Bugzilla – Bug 112820
udev error messages during boot
Last modified: 2005-09-03 14:44:14 UTC
After fresh installation (tip from aj) there are always many messages from udev in /var/log/boot.msg: <3>udev[901]: run_program: exec of program '/usr/sbin/hal.hotplug' failed <3>udev[902]: run_program: exec of program '/usr/sbin/hal.hotplug' failed /usr/sbin/hal.hotplug is present but called befor HAL started up. I thinks this was not happen on Beta2 and Preview. What changed in udev? I attach boot.msg
Created attachment 47460 [details] boot.msg with errors from udev
I just added the file name to the log, so that we see the failing program. the error is from initrd, where we have the same udev rules as in the normal root, but no HAL. It should already be fixed with the new udev package after beta3.
What is the effect of this? Will this break anything? Will something not work?
Is this the fix? Wed Aug 24 15:45:37 CEST 2005 - zoz@suse.de - submitted the wrong set of *rc files last time. Fixed that.
No, that was another problem. The wrong *rc files were not so importatnt for beta3, becaause these are rarely used. But, there are much more programs called that reside in /usr. I will list them. Further i forgot to change the rule for firmware loading, that effects some wlan cards. Will fix that now.
Wrong firmware.sh is fixed and submitted. udev needs these files from /usr /usr/bin/us428control /usr/sbin/alsactl /usr/sbin/hal.hotplug /usr/sbin/openct-control /usr/share/alsa/firmware/usx2yloader/us122fw.ihx /usr/share/alsa/firmware/usx2yloader/us224fw.ihx /usr/share/alsa/firmware/usx2yloader/us428fw.ihx The alsa firmware could go to /lib/firmware. And failing to call hal.hotplug is no problem, since it is then not running anyway. What about the rest. For initrd/initramfs it would be ok to ignore these failures. But it would be better to drop these rules there and get proper error messages if something else does not work, that should work.
Created attachment 47484 [details] comment out binaries in /usr How about this? It comments out the useless binaries in the uder rules in initrd.
Yes, thats fine for initrd. And what about later? boot.udev requires boot.localfs (via boot.scsidev).
*** Bug 112915 has been marked as a duplicate of this bug. ***
I'm not sure, whether this one is a duplicate of bug 112915
I'm afraid that we need to move the whole ALSA stuff to /lib, even if they are stripped from udev rules in mkinitrd. Or, can files under /usr be accessible at the stage after mkinitrd safely? The firmware files are easy to move, but others are a bit tough.
The openct stuff no longer tries to do anything hotplug related. I just submitted a new package that moves the whole thing to HAL.
to comment 12: If there are not many events that match on these rules, then may call a wrapper script, that checks if /usr is available and waits some time if not. to comment 13: So i remove all rules regarding openct.
openct rules removed from 50-udev.rules
*** Bug 112955 has been marked as a duplicate of this bug. ***
*** Bug 113153 has been marked as a duplicate of this bug. ***
*** Bug 113077 has been marked as a duplicate of this bug. ***
mkinitrd changes from comment #7 submitted to STABLE.
*** Bug 113292 has been marked as a duplicate of this bug. ***
I'm done with mkinitrd, so assigning to Kay.
When loading ipw2100 module, it was looking for firmware.hotplug.sh - but that file has now been renamed to udev.firmware.sh I had to create a symlink 'firmware.hotplug.sh -> udev.firmware.sh' and module loads properly now. Is this related to this bug report?
See Comment #05, it should be fixed.
*** Bug 113694 has been marked as a duplicate of this bug. ***
*** Bug 113695 has been marked as a duplicate of this bug. ***
*** Bug 113765 has been marked as a duplicate of this bug. ***
*** Bug 114028 has been marked as a duplicate of this bug. ***
*** Bug 114109 has been marked as a duplicate of this bug. ***
*** Bug 114596 has been marked as a duplicate of this bug. ***
The release versioned 068git20050826 seems to have taken care of most of these, however a few remain: udev[813]: main: action, subsystem or devpath missing udev[859]: run_program: exec of program '/sbin/udev.mount.sh' failed udev[860]: run_program: exec of program '/sbin/udev.mount.sh' failed udev[861]: run_program: exec of program '/sbin/udev.mount.sh' failed udev[864]: run_program: exec of program '/sbin/udev.mount.sh' failed Are these worth their own bug?
OK, the /sbin/udev.mount.sh issues are in bug 114596, mea culpa.
we continue on #115133 *** This bug has been marked as a duplicate of 115133 ***