lp:ubuntu/oneiric-proposed/udev

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

210. By Herton R. Krzesinski on 2012-07-16

[ Herton Ronaldo Krzesinski ]
* Refresh debian/patches/avoid-exit-deadlock-for-dm_cookie.patch:
  avoid serialization issues in handling of events with DM_COOKIE
  set (LP: #1017715).

[ Eric Miao ]
* fix-race-condition-of-serio-driver-module-not-loading.patch:
  Work around race condition in PS/2 keyboard/mouse handling. This resulted
  in udev sometimes failing to load the psmouse driver, causing touchpads to
  occasionally not work after boot up. (LP: #1014460)

209. By Steve Langasek on 2012-03-05

Add debian/patches/avoid-exit-deadlock-for-dm_cookie.patch,
debian/libudev0.symbols: do not exit across a pending DM_COOKIE
event to avoid vgchange deadlocks, thanks to Herton R. Krzesinski
(LP: #802626).

208. By Steve Langasek on 2011-12-16

[ Andy Whitcroft ]
* debian/udev.initramfs-bottom: ignore timeout errors from udevadm we
  expect (and indeed requested) in certain failure modes. If we do not we
  will bail out early and not transfer /dev into /root which is always
  fatal leading to unbootable machines. (LP: #818177)
* avoid-exit-deadlock-for-timely-events: avoid deadlock when exiting
  by continuing to handle events with timeliness requirements.
  The timeliness requirement will be violated if we ignore them which
  is highly undesirable. Also these events are typically dependant
  events and may well block the events we are waiting on leading to
  boot delays and uninitialised devices. (LP: #842560)
* debian/udev.initramfs-bottom: increase the client-side timeout to
  better cope with potential timeout extension issues in udev. We very
  much would prefer udev to time itself out and guarantee to have
  completed than take action ourselves. Very worst case the timeout may
  be doubled from the default of 60s so increase ours accordingly. Note,
  we should only ever trip this timeout when we are already in severe
  trouble. (LP: #818177)

207. By Steve Langasek on 2011-10-08

* Set a timeout of 61 seconds when calling udevadm control --exit, to
  ensure that udevd always reaches its *own* timeout first. If udevadm
  times out before udevd, it exits non-zero, breaking the move of the /dev
  mount; and if we ignore the udevadm failure then udevd might still be
  running, which is the original issue that the patch for bug #624469
  was trying to address. Raising the timeout doesn't make the boot any
  slower, it just makes sure we let udevd count the 60 seconds instead of
  udevadm. LP: #818177.
* debian/patches/race-condition-on-exit: Process events before signals
  in the worker thread, so we don't sit around waiting 60 seconds for an
  already-dead process.

206. By Andy Whitcroft on 2011-08-23

Include input_id in the initramfs to prevent errors when enumerating
devices in early init (LP: #831516).

205. By Felix Geyer on 2011-08-16

Add debian/patches/move-input_id-to-default-rules.patch: Fix bluetooth
mice and keyboards in the X server. Patch cherry-picked from upstream.
(LP: #827489)

204. By Martin Pitt on 2011-08-15

[ Martin Pitt ]
* New upstream bug fix release.
* debian/rules: Configure with --enable-udev_acl, the default changed in
  this upstream version because of newer systemd versions.
* debian/udev.examples: Drop scsi_id.config, this file has been obsolete for
  a long time.

[ David Henningsson ]
* To enable jack detection in PulseAudio, make jack detection input
  devices world-readable, just as being done for joysticks (LP: #826662).

203. By Colin Watson on 2011-07-28

The kernel won't mount devtmpfs automatically when /init exists, which
is always the case in the installer, so mount it manually (LP: #817443).

202. By Colin Watson on 2011-07-25

debian/udev.installer-startup: Create /dev/pts before mounting it
(LP: #815962).

201. By Martin Pitt on 2011-07-14

* debian/udev-udeb.dirs, debian/udev.dirs, debian/udev.postinst: Get rid of
  the remaining /lib/udev/devices/ stuff:
  - shm: will be symlinked to /run/shm/ by mountall, and avoid having a
    non-world-writable shm (which breaks e. g. ecryptfs)
  - net: created by devtmpfs already
  - pts: mounted as devpts by mountall
* Drop debian/udev.init. It has a ton of deprecated stuff in it, and with
  upstream supporting systemd and us having upstart scripts there won't ever
  be a time when we go back to a sysvinit script.
* debian/udev.installer-startup:
  - Drop copying of /lib/udev/devices/. udevd already does this, and it's
    obsolete with devtmpfs anyway.
  - Drop unconditional mounting of a tmpfs on /dev and mknod'ing stuff. We
    already have a devtmpfs there. Only mount devtmpfs if /dev/null doesn't
    exist already.