5 DRIVER== will match only for devices that actually have a real
6 driver. DRIVERS== must be used, if parent devices should be
9 Libvolume_id's "linux_raid" detection needed another fix.
17 Add additional check to volume_id detection of via_raid, cause
18 some company decided to put a matching pattern all over the empty
19 storage area of their music players.
23 Fix path_id for SAS devices.
27 The udev daemon can be started with --debug-trace now, which will
28 execute all events serialized to get a chance to catch a possible
29 action that crashes the box.
31 A warning is logged, if PHYSDEV* keys, the "device" link, or a parent
32 device attribute like $attr{../file} is used, only WAIT_FOR_SYSFS rules
33 are excluded from the warning. Referencing parent attributes directly
34 may break when something in the kernel driver model changes. Udev will
35 just find the attribute by walking up the parent chain.
37 Udevtrigger now sorts the list of devices depending on the device
38 dependency, so a "usb" device is triggered after the parent "pci"
43 Revert persistent-storage ata-serial '_' '-' replacement.
49 Udevtrigger can now filter the list of devices to be triggered. Matches
50 for subsystems or sysfs attributes can be specified.
52 The entries in /dev/.udev/queue and /dev/.udev/failed have changed to
53 zero-sized files to avoid pointing to /sys and confuse broken tools which
54 scan the /dev directory. To retry failed events, udevtrigger --retry-failed
57 The rules and scripts to create udev rules for persistent network
58 devices and optical drives are in the extras/rules_generator directory
59 now. If you use something similar, please consider replacing your own
60 version with this, to share the support effort. The rule_generator
61 installs its own rules into /etc/udev/rules.d.
63 The cdrom_id tool installs its own rule now in /etc/udev/rules.d, cause
64 the rule_generator depends on cdrom_id to be called in an earlier rule.
70 Renaming of some key names (the old names still work):
71 BUS -> SUBSYSTEMS, ID -> KERNELS, SYSFS -> ATTRS, DRIVER -> DRIVERS.
72 (The behavior of the key DRIVER will change soon in one of the next
73 releases, to match only the event device, please switch to DRIVERS
74 instead. If DRIVER is used, it will behave like DRIVERS, but an error
76 With the new key names, we have a more consistent and simpler scheme.
77 We can match the properties of the event device only, with: KERNEL,
78 SUBSYSTEM, ATTR, DRIVER. Or include all the parent devices in the match,
79 with: KERNELS, SUBSYSTEMS, ATTRS, DRIVERS. ID, BUS, SYSFS, DRIVER are no
80 longer mentioned in the man page and should be switched in the rule
83 ATTR{file}="value" can be used now, to write to a sysfs file of the
84 event device. Instead of:
85 ..., SYSFS{type}=="0|7|14", RUN+="/bin/sh -c 'echo 60 > /sys$$DEVPATH/timeout'"
87 ..., ATTR{type}=="0|7|14", ATTR{timeout}="60"
89 All the PHYSDEV* keys are deprecated and will be removed from a
91 PHYDEVPATH - is the path of a parent device and should not be
93 PHYSDEVBUS - is just a SUBSYSTEM value of a parent, and can be
94 matched with SUBSYSTEMS==
95 PHYSDEVDRIVER - for bus devices it is available as ENV{DRIVER}.
96 Newer kernels will have DRIVER in the environment,
97 for older kernels udev puts in. Class device will
98 no longer carry this property of a parent and
99 DRIVERS== can be used to match such a parent value.
100 Note that ENV{DRIVER} is only available for a few bus devices, where
101 the driver is already bound at device event time. On coldplug, the
102 events for a lot devices are already bound to a driver, and they will have
103 that value set. But on hotplug, at the time the kernel creates the device,
104 it can't know what driver may claim the device after that, therefore
105 in most cases it will be empty.
107 Failed events should now be re-triggered with:
108 udevtrigger --retry-failed.
109 Please switch to this command, so we keep the details of the /dev/.udev/failed/
110 files private to the udev tools. We may need to switch the current symlink
111 target, cause some obviously broken tools try to scan all files in /dev
112 including /dev/.udev/, find the links to /sys and end up stat()'ing sysfs files
113 million times. This takes ages on slow boxes.
115 The udevinfo attribute walk (-a) now works with giving a device node
116 name (-n) instead of a devpath (-p). The query now always works, also when
117 no database file was created by udev.
119 The built-in /etc/passwd /etc/group parser is removed, we always depend on
120 getpwnam() and getgrnam() now. One of the next releases will depend on
121 fnmatch() and may use getopt_long().
125 Bugfixes and small improvements.
129 Fix path_id for recent kernels.
135 Added support for swapping network interface names, by temporarily
136 renaming the device and wait for the target name to become free.
140 The built-in MODALIAS key and substitution is removed.
144 The binary firmware helper is replaced by the usual simple
145 shell script. Udevsend is removed from the tree.
153 Some more keys require the correct use of '==' and '=' depending
154 on the kind of operation beeing an assignment or a match. Rules
155 with invalid operations are skipped and logged to syslog. Please
156 test with udevtest if the parsing of your rules throws errors and
157 fix possibly broken rules.
161 Provide "udevsettle" to wait for all current udev events to finish.
162 It also watches the current kernel netlink queue by comparing the
163 even sequence number to make sure that there are no current pending
164 events that have not already arrived in the daemon.
168 Fix rule to skip persistent rules for removable IDE devices, which
169 also skipped optical IDE drives.
171 All *_id program are installed in /lib/udev/ by default now.
173 No binary is stripped anymore as this should be done in the
174 packaging process and not at build time.
176 libvolume_id is provided as a shared library now and vol_id is
177 linked against it. Also one of the next HAL versions will require
178 this library, and the HAL build process will also require the
179 header file to be installed. The copy of the same code in HAL will
180 be removed to have only a single copy left on the system.
184 Add persistent links for SCSI tapes. The rules file is renamed
185 to 60-persistent-storage.rules.
187 Create persistent path for usb devices. Can be used for all sorts
188 of devices that can't be distinguished by other properties like
189 multiple identical keyboards and mice connected to the same box.
191 Provide "udevtrigger" program to request events on coldplug. The
192 shell script is much too slow with thousends of devices.
196 Fix persistent disk rules to exclude removable IDE drives.
198 Warn if %e, $modalias or MODALIAS is used.
202 Fix queue export, which wasn't correct for subsequent add/remove
203 events for the same device.
207 Fix cramfs detection on big endian.
209 Make WAIT_FOR_SYSFS usable in "normal" rules and silent if the whole
214 If BUS== and SYSFS{}== have been used in the same rule, the sysfs
215 attributes were only checked at the parent device that matched the
216 by BUS requested subsystem. Fix it to also look at the device we
217 received the event for.
219 Build variable CROSS has changed to CROSS_COMPILE to match the kernel
224 Fix a bug where NAME="" would prevent RUN from beeing executed.
226 RUN="/bin/program" does not longer automatically add the subsystem
227 as the first parameter. This is from the days of /sbin/hotplug
228 which is dead now and it's just confusing to need to add a space at
229 the end of the program name to prevent this.
230 If you use rules that need the subsystem as the first parameter,
231 like the old "udev_run_hotlugd" and "udev_run_devd", add the subsystem
232 to the key like RUN+="/bin/program $env{SUBSYSTEM}".
236 The udev man page has moved to udev(7) as it does not describe a command
237 anymore. The programs udev, udevstart and udevsend are no longer installed
238 by default and must be copied manually, if they should be installed or
239 included in a package.
241 Fix a bug where "ignore_device" could run earlier collected RUN keys before
242 the ignore rule was applied.
244 More preparation for future sysfs changes. usb_id and scsi_id no longer
245 depend on a magic order of devices in the /devices chain. Specific devices
246 should be requested by their subsytem.
248 This will always find the scsi parent device without depending on a specific
250 dev = sysfs_device_get(devpath);
251 dev_usb = sysfs_device_get_parent_with_subsystem(dev, "scsi");
253 The "device" link in the current sysfs layout will be automatically
254 _resolved_ as a parent and in the new sysfs layout it will just _be_ the
255 parent in the devpath. If a device is requested by it's symlink, like all
256 class devices in the new sysfs layout will look like, it gets automatically
257 resolved and substituted with the real devpath and not the symlink path.
260 A similar logic must be applied to _all_ sysfs users, including
261 scripts, that search along parent devices in sysfs. The explicit use of
262 the "device" link must be avoided. With the future sysfs layout all
263 DEVPATH's will start with /devices/ and have a "subsystem" symlink poiting
264 back to the "class" or the "bus". The layout of the parent devices in
265 /devices is not necessarily expected to be stable across kernel releases and
266 searching for parents by their subsystem should make sysfs users tolerant
267 for changed parent chains.
271 Prepare udev to work with the experimental kernel patch, that moves
272 /sys/class devices to /sys/devices and /sys/block to /sys/class/block.
274 Clarify BUS, ID, $id usage and fix $id behavior. This prepares for
275 moving the class devices to /sys/devices.
277 Thanks again to Marco for help finding a hopefully nice compromise
278 to make %b simpler and working again.
282 Complete removal of libsysfs, replaced by simple helper functions
283 which are much simpler and a bit faster. The udev daemon operatesentirely
284 on event parameters and does not use sysfs for simple rules anymore.
285 Please report any new bugs/problems, that may be caused by this big
286 change. They will be fixed immediately.
288 The enumeration format character '%e' is deprecated and will be
289 removed sometimes from a future udev version. It never worked correctly
290 outside of udevstart, so we can't use it with the new parallel
291 coldplug. A simple enumeration is as useless as the devfs naming
292 scheme, just get rid of both if you still use it.
294 MODALIAS and $modalias is not needed and will be removed from one of
295 the next udev versions, replace it in all rules with ENV{MODALIAS} or
296 the sysfs "modalias" value.
298 Thanks a lot to Marco for all his help on finding and fixing bugs.
302 Let scsi_id request libata drive serial numbers from page 0x80.
304 Renamed etc/udev/persistent.rules to persistent-disk.rules and
305 added /dev/disk/by-name/* for device mapper device names.
307 Removed %e from the man page. It never worked reliably outside
308 of udevstart and udevstart is no longer recommended to use.
312 Symlinks are now exported to the event environment. Hopefully it's no
313 longer needed to run udevinfo from an event process, like it was
314 mentioned on the hotplug list:
315 UDEV [1134776873.702967] add@/block/sdb
318 DEVLINKS=/dev/disk/by-id/usb-IBM_Memory_Key_0218B301030027E8 /dev/disk/by-path/usb-0218B301030027E8:0:0:0
322 Fix a problem if udevsend is used as the hotplug handler and tries to use
323 syslog, which causes a "vc" event loop. 2.6.15 will make udevsend obsolete
324 and this kind of problems will hopefully go away soon.
328 All built-in logic to work around bad sysfs timing is removed with this
329 version. The need to wait for sysfs files is almost fixed with a kernel
330 version that doesn't work with this udev version anyway. Until we fix
331 the timing of the "bus" link creation, the former integrated logic should
332 be emulated by a rule placed before all other rules:
333 ACTION=="add", DEVPATH=="/devices/*", ENV{PHYSDEVBUS}=="?*", WAIT_FOR_SYSFS="bus"
335 The option "udev_db" does no longer exist. All udev state will be in
336 /$udev_root/.udev/ now, there is no longer an option to set this
338 If the init script or something else used this value, just depend on
339 this hardcoded path. But remember _all_content_ of this directory is
340 still private to udev and can change at any time.
342 Default location for rule sripts and helper programs is now: /lib/udev/.
343 Everything that is not useful on the commandline should go into this
344 directory. Some of the helpers in the extras folder are installed there
345 now. The rules need to be changed, to find the helpers there.
347 Also /lib/udev/devices is recommended as a directory where packages or
348 the user can place real device nodes, which get copied over to /dev at
349 every boot. This should replace the various solutions with custom config
352 Udevsend does no longer start the udev daemon. This must be done with
353 the init script that prepares /dev on tmpfs and creates the initial nodes,
354 before starting the daemon.
358 Silent a too verbose error logging for the old hotplug.d/ dev.d/
361 The copy of klibc is removed. A systemwide installed version of klibc
362 should be used to build a klibc udev now.
366 NAME="" will not create any nodes, but execute RUN keys. To completely
367 ignore an event the OPTION "ignore_device" should be used.
369 After removal of the reorder queue, events with a TIMEOUT can be executed
370 without any queuing now.
374 Fixed bug in udevd, if inotify is not available. We depend on netlink
375 uevents now, kernels without that event source will not work with that
376 version of udev anymore.
380 The rule parsing happens now in the daemon once at startup, all udev
381 event processes inherit the already parsed rules from the daemon.
382 It is shipped with SUSE10.0 and reduces heavily the system load at
383 startup. The option to save precompiled rules and let the udev process
384 pick the them up is removed, as it's no longer needed.
386 Kernel 2.6.15 will have symlinks at /class/input pointing to the real
387 device. Libsysfs is changed to "translate" the requested link into the
388 real device path, as it would happen with the hotplug event. Otherwise
389 device removal and the udev database will not work.
391 Using 'make STRIPCMD=' will leave the binaries unstripped for debugging
394 A few improvements for vol_id, the filesytem probing code.
398 Fix a stupid typo in extras/run_directory for "make install".
400 scsi_id creates the temporary devnode now in /dev for usage with a
401 non-writable /tmp directory.
403 The uevent kernel socket buffer can carry app. 50.000 events now,
404 let's see who can break this again. :)
406 The upcoming kernel will have a new input driver core integration.
407 Some class devices are now symlinks to the real device. libsysfs
408 needs a fix for this to work correctly. Udevstart of older udev
409 versions will _not_ create these devices!
413 Fix a 'install' target in the Makefile, that prevents EXTRAS from
418 A bunch of mostly trivial bugfixes. From now on no node name or
419 symlink name can contain any character than plain whitelisted ascii
420 characters or validated utf8 byte-streams. This is needed for the
421 /dev/disk/by-label/* links, because we import untrusted data and
422 export it to the filesystem.
426 More bugfixes. If udevd was started from the kernel, we don't
427 have stdin/stdout/stderr, which broke the forked tools in some
432 Bugfix. udevstart event ordering was broken for a long time.
433 The new run_program() uncovered it, because /dev/null was not
434 available while we try to run external programs.
435 Now udevstart should create it before we run anything.
439 Minor bugfixes and some distro rules updates. If you don't have the
440 persistent disk rules in /dev/disk/by-*/* on your distro, just
441 grab it from here. :)
445 We can use socket communication now to pass events from udev to
447 RUN+="socket:/org/freedesktop/hal/udev_event"
448 will pass the whole udev event to the HAL daemon without the need
449 for a forked helper. (See ChangeLog for udevmonitor, as an example)
453 Mostly bugfixes and see ChangeLog.
455 The test for the existence of an environment value should be
457 ENV{KEY}=="*" to ENV{KEY}=="?*"
458 because "*" will not fail anymore, if the key does not exist or
463 Bugfixes and a few tweaks described in the ChangeLog.
467 Mostly a Bugfix release.
469 Added WAIT_FOR_SYSFS="<attribute>" to be able to fight against the sysfs
470 timing with custom rules.
474 We changed the internal rule storage format. Our large rule files took
475 2 MB of RAM, with the change we are down to 99kB.
477 If the device-node has been created with default name and no symlink or
478 options are to remenber, it is not longer stored in the udevdb. HAL will
479 need to be updated to work correctly with that change.
481 To overrride optimization flags, OPTFLAGS may be used now.
489 Major changes happened with this release. The goal is to take over the
490 complete kernel-event handling and provide a more efficient way to dispatch
491 kernel events. Replacing most of the current shell script logic and the
492 kernel forked helper with a netlink-daemon and a rule-based event handling.
494 o udevd listens to netlink events now. The first valid netlink event
495 will make udevd ignore any message from udevsend that contains a
496 SEQNUM, to avoid duplicate events. The forked events can be disabled
498 echo "" > /proc/sys/kernel/hotplug
499 For full support, the broken input-subsytem needs to be fixed, not to
500 bypass the driver core.
502 o /etc/dev.d/ + /etc/hotplug.d/ directory multiplexing is completely
503 removed from udev itself and must be emulated by calling small
504 helper binaries provided in the extras folder:
505 make EXTRAS=extras/run_directory/
506 will build udev_run_devd and udev_run_hotplugd, which can be called
507 from a rule if needed:
508 RUN+="/sbin/udev_run_hotplugd"
509 The recommended way to handle this is to convert all the calls from
510 the directories to explicit udev rules and get completely rid of the
511 multiplexing. (To catch a ttyUSB event, you now no longer need to
512 fork and exit 300 tty script instances you are not interested in, it
513 is just one rule that matches exactly the device.)
515 o udev handles now _all_ events not just events for class and block
516 devices, this way it is possible to control the complete event
517 behavior with udev rules. Especially useful for rules like:
518 ACTION="add", DEVPATH="/devices/*", MODALIAS=="?*", RUN+="/sbin/modprobe $modalias"
520 o As used in the modalias rule, udev supports now textual
521 substitution placeholder along with the usual format chars. This
522 needs to be documented, for now it's only visible in udev_rules_parse.c.
524 o The rule keys support now more operations. This is documented in the
525 man page. It is possible to add values to list-keys like the SYMLINK
526 and RUN list with KEY+="value" and to clear the list by assigning KEY="".
527 Also "final"-assignments are supported by using KEY:="value", which will
528 prevent changing the key by any later rule.
530 o kernel 2.6.12 has the "detached_state" attribute removed from
531 sysfs, which was used to recognize sysfs population. We switched that
532 to wait for the "bus" link, which is only available in kernels after 2.6.11.
533 Running this udev version on older kernels may cause a short delay for
536 o To provide infrastructure for persistent device naming, the id programs:
537 scsi_id, vol_id (former udev_volume_id), and ata_id (new) are able now
538 to export the probed data in environment key format:
539 pim:~ # /sbin/ata_id --export /dev/hda
540 ID_MODEL=HTS726060M9AT00
541 ID_SERIAL=MRH401M4G6UM9B
545 KERNEL="hd*[!0-9]", IMPORT="/sbin/ata_id --export $tempnode"
546 KERNEL="hd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_MODEL}_$env{ID_SERIAL}"
549 kay@pim:~> tree /dev/disk
552 | |-- HTS726060M9AT00_MRH401M4G6UM9B -> ../../hda
553 | `-- IBM-Memory_Key -> ../../sda
555 | |-- swap -> ../../hda1
556 | |-- date -> ../../sda1
557 | `-- home -> ../../hda3
559 |-- 2E08712B0870F2E7 -> ../../hda3
560 |-- 9352cfef-7687-47bc-a2a3-34cf136f72e1 -> ../../hda1
561 |-- E845-7A89 -> ../../sda1
562 `-- b2a61681-3812-4f13-a4ff-920d70604299 -> ../../hda2
564 The IMPORT= operation will import these keys in the environment and make
565 it available for later PROGRAM= and RUN= executed programs. The keys are
566 also stored in the udevdb and can be queried from there with one of the
569 o A few binaries are silently added to the repository, which can be used
570 to replay kernel events from initramfs instead of using coldplug. udevd
571 can be instructed now to queue-up events while the stored events from
572 initramfs are filled into the udevd-queue. This code is still under
573 development and there is no documentation now besides the code itself.
574 The additional binaries get compiled, but are not installed by default.
576 o There is also a temporary fix for a performance problem where too many
577 events happen in parallel and every event needs to parse the rules.
578 udev can now read precompiled rules stored on disk. This is likely to be
579 replaced by a more elegant solution in a future udev version.
583 With kernel version 2.6.12, the sysfs file "detached_state" was removed.
584 Fix for libsysfs not to expect this file was added.
588 All rules are applied now, but only the first matching rule with a NAME-key
589 will be applied. All later rules with NAME-key are completely ignored. This
590 way system supplied symlinks or permissions gets applied to user-defined
594 Please check your rules setup, if you may need to add OPTIONS="last_rule"
595 to some rules, to keep the old behavior.
597 The rules are read on "remove"-events too. That makes is possible to match
598 with keys that are available on remove (KERNEL, SUBSYSTEM, ID, ENV, ...) to
599 instruct udev to ignore an event (OPTIONS="ignore_device").
600 The new ACTION-key may be used to let a rule act only at a "remove"-event.
602 The new RUN-key supports rule-based execution of programs after device-node
603 handling. This is meant as a general replacement for the dev.d/-directories
604 to give fine grained control over the execution of programs.
606 The %s{}-sysfs format char replacement values are searched at any of the
607 devices in the device chain now, not only at the class-device.
609 We support log priority levels now. The value udev_log in udev.conf is used
610 to determine what is printed to syslog. This makes it possible to
611 run a version with compiled-in debug messages in a production environment
612 which is sometimes needed to find a bug.
613 It is still possible to supress the inclusion of _any_ syslog usage with
614 USE_LOG=false to create the smallest possible binaries if needed.
615 The configured udev_log value can be overridden with the environment variable
620 Possible use of a system-wide klibc:
621 make USE_KLIBC=true KLCC=/usr/bin/klcc all
622 will link against an external klibc and our own version will be ignored.
626 We support an unlimited count of symlinks now.
628 If USE_STATIC=true is passed to a glibc build, we link statically and use
629 a built-in userdb parser to resolve user and group names.
631 The PLACE= key is gone. It can be replaced by an ID= for a long time, because
632 we walk up the chain of physical devices to find a match.
634 The KEY="<value>" format supports '=', '==', '!=,' , '+=' now. This makes it
635 easy to skip certain attribute matches without composing rules with weird
636 character class negations like:
637 KERNEL="[!s][!c][!d]*"
638 this can now be replaced with:
640 The current simple '=' is still supported, and should work as it does today,
641 but existing rules should be converted if possible, to be better readable.
643 We have new ENV{}== key now, to match against a maximum of 5 environment
646 udevstart is its own binary again, because we don't need co carry this araound
647 with every forked event.