On Mon, Aug 04, 2003 at 07:01:33PM +0100, Russell King wrote:
> On Mon, Aug 04, 2003 at 10:18:59AM -0700, Jeremy T. Bouse wrote:
> > [1.]
> > Problem with Wireless PCMCIA card insertion.
>
> PCMCIA or CardBus?
>
My bad it is a CardBus bridge using the yenta socket
CONFIG_CARDBUS=y
> > [5.]
> > Jul 28 18:02:10 vaio kernel: airo: Probing for PCI adapters
> > Jul 28 18:02:10 vaio kernel: kobject_register failed for airo (-17)
>
> This looks like an error in airo. For starters, it doesn't unregister
> its PCI driver structure when the module removed, so when it is
> inserted the next time, you get this complaint.
>
> (You can only have one driver called "airo" registered with the device
> model.)
>
I am currently compiling -test2-bk4 to test it out and see if
there is any change... I did notice a patch to airo.c in it... I'll
advise if there is any change good or bad...
>
> Oops, it seems to be scheduling from timer context. That's pretty bad.
> I guess this is the culpret:
>
> static u16 issuecommand(struct airo_info *ai, Cmd *pCmd, Resp *pRsp) {
> ...
> if (!in_interrupt() && (max_tries & 255) == 0)
> schedule();
> ...
> }
>
This must be active only if CONFIG_PREEMPT is set as the atomic
errors did seem to cease after recompiling with CONFIG_PREEMPT not being
set...
Regards,
Jeremy T. Bouse
This archive was generated by hypermail 2b29 : Thu Aug 07 2003 - 22:00:24 EST