kernel 2.4.26 oops (maybe solved)

kernel 2.4.26 oops (maybe solved)

[Peter Santoro]

After trying many h/w and s/w configurations, I've apparently isolated 
my instability issues to using the following the linux kernel highmem 
options: CONFIG_HIGHMEM4G=y, CONFIG_HIGHMEM=y, CONFIG_HIGHMEMIO=y.  I 
have 1GB ram, so maybe one of my dimms is bad or maybe there's a highmem 
bug in the 2.4.X kernel.

The crashes in my previous emails today were due to using the latest 
alsa modules (loaded, but not used by any application) with a HIGHMEM 
enabled kernel.  I appear to have no problem using alsa when HIGHMEM is 
disabled.  Apparently, I'm not the only one having problems with alsa 
and highmem 
(http://www.mail-archive.com/alsa-user@lists.sourceforge.net/msg13918.html).

I would be willing to work with a kernel developer to better isolate 
this problem and test a patch.

Thank you,


Peter Santoro

Re: kernel 2.4.26 oops (maybe solved)

[Alex Lyashkov]

В Чтв, 22.07.2004, в 06:26, Peter Santoro пишет:
> After trying many h/w and s/w configurations, I've apparently isolated 
> my instability issues to using the following the linux kernel highmem 
> options: CONFIG_HIGHMEM4G=y, CONFIG_HIGHMEM=y, CONFIG_HIGHMEMIO=y.  I 
> have 1GB ram, so maybe one of my dimms is bad or maybe there's a highmem 
> bug in the 2.4.X kernel.
> 
> The crashes in my previous emails today were due to using the latest 
> alsa modules (loaded, but not used by any application) with a HIGHMEM 
> enabled kernel.  I appear to have no problem using alsa when HIGHMEM is 
> disabled.  Apparently, I'm not the only one having problems with alsa 
> and highmem 
> (http://www.mail-archive.com/alsa-user@lists.sourceforge.net/msg13918.html).
> 
> I would be willing to work with a kernel developer to better isolate 
> this problem and test a patch.
> 
> Thank you,
> 
> 
> Peter Santoro
> -
sound like bad work with pte mapping. As possible use pte_*_map without
pte_unmap or same.
 

-- 
Alex Lyashkov <shadow@psoft.net>
PSoft

Re: kernel 2.4.26 oops (maybe solved)

[Marcelo Tosatti]

On Wed, Jul 21, 2004 at 11:26:22PM -0400, Peter Santoro wrote:
> After trying many h/w and s/w configurations, I've apparently isolated 
> my instability issues to using the following the linux kernel highmem 
> options: CONFIG_HIGHMEM4G=y, CONFIG_HIGHMEM=y, CONFIG_HIGHMEMIO=y.  I 
> have 1GB ram, so maybe one of my dimms is bad or maybe there's a highmem 
> bug in the 2.4.X kernel.
> 
> The crashes in my previous emails today were due to using the latest 
> alsa modules (loaded, but not used by any application) with a HIGHMEM 
> enabled kernel.  I appear to have no problem using alsa when HIGHMEM is 
> disabled.  Apparently, I'm not the only one having problems with alsa 
> and highmem 
> (http://www.mail-archive.com/alsa-user@lists.sourceforge.net/msg13918.html).
> 
> I would be willing to work with a kernel developer to better isolate 
> this problem and test a patch.

Hi Peter, 

Care to disable ALSA to see if the problem really only happens
when running ALSA+HIGHMEM ?

Your previous oopses in core VM code seem to be hardware problems
to me, I haven't done any deep analysis though.

Re: kernel 2.4.26 oops (maybe solved)

[Peter Santoro]

Marcelo Tosatti wrote:
> Hi Peter, 
> 
> Care to disable ALSA to see if the problem really only happens
> when running ALSA+HIGHMEM ?
> 
> Your previous oopses in core VM code seem to be hardware problems
> to me, I haven't done any deep analysis though.
> 

Marcelo,

Thank you for responding to my post.  I know how busy you and your 
kernel colleagues are and I truly appreciate all your efforts.  The last 
thing I want to do is waste anyone's time.  (Earlier this year, I 
reported a Lexar usb thumb drive crash using an ibm pc to the usb kernel 
developers and worked with Pete Zaitcev to test his 2.4.27 patch.)

Although my asus system has been very stable for about 4 years, I do 
realize that my current problems could be due to failing hardware. 
Given that I've also experienced a few random reboots over the last 4 
months, I did replace the psu with a new Antec 430W Truepower psu.  I 
also tried replacing my cpu with an Intel P3 600e and have run all 
memtest86 tests a few times without errors.  The power to the room for 
this computer is on a separate circuit, but I suppose I could try using 
a ups.  Or maybe it's time for a new motherboard?

My original crashes started occurring around the time I upgraded to 
2.4.26 and started using HIGHMEM (no alsa) - of course, this could just 
be a coincidence.  Adding alsa simply made the system more unstable.  I 
have been running without a crash for about 5 days now (alsa without 
HIGHMEM) - until today (2 random reboots: alsa without HIGHMEM).

I am now running without HIGHMEM and without alsa, to see if it makes 
any difference.  I will probably also try running memtest86 for a day or 
two more.  If you have any other suggestions, I'd be happy to listen.


Sincerely,


Peter Santoro