182623909 >> (i-1) * 5&31|!!(i<7)<<6) && main(++i);
> }