Am 12.10.2012 12:33, schrieb Michael Schmitt:
Hi Jonathan,
Am 11.10.2012 02:11, schrieb Jonathan Nieder:
Michael Schmitt wrote:
Am 11.10.2012 00:23, schrieb Jonathan Nieder:
- What was the first non-working kernel? If you have time to bisect
through the pre-compiled kernels at
http://snapshot.debian.org/package/linux/ to find the oldest
broken
and newest non-broken one, that would be very useful.
Dunno if I got you right there, but I will try to test all
kernel-packages
between the working and non-working one tomorrow which are available at
snapshot.d.o. What I can tell now: 3.4.1-1~experimental.1 (working);
3.5.2-1~experimental.1 (broken). And I realize now the 3.4 was from
experimental as well and not sid... but I guess that doesn't matter
much.
By "bisect" I meant one can try the package halfway between the newest
working and oldest broken version and repeat as necessary, which takes
O(log n) reboots instead of O(n). But since there are only two
kernels in that range to test the distinction doesn't make a
difference. So you understood correctly. :)
Glad to hear! :) And done as requested, see the "full-cycle" dmesg
logs (cold-boot, suspend, resume, modprobe) attached. With
3.5-1~experimental.1 I did test it without X too. After the first
resume, I did suspend again, resumed, stopped X and did check for any
output from /dev/input/mice while moving the mouse >> none. After
modprobing hid_logitech it did work again. And with
3.4.4-1~experimental.1 no issue. So the issue was introduced between
3.4.4-1~experimental.1 and 3.5-1~experimental.1. And the only message
that did keep popping up that looks suspicious to me, "usbhid
2-1.3:1.1: >reset_resume error 1" as noted in the first message already.
Thanks for the kernel log.
Ciao,
Jonathan
A little digging on my own revealed
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1049623
especially
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1049623/comments/5 (oh
I hate ubuntu bugreports.... they get so damn noisy) and
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=dc3c78e43469063c5bf4b744214508f94c4129f9
for a possible fix.
Anything more you need?
regards
Michael
Err... silly me, the link on git.kernel.org was the possible commit that
introduced that issue, the possible fix is in the ubuntu-bugreport in
comment-#6.
regards
Michael
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
