Re: 2.6.39-rc5-git2 boot crashs

[Linus Torvalds]

On Tue, May 3, 2011 at 12:08 PM, Ingo Molnar <mingo@xxxxxxx> wrote:
>
> I have no smart ideas straight away - just an observation: i keep testing
> CONFIG_X86_ELAN=y on real hardware, and it's enabled in about 4% of my configs:

So how often do you do more than just boot?

Because this problem doesn't happen at boot-time, the "bad IO" issue
seems to happen only after some real filesystem work (ie the "unzip a
140MB file" etc). In one of the logs, the SATA driver issue ended up
happening four hours after boot (but I think Werner could trigger it
in minutes by doing heavy FS work).

That's why I'd suspect some subtle race or other and/or memory
allocation re-use.

The subject line "boot crashs" is misleading, because that's the
unrelated logfs thing that I have a tentative patch for.

That said, I don't really see why ELAN would be so special.

Werner: can you test without X86_EXTENDED_PLATFORM, but then picking
the i486 config instead? That should approximate the ELAN
configuration pretty closely, at least in things like compiler flags..

                   Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/