1. 24 Apr, 2011 4 commits
  2. 23 Apr, 2011 6 commits
    • Dave Reisner's avatar
      Use Makefile to replace install.sh · a2b48109
      Dave Reisner authored
      If we're going to use a Makefile for minilogd, we might as well use it
      for everything. This has some moving pieces:
      * fix minilogd rule: minilogd.o is a target dep not a build rule
      * rename adjtime.cron => adjtime
      * fixup PKGBUILD to account for changes
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      remount root very early · 5fcb5afe
      Tom Gundersen authored
      This remounts root read-only as early as possible.
      In FS#23845 we saw an example of what might happen if we boot with
      root mounted rw and something manages to opens a file read-write
      on root before we remount read-only.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      create /run/lock · eda191cf
      Tom Gundersen authored
      This allows /var/run and /var/lock to be symlinked to /run and
      /run/lock, respectively. This would be the final step in fixing
      With /run/lock created early in boot it would be possible for tools
      like mdadm and lvm2 to use this directory for their locking, and
      maybe avoid the --sysinit parameter. This would have to be implemented
      in the respective packages though.
      Finally, /run/lock is created 1777 just like /var/lock is now.
      For the future, it might make sense to follow systemd's lead and make
      it 0755, and then create a subdir /run/lock/lockdev owned by root:lock
      with permissions 0755. This would improve security and fend agains DoS
      attacks, but we would have to check carefully what packages are affected.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      bootlogd: log messages to console during early boot · a5bdefd4
      Tom Gundersen authored
      This saves all the output during boot to /var/log/boot, which makes
      debugging boot problems much simpler.
      The logfile contains all the control (color) characters from boot, so
      some post processing is in order before the log can be read in a nice
      format (e.g. less should be able to show this nicely in raw mode).
      Start bootlogd as soon as the required /dev/pts is mounted.
      When boot is finished (either at the end of rc.multi or rc.single)
      create the required lof file (/var/log/boot) if it does not exist. This
      causes bootlogd to flush the log to disk.
      Finally, send SIGTERM to bootlogd and clean up after it.
      Idea-and-original-patch-by: default avatarSøren Poulsen <nikorpoulsen@gmail.com>
      Tested-by: default avatarDavid Goldstein <djg1971@gmail.com>
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      /dev: mount /dev/pts and /dev/shm during early boot · 09c27ea2
      Tom Gundersen authored
      This has two purposes:
      1) it allows boot services like bootlogd (that needs /dev/pts) to be
      started very early in boot, and
      2) by setting the standard options in rc.sysinit rather than in fstab we
      have control over them and can easily integrate them with other packages
      and make adjustments to fix bugs. This brings /dev/{pts,shm} into line
      with how we handle /dev, /proc, /sys and /run.
      With this patch /dev/{pts,shm} can (and probably should) be removed from
      the default fstab file, and fstab should only contain entries for the
      mountpoints the user has specified during install, like /boot, /home and
      /var. It will now be possible to boot a working system with an empty fstab.
      If /dev/{pts,shm} is in fstab the mountoptions there will take precedenc
      over the standard ones.
      Cc: Pierre Schmitz <pierre@archlinux.org>
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Gerardo Exequiel Pozzi's avatar
      umount tmpfs on shutdown if devtmpfs is used · 9ce42645
      Gerardo Exequiel Pozzi authored
      This fixes FS#19783.
      some adjustments:
       - always umount ramfs as we never mount /dev as ramfs; and
       - don't umount /dev/pts as it might still be useful.]
      Signed-off-by: default avatarGerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar>
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
  3. 22 Apr, 2011 1 commit
  4. 21 Apr, 2011 6 commits
  5. 16 Apr, 2011 2 commits
  6. 15 Apr, 2011 2 commits
  7. 13 Apr, 2011 4 commits
  8. 11 Apr, 2011 1 commit
  9. 03 Apr, 2011 6 commits
    • Tom Gundersen's avatar
      depmod: no longer update module dependencies on boot · 42d52ae2
      Tom Gundersen authored
      This should not be necessary as the packages installing modules should
      be responsible for updating the dependencies.
      Furthermore, as modules can be loaded very early in boot, but depmod can
      only be called after root is mounted rw, we can not rely on depmod
      fixing broken module dependencies.
      Fixes: FS#20174.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      dev: create pts and shm directories · db3ba801
      Tom Gundersen authored
      Creates the mountpoints for the standard api filesystems in /etc/fstab.
      At the moment this is already done by /lib/udev/devices in our udev
      package. However, with this change it is possible to boot a fully
      working system with the udev unaltered from upstream.
      For the future, this patch allows us to get rid of the /lib/udev/devices
      directory from the udev package alltogether.
      v2: simplified after suggestion from Dave.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      workaround: do not set property=STARTUP=1 when calling udevadm · c3063d04
      Tom Gundersen authored
      This was needed to properly deal with dm devices being created in
      initrd. However, the lvm2/devicemapper rule files no longer use this
      property and now contain a proper fix.
      By grepping through /lib/udev/rules.d I could not find any other rule
      file using this property. Finally, the startup script submitted to
      systemd by the udev maintainer does not set the property, and there has
      not been any bug complaints about this.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      workaround: do not unset NIS on shutdown · 25c74161
      Tom Gundersen authored
      If this bug is still present (after more than 4 years) it should be
      reported and fixed in syslog-ng.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      workaround: do not rmmod psmouse · 2955415f
      Tom Gundersen authored
      If this bug is still present (after more than 4 years) it should be
      reported and fixed in the kernel.
      Dave points out that this was fixed by kernel commit a1cec06177386
      (which is in 2.6.21 and was backported to at least
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
    • Tom Gundersen's avatar
      mtab: don't write to mtab if it is a symlink · 71a1b74e
      Tom Gundersen authored
      In the future /etc/mtab should be a symlink to /proc/self/mounts, add
      support for this in rc.sysinit to be prepared. While strictly not
      needed, this makes it clear that the functionality can be removed once
      we move over to using symlinks.
      This functionality is expected with the next release of util-linux
      (2.20), but is already available in the current version if the
      experimental libmount support is enabled.
      v2: simplified after suggestion from Seblu.
      Signed-off-by: default avatarTom Gundersen <teg@jklm.no>
  10. 01 Apr, 2011 3 commits
  11. 30 Mar, 2011 3 commits
  12. 29 Mar, 2011 2 commits