I was wrong. It's not the x86 UAPI split, it's the DT pull. More people added.
Grant and Guennadi - it looks like some nasty memory corruption,
because now it didn't cause infinite scrolling, now it caused an oops
in unmap_single_vma() during exit_vmas(). In udevd. Followed by a
hang. I don't know *which* commit it is yet, but commit 4939e27d46fe
("Merge tag 'devicetree-for-linus' ...") is bad, and the previous
merge (the ARM mvebu one is fine).
I'll bisect to the commit, but wanted to let David off the hook, and
put the DT people on the hook. Maybe it's some odd merge artifact,
although I don't think that one had any conflicts at all.
Linus
On Fri, Dec 14, 2012 at 5:16 PM, Linus Torvalds
<torva...@linux-foundation.org> wrote:
>
> Ugh. This doesn't seem to work for me at all. It causes infinite
> scrolling of some text that I have no idea about.
>
> I started bisecting (because I thought it might be something else and
> I hadn't booted after every pull), but by now the only thing I have
> left is ARM and a couple of tiny OF patches .. and the x86 UAPI split.
>
> The split *should* have been safe, since it's mostly a "compile or
> not" thing like Peter said, but we had similar problems on other
> architectures, when things compiled but didn't actually work due to
> missing #define's and #ifdef handling. Things like
> architecture-specific macros that have default versions available when
> the macro is missing etc.
>
> Now, maybe it's some of the other remaining commits after all, but
> from where I am in the bisection it really looks like the uapi patch
> is the most likely culprit. So I thought I'd let people know...
>
> Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
