Sorry to reply to my own message, I am wondering if it is now OK to just turn user space notification of hotplugging off the way the script originally did, but putting /bin/true as the usermodehelper. I know David Z has put a patch in hal to deal with missing hotplug events, but should the networking scripts actively try to shut them off. -- I don't think that is a solution. I would put this into your BZ entry Harold, but I don't have my bz password with me. Andrew > -----Original Message----- > From: Andrew [mailto:cmkrnl@xxxxxxxxxxxxx] > Sent: Friday, October 22, 2004 04:18 PM > To: 'For testers of Fedora Core development releases' > Subject: Re: Prism54 firmware load failure only during boot > > Hi, > > I just read through your bz. I think that RH should reconsider that patch to the network-functions. There is more than just one agent in the hotplug world. What about the hotplug event that fire while that is running. without normal /sbin/hotplug script directing the flow to the right userspace stuff. I can see lots of possible breakage from the way it is right now. > Here is the patch I am referring to > --- network-functions 2004-08-25 17:52:38.000000000 +0100 > +++ /etc/sysconfig/network-scripts/network-functions 2004-08-25 17:53:52.000000000 +0100 > @@ -179,7 +179,7 @@ > return 2 > fi > HOTPLUG=`cat /proc/sys/kernel/hotplug` > - echo "/bin/true" > /proc/sys/kernel/hotplug > + echo "/etc/hotplug/firmware.agent" > /proc/sys/kernel/hotplug > modprobe $1 > /dev/null 2>&1 || { > echo "$HOTPLUG" > /proc/sys/kernel/hotplug > return 1 > I don't think this is the right thing to do and HMMM I was missing hotplug events getting to HAL. -- I think we might have found where the real issue is -- HAL or anything else can't see hotplug events that hijacked like this patch does. > > Andrew > > Andrew > > -----Original Message----- > > From: Harold Campbell [mailto:hcamp@xxxxxxxxxx] > > Sent: Friday, October 22, 2004 03:38 PM > > To: fedora-test-list@xxxxxxxxxx > > Subject: Re: Prism54 firmware load failure only during boot > > > > On Fri, 2004-10-22 at 09:05, Neil Gierman wrote: > > > > > Now in FC3test3, the prism54 module loads and complains about not > > > being able to load the firmware (the firmware is in the correct place > > > according to /etc/hotplug/firmware.agent). As soon as get a shell I > > > rmmod prism54 and modprobe prism54 and it loads and is able to upload > > > the firmware just fine (which I have now added the rmmod and modprobe > > > to rc.local so I can reboot remotely and still gain access after > > > reboot). I updated all packages (including kernel) that were available > > > from yum and the same symptoms are there. I searched bugzilla and > > > didn't see anything on the prism54. > > > > > > Has anyone seen this, or should it go to bugzilla? > > > > > > > I opened bug #129155 some time ago. It applies to FC2 and a minipci card > > (prism54 chipset). Perhaps the same problem has been extended to FC3 and > > cardbus. > > > > -- > > Harold Campbell <hcamp@xxxxxxxxxx> > > > > -- > > fedora-test-list mailing list > > fedora-test-list@xxxxxxxxxx > > To unsubscribe: > > http://www.redhat.com/mailman/listinfo/fedora-test-list > > > > > > -- > fedora-test-list mailing list > fedora-test-list@xxxxxxxxxx > To unsubscribe: > http://www.redhat.com/mailman/listinfo/fedora-test-list >