[10:27] <wget> Hi guys. Wrt a question I posted here during the week-end, but as you were absent, let me ask it again.
[10:27] <wget> Guys: I had a discussion ages ago with one of your kernel maintainer about this topic, and it's still not resolved and still impacting Ubuntu (while haven't tested 15.x nor 16.x series).
[10:27] <wget> https://bugzilla.kernel.org/show_bug.cgi?id=75381
[10:27] <wget> that bug is related to this one: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1269883?comments=all
[10:27] <wget> DO you have any news about this?
[10:36] <davmor2> wget: is it using kernel mod alx by any chance?
[10:36] <davmor2> wget: I ask due to this issue http://askubuntu.com/questions/616182/alx-on-ar8191-does-not-appear-to-be-working-15-04/620489#620489
[10:38] <wget> davmor2:  lsmod |grep alx doesn't report anything when the adapter is connected
[10:38] <wget> so I assume it isn't using that mod
[10:38] <wget> the device is not a quallcomm atheros
[10:39] <davmor2> wget: ah okay no worries 
[10:39] <wget> mine is an Asix ax88179_178a
[10:39] <wget> like all USB3.0 to Ethernet you can find in Europe
[11:01] <TJ-> I've got a live issue here right now with 16.04, wonder if anyone has some ideas? GUI user with $HOME/.Xauthority file, which is subsequently "mount --bind /dev/shm/user/.Xauthority $HOME/.Xauthority" .. "xauth -f ... list" shows the MIT_MAGIC_COOKIE on both 'sides' identical... but then, after doing "xauth -f /dev/shm/user/.Xauthority generate :0 ." the cookies on each side of the bind mount differ
[11:01] <TJ-> and a hash of the file confirms the actual files are different - the 'target' doesn't reflect the content of the updated 'source'. Is this a weird kernel issue?
[11:07] <apw> TJ-, you are bind mounting a file ?
[11:08] <apw> TJ-, it looks like so, if so, it could well be linking the new name to the original inode
[11:08] <TJ-> apw: yes, correct, ahhhh and xauth (I read the source!) does a hard-link rename!
[11:09] <TJ-> apw: THANK you!! This has been causing some automation bugs that have been very annoying
[11:09] <apw> you can prolly prove that by linking the original file to a temp name before you replace the auth
[11:09] <TJ-> so, need to figure out a workaround
[11:10] <apw> and then catting into the extra name, and see if the one on the real home then changes
[11:10] <TJ-> yeah, that makes sense now. OK, so we just need to take care of the timing of the --bind 
[11:11] <apw> symlink the .Xauthority into a like .config/Xauthority/.Xauthority
[11:11] <apw> and bind mount the .config/Xauthority perhaps ?
[11:11] <apw> or .config
[11:26] <TJ-> apw: that did it, thanks :)
[12:44] <spice> Hi all! I have to following issue: I develop a Linux network driver. The driver compiled and worked without any issues. After I upgraded the gcc package on my Ubuntu 14.04.4 LTS machine, the kernel module produces paging bugs.
[12:56] <god> spice, what did you upgrade from, to?
[13:00] <spice> that is a good question ... I think it was a rahter minor gcc upgrade
[13:01] <spice> right now it is:  4.8.4-2ubuntu1~14.04.3 Before it might have been 4.8.4-1ubuntu.... or so.
[13:01] <spice> Is there any way to reinstall previous version of packages to make sure my claim is correct?
[13:04] <apw> spice, yes, all old versions are available via the launchpad librarian
[13:04] <apw> https://launchpad.net/ubuntu/+source/gcc-4.8/+publishinghistory
[13:07] <spice> okay, i will try that and see if this is definitely the culprit of my problem. Thank you for now!
[13:08] <apw> spice, it is not uncommon for a gcc upgrade to expose existing bugs through changes in optmisation, the -fpie becoming default in yakkety is finding all sorts of latent bugs
[14:38] <tseliot> rtg: all the nvidia drivers should be fixed in 16.10 now
[14:39] <tseliot> mamarley: ^
[14:41] <rtg> tseliot, thanks, I'll get the kernel ADT tests re-ran
[14:42] <tseliot> rtg: yes, as soon as the nvidias are published
[14:42] <rtg> right
[14:42] <tseliot> (I've just uploaded the fixes)
[14:45] <mamarley> \o/