udevid segfaulting

Bryan Kadzban bryan at kadzban.is-a-geek.net
Sat Oct 27 17:43:04 PDT 2007


-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Bruce Dubbs wrote:
> Bruce Dubbs wrote:
>> Matthew Burgess wrote:
>>> Hi,
>>> 
>>> I've just built an SVN version of LFS, along with several package
>>> updates (everything in Trac besides grep, texinfo and udev).
>>> I'm getting the following on bootup:
>>> 
>>> Has anyone seen anything similar before, or might this be down to
>>> GCC-4.2.2 or Glibc-2.7 upgrades?  Is there any way of debugging 
>>> these segfaults earlier (for example, can I check what arguments 
>>> udev_id was being passed to so I might be able to attach gdb to 
>>> the problematic process and get a full stacktrace)?
>> 
>> I haven't, but the internet has.  Google for: kernel usb_id 
>> segfault
>> 
>> It looks like a kernel driver problem.
> 
> After a little more looking, the kernel guys think it us a udev 
> problem. Try backing off on the udev version.

It is a udev problem, but backing off to an earlier version won't help.
There was a thread on linux-hotplug-devel about this issue, and the
reason the bug showed up there was that the sysfs layout in the newer
kernel changed (and CONFIG_SYSFS_DEPRECATED was off?), I think.  Anyway
the thread is here:

http://marc.info/?t=119245141800002&r=1&w=2

Specifically, this post:

http://marc.info/?l=linux-hotplug-devel&m=119245213012211&w=2

where Kay is saying usb_id needs to be fixed, and this udev git commit:

http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;h=963033472f8d2aab0bf77147b916de8f3b0b5450

which is supposed to be the fix.  Try adding that patch to your udev
build.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHI9sXS5vET1Wea5wRAzUdAKDHxM+IK+4iBJyQqcxnTofZ9kEEsQCdEquN
AeVhV4ZtTzzwTseSvqt52lw=
=BXEE
-----END PGP SIGNATURE-----



More information about the lfs-dev mailing list