Simon Geard wrote: > On Sun, 2011-05-01 at 13:17 +1200, Simon Geard wrote: >> On Sun, 2011-05-01 at 07:25 +0800, xinglp wrote: >>> On my X86_64 lfs (SVN-20110427) system . >>> "/dev/disk/by-label/xxx" has not turnd up when "mountfs" started , so >>> I can't use "LABEL=xxxx" in fstab . >>> On 32bits system , this works well. >> Maybe related, I've noticed problems with a 32-bit system using the same >> build, where partitions aren't always appearing in time for boot scripts >> to use them. For example, on one bootup, swapon failed because /dev/sda3 >> (my swap partition) didn't exist at the time. On another, /home didn't >> get mounted, presumably for similar reason. >> >> Anyone else noticed this kind of thing on SVN-20110427 (or thereabouts)? > > Just tried again on a clean build, same thing is happening. Looking at > console output, there's a lot of stuff coming from drivers registering > hardware, even after the udev step has OKed and subsequent steps (like > mounting swap) have run. > > I'd have expected 'udevadm settle' to block while this stuff's still > happening - is that not the case, maybe with 168?
It's not a good option, but perhaps a sleep( 2 ) before or after after '/sbin/udevadm settle' in the udev script would isolate the problem. Do you have a lot of kernel modules or a lot of drivers built in? That may or may not be an issue. It sounds like a race condition between udev and the scripts. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page