3 Bugfixes. DRIVER== will match only for devices that actually have
4 a driver. DRIVERS== must be used, to include the parent devices
13 Add additional check to volume_id detection of via_raid, cause
14 some company decided to put a matching pattern all over the empty
15 storage area of their music players.
19 Fix path_id for SAS devices.
23 The udev daemon can be started with --debug-trace now, which will
24 execute all events serialized to get a chance to catch a possible
25 action that crashes the box.
27 A warning is logged, if PHYSDEV* keys, the "device" link, or a parent
28 device attribute like $attr{../file} is used, only WAIT_FOR_SYSFS rules
29 are excluded from the warning. Referencing parent attributes directly
30 may break when something in the kernel driver model changes. Udev will
31 just find the attribute by walking up the parent chain.
33 Udevtrigger now sorts the list of devices depending on the device
34 dependency, so a "usb" device is triggered after the parent "pci"
39 Revert persistent-storage ata-serial '_' '-' replacement.
45 Udevtrigger can now filter the list of devices to be triggered. Matches
46 for subsystems or sysfs attributes can be specified.
48 The entries in /dev/.udev/queue and /dev/.udev/failed have changed to
49 zero-sized files to avoid pointing to /sys and confuse broken tools which
50 scan the /dev directory. To retry failed events, udevtrigger --retry-failed
53 The rules and scripts to create udev rules for persistent network
54 devices and optical drives are in the extras/rules_generator directory
55 now. If you use something similar, please consider replacing your own
56 version with this, to share the support effort. The rule_generator
57 installs its own rules into /etc/udev/rules.d.
59 The cdrom_id tool installs its own rule now in /etc/udev/rules.d, cause
60 the rule_generator depends on cdrom_id to be called in an earlier rule.
66 Renaming of some key names (the old names still work):
67 BUS -> SUBSYSTEMS, ID -> KERNELS, SYSFS -> ATTRS, DRIVER -> DRIVERS.
68 (The behavior of the key DRIVER will change soon in one of the next
69 releases, to match only the event device, please switch to DRIVERS
70 instead. If DRIVER is used, it will behave like DRIVERS, but an error
72 With the new key names, we have a more consistent and simpler scheme.
73 We can match the properties of the event device only, with: KERNEL,
74 SUBSYSTEM, ATTR, DRIVER. Or include all the parent devices in the match,
75 with: KERNELS, SUBSYSTEMS, ATTRS, DRIVERS. ID, BUS, SYSFS, DRIVER are no
76 longer mentioned in the man page and should be switched in the rule
79 ATTR{file}="value" can be used now, to write to a sysfs file of the
80 event device. Instead of:
81 ..., SYSFS{type}=="0|7|14", RUN+="/bin/sh -c 'echo 60 > /sys$$DEVPATH/timeout'"
83 ..., ATTR{type}=="0|7|14", ATTR{timeout}="60"
85 All the PHYSDEV* keys are deprecated and will be removed from a
87 PHYDEVPATH - is the path of a parent device and should not be
89 PHYSDEVBUS - is just a SUBSYSTEM value of a parent, and can be
90 matched with SUBSYSTEMS==
91 PHYSDEVDRIVER - for bus devices it is available as ENV{DRIVER}.
92 Newer kernels will have DRIVER in the environment,
93 for older kernels udev puts in. Class device will
94 no longer carry this property of a parent and
95 DRIVERS== can be used to match such a parent value.
96 Note that ENV{DRIVER} is only available for a few bus devices, where
97 the driver is already bound at device event time. On coldplug, the
98 events for a lot devices are already bound to a driver, and they will have
99 that value set. But on hotplug, at the time the kernel creates the device,
100 it can't know what driver may claim the device after that, therefore
101 in most cases it will be empty.
103 Failed events should now be re-triggered with:
104 udevtrigger --retry-failed.
105 Please switch to this command, so we keep the details of the /dev/.udev/failed/
106 files private to the udev tools. We may need to switch the current symlink
107 target, cause some obviously broken tools try to scan all files in /dev
108 including /dev/.udev/, find the links to /sys and end up stat()'ing sysfs files
109 million times. This takes ages on slow boxes.
111 The udevinfo attribute walk (-a) now works with giving a device node
112 name (-n) instead of a devpath (-p). The query now always works, also when
113 no database file was created by udev.
115 The built-in /etc/passwd /etc/group parser is removed, we always depend on
116 getpwnam() and getgrnam() now. One of the next releases will depend on
117 fnmatch() and may use getopt_long().
121 Bugfixes and small improvements.
125 Fix path_id for recent kernels.
131 Added support for swapping network interface names, by temporarily
132 renaming the device and wait for the target name to become free.
136 The built-in MODALIAS key and substitution is removed.
140 The binary firmware helper is replaced by the usual simple
141 shell script. Udevsend is removed from the tree.
149 Some more keys require the correct use of '==' and '=' depending
150 on the kind of operation beeing an assignment or a match. Rules
151 with invalid operations are skipped and logged to syslog. Please
152 test with udevtest if the parsing of your rules throws errors and
153 fix possibly broken rules.
157 Provide "udevsettle" to wait for all current udev events to finish.
158 It also watches the current kernel netlink queue by comparing the
159 even sequence number to make sure that there are no current pending
160 events that have not already arrived in the daemon.
164 Fix rule to skip persistent rules for removable IDE devices, which
165 also skipped optical IDE drives.
167 All *_id program are installed in /lib/udev/ by default now.
169 No binary is stripped anymore as this should be done in the
170 packaging process and not at build time.
172 libvolume_id is provided as a shared library now and vol_id is
173 linked against it. Also one of the next HAL versions will require
174 this library, and the HAL build process will also require the
175 header file to be installed. The copy of the same code in HAL will
176 be removed to have only a single copy left on the system.
180 Add persistent links for SCSI tapes. The rules file is renamed
181 to 60-persistent-storage.rules.
183 Create persistent path for usb devices. Can be used for all sorts
184 of devices that can't be distinguished by other properties like
185 multiple identical keyboards and mice connected to the same box.
187 Provide "udevtrigger" program to request events on coldplug. The
188 shell script is much too slow with thousends of devices.
192 Fix persistent disk rules to exclude removable IDE drives.
194 Warn if %e, $modalias or MODALIAS is used.
198 Fix queue export, which wasn't correct for subsequent add/remove
199 events for the same device.
203 Fix cramfs detection on big endian.
205 Make WAIT_FOR_SYSFS usable in "normal" rules and silent if the whole
210 If BUS== and SYSFS{}== have been used in the same rule, the sysfs
211 attributes were only checked at the parent device that matched the
212 by BUS requested subsystem. Fix it to also look at the device we
213 received the event for.
215 Build variable CROSS has changed to CROSS_COMPILE to match the kernel
220 Fix a bug where NAME="" would prevent RUN from beeing executed.
222 RUN="/bin/program" does not longer automatically add the subsystem
223 as the first parameter. This is from the days of /sbin/hotplug
224 which is dead now and it's just confusing to need to add a space at
225 the end of the program name to prevent this.
226 If you use rules that need the subsystem as the first parameter,
227 like the old "udev_run_hotlugd" and "udev_run_devd", add the subsystem
228 to the key like RUN+="/bin/program $env{SUBSYSTEM}".
232 The udev man page has moved to udev(7) as it does not describe a command
233 anymore. The programs udev, udevstart and udevsend are no longer installed
234 by default and must be copied manually, if they should be installed or
235 included in a package.
237 Fix a bug where "ignore_device" could run earlier collected RUN keys before
238 the ignore rule was applied.
240 More preparation for future sysfs changes. usb_id and scsi_id no longer
241 depend on a magic order of devices in the /devices chain. Specific devices
242 should be requested by their subsytem.
244 This will always find the scsi parent device without depending on a specific
246 dev = sysfs_device_get(devpath);
247 dev_usb = sysfs_device_get_parent_with_subsystem(dev, "scsi");
249 The "device" link in the current sysfs layout will be automatically
250 _resolved_ as a parent and in the new sysfs layout it will just _be_ the
251 parent in the devpath. If a device is requested by it's symlink, like all
252 class devices in the new sysfs layout will look like, it gets automatically
253 resolved and substituted with the real devpath and not the symlink path.
256 A similar logic must be applied to _all_ sysfs users, including
257 scripts, that search along parent devices in sysfs. The explicit use of
258 the "device" link must be avoided. With the future sysfs layout all
259 DEVPATH's will start with /devices/ and have a "subsystem" symlink poiting
260 back to the "class" or the "bus". The layout of the parent devices in
261 /devices is not necessarily expected to be stable across kernel releases and
262 searching for parents by their subsystem should make sysfs users tolerant
263 for changed parent chains.
267 Prepare udev to work with the experimental kernel patch, that moves
268 /sys/class devices to /sys/devices and /sys/block to /sys/class/block.
270 Clarify BUS, ID, $id usage and fix $id behavior. This prepares for
271 moving the class devices to /sys/devices.
273 Thanks again to Marco for help finding a hopefully nice compromise
274 to make %b simpler and working again.
278 Complete removal of libsysfs, replaced by simple helper functions
279 which are much simpler and a bit faster. The udev daemon operatesentirely
280 on event parameters and does not use sysfs for simple rules anymore.
281 Please report any new bugs/problems, that may be caused by this big
282 change. They will be fixed immediately.
284 The enumeration format character '%e' is deprecated and will be
285 removed sometimes from a future udev version. It never worked correctly
286 outside of udevstart, so we can't use it with the new parallel
287 coldplug. A simple enumeration is as useless as the devfs naming
288 scheme, just get rid of both if you still use it.
290 MODALIAS and $modalias is not needed and will be removed from one of
291 the next udev versions, replace it in all rules with ENV{MODALIAS} or
292 the sysfs "modalias" value.
294 Thanks a lot to Marco for all his help on finding and fixing bugs.
298 Let scsi_id request libata drive serial numbers from page 0x80.
300 Renamed etc/udev/persistent.rules to persistent-disk.rules and
301 added /dev/disk/by-name/* for device mapper device names.
303 Removed %e from the man page. It never worked reliably outside
304 of udevstart and udevstart is no longer recommended to use.
308 Symlinks are now exported to the event environment. Hopefully it's no
309 longer needed to run udevinfo from an event process, like it was
310 mentioned on the hotplug list:
311 UDEV [1134776873.702967] add@/block/sdb
314 DEVLINKS=/dev/disk/by-id/usb-IBM_Memory_Key_0218B301030027E8 /dev/disk/by-path/usb-0218B301030027E8:0:0:0
318 Fix a problem if udevsend is used as the hotplug handler and tries to use
319 syslog, which causes a "vc" event loop. 2.6.15 will make udevsend obsolete
320 and this kind of problems will hopefully go away soon.
324 All built-in logic to work around bad sysfs timing is removed with this
325 version. The need to wait for sysfs files is almost fixed with a kernel
326 version that doesn't work with this udev version anyway. Until we fix
327 the timing of the "bus" link creation, the former integrated logic should
328 be emulated by a rule placed before all other rules:
329 ACTION=="add", DEVPATH=="/devices/*", ENV{PHYSDEVBUS}=="?*", WAIT_FOR_SYSFS="bus"
331 The option "udev_db" does no longer exist. All udev state will be in
332 /$udev_root/.udev/ now, there is no longer an option to set this
334 If the init script or something else used this value, just depend on
335 this hardcoded path. But remember _all_content_ of this directory is
336 still private to udev and can change at any time.
338 Default location for rule sripts and helper programs is now: /lib/udev/.
339 Everything that is not useful on the commandline should go into this
340 directory. Some of the helpers in the extras folder are installed there
341 now. The rules need to be changed, to find the helpers there.
343 Also /lib/udev/devices is recommended as a directory where packages or
344 the user can place real device nodes, which get copied over to /dev at
345 every boot. This should replace the various solutions with custom config
348 Udevsend does no longer start the udev daemon. This must be done with
349 the init script that prepares /dev on tmpfs and creates the initial nodes,
350 before starting the daemon.
354 Silent a too verbose error logging for the old hotplug.d/ dev.d/
357 The copy of klibc is removed. A systemwide installed version of klibc
358 should be used to build a klibc udev now.
362 NAME="" will not create any nodes, but execute RUN keys. To completely
363 ignore an event the OPTION "ignore_device" should be used.
365 After removal of the reorder queue, events with a TIMEOUT can be executed
366 without any queuing now.
370 Fixed bug in udevd, if inotify is not available. We depend on netlink
371 uevents now, kernels without that event source will not work with that
372 version of udev anymore.
376 The rule parsing happens now in the daemon once at startup, all udev
377 event processes inherit the already parsed rules from the daemon.
378 It is shipped with SUSE10.0 and reduces heavily the system load at
379 startup. The option to save precompiled rules and let the udev process
380 pick the them up is removed, as it's no longer needed.
382 Kernel 2.6.15 will have symlinks at /class/input pointing to the real
383 device. Libsysfs is changed to "translate" the requested link into the
384 real device path, as it would happen with the hotplug event. Otherwise
385 device removal and the udev database will not work.
387 Using 'make STRIPCMD=' will leave the binaries unstripped for debugging
390 A few improvements for vol_id, the filesytem probing code.
394 Fix a stupid typo in extras/run_directory for "make install".
396 scsi_id creates the temporary devnode now in /dev for usage with a
397 non-writable /tmp directory.
399 The uevent kernel socket buffer can carry app. 50.000 events now,
400 let's see who can break this again. :)
402 The upcoming kernel will have a new input driver core integration.
403 Some class devices are now symlinks to the real device. libsysfs
404 needs a fix for this to work correctly. Udevstart of older udev
405 versions will _not_ create these devices!
409 Fix a 'install' target in the Makefile, that prevents EXTRAS from
414 A bunch of mostly trivial bugfixes. From now on no node name or
415 symlink name can contain any character than plain whitelisted ascii
416 characters or validated utf8 byte-streams. This is needed for the
417 /dev/disk/by-label/* links, because we import untrusted data and
418 export it to the filesystem.
422 More bugfixes. If udevd was started from the kernel, we don't
423 have stdin/stdout/stderr, which broke the forked tools in some
428 Bugfix. udevstart event ordering was broken for a long time.
429 The new run_program() uncovered it, because /dev/null was not
430 available while we try to run external programs.
431 Now udevstart should create it before we run anything.
435 Minor bugfixes and some distro rules updates. If you don't have the
436 persistent disk rules in /dev/disk/by-*/* on your distro, just
437 grab it from here. :)
441 We can use socket communication now to pass events from udev to
443 RUN+="socket:/org/freedesktop/hal/udev_event"
444 will pass the whole udev event to the HAL daemon without the need
445 for a forked helper. (See ChangeLog for udevmonitor, as an example)
449 Mostly bugfixes and see ChangeLog.
451 The test for the existence of an environment value should be
453 ENV{KEY}=="*" to ENV{KEY}=="?*"
454 because "*" will not fail anymore, if the key does not exist or
459 Bugfixes and a few tweaks described in the ChangeLog.
463 Mostly a Bugfix release.
465 Added WAIT_FOR_SYSFS="<attribute>" to be able to fight against the sysfs
466 timing with custom rules.
470 We changed the internal rule storage format. Our large rule files took
471 2 MB of RAM, with the change we are down to 99kB.
473 If the device-node has been created with default name and no symlink or
474 options are to remenber, it is not longer stored in the udevdb. HAL will
475 need to be updated to work correctly with that change.
477 To overrride optimization flags, OPTFLAGS may be used now.
485 Major changes happened with this release. The goal is to take over the
486 complete kernel-event handling and provide a more efficient way to dispatch
487 kernel events. Replacing most of the current shell script logic and the
488 kernel forked helper with a netlink-daemon and a rule-based event handling.
490 o udevd listens to netlink events now. The first valid netlink event
491 will make udevd ignore any message from udevsend that contains a
492 SEQNUM, to avoid duplicate events. The forked events can be disabled
494 echo "" > /proc/sys/kernel/hotplug
495 For full support, the broken input-subsytem needs to be fixed, not to
496 bypass the driver core.
498 o /etc/dev.d/ + /etc/hotplug.d/ directory multiplexing is completely
499 removed from udev itself and must be emulated by calling small
500 helper binaries provided in the extras folder:
501 make EXTRAS=extras/run_directory/
502 will build udev_run_devd and udev_run_hotplugd, which can be called
503 from a rule if needed:
504 RUN+="/sbin/udev_run_hotplugd"
505 The recommended way to handle this is to convert all the calls from
506 the directories to explicit udev rules and get completely rid of the
507 multiplexing. (To catch a ttyUSB event, you now no longer need to
508 fork and exit 300 tty script instances you are not interested in, it
509 is just one rule that matches exactly the device.)
511 o udev handles now _all_ events not just events for class and block
512 devices, this way it is possible to control the complete event
513 behavior with udev rules. Especially useful for rules like:
514 ACTION="add", DEVPATH="/devices/*", MODALIAS=="?*", RUN+="/sbin/modprobe $modalias"
516 o As used in the modalias rule, udev supports now textual
517 substitution placeholder along with the usual format chars. This
518 needs to be documented, for now it's only visible in udev_rules_parse.c.
520 o The rule keys support now more operations. This is documented in the
521 man page. It is possible to add values to list-keys like the SYMLINK
522 and RUN list with KEY+="value" and to clear the list by assigning KEY="".
523 Also "final"-assignments are supported by using KEY:="value", which will
524 prevent changing the key by any later rule.
526 o kernel 2.6.12 has the "detached_state" attribute removed from
527 sysfs, which was used to recognize sysfs population. We switched that
528 to wait for the "bus" link, which is only available in kernels after 2.6.11.
529 Running this udev version on older kernels may cause a short delay for
532 o To provide infrastructure for persistent device naming, the id programs:
533 scsi_id, vol_id (former udev_volume_id), and ata_id (new) are able now
534 to export the probed data in environment key format:
535 pim:~ # /sbin/ata_id --export /dev/hda
536 ID_MODEL=HTS726060M9AT00
537 ID_SERIAL=MRH401M4G6UM9B
541 KERNEL="hd*[!0-9]", IMPORT="/sbin/ata_id --export $tempnode"
542 KERNEL="hd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_MODEL}_$env{ID_SERIAL}"
545 kay@pim:~> tree /dev/disk
548 | |-- HTS726060M9AT00_MRH401M4G6UM9B -> ../../hda
549 | `-- IBM-Memory_Key -> ../../sda
551 | |-- swap -> ../../hda1
552 | |-- date -> ../../sda1
553 | `-- home -> ../../hda3
555 |-- 2E08712B0870F2E7 -> ../../hda3
556 |-- 9352cfef-7687-47bc-a2a3-34cf136f72e1 -> ../../hda1
557 |-- E845-7A89 -> ../../sda1
558 `-- b2a61681-3812-4f13-a4ff-920d70604299 -> ../../hda2
560 The IMPORT= operation will import these keys in the environment and make
561 it available for later PROGRAM= and RUN= executed programs. The keys are
562 also stored in the udevdb and can be queried from there with one of the
565 o A few binaries are silently added to the repository, which can be used
566 to replay kernel events from initramfs instead of using coldplug. udevd
567 can be instructed now to queue-up events while the stored events from
568 initramfs are filled into the udevd-queue. This code is still under
569 development and there is no documentation now besides the code itself.
570 The additional binaries get compiled, but are not installed by default.
572 o There is also a temporary fix for a performance problem where too many
573 events happen in parallel and every event needs to parse the rules.
574 udev can now read precompiled rules stored on disk. This is likely to be
575 replaced by a more elegant solution in a future udev version.
579 With kernel version 2.6.12, the sysfs file "detached_state" was removed.
580 Fix for libsysfs not to expect this file was added.
584 All rules are applied now, but only the first matching rule with a NAME-key
585 will be applied. All later rules with NAME-key are completely ignored. This
586 way system supplied symlinks or permissions gets applied to user-defined
590 Please check your rules setup, if you may need to add OPTIONS="last_rule"
591 to some rules, to keep the old behavior.
593 The rules are read on "remove"-events too. That makes is possible to match
594 with keys that are available on remove (KERNEL, SUBSYSTEM, ID, ENV, ...) to
595 instruct udev to ignore an event (OPTIONS="ignore_device").
596 The new ACTION-key may be used to let a rule act only at a "remove"-event.
598 The new RUN-key supports rule-based execution of programs after device-node
599 handling. This is meant as a general replacement for the dev.d/-directories
600 to give fine grained control over the execution of programs.
602 The %s{}-sysfs format char replacement values are searched at any of the
603 devices in the device chain now, not only at the class-device.
605 We support log priority levels now. The value udev_log in udev.conf is used
606 to determine what is printed to syslog. This makes it possible to
607 run a version with compiled-in debug messages in a production environment
608 which is sometimes needed to find a bug.
609 It is still possible to supress the inclusion of _any_ syslog usage with
610 USE_LOG=false to create the smallest possible binaries if needed.
611 The configured udev_log value can be overridden with the environment variable
616 Possible use of a system-wide klibc:
617 make USE_KLIBC=true KLCC=/usr/bin/klcc all
618 will link against an external klibc and our own version will be ignored.
622 We support an unlimited count of symlinks now.
624 If USE_STATIC=true is passed to a glibc build, we link statically and use
625 a built-in userdb parser to resolve user and group names.
627 The PLACE= key is gone. It can be replaced by an ID= for a long time, because
628 we walk up the chain of physical devices to find a match.
630 The KEY="<value>" format supports '=', '==', '!=,' , '+=' now. This makes it
631 easy to skip certain attribute matches without composing rules with weird
632 character class negations like:
633 KERNEL="[!s][!c][!d]*"
634 this can now be replaced with:
636 The current simple '=' is still supported, and should work as it does today,
637 but existing rules should be converted if possible, to be better readable.
639 We have new ENV{}== key now, to match against a maximum of 5 environment
642 udevstart is its own binary again, because we don't need co carry this araound
643 with every forked event.