chiark / gitweb /
elogind.git
18 years agovolume_id: add Veritas fs
Kay Sievers [Sun, 23 Oct 2005 14:31:13 +0000 (16:31 +0200)]
volume_id: add Veritas fs

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agovolume_id: add OVFS Version 1
Kay Sievers [Sun, 23 Oct 2005 14:08:56 +0000 (16:08 +0200)]
volume_id: add OVFS Version 1

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years ago071 release
Kay Sievers [Wed, 19 Oct 2005 22:28:34 +0000 (00:28 +0200)]
071 release

18 years agolibsysfs: accept sysmlinks to directories instead of real directories
Kay Sievers [Wed, 19 Oct 2005 21:59:54 +0000 (23:59 +0200)]
libsysfs: accept sysmlinks to directories instead of real directories

Recent changes to the sysfs layout introduced class-devices pointed to
by a symlink, instead of the real object at that location.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix selinux compilation
Kay Sievers [Sun, 16 Oct 2005 12:16:43 +0000 (14:16 +0200)]
fix selinux compilation

Taken from Red Hat CVS.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevd: move logging from err to info for non-hotplug uevent
Kay Sievers [Wed, 12 Oct 2005 09:55:49 +0000 (11:55 +0200)]
udevd: move logging from err to info for non-hotplug uevent

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoforce event socket buffer size to 16MB
Kay Sievers [Tue, 11 Oct 2005 19:06:50 +0000 (21:06 +0200)]
force event socket buffer size to 16MB

Netlink events get lost when the kernel creates thousends of events
faster than udevd reads it. The default is 128 KB, which can carry
app. 500 events. Set it to 16 MB now.

  I have 4000 fibrechannel LUNs connected to my system.  There are two
  paths to the devices and two ports on the host connected via a switch.
  This gives 16000 when probed.
  I have had problems getting all of the entries in /dev created.
  -- Mark Haverkamp <markh@osdl.org>

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevd: fix initial timeout handling
Kay Sievers [Mon, 3 Oct 2005 14:36:52 +0000 (16:36 +0200)]
udevd: fix initial timeout handling

UDEVD_EVENT_TIMEOUT=0 didn't work directly after udevd startup.
The whole event timeout handling is not needed since we use netlink.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoclarify README
Kay Sievers [Mon, 3 Oct 2005 14:28:11 +0000 (16:28 +0200)]
clarify README

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoMerge branch 'master' from gregkh@master.kernel.org:/pub/scm/linux/hotplug/udev
<gregkh@suse.de> [Tue, 27 Sep 2005 15:12:09 +0000 (08:12 -0700)]
Merge branch 'master' from gregkh@/linux/hotplug/udev

18 years agoRemove the udev.spec file as no one uses it anymore
<gregkh@suse.de> [Tue, 27 Sep 2005 15:11:31 +0000 (08:11 -0700)]
Remove the udev.spec file as no one uses it anymore

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years agoupdate Debian persistent rules
Kay Sievers [Tue, 27 Sep 2005 14:43:27 +0000 (16:43 +0200)]
update Debian persistent rules

18 years agoadd edd-*-part%n to the persistent.rules
Kay Sievers [Tue, 27 Sep 2005 14:38:43 +0000 (16:38 +0200)]
add edd-*-part%n to the persistent.rules

Thanks to:
  Marco d'Itri <md@Linux.IT>

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoswitch tools and volume_id from LGPL to GPLv2
Kay Sievers [Tue, 27 Sep 2005 14:27:35 +0000 (16:27 +0200)]
switch tools and volume_id from LGPL to GPLv2

After beeing hit by proprietary applications which statically
link the LGPL'd libusb, which needs a patch to reflect the recent
kernel changes, I decided not to provide LGPL code anymore.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoremove no longer needed includes
Kay Sievers [Tue, 27 Sep 2005 13:05:34 +0000 (15:05 +0200)]
remove no longer needed includes

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate Debian rules
Kay Sievers [Fri, 23 Sep 2005 07:51:49 +0000 (09:51 +0200)]
update Debian rules

18 years agoupdate Fedora rules
Kay Sievers [Sun, 18 Sep 2005 03:37:33 +0000 (05:37 +0200)]
update Fedora rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate Debian rules
Kay Sievers [Sun, 18 Sep 2005 03:34:05 +0000 (05:34 +0200)]
update Debian rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate devfs rules header
Kay Sievers [Sun, 18 Sep 2005 03:08:52 +0000 (05:08 +0200)]
update devfs rules header

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agovolume_id: set reiser instead of reiserfs for filesystem type
Kay Sievers [Sun, 18 Sep 2005 03:06:10 +0000 (05:06 +0200)]
volume_id: set reiser instead of reiserfs for filesystem type

Thanks to Theodore Ts'o <tytso@mit.edu> for the hint and a test image.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoscsi_id: switch temporary node creation to /dev
Kay Sievers [Sun, 18 Sep 2005 02:49:23 +0000 (04:49 +0200)]
scsi_id: switch temporary node creation to /dev

/tmp is not writable on most systems, so just use /dev to create a
temporary node. If called from a udev rule, "-d $tempnode" should be
used and udev will create a temporary node and pass the name before
calling scsi_id.
Also remove gen_scsi_id_udev_rules.sh per Patrick's request, as it's
no longer needed with the persistent disk links.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd "Persistent Device Naming" rules file for disks
Kay Sievers [Sun, 18 Sep 2005 02:41:09 +0000 (04:41 +0200)]
add "Persistent Device Naming" rules file for disks

Based on:
  "Linux Persistent Device Naming", 2004, Hannes Reinecke <hare@suse.de>

18 years agoata_id: open volume O_NONBLOCK
Kay Sievers [Fri, 16 Sep 2005 19:23:36 +0000 (21:23 +0200)]
ata_id: open volume O_NONBLOCK

Opticals like that and should be ok for disks too, as long as we
don't read().

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoedd_id: check that EDD id is unique
John Hull [Fri, 16 Sep 2005 19:10:12 +0000 (21:10 +0200)]
edd_id: check that EDD id is unique

The attached patch adds a check to edd_id.c to verify that the MBR
signature on the device node passed to the program is unique to only
that disk.

Signed-off-by: John Hull <john_hull@dell.com>
18 years agorun_directory: fix typo in "make install"
Marco d'Itri [Fri, 16 Sep 2005 19:05:15 +0000 (21:05 +0200)]
run_directory: fix typo in "make install"

18 years ago070 release
Greg KH [Wed, 14 Sep 2005 23:19:07 +0000 (16:19 -0700)]
070 release

18 years agoupdate RELEASE-NOTES
Kay Sievers [Wed, 14 Sep 2005 19:12:08 +0000 (21:12 +0200)]
update RELEASE-NOTES

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoscsi_id: derive a UID for a SCSI-2 not compliant with the page 83
Edward Goggin [Wed, 14 Sep 2005 18:23:48 +0000 (20:23 +0200)]
scsi_id: derive a UID for a SCSI-2 not compliant with the page 83

This patch is to enable the use of scsi_id to derive a UID for a SCSI-2
device which is not compliant with the page 83 inquiry reply format for
either SPC-2 or SPC-3.  In this case, the page 83 reply does not
contain a list of  Identification descriptors but a single binary encoded
hexa-decimal Vendor Specified Identifier.

The update is being driven by the need for scsi_id to support older
model EMC Symmetrix hardware, that is, models 4, 5, and 6.

18 years agoadd install test to 'make buildtest'
Kay Sievers [Wed, 14 Sep 2005 15:29:27 +0000 (17:29 +0200)]
add install test to 'make buildtest'

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoEXTRAS: cleanup and sync all Makefiles
Kay Sievers [Wed, 14 Sep 2005 15:00:02 +0000 (17:00 +0200)]
EXTRAS: cleanup and sync all Makefiles

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agovolume_id: fix error handling with failing read()
Kay Sievers [Wed, 14 Sep 2005 12:29:59 +0000 (14:29 +0200)]
volume_id: fix error handling with failing read()

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix a debug text typo in udev_rules.c
Olivier Blin [Wed, 14 Sep 2005 12:28:59 +0000 (14:28 +0200)]
fix a debug text typo in udev_rules.c

18 years agoudevd: fix udevd read() calls to leave room for null byte
Amir Shalem [Wed, 14 Sep 2005 12:28:13 +0000 (14:28 +0200)]
udevd: fix udevd read() calls to leave room for null byte

18 years agoEXTRAS/Makefile: fix install targets to match main Makefile
Jürg Billeter [Wed, 14 Sep 2005 12:27:28 +0000 (14:27 +0200)]
EXTRAS/Makefile: fix install targets to match main Makefile

18 years agofix nbd error messages with a gentoo rule hack
Greg KH [Tue, 13 Sep 2005 21:10:27 +0000 (14:10 -0700)]
fix nbd error messages with a gentoo rule hack

18 years agofix scsi_id rule in gentoo config file
Greg KH [Tue, 13 Sep 2005 17:39:07 +0000 (10:39 -0700)]
fix scsi_id rule in gentoo config file

18 years ago069 release
Greg KH [Tue, 13 Sep 2005 17:24:51 +0000 (10:24 -0700)]
069 release

18 years agoFAQ: update things that have changed
Kay Sievers [Mon, 12 Sep 2005 14:00:09 +0000 (16:00 +0200)]
FAQ: update things that have changed

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofic gcov use and move it into the Makefile
Kay Sievers [Mon, 12 Sep 2005 13:26:33 +0000 (15:26 +0200)]
fic gcov use and move it into the Makefile

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoremove chassis_id program
Kay Sievers [Mon, 12 Sep 2005 12:28:59 +0000 (14:28 +0200)]
remove chassis_id program

Broken and unmaintained!

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoMakefile: cleanup install targets
Kay Sievers [Mon, 12 Sep 2005 12:21:14 +0000 (14:21 +0200)]
Makefile: cleanup install targets

Split into individual target and never remove any config file.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoMakefile: fix prerequisits for $(PROGRAMS)
Kay Sievers [Sat, 10 Sep 2005 21:35:30 +0000 (23:35 +0200)]
Makefile: fix prerequisits for $(PROGRAMS)

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofirmware_helper: fix write count
Duncan Sands [Fri, 9 Sep 2005 10:38:58 +0000 (12:38 +0200)]
firmware_helper: fix write count

18 years agoupdate TODO
Kay Sievers [Thu, 8 Sep 2005 10:45:09 +0000 (12:45 +0200)]
update TODO

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix useless but funny name_cdrom.pl script to work again
Kay Sievers [Thu, 8 Sep 2005 10:24:18 +0000 (12:24 +0200)]
fix useless but funny name_cdrom.pl script to work again

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate SUSE rules
Kay Sievers [Thu, 8 Sep 2005 01:53:54 +0000 (03:53 +0200)]
update SUSE rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd "totally outdated" header to docs/overview :)
Kay Sievers [Thu, 8 Sep 2005 01:49:23 +0000 (03:49 +0200)]
add "totally outdated" header to docs/overview :)

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agodelete old Fedora html page
Kay Sievers [Thu, 8 Sep 2005 01:38:06 +0000 (03:38 +0200)]
delete old Fedora html page

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agodelete docs/persistent_naming - obsoleted by persistent disk names
Kay Sievers [Thu, 8 Sep 2005 01:36:45 +0000 (03:36 +0200)]
delete docs/persistent_naming - obsoleted by persistent disk names

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agodelete cdromsymlinks* - obsoleted by cdrom_id and IMPORT rules
Kay Sievers [Thu, 8 Sep 2005 01:34:34 +0000 (03:34 +0200)]
delete cdromsymlinks* - obsoleted by cdrom_id and IMPORT rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoklibc: update to version 1.1.1
Kay Sievers [Tue, 6 Sep 2005 20:57:07 +0000 (22:57 +0200)]
klibc: update to version 1.1.1

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agomove and update libsysfs.txt
Kay Sievers [Tue, 6 Sep 2005 18:33:59 +0000 (20:33 +0200)]
move and update libsysfs.txt

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd edd_id tool to match BIOS EDD disk information
Kay Sievers [Tue, 6 Sep 2005 10:18:04 +0000 (12:18 +0200)]
add edd_id tool to match BIOS EDD disk information

Thanks to:
    John Hull <John_Hull@Dell.com>
    Matt Domsch <Matt_Domsch@Dell.com>

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoscsi_id: rename SYSFS to LIBSYSFS
Kay Sievers [Wed, 31 Aug 2005 21:36:25 +0000 (23:36 +0200)]
scsi_id: rename SYSFS to LIBSYSFS

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevinfo: restore -d option
Kay Sievers [Wed, 31 Aug 2005 00:35:35 +0000 (02:35 +0200)]
udevinfo: restore -d option

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoalso forgot udevmonitor.8
Kay Sievers [Tue, 30 Aug 2005 22:03:20 +0000 (00:03 +0200)]
also forgot udevmonitor.8

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd missing udevsend man page
Kay Sievers [Tue, 30 Aug 2005 22:00:22 +0000 (00:00 +0200)]
add missing udevsend man page

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoall man pages rewritten to use DocBook XML
Kay Sievers [Tue, 30 Aug 2005 21:30:52 +0000 (23:30 +0200)]
all man pages rewritten to use DocBook XML

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoMakefile: remove all the duplicated rules
Kay Sievers [Tue, 30 Aug 2005 15:03:25 +0000 (17:03 +0200)]
Makefile: remove all the duplicated rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agotest the automatic man page rebuild and checkin
Kay Sievers [Tue, 30 Aug 2005 03:02:11 +0000 (05:02 +0200)]
test the automatic man page rebuild and checkin

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevinfo: rename dump() to export()
Kay Sievers [Tue, 30 Aug 2005 02:56:24 +0000 (04:56 +0200)]
udevinfo: rename dump() to export()

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agowrite man page masters in DocBook XML
Kay Sievers [Tue, 30 Aug 2005 02:41:42 +0000 (04:41 +0200)]
write man page masters in DocBook XML

The troff man pages will still be checked into the tree but the
source is DocBook XML format living in the docs/ directory now.
Start with the easy ones, the main udev page is still left to
rewrite.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevinfo: add database export
Kay Sievers [Mon, 29 Aug 2005 13:01:08 +0000 (15:01 +0200)]
udevinfo: add database export

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate RELEASE-NOTES
Kay Sievers [Mon, 29 Aug 2005 01:49:40 +0000 (03:49 +0200)]
update RELEASE-NOTES

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agocheck the udevdb before assigning a new %e
Kay Sievers [Mon, 29 Aug 2005 01:48:17 +0000 (03:48 +0200)]
check the udevdb before assigning a new %e

A second "add" event created a new enumeration with %e. Check the
udevdb first and if a matching name is already used.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudev_test.pl: we replace untrusted chars with '_'
Kay Sievers [Mon, 29 Aug 2005 00:19:36 +0000 (02:19 +0200)]
udev_test.pl: we replace untrusted chars with '_'

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoremove special TIMEOUT handling from incoming queue
Kay Sievers [Sun, 28 Aug 2005 23:19:02 +0000 (01:19 +0200)]
remove special TIMEOUT handling from incoming queue

Moving events directly to the exec queue instead of the reordering
incoming queue, leaves holes in the sequence, that lead to timeouts for
all other events. Remove that part of the special handling.

(With netlink, events can't get out-of-order and the maximum timeout is 5
seconds and should not cause any trouble with the 10 sec timout for the
firmware class anyway. Events with timeouts are still prioritized for
execution, but don't bypass the incoming queue anymore.)

Many thanks to:
  Uberto Barbini <uberto@ubiland.net>

for his endless debugging and sending all the traces, that showed this
failure with his DVB device:
  UEVENT[1124474094] add@/module/stv0299
  UEVENT[1124474094] add@/module/ves1x93
  UEVENT[1124474094] add@/module/ttpci_eeprom
  UEVENT[1124474094] add@/module/saa7146
  UEVENT[1124474094] add@/module/video_buf
  UEVENT[1124474094] add@/module/saa7146_vv
  UEVENT[1124474094] add@/module/dvb_core
  UEVENT[1124474094] add@/module/dvb_ttpci
  UEVENT[1124474094] add@/bus/pci/drivers/dvb
  UEVENT[1124474094] add@/class/firmware/0000:00:14.0
  UDEV  [1124474094] add@/module/dvb_core
  UDEV  [1124474094] add@/module/saa7146_vv
  UDEV  [1124474094] add@/module/dvb_ttpci
  UDEV  [1124474094] add@/module/ves1x93
  UDEV  [1124474094] add@/module/ttpci_eeprom
  UDEV  [1124474094] add@/module/saa7146
  UDEV  [1124474094] add@/module/stv0299
  UDEV  [1124474094] add@/module/video_buf
  UDEV  [1124474094] add@/bus/pci/drivers/dvb
  UEVENT[1124474094] remove@/class/firmware/0000:00:14.0  <- event with TIMEOUT will leave a hole in the incoming
  UDEV  [1124474094] add@/class/firmware/0000:00:14.0        sequence, which will cause a wait for the alarm()
  UEVENT[1124474094] add@/class/i2c-adapter/i2c-1            that flushes the queue
  UEVENT[1124474094] add@/class/i2c-dev/i2c-1
  UDEV  [1124474094] remove@/class/firmware/0000:00:14.0  <- event also has TIMEOUT and is executed immediately
  UEVENT[1124474095] add@/class/dvb/dvb0.demux0
  UEVENT[1124474095] add@/class/dvb/dvb0.dvr0
  UEVENT[1124474095] add@/class/dvb/dvb0.video0
  UEVENT[1124474095] add@/class/dvb/dvb0.audio0
  UEVENT[1124474095] add@/class/dvb/dvb0.ca0
  UEVENT[1124474095] add@/class/dvb/dvb0.osd0
  UEVENT[1124474095] add@/class/dvb/dvb0.net0
  UEVENT[1124474095] add@/class/video4linux/video1
  UEVENT[1124474095] add@/class/dvb/dvb0.frontend0
  UDEV  [1124474099] add@/class/i2c-adapter/i2c-1         <- all others have 5 seconds delay cause of the missing event
  UDEV  [1124474099] add@/class/dvb/dvb0.ca0                 missing events
  UDEV  [1124474099] add@/class/dvb/dvb0.osd0
  UDEV  [1124474099] add@/class/video4linux/video1
  UDEV  [1124474099] add@/class/dvb/dvb0.frontend0
  UDEV  [1124474099] add@/class/dvb/dvb0.video0
  UDEV  [1124474099] add@/class/dvb/dvb0.audio0
  UDEV  [1124474099] add@/class/i2c-dev/i2c-1
  UDEV  [1124474099] add@/class/dvb/dv

My test program that simulates a similar sequence, runs without any delay
now. (With one of the next versions we will make netlink mandatory, then
we can remove the whole input queue crap with the reordering anyway.)

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevd: move some logging to "info" and "err"
Kay Sievers [Sun, 28 Aug 2005 21:19:45 +0000 (23:19 +0200)]
udevd: move some logging to "info" and "err"

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agovol_id: ID_LABEL_SAFE will no longer contain fancy characters
Kay Sievers [Sun, 28 Aug 2005 21:16:56 +0000 (23:16 +0200)]
vol_id: ID_LABEL_SAFE will no longer contain fancy characters

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agorename ...trailing_char() to ...trailing_chars()
Kay Sievers [Sun, 28 Aug 2005 21:15:51 +0000 (23:15 +0200)]
rename ...trailing_char() to ...trailing_chars()

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agodon't reenter get_udevd_msg() if message is ignored
Kay Sievers [Sun, 28 Aug 2005 21:14:59 +0000 (23:14 +0200)]
don't reenter get_udevd_msg() if message is ignored

Check all event sources with every iteration.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoremove not needed sig_flag for state of signal_pipe
Kay Sievers [Sun, 28 Aug 2005 21:13:14 +0000 (23:13 +0200)]
remove not needed sig_flag for state of signal_pipe

Less internal state is better. And, we don't need to call
read() twice with every signal handling.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agouse WRITE_END/READ_END for the pipe index
Kay Sievers [Sun, 28 Aug 2005 18:25:37 +0000 (20:25 +0200)]
use WRITE_END/READ_END for the pipe index

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agostart to enforce plain ascii or valid utf8
Kay Sievers [Sun, 28 Aug 2005 13:55:58 +0000 (15:55 +0200)]
start to enforce plain ascii or valid utf8

No device node or symlink can have other characters as plain
readable ascii or valid utf8. The /dev/disk/by-label/* symlinks
can no longer contain weird stuff read from untrusted sources.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd uft8 validation for safe volume label exporting
Kay Sievers [Sun, 28 Aug 2005 11:49:32 +0000 (13:49 +0200)]
add uft8 validation for safe volume label exporting

We will not support any other character encoding than plain ascii
or utf8 for volume labels. All invalid utf8 and non-ascii characters
are substituted for security reasons. No options, no fancy heuristics.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agodon't store devpath in udevdb, we don't need it
Kay Sievers [Sat, 27 Aug 2005 21:27:43 +0000 (23:27 +0200)]
don't store devpath in udevdb, we don't need it

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd and use name_list_cleanup() for cleaning up the string lists
Kay Sievers [Sat, 27 Aug 2005 14:15:41 +0000 (16:15 +0200)]
add and use name_list_cleanup() for cleaning up the string lists

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agochange parameter order of udev_db_search_name()
Kay Sievers [Sat, 27 Aug 2005 13:46:58 +0000 (15:46 +0200)]
change parameter order of udev_db_search_name()

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoprint persistent data with "udevinfo -q all"
Kay Sievers [Sat, 27 Aug 2005 13:37:03 +0000 (15:37 +0200)]
print persistent data with "udevinfo -q all"

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoprepare for new HAL udevdb dump
Kay Sievers [Sat, 27 Aug 2005 00:59:20 +0000 (02:59 +0200)]
prepare for new HAL udevdb dump

HAL soon wants to read the whole content of the udevdatabase while
starting up. This makes the whole udev structure available to the
udevinfo "dump".

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix lookup for name in the udevdb, it should return the devpath
Kay Sievers [Fri, 26 Aug 2005 20:43:49 +0000 (22:43 +0200)]
fix lookup for name in the udevdb, it should return the devpath

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agostore ENV{key}="value" exported keys in the database
Kay Sievers [Fri, 26 Aug 2005 19:38:18 +0000 (21:38 +0200)]
store ENV{key}="value" exported keys in the database

Just like the IMPORT values are stored in the udev database the
exported keys are stored in the udev database. The database is also
imported on a remove event before the rules are matched. So it
is possible to do:
  ACTION=="add", KERNEL=="sda", ENV{MY_DEVICE}="1234", RUN+="/bin/logger add $devpath"
  ACTION=="remove", ENV{MY_DEVICE}=="1234", RUN+="/bin/logger remove $devpath"

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoswitch to '==' in raid-devfs.sh
Thierry Vignaud [Fri, 26 Aug 2005 12:52:57 +0000 (14:52 +0200)]
switch to '==' in raid-devfs.sh

18 years agoupdate SUSE rules
Kay Sievers [Fri, 26 Aug 2005 12:27:13 +0000 (14:27 +0200)]
update SUSE rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd pci_express to bus list
Kay Sievers [Thu, 25 Aug 2005 01:37:38 +0000 (03:37 +0200)]
add pci_express to bus list

Hopefully that wait_for_sysfs crap can go away some day...

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoupdate SUSE rules
Kay Sievers [Thu, 25 Aug 2005 00:30:01 +0000 (02:30 +0200)]
update SUSE rules

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agopath_id: remove SUSE specific PATH
Kay Sievers [Wed, 24 Aug 2005 22:40:41 +0000 (00:40 +0200)]
path_id: remove SUSE specific PATH

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix typo in firmware_helper
Amir Shalem [Wed, 24 Aug 2005 22:39:45 +0000 (00:39 +0200)]
fix typo in firmware_helper

The wrong string is being nullifed.

18 years agoudevmonitor: cleanup on exit
Kay Sievers [Wed, 24 Aug 2005 22:38:25 +0000 (00:38 +0200)]
udevmonitor: cleanup on exit

We want to write the output buffer before
we exit, otherwise redirection to a file
does not work correctly.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agovolume_id: don't probe for mac_partition_maps
Kay Sievers [Wed, 24 Aug 2005 22:37:46 +0000 (00:37 +0200)]
volume_id: don't probe for mac_partition_maps

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoudevd: keep the right order for messages without SEQNUM
Kay Sievers [Mon, 22 Aug 2005 22:57:46 +0000 (00:57 +0200)]
udevd: keep the right order for messages without SEQNUM

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix exit code of udevinitsend and udevmonitor
Kay Sievers [Mon, 22 Aug 2005 10:01:55 +0000 (12:01 +0200)]
fix exit code of udevinitsend and udevmonitor

Thanks to: Marco d'Itri <md@Linux.IT> for the initial patch.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoadd program name to logged error
Kay Sievers [Mon, 22 Aug 2005 09:42:58 +0000 (11:42 +0200)]
add program name to logged error

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years ago*_id: fix zero length in set_str()
Kay Sievers [Mon, 22 Aug 2005 09:37:12 +0000 (11:37 +0200)]
*_id: fix zero length in set_str()

Thanks to Olaf Hering <olh@suse.de>

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years ago068 release
Greg KH [Thu, 18 Aug 2005 17:04:49 +0000 (10:04 -0700)]
068 release

18 years agoadd EXTRAS documentation to the README file.
Greg KH [Thu, 18 Aug 2005 17:02:27 +0000 (10:02 -0700)]
add EXTRAS documentation to the README file.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years agofix printed udevmonitor header
Kay Sievers [Thu, 18 Aug 2005 11:43:27 +0000 (13:43 +0200)]
fix printed udevmonitor header

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agobring std(in|out|err) fd's in a sane state
Kay Sievers [Thu, 18 Aug 2005 11:07:15 +0000 (13:07 +0200)]
bring std(in|out|err) fd's in a sane state

If the kernel forks us as an usermodhelper, we don't have any of
the standard fd's and the first open() will start with fd=0.
This is inherited to all forked childs and confuses later forked
helpers where we want to read from a pipe connected to the helpers
stdout/stderr.

  # ls -l /proc/$(pidof udevd)/fd
  total 6
  dr-x------  2 root root  0 2005-08-18 12:44 .
  dr-xr-xr-x  4 root root  0 2005-08-18 12:44 ..
  lrwx------  1 root root 64 2005-08-18 12:44 0 -> /dev/null
  lrwx------  1 root root 64 2005-08-18 12:44 1 -> socket:[1274617]
  lr-x------  1 root root 64 2005-08-18 12:44 2 -> pipe:[1274618]
  l-wx------  1 root root 64 2005-08-18 12:44 3 -> pipe:[1274618]
  lrwx------  1 root root 64 2005-08-18 12:44 4 -> socket:[1274619]
  lrwx------  1 root root 64 2005-08-18 12:44 5 -> socket:[1274620]

Ouch! This will obviously not redirect sterr, it will kill the pipe
we established between the parent and the child:

  devnull = open("/dev/null", O_RDWR);
  dup2(devnull, STDERR_FILENO);

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agofix typo in cdrom_id syslog
Kay Sievers [Thu, 18 Aug 2005 02:54:59 +0000 (04:54 +0200)]
fix typo in cdrom_id syslog

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
18 years agoMerge rsync://rsync.kernel.org/pub/scm/linux/hotplug/udev
Kay Sievers [Thu, 18 Aug 2005 02:48:43 +0000 (04:48 +0200)]
Merge rsync.kernel.org/pub/scm/linux/hotplug/udev