Bugzilla – Bug 106708
PCMCIA network does not start at system boot, nor at apm resume
Last modified: 2007-06-05 11:20:29 UTC
I have a PCMCIA wireless network card that works just fine if I manually eject and re-insert the card after my laptop has booted. However, as summary says, unlike previous versions of SuSE (and other distributions that used pcmcia's cardmgr), this card and network devices are not automatically started at boot- up. In addition, if I do an apm suspend/resume, the card does not get stopped/ started again, which means that when I suspend, leave work, go home, and resume, I again have to manually eject and re-insert the card to get it to reconnect to the new AP and get an IP address. Please tell me what else I can provide for help in debugging this. Thank you! =:)
For the first problem: Since beta3 we have /etc/sysconfig/pcmcia again. There is a variable PCMCIA_COLDPLUG_REINSERT=no. Set this to yes. I also have one card that need that special treatment; all my other cards don't need that. Please tell me if that helps. For the second problem: Check the configuration possibilities of your power management daemon and let this deaemon call 'pccardctl eject/insert' at suspend/resume.
Hi Christian, Thanks for the quick attention! =:) I will update to beta3 as soon as I have an installation source for it. Am I correct in understanding you as saying that I can't get this fix until beta3 is released? Also, will there be a Yast2 section that allows this workaround to be configured? Or is this going to rely on the user manually editting a config file by hand? I sure hope it's the former, since the latter is not very best- desktop-ish. =:) As to the second problem... I use apm since acpi does not properly suspend and resume my laptop (IBM Thinkpad A31). Is there a place in Yast2 that allows for services/subsystems/additional configuration to take place for when my laptop is suspended/resumed? Or is this another thing that the user will have to figure out on his own and go to the command line for? Thanks again for your help Christian! =:)
Yes, please wait for beta3. You can change that variable in YaST Sysconfig Editor. The suspend/resume issue is a configuration problem, not a bug. Would be nice, to make it easy to configure such stuff, yes. We are working on things like that.
Hi Christian, Okay, I'm running beta3 and I tried setting PCMCIA_COLDPLUG_REINSERT=yes and it does seem to work. But I noticed that I had also enabled NetworkManager and I think that that might have been causing conflicts with the regular networking stuff. So I disabled networkmanager from starting up, re-set PCMCIA_COLDPLUG_REINSERT=no, and my laptop seems to start up correctly now with PCMCIA networking. So, all in all, I don't know which of the variables it is that I've changed that have cause pcmcia networking to now work: - I've switched from acpi to apm (acpi doesn't work on my thinkpad A31) - I've upgraded from beta1 to beta3 (maybe something in the base system has been fixed since then?) - I've turned on and off networkmanager At any rate, thank you very much for your help. I will mark this as FIXED. =:)
Blah. Christian, I'm not sure if I should reopen this or not. I guess I will since it's still not working. My PCMCIA network card does correctly get started at boot time now with beta3, but it does not eject/insert on suspend/resume-- even after I have these set: ## Path: System/Powermanagement/apmd/Suspend actions ## Description: Some actions that may be executed before suspend ## Type: yesno ## Default: no # If PCMCIA is compiled with APM-Support, cards are normally suspended # before your system suspends. If you don't have APM-Support in PCMCIA, # you can let apmd do this job. The SuSE PCMCIA package does note have # APM support, because it is not reliable. Set it here, if you need it. APMD_PCMCIA_SUSPEND_ON_SUSPEND="yes" ## Path: System/Powermanagement/apmd/Suspend actions ## Type: yesno ## Default: no # PCMCIA cards can be more or less amenable to an APM suspend event. # If you have a card that cannot be suspended properly (such as a SCSI # card), then it should be "ejected" before entering suspend mode. The # cards are not physically ejected; rather, the power is turned off to # them via the "cardctl eject" command, and is reactivated upon resume. APMD_PCMCIA_EJECT_ON_SUSPEND="yes" I notice that the comment says "cardctl eject", while that command doesn't exist--it's now pccardctl. Not sure if that makes a difference, nor can I see where this setting is actually looked for/used. Also, you mentioned that this is a configuration problem. Can you please tell me where I should manually add the pccardctl eject and pccardctl insert lines if I need to do it myself? I would think that that is the purpose of the above 2 configuration variables. Thanks!!
i'll take this one
first question: is there a reason why you are using apmd instead of powersaved?
Sorry--I mis-represented this. I'm not using apmd. powersaved is running. What I have had to do is add "acpi=off" to my kernel line. My laptop (IBM Thinkpad A31) does not successfully suspend/resume with ACPI, unfortunately. But you're right--I'm not using apmd, but rather the default powersaved. Sorry about the confusion. =:)
ok. This is why the above settings do nothing for you. Unfortunately there is not much you can do out of the box to fix this with powersaved, other than pulling out the card and reinserting it after resume (i would even pull it out before suspend and reinsert it after resume - some of the pccard drivers are very old and buggy wrt suspend). I will fix this after 10.0, so you get a "EJECT_PCMCIA" variable in sysconfig/powersave/sleep or something like that, but it probably won't make it into 10.0 :-( If you really want to automate this, you can have a look at custom event scripts as described in http://forge.novell.com/modules/xfmod/cvs/cvsbrowse.php/powersave/powersave/docs/README.events?rev=HEAD&content-type=text/vnd.viewcvs-markup and http://forge.novell.com/modules/xfmod/cvs/cvsbrowse.php/powersave/powersave/docs/README.scripts?rev=HEAD&content-type=text/vnd.viewcvs-markup and in the /usr/share/doc/packages/powersave/contrib directory and create a script that ejects the card before suspend and another script that reinserts it on resume, but this is clearly more work than just pulling out the card and putting it back in ;-)
i have added SUSPEND2DISK_EJECT_PCMCIA="" SUSPEND2RAM_EJECT_PCMCIA="" STANDBY_EJECT_PCMCIA="" for the next beta
Excellent, thank you! This will be in RC1 then?
yes, but the default is "no", so you have to set it to "yes" manually in /etc/sysconfig/powersave/sleep (if it was enabled by default, i would change well-tested behaviour - and you were the only one reporting this problem, so i do not want to break it for everybody else :-)
Excellent. I've just tried this with RC1 and confirmed that it does seem to work by using SUSPEND2RAM_EJECT_PCMCIA. I've not tried suspending, switching AP's and then resuming (suspend, go home, resume), but here's to hoping that doesn't cause this to not work. =:) Thanks for your help Stefan!! =:) Very well done! =:)