Kernel Oops Reports

[Andy Whitcroft]

On Thu, Sep 30, 2010 at 03:15:47PM -0700, Brian Murray wrote:
> With the end goal of trying to consolidate some of the duplicate Oops
> reports I've been poking at Oops bugs some and haven written a few
> scripts for finding the tops Oops reports for example 23 Oops'es are
> about:
> 
> WARNING: at
> /build/buildd/linux-2.6.35/drivers/gpu/drm/radeon/radeon_fence.c:235
> radeon_fence_wait
> 
> Talking with Brad about this he indicated that finding out the hardware
> involved with this bug reports would be quite useful.  So I've something
> like this[1] now:
> 
> WARNING: at
> /build/buildd/linux-2.6.35/drivers/gpu/drm/radeon/radeon_fence.c:235
> 
> This device: ATI Technologies Inc Mobility Radeon HD 3400 Series
> [1002:95c4] (prog-if 00 [VGA controller])
>         LP: #642088 is New, has 0 dupes
>         LP: #620052 is New, has 0 dupes
> 
> With my limited kernel knowledge I'd guess bug 642088 and bug 620052 are
> duplicates.  However, not being a kernel developer I'd like input as to
> what other properties of the bugs should be checked to ensure that they
> are in fact duplicates.  Alternatively, what type of report would you
> like to have to facilitate your consolidation of likely duplicate
> Oops'es.
> 
> [1] The full output of warning results for that Oops is attached

Well that report is pretty handy, perhaps we could use something like
that to feed into our weekly review.  As a general form it looks handy
to me as a consumer, obviously for other types of oops we need different
types of hardware information.  It might make sense to generalise the
report so it brings base system information and graphics information.

JFo, perhaps we should try reviewing a couple of these monday.

-apw