Re: oops in :snd_pcm_oss:resample_expand+0x19c/0x1f0

[Jean-Marc Saffroy <saffroy@gmail.com>]

On Sun, 24 Sep 2006, Jean-Marc Saffroy wrote:

> On Sun, 2006-09-24 at 19:01 +0200, Jean-Marc Saffroy wrote:
>> Yes yes, I know the kernel is tainted "P" (courtesy of the infamous
>> Nvidia module), so flame me if you want, but some investigation leads
>> me to think it could be an issue in core kernel modules, so read on if
>> you still want the gory details. ;-)
>
> Please don't post tainted Oops reports to LKML.  It's not a political 
> issue, it's a technical one - a tainted kernel CANNOT be debugged.

Thank you for flaming. ;-) I thought the lines above would lead some 
people to read on and post an enlightening answer to the implicit question 
that followed, so I'll restate it below.

> Reproduce with an untainted kernel and repost.

Ok, I'll try.

In the mean time, maybe someone can try to answer my original question, 
that was: is there a race between snd_pcm_oss_change_params() reallocating 
a buffer and resample_expand() using another (possibly the same)?

Now I hope that my mail client does not taint the question as well. ;-)


Cheers,

-- 
saffroy@gmail.com