Re: [ANNOUNCE] udev 046 release

From: Mathieu Segaud
Date: Fri Nov 19 2004 - 06:27:55 EST


Greg KH <greg@xxxxxxxxx> disait dernièrement que :

> I've released the 046 version of udev. It can be found at:
> kernel.org/pub/linux/utils/kernel/hotplug/udev-046.tar.gz
>
> (yes, there has also been releases for versions 040, 041, 042, 043, 044,
> and 045, I just forgot to announce them...)
>
> udev allows users to have a dynamic /dev and provides the ability to
> have persistent device names. It uses sysfs and /sbin/hotplug and runs
> entirely in userspace. It requires a 2.6 kernel with CONFIG_HOTPLUG
> enabled to run. Please see the udev FAQ for any questions about it:
> kernel.org/pub/linux/utils/kernel/hotplug/udev-FAQ
>
> For any udev vs devfs questions anyone might have, please see:
> kernel.org/pub/linux/utils/kernel/hotplug/udev_vs_devfs
>
> And there is a general udev web page at:
> http://www.kernel.org/pub/linux/utils/kernel/hotplug/udev.html
>
> This release has a completly new backend database. If you have had
> _any_ problems with udev lockups, 100% cpu utilitization, or issues with
> large machines having a zillion udev processes hanging around, please
> try this release.
>
> For those of you without any udev issues, I recommend the 045 release
> at:
> kernel.org/pub/linux/utils/kernel/hotplug/udev-045.tar.gz
>
> It's the same backend database, and has lots of bugfixes and tweaks
> since the 039 release.
>
> Major changes in this release from the 045 release:
> - no more tdb backend, we now have 1 file per sysfs entry in
> /dev/.udevdb that holds the information we need. Yes, this is
> a bit more memory than previously used, but we get the locking
> issues fixed for free as we rely on the kernel file locking
> rules to get it correct. Anyway, you should have /dev mounted
> on a tmpfs not ramfs so the memory is swapable.
> - new keywords to match on (DRIVER and SUBSYSTEM)
> - lots of good wait_for_sysfs fixes (we now take advantage of
> the new 2.6.10-rc2 kernel information and don't have to guess
> at what files we might have to wait for.)
> - loads of bugfixes and rule files tweaks.

seems like these changes broke something in rules applying to eth* devices.
the rules put and still working with udev 045 have no effect, now....
not so inconvenient now that I've got just one card in my box, but I guess
it could be a show-stopper for laptop users.

My rules which can be found at the end of /etc/udev/rules.d/50-udev.rules are:

KERNEL="eth*", SYSFS{address}="00:10:5a:49:36:d8", NAME="external"
KERNEL="eth*", SYSFS{address}="00:50:04:69:db:56", NAME="private"
KERNEL="eth*", SYSFS{address}="00:0c:6e:e4:2c:81", NAME="dmz"

Regards,


>
> Thanks to Kay Sievers for all the work he's been doing on udev lately,
> especially the database backend fixes. I really appreciate it.
>
> Also thanks to everyone who has send me patches for the recent releases,
> a full list of everyone, and their changes is below.
>
> udev development is done in a BitKeeper repository located at:
> bk://linuxusb.bkbits.net/udev
>
> Daily snapshots of udev from the BitKeeper tree can be found at:
> http://www.codemonkey.org.uk/projects/bitkeeper/udev/
> If anyone ever wants a tarball of the current bk tree, just email me.
>
> thanks,
>
> greg k-h
>
>
> Summary of changes from v045 to v046
> ============================================
>
> Greg Kroah-Hartman:
> o make spotless for releases
>
> Kay Sievers:
> o Don't try to print major/minor for devices without a dev file
> o remove get_device_type and merge that into udev_set_values()
> o prevent udevd crash if DEVPATH is not set
> o add ippp and bcrypt to the exception lists of wait_for_sysfs
> o let klibc add the trailing newline to syslog conditionally
> o disable logging for udevstart
> o add NAME{ignore_remove} attribute
> o remove historical SYSFS_attr="value" format
> o don't wait for sysfs if the kernel(2.6.10-rc2) tells us what not to expect
> o change key names in udevinfo sysfs walk to match the kernel
> o support DRIVER as a rule key
> o support SUBSYSTEM as a rule key
> o rename udevdb* to udev_db*
> o Make dev.d/ handling a separate processing stage
> o make the udev object available to more processing stages
> o remove udev_lib dependency from udevsend, which makes it smaller
> o add ACTION to udev object to expose it to the whole process
> o make udevinfo's -r option also workimg for symlink queries
> o let udev act as udevstart if argv[1] == "udevstart"
> o improve udevinfo sysfs info walk
> o add sysfs info walk to udevinfo
> o pass the whole event environment to udevd
> o replace tdb database by simple lockless file database
>
>
> Summary of changes from v044 to v045
> ============================================
>
> Martin Schlemmer:
> o Some updates for Gentoo's udev rules
>
>
> Summary of changes from v043 to v044
> ============================================
>
> Greg Kroah-Hartman:
> o add cdsymlinks.sh support to gentoo rules file
> o fix gentoo legacy tty rule
> o remove 'sudo' usage from the Makefile
> o make udev-test.pl test for root permissions before running
>
> Kay Sievers:
> o reduce syslog noise of udevsend if multiple instances try to start udevd
> o add i2c-dev to the list of devices without a bus
>
>
> Summary of changes from v042 to v043
> ============================================
>
> Greg Kroah-Hartman:
> o add test target to makefile
> o add dumb script to show all sysfs devices in the system
>
> Kay Sievers:
> o Shut up wait_for_sysfs class/net failure messages, as it's not possible to
> get that right for all net devices. Kernels later than 2.6.10-rc1 will
> handle that by carrying the neccessary information in the hotplug event.
> o wait() for specific pid to return from fork()
> o Don't use any syslog() in signal handler, cause it may deadlock
> o Add support for highpoint ataraid to volume_id to suppress label reading on raid set members.
> o Add a bunch of devices without "device" symlinks
> o Exit, if udevtest cannot open the device (segfault)
> o Patches from Harald Hoyer <harald@xxxxxxxxxx>
> o Apply the default permissions even if we found a entry in the permissions
> file. Correct one test, as the default is applied correctly now and the
> mode will no longer be 0000.
> o add test for format chars in multiple symlinks to replace
> o Add net/vmnet and class/zaptel to the list of devices without physical device
>
>
> Summary of changes from v040 to v042
> ============================================
>
> Greg Kroah-Hartman:
> o add inotify to the rules for gentoo
>
> Kay Sievers:
> o skip waiting for device if we get a bad event for class creation and not for a device underneath it
> o add net/pan and net/bnep handling
> o switch wait for bus_file to stat() instead of open() add net/tun device handling add ieee1394 device handling
> o Remove the last klibc specific line from the main udev code Move _KLIBC_HAS_ARCH_SIG_ATOMIC_T to the fixup file which is automatically included by the Makefile is we build with klibc
> o ignore *.rej files from failed patches
> o update to libsysfs 1.2.0 and add some stuff klib_fixup Now we have only the sysfs.h file different from the upstream version to map our dbg() macro.
> o improve klibc fixup integration
> o cleanup udevd/udevstart
> o expose sysfs functions for sharing it
>
>
> Summary of changes from v039 to v040
> ============================================
>
> <jk:blackdown.de>:
> o wait_for_sysfs update for dm devices
>
> Greg Kroah-Hartman:
> o sparse cleanups on the tree
> o fix stupid cut-and-paste error for msr devices on gentoo boxes
> o add *~ to bk ignore list
> o delete udevruler.c as per Kay's request
> o fix up the wait_for_sysfs_test script a bit
>
> Kay Sievers:
> o fix debug in volume id / fix clashing global var name
> o volume_id fix
> o $local user
> o cleanup netif handling and netif-dev.d/ events
> o big cleanup of internal udev api
> o don't wait for dummy devices
> o close the syslog
> o Fix ppp net devices in wait_for_sysfs
> o Fix wait_for_sysfs messages (more debugging info)
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: InterSystems CACHE
> FREE OODBMS DOWNLOAD - A multidimensional database that combines
> robust object and relational technologies, making it a perfect match
> for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
> _______________________________________________
> Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
> Linux-hotplug-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel
>

--
<riel> google rules
<google> rules: http://www.law.cornell.edu/rules/fre/overview.html

- Rik van Riel chatting with the bots on #kernelnewbies

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/