5 Control characters in filesystem label strings are no longer silenty
6 removed, but hex-encoded, to be able to uniquely identify the device
7 by its symlink in /dev/disk/by-label/.
8 If libvolume_id is used by mount(8), LABEL= will work as expected,
9 if slashes or other characters are used in the label string.
11 To test the existence of a file, TEST=="<file>" and TEST!="<file>"
12 can be specified now. The TEST key accepts an optional mode mask
13 TEST{0100}=="<is executable file>".
15 Scsi_id now supports a mode without expecting scsi-specific sysfs
16 entries to allow the extraction of cciss-device persistent properties.
22 In the future, we may see uuid's which are just simple character
23 strings (see the DDF Raid Specification). For that reason vol_id now
24 exports ID_FS_UUID_SAFE, just like ID_FS_LABEL_SAFE. For things like
25 the creation of symlinks, the *_SAFE values ensure, that no control
26 or whitespace characters are used in the filename.
28 Possible users of libvolume_id, please use the volume_id_get_* functions.
29 The public struct will go away in a future release of the library.
35 Removal of useless extras/eventrecorder.sh.
45 The directory multiplexer for dev.d/ and hotplug.d are finally removed
46 from the udev package.
52 Symlinks can have priorities now, the priority is assigned to the device
53 and specified with OPTIONS="link_priority=100". Devices with higher
54 priorities overwrite the symlinks of devices with lower priorities.
55 If the device that currently owns the link, goes away, the symlink
56 will be removed, and recreated, pointing to the next device with the
57 highest actual priority. This should make /dev/disk/by-{label,uuid,id}
58 more reliable, if multiple devices contain the same metadata and overwrite
61 The dasd_id program is removed from the udev tree, and dasdinfo, with the
62 needed rules, are part of the s390-tools now.
64 Please add KERNEL=="[0-9]*:[0-9]*" to the scsi wait-for-sysfs rule,
65 we may get the scsi sysfs mess fixed some day, and this will only catch
66 the devices we are looking for.
68 USB serial numbers for storage devices have the target:lun now appended,
69 to make it possibble to distinguish broken multi-lun devices with all
70 the same SCSI identifiers.
72 Note: The extra "run_directory" which searches and executes stuff in
73 /etc/hotplug.d/ and /etc/dev.d/ is long deprecated, and will be removed
74 with the next release. Make sure, that you don't use it anymore, or
75 provides your own implementation of that inefficient stuff.
76 We are tired of reports about a "slow udev", because these directories
77 contain stuff, that runs with _every_ event, instead of using rules,
78 that run programs only for the matching events.
88 DRIVER== will match only for devices that actually have a real
89 driver. DRIVERS== must be used, if parent devices should be
90 included in the match.
92 Libvolume_id's "linux_raid" detection needed another fix.
100 Add additional check to volume_id detection of via_raid, cause
101 some company decided to put a matching pattern all over the empty
102 storage area of their music players.
106 Fix path_id for SAS devices.
110 The udev daemon can be started with --debug-trace now, which will
111 execute all events serialized to get a chance to catch a possible
112 action that crashes the box.
114 A warning is logged, if PHYSDEV* keys, the "device" link, or a parent
115 device attribute like $attr{../file} is used, only WAIT_FOR_SYSFS rules
116 are excluded from the warning. Referencing parent attributes directly
117 may break when something in the kernel driver model changes. Udev will
118 just find the attribute by walking up the parent chain.
120 Udevtrigger now sorts the list of devices depending on the device
121 dependency, so a "usb" device is triggered after the parent "pci"
126 Revert persistent-storage ata-serial '_' '-' replacement.
132 Udevtrigger can now filter the list of devices to be triggered. Matches
133 for subsystems or sysfs attributes can be specified.
135 The entries in /dev/.udev/queue and /dev/.udev/failed have changed to
136 zero-sized files to avoid pointing to /sys and confuse broken tools which
137 scan the /dev directory. To retry failed events, udevtrigger --retry-failed
140 The rules and scripts to create udev rules for persistent network
141 devices and optical drives are in the extras/rules_generator directory
142 now. If you use something similar, please consider replacing your own
143 version with this, to share the support effort. The rule_generator
144 installs its own rules into /etc/udev/rules.d.
146 The cdrom_id tool installs its own rule now in /etc/udev/rules.d, cause
147 the rule_generator depends on cdrom_id to be called in an earlier rule.
153 Renaming of some key names (the old names still work):
154 BUS -> SUBSYSTEMS, ID -> KERNELS, SYSFS -> ATTRS, DRIVER -> DRIVERS.
155 (The behavior of the key DRIVER will change soon in one of the next
156 releases, to match only the event device, please switch to DRIVERS
157 instead. If DRIVER is used, it will behave like DRIVERS, but an error
159 With the new key names, we have a more consistent and simpler scheme.
160 We can match the properties of the event device only, with: KERNEL,
161 SUBSYSTEM, ATTR, DRIVER. Or include all the parent devices in the match,
162 with: KERNELS, SUBSYSTEMS, ATTRS, DRIVERS. ID, BUS, SYSFS, DRIVER are no
163 longer mentioned in the man page and should be switched in the rule
166 ATTR{file}="value" can be used now, to write to a sysfs file of the
167 event device. Instead of:
168 ..., SYSFS{type}=="0|7|14", RUN+="/bin/sh -c 'echo 60 > /sys$$DEVPATH/timeout'"
170 ..., ATTR{type}=="0|7|14", ATTR{timeout}="60"
172 All the PHYSDEV* keys are deprecated and will be removed from a
174 PHYDEVPATH - is the path of a parent device and should not be
176 PHYSDEVBUS - is just a SUBSYSTEM value of a parent, and can be
177 matched with SUBSYSTEMS==
178 PHYSDEVDRIVER - for bus devices it is available as ENV{DRIVER}.
179 Newer kernels will have DRIVER in the environment,
180 for older kernels udev puts in. Class device will
181 no longer carry this property of a parent and
182 DRIVERS== can be used to match such a parent value.
183 Note that ENV{DRIVER} is only available for a few bus devices, where
184 the driver is already bound at device event time. On coldplug, the
185 events for a lot devices are already bound to a driver, and they will have
186 that value set. But on hotplug, at the time the kernel creates the device,
187 it can't know what driver may claim the device after that, therefore
188 in most cases it will be empty.
190 Failed events should now be re-triggered with:
191 udevtrigger --retry-failed.
192 Please switch to this command, so we keep the details of the /dev/.udev/failed/
193 files private to the udev tools. We may need to switch the current symlink
194 target, cause some obviously broken tools try to scan all files in /dev
195 including /dev/.udev/, find the links to /sys and end up stat()'ing sysfs files
196 million times. This takes ages on slow boxes.
198 The udevinfo attribute walk (-a) now works with giving a device node
199 name (-n) instead of a devpath (-p). The query now always works, also when
200 no database file was created by udev.
202 The built-in /etc/passwd /etc/group parser is removed, we always depend on
203 getpwnam() and getgrnam() now. One of the next releases will depend on
204 fnmatch() and may use getopt_long().
208 Bugfixes and small improvements.
212 Fix path_id for recent kernels.
218 Added support for swapping network interface names, by temporarily
219 renaming the device and wait for the target name to become free.
223 The built-in MODALIAS key and substitution is removed.
227 The binary firmware helper is replaced by the usual simple
228 shell script. Udevsend is removed from the tree.
236 Some more keys require the correct use of '==' and '=' depending
237 on the kind of operation beeing an assignment or a match. Rules
238 with invalid operations are skipped and logged to syslog. Please
239 test with udevtest if the parsing of your rules throws errors and
240 fix possibly broken rules.
244 Provide "udevsettle" to wait for all current udev events to finish.
245 It also watches the current kernel netlink queue by comparing the
246 even sequence number to make sure that there are no current pending
247 events that have not already arrived in the daemon.
251 Fix rule to skip persistent rules for removable IDE devices, which
252 also skipped optical IDE drives.
254 All *_id program are installed in /lib/udev/ by default now.
256 No binary is stripped anymore as this should be done in the
257 packaging process and not at build time.
259 libvolume_id is provided as a shared library now and vol_id is
260 linked against it. Also one of the next HAL versions will require
261 this library, and the HAL build process will also require the
262 header file to be installed. The copy of the same code in HAL will
263 be removed to have only a single copy left on the system.
267 Add persistent links for SCSI tapes. The rules file is renamed
268 to 60-persistent-storage.rules.
270 Create persistent path for usb devices. Can be used for all sorts
271 of devices that can't be distinguished by other properties like
272 multiple identical keyboards and mice connected to the same box.
274 Provide "udevtrigger" program to request events on coldplug. The
275 shell script is much too slow with thousends of devices.
279 Fix persistent disk rules to exclude removable IDE drives.
281 Warn if %e, $modalias or MODALIAS is used.
285 Fix queue export, which wasn't correct for subsequent add/remove
286 events for the same device.
290 Fix cramfs detection on big endian.
292 Make WAIT_FOR_SYSFS usable in "normal" rules and silent if the whole
297 If BUS== and SYSFS{}== have been used in the same rule, the sysfs
298 attributes were only checked at the parent device that matched the
299 by BUS requested subsystem. Fix it to also look at the device we
300 received the event for.
302 Build variable CROSS has changed to CROSS_COMPILE to match the kernel
307 Fix a bug where NAME="" would prevent RUN from beeing executed.
309 RUN="/bin/program" does not longer automatically add the subsystem
310 as the first parameter. This is from the days of /sbin/hotplug
311 which is dead now and it's just confusing to need to add a space at
312 the end of the program name to prevent this.
313 If you use rules that need the subsystem as the first parameter,
314 like the old "udev_run_hotlugd" and "udev_run_devd", add the subsystem
315 to the key like RUN+="/bin/program $env{SUBSYSTEM}".
319 The udev man page has moved to udev(7) as it does not describe a command
320 anymore. The programs udev, udevstart and udevsend are no longer installed
321 by default and must be copied manually, if they should be installed or
322 included in a package.
324 Fix a bug where "ignore_device" could run earlier collected RUN keys before
325 the ignore rule was applied.
327 More preparation for future sysfs changes. usb_id and scsi_id no longer
328 depend on a magic order of devices in the /devices chain. Specific devices
329 should be requested by their subsytem.
331 This will always find the scsi parent device without depending on a specific
333 dev = sysfs_device_get(devpath);
334 dev_usb = sysfs_device_get_parent_with_subsystem(dev, "scsi");
336 The "device" link in the current sysfs layout will be automatically
337 _resolved_ as a parent and in the new sysfs layout it will just _be_ the
338 parent in the devpath. If a device is requested by it's symlink, like all
339 class devices in the new sysfs layout will look like, it gets automatically
340 resolved and substituted with the real devpath and not the symlink path.
343 A similar logic must be applied to _all_ sysfs users, including
344 scripts, that search along parent devices in sysfs. The explicit use of
345 the "device" link must be avoided. With the future sysfs layout all
346 DEVPATH's will start with /devices/ and have a "subsystem" symlink poiting
347 back to the "class" or the "bus". The layout of the parent devices in
348 /devices is not necessarily expected to be stable across kernel releases and
349 searching for parents by their subsystem should make sysfs users tolerant
350 for changed parent chains.
354 Prepare udev to work with the experimental kernel patch, that moves
355 /sys/class devices to /sys/devices and /sys/block to /sys/class/block.
357 Clarify BUS, ID, $id usage and fix $id behavior. This prepares for
358 moving the class devices to /sys/devices.
360 Thanks again to Marco for help finding a hopefully nice compromise
361 to make %b simpler and working again.
365 Complete removal of libsysfs, replaced by simple helper functions
366 which are much simpler and a bit faster. The udev daemon operatesentirely
367 on event parameters and does not use sysfs for simple rules anymore.
368 Please report any new bugs/problems, that may be caused by this big
369 change. They will be fixed immediately.
371 The enumeration format character '%e' is deprecated and will be
372 removed sometimes from a future udev version. It never worked correctly
373 outside of udevstart, so we can't use it with the new parallel
374 coldplug. A simple enumeration is as useless as the devfs naming
375 scheme, just get rid of both if you still use it.
377 MODALIAS and $modalias is not needed and will be removed from one of
378 the next udev versions, replace it in all rules with ENV{MODALIAS} or
379 the sysfs "modalias" value.
381 Thanks a lot to Marco for all his help on finding and fixing bugs.
385 Let scsi_id request libata drive serial numbers from page 0x80.
387 Renamed etc/udev/persistent.rules to persistent-disk.rules and
388 added /dev/disk/by-name/* for device mapper device names.
390 Removed %e from the man page. It never worked reliably outside
391 of udevstart and udevstart is no longer recommended to use.
395 Symlinks are now exported to the event environment. Hopefully it's no
396 longer needed to run udevinfo from an event process, like it was
397 mentioned on the hotplug list:
398 UDEV [1134776873.702967] add@/block/sdb
401 DEVLINKS=/dev/disk/by-id/usb-IBM_Memory_Key_0218B301030027E8 /dev/disk/by-path/usb-0218B301030027E8:0:0:0
405 Fix a problem if udevsend is used as the hotplug handler and tries to use
406 syslog, which causes a "vc" event loop. 2.6.15 will make udevsend obsolete
407 and this kind of problems will hopefully go away soon.
411 All built-in logic to work around bad sysfs timing is removed with this
412 version. The need to wait for sysfs files is almost fixed with a kernel
413 version that doesn't work with this udev version anyway. Until we fix
414 the timing of the "bus" link creation, the former integrated logic should
415 be emulated by a rule placed before all other rules:
416 ACTION=="add", DEVPATH=="/devices/*", ENV{PHYSDEVBUS}=="?*", WAIT_FOR_SYSFS="bus"
418 The option "udev_db" does no longer exist. All udev state will be in
419 /$udev_root/.udev/ now, there is no longer an option to set this
421 If the init script or something else used this value, just depend on
422 this hardcoded path. But remember _all_content_ of this directory is
423 still private to udev and can change at any time.
425 Default location for rule sripts and helper programs is now: /lib/udev/.
426 Everything that is not useful on the commandline should go into this
427 directory. Some of the helpers in the extras folder are installed there
428 now. The rules need to be changed, to find the helpers there.
430 Also /lib/udev/devices is recommended as a directory where packages or
431 the user can place real device nodes, which get copied over to /dev at
432 every boot. This should replace the various solutions with custom config
435 Udevsend does no longer start the udev daemon. This must be done with
436 the init script that prepares /dev on tmpfs and creates the initial nodes,
437 before starting the daemon.
441 Silent a too verbose error logging for the old hotplug.d/ dev.d/
444 The copy of klibc is removed. A systemwide installed version of klibc
445 should be used to build a klibc udev now.
449 NAME="" will not create any nodes, but execute RUN keys. To completely
450 ignore an event the OPTION "ignore_device" should be used.
452 After removal of the reorder queue, events with a TIMEOUT can be executed
453 without any queuing now.
457 Fixed bug in udevd, if inotify is not available. We depend on netlink
458 uevents now, kernels without that event source will not work with that
459 version of udev anymore.
463 The rule parsing happens now in the daemon once at startup, all udev
464 event processes inherit the already parsed rules from the daemon.
465 It is shipped with SUSE10.0 and reduces heavily the system load at
466 startup. The option to save precompiled rules and let the udev process
467 pick the them up is removed, as it's no longer needed.
469 Kernel 2.6.15 will have symlinks at /class/input pointing to the real
470 device. Libsysfs is changed to "translate" the requested link into the
471 real device path, as it would happen with the hotplug event. Otherwise
472 device removal and the udev database will not work.
474 Using 'make STRIPCMD=' will leave the binaries unstripped for debugging
477 A few improvements for vol_id, the filesytem probing code.
481 Fix a stupid typo in extras/run_directory for "make install".
483 scsi_id creates the temporary devnode now in /dev for usage with a
484 non-writable /tmp directory.
486 The uevent kernel socket buffer can carry app. 50.000 events now,
487 let's see who can break this again. :)
489 The upcoming kernel will have a new input driver core integration.
490 Some class devices are now symlinks to the real device. libsysfs
491 needs a fix for this to work correctly. Udevstart of older udev
492 versions will _not_ create these devices!
496 Fix a 'install' target in the Makefile, that prevents EXTRAS from
501 A bunch of mostly trivial bugfixes. From now on no node name or
502 symlink name can contain any character than plain whitelisted ascii
503 characters or validated utf8 byte-streams. This is needed for the
504 /dev/disk/by-label/* links, because we import untrusted data and
505 export it to the filesystem.
509 More bugfixes. If udevd was started from the kernel, we don't
510 have stdin/stdout/stderr, which broke the forked tools in some
515 Bugfix. udevstart event ordering was broken for a long time.
516 The new run_program() uncovered it, because /dev/null was not
517 available while we try to run external programs.
518 Now udevstart should create it before we run anything.
522 Minor bugfixes and some distro rules updates. If you don't have the
523 persistent disk rules in /dev/disk/by-*/* on your distro, just
524 grab it from here. :)
528 We can use socket communication now to pass events from udev to
530 RUN+="socket:/org/freedesktop/hal/udev_event"
531 will pass the whole udev event to the HAL daemon without the need
532 for a forked helper. (See ChangeLog for udevmonitor, as an example)
536 Mostly bugfixes and see ChangeLog.
538 The test for the existence of an environment value should be
540 ENV{KEY}=="*" to ENV{KEY}=="?*"
541 because "*" will not fail anymore, if the key does not exist or
546 Bugfixes and a few tweaks described in the ChangeLog.
550 Mostly a Bugfix release.
552 Added WAIT_FOR_SYSFS="<attribute>" to be able to fight against the sysfs
553 timing with custom rules.
557 We changed the internal rule storage format. Our large rule files took
558 2 MB of RAM, with the change we are down to 99kB.
560 If the device-node has been created with default name and no symlink or
561 options are to remenber, it is not longer stored in the udevdb. HAL will
562 need to be updated to work correctly with that change.
564 To overrride optimization flags, OPTFLAGS may be used now.
572 Major changes happened with this release. The goal is to take over the
573 complete kernel-event handling and provide a more efficient way to dispatch
574 kernel events. Replacing most of the current shell script logic and the
575 kernel forked helper with a netlink-daemon and a rule-based event handling.
577 o udevd listens to netlink events now. The first valid netlink event
578 will make udevd ignore any message from udevsend that contains a
579 SEQNUM, to avoid duplicate events. The forked events can be disabled
581 echo "" > /proc/sys/kernel/hotplug
582 For full support, the broken input-subsytem needs to be fixed, not to
583 bypass the driver core.
585 o /etc/dev.d/ + /etc/hotplug.d/ directory multiplexing is completely
586 removed from udev itself and must be emulated by calling small
587 helper binaries provided in the extras folder:
588 make EXTRAS=extras/run_directory/
589 will build udev_run_devd and udev_run_hotplugd, which can be called
590 from a rule if needed:
591 RUN+="/sbin/udev_run_hotplugd"
592 The recommended way to handle this is to convert all the calls from
593 the directories to explicit udev rules and get completely rid of the
594 multiplexing. (To catch a ttyUSB event, you now no longer need to
595 fork and exit 300 tty script instances you are not interested in, it
596 is just one rule that matches exactly the device.)
598 o udev handles now _all_ events not just events for class and block
599 devices, this way it is possible to control the complete event
600 behavior with udev rules. Especially useful for rules like:
601 ACTION="add", DEVPATH="/devices/*", MODALIAS=="?*", RUN+="/sbin/modprobe $modalias"
603 o As used in the modalias rule, udev supports now textual
604 substitution placeholder along with the usual format chars. This
605 needs to be documented, for now it's only visible in udev_rules_parse.c.
607 o The rule keys support now more operations. This is documented in the
608 man page. It is possible to add values to list-keys like the SYMLINK
609 and RUN list with KEY+="value" and to clear the list by assigning KEY="".
610 Also "final"-assignments are supported by using KEY:="value", which will
611 prevent changing the key by any later rule.
613 o kernel 2.6.12 has the "detached_state" attribute removed from
614 sysfs, which was used to recognize sysfs population. We switched that
615 to wait for the "bus" link, which is only available in kernels after 2.6.11.
616 Running this udev version on older kernels may cause a short delay for
619 o To provide infrastructure for persistent device naming, the id programs:
620 scsi_id, vol_id (former udev_volume_id), and ata_id (new) are able now
621 to export the probed data in environment key format:
622 pim:~ # /sbin/ata_id --export /dev/hda
623 ID_MODEL=HTS726060M9AT00
624 ID_SERIAL=MRH401M4G6UM9B
628 KERNEL="hd*[!0-9]", IMPORT="/sbin/ata_id --export $tempnode"
629 KERNEL="hd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_MODEL}_$env{ID_SERIAL}"
632 kay@pim:~> tree /dev/disk
635 | |-- HTS726060M9AT00_MRH401M4G6UM9B -> ../../hda
636 | `-- IBM-Memory_Key -> ../../sda
638 | |-- swap -> ../../hda1
639 | |-- date -> ../../sda1
640 | `-- home -> ../../hda3
642 |-- 2E08712B0870F2E7 -> ../../hda3
643 |-- 9352cfef-7687-47bc-a2a3-34cf136f72e1 -> ../../hda1
644 |-- E845-7A89 -> ../../sda1
645 `-- b2a61681-3812-4f13-a4ff-920d70604299 -> ../../hda2
647 The IMPORT= operation will import these keys in the environment and make
648 it available for later PROGRAM= and RUN= executed programs. The keys are
649 also stored in the udevdb and can be queried from there with one of the
652 o A few binaries are silently added to the repository, which can be used
653 to replay kernel events from initramfs instead of using coldplug. udevd
654 can be instructed now to queue-up events while the stored events from
655 initramfs are filled into the udevd-queue. This code is still under
656 development and there is no documentation now besides the code itself.
657 The additional binaries get compiled, but are not installed by default.
659 o There is also a temporary fix for a performance problem where too many
660 events happen in parallel and every event needs to parse the rules.
661 udev can now read precompiled rules stored on disk. This is likely to be
662 replaced by a more elegant solution in a future udev version.
666 With kernel version 2.6.12, the sysfs file "detached_state" was removed.
667 Fix for libsysfs not to expect this file was added.
671 All rules are applied now, but only the first matching rule with a NAME-key
672 will be applied. All later rules with NAME-key are completely ignored. This
673 way system supplied symlinks or permissions gets applied to user-defined
677 Please check your rules setup, if you may need to add OPTIONS="last_rule"
678 to some rules, to keep the old behavior.
680 The rules are read on "remove"-events too. That makes is possible to match
681 with keys that are available on remove (KERNEL, SUBSYSTEM, ID, ENV, ...) to
682 instruct udev to ignore an event (OPTIONS="ignore_device").
683 The new ACTION-key may be used to let a rule act only at a "remove"-event.
685 The new RUN-key supports rule-based execution of programs after device-node
686 handling. This is meant as a general replacement for the dev.d/-directories
687 to give fine grained control over the execution of programs.
689 The %s{}-sysfs format char replacement values are searched at any of the
690 devices in the device chain now, not only at the class-device.
692 We support log priority levels now. The value udev_log in udev.conf is used
693 to determine what is printed to syslog. This makes it possible to
694 run a version with compiled-in debug messages in a production environment
695 which is sometimes needed to find a bug.
696 It is still possible to supress the inclusion of _any_ syslog usage with
697 USE_LOG=false to create the smallest possible binaries if needed.
698 The configured udev_log value can be overridden with the environment variable
703 Possible use of a system-wide klibc:
704 make USE_KLIBC=true KLCC=/usr/bin/klcc all
705 will link against an external klibc and our own version will be ignored.
709 We support an unlimited count of symlinks now.
711 If USE_STATIC=true is passed to a glibc build, we link statically and use
712 a built-in userdb parser to resolve user and group names.
714 The PLACE= key is gone. It can be replaced by an ID= for a long time, because
715 we walk up the chain of physical devices to find a match.
717 The KEY="<value>" format supports '=', '==', '!=,' , '+=' now. This makes it
718 easy to skip certain attribute matches without composing rules with weird
719 character class negations like:
720 KERNEL="[!s][!c][!d]*"
721 this can now be replaced with:
723 The current simple '=' is still supported, and should work as it does today,
724 but existing rules should be converted if possible, to be better readable.
726 We have new ENV{}== key now, to match against a maximum of 5 environment
729 udevstart is its own binary again, because we don't need co carry this araound
730 with every forked event.