Closed GoogleCodeExporter closed 9 years ago
Sometimes dkms does not recompile module after upgrading the kernel;
reinstalling the .deb package is recommended
Original comment by SergeyK...@gmail.com
on 26 Dec 2010 at 11:37
Recommended, but, alas, not doing the job.
I did apt-get remove --purge && apt-get install for easy-slow-down and
samsung-tools, but the situation remains the same: Full power
backlight, and the slider dysfunctional.
Linux 2.6.35-24-generic #43~ppa1~loms~maverick-Ubuntu SMP Fri Dec 24
18:15:40 UTC 2010 i686 GNU/Linux
What next?
Uwe
Original comment by udip...@gmail.com
on 27 Dec 2010 at 3:37
Though I did the purge && install with 2.6.24, and it doesn't work with it,
when I boot to the older version,
2.6.35-23-generic #42~ppa1~loms~maverick-Ubuntu SMP Tue Nov 30 02:35:28 UTC
2010 i686 GNU/Linux
it works splendidly.
You need anything from my side for debugging?
Uwe
Original comment by udip...@gmail.com
on 1 Jan 2011 at 10:22
- what is the output of 'dkms status'?
- do you have samsung-backlight.ko in
/lib/modules/2.6.35-24-generic/updates/dkms/
- what is the output of 'sudo insmod
/lib/modules/2.6.35-24-generic/updates/dkms/samsung-backlight.ko' under the new
kernel?
- do you have a boot option acpi_backlight=vendor i915.modeset=1 with the new
kernel?
- do you have any backlight module related messages in log (dmesg | grep
Samsung)?
Original comment by SergeyK...@gmail.com
on 7 Jan 2011 at 8:14
Do you use voria kernel?
Original comment by SergeyK...@gmail.com
on 7 Jan 2011 at 10:31
No, (I don't know what voria is supposed to be), just the plain,
standard, ubuntu kernel for netbook remix.
Uwe
Original comment by udip...@gmail.com
on 7 Jan 2011 at 12:19
easy_slow_down_manager, 0.13.3, 2.6.35-24-generic, i686: installed
samsung_backlight, 0.13.3, 2.6.35-22-generic, i686: built
samsung_backlight, 0.13.3, 2.6.35-24-generic, i686: installed
samsung_backlight, 0.13.3, 2.6.35-23-generic, i686: installed
-rw-r--r-- 1 root root 9572 2010-12-27 11:29 easy_slow_down_manager.ko
-rw-r--r-- 1 root root 14092 2010-12-22 08:52 samsung-backlight.ko
insmod: error inserting
'/lib/modules/2.6.35-24-generic/updates/dkms/samsung-backlight.ko': -1
File exists
(I don't understand)
[ 15.282885] Samsung-backlight: checking for SABI support.
[ 15.283074] Samsung-backlight: SABI is supported (f5191)
Please ask, whenever you need more!
Uwe
Original comment by udip...@gmail.com
on 7 Jan 2011 at 1:47
Hi, i have the same problem like udippel. Any solution?
Original comment by diegofo...@gmail.com
on 7 Jan 2011 at 3:52
ok, module seems to be installed properly and loaded. Could you attach Xserver
log: /var/log/Xorg.0.log ?
Also, which backlight interfaces do you have? (ls -l /sys/class/backlight)
What is the maximum brightness level? (cat
/sys/class/backlight/samsung/max_brightness)
Original comment by SergeyK...@gmail.com
on 7 Jan 2011 at 4:09
backlight interfaces
lrwxrwxrwx 1 root root 0 2011-01-07 14:15 acpi_video0 ->
../../devices/virtual/backlight/acpi_video0
lrwxrwxrwx 1 root root 0 2011-01-07 14:15 samsung ->
../../devices/virtual/backlight/samsung
maximum brightness level = 7
Original comment by diegofo...@gmail.com
on 7 Jan 2011 at 5:20
Attachments:
Thanks for your help!
Original comment by diegofo...@gmail.com
on 7 Jan 2011 at 5:23
the problem is that you have two interfaces for brightness control: samsung and
acpi and X server chooses the wrong one (acpi);
edit /etc/default/grub by adding (or editing) the line
GRUB_CMDLINE_LINUX_DEFAULT="acpi_backlight=vendor"
run update-grub and reboot
Original comment by SergeyK...@gmail.com
on 7 Jan 2011 at 6:53
Hi SergeyKo81 thanks for you help, a made the changes that you said and now i
can change the bright with Fn + UP but when i star GNOME i have blingking about
30 sec. I looked some forums and i found this
http://www.voria.org/forum/viewtopic.php?f=3&t=625&start=15
Do you have any idea about this? Thanks in advance
Original comment by diegofo...@gmail.com
on 7 Jan 2011 at 7:43
- try playing with the GNOME powersave option 'dim display when idle'
- try disabling brightness change on battery state changes
I had such weird things with xneur (automatic key translation tool), maybe
you're using something similar?
Original comment by SergeyK...@gmail.com
on 7 Jan 2011 at 8:18
Done
lrwxrwxrwx 1 root root 0 2011-01-08 09:05 acpi_video0 ->
../../devices/virtual/backlight/acpi_video0
lrwxrwxrwx 1 root root 0 2011-01-08 09:05 samsung ->
../../devices/virtual/backlight/samsung
7
Hope this helps,
Uwe
Original comment by udip...@gmail.com
on 8 Jan 2011 at 1:15
Uwe, please, see comment #12: the same solution applies to you.
Original comment by SergeyK...@gmail.com
on 8 Jan 2011 at 9:00
When I boot back to the previous 2.6.35-23 kernel, on this same
machine, this works perfectly well and correct:
$ uname -a
Linux 2.6.35-23-generic #42~ppa1~loms~maverick-Ubuntu SMP Tue Nov 30
02:35:28 UTC 2010 i686 GNU/Linux
$ ls -l /sys/class/backlight/
total 0
lrwxrwxrwx 1 root root 0 2011-01-08 18:22 samsung ->
../../devices/virtual/backlight/samsung
$
What do I make from this? Sorry to contradict, but the problem does
not seem to be a missing boot / grub option, but a different behaviour
of the kernel w.r.t. acpi.
Since all previous kernels have behaved well, and only this one
behaves differently, I am very tempted to point fingers at the
compilation of the bumped-up 2.6.35-23 to 2.6.35-24 kernel. Except
that the difference, the patch of a vulnerability, had anything to do
with acpi.
Maybe you might want to ask the maintainer,
Ubuntu Kernel Team <kernel-team@lists.ubuntu.com>
about this change of behaviour?
I don't want to dodge the task, but I don't know enough about your
software to really start arguing with them.
As of now, instead of fiddling with grub, I for one would suggest to
stick with 2.6.35-23 on Ubuntu Samsung netbooks, at least the N-150
family.
To me this looks conspicuously like a bug, a failed compilation option
or setting, or something to that behalf. Again, had not all previous
kernels behaved properly, I'd agree with your solution as suggested in
# 12 and # 16.
A vulnerability patch is not supposed to change behaviour in another
subsystem, though.
My 2 humble sen as a not-kernel-developer,
Uwe
P.S.: Of course, X can go for the wrong API, if the kernel *offers*
it; like in 2.6.35-24
Original comment by udip...@gmail.com
on 8 Jan 2011 at 10:54
The module is intended to replace the default implementation of brightness
control. For this purpose the kernel offers special boot parameter, and it is
the way the module is supposed to be used (as indicated in the readme file on
the donwloads page). From this point of view I see no reasons talking to Ubuntu
developers.
Moreover, I would consider missing acpi brightness interface as a bug on the
previous versions of the kernel (despite it does not work)
Original comment by SergeyK...@gmail.com
on 9 Jan 2011 at 3:45
Okay, I very much appreciate your comment!
So this means, we have to live with yet another addition to the
easy-slow-down-manager; that is preparing grub to tell the kernel the
correct interface to use?
I can agree, this is kind of logical, the first one to be expected
would be the acpi, to be available, in any case.
And still, there is one more argument; and that's an ancient
Unix-concept: override. Only if the user has no ~/.xyz.conf, will
/etc/xyz.conf be used as system-wide config, and applied to the user.
acpi is the default, and should be there. However, a user- or
admin-supplied backlight interface should take precedence. acpi should
only kick in as 'last resort'. grub ought not be overloaded with all
sorts of options defining precedences.
Thanks so much for this software anyway!
Uwe
Original comment by udip...@gmail.com
on 9 Jan 2011 at 4:12
Original comment by SergeyK...@gmail.com
on 29 Jul 2011 at 5:55
Original issue reported on code.google.com by
udip...@gmail.com
on 22 Dec 2010 at 11:57