Ejecting a CardBus device

Ejecting a CardBus device

[Felipe Alfaro Solana]

Hi All!

How can I tell the CardBus subsystem to eject my CardBus NIC by software
with 2.6.0 kernels? In 2.4 I could use "cardctl eject", but I don't know
how to do the same on 2.6.0-test5-mm4.

I need to eject my CardBus NIC if I want to be able to suspend the
machine using APM. Resuming from APM when the "yenta_socket" and
"pcmcia_core" modules are loaded causes a deadlock in the kernel during
resume, and the machine never comes back completely. Thus, before
suspending, I need to rmmod "pcmcia_core" and "yenta_socket" (well, and
uhci-hcd and the sound modules).

Currently, I need to manually eject the card by pushing the eject button
on the side of the laptop if I want to to rmmod "pcmcia_core" (not doing
so, causes pcmcia_core to complain that it's busy). So, I wonder if
there's a way to tell pcmcia_core to "eject" the CardBus NIC (without
physically ejecting the card from the socket) in order to being able to
rmmod it before trying to suspend.

Thanks!

   Felipe Alfaro Solana

Re: Ejecting a CardBus device

[Russell King]

On Sat, Sep 27, 2003 at 04:00:16AM +0200, Felipe Alfaro Solana wrote:
> How can I tell the CardBus subsystem to eject my CardBus NIC by software
> with 2.6.0 kernels? In 2.4 I could use "cardctl eject", but I don't know
> how to do the same on 2.6.0-test5-mm4.

The same works with 2.6.0-test5.

> I need to eject my CardBus NIC if I want to be able to suspend the
> machine using APM. Resuming from APM when the "yenta_socket" and
> "pcmcia_core" modules are loaded causes a deadlock in the kernel during
> resume, and the machine never comes back completely.

It would be nice to solve this problem.  Since APM suspend with 2.6
kernels seems to be a complete dead loss with my laptop, don't look
to me to diagnose this one.

-- 
Russell King (rmk@arm.linux.org.uk)	http://www.arm.linux.org.uk/personal/
      Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
      maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                      2.6 Serial core

Re: Ejecting a CardBus device

[Felipe Alfaro Solana]

On Sat, 2003-09-27 at 09:28, Russell King wrote:
> On Sat, Sep 27, 2003 at 04:00:16AM +0200, Felipe Alfaro Solana wrote:
> > How can I tell the CardBus subsystem to eject my CardBus NIC by software
> > with 2.6.0 kernels? In 2.4 I could use "cardctl eject", but I don't know
> > how to do the same on 2.6.0-test5-mm4.
> 
> The same works with 2.6.0-test5.

I doesn't seem to work: "cardctl eject" complains that no pcmcia driver
appears in /proc/devices. Any ideas?

Meanwhile, I'll take a look at the source code of cardctl to see if I
can learn how to force the eject by software.

Re: Ejecting a CardBus device

[Russell King]

On Sat, Sep 27, 2003 at 12:18:09PM +0200, Felipe Alfaro Solana wrote:
> On Sat, 2003-09-27 at 09:28, Russell King wrote:
> > On Sat, Sep 27, 2003 at 04:00:16AM +0200, Felipe Alfaro Solana wrote:
> > > How can I tell the CardBus subsystem to eject my CardBus NIC by software
> > > with 2.6.0 kernels? In 2.4 I could use "cardctl eject", but I don't know
> > > how to do the same on 2.6.0-test5-mm4.
> > 
> > The same works with 2.6.0-test5.
> 
> I doesn't seem to work: "cardctl eject" complains that no pcmcia driver
> appears in /proc/devices. Any ideas?

That'll be because ds got unloaded.  Although ds isn't required for
cardbus cards, it does provide the interface to cardctl.

-- 
Russell King (rmk@arm.linux.org.uk)	http://www.arm.linux.org.uk/personal/
      Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
      maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                      2.6 Serial core

Re: Ejecting a CardBus device

[Felipe Alfaro Solana]

On Sat, 2003-09-27 at 12:20, Russell King wrote:
> > 
> > I doesn't seem to work: "cardctl eject" complains that no pcmcia driver
> > appears in /proc/devices. Any ideas?
> 
> That'll be because ds got unloaded.  Although ds isn't required for
> cardbus cards, it does provide the interface to cardctl.

Bingo! Now APM suspend is working flawlessly for the very first time in
my life... Thanks.