Opened 15 years ago
Closed 14 years ago
#2685 closed defect (worksforme)
Does udev populate /dev correctly?
Reported by: | Owned by: | ||
---|---|---|---|
Priority: | normal | Milestone: | 6.7 |
Component: | Bootscripts | Version: | SVN |
Severity: | normal | Keywords: | |
Cc: |
Description
From Andy Benton on lfs-dev:
I haven't updated my bootscripts lately but today I decided to rectify that, I downloaded lfs-bootscripts-20100529 and installed them on a new system. They worked fine apart from the mountfs script which returned a fail.
When I logged in and ran mount -a it complained that /dev/pts and /dev/shm did not exist. It used to be that the udev bootscript copied /lib/udev/devices/{pts,shm} to /dev but looking at it now, the only thing it copies is /lib/udev/devices/null
On the udev page in chapter 6 it still says to install /lib/{firmware,udev/devices/{pts,shm}}. If the udev bootscript doesn't copy /lib/udev/devices/{pts,shm} to /dev, how are /dev/{pts,shm} going to be created?
Change History (4)
comment:1 by , 15 years ago
comment:2 by , 15 years ago
Hi,Andrew Benton
I saw your reply on lfs-dev, I haven't subscribed to lfs-dev mailing list, so I reply here.
Here is what my system shows:
[hongxun@HouLFS-09:17:11 ~]$ ls -l /lib/udev/devices total 12 crw-rw-rw- 1 root root 1, 3 Jun 18 2010 null1 drwxr-xr-x 2 root root 4096 Mar 25 15:23 pts drwxr-xr-x 2 root root 4096 Mar 25 15:23 shm drwxr-xr-x 2 root root 4096 Jun 18 09:04 testdir [hongxun@HouLFS-09:14:08 devices]$ ls -dl /dev/{pts,shm,testdir,null1} crw-rw-rw- 1 root root 1, 3 Jun 18 2010 /dev/null1 drwxr-xr-x 2 root root 0 Jun 18 2010 /dev/pts drwxrwxrwt 2 root root 40 Jun 18 09:11 /dev/shm drwxr-xr-x 2 root root 40 Jun 18 2010 /dev/testdir [hongxun@HouLFS-09:26:51 ~]$ diff -u lfs-bootscripts-20100529/lfs/init.d/udev /etc/rc.d/init.d/udev --- lfs-bootscripts-20100529/lfs/init.d/udev 2010-05-29 18:41:33.000000000 +0800 +++ /etc/rc.d/init.d/udev 2010-06-18 09:26:51.207638349 +0800 @@ -34,28 +34,26 @@ # made or removed during this boot don't affect the next one. # The reason we don't write to mtab is because we don't ever # want /dev to be unavailable (such as by `umount -a'). - if ! mountpoint /dev > /dev/null; then - mount -n -t tmpfs tmpfs /dev -o mode=755 - fi - if [ ${?} != 0 ]; then - echo_failure - boot_mesg -n "FAILURE:\n\nCannot mount a tmpfs" ${FAILURE} - boot_mesg -n " onto /dev, this system will be halted." - boot_mesg -n "\n\nAfter you press Enter, this system" - boot_mesg -n " will be halted and powered off." - boot_mesg -n "\n\nPress Enter to continue..." ${INFO} - boot_mesg "" ${NORMAL} - read ENTER - /etc/rc.d/init.d/halt stop - fi + #if ! mountpoint /dev > /dev/null; then + # mount -n -t tmpfs tmpfs /dev -o mode=755 + #fi +# if [ ${?} != 0 ]; then +# echo_failure +# boot_mesg -n "FAILURE:\n\nCannot mount a tmpfs" ${FAILURE} +# boot_mesg -n " onto /dev, this system will be halted." +# boot_mesg -n "\n\nAfter you press Enter, this system" +# boot_mesg -n " will be halted and powered off." +# boot_mesg -n "\n\nPress Enter to continue..." ${INFO} +# boot_mesg "" ${NORMAL} +# read ENTER +# /etc/rc.d/init.d/halt stop +# fi # Udev handles uevents itself, so we don't need to have # the kernel call out to any binary in response to them - echo > /proc/sys/kernel/hotplug + #echo > /proc/sys/kernel/hotplug - # Copy the only static device node that Udev >= 155 doesn't - # handle to /dev - cp -a /lib/udev/devices/null /dev + # Copy static device nodes to /dev # Start the udev daemon to continually watch for, and act on, # uevents [hongxun@HouLFS-09:26:53 ~]$ udevd -V 157
It's no doubt that udev daemon will copy the files for you.
Do {pts,shm} exist under your /lib/udev/devices directory?
comment:3 by , 15 years ago
I see the same behaviour as hohoxu_hao115; udev-156 correctly copies the contents of /lib/udev/devices for me, rather than requiring the bootscript to do it. Not sure what's going wrong for you, Andy, but I can't reproduce here. Could you confirm that your /lib/udev/devices contains the following please?
root:/# ls -l /lib/udev/devices/ total 8 crw-rw-rw- 1 root root 1, 3 May 29 10:18 null drwxr-xr-x 2 root root 4096 May 29 10:18 pts drwxr-xr-x 2 root root 4096 May 29 10:18 shm
comment:4 by , 14 years ago
Resolution: | → worksforme |
---|---|
Status: | new → closed |
From udev version 155(if i remember correctly), udev will copy content in /lib/udev/devices directory for us. I think Andy Benton didn't upgrade his udev package.
Further more, if we enable kernel devtmpfs, we even don't need null device exists in /lib/udev/devices directory. And according to what udev changelog said, enabling kernel devtmpfs is recommended by udev developers, in future udev release , it will be mandatory.