axnet_cs and udev

Russell King rmk+pcmcia at arm.linux.org.uk
Mon Mar 8 19:25:09 GMT 2004


On Mon, Mar 08, 2004 at 07:52:46PM +0100, Stefan Wolf wrote:
> On Mon, 8 Mar 2004, Russell King wrote:
> > On Mon, Mar 08, 2004 at 04:41:52PM +0100, Stefan Wolf wrote:
> > > But I still have the race condition between cardmgr and an
> > > other service which both are starting my network script.
> >
> > That'll be a race between cardmgr and hotplug.  Hotplug now has the
> > responsibility to bring up network devices, so your PCMCIA network
> > script (/etc/pcmcia/network) should ignore the 'start' action.
> >
> But the stop action isn't handled by hotplug? Why is this so?
> At least in my system hotplug starts the network, but doesn't stop it!

Well, under 2.6, I'm running with this:

lrwxrwxrwx    1 root     root           14 Jul 16  2003 /etc/pcmcia/network -> ../../bin/true

and everything seems happy, but then I did put this test box together
from scratch.

The only reason I didn't suggest doing the same for the 'stop' action
is that the Fedora Core 1 scripts don't do so, neither do the Red Hat
scripts.  Maybe the problem is that 2.4 kernels don't give the
required event... shrug.

I think you may need to do some investigation yourself to discover
whether gentoo is right in keeping their changes in /etc/pcmcia/network
or whether that's a long-standing bug.

-- 
Russell King
 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



More information about the linux-pcmcia mailing list