2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

[Jos Hulzink]

[-- Attachment #1: Type: text/plain, Size: 507 bytes --]

Hi,

Here the early oops I'm stuck with in 2.5.65. Had to write it down manually, 
so the report is not complete. If I missed something crucial, please ask. The 
bug is easy to reproduce, so I can write anything down you need.

exact oops message scrolled away, console isn't set up yet, so I can't scroll 
back.

EIP at current_kernel_time +0x0f/0x40

called from:

sysfs_new_inode
sysfs_fill_super
get_sb_single
do_kern_mount
sysfs_fill_super
rest_init
kern_mount
rest_init

My config is attached.

Jos



[-- Attachment #2: config.gz --]
[-- Type: application/x-gzip, Size: 5472 bytes --]

Re: 2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

[Andrew Morton]

Jos Hulzink <josh@stack.nl> wrote:
>
> Hi,
> 
> Here the early oops I'm stuck with in 2.5.65. Had to write it down manually, 
> so the report is not complete. If I missed something crucial, please ask. The 
> bug is easy to reproduce, so I can write anything down you need.
> 
> exact oops message scrolled away, console isn't set up yet, so I can't scroll 
> back.
> 
> EIP at current_kernel_time +0x0f/0x40
> 

That might be an lfence instruction.  I suspect you've chosen the wrong
CPU type?

Re: 2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

[Jos Hulzink]

On Saturday 22 March 2003 13:17, Andrew Morton wrote:
> Jos Hulzink <josh@stack.nl> wrote:
> > Hi,
> >
> > Here the early oops I'm stuck with in 2.5.65. Had to write it down
> > manually, so the report is not complete. If I missed something crucial,
> > please ask. The bug is easy to reproduce, so I can write anything down
> > you need.
> >
> > exact oops message scrolled away, console isn't set up yet, so I can't
> > scroll back.
> >
> > EIP at current_kernel_time +0x0f/0x40
>
> That might be an lfence instruction.  I suspect you've chosen the wrong
> CPU type?

Gee, you guys are good. Now I'm only left wondering how this setting could be 
reset to PIV, while I'm sure I set it to PII (as you can see from earlier 
postings, I managed to get it running till a SCSI lockup caused by probably 
IRQ prioblems). For now I'll not blame the kernel and tell myself I'm stupid 
;-)

Jos

Re: 2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

[William Lee Irwin III]

On Sat, Mar 22, 2003 at 01:34:41PM +0100, Jos Hulzink wrote:
> Gee, you guys are good. Now I'm only left wondering how this setting could be 
> reset to PIV, while I'm sure I set it to PII (as you can see from earlier 
> postings, I managed to get it running till a SCSI lockup caused by probably 
> IRQ prioblems). For now I'll not blame the kernel and tell myself I'm stupid 
> ;-)

Which SCSI? I've been seeing numerous aic7xxx badnesses ca. whatever bk
snapshot 2.5.65-mm2 was based on. Still looking for the use-after-free...


-- wli

Re: 2.5.65: oops: EIP at current_kernel_time +0x0f/0x40

[Jos Hulzink]

On Saturday 22 March 2003 13:46, William Lee Irwin III wrote:
> On Sat, Mar 22, 2003 at 01:34:41PM +0100, Jos Hulzink wrote:
> > Gee, you guys are good. Now I'm only left wondering how this setting
> > could be reset to PIV, while I'm sure I set it to PII (as you can see
> > from earlier postings, I managed to get it running till a SCSI lockup
> > caused by probably IRQ prioblems). For now I'll not blame the kernel and
> > tell myself I'm stupid ;-)
>
> Which SCSI? I've been seeing numerous aic7xxx badnesses ca. whatever bk
> snapshot 2.5.65-mm2 was based on. Still looking for the use-after-free...

00:06.0 SCSI storage controller: Adaptec AIC-7880U

Onboard SCSI controller, MPS 1.4 enabled Intel 440 LX chipset (i.e. IRQ 
rerouting and stuff) Kernel is recompiled now (processor type change takes a 
while, for you need to recompile everything from scatch) to test the 
different IRQ settings, have dealt with problems like these before.

Note that I'm testing the official,plain Linus-2.5.65

Jos