3 Symlinks are now exported to the event environment. Hopefully it's no
4 longer needed to run udevinfo from an event process, like it was
5 mentioned on the hotplug list:
6 UDEV [1134776873.702967] add@/block/sdb
9 DEVLINKS=/dev/disk/by-id/usb-IBM_Memory_Key_0218B301030027E8 /dev/disk/by-path/usb-0218B301030027E8:0:0:0
13 Fix a problem if udevsend is used as the hotplug handler and tries to use
14 syslog, which causes a "vc" event loop. 2.6.15 will make udevsend obsolete
15 and this kind of problems will hopefully go away soon.
19 All built-in logic to work around bad sysfs timing is removed with this
20 version. The need to wait for sysfs files is almost fixed with a kernel
21 version that doesn't work with this udev version anyway. Until we fix
22 the timing of the "bus" link creation, the former integrated logic should
23 be emulated by a rule placed before all other rules:
24 ACTION=="add", DEVPATH=="/devices/*", ENV{PHYSDEVBUS}=="?*", WAIT_FOR_SYSFS="bus"
26 The option "udev_db" does no longer exist. All udev state will be in
27 /$udev_root/.udev/ now, there is no longer an option to set this
29 If the init script or something else used this value, just depend on
30 this hardcoded path. But remember _all_content_ of this directory is
31 still private to udev and can change at any time.
33 Default location for rule sripts and helper programs is now: /lib/udev/.
34 Everything that is not useful on the commandline should go into this
35 directory. Some of the helpers in the extras folder are installed there
36 now. The rules need to be changed, to find the helpers there.
38 Also /lib/udev/devices is recommended as a directory where packages or
39 the user can place real device nodes, which get copied over to /dev at
40 every boot. This should replace the various solutions with custom config
43 Udevsend does no longer start the udev daemon. This must be done with
44 the init script that prepares /dev on tmpfs and creates the initial nodes,
45 before starting the daemon.
49 Silent a too verbose error logging for the old hotplug.d/ dev.d/
52 The copy of klibc is removed. A systemwide installed version of klibc
53 should be used to build a klibc udev now.
57 NAME="" will not create any nodes, but execute RUN keys. To completely
58 ignore an event the OPTION "ignore_device" should be used.
60 After removal of the reorder queue, events with a TIMEOUT can be executed
61 without any queuing now.
65 Fixed bug in udevd, if inotify is not available. We depend on netlink
66 uevents now, kernels without that event source will not work with that
67 version of udev anymore.
71 The rule parsing happens now in the daemon once at startup, all udev
72 event processes inherit the already parsed rules from the daemon.
73 It is shipped with SUSE10.0 and reduces heavily the system load at
74 startup. The option to save precompiled rules and let the udev process
75 pick the them up is removed, as it's no longer needed.
77 Kernel 2.6.15 will have symlinks at /class/input pointing to the real
78 device. Libsysfs is changed to "translate" the requested link into the
79 real device path, as it would happen with the hotplug event. Otherwise
80 device removal and the udev database will not work.
82 Using 'make STRIPCMD=' will leave the binaries unstripped for debugging
85 A few improvements for vol_id, the filesytem probing code.
89 Fix a stupid typo in extras/run_directory for "make install".
91 scsi_id creates the temporary devnode now in /dev for usage with a
92 non-writable /tmp directory.
94 The uevent kernel socket buffer can carry app. 50.000 events now,
95 let's see who can break this again. :)
97 The upcoming kernel will have a new input driver core integration.
98 Some class devices are now symlinks to the real device. libsysfs
99 needs a fix for this to work correctly. Udevstart of older udev
100 versions will _not_ create these devices!
104 Fix a 'install' target in the Makefile, that prevents EXTRAS from
109 A bunch of mostly trivial bugfixes. From now on no node name or
110 symlink name can contain any character than plain whitelisted ascii
111 characters or validated utf8 byte-streams. This is needed for the
112 /dev/disk/by-label/* links, cause we import untrusted data and
113 export it to the filesystem.
117 More bugfixes. If udevd was started from the kernel, we don't
118 have stdin/stdout/stderr, which broke the forked tools in some
123 Bugfix. udevstart event ordering was broken for a long time.
124 The new run_program() uncovered it, cause /dev/null was not
125 available while we try to run external programs.
126 Now udevstart should create it before we run anything.
130 Minor bugfixes and some distro rules updates. If you don't have the
131 persistent disk rules in /dev/disk/by-*/* on your distro, just
132 grab it from here. :)
136 We can use socket communication now to pass events from udev to
138 RUN+="socket:/org/freedesktop/hal/udev_event"
139 will pass the whole udev event to the HAL daemon without the need
140 for a forked helper. (See ChangeLog for udevmonitor, as an example)
144 Mostly bugfixes and see ChangeLog.
146 The test for the existence of an environment value should be
148 ENV{KEY}=="*" to ENV{KEY}=="?*"
149 cause "*" will not fail anymore, if the key does not exist or
154 Bugfixes and a few tweaks described in the ChangeLog.
158 Mostly a Bugfix release.
160 Added WAIT_FOR_SYSFS="<attribute>" to be able to fight against the sysfs
161 timing with custom rules.
165 We changed the internal rule storage format. Our large rule files took
166 2 MB of RAM, with the change we are down to 99kB.
168 If the device-node has been created with default name and no symlink or
169 options are to remenber, it is not longer stored in the udevdb. HAL will
170 need to be updated to work correctly with that change.
172 To overrride optimization flags, OPTFLAGS may be used now.
180 Major changes happened with this release. The goal is to take over the
181 complete kernel-event handling and provide a more efficient way to dispatch
182 kernel events. Replacing most of the current shell script logic and the
183 kernel forked helper with a netlink-daemon and a rule-based event handling.
185 o udevd listens to netlink events now. The first valid netlink event
186 will make udevd ignore any message from udevsend that contains a
187 SEQNUM, to avoid duplicate events. The forked events can be disabled
189 echo "" > /proc/sys/kernel/hotplug
190 For full support, the broken input-subsytem needs to be fixed, not to
191 bypass the driver core.
193 o /etc/dev.d/ + /etc/hotplug.d/ directory multiplexing is completely
194 removed from udev itself and must be emulated by calling small
195 helper binaries provided in the extras folder:
196 make EXTRAS=extras/run_directory/
197 will build udev_run_devd and udev_run_hotplugd, which can be called
198 from a rule if needed:
199 RUN+="/sbin/udev_run_hotplugd"
200 The recommended way to handle this is to convert all the calls from
201 the directories to explicit udev rules and get completely rid of the
202 multiplexing. (To catch a ttyUSB event, you now no longer need to
203 fork and exit 300 tty script instances you are not interested in, it
204 is just one rule that matches exactly the device.)
206 o udev handles now _all_ events not just events for class and block
207 devices, this way it is possible to control the complete event
208 behavior with udev rules. Especially useful for rules like:
209 ACTION="add", DEVPATH="/devices/*", MODALIAS=="?*", RUN+="/sbin/modprobe $modalias"
211 o As used in the modalias rule, udev supports now textual
212 substitution placeholder along with the usual format chars. This
213 needs to be documented, for now it's only visible in udev_rules_parse.c.
215 o The rule keys support now more operations. This is documented in the
216 man page. It is possible to add values to list-keys like the SYMLINK
217 and RUN list with KEY+="value" and to clear the list by assigning KEY="".
218 Also "final"-assignments are supported by using KEY:="value", which will
219 prevent changing the key by any later rule.
221 o kernel 2.6.12 has the "detached_state" attribute removed from
222 sysfs, which was used to recognize sysfs population. We switched that
223 to wait for the "bus" link, which is only available in kernels after 2.6.11.
224 Running this udev version on older kernels may cause a short delay for
227 o To provide infrastructure for persistent device naming, the id programs:
228 scsi_id, vol_id (former udev_volume_id), and ata_id (new) are able now
229 to export the probed data in environment key format:
230 pim:~ # /sbin/ata_id --export /dev/hda
231 ID_MODEL=HTS726060M9AT00
232 ID_SERIAL=MRH401M4G6UM9B
236 KERNEL="hd*[!0-9]", IMPORT="/sbin/ata_id --export $tempnode"
237 KERNEL="hd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_MODEL}_$env{ID_SERIAL}"
240 kay@pim:~> tree /dev/disk
243 | |-- HTS726060M9AT00_MRH401M4G6UM9B -> ../../hda
244 | `-- IBM-Memory_Key -> ../../sda
246 | |-- swap -> ../../hda1
247 | |-- date -> ../../sda1
248 | `-- home -> ../../hda3
250 |-- 2E08712B0870F2E7 -> ../../hda3
251 |-- 9352cfef-7687-47bc-a2a3-34cf136f72e1 -> ../../hda1
252 |-- E845-7A89 -> ../../sda1
253 `-- b2a61681-3812-4f13-a4ff-920d70604299 -> ../../hda2
255 The IMPORT= operation will import these keys in the environment and make
256 it available for later PROGRAM= and RUN= executed programs. The keys are
257 also stored in the udevdb and can be queried from there with one of the
260 o A few binaries are silently added to the repository, which can be used
261 to replay kernel events from initramfs instead of using coldplug. udevd
262 can be instructed now to queue-up events while the stored events from
263 initramfs are filled into the udevd-queue. This code is still under
264 development and there is no documentation now besides the code itself.
265 The additional binaries get compiled, but are not installed by default.
267 o There is also a temporary fix for a performance problem where too many
268 events happen in parallel and every event needs to parse the rules.
269 udev can now read precompiled rules stored on disk. This is likely to be
270 replaced by a more elegant solution in a future udev version.
274 With kernel version 2.6.12, the sysfs file "detached_state" was removed.
275 Fix for libsysfs not to expect this file was added.
279 All rules are applied now, but only the first matching rule with a NAME-key
280 will be applied. All later rules with NAME-key are completely ignored. This
281 way system supplied symlinks or permissions gets applied to user-defined
285 Please check your rules setup, if you may need to add OPTIONS="last_rule"
286 to some rules, to keep the old behavior.
288 The rules are read on "remove"-events too. That makes is possible to match
289 with keys that are available on remove (KERNEL, SUBSYSTEM, ID, ENV, ...) to
290 instruct udev to ignore an event (OPTIONS="ignore_device").
291 The new ACTION-key may be used to let a rule act only at a "remove"-event.
293 The new RUN-key supports rule-based execution of programs after device-node
294 handling. This is meant as a general replacement for the dev.d/-directories
295 to give fine grained control over the execution of programs.
297 The %s{}-sysfs format char replacement values are searched at any of the
298 devices in the device chain now, not only at the class-device.
300 We support log priority levels now. The value udev_log in udev.conf is used
301 to determine what is printed to syslog. This makes it possible to
302 run a version with compiled-in debug messages in a production environment
303 which is sometimes needed to find a bug.
304 It is still possible to supress the inclusion of _any_ syslog usage with
305 USE_LOG=false to create the smallest possible binaries if needed.
306 The configured udev_log value can be overridden with the environment variable
311 Possible use of a system-wide klibc:
312 make USE_KLIBC=true KLCC=/usr/bin/klcc all
313 will link against an external klibc and our own version will be ignored.
317 We support an unlimited count of symlinks now.
319 If USE_STATIC=true is passed to a glibc build, we link statically and use
320 a built-in userdb parser to resolve user and group names.
322 The PLACE= key is gone. It can be replaced by an ID= for a long time, cause
323 we walk up the chain of physical devices to find a match.
325 The KEY="<value>" format supports '=', '==', '!=,' , '+=' now. This makes it
326 easy to skip certain attribute matches without composing rules with weird
327 character class negations like:
328 KERNEL="[!s][!c][!d]*"
329 this can now be replaced with:
331 The current simple '=' is still supported, and should work as it does today,
332 but existing rules should be converted if possible, to be better readable.
334 We have new ENV{}== key now, to match against a maximum of 5 environment
337 udevstart is its own binary again, cause we don't need co carry this araound
338 with every forked event.