chiark / gitweb /
elogind.git
16 years ago[PATCH] Add infiniband to gentoo rules.
greg@kroah.com [Wed, 9 Feb 2005 05:39:54 +0000 (21:39 -0800)]
[PATCH] Add infiniband to gentoo rules.

16 years ago[PATCH] Another gentoo fix, adding dvb support.
greg@kroah.com [Wed, 9 Feb 2005 05:32:55 +0000 (21:32 -0800)]
[PATCH] Another gentoo fix, adding dvb support.

16 years ago[PATCH] Fix gentoo bug #76056 (fb device group permissions.)
greg@kroah.com [Wed, 9 Feb 2005 05:27:02 +0000 (21:27 -0800)]
[PATCH] Fix gentoo bug #76056 (fb device group permissions.)

16 years ago[PATCH] Fix gentoo bug #81102, device nodes for the pktcdvd device.
greg@kroah.com [Wed, 9 Feb 2005 05:24:29 +0000 (21:24 -0800)]
[PATCH] Fix gentoo bug #81102, device nodes for the pktcdvd device.

16 years ago[PATCH] create /block/*/range count of partitons for all_partitions
kay.sievers@vrfy.org [Wed, 9 Feb 2005 07:43:18 +0000 (08:43 +0100)]
[PATCH] create /block/*/range count of partitons for all_partitions

16 years ago[PATCH] provide temporary device node for callouts to access the device
kay.sievers@vrfy.org [Wed, 9 Feb 2005 03:37:32 +0000 (04:37 +0100)]
[PATCH] provide temporary device node for callouts to access the device

%N will create a temporary node for a callout a be sustituted with the
name of the node.

16 years ago[PATCH] udev_volume_id: fix dasd disklabel reading with -l option
kay.sievers@vrfy.org [Wed, 9 Feb 2005 00:15:35 +0000 (01:15 +0100)]
[PATCH] udev_volume_id: fix dasd disklabel reading with -l option

16 years ago[PATCH] udev_volume_id: volume_id version 034
kay.sievers@vrfy.org [Wed, 9 Feb 2005 00:02:18 +0000 (01:02 +0100)]
[PATCH] udev_volume_id: volume_id version 034

16 years ago[PATCH] trivial: remove _all_ trailing slashes with no_trailing_slash()
mbuesch@freenet.de [Sun, 6 Feb 2005 02:23:43 +0000 (03:23 +0100)]
[PATCH] trivial: remove _all_ trailing slashes with no_trailing_slash()

This changes no_trailing_slash() to remove all trailing slashes
and not just the last one.

16 years ago[PATCH] trivial: fix signedness
mbuesch@freenet.de [Sun, 6 Feb 2005 00:39:25 +0000 (01:39 +0100)]
[PATCH] trivial: fix signedness

We are already using %u in the sscanf() format string.
The patch replaces the int for the kernel release number with
unsigned int, as we can never have negative release numbers.

16 years ago[PATCH] debian: update rules files
md@Linux.IT [Sun, 6 Feb 2005 00:08:31 +0000 (01:08 +0100)]
[PATCH] debian: update rules files

16 years ago[PATCH] raid-devfs.sh: devfs names for hardware RAID controllers
md@Linux.IT [Sun, 6 Feb 2005 00:07:13 +0000 (01:07 +0100)]
[PATCH] raid-devfs.sh: devfs names for hardware RAID controllers

16 years ago[PATCH] scsi_id: when udevstart is started, /tmp is not writeable.
md@Linux.IT [Sun, 6 Feb 2005 00:01:29 +0000 (01:01 +0100)]
[PATCH] scsi_id: when udevstart is started, /tmp is not writeable.

16 years ago[PATCH] cdsymlinks.sh: trivial fix, the variable is initialized to '', not 0
md@Linux.IT [Sun, 6 Feb 2005 00:00:04 +0000 (01:00 +0100)]
[PATCH] cdsymlinks.sh: trivial fix, the variable is initialized to '', not 0

16 years ago[PATCH] gentoo/udev.rules: add default permissions for sound devices
sschweizer@gmail.com [Sat, 5 Feb 2005 23:41:53 +0000 (00:41 +0100)]
[PATCH] gentoo/udev.rules: add default permissions for sound devices

16 years ago[PATCH] namdev: allow symlink-only rules to specify node permissions
mbuesch@freenet.de [Sat, 5 Feb 2005 23:13:18 +0000 (00:13 +0100)]
[PATCH] namdev: allow symlink-only rules to specify node permissions

First case without the patch:
KERNEL="hd*", GROUP="disk", MODE="0660"
KERNEL="hdc", SYMLINK="cdrecorder", MODE="0666"
brw-rw----  1 root disk 22, 0 Feb  5 14:08 /dev/hdc

Second case without the patch:
KERNEL="hd*", GROUP="disk", MODE="0660"
KERNEL="hdc", MODE="0666"
brw-rw-rw-  1 root disk 22, 0 Feb  5 14:08 /dev/hdc

16 years ago[PATCH] udevd: fix valgrind warning
mbuesch@freenet.de [Sat, 5 Feb 2005 23:09:34 +0000 (00:09 +0100)]
[PATCH] udevd: fix valgrind warning

==4304== Syscall param sigaction(act) contains uninitialised or unaddressable byte(s)
==4304==    at 0x804A37B: (within /sbin/udevd)
==4304==  Address 0x52BFE7FC is on thread 1's stack

16 years ago[PATCH] udev_volume_id: rename probe_ibm into probe_dasd
kay.sievers@vrfy.org [Sat, 5 Feb 2005 04:23:26 +0000 (05:23 +0100)]
[PATCH] udev_volume_id: rename probe_ibm into probe_dasd

16 years ago[PATCH] udev_volume_id: volume_id version 032
kay.sievers@vrfy.org [Sat, 5 Feb 2005 03:10:48 +0000 (04:10 +0100)]
[PATCH] udev_volume_id: volume_id version 032

16 years ago[PATCH] Makefile: add some more warnings and prepare for clean gcc4 compile
kay.sievers@vrfy.org [Sat, 5 Feb 2005 01:44:37 +0000 (02:44 +0100)]
[PATCH] Makefile: add some more warnings and prepare for clean gcc4 compile

16 years ago[PATCH] Makefile: cleanup conditional config option sections
kay.sievers@vrfy.org [Sat, 5 Feb 2005 01:43:16 +0000 (02:43 +0100)]
[PATCH] Makefile: cleanup conditional config option sections

16 years ago[PATCH] fix -Wsign-compare warnings
kay.sievers@vrfy.org [Sat, 5 Feb 2005 01:38:56 +0000 (02:38 +0100)]
[PATCH] fix -Wsign-compare warnings

16 years ago[PATCH] chassis_id: clean compilation and fix bad function parameter passing
kay.sievers@vrfy.org [Sat, 5 Feb 2005 01:36:54 +0000 (02:36 +0100)]
[PATCH] chassis_id: clean compilation and fix bad function parameter passing

Adding prototypes for functions resulted in an error, cause:
  table_find_disk(disk_snum, &chassis_num, &slot_num);

is called but the function is defined as:
  int table_find_disk(char *serialnumber , int *host_num, int *chassis_num, int *slot_num)

which can obviously not work correctly.

Using popen() is not klibc compatible, so skip the compilation if
a klibc compile is requested.

16 years ago[PATCH] simple_build_check: make it possible to pass KERNEL_DIR
kay.sievers@vrfy.org [Fri, 4 Feb 2005 20:57:39 +0000 (21:57 +0100)]
[PATCH] simple_build_check: make it possible to pass KERNEL_DIR

16 years ago[PATCH] selinux: cleanup udev integration
kay.sievers@vrfy.org [Fri, 4 Feb 2005 20:38:26 +0000 (21:38 +0100)]
[PATCH] selinux: cleanup udev integration

Move code into a .c-file instead of big inline functions in a header file.
Pass the device name down instead of relying that the node name is equal
to the kernel name.

16 years ago[PATCH] 051 release 051
greg@kroah.com [Thu, 3 Feb 2005 08:24:38 +0000 (00:24 -0800)]
[PATCH] 051 release

16 years ago[PATCH] gentoo fb permission fix.
greg@kroah.com [Thu, 3 Feb 2005 08:22:31 +0000 (00:22 -0800)]
[PATCH] gentoo fb permission fix.

16 years ago[PATCH] The attached patch fixes the code path if namedev_name_device() fails.
mbuesch@freenet.de [Tue, 1 Feb 2005 10:12:32 +0000 (11:12 +0100)]
[PATCH] The attached patch fixes the code path if namedev_name_device() fails.

16 years ago[PATCH] udev_volume_id: new version of volume_id
kay.sievers@vrfy.org [Mon, 31 Jan 2005 04:28:44 +0000 (05:28 +0100)]
[PATCH] udev_volume_id: new version of volume_id

Every filesystem has its own subdirectory.
The IBM dasd label reading is included into volume_id.

16 years ago[PATCH] udev_volume_id: fix -d option
cborntra@de.ibm.com [Sun, 30 Jan 2005 12:09:38 +0000 (13:09 +0100)]
[PATCH] udev_volume_id: fix -d option

The -d option in udev_allows to go from a partition to the underlying disk
for s390 dasd labels. If the device is already the disk itself, finding the
parent will fail, therefore -d on /sys/block/dasda/ for example gives no
result at all.

16 years ago[PATCH] klibc: update to version 0.198
kay.sievers@vrfy.org [Fri, 28 Jan 2005 06:53:22 +0000 (07:53 +0100)]
[PATCH] klibc: update to version 0.198

16 years ago[PATCH] remove the lfs startup script on request of the author.
greg@kroah.com [Thu, 20 Jan 2005 06:04:54 +0000 (22:04 -0800)]
[PATCH] remove the lfs startup script on request of the author.

The distro now has its own version, so this one is not needed.

16 years ago[PATCH] clean up the aoe char device rules, and delete the block one as it's not...
greg@kroah.com [Thu, 13 Jan 2005 01:46:50 +0000 (17:46 -0800)]
[PATCH] clean up the aoe char device rules, and delete the block one as it's not needed.

16 years ago[PATCH] fix gcc 2.96 issue in libsysfs
greg@bucket.kroah.org [Wed, 12 Jan 2005 16:15:22 +0000 (08:15 -0800)]
[PATCH] fix gcc 2.96 issue in libsysfs

16 years ago[PATCH] rearrange link order in Makefile
hare@suse.de [Thu, 20 Jan 2005 03:44:36 +0000 (04:44 +0100)]
[PATCH] rearrange link order in Makefile

The link-order in the main Makefile is not quite correct (sysfs should
be linked after udev.a) which might cause the build to fail.

16 years ago[PATCH] udev_volume_id: fix FAT label reading
kay.sievers@vrfy.org [Tue, 18 Jan 2005 08:13:01 +0000 (09:13 +0100)]
[PATCH] udev_volume_id: fix FAT label reading

16 years ago[PATCH] klibc: update to version 0.196
kay.sievers@vrfy.org [Mon, 17 Jan 2005 08:26:07 +0000 (09:26 +0100)]
[PATCH] klibc: update to version 0.196

16 years ago[PATCH] I tried to compile udev 050plus with the GCC 4.0 snapshot 200412119 and
tklauser@access.unizh.chbk [Mon, 17 Jan 2005 02:54:26 +0000 (03:54 +0100)]
[PATCH] I tried to compile udev 050plus with the GCC 4.0 snapshot 200412119 and

16 years ago[PATCH] udevd: throttle the forking of processes
kay.sievers@vrfy.org [Sun, 16 Jan 2005 23:53:08 +0000 (00:53 +0100)]
[PATCH] udevd: throttle the forking of processes

If the system reaches a defined limit of processes in running state, udevd
starts to count its own processes in running state from its session (all
forked hotplug child processes, subprocesses and callouts) and throttles
further process forking if the limit is reached.

This should help setups with hundreds of events emitted hotplug events
in parallel with hundreds of processes in "R" state. which makes the machine
unresponsible.

I placed a 100% cpu time consuming program in /etc/hotplug.d/ which runs for 5
seconds. With this patch I can load "scsi_debug add_host=100" without any major
problem. Without the patch the box is completly unresponsible for many minutes.

16 years ago[PATCH] udevd: add possible initialization of expected_seqnum
kay.sievers@vrfy.org [Sun, 16 Jan 2005 04:53:29 +0000 (05:53 +0100)]
[PATCH] udevd: add possible initialization of expected_seqnum

UDEVD_EXPECTED_SEQNUM=$[`cat /sys/kernel/hotplug_seqnum` + 1] ./udevd -d
will init udevd, so the next event will not wait for any timeout.

16 years ago[PATCH] udevd: it's obviously not the brightest idea to exit a device node manager...
kay.sievers@vrfy.org [Sun, 16 Jan 2005 03:39:02 +0000 (04:39 +0100)]
[PATCH] udevd: it's obviously not the brightest idea to exit a device node manager if it doesn't find /dev/null

16 years ago[PATCH] udevd: separate socket handling to prepare for other event sources
kay.sievers@vrfy.org [Sun, 16 Jan 2005 03:08:54 +0000 (04:08 +0100)]
[PATCH] udevd: separate socket handling to prepare for other event sources

16 years ago[PATCH] udevd: support -d switch to become a daemon
kay.sievers@vrfy.org [Sun, 16 Jan 2005 03:06:22 +0000 (04:06 +0100)]
[PATCH] udevd: support -d switch to become a daemon

16 years ago[PATCH] add aoe block and char device rules to the gentoo rule file.
greg@kroah.com [Wed, 12 Jan 2005 15:38:49 +0000 (07:38 -0800)]
[PATCH] add aoe block and char device rules to the gentoo rule file.

16 years ago[PATCH] udev_volume_id: version 27
kay.sievers@vrfy.org [Wed, 12 Jan 2005 09:44:52 +0000 (10:44 +0100)]
[PATCH] udev_volume_id: version 27

Fix FAT label reading bug for very large volumes.
Recognize FAT label at Win98 formatted volumes.
Read iso9660 joliet descriptor for unicode labels.
Support uuid/label of swap partitions.

16 years ago[PATCH] This fixes a silly mistake in how udevinfo prints the major and minor
roland@digitalvampire.org [Wed, 12 Jan 2005 09:34:02 +0000 (10:34 +0100)]
[PATCH] This fixes a silly mistake in how udevinfo prints the major and minor

16 years ago[PATCH] udevd: split up message receiving an queueing
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:37:50 +0000 (05:37 +0100)]
[PATCH] udevd: split up message receiving an queueing

16 years ago[PATCH] remove useless warning if udev.conf contains keys not read by udev itself
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:37:01 +0000 (05:37 +0100)]
[PATCH] remove useless warning if udev.conf contains keys not read by udev itself

16 years ago[PATCH] improve event sequence serialization
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:35:24 +0000 (05:35 +0100)]
[PATCH] improve event sequence serialization

We delay the execution of events if there is already an event running for
the same device, a parent or a child device. "add" events with a reference
to a physical device will also wait for the physical device to finish.

Here we fix the devpath comparison logic to return "busy" only if a complete
device names is matching. /block/sdaa will not longer block a /block/sda event.

16 years ago[PATCH] remove udevsend syslog noise on udevd startup
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:33:49 +0000 (05:33 +0100)]
[PATCH] remove udevsend syslog noise on udevd startup

16 years ago[PATCH] limit the initial timeout of the udevd event handling
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:33:26 +0000 (05:33 +0100)]
[PATCH] limit the initial timeout of the udevd event handling

Define an inititialization phase, where we delay the events only
for a short time to eventually recollect the event sequence instead
of delaying the very first events for 10 seconds, which breaks the
firmware loader.

16 years ago[PATCH] correct detection of hotplug.d/ udevsend loop
kay.sievers@vrfy.org [Wed, 5 Jan 2005 04:31:42 +0000 (05:31 +0100)]
[PATCH] correct detection of hotplug.d/ udevsend loop

If udevsend is called from the commandline the env may be too
big to fit in our buffer and we tag the message.

16 years ago[PATCH] correct log statement
kay.sievers@vrfy.org [Wed, 22 Dec 2004 00:41:03 +0000 (01:41 +0100)]
[PATCH] correct log statement

16 years ago[PATCH] remove default_* permissions from udev.conf file
kay.sievers@vrfy.org [Tue, 21 Dec 2004 05:10:57 +0000 (06:10 +0100)]
[PATCH] remove default_* permissions from udev.conf file

With the "permissions only rules" we can just place:

   MODE="0660", OWNER="root", GROUP="root"

at the beginning of the rules file and get exactly the same behavior.
If no values are given the compiled-in defaults are used.

16 years ago[PATCH] update Fedora config files and add some more tests
kay.sievers@vrfy.org [Mon, 20 Dec 2004 23:19:34 +0000 (00:19 +0100)]
[PATCH] update Fedora config files and add some more tests

16 years ago[PATCH] allow permissions only rules
kay.sievers@vrfy.org [Mon, 20 Dec 2004 22:44:57 +0000 (23:44 +0100)]
[PATCH] allow permissions only rules

On Mon, 2004-12-20 at 15:03 +0100, Marco d'Itri wrote:
> +# all block devices
> > +SUBSYSTEM="block",   NAME="%k", GROUP="disk"
> > +
> NAME="%k" here and in similar rules does not add any new information,
> what about making it optional, like it is for SYMLINK-only rules?

16 years ago[PATCH] add SUBSYSTEM rule to catch all block devices and apply the disk permissions
kay.sievers@vrfy.org [Mon, 20 Dec 2004 21:24:19 +0000 (22:24 +0100)]
[PATCH] add SUBSYSTEM rule to catch all block devices and apply the disk permissions

16 years ago[PATCH] update Fedora config files
kay.sievers@vrfy.org [Mon, 20 Dec 2004 08:53:32 +0000 (09:53 +0100)]
[PATCH] update Fedora config files

16 years ago[PATCH] handle renamed network interfaces properly if we manage hotplug.d/
kay.sievers@vrfy.org [Mon, 20 Dec 2004 07:57:31 +0000 (08:57 +0100)]
[PATCH] handle renamed network interfaces properly if we manage hotplug.d/

If we take over the hotplug call and manage the events we don't need
to call the event fake script in dev.d/. Just set all expected values
to the new network interface name and call hotplug.d/. This way the
device renaming is completely handled inside of udev and userspace
can't get confused.

16 years ago[PATCH] allow multiline rules by backslash at the end of the line
kay.sievers@vrfy.org [Mon, 20 Dec 2004 06:38:33 +0000 (07:38 +0100)]
[PATCH] allow multiline rules by backslash at the end of the line

On Sun, 2004-12-19 at 18:31 +0100, Marco d'Itri wrote:
> > On Dec 19, Kay Sievers <kay.sievers@vrfy.org> wrote:
>
> > (Feature request: would it be possible to extend the rules files parser
> > to support continuation lines? I'd like it to consider lines starting
> > with white space as part of the previous line.)
>
> How about the usual backslash at the end of the line. Here is a simple
> patch.

16 years ago[PATCH] add OnStream tape drive rules
kay.sievers@vrfy.org [Mon, 20 Dec 2004 02:11:22 +0000 (03:11 +0100)]
[PATCH] add OnStream tape drive rules

On Sun, 2004-12-19 at 17:36 +0000, Willem Riede wrote:
Would you mind adding OnStream tape drives?
>
> KERNEL="osst*",         NAME="%k", GROUP="tape", MODE="0660"
> KERNEL="nosst*",        NAME="%k", GROUP="tape", MODE="0660"

16 years ago[PATCH] simplify rules file by setting default mode to 0660
kay.sievers@vrfy.org [Mon, 20 Dec 2004 02:04:11 +0000 (03:04 +0100)]
[PATCH] simplify rules file by setting default mode to 0660

On Sun, 2004-12-19 at 18:31 +0100, Marco d'Itri wrote:
> It may be better to use default_mode="0660", I do not think that there
> are any mode 600 devices. This would allow setting only the group for
> most of them.

16 years ago[PATCH] simplify permission application
kay.sievers@vrfy.org [Mon, 20 Dec 2004 00:31:56 +0000 (01:31 +0100)]
[PATCH] simplify permission application

We have only one source of permissions now, so apply the default
permissions at udev init.

16 years ago[PATCH] I broke the extras/ again. Add simple build test script now.
kay.sievers@vrfy.org [Sun, 19 Dec 2004 01:39:52 +0000 (02:39 +0100)]
[PATCH] I broke the extras/ again. Add simple build test script now.

16 years ago[PATCH] initial merge of fedora udev.permissions into udev.rules
kay.sievers@vrfy.org [Sat, 18 Dec 2004 21:59:21 +0000 (22:59 +0100)]
[PATCH] initial merge of fedora udev.permissions into udev.rules

16 years ago[PATCH] remove permissions file mentioning from the udev man page
kay.sievers@vrfy.org [Sat, 18 Dec 2004 21:58:37 +0000 (22:58 +0100)]
[PATCH] remove permissions file mentioning from the udev man page

16 years ago[PATCH] fix some typos in gentoo's udev.rules introduced by the merge
kay.sievers@vrfy.org [Sat, 18 Dec 2004 21:57:49 +0000 (22:57 +0100)]
[PATCH] fix some typos in gentoo's udev.rules introduced by the merge

16 years ago[PATCH] initial merge of gentoo udev.permissions into udev.rules
kay.sievers@vrfy.org [Sat, 18 Dec 2004 11:22:16 +0000 (12:22 +0100)]
[PATCH] initial merge of gentoo udev.permissions into udev.rules

16 years ago[PATCH] update Fedora config files
kay.sievers@vrfy.org [Sat, 18 Dec 2004 11:21:21 +0000 (12:21 +0100)]
[PATCH] update Fedora config files

16 years ago[PATCH] update Fedora config files
kay.sievers@vrfy.org [Sat, 18 Dec 2004 11:20:41 +0000 (12:20 +0100)]
[PATCH] update Fedora config files

16 years ago[PATCH] complete removal of explicit udev permissions config file
kay.sievers@vrfy.org [Sat, 18 Dec 2004 10:34:17 +0000 (11:34 +0100)]
[PATCH] complete removal of explicit udev permissions config file

16 years ago[PATCH] I made some more changes to the manpage of udev including:
tklauser@access.unizh.ch [Fri, 17 Dec 2004 23:19:35 +0000 (00:19 +0100)]
[PATCH] I made some more changes to the manpage of udev including:

Patch-From: Tobias Klauser <tklauser@access.unizh.ch>

- Move the description of the environment variables to a new section
  called  "ENVIRONMENT" (as stated in man(7))
- Pointer to the "official" udev homepage
- Fixes for various typos and whitespace damages

16 years ago[PATCH] fix udev_volume_id build error.
greg@kroah.com [Sat, 18 Dec 2004 14:19:12 +0000 (06:19 -0800)]
[PATCH] fix udev_volume_id build error.

bah, when will I remember to check the extras _before_ I do a release...

16 years ago[PATCH] 050 release 050
greg@kroah.com [Sat, 18 Dec 2004 13:52:25 +0000 (05:52 -0800)]
[PATCH] 050 release

16 years ago[PATCH] selinux patch
harald@redhat.com [Fri, 17 Dec 2004 02:59:22 +0000 (03:59 +0100)]
[PATCH] selinux patch

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=142713

/sbin/udevstart segfaults on an ATIIXP chipset which is not supported
well by the kernel yet. There, /proc/ide/hda/media can not be read
(EIO error) and udevstart seems to give a null-pointer to an SELinux
function checking the media-type.

16 years ago[PATCH] update libsysfs to CVS version and fix segfaulting attribute reading
kay.sievers@vrfy.org [Thu, 16 Dec 2004 15:05:01 +0000 (16:05 +0100)]
[PATCH] update libsysfs to CVS version and fix segfaulting attribute reading

16 years ago[PATCH] klibc supports LOG_PID now, so remove our own implementation
kay.sievers@vrfy.org [Wed, 15 Dec 2004 10:27:10 +0000 (11:27 +0100)]
[PATCH] klibc supports LOG_PID now, so remove our own implementation

16 years agobleah, more merge fixes...
Greg KH [Wed, 27 Apr 2005 06:15:56 +0000 (23:15 -0700)]
bleah, more merge fixes...

16 years ago[PATCH] avoid building klibc test programs and pass SUBDIRS= to klibc clean
kay.sievers@vrfy.org [Wed, 15 Dec 2004 10:26:05 +0000 (11:26 +0100)]
[PATCH] avoid building klibc test programs and pass SUBDIRS= to klibc clean

16 years ago[PATCH] 049 release 049
greg@kroah.com [Wed, 15 Dec 2004 07:00:33 +0000 (23:00 -0800)]
[PATCH] 049 release

16 years ago[PATCH] fix 'make clean' error in klibc
greg@kroah.com [Wed, 15 Dec 2004 01:51:58 +0000 (17:51 -0800)]
[PATCH] fix 'make clean' error in klibc

16 years ago[PATCH] update klibc to 0.194
kay.sievers@vrfy.org [Tue, 14 Dec 2004 10:42:06 +0000 (11:42 +0100)]
[PATCH] update klibc to 0.194

16 years ago[PATCH] export DEVNAME regardless of the state of udev_dev_d
kay.sievers@vrfy.org [Tue, 14 Dec 2004 10:06:49 +0000 (11:06 +0100)]
[PATCH] export DEVNAME regardless of the state of udev_dev_d

16 years ago[PATCH] remove leftover from udevinfo's -d option
azarah@nosferatu.za.org [Tue, 14 Dec 2004 08:42:20 +0000 (09:42 +0100)]
[PATCH] remove leftover from udevinfo's -d option

16 years ago[PATCH] add class specific files for class/spi_transport and class/spi_host
kay.sievers@vrfy.org [Sun, 12 Dec 2004 07:34:46 +0000 (08:34 +0100)]
[PATCH] add class specific files for class/spi_transport and class/spi_host

16 years ago[PATCH] udevd-test.pl: remove wrong date calculation
kay.sievers@vrfy.org [Sun, 12 Dec 2004 02:35:39 +0000 (03:35 +0100)]
[PATCH] udevd-test.pl: remove wrong date calculation

A hour wrap during the test run does not work :) Just remove
all the useless date conversions as we are only interested in the
seconds it takes to process.

[root@pim udevd-test]# ./udevd-test.pl 9
   ...
   device: /class/tty/console, action: remove
   forking udev time:     Sat Dec 11 18:59:57 2004
   the delay time is:     3 s

   device: /class/tty/ptmx, action: remove
   forking udev time:     Sat Dec 11 19:00:07 2004
   the delay time is:     4013 s

   the delay time is: 4013
   udevd doesn't act properly.

16 years ago[PATCH] check earlier if we should run as udevstart
kay.sievers@vrfy.org [Sun, 12 Dec 2004 02:29:59 +0000 (03:29 +0100)]
[PATCH] check earlier if we should run as udevstart

16 years ago[PATCH] remove double initialization
kay.sievers@vrfy.org [Sun, 12 Dec 2004 01:54:55 +0000 (02:54 +0100)]
[PATCH] remove double initialization

16 years ago[PATCH] include missing header to udevtest.c
kay.sievers@vrfy.org [Sun, 12 Dec 2004 01:32:25 +0000 (02:32 +0100)]
[PATCH] include missing header to udevtest.c

16 years ago[PATCH] add -V option to udev to print the version number
kay.sievers@vrfy.org [Sun, 12 Dec 2004 01:27:50 +0000 (02:27 +0100)]
[PATCH] add -V option to udev to print the version number

On Sat, 2004-12-11 at 18:44 +0200, Martin Schlemmer [c] wrote:
>
> Any suggestions to determining the version of the installed udev?
> This is now during startup, to see if we can make use of using
> udevsend as hotplug agent.  If the system was up, udevinfo could
> be used, but that is in /usr/bin that might be on a seperate /usr.
> I know we might move udevinfo to /bin, but that might be an issue
> for some, and adding a -V switch to /sbin/udev might be a better
> choice.

16 years ago[PATCH] prevent udev node creatinon for "class" registration
kay.sievers@vrfy.org [Sun, 12 Dec 2004 01:03:14 +0000 (02:03 +0100)]
[PATCH] prevent udev node creatinon for "class" registration

I've found a /dev/video4linux node and just realized, that libsysfs
searches all subdirs for an attribute name.
So it found /class/video4linux/video0/dev for the videodev class
creation event /class/video4linux and created a node.

Just ignore the SUBSYSTEM="class" events now.

16 years ago[PATCH] udevd: serialization of the event sequence of a chain of devices
kay.sievers@vrfy.org [Sat, 11 Dec 2004 20:43:08 +0000 (21:43 +0100)]
[PATCH] udevd: serialization of the event sequence of a chain of devices

Currently udevd delays only events for the same DEVPATH.

Example of an "add" event sequence:
  /block/sda
  /block/sda/sda1

With this change, we make sure, that the udev process handling
/block/sda has finished its work (waited for all attributes,
created the node) before we fork the udev event for /block/sda/sda1.
This way the event for sda1 can be sure, that the node for the
main device is already created (may be useful for disk labels).

It will not affect any parallel device handling, only the sequence
of the devices directory chain is serialized. The 10.000 disks
plugged in will still run as parallel events. :)

The main motivation to do this is the program execution of the
dev.d/ and hotplug.d/ directory. If we don't wait for the parent
event to exit, we can't be sure that the executed scripts are
run in the right order.

On Thu, Dec 09, 2004 at 09:18:28AM +0100, Kay Sievers wrote:
> On Wed, 2004-12-08 at 19:07 -0800, David Brownell wrote:
> > Could that argument apply to the underlying hardware, too?

> We now make sure that the sequence of events for a device
> is serialized for every device chain and the class/block
> devices which have a "device" link to a physical device are
> handled after the physical device is fully populated and
> notified to userspace. It will only work this way on kernels
> later than 2.6.10-rc1 cause it depends on the PHYSDEVPATH
> value in the hotplug environment.

16 years ago[PATCH] add a class/fc_host file to the list of what to wait for
kay.sievers@vrfy.org [Sat, 11 Dec 2004 05:07:31 +0000 (06:07 +0100)]
[PATCH] add a class/fc_host file to the list of what to wait for

16 years ago[PATCH] udev_volume_id: links sysfs.a instead of all objects
kay.sievers@vrfy.org [Thu, 9 Dec 2004 05:06:17 +0000 (06:06 +0100)]
[PATCH] udev_volume_id: links sysfs.a instead of all objects

cuts down the size from:
  dynamically linked: 48.030 -> 45.187
  statically linked:  149.177 ->  142.362

16 years ago[PATCH] 048 release 048
greg@kroah.com [Thu, 9 Dec 2004 03:42:47 +0000 (19:42 -0800)]
[PATCH] 048 release

16 years ago[PATCH] fix udev_volume_id so it will now build properly.
greg@kroah.com [Thu, 9 Dec 2004 03:35:29 +0000 (19:35 -0800)]
[PATCH] fix udev_volume_id so it will now build properly.

16 years ago[PATCH] fix scsi_id build errors due to changes in the main udev makefile
greg@kroah.com [Thu, 9 Dec 2004 03:30:41 +0000 (19:30 -0800)]
[PATCH] fix scsi_id build errors due to changes in the main udev makefile

also fixed up the QUIET issue when it was disabled for the extras programs

16 years ago[PATCH] 047 release 047
greg@kroah.com [Thu, 9 Dec 2004 02:29:04 +0000 (18:29 -0800)]
[PATCH] 047 release

16 years ago[PATCH] add more debian rules files.
greg@kroah.com [Thu, 9 Dec 2004 01:09:40 +0000 (17:09 -0800)]
[PATCH] add more debian rules files.

16 years ago[PATCH] move distro specific config files into their own directories
greg@kroah.com [Thu, 9 Dec 2004 01:08:20 +0000 (17:08 -0800)]
[PATCH] move distro specific config files into their own directories