2003-07-16 23:24:21

by Steven Newbury

[permalink] [raw]
Subject: lots of oopses with recent kernels

I am using gcc version 3.3 20030623 (Red Hat Linux 3.3-12)
so that may be to blame...

...but I have received the bellow oopses while copying files to a "Device-Mapper" striped device with
'tar cf - . | tar - -C /mnt/tmp'

oops1: reiserfs format device (with BadRAM patch)
oops2: ext2/3 format device (with BadRAM patch)
oops3: reiserfs format device (with mem=320M BadRAM patch
removed.

Kernel 2.6.0-test1-mm1 (with latest device-mapper patches)
was used with these oopses however I have had a similar
oops though while using device-mapper (I was compiling
glibc) with 2.6.0-test1:

2.6.0-test1.oops

While using 2.6.0-test1-ac1 to attempt the same procedure as described above USB was killed:

uhci-bad

Also attached dmesg and .config used for 2.6.0-test1-mm1

Now having said all of that, I should add that the system runs completely stable under a modified Red Hat 2.4.20-9 (ACPI patch amoung other things) also compiled with the same compiler.

The hardware in this machine is:

Intel P3 (coppermine stepping 10) 933MHz
BIOSTAR M6VBE-A (VIA Apollo Pro 133 based)
384MiB 256+128 (with defect in last 64MiB of 128MiB) 133MHz


-----------------------------------------
Email provided by http://www.ntlhome.com/


Attachments:
oops1 (1.01 kB)
oops2 (1.01 kB)
oops3 (840.00 B)
2.6.0-test1.oops (3.66 kB)
uhci-bad (217.00 B)
dmesg (10.69 kB)
.config (23.77 kB)
Download all attachments

2003-07-17 09:58:47

by Oleg Drokin

[permalink] [raw]
Subject: Re: lots of oopses with recent kernels

Hello!

On Wed, Jul 16, 2003 at 11:35:55PM +0000, [email protected] wrote:

> I am using gcc version 3.3 20030623 (Red Hat Linux 3.3-12)
> so that may be to blame...
> ...but I have received the bellow oopses while copying files to a "Device-Mapper" striped device with
> 'tar cf - . | tar - -C /mnt/tmp'
> oops1: reiserfs format device (with BadRAM patch)
> oops2: ext2/3 format device (with BadRAM patch)
> oops3: reiserfs format device (with mem=320M BadRAM patch
> removed.

Sigh, I see lots of times your kernel jumps at random addresses for no good reason.
(like in 2 out of your four oopses you attached).
Have you tried to compile without CONFIG_PREEMPT just to see if it will be any better?

Bye,
Oleg

2003-07-18 01:04:36

by Steven Newbury

[permalink] [raw]
Subject: Re: Re: lots of oopses with recent kernels


>
> From: Oleg Drokin <[email protected]>
> Date: 2003/07/17 Thu AM 10:13:38 GMT
> To: [email protected]
> CC: [email protected]
> Subject: Re: lots of oopses with recent kernels
>
> Hello!
>
> On Wed, Jul 16, 2003 at 11:35:55PM +0000, [email protected] wrote:
>
> > I am using gcc version 3.3 20030623 (Red Hat Linux 3.3-12)
> > so that may be to blame...
> > ...but I have received the bellow oopses while copying files to a "Device-Mapper" striped device with
> > 'tar cf - . | tar - -C /mnt/tmp'
> > oops1: reiserfs format device (with BadRAM patch)
> > oops2: ext2/3 format device (with BadRAM patch)
> > oops3: reiserfs format device (with mem=320M BadRAM patch
> > removed.
>
> Sigh, I see lots of times your kernel jumps at random addresses for no good reason.
> (like in 2 out of your four oopses you attached).
> Have you tried to compile without CONFIG_PREEMPT just to see if it will be any better?
>
> Bye,
> Oleg
>

Part of the problem is due to a bug in my BadRAM patch though I am unable to find it. :-(

Any chance any of the oopses pointed to anything w.r.t. the BadRAM patch?

I am now running with CONFIG_PREEMT=n and without BadRAM. I don't trust my Promise IDE controller, it might also be causing problems, I am getting IDE errors off of it with the Barracuda's. I have ordered a new one, which should arrive in the morning (Friday). I will let you know how it goes...


-----------------------------------------
Email provided by http://www.ntlhome.com/