Udev-108

[Matthew Burgess]

New version.

udev 107
========

Bugfixes.

Symlinks can have priorities now, the priority is assigned to the device and specified with OPTIONS="link_priority=100". Devices with higher priorities overwrite the symlinks of devices with lower priorities. If the device that currently owns the link, goes away, the symlink will be removed, and recreated, pointing to the next device with the highest actual priority. This should make /dev/disk/by-{label,uuid,id} more reliable, if multiple devices contain the same metadata and overwrite these symlinks.

The dasd_id program is removed from the udev tree, and dasdinfo, with the needed rules, are part of the s390-tools now.

Please add KERNEL=="[0-9]*:[0-9]*" to the scsi wait-for-sysfs rule, we may get the scsi sysfs mess fixed some day, and this will only catch the devices we are looking for.

USB serial numbers for storage devices have the target:lun now appended, to make it possibble to distinguish broken multi-lun devices with all the same SCSI identifiers.

Note: The extra "run_directory" which searches and executes stuff in /etc/hotplug.d/ and /etc/dev.d/ is long deprecated, and will be removed with the next release. Make sure, that you don't use it anymore, or provides your own implementation of that inefficient stuff. We are tired of reports about a "slow udev", because these directories contain stuff, that runs with _every_ event, instead of using rules, that run programs only for the matching events.

[Matthew Burgess]

Now Udev-108. The changelog just mentions bugfixes and the removal, finally, of the directory multiplexer for dev.d/ and hotplug.d/.

[alexander@…]

Since there are no longer any non-needed helpers, I propose to simplify the compilation instructions:

make EXTRAS="`echo extras/*/`"
make DESTDIR=/ EXTRAS="`echo extras/*/`" install

[bryan@linuxfromscratch.org]

The changelog (​http://www.kernel.org/git/?p=linux/hotplug/udev.git;a=blob;hb=HEAD;f=ChangeLog -- if you get an XML well-formed-ness error from that URL, refresh the page once or twice) mentions one thing that will affect us. The write_cd_rules script was changed to use by-id by default for USB and Firewire CD drives, instead of by-path.

One possible issue is that this is based on $ID_BUS; that value is only set by usb_id (sets it to usb), ata_id (sets it to ata), and scsi_id (sets it to scsi). I don't see anything that sets it to "ieee1394", which is one value that the script looks for. The other issue is that I'm not sure that usb_id / ata_id / scsi_id are run before write_cd_rules. I'll have to look into it more closely.

Anyway, the current book (in section 7.12) says the default is by-path, and the user should override it if they want by-id (by running a sed on the rules file that adds "by-id" to the write_cd_rules arguments). To make it work with -108, we probably need to send the user to one of two subsections depending on what kind of CD drive they have; if it's USB or Firewire, tell them the default is by-id and if they want by-path (not recommended), they should add that to the rule. Otherwise, if the device isn't USB or Firewire, then tell them the default is by-path and if they want by-id, add that to the rule.

[Matthew Burgess]

Bryan, is the attached patch accurate?

[bryan@linuxfromscratch.org]

Looks fine to me after a simple read-through. :-)

(I'd probably have used the <quote> element instead of manual &quot; attributes personally, because I don't think the &quot; attributes will get "angled" quote marks, while the <quote> element will. But that doesn't matter all that much, either: the real contents of the patch are fine.)

[Matthew Burgess]

Thanks, Bryan. I've updated it to use <quote> now, I always forget about that! Now, as we're right in the area that #1972 will be touching, we may as well fix that up while we're here. I'll postpone applying this patch until we've sorted out the details in that ticket.

[Matthew Burgess]

Fixed in r8064.

[Jeremy Huntwork]

Milestone 6.3 deleted