-
udev (079-0ubuntu35.1) dapper-security; urgency=low
* SECURITY UPDATE: root privilege escalation via udev event spoofing.
- Add debian/patches/81-netlink-owner-check.patch: upstream
fixes.
-- Kees Cook <email address hidden> Wed, 08 Apr 2009 17:06:57 -0700
-
udev (079-0ubuntu35) dapper-proposed; urgency=low
* Include firmware_helper in initramfs
- LP: #74004
-- Ben Collins <email address hidden> Wed, 17 Oct 2007 10:22:11 -0400
-
udev (079-0ubuntu34) dapper; urgency=low
* Don't write the empty string to /proc/sys/kernel/hotplug if it doesn't
exist (ie. the kernel wasn't compiled with CONFIG_HOTPLUG). We don't
use or need this for anything, we were just being tidy.
-- Scott James Remnant <email address hidden> Mon, 22 May 2006 11:07:09 +0100
-
udev (079-0ubuntu33) dapper; urgency=low
* Only attempt to rename network interfaces that have an associated
driver, e.g. not bridges and vlans.
-- Scott James Remnant <email address hidden> Tue, 16 May 2006 17:24:50 +0100
-
udev (079-0ubuntu32) dapper; urgency=low
* Add MINKVER="2.6.15" to our hook script, so mkinitramfs will refuse
to include our new udev in an old kernel's initramfs. Depend on a
new enough version of initramfs-tools so that this hack works.
-- Adam Conrad <email address hidden> Tue, 16 May 2006 21:13:21 +1000
-
udev (079-0ubuntu31) dapper; urgency=low
* Add vbi symlink for v4l* application that otherwise will complain about
it and makes the X world more happy (including the maintainer that can
watch^Wtest TV):
- Update debian/60-symlinks.rules.
-- Fabio M. Di Nitto <email address hidden> Fri, 12 May 2006 12:22:07 +0200
-
udev (079-0ubuntu30) dapper; urgency=low
* Drop the /etc/iftab fixing code, there are just too many little problems
with doing this. If after a breezy->dapper upgrade your network devices
are incorrectly named, fix /etc/iftab manually. We'll work on a better
solution later.
-- Scott James Remnant <email address hidden> Mon, 8 May 2006 14:17:26 +0100
-
udev (079-0ubuntu29) dapper; urgency=low
* Add the address of any UP interface to /etc/iftab if not already
present to ensure they don't change names either.
-- Scott James Remnant <email address hidden> Fri, 5 May 2006 18:08:22 +0100
-
udev (079-0ubuntu28) dapper; urgency=low
* Ignore device-mapper devices and let LVM take care of creating them.
-- Scott James Remnant <email address hidden> Fri, 28 Apr 2006 15:42:37 +0100
-
udev (079-0ubuntu27) dapper; urgency=low
* Improve the iftab helper somewhat:
- add support for driver, bus and businfo selectors
- add support for any sysfs attribute with a SYSFS{...} selector
* On package upgrade fix any /etc/iftab entries that appear to be incorrect,
this is a rather simple fix but should work for the majority of people.
-- Scott James Remnant <email address hidden> Mon, 24 Apr 2006 15:48:04 +0100
-
udev (079-0ubuntu26) dapper; urgency=low
* Add missing conflicts on ifrename; all of the work done by that tool
is now performed by udev itself.
-- Scott James Remnant <email address hidden> Wed, 19 Apr 2006 09:26:50 +0100
-
udev (079-0ubuntu25) dapper; urgency=low
* Rework Andrey Borzenkov's patches applied to upstream udev into
75-format-run-rules-later.patch; this causes any formatting added to
RUN rules to be applied after the NAME has been changed, and fixes
Ubuntu: #38787.
-- Scott James Remnant <email address hidden> Tue, 18 Apr 2006 08:47:14 +0100
-
udev (079-0ubuntu24) dapper; urgency=low
* Fixed typo (PREREQS -> PREREQ) in initramfs hook, purely cosmetic
and didn't cause any problems. Ubuntu: #36962.
* Provide /dev/video and /dev/radio symlinks for compatibility with
older software. Ubuntu: #36860
- as with all /dev symlinks, there's no guarantee which device they'll
point to if you have multiple devices. This is not considered a bug,
software should be modified to use HAL and present a list of available
devices rather than rely on symlinks.
* Changed udevplug to write "add" to the uevent file, instead of a silly
string; future kernel versions may support multiple commands.
-- Scott James Remnant <email address hidden> Mon, 3 Apr 2006 16:23:37 +0100
-
udev (079-0ubuntu23) dapper; urgency=low
* Install udev.initramfs-local into casper-premount so we get to see CD
drives, etc on the live CD.
-- Tollef Fog Heen <email address hidden> Mon, 27 Mar 2006 20:44:32 +0200
-
udev (079-0ubuntu22) dapper; urgency=low
"He takes his food with a horrid zest,
he eats one half and he wears the rest!"
* Split the initramfs script out even more, instead of a case statement
in init-premount move the module loading to local-top and nfs-top.
This fixes the long-standing bug where changing boot= on the kernel
command line has no effect.
* Drop the "|| true"s from usplash_write calls, infinity says they're
unnecessary.
* Actually load modules for intelligent devices like we say we do in the
comments.
-- Scott James Remnant <email address hidden> Thu, 23 Mar 2006 17:44:13 +0000
-
udev (079-0ubuntu21) dapper; urgency=low
* Split the loop-and-wait part of the initramfs script out; it'll
go in the initramfs local script itself. Ubuntu: #36010
-- Scott James Remnant <email address hidden> Wed, 22 Mar 2006 15:33:12 +0000
-
udev (079-0ubuntu20) dapper; urgency=low
"I'm just the fastest thing you'll ever see.
That streak of lightning you just missed was me."
* Load ide-generic in all cases after loading PCI storage controller
drivers. This reduces the complexity of the root filesystem mounting
to just a single path for both IDE and everything else and opens up to
using a UUID for everything.
* Set a global usplash timeout for the entire initramfs script.
Ubuntu: #35104.
-- Scott James Remnant <email address hidden> Tue, 21 Mar 2006 19:47:49 +0000
-
udev (079-0ubuntu19) dapper; urgency=low
"For I am ordinary, unimportant, and undeserving of such attention."
* Place joysticks in the "plugdev" group, for programs that access the
joystick device directly and not through X. Ubuntu: #33354.
* The storage device enumerator used by the installer was a bit too
efficient; it would generate new symlink names every time an add event
for a device was received. The installer causes many add events to occur,
each time hw-detect and the partitioner is run. Fix the script to
return an existing symlink name if one exists in the udevdb.
Ubuntu: #27926.
* Fix iftab_helper segmentation fault when /etc/iftab doesn't exist,
caused by returning NULL rather than the suggested name; now returns
whatever name the kernel gave. Ubuntu: #34078.
-- Scott James Remnant <email address hidden> Wed, 15 Mar 2006 18:28:39 +0000
-
udev (079-0ubuntu18) dapper; urgency=low
"Just a little change; Small, to say the least;
Both a little scared; Neither one prepared."
* Fix ide_media helper to construct IDE device names above 'hdf'
properly. Ubuntu: #32679.
* Hard-code the location of usplash_write and call directly, as
we don't have "type" in busybox. This corrects usplash timing out
early while waiting for SCSI root filesystem to show up. Ubuntu: #28357.
* Set the all_partitions options for removable IOMEGA_ZIP* devices.
Ubuntu: #27752.
-- Scott James Remnant <email address hidden> Thu, 2 Mar 2006 13:36:49 +0100
-
udev (079-0ubuntu17) dapper; urgency=low
"The greatest thing you'll ever learn, is just to be loved, and be loved
in return."
* Fix build failure caused by accidental double-inclusion of the
SORECVBUFFORCE patch.
* As the ifrename patch generally seems to be working, drop the "_clashed"
moniker and instead just use the next free interface number available to
us. So instead of eth0_clashed, you'll see e.g. eth2 now. If you wanted
it as eth0, you need to change your /etc/iftab file.
-- Scott James Remnant <email address hidden> Wed, 1 Mar 2006 13:06:47 +0100
-
udev (079-0ubuntu16) dapper; urgency=low
"You better shape up, 'cause I need a man, and my heart is set on you."
* Define the SORCVUBUFFORCE socket option correctly on Sparc (and other
ancient platforms).
-- Scott James Remnant <email address hidden> Wed, 1 Mar 2006 11:23:41 +0100
-
udev (079-0ubuntu15) dapper; urgency=low
"Rembrandts! El Grecos! Toulouse-Letrec-os!
Painted last week on the banks of the Thames!"
* We also need to "catch up" events of the scsi_device, usb_device and
usb_host classes just in case the user compiled things into their kernel
or put their SATA driver into /etc/mkinitramfs/modules (yes, you,
crispin).
-- Scott James Remnant <email address hidden> Mon, 27 Feb 2006 14:05:25 +0100
-
udev (079-0ubuntu14) dapper; urgency=low
"With tuppence for paper and strings, you can have your own set of wings."
* Log the results of performing the udevplug to /dev/.udev.log by using
udevmonitor. This log file will be moved to /var/log/udev later in the
boot sequence. Note that this won't work if /usr is on a separate
filesystem because of the location of udevmonitor, but that's a small
loss.
-- Scott James Remnant <email address hidden> Tue, 21 Feb 2006 16:22:33 +0000
-
udev (079-0ubuntu13) dapper; urgency=low
"But when I'm way up here, it's crystal clear;
That now i'm in a whole new world."
* Remove the blanket loading of ide-generic after any IDE device, by
this time they should have been claimed by a driver already. I kept
this around for historical reasons, but it breaks people with multiple
IDE controllers in their machine (hi, Fabio!). If this causes new bugs,
it can go back in.
-- Scott James Remnant <email address hidden> Mon, 20 Feb 2006 11:01:23 +0000
-
udev (079-0ubuntu12) dapper; urgency=low
"Should we blame the government? Or blame society?
Or should we blame the images on TV? No, Blame Canada!"
* Drink coffee.
* Read 80-extras-iftab_helper.patch again.
* Stare in disbelief.
* Drink more coffee.
* Add "&& !found" to the "rename to clashed" if condition.
* Close bug #31188.
* Hide in shame.
-- Scott James Remnant <email address hidden> Tue, 14 Feb 2006 11:28:17 +0000
-
udev (079-0ubuntu11) dapper; urgency=low
"How do you document real life when real life's getting more like
fiction each day?
* More updates to ifrename support:
- iftab_helper now supports the common "mac" selector, used to
distinguish different interfaces created by the same driver.
-- Scott James Remnant <email address hidden> Mon, 13 Feb 2006 21:19:19 +0000
-
udev (079-0ubuntu10) dapper; urgency=low
"Near, far, in our motor car; Oh what a happy time we'll spend."
* Fixes to a couple of think-os in the previous ifrename patch:
- after renaming to the temporary interface name, make sure we copy
that into ifr_name otherwise we'll be trying to rename the interface
we've just renamed. Ubuntu #31040.
- use "OLDNAME_ifrename" as the temporary name rather than "__NEWNAME"
to avoid clashes and just generally make it more obvious.
- use "NAME_clashed" as the iftab clash name rather than "_NAME" to make
it more obvious
- don't write an initial /etc/iftab on install or upgrade; leave it to
the installer.
-- Scott James Remnant <email address hidden> Fri, 10 Feb 2006 14:37:02 +0000
-
udev (079-0ubuntu9) dapper; urgency=low
"Dancing freaks and sycophants, Westwood heels and shoulder pads."
* Rework the initramfs premount script a little to try and make the root
filesystem order a little more predictable. In particular make sure we
don't load USB/firewire/etc. drivers until we've loaded IDE and SCSI
drivers otherwise sda could be won by a fast USB key over your SATA
root filesystem. Ubuntu #28150, #30418.
* Include support for renaming of network interfaces. This is done with
a new iftab_helper that can parse a subset of the ifrename /etc/iftab
file and return the expected name of the network interface. This is
called from a udev rule that uses the internal support for network
interface renaming. Interface swapping is supported both by renaming
network interfaces that would clash with one specified in the file and
by waiting for an interface name to become free in udev itself.
Ubuntu: #7050, #30240.
-- Scott James Remnant <email address hidden> Thu, 9 Feb 2006 22:04:10 +0000
-
udev (079-0ubuntu8) dapper; urgency=low
"Forget about your algebra and calculus, You can always do your homework
on the morning bus."
* Change the udev message to "Loading hardware drivers" to stop people
thinking this does any kind of hardware detection, which it doesn't.
-
udev (079-0ubuntu7) dapper; urgency=low
"And someday when you're old and bent, Think of those you might have spent,
With a bad, bad man!"
* Include 65-persistent-disk.rules and helpers in the udeb so the
installer can use /dev/disk/by-*.
-
udev (079-0ubuntu6) dapper; urgency=low
"Some little token of esteem is needed, like a dish of cream."
* Don't check whether a device is claimed by an already loaded driver,
while clever and theoretically time-saving it turns out that various
forms of devices can be claimed by multiple drivers
Ubuntu #28654, #30377.
* Load mmc_block when an mmc device is inserted; will improve once
we get a way to determine exactly which mmc sub-module we need, but
this will give us out-of-the-box SD/MMC card support for now.
* Also load i2o_block for all i2o devices, again this is a temporary
workaround for a bad kernel subsystem.
-- Scott James Remnant <email address hidden> Fri, 3 Feb 2006 17:50:19 +0000
-
udev (079-0ubuntu5) dapper; urgency=low
"We was full of hope, until I got addicted to crack and dope!"
* Switch back to using "cp -a -f" in the init script and installer
startup; that works fine with busybox and with our patched coreutils.
-- Scott James Remnant <email address hidden> Thu, 2 Feb 2006 09:52:00 +0000
-
udev (079-0ubuntu3) dapper; urgency=low
"There's gonna be a quiz at your ascension, not to mention any
threat of hell."
* Wait time for SCSI devices should be 30 seconds, not 3 seconds.
Accidentally broken while fixing usplash timeouts.
-- Scott James Remnant <email address hidden> Tue, 10 Jan 2006 10:54:42 +0000
-
udev (079-0ubuntu2) dapper; urgency=low
"I'm brushing up on looking down."
* Probe for PCI, ISA, PCMCIA, etc. brides, USB and Firewire controllers
and docking stations during initramfs, to support more distant root disks.
* Also probe for input devices so when it all goes wrong, you can use that
USB keyboard to fix it.
* Build the udeb separately without SELinux support, as that's not
required during the installer.
* Don't allow usplash to timeout while we run udevplug in the init script
or while we wait for SCSI devices to get out of bed.
-- Scott James Remnant <email address hidden> Fri, 6 Jan 2006 10:10:01 +0000
-
udev (079-0ubuntu1) dapper; urgency=low
"You know, some guys just can't hold their arsenic."
* New upstream release:
- %e (enumerate) in rules has been deprecated.
- event environment now lists symlinks to device in DEVLINKS variable
- device-mapper support for persistent disk rules
* Dropped 10-udev-conf.patch, now included upstream.
* Dropped 02-no-sepol.patch, as we now depend on that.
* Added 10-selinux-include-udev-h.patch to fix build failure with selinux.
* Increased versioned dependency on libselinux1-dev to that which includes
matchpathcon_init_prefix.
* Fixed dvb device naming to create correct format names, bugs in both
udev rules to match the device and program to generate the right names.
Ubuntu #20874.
* Fixed a bug in udevplug that caused it to wait for an event to complete
that it never tickled. Ubuntu #20943.
* Included firmware loading in the udeb so network cards needing firmware
will work in the installer. Ubuntu #20993
* Plug devices in serial during the initramfs, provides a little more
predictability of device names of the root filesystem.
* Rule changes:
- wait for the address attribute of network devices to appear in sysfs
before processing.
Rationale: solves race not yet fixed in kernel.
- dropped %e from symlink rules; this means you will only have one
/dev/cdrom symlink no matter how many devices you have - and it isn't
necessarily predictable which one it is. Software should use HAL or
similar to present "human names" for devices, and sysadmins should
use the /dev/disk/* names.
Rationale: upstream.
- /etc/udev/rules.d/65-persistent.rules renamed to 65-persistent-disk.rules
Rationale: upstream.
- place /dev/nvram in the nvram group.
Rationale: Ubuntu #21571 and breezy.
- load sg module for all SCSI devices.
Rationale: Ubuntu #12434, SuSE rules.
- place SCSI processors (SYSFS{type}==3) from HP in the scanner group.
Rationale: Ubuntu #12434.
- change permissions of removable devices (floppies, usb, ieee1394, etc.)
to 0660 (the default) instead of overriding it down to 0640.
Rationale: allows formatting of the media, and the groups are intended
to imply physical access anyway.
-- Scott James Remnant <email address hidden> Wed, 4 Jan 2006 07:58:45 +0000
-
udev (077-0ubuntu5) dapper; urgency=low
"And if you're lucky, then the god's a she."
* Spell firewire correctly (with the extra e).
* Spell midi correctly.
* Remove strange "0" floating in the IDE device permissions rules.
* Explicitly state cdrom modes as 0660 in case they've been dropped to
0640 because they're removable.
-- Scott James Remnant <email address hidden> Mon, 12 Dec 2005 13:05:53 +0000
