Open GoogleCodeExporter opened 9 years ago
Sorry, I just noticed that "slowing down" works fine. But it would be still
very good to have a brightness adjust control.
Original comment by whitequ...@gmail.com
on 20 Nov 2010 at 7:03
did you use kernel option 'acpi_backlight=vendor'? did you have something in
/sys/class/backlight/ and /sys/devices/virtual/backlight/?
Original comment by SergeyK...@gmail.com
on 27 Nov 2010 at 1:35
There is a /sys/class/backlight/samsung/ directory/symlink, and currently (with
acpi_backlight=vendor option) it works incorrectly: bl_power does not work at
all, and modifying brightness from 0 to 7 actually makes the screen a little
bit brighter or darker, somewhere in the middle of its possible range. (Not to
mention that max_brightness is 7. _Seven_, while I currently use about 20
levels with a handwritten script).
Also, I think that script can be interesting to you: there's a linear mapping
between button presses and brightness changes which turned out to be very
convenient and smooth. Here it is (set_brightness does the setpci hack and
get_brightness returns the value previously saved by get_brightness):
adjust_brightness() {
local value=$(get_brightness)
local dir=$1
local step=$(( value / 10 + 3))
if [ "$dir" == "up" ]; then
value=$(( value + step ))
elif [ "$dir" == "down" ]; then
value=$(( value - step ))
else
echo "Invalid direction $dir" >&2
exit 1
fi
set_brightness $value
}
Original comment by whitequ...@gmail.com
on 27 Nov 2010 at 7:54
could you try to load the module samsung_backlight with 'use_sabi=0 force=1'
options? this way the driver should go by the same setpci trick.
How many brightness levels do you have in Windows?
Original comment by SergeyK...@gmail.com
on 28 Nov 2010 at 8:18
With 'use_sabi=0 force=1' parameters it works just fine, with 255 of brightness
levels. 'bl_power' still do not work, but 'echo 0 >backlight' works.
Don't know. I didn't even booted that operating system. (Actually I tried, but
waiting half a hour for the _preinstalled_ OS to do something is frustrating.)
Original comment by whitequ...@gmail.com
on 1 Dec 2010 at 5:16
ok, seems that samsung has changed number of brightness levels, supported by
BIOS, for your laptop model. I'll add your model to the list in the next
revision and add some parameters to the module, so you could tune up number of
brightness levels
Original comment by SergeyK...@gmail.com
on 1 Dec 2010 at 10:05
The latest version should work with parameter usa_sabi=0;
otherwise, you could play with the options maxbright and levels (for example,
try maxbright=16 levels=16);
(bl_power not supported yet)
Original comment by SergeyK...@gmail.com
on 24 Jan 2011 at 6:23
Thanks, I'll check it in a few days. What about X integration? (IIRC, X just
uses kernel backlight interface, so this is about porting it to a specific
sysfs API)
Original comment by whitequ...@gmail.com
on 24 Jan 2011 at 6:35
X brightness control works properly, if the driver is installed correctly (see
boot args in readme)
Original comment by SergeyK...@gmail.com
on 29 Jan 2011 at 8:17
I've tested that. With use_sabi=1 (the default) it works in a very strange way;
while I'm not sure this is what really happens, I'd try to describe it as such:
it queues backlight level change requests somewhere, and instead of updating
with my changes fetches one element from that queue and executes it. For
example, I may write several different values to `brightness`, and each time
the actual level would change by 1, and then I would write several 1's or 0's
to bl_power, and it will suddenly change the level several times in a manner
looking unrelated to what I write.
It works with use_sabi=0, through also force=1 option is required. Sadly, the X
backlight control does not work:
$ grep -i backlight /var/log/Xorg.0.log
Kernel command line: BOOT_IMAGE=/vmlinuz-2.6.32-5-686
root=/dev/mapper/mannaz-root ro quiet splash acpi_backlight=vendor
(II) intel(0): found backlight control interface /sys/class/backlight/samsung
But nothing happens when I press backlight change keys.
Original comment by whitequ...@gmail.com
on 5 Feb 2011 at 12:31
Original issue reported on code.google.com by
whitequ...@gmail.com
on 19 Nov 2010 at 11:30