Lock-up on PPC64

Benjamin Herrenschmidt benh at kernel.crashing.org
Tue Jan 6 23:05:25 EST 2009


On Mon, 2009-01-05 at 23:28 +1100, Michael Ellerman wrote:
> I'm confused. Which code never exercises which path, and so what
> deserves a better look?

Well, the thing is with the fix that went into 2.6.30, some error path
will never bit hit, or pretty much.

Without the fix, what can happen is that a syscall incorrectly is
interpreted as returning an error, thus triggering rarely used error
path in Mono. However, that shouldn't lockup the whole machine :-)

So It's still worth investigating what's happening when taking the fix
out, as we may have another bug somewhere lurking.

Cheers,
Ben.




More information about the Linuxppc-dev mailing list