LinuxQuestions.org
Visit Jeremy's Blog.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices

Reply
 
Search this Thread
Old 07-17-2005, 05:38 PM   #16
MMYoung
Member
 
Registered: Apr 2004
Location: Arkansas
Distribution: Ubuntu 8.10
Posts: 365

Rep: Reputation: 30

Quote:
Originally posted by dhave
One question, though: what's the downside of just not updating udev, at least for now? Are there significant new features in the post-54 releases of udev that make updating worthwhile?
I don't see any "downside" for not updating udev, other than the downside of your alsa devices actually working if you don't . Any other "significat features" would be made "moot" (to me anyway) if sound didn't work. However, there may be some enhancements that would be good for a sever, where sound (much less alsa) would be of less importance.

Later,
MMYoung
 
Old 07-17-2005, 06:20 PM   #17
jong357
Senior Member
 
Registered: May 2003
Location: Columbus, OH
Distribution: DIYSlackware
Posts: 1,914

Rep: Reputation: 52
Well, too often people just upgrade to upgrade and don't know if they need to or not. I'm guilty of occassionally doing "Ooooo, Look! A new version of xyz" where xyz is some back water irrelevant package even.... I suppose udev is along those lines. 054 was working just fine. This post and the changelog made me curious tho... So, I had to try and upgrade for no good reason really... The changelog is quite extensive, but as with most things, will probably only effect a few people....
Code:
Summary of changes from v062 to v063
============================================

Anton Farygin:
  fix typo in GROUP value application

Greg Kroah-Hartman:
  add 'make tests' as I'm always typing that one wrong...
  Really commit the udev_run_devd changes...
  Fixed udev_run_devd to run the /etc/dev.d/DEVNAME/ files too
  fix position of raw rules in gentoo config file

Hannes Reinecke:
  dasd_id: add s390 disk-label prober
  fix usb_id and let scsi_id ignore "illegal request"

Kay Sievers:
  volume_id: remove s390 dasd handling, it is dasd_id now
  trivial fixes for *_id programs
  IMPORT: add {parent} to import the persistent data of the parent device
  allow multiple values to be matched with KEY=="value1|value2"
  udevd: set incoming socket buffer SO_RCVBUF to maximum
  remember mapped rules state
  ata_id: check for empty serial number
  compile dasd only on s390

Ville Skyttä:
  correct default mode documentation in udev


Summary of changes from v061 to v062
============================================

Kay Sievers:
  fix symlink values separated by multiple spaces
  update RELEASE-NOTES
  fix typo in group assignment
  fix default-name handling and NAME="" rules
  add WAIT_FOR_SYSFS key to loop until a file in sysfs arrives
  fix unquoted strings in udevinitsend

Summary of changes from v060 to v061
============================================

Greg Kroah-Hartman:
  Sync up the Debian rules files
  fix cdrom symlink problem in gentoo rules
  Fix ChangeLog titles

Kay Sievers:
  update RELEASE-NOTES
  we want to provide OPTFLAGS
  rename ALARM_TIMEOUT to UDEV_ALARM_TIMEOUT
  udevd: optimize env-key parsing
  don't resolve OWNER, GROUP on precompile if string contains %, $
  set default device node to /dev
  create udevdb files only if somehting interesting happened
  pack parsed rules list
  replace useless defines by inline text
  move rule matches to function
  add usb_id program to generate usb-storage device identifiers
  add IEEE1394 rules to the gentoo rule file
  fake also kernel-name if we renamed a netif
  allow OPTIONS to be recognized for /sys/modules /sys/devices events
  switch gentoo rules to new operators


Summary of changes from v059 to v060
============================================

Greg Kroah-Hartman:
  Fix the gentoo udev rules to allow the box to boot properly

Gustavo Zacarias:
  Udev doesn't properly build with $CROSS

Kay Sievers:
  Keep udevstart from skipping devices without a 'dev' file

Marco d'Itri:
  #define NETLINK_KOBJECT_UEVENT


Summary of changes from v058 to v059
============================================

Greg Kroah-Hartman:
  Update the gentoo rule file
  Fix udevinfo for empty sysfs directories
  Fix makefile to allow 'make release' to work with git

Hannes Reinecke:
  udev: fix netdev RUN handling
  udevcontrol: fix exit code

Kay Sievers:
  prepare RELEASE-NOTES
  add ID_TYPE to the id probers
  add -x to scsi_id to export the queried values in env format
  store the imported device information in the udevdb
  rename udev_volume_id to vol_id and add --export option
  add ata_id to read serial numbers from ATA drives
  IMPORT allow to import program returned keys into the env
  unify execute_command() and execute_program()
  IMPORT=<file> allow to import a shell-var style config-file
  allow rules to be compiled to one binary file
  fix the fix and change the file to wait for to the "bus" link
  fix udevstart and let all events trvel trough udev
  prepare for module loading rules and add MODALIAS key
  remove device node, when type block/char has changed
  Makefile: remove dev.d/ hotplug.d/ from install target
  udevcontrol: add max_childs command
  udevd: control log-priority of the running daemon with udevcontrol
  udeveventrecorder: add small program that writes an event to disk
  klibc: add missing files
  udevinitsend: handle replay messages correctly
  udev man page: add operators
  udevd: allow starting of udevd with stopped exec-queue
  klibc: version 1.0.14
  udev: handle all events - not only class and block devices
  volume_id: use udev-provided log-level
  udev: clear lists if a new value is assigned
  udev: move dev.d/ handling to external helper
  udev: allow final assignments :=
  udevd: improve timeout handling
  Makefile: fix DESTDIR
  udevd: add initsend
  udevd: add udevcontrol
  udevd: listen for netlink events

Stefan Schweizer:
  Dialout group fix for capi devices in the gentoo rules file

Summary of changes from v057 to v058
============================================

Daniel Drake:
  o Writing udev rules docs update

Darren Salt:
  o update cdsymlinks to latest version

Greg Kroah-Hartman:
  o remove detach_state files from the sysfs test tree
  o Update permissions on test scripts so they will run properly now
  o hopefully fix up the symlinks in the test directory
  o Removed klibc/klibc.spec as it is autogenerated
  o Added symlinks thanks to Kay's script and git hacking
  o add Red Hat/Fedora html documenation
  o Update Red Hat default udev rules

Kay Sievers:
  o selinux: fix handling during creation of symlinks
  o Fedora udev.rules update
  o libsysfs: version 2.0
  o klibc: version 1.0.7

Masanao Igarashi:
  o Fix libsysfs issue with relying on the detach_state file to be

Summary of changes from v056 to v057
============================================

<tklauser:access.unizh.ch>:
  o fix stupid all_partitions bug

Kay Sievers:
  o add test for make -j4 to build-check
  o klibc: version 1.0.6
  o update Debian rules
  o apply default permissions only for devices that will need it
  o adapt RELEASE-NOTES
  o udev_volume_id: fix endianess macros
  o udev-test.pl: add test for DEVNAME export to RUN environment
  o update the man page to reflect the recent changes
  o export DEVNAME to RUN-key executed programs
  o fix make -j4 and the local klibc-install
  o update RELEASE-NOTES
  o add RUN key to be able to run rule based notification
  o fix udevtest to print the error if logging is disabled
  o move execute_program to utils + add action to init_device
  o correct correction for error path for PROGRAM execution
  o correct error path for PROGRAM execution
  o klibc: version 1.0.5
  o check for strlen()==0 before accessing strlen()-1
  o allow to match against empty key values
  o read %s{}-sysfs values at any device in the chain
  o udev_rules.c: don't change sysfs_device while walking up the device chain
  o klibc: strlcpy/strlcat - don't alter destination if size == 0
  o fix klibc's broken strlcpy/strlcat
  o udevinfo: print SYSFS attribute the same way we match it
  o remove untrusted chars read from sysfs-values or returned by PROGRAM
  o udevinfo: print errors to stderr instead of stdout
  o klibc: version 1.0.4
  o support log-priority levels in udev.conf
  o test-suite: remove UDEV_TEST, it's not needed anymore
  o libsysfs: remove trailing slash on SYSFS_PATH override


Summary of changes from v055 to v056
============================================

<tklauser:access.unizh.ch>:
  o fix header paths in udev_libc_wrapper.c

Kay Sievers:
  o udev-test.pl: use more common user/group names
  o klibc: remove SCCS directories from the temporary klibc install
  o udev-test.pl: add a test where the group cannot be found in /etc/passwd
  o udev-test.pl: add check for textual uid/gid
  o fix bad typo that prevents the GROUP to be applied
  o udevd: don't delay events with TIMEOUT in the environment
  o klibc: use klcc wrapper instead of our own Makefile
  o change call_foreach_file to return a list


Summary of changes from v054 to v055
============================================

<jkluebs:luebsphoto.com>:
  o This patch causes the remove handler to check that each symlink actually points to the correct devnode and skip it if it does not.

<pebenito:gentoo.org>:
  o udev selinux fix

<tklauser:access.unizh.ch>:
  o The following patch fixes some warnings when compiling volume_id from udev with the -Wall compiler flag. Define _GNU_SOURCE for strnlen() and correct the path to logging.h
  o The following patch fixes a warning when compiling chassis_id from udev with the -Wall compiler flag. There are too much conversions in the format string of sscanf(). One %d can be dropped.

Greg Kroah-Hartman:
  o fix raid rules
  o added frugalware udev ruleset
  o merge selinux and Kay's symlink fixes together

Hannes Reinecke:
  o volume_id: Fix label/uuid reading for reiserfs

Kay Sievers:
  o add udevstart to the RELEASE-NOTES
  o volume_id: version 43
  o clarify the shortcomings of %e
  o correct rule match for devices without a physical device
  o remove unneeded code, libsysfs does this for us
  o add final release note
  o add ENV{} key to match agains environment variables
  o simplify sysfs_pair handling
  o add a test and simplify debug statement
  o support =, ==, !=, += for the key match and assignment
  o add OPTION="last_rule" to skip any later rule
  o rename namedev_dev to udev_rule
  o correct enum device_type
  o remove udevstart on make clean
  o volume_id: version 42
  o volume_id: version 41
  o remove unneeded include
  o The path to dlist.h is not correct
  o udevinfo -d: use '=' as separator, cause ':' may be a part of the devpath
  o klibc: version 1.0.3
  o add RELEASE-NOTES file
  o test suite: move "driver" link to physical device
  o remove PLACE key match
  o don't lookup "root" in the userdb
  o fix ia64 compile
  o fix segfaulting udev while DRIVER matching
  o cleanup list.h
  o klibc: version 0.214
  o rename device_list->list to device_list->node
  o replace strncpy()/strncat() by strlcpy()/strlcat()
  o split udev and udevstart
  o udev_volume_id: version 39
  o rename LOG to USE_LOG in all places
  o remove Makefile magic for klibc integration
  o klibc_fixups: remove no longer needed stuff
  o udev_volume_id: volume_id v38
  o use numeric owner/group as default values to avoid parsing userdb
  o fix up segfaulting binaries with new klibc
  o udevinfo -d: speed-up device dump
  o klibc: version 0.211
  o klibc_fixups: remove unneeded stuff
  o replace weird defines by real code
  o udev-test.pl: remove useless tests
  o allow unlimitied count of symlinks
  o unmap db-file after use
  o remove typedef for call_foreach_file() handler function
  o correct udev_init_device
  o rename attributes to options
  o kill stupid gcc4 warning
  o trivial clenaup of namedev code
  o klibc: check for gcc4
  o klibc: update v0.205

Thierry Vignaud:
  o gentoo rule update for raid devices
 
Old 07-17-2005, 06:35 PM   #18
dhave
Member
 
Registered: Oct 2004
Location: still outside the Matrix
Distribution: Arch, formerly Gentoo and Slackware
Posts: 438

Original Poster
Rep: Reputation: 31
Quote:
Originally posted by jong357
Well, too often people just upgrade to upgrade and don't know if they need to or not. I'm guilty of occassionally doing "Ooooo, Look! A new version of xyz" where xyz is some back water irrelevant package even....
Yep, that's my case, too. I usually try to heed PV's warnings, but sometimes I don't, or sometimes I don't fully understand. So, rather than holding off until I get full info, I usually forge ahead -- and then have to come here to get help cleaning up the mess afterward. But that's how you learn, right?
Quote:
I suppose udev is along those lines. 054 was working just fine. This post and the changelog made me curious tho... So, I had to try and upgrade for no good reason really... The changelog is quite extensive, but as with most things, will probably only effect a few people....
Nothing really jumped off the page at me, but I haven't gone over it carefully enough. I'm sure in good time the kinks will be worked out. Meanwhile, on this one I can be patient. I think.

BTW, thanks for the reminder that Current isn't really a 2.6.x kernel config. That's very easy to forget.
 
Old 07-18-2005, 05:33 AM   #19
GlowGlow
Member
 
Registered: Jun 2005
Posts: 111

Rep: Reputation: 15
Quote:
Originally posted by MMYoung
Do you think that "wanting to keep udev out of the loop" is more of a resistance to change type thing, or is it due to some type of hardware problem? I'm really interested to know, because for me, personally, udev it a much better way of handling /dev population than the old way of just pretty much putting something in /dev for every device known to man. Not saying it didn't work, as it work well, but udev is more "elegant", as you put it, IMHO.
Good question. For me personally, it deviates too much from the 'keep it simple stupid' paradigm. udev adds an unecessary layer of complexity. Besides that I do not feel that it is finished yet, e.g. the rule syntax changes every few weeks/months. But I can see that others find it comfortable, so I do not necessarily object to having it as a part of Slackware, but I would object to having it as a mandatory aspect of the system (and would write an e-mail to PV to argue for making it optional again).
 
Old 07-18-2005, 03:28 PM   #20
granth
Member
 
Registered: Jul 2004
Location: USA
Distribution: Slackware64
Posts: 210

Rep: Reputation: 39
I like udev but it gave me some problems when trying to use alsa as modules in my kernel. I got sound working simply by building it into the kernel. (monolithic) I'm really having problems with gstreamer though. I cant change tracks on cd's without gnome-cd freezing. Also when I change my multimedia prefs. to alsasink, the test button produces a very choppy test tone.
 
Old 07-18-2005, 07:41 PM   #21
jong357
Senior Member
 
Registered: May 2003
Location: Columbus, OH
Distribution: DIYSlackware
Posts: 1,914

Rep: Reputation: 52
Quote:
Originally posted by GlowGlow
Good question. For me personally, it deviates too much from the 'keep it simple stupid' paradigm. udev adds an unecessary layer of complexity. Besides that I do not feel that it is finished yet
Right... He didn't come out and say it in his email, but I could tell that Pat felt the same way but ended with a "perhaps that's the way of the future" type of comment. I really don't think you'll see udev in the drivers seat in Slackware for quite some time. No worries... If you've used Slack for any length of time and follow the changelog and various comments made by him, it's evident he doesn't jump the gun on anything.

Personally, I have no problem with using it. It's young, yes. Constantly changing, yes. It still works tho and thats good enough for me.
 
Old 07-20-2005, 11:33 AM   #22
dive
Senior Member
 
Registered: Aug 2003
Location: UK
Distribution: Slackware
Posts: 3,199

Rep: Reputation: 290Reputation: 290Reputation: 290
How do I tell my version of udev? I recently used slapt-get upgrade and am having some sound problems
 
Old 07-20-2005, 11:36 AM   #23
keefaz
Senior Member
 
Registered: Mar 2004
Distribution: Slackware
Posts: 4,286

Rep: Reputation: 68
Code:
udevinfo -V
 
Old 07-20-2005, 07:21 PM   #24
dhave
Member
 
Registered: Oct 2004
Location: still outside the Matrix
Distribution: Arch, formerly Gentoo and Slackware
Posts: 438

Original Poster
Rep: Reputation: 31
I guess you guys have seen that Pat has switched to udev-058 in today's Current updates. This is his changelog entry:
Quote:
a/udev-058-i486-1.tgz: Switched to udev-058, as newer versions still have
problems (these are probably caused by the elimination of the
/etc/hotplug.d/ directory, as this used to contain a link to udevstart).
It was pointed out that udev-062 and udev-063 do create the missing
devices if you run udevstart after boot (and possibly after plugging in
new devices), but udev-058 is working fine without any kludges and seems
to be the most stable version to use with 2.6.12.* kernels. Also, made
a fix in /etc/udev/scripts/make_extra_nodes to set a default LANG before
calling /bin/ls to look for cdrom and dvd devices (not all LANG settings
will produce the same number of fields with ls, which can break cd/dvd
symlinks). Thanks to Lukasz Stelmach for pointing out this bug.
 
Old 07-20-2005, 08:15 PM   #25
dhave
Member
 
Registered: Oct 2004
Location: still outside the Matrix
Distribution: Arch, formerly Gentoo and Slackware
Posts: 438

Original Poster
Rep: Reputation: 31
Quote:
Originally posted by dhave I guess you guys have seen that Pat has switched to udev-058 in today's Current updates.
Unfortunately, alsa doesn't like udev-058, either -- at least not on my system. I'm back with good old reliable udev-054.

FWIW, I'm running kernel 2.6.12.2 with alsa-driver-1.0.9b_2.6.12.2-i486-1.
 
Old 07-21-2005, 04:30 AM   #26
GlowGlow
Member
 
Registered: Jun 2005
Posts: 111

Rep: Reputation: 15
Quote:
Originally posted by jong357
Right... He didn't come out and say it in his email, but I could tell that Pat felt the same way but ended with a "perhaps that's the way of the future" type of comment. I really don't think you'll see udev in the drivers seat in Slackware for quite some time. No worries...
I do not worry about Patrick's choices. Rather the upstream depracation of hotplug.d and the new policy that udev should handle hotplug events. Then there is hotplug-ng, almost completely written in C, throwing away the flexibility of a scripted hotplug. I do not really like the upstream policy, an I am not exactly sure what they want to turn it in.

Quote:
If you've used Slack for any length of time and follow the changelog and various comments made by him, it's evident he doesn't jump the gun on anything.
I have used Slackware since 1994, so I have no worries about that .
 
Old 07-21-2005, 05:13 AM   #27
carboncopy
Senior Member
 
Registered: Jan 2003
Location: Malaysia
Distribution: Fedora Core, Slackware, Mac OS X, Debian, OpenSUSE
Posts: 1,210
Blog Entries: 4

Rep: Reputation: 45
May I know where can I get udev-054-i486-3.tgz?

udev-58 does not work for my sound. But USB drive works fine
udev-62 works fine for my sound. But not for USB drive.

On udev-58, me get this error when execute alsamixer.
Code:
alsamixer: function snd_ctl_open failed for default: No such file or directory
Does this have anything to do with the "new" glibc?
 
Old 07-21-2005, 05:41 AM   #28
dhave
Member
 
Registered: Oct 2004
Location: still outside the Matrix
Distribution: Arch, formerly Gentoo and Slackware
Posts: 438

Original Poster
Rep: Reputation: 31
Quote:
Originally posted by carboncopy
May I know where can I get udev-054-i486-3.tgz?
After a bit of Googling, I found it here:

http://linux.jetstreamgames.co.nz/sl...t/slackware/a/

You'd better hurry, though, because if this site gets updated, udev-54 will disappear.
 
Old 07-21-2005, 05:47 AM   #29
carboncopy
Senior Member
 
Registered: Jan 2003
Location: Malaysia
Distribution: Fedora Core, Slackware, Mac OS X, Debian, OpenSUSE
Posts: 1,210
Blog Entries: 4

Rep: Reputation: 45
Thankd dhave.

Sound is back BUT USB drive is gone.

How can I update the udev rule for the following device:
Code:
 Bus 001 Device 006: ID ffff:ffff  
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               2.00
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0xffff 
  idProduct          0xffff 
  bcdDevice            1.00
  iManufacturer           0 
  iProduct                1 USB Flash Drive
  iSerial                 2 123456789ABCDEF
  bNumConfigurations      1
  Configuration Descriptor:
    bLength                 9
    bDescriptorType         2
    wTotalLength           32
    bNumInterfaces          1
    bConfigurationValue     1
    iConfiguration          0 
    bmAttributes         0x80
    MaxPower              500mA
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        0
      bAlternateSetting       0
      bNumEndpoints           2
      bInterfaceClass         8 Mass Storage
      bInterfaceSubClass      6 SCSI
      bInterfaceProtocol     80 Bulk (Zip)
      iInterface              0 
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x81  EP 1 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  bytes 64 once
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x02  EP 2 OUT
        bmAttributes            2
 
Old 07-22-2005, 03:43 PM   #30
dhave
Member
 
Registered: Oct 2004
Location: still outside the Matrix
Distribution: Arch, formerly Gentoo and Slackware
Posts: 438

Original Poster
Rep: Reputation: 31
Here's the latest from Patrick regarding udev, from today's (22july05) changelog:

Quote:
a/udev-058-i486-2.tgz: Added a line to udev.rules to (hopefully) help
with the ALSA issues:
KERNEL="controlC[0-9]", NAME="snd/%k", MODE="0666"
Now, it would seem to me that the already-existing line:
KERNEL="controlC[0-9]*", NAME="snd/%k", MODE="0666"
...should have already covered this. It works with previous versions
of udev just fine, and this seems to me to be a udev bug. Oh well,
give it a test and let me know if it's still causing any problems, in
which case I'll probably go back to 054 for the Slackware 10.2 release.
I'd rather not spend the next couple of months dorking around with
udev problems and not getting a Slackware release out because of it.
Thanks to Andris Pavenis for the one line udev.rules fix.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Gentoo update udev = no more X for users teek Linux - Distributions 8 03-14-2006 09:22 PM
Upgrading aaa-elflibs in current changelog Changeling Slackware 1 09-09-2005 10:34 AM
FC3 ndiswrapper - not working after udev update? trey85stang Fedora 0 11-23-2004 09:21 PM
Hehe- Current slack changelog cav Slackware 1 09-11-2003 10:43 PM
Mirror of current changelog? Coldfirex Slackware 3 02-04-2003 07:30 AM


All times are GMT -5. The time now is 05:11 PM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration