The network comes up fine, but the errors are annoying: cardmgr[3841]: watching 2 sockets cardmgr[3841]: Card Services release does not match cardmgr[3841]: starting, version is 3.1.33 cardmgr[3841]: socket 0: Lucent Technologies Wavelan/IEEE cardmgr[3841]: executing: 'modprobe hermer' cardmgr[3841]: executing: 'modprobe orinoco_cs' cardmgr[3841]: executing: './network start eth1' cardmgr[3841]: + stty: standard input: Invalid argument cardmgr[3841]: + 7G [ ok ]g eth1 up... cardmgr[3841]: + stty: standard input: Invalid argument cardmgr[3841]: exiting
yep. Once Azarah's latest init scripts are more widely distributed, I will switch over to the quiet init scripts, so the control codes don't make cardmgr spew.
I'd like to add to this, There are also some error messages on shutdown pertaining to the scripts either not running ./network stop eth1 or the start-stop-daemon suddenly just killing the pid of cardmgr. I get some dhcpStop errors to the tune of something like: SIOCTLADDR: inappropriate ioctl to device... two of them to be more exact. I also get a message regarding the ./network script.. showing it's options start stop restart reload.. meaning somewhere in some script it's calling ./network with no options, or invalid options.... otherwise it wouldn't show the options per running /etc/pcmcia/network with no options. Blargh..... P.S those stty input errors on bootup are frustrating.. being how I see gentoo as a more 'perfectionist' os.. i hate seeing things like that.
If you setup your logger to load before pcmcia, i.e., in the boot runlevel, then you won't see those errors. They will still go into your log, though.
They shouldn't throw out errors like that in the first place... What about the shutdown errors? And how would I start logger before pcmcia?
I am not sure about the shutdown errors, but I get them too. To make the logger start before pcmcia, I suggest that you do a rc-update del [loggername] rc-update add [loggername] boot to add the logger to the boot runlevel. If you don't want to do that for some reason, then you will need to edit /etc/init.d/pcmcia and add "before [loggername]" to the depends() function, but realize that, in general, editing init scripts is not a good idea.
Hi, I'm adding this here, as opposed to a new entry, due to Ryan Wheeler's comments with regard to cardctl's execution of /etc/pcmcia/network resulting in errors. I WAS experiencing this too and discovered it was a result of the 'stab' file not existing on the filesystem (sys-apps/baselayout-1.7.9-r1, sys-apps/pcmcia-cs-3.1.33-r3). This becomes a serious problem when attempting to use 'cardctl' from userland, and also for cardmgr when executing any of the action scripts in /etc/pcmcia/. So, long story short, I modified /etc/conf.d/pcmcia entry CARDMGR_OPTS to be empty. The -f flag was causing cardmgr to not create the stab file, (according to the changelog this option was removed -- however it appears in files/pcmcia.conf. Also I've not tested this with dhcp). Further to the stab location: it's actually not necessary to specify /var/run on the command line with the -s switch since it's the alternate location for the stab file according to stab(5). hth. Thanks for the pcmcia reworking to use gentoo init process.
I think the stty error is due to the fact that cardmgr runs without a tty attached. As a result the stty program returns an error. The stty program is used to compute the last colum in order to display the [ ok ] or [ !! ] at the right of the screen (in /sbin/functions.sh). When there is no tty this value is not correctly computed and thus the [ ok*] is written above the previous text ( which gives 7G [ ok ]g eth1 up... ). Maybe it would be wiser to test if there is a tty using /usr/bin/tty.
Now that this bug is 6 months old, does it still exist? I don't think I have seen it in my logs lately, but I could be wrong. I am aware of the init script trying to shut down networking twice on system shutdown, but that is a different issue.
I don't see this anymore.
Okay, I'm closing this thing.