Repeatable hang on boot with PCMCIA card present

Repeatable hang on boot with PCMCIA card present

[Steve Smith]

[I sent the following to the person responsible for the patch but
haven't heard anything so I assume he's unavailable...]

Hi,

With recent kernel releases I have started seeing consistent hangs
during boot when a PCMCIA card is present in the slot (the card in
question is a Linksys wireless-B card).  The symptoms are:

    If the card is present during boot an error of "Unknown interrupt
    or fault at EIP ..." appears.

    If the card is not present there is no error.

    The card can be plugged-in post-boot without problems.

Using git-bisect I have narrowed down the error to one commit, namely
"use bitfield instead of p_state and state".  The commit# is

    e2d4096365e06b9a3799afbadc28b4519c0b3526

However I am still seeing this problem with the latest -RC releases.

Additional information:

   Machine: Dell Inspiron 8100
   Distribution: Ubuntu Dapper

Please let me know if you need any more information.

Regards,
Steve Smith

Re: Repeatable hang on boot with PCMCIA card present

[Andrew Morton]

On Sat, 16 Sep 2006 15:03:31 +1000
tarka@internode.on.net (Steve Smith) wrote:

> [I sent the following to the person responsible for the patch but
> haven't heard anything so I assume he's unavailable...]
> 
> Hi,
> 
> With recent kernel releases I have started seeing consistent hangs
> during boot when a PCMCIA card is present in the slot (the card in
> question is a Linksys wireless-B card).  The symptoms are:
> 
>     If the card is present during boot an error of "Unknown interrupt
>     or fault at EIP ..." appears.
> 
>     If the card is not present there is no error.
> 
>     The card can be plugged-in post-boot without problems.
> 
> Using git-bisect I have narrowed down the error to one commit, namely
> "use bitfield instead of p_state and state".  The commit# is
> 
>     e2d4096365e06b9a3799afbadc28b4519c0b3526
>
> However I am still seeing this problem with the latest -RC releases.

Thanks for doing that.

Damn, that was a huge patch.  Have you been able to grab
a copy of the oops output?  It would really help.  Even a photo of
the screen..

Re: Repeatable hang on boot with PCMCIA card present

[Steve Smith]

On Mon, Sep 18, 2006 at 07:09:02PM -0700, Andrew Morton wrote:
> Damn, that was a huge patch.  Have you been able to grab
> a copy of the oops output?  It would really help.  Even a photo of
> the screen..

No oops I'm afraid, just the hang and EIP message.  The numbers
with the message change each time but I can send them when I get to
the machine later.  Is there another way to coax more information out
of the kernel?

Thanks,
Steve

Re: Repeatable hang on boot with PCMCIA card present

[Andrew Morton]

On Tue, 19 Sep 2006 12:25:41 +1000
tarka@internode.on.net (Steve Smith) wrote:

> On Mon, Sep 18, 2006 at 07:09:02PM -0700, Andrew Morton wrote:
> > Damn, that was a huge patch.  Have you been able to grab
> > a copy of the oops output?  It would really help.  Even a photo of
> > the screen..
> 
> No oops I'm afraid, just the hang and EIP message.  The numbers
> with the message change each time but I can send them when I get to
> the machine later.

hm, ugly.

>  Is there another way to coax more information out
> of the kernel?

I don't expect so - if it doesn't print anything useful to the console
we're rather stuck.  You probably don't have a serial port.  If you have a
lan port then netconsole _might_ work, but I'd say it's unlikely.

Dunno.  You _might_ be able to get a better trace by changing settings such
as CONFIG_FRAME_POINTER, CONFIG_4KSTACKS, CONFIG_UNWIND_INFO, and
CONFIG_DEBUG_*.

Re: Repeatable hang on boot with PCMCIA card present

[Dmitry Torokhov]

On Monday 18 September 2006 22:09, Andrew Morton wrote:
> On Sat, 16 Sep 2006 15:03:31 +1000
> tarka@internode.on.net (Steve Smith) wrote:
> 
> > [I sent the following to the person responsible for the patch but
> > haven't heard anything so I assume he's unavailable...]
> > 
> > Hi,
> > 
> > With recent kernel releases I have started seeing consistent hangs
> > during boot when a PCMCIA card is present in the slot (the card in
> > question is a Linksys wireless-B card).  The symptoms are:
> > 
> >     If the card is present during boot an error of "Unknown interrupt
> >     or fault at EIP ..." appears.
> > 
> >     If the card is not present there is no error.
> > 
> >     The card can be plugged-in post-boot without problems.
> > 
> > Using git-bisect I have narrowed down the error to one commit, namely
> > "use bitfield instead of p_state and state".  The commit# is
> > 
> >     e2d4096365e06b9a3799afbadc28b4519c0b3526
> >
> > However I am still seeing this problem with the latest -RC releases.
> 
> Thanks for doing that.
> 
> Damn, that was a huge patch.  Have you been able to grab
> a copy of the oops output?  It would really help.  Even a photo of
> the screen..
> 

Hmm, not sure why you CCed me unless you remembered I have Inspiron 8100.
What chipset does that Linksys card use? I just tried one of my PCMCIA
cards with orinoco_cs and it booted fine on today's pull from Linus...
 
-- 
Dmitry

Re: Repeatable hang on boot with PCMCIA card present

[Andrew Morton]

On Mon, 18 Sep 2006 23:37:52 -0400
Dmitry Torokhov <dtor@insightbb.com> wrote:

> On Monday 18 September 2006 22:09, Andrew Morton wrote:
> > On Sat, 16 Sep 2006 15:03:31 +1000
> > tarka@internode.on.net (Steve Smith) wrote:
> > 
> > > [I sent the following to the person responsible for the patch but
> > > haven't heard anything so I assume he's unavailable...]
> > > 
> > > Hi,
> > > 
> > > With recent kernel releases I have started seeing consistent hangs
> > > during boot when a PCMCIA card is present in the slot (the card in
> > > question is a Linksys wireless-B card).  The symptoms are:
> > > 
> > >     If the card is present during boot an error of "Unknown interrupt
> > >     or fault at EIP ..." appears.
> > > 
> > >     If the card is not present there is no error.
> > > 
> > >     The card can be plugged-in post-boot without problems.
> > > 
> > > Using git-bisect I have narrowed down the error to one commit, namely
> > > "use bitfield instead of p_state and state".  The commit# is
> > > 
> > >     e2d4096365e06b9a3799afbadc28b4519c0b3526
> > >
> > > However I am still seeing this problem with the latest -RC releases.
> > 
> > Thanks for doing that.
> > 
> > Damn, that was a huge patch.  Have you been able to grab
> > a copy of the oops output?  It would really help.  Even a photo of
> > the screen..
> > 
> 
> Hmm, not sure why you CCed me unless you remembered I have Inspiron 8100.

oop, I get my Dominiks and Dmitrys mixed up for some reason, sorry.

> What chipset does that Linksys card use? I just tried one of my PCMCIA
> cards with orinoco_cs and it booted fine on today's pull from Linus...
>  
> -- 
> Dmitry

Re: Repeatable hang on boot with PCMCIA card present

[Steve Smith]

On Mon, Sep 18, 2006 at 11:37:52PM -0400, Dmitry Torokhov wrote:
> Hmm, not sure why you CCed me unless you remembered I have Inspiron 8100.
> What chipset does that Linksys card use?

It's a WPC11, one of the prism family ones.  It's using the HostAP
driver.

> I just tried one of my PCMCIA cards with orinoco_cs and it booted
> fine on today's pull from Linus...

OK, I'll try it with the orinoco driver when I get a chance.

Cheers,
Steve