5 Final merge of patches/features from the Ubuntu package.
11 Control characters in filesystem label strings are no longer silenty
12 removed, but hex-encoded, to be able to uniquely identify the device
13 by its symlink in /dev/disk/by-label/.
14 If libvolume_id is used by mount(8), LABEL= will work as expected,
15 if slashes or other characters are used in the label string.
17 To test the existence of a file, TEST=="<file>" and TEST!="<file>"
18 can be specified now. The TEST key accepts an optional mode mask
19 TEST{0100}=="<is executable file>".
21 Scsi_id now supports a mode without expecting scsi-specific sysfs
22 entries to allow the extraction of cciss-device persistent properties.
28 In the future, we may see uuid's which are just simple character
29 strings (see the DDF Raid Specification). For that reason vol_id now
30 exports ID_FS_UUID_SAFE, just like ID_FS_LABEL_SAFE. For things like
31 the creation of symlinks, the *_SAFE values ensure, that no control
32 or whitespace characters are used in the filename.
34 Possible users of libvolume_id, please use the volume_id_get_* functions.
35 The public struct will go away in a future release of the library.
41 Removal of useless extras/eventrecorder.sh.
51 The directory multiplexer for dev.d/ and hotplug.d are finally removed
52 from the udev package.
58 Symlinks can have priorities now, the priority is assigned to the device
59 and specified with OPTIONS="link_priority=100". Devices with higher
60 priorities overwrite the symlinks of devices with lower priorities.
61 If the device that currently owns the link, goes away, the symlink
62 will be removed, and recreated, pointing to the next device with the
63 highest actual priority. This should make /dev/disk/by-{label,uuid,id}
64 more reliable, if multiple devices contain the same metadata and overwrite
67 The dasd_id program is removed from the udev tree, and dasdinfo, with the
68 needed rules, are part of the s390-tools now.
70 Please add KERNEL=="[0-9]*:[0-9]*" to the scsi wait-for-sysfs rule,
71 we may get the scsi sysfs mess fixed some day, and this will only catch
72 the devices we are looking for.
74 USB serial numbers for storage devices have the target:lun now appended,
75 to make it possibble to distinguish broken multi-lun devices with all
76 the same SCSI identifiers.
78 Note: The extra "run_directory" which searches and executes stuff in
79 /etc/hotplug.d/ and /etc/dev.d/ is long deprecated, and will be removed
80 with the next release. Make sure, that you don't use it anymore, or
81 provides your own implementation of that inefficient stuff.
82 We are tired of reports about a "slow udev", because these directories
83 contain stuff, that runs with _every_ event, instead of using rules,
84 that run programs only for the matching events.
94 DRIVER== will match only for devices that actually have a real
95 driver. DRIVERS== must be used, if parent devices should be
96 included in the match.
98 Libvolume_id's "linux_raid" detection needed another fix.
106 Add additional check to volume_id detection of via_raid, cause
107 some company decided to put a matching pattern all over the empty
108 storage area of their music players.
112 Fix path_id for SAS devices.
116 The udev daemon can be started with --debug-trace now, which will
117 execute all events serialized to get a chance to catch a possible
118 action that crashes the box.
120 A warning is logged, if PHYSDEV* keys, the "device" link, or a parent
121 device attribute like $attr{../file} is used, only WAIT_FOR_SYSFS rules
122 are excluded from the warning. Referencing parent attributes directly
123 may break when something in the kernel driver model changes. Udev will
124 just find the attribute by walking up the parent chain.
126 Udevtrigger now sorts the list of devices depending on the device
127 dependency, so a "usb" device is triggered after the parent "pci"
132 Revert persistent-storage ata-serial '_' '-' replacement.
138 Udevtrigger can now filter the list of devices to be triggered. Matches
139 for subsystems or sysfs attributes can be specified.
141 The entries in /dev/.udev/queue and /dev/.udev/failed have changed to
142 zero-sized files to avoid pointing to /sys and confuse broken tools which
143 scan the /dev directory. To retry failed events, udevtrigger --retry-failed
146 The rules and scripts to create udev rules for persistent network
147 devices and optical drives are in the extras/rules_generator directory
148 now. If you use something similar, please consider replacing your own
149 version with this, to share the support effort. The rule_generator
150 installs its own rules into /etc/udev/rules.d.
152 The cdrom_id tool installs its own rule now in /etc/udev/rules.d, cause
153 the rule_generator depends on cdrom_id to be called in an earlier rule.
159 Renaming of some key names (the old names still work):
160 BUS -> SUBSYSTEMS, ID -> KERNELS, SYSFS -> ATTRS, DRIVER -> DRIVERS.
161 (The behavior of the key DRIVER will change soon in one of the next
162 releases, to match only the event device, please switch to DRIVERS
163 instead. If DRIVER is used, it will behave like DRIVERS, but an error
165 With the new key names, we have a more consistent and simpler scheme.
166 We can match the properties of the event device only, with: KERNEL,
167 SUBSYSTEM, ATTR, DRIVER. Or include all the parent devices in the match,
168 with: KERNELS, SUBSYSTEMS, ATTRS, DRIVERS. ID, BUS, SYSFS, DRIVER are no
169 longer mentioned in the man page and should be switched in the rule
172 ATTR{file}="value" can be used now, to write to a sysfs file of the
173 event device. Instead of:
174 ..., SYSFS{type}=="0|7|14", RUN+="/bin/sh -c 'echo 60 > /sys$$DEVPATH/timeout'"
176 ..., ATTR{type}=="0|7|14", ATTR{timeout}="60"
178 All the PHYSDEV* keys are deprecated and will be removed from a
180 PHYDEVPATH - is the path of a parent device and should not be
182 PHYSDEVBUS - is just a SUBSYSTEM value of a parent, and can be
183 matched with SUBSYSTEMS==
184 PHYSDEVDRIVER - for bus devices it is available as ENV{DRIVER}.
185 Newer kernels will have DRIVER in the environment,
186 for older kernels udev puts in. Class device will
187 no longer carry this property of a parent and
188 DRIVERS== can be used to match such a parent value.
189 Note that ENV{DRIVER} is only available for a few bus devices, where
190 the driver is already bound at device event time. On coldplug, the
191 events for a lot devices are already bound to a driver, and they will have
192 that value set. But on hotplug, at the time the kernel creates the device,
193 it can't know what driver may claim the device after that, therefore
194 in most cases it will be empty.
196 Failed events should now be re-triggered with:
197 udevtrigger --retry-failed.
198 Please switch to this command, so we keep the details of the /dev/.udev/failed/
199 files private to the udev tools. We may need to switch the current symlink
200 target, cause some obviously broken tools try to scan all files in /dev
201 including /dev/.udev/, find the links to /sys and end up stat()'ing sysfs files
202 million times. This takes ages on slow boxes.
204 The udevinfo attribute walk (-a) now works with giving a device node
205 name (-n) instead of a devpath (-p). The query now always works, also when
206 no database file was created by udev.
208 The built-in /etc/passwd /etc/group parser is removed, we always depend on
209 getpwnam() and getgrnam() now. One of the next releases will depend on
210 fnmatch() and may use getopt_long().
214 Bugfixes and small improvements.
218 Fix path_id for recent kernels.
224 Added support for swapping network interface names, by temporarily
225 renaming the device and wait for the target name to become free.
229 The built-in MODALIAS key and substitution is removed.
233 The binary firmware helper is replaced by the usual simple
234 shell script. Udevsend is removed from the tree.
242 Some more keys require the correct use of '==' and '=' depending
243 on the kind of operation beeing an assignment or a match. Rules
244 with invalid operations are skipped and logged to syslog. Please
245 test with udevtest if the parsing of your rules throws errors and
246 fix possibly broken rules.
250 Provide "udevsettle" to wait for all current udev events to finish.
251 It also watches the current kernel netlink queue by comparing the
252 even sequence number to make sure that there are no current pending
253 events that have not already arrived in the daemon.
257 Fix rule to skip persistent rules for removable IDE devices, which
258 also skipped optical IDE drives.
260 All *_id program are installed in /lib/udev/ by default now.
262 No binary is stripped anymore as this should be done in the
263 packaging process and not at build time.
265 libvolume_id is provided as a shared library now and vol_id is
266 linked against it. Also one of the next HAL versions will require
267 this library, and the HAL build process will also require the
268 header file to be installed. The copy of the same code in HAL will
269 be removed to have only a single copy left on the system.
273 Add persistent links for SCSI tapes. The rules file is renamed
274 to 60-persistent-storage.rules.
276 Create persistent path for usb devices. Can be used for all sorts
277 of devices that can't be distinguished by other properties like
278 multiple identical keyboards and mice connected to the same box.
280 Provide "udevtrigger" program to request events on coldplug. The
281 shell script is much too slow with thousends of devices.
285 Fix persistent disk rules to exclude removable IDE drives.
287 Warn if %e, $modalias or MODALIAS is used.
291 Fix queue export, which wasn't correct for subsequent add/remove
292 events for the same device.
296 Fix cramfs detection on big endian.
298 Make WAIT_FOR_SYSFS usable in "normal" rules and silent if the whole
303 If BUS== and SYSFS{}== have been used in the same rule, the sysfs
304 attributes were only checked at the parent device that matched the
305 by BUS requested subsystem. Fix it to also look at the device we
306 received the event for.
308 Build variable CROSS has changed to CROSS_COMPILE to match the kernel
313 Fix a bug where NAME="" would prevent RUN from beeing executed.
315 RUN="/bin/program" does not longer automatically add the subsystem
316 as the first parameter. This is from the days of /sbin/hotplug
317 which is dead now and it's just confusing to need to add a space at
318 the end of the program name to prevent this.
319 If you use rules that need the subsystem as the first parameter,
320 like the old "udev_run_hotlugd" and "udev_run_devd", add the subsystem
321 to the key like RUN+="/bin/program $env{SUBSYSTEM}".
325 The udev man page has moved to udev(7) as it does not describe a command
326 anymore. The programs udev, udevstart and udevsend are no longer installed
327 by default and must be copied manually, if they should be installed or
328 included in a package.
330 Fix a bug where "ignore_device" could run earlier collected RUN keys before
331 the ignore rule was applied.
333 More preparation for future sysfs changes. usb_id and scsi_id no longer
334 depend on a magic order of devices in the /devices chain. Specific devices
335 should be requested by their subsytem.
337 This will always find the scsi parent device without depending on a specific
339 dev = sysfs_device_get(devpath);
340 dev_usb = sysfs_device_get_parent_with_subsystem(dev, "scsi");
342 The "device" link in the current sysfs layout will be automatically
343 _resolved_ as a parent and in the new sysfs layout it will just _be_ the
344 parent in the devpath. If a device is requested by it's symlink, like all
345 class devices in the new sysfs layout will look like, it gets automatically
346 resolved and substituted with the real devpath and not the symlink path.
349 A similar logic must be applied to _all_ sysfs users, including
350 scripts, that search along parent devices in sysfs. The explicit use of
351 the "device" link must be avoided. With the future sysfs layout all
352 DEVPATH's will start with /devices/ and have a "subsystem" symlink poiting
353 back to the "class" or the "bus". The layout of the parent devices in
354 /devices is not necessarily expected to be stable across kernel releases and
355 searching for parents by their subsystem should make sysfs users tolerant
356 for changed parent chains.
360 Prepare udev to work with the experimental kernel patch, that moves
361 /sys/class devices to /sys/devices and /sys/block to /sys/class/block.
363 Clarify BUS, ID, $id usage and fix $id behavior. This prepares for
364 moving the class devices to /sys/devices.
366 Thanks again to Marco for help finding a hopefully nice compromise
367 to make %b simpler and working again.
371 Complete removal of libsysfs, replaced by simple helper functions
372 which are much simpler and a bit faster. The udev daemon operatesentirely
373 on event parameters and does not use sysfs for simple rules anymore.
374 Please report any new bugs/problems, that may be caused by this big
375 change. They will be fixed immediately.
377 The enumeration format character '%e' is deprecated and will be
378 removed sometimes from a future udev version. It never worked correctly
379 outside of udevstart, so we can't use it with the new parallel
380 coldplug. A simple enumeration is as useless as the devfs naming
381 scheme, just get rid of both if you still use it.
383 MODALIAS and $modalias is not needed and will be removed from one of
384 the next udev versions, replace it in all rules with ENV{MODALIAS} or
385 the sysfs "modalias" value.
387 Thanks a lot to Marco for all his help on finding and fixing bugs.
391 Let scsi_id request libata drive serial numbers from page 0x80.
393 Renamed etc/udev/persistent.rules to persistent-disk.rules and
394 added /dev/disk/by-name/* for device mapper device names.
396 Removed %e from the man page. It never worked reliably outside
397 of udevstart and udevstart is no longer recommended to use.
401 Symlinks are now exported to the event environment. Hopefully it's no
402 longer needed to run udevinfo from an event process, like it was
403 mentioned on the hotplug list:
404 UDEV [1134776873.702967] add@/block/sdb
407 DEVLINKS=/dev/disk/by-id/usb-IBM_Memory_Key_0218B301030027E8 /dev/disk/by-path/usb-0218B301030027E8:0:0:0
411 Fix a problem if udevsend is used as the hotplug handler and tries to use
412 syslog, which causes a "vc" event loop. 2.6.15 will make udevsend obsolete
413 and this kind of problems will hopefully go away soon.
417 All built-in logic to work around bad sysfs timing is removed with this
418 version. The need to wait for sysfs files is almost fixed with a kernel
419 version that doesn't work with this udev version anyway. Until we fix
420 the timing of the "bus" link creation, the former integrated logic should
421 be emulated by a rule placed before all other rules:
422 ACTION=="add", DEVPATH=="/devices/*", ENV{PHYSDEVBUS}=="?*", WAIT_FOR_SYSFS="bus"
424 The option "udev_db" does no longer exist. All udev state will be in
425 /$udev_root/.udev/ now, there is no longer an option to set this
427 If the init script or something else used this value, just depend on
428 this hardcoded path. But remember _all_content_ of this directory is
429 still private to udev and can change at any time.
431 Default location for rule sripts and helper programs is now: /lib/udev/.
432 Everything that is not useful on the commandline should go into this
433 directory. Some of the helpers in the extras folder are installed there
434 now. The rules need to be changed, to find the helpers there.
436 Also /lib/udev/devices is recommended as a directory where packages or
437 the user can place real device nodes, which get copied over to /dev at
438 every boot. This should replace the various solutions with custom config
441 Udevsend does no longer start the udev daemon. This must be done with
442 the init script that prepares /dev on tmpfs and creates the initial nodes,
443 before starting the daemon.
447 Silent a too verbose error logging for the old hotplug.d/ dev.d/
450 The copy of klibc is removed. A systemwide installed version of klibc
451 should be used to build a klibc udev now.
455 NAME="" will not create any nodes, but execute RUN keys. To completely
456 ignore an event the OPTION "ignore_device" should be used.
458 After removal of the reorder queue, events with a TIMEOUT can be executed
459 without any queuing now.
463 Fixed bug in udevd, if inotify is not available. We depend on netlink
464 uevents now, kernels without that event source will not work with that
465 version of udev anymore.
469 The rule parsing happens now in the daemon once at startup, all udev
470 event processes inherit the already parsed rules from the daemon.
471 It is shipped with SUSE10.0 and reduces heavily the system load at
472 startup. The option to save precompiled rules and let the udev process
473 pick the them up is removed, as it's no longer needed.
475 Kernel 2.6.15 will have symlinks at /class/input pointing to the real
476 device. Libsysfs is changed to "translate" the requested link into the
477 real device path, as it would happen with the hotplug event. Otherwise
478 device removal and the udev database will not work.
480 Using 'make STRIPCMD=' will leave the binaries unstripped for debugging
483 A few improvements for vol_id, the filesytem probing code.
487 Fix a stupid typo in extras/run_directory for "make install".
489 scsi_id creates the temporary devnode now in /dev for usage with a
490 non-writable /tmp directory.
492 The uevent kernel socket buffer can carry app. 50.000 events now,
493 let's see who can break this again. :)
495 The upcoming kernel will have a new input driver core integration.
496 Some class devices are now symlinks to the real device. libsysfs
497 needs a fix for this to work correctly. Udevstart of older udev
498 versions will _not_ create these devices!
502 Fix a 'install' target in the Makefile, that prevents EXTRAS from
507 A bunch of mostly trivial bugfixes. From now on no node name or
508 symlink name can contain any character than plain whitelisted ascii
509 characters or validated utf8 byte-streams. This is needed for the
510 /dev/disk/by-label/* links, because we import untrusted data and
511 export it to the filesystem.
515 More bugfixes. If udevd was started from the kernel, we don't
516 have stdin/stdout/stderr, which broke the forked tools in some
521 Bugfix. udevstart event ordering was broken for a long time.
522 The new run_program() uncovered it, because /dev/null was not
523 available while we try to run external programs.
524 Now udevstart should create it before we run anything.
528 Minor bugfixes and some distro rules updates. If you don't have the
529 persistent disk rules in /dev/disk/by-*/* on your distro, just
530 grab it from here. :)
534 We can use socket communication now to pass events from udev to
536 RUN+="socket:/org/freedesktop/hal/udev_event"
537 will pass the whole udev event to the HAL daemon without the need
538 for a forked helper. (See ChangeLog for udevmonitor, as an example)
542 Mostly bugfixes and see ChangeLog.
544 The test for the existence of an environment value should be
546 ENV{KEY}=="*" to ENV{KEY}=="?*"
547 because "*" will not fail anymore, if the key does not exist or
552 Bugfixes and a few tweaks described in the ChangeLog.
556 Mostly a Bugfix release.
558 Added WAIT_FOR_SYSFS="<attribute>" to be able to fight against the sysfs
559 timing with custom rules.
563 We changed the internal rule storage format. Our large rule files took
564 2 MB of RAM, with the change we are down to 99kB.
566 If the device-node has been created with default name and no symlink or
567 options are to remenber, it is not longer stored in the udevdb. HAL will
568 need to be updated to work correctly with that change.
570 To overrride optimization flags, OPTFLAGS may be used now.
578 Major changes happened with this release. The goal is to take over the
579 complete kernel-event handling and provide a more efficient way to dispatch
580 kernel events. Replacing most of the current shell script logic and the
581 kernel forked helper with a netlink-daemon and a rule-based event handling.
583 o udevd listens to netlink events now. The first valid netlink event
584 will make udevd ignore any message from udevsend that contains a
585 SEQNUM, to avoid duplicate events. The forked events can be disabled
587 echo "" > /proc/sys/kernel/hotplug
588 For full support, the broken input-subsytem needs to be fixed, not to
589 bypass the driver core.
591 o /etc/dev.d/ + /etc/hotplug.d/ directory multiplexing is completely
592 removed from udev itself and must be emulated by calling small
593 helper binaries provided in the extras folder:
594 make EXTRAS=extras/run_directory/
595 will build udev_run_devd and udev_run_hotplugd, which can be called
596 from a rule if needed:
597 RUN+="/sbin/udev_run_hotplugd"
598 The recommended way to handle this is to convert all the calls from
599 the directories to explicit udev rules and get completely rid of the
600 multiplexing. (To catch a ttyUSB event, you now no longer need to
601 fork and exit 300 tty script instances you are not interested in, it
602 is just one rule that matches exactly the device.)
604 o udev handles now _all_ events not just events for class and block
605 devices, this way it is possible to control the complete event
606 behavior with udev rules. Especially useful for rules like:
607 ACTION="add", DEVPATH="/devices/*", MODALIAS=="?*", RUN+="/sbin/modprobe $modalias"
609 o As used in the modalias rule, udev supports now textual
610 substitution placeholder along with the usual format chars. This
611 needs to be documented, for now it's only visible in udev_rules_parse.c.
613 o The rule keys support now more operations. This is documented in the
614 man page. It is possible to add values to list-keys like the SYMLINK
615 and RUN list with KEY+="value" and to clear the list by assigning KEY="".
616 Also "final"-assignments are supported by using KEY:="value", which will
617 prevent changing the key by any later rule.
619 o kernel 2.6.12 has the "detached_state" attribute removed from
620 sysfs, which was used to recognize sysfs population. We switched that
621 to wait for the "bus" link, which is only available in kernels after 2.6.11.
622 Running this udev version on older kernels may cause a short delay for
625 o To provide infrastructure for persistent device naming, the id programs:
626 scsi_id, vol_id (former udev_volume_id), and ata_id (new) are able now
627 to export the probed data in environment key format:
628 pim:~ # /sbin/ata_id --export /dev/hda
629 ID_MODEL=HTS726060M9AT00
630 ID_SERIAL=MRH401M4G6UM9B
634 KERNEL="hd*[!0-9]", IMPORT="/sbin/ata_id --export $tempnode"
635 KERNEL="hd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_MODEL}_$env{ID_SERIAL}"
638 kay@pim:~> tree /dev/disk
641 | |-- HTS726060M9AT00_MRH401M4G6UM9B -> ../../hda
642 | `-- IBM-Memory_Key -> ../../sda
644 | |-- swap -> ../../hda1
645 | |-- date -> ../../sda1
646 | `-- home -> ../../hda3
648 |-- 2E08712B0870F2E7 -> ../../hda3
649 |-- 9352cfef-7687-47bc-a2a3-34cf136f72e1 -> ../../hda1
650 |-- E845-7A89 -> ../../sda1
651 `-- b2a61681-3812-4f13-a4ff-920d70604299 -> ../../hda2
653 The IMPORT= operation will import these keys in the environment and make
654 it available for later PROGRAM= and RUN= executed programs. The keys are
655 also stored in the udevdb and can be queried from there with one of the
658 o A few binaries are silently added to the repository, which can be used
659 to replay kernel events from initramfs instead of using coldplug. udevd
660 can be instructed now to queue-up events while the stored events from
661 initramfs are filled into the udevd-queue. This code is still under
662 development and there is no documentation now besides the code itself.
663 The additional binaries get compiled, but are not installed by default.
665 o There is also a temporary fix for a performance problem where too many
666 events happen in parallel and every event needs to parse the rules.
667 udev can now read precompiled rules stored on disk. This is likely to be
668 replaced by a more elegant solution in a future udev version.
672 With kernel version 2.6.12, the sysfs file "detached_state" was removed.
673 Fix for libsysfs not to expect this file was added.
677 All rules are applied now, but only the first matching rule with a NAME-key
678 will be applied. All later rules with NAME-key are completely ignored. This
679 way system supplied symlinks or permissions gets applied to user-defined
683 Please check your rules setup, if you may need to add OPTIONS="last_rule"
684 to some rules, to keep the old behavior.
686 The rules are read on "remove"-events too. That makes is possible to match
687 with keys that are available on remove (KERNEL, SUBSYSTEM, ID, ENV, ...) to
688 instruct udev to ignore an event (OPTIONS="ignore_device").
689 The new ACTION-key may be used to let a rule act only at a "remove"-event.
691 The new RUN-key supports rule-based execution of programs after device-node
692 handling. This is meant as a general replacement for the dev.d/-directories
693 to give fine grained control over the execution of programs.
695 The %s{}-sysfs format char replacement values are searched at any of the
696 devices in the device chain now, not only at the class-device.
698 We support log priority levels now. The value udev_log in udev.conf is used
699 to determine what is printed to syslog. This makes it possible to
700 run a version with compiled-in debug messages in a production environment
701 which is sometimes needed to find a bug.
702 It is still possible to supress the inclusion of _any_ syslog usage with
703 USE_LOG=false to create the smallest possible binaries if needed.
704 The configured udev_log value can be overridden with the environment variable
709 Possible use of a system-wide klibc:
710 make USE_KLIBC=true KLCC=/usr/bin/klcc all
711 will link against an external klibc and our own version will be ignored.
715 We support an unlimited count of symlinks now.
717 If USE_STATIC=true is passed to a glibc build, we link statically and use
718 a built-in userdb parser to resolve user and group names.
720 The PLACE= key is gone. It can be replaced by an ID= for a long time, because
721 we walk up the chain of physical devices to find a match.
723 The KEY="<value>" format supports '=', '==', '!=,' , '+=' now. This makes it
724 easy to skip certain attribute matches without composing rules with weird
725 character class negations like:
726 KERNEL="[!s][!c][!d]*"
727 this can now be replaced with:
729 The current simple '=' is still supported, and should work as it does today,
730 but existing rules should be converted if possible, to be better readable.
732 We have new ENV{}== key now, to match against a maximum of 5 environment
735 udevstart is its own binary again, because we don't need co carry this araound
736 with every forked event.