From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Zeuthen Date: Mon, 16 Aug 2004 20:44:50 +0000 Subject: Re: Udev infinite loop Message-Id: <1092689091.3950.28.camel@localhost.localdomain> List-Id: References: <411C9E97.4050707@yahoo.fr> In-Reply-To: <411C9E97.4050707@yahoo.fr> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org On Mon, 2004-08-16 at 10:43 +0200, David Zeuthen wrote: > On Mon, 2004-08-16 at 10:12 +0200, Guillaume Chazarain wrote: > > That's bad news, it means you cannot reproduce it :( > > Here is a tar.bz2 containing strace logs of dbus-daemon-1 and udevd. > > There is also a dmesg > > output, my kernel .config, /etc/udev /sys and kernel logs. > > I see nothing wrong in them, but hopefully you will. > > Guillaume: I believe I've fixed this problem in the just released hal 0.2.97, at least it works nicely with my PCMCIA card reader which showed the same issues. I see that you are running Fedora Core development; updated hal-0.2.97 packages should appear in rawhide soon [1]. It would be great if you can test these packages ('rpm -q hal' should report hal-0.2.97-1 or higher) and if it fails then please report the bug to http://bugzilla.redhat.com or the hal@freedesktop.org mailing list. If it just works, also feel free to mail me personally; I would appreciate it! Thanks. > This is not an udev problem; this is caused by the fact that hal polls > your device for media; fwiw, I've seen this with the ide-cs driver as > well when using a PCMCIA CF card reader and I'm looking into > blacklisting such devices or at least working around issues. Please > state your version of hal and include the output of 'lshal' so we can > get this resolved. > > To me, it seems that if you do a 'dd if=/dev/hde of=/tmp/foo bsQ2 > count' then it causes a all partitions on /dev/hde to be hot > unplugged and then hot replugged. Uhm, isn't that some kind of bug in > the driver or what. Hotplug folks: I'd just like to note that this problem is not exclusive to hal at all, it even applies to mount(1) if the filesystem given is 'auto', cause then mount(1) access the device file and this will cause a hotplug removal and then a hotplug add. Specifically, using udev callouts to get the drive serial number or volume lanel will also cause infinite loops. Can anyone please confirm this is a bug in the IDE subsystem? Thanks, David [1] : Alternatively download the (S)RPMs from here http://freedesktop.org/~david/rpms/ ------------------------------------------------------- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 _______________________________________________ Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net Linux-hotplug-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel