I have a ThinkPad X61 running Fedora 8 with kernel 2.6.23.9. This
includes thinkpad_acpi module version 0.16 and a video module with no
version number other than that of the kernel.
With both the thinkpad_acpi and video kernel modules loaded, I see three
representations of the display backlight under /sys/class/backlight:
acpi_video0
acpi_video1
thinkpad_screen
Attempting to set brightness levels using each of these gives the
following results:
acpi_video0/max_brightness is "100". Echoing values from 0-100 into
acpi_video0/brightness has no observable effect on either the integrated
LCD display or on an externally-connected VGA monitor. The value in
acpi_video0/actual_brightness does change, though at coarser steps.
acpi_video1/max_brightness is "100". Echoing values from 0-100 into
acpi_video1/brightness changes the brightness of the integrated LCD
display. The display goes through sixteen discrete brightness levels.
This seems correct, as the same laptop offers sixteen brightness levels
when running Windows Vista. However, it is slightly odd that values
from 0-19 inclusive have no effect on the display brightness. Values
20-24 set the backlight to its darkest available setting of 20; values
0-19 do nothing at all.
thinkpad_screen/max_brightness is "7". This seems incorrect, as the
same laptop offers sixteen brightness levels when running Windows Vista.
Echoing values 0-7 into thinkpad_screen/brightness changes the
brightness of the integrated LCD display. However, even at brightness
level 0 the display is still brighter than it can be. It appears that
this interface only steps the backlight through the brightest 8 of its
16 possible values. Also, after echoing 0 into
thinkpad_screen/brightness, acpi_video1/actual_brightness reports
brightness level 60 instead of the minimum level of 20 we can get by
using acpi_video1/brightness to set the level.
I'll also note that
<http://people.freedesktop.org/~hughsient/quirk/quirk-backlight-index.html>
states that "The video kernel acpi driver is buggy on quite a few IBM
and Lenovo laptops" and recommends unloading or blacklisting it entirely
to fix the problem of multiple laptop panel devices showing up in HAL.
To summarize, I'm seeing the following glitches here:
- three devices in /sys/class/backlight for one piece of hardware
- acpi_video0 appears to do nothing useful at all
- acpi_video1 works well, except that levels 0-19 are no-ops
- thinkpad_screen only exposes the brightest 8 of 16 levels
HAL is the main consumer of this information in my working environment,
and HAL has good facilities for working around quirky hardware. I'd
appreciate feedback on which of the above glitches should be considered
true thinkpad_acpi or video module bugs and which I should work around
in user-space by defining various HAL quirks.
The lack of an X61 row in the compatibility table at
<http://ibm-acpi.sourceforge.net/> suggests I'm in uncharted territory.
I'm happy to help fill that table in of there are specific tests you'd
like me to run. Or is this a more informal process of doing whatever it
is I think I should do to test each feature column? Let me know; I'd
like to help.
Regards,
Ben Liblit
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
ibm-acpi-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
