In lk 2.5.9 building a kernel with the ide subsystem and
ide-disk built in but no ATA disks oopses around mount
time (of the scsi disk) in the boot sequence.
This worked fine in lk 2.5.8 and lk 2.5.8-dj1 . Going into
the BIOS and disabling the on board IDE chipsets makes the
oops go away:
$ uname -a
Linux frig 2.5.9 #21 Mon Apr 22 20:51:46 EDT 2002 i686 unknown
$ df
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/sda3 8499168 6437568 1629856 80% /
/dev/sda1 31079 16949 12526 58% /boot
Hand decoding the oops gave this stack backtrace:
__ide_end_request
ide_end_request
ide_intr
handle_IRQ
do_IRQ
default_idle
Doug Gilbert
Douglas Gilbert wrote:
> In lk 2.5.9 building a kernel with the ide subsystem and
> ide-disk built in but no ATA disks oopses around mount
> time (of the scsi disk) in the boot sequence.
>
> This worked fine in lk 2.5.8 and lk 2.5.8-dj1 . Going into
> the BIOS and disabling the on board IDE chipsets makes the
> oops go away:
> $ uname -a
> Linux frig 2.5.9 #21 Mon Apr 22 20:51:46 EDT 2002 i686 unknown
> $ df
> Filesystem 1k-blocks Used Available Use% Mounted on
> /dev/sda3 8499168 6437568 1629856 80% /
> /dev/sda1 31079 16949 12526 58% /boot
>
> Hand decoding the oops gave this stack backtrace:
> __ide_end_request
> ide_end_request
> ide_intr
> handle_IRQ
> do_IRQ
> default_idle
>
I will have to check whatever the tagged command queue get's initialized properl
during module loading. Propably not :-(.