Bug#482100: linux-image-2.6.25-3-686-bigmem : Complete system freeze
On Tue, Jul 01, 2008 at 01:57:10AM +0100, thveillon@gmail.com wrote:
> >
> It's a bit too early to tell, but seems like 2.6.26-rc8-amd64 from
> Debian is working as fine as is the vanilla kernel. It's currently
> running on two clones of the "problematic" system and no freeze so far.
> One is stock kernel just installed from trunk repository, running in
> console mode without x started. The other one has been rebuilt (but with
> strictly the stock 2.6.26-rc8-amd64 configuration file) from trunk
> linux-source-2.6.26-rc8 with gcc 4.3 (Lenny default), and is running
> with x and nvidia's latest beta driver, doing equally good.
> Looks like the problem is strictly a 2.6.25 issue. I would really like
> to understand what's going wrong in the equation Debian Lenny - my
> hardware - 2.6.25 kernel ... I tried once again the 2.6.25 , using the
> working 2.6.26 config to build it, but only to freeze the system after a
> mere 30 minutes. The cpu seems to be idling even though there is
> activity on the box, and never coming up again. Nothing works, keyboard
> (ps2) and "magic" keys are useless, just a message telling "cpu stuck
> for xx seconds" for a while, then even this freezes and I have to
> hard-reboot.
> Exact same system and hardware, I really can't settle for a
> configuration problem or some weird cpufreq governor ill behavior since
> it's working perfectly with 2.6.24 and 2.6.26... and even the vanilla
> 2.6.25 is crashing the same way.
>
> Anyway, I can live with not knowing as long as my system works again and
> I am able to upgrade the kernel when needed.
>
> If you need anything else, or if there's a way I can rebuild a 2.6.25
> with debug on to help gathering information I can do it, but I doubt it
> would be of any interest since we've already moved on to 2.6.26 . Of
> course if 2.6.25 is to be the next Stable default kernel you may come
> across weird problems like mine again...
>
> Best regards,
>
> Tom
Lenny will ship with 2.6.26 as long as it is fixed there and not
happening on 2.6.24 (etch + half kernel) i'm more than happy.
bug will be closed on upload of 2.6.26 too unstable.
thanks for the feedback
--
maks
Reply to: