About a week ago, I blindly (and in retrospect, foolishly) agreed to an automatic update to the kernel on my OpenSUSE 10.2 box. After this update, it appears that my LIRC installation no longer works, meaning my Remote Control is not currently functioning in Linux. From what I can tell, others have had similar problems in the past, having to do with lirc modules that are no longer compatible with the kernel or something. It looks like I might have to uninstall, then download the source, build the modules myself and try to install that way. Ugh... Does anybody have any pointers?

I'm pasting some debug-type info here.

schiller@helium:~> uname -r
2.6.18.8-0.1-default
schiller@helium:~> lsmod | grep lirc
lirc_atiusb 21560 1
lirc_dev 18548 1 lirc_atiusb
usbcore 115024 3 lirc_atiusb,ehci_hcd,ohci_hcd
schiller@helium:~> cat /etc/modprobe.conf.local
#
# please add local extensions to this file
#
alias char-major-61 lirc_atiusb
alias lirc_dev lirc_atiusb

schiller@helium:~> ps -aef | grep lirc
root 2189 2188 0 08:11 ? 00:00:00 /sbin/modprobe -q -- lirc_dev
schiller 8788 7879 0 13:18 pts/5 00:00:00 grep lirc
schiller@helium:~>

I just noticed this page which says the lirc-kmp-default package needs updates. I'll have to see if I have this version or not, but this could very well be my problem.

§346 · March 13, 2007 · Linux, Software, Technology · · [Print]

Leave a Comment to “Lack of Control”

  1. Mauriat says:

    Arrggh!!! Never ever do blind kernel updates unless every module dependancy comes with it. This plagues the Fedora world, and apparently can hit the Suse world to. 😛

    First, check if the ‘lirc_dev’ module matches your kernel version. Typically if they don’t match it won’t even load, but just to be 100% sure:
    # /sbin/modinfo lirc_dev | grep vermagic
    this should match the output of ‘uname -rm’

    Next check if the module can be loaded and unloaded properly:
    if already loaded:
    # /sbin/modprobe -rv lirc_dev
    to load:
    # /sbin/modprobe -v lirc_dev
    ( I would recommend killing that process 2189 before doing the above: ‘killall modprobe’ as root)

    If it loaded correctly, there should be some sort of message in the system log.
    Either run ‘dmesg’ and see the output at the very end or open ‘/var/log/messages’.

    This won’t solve your problem, but it is my general procedure to better debug these problems.
    Hope it helps.

  2. Mauriat,

    Yes, now I know better. Anyway, I removed lirc and built it from scratch, but in the end I had the exact same problem. Then I found out that killing the “modprobe lirc_dev” process and then manually invoking lircd seemed to fix the issue. I’m not 100% certain what to do about that strange modprobe issue yet, but at least I have a stopgap solution until I can dig further into it.

    Jeff