Re: Brightness not quite tracked on X60s, thinkpad-acpi-0.15-20070815_v2.6.22.2.patch

On Sat, Aug 25, 2007 at 02:08:14AM -0300, Henrique de Moraes Holschuh wrote:
> On Fri, 17 Aug 2007, Josh Mandel wrote:
> > is still fooled when I close the laptop's lid or run "xset dpms force
> > off".  In these cases, the screen brightness (once restored from the
> > off state) shoots up to 100%, but the thinkpad_acpi continues to
> > report the previous value.  I should mention that I'm running BIOS
> > 2.12 and the "video" module is loaded on my system.  (Is loading this
> > module a mistake?)
> 
> THAT might also be Lenovo doing a botched up job in the DSDT. If this is
> true, it simply means you can say bye-bye to brightness support on
> thinkpad-acpi for the x61/t61 and later, there is only so much breakage I
> will tolerate before I decide they do NOT care about it and will keep
> breaking things further and further, and just drop the support.

I think this is actually an issue with the -intel X driver, which pokes 
the chipset backlight registers directly and overrides the platform 
legacy backlight controller.

Still, I suspect that the best approach to handling the backlight stuff 
on new Thinkpads is just to leave it up to the video extension. Just 
check if _BCL, _BCM and _BQC are implemented and then refuse to create 
the backlight?

--

-- 
Matthew Garrett | mjg59@...

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.

Re: Brightness not quite tracked on X60s, thinkpad-acpi-0.15-20070815_v2.6.22.2.patch

On Mon, 01 Oct 2007, Matthew Garrett wrote:
> On Sat, Aug 25, 2007 at 02:08:14AM -0300, Henrique de Moraes Holschuh wrote:
> > On Fri, 17 Aug 2007, Josh Mandel wrote:
> > > is still fooled when I close the laptop's lid or run "xset dpms force
> > > off".  In these cases, the screen brightness (once restored from the
> > > off state) shoots up to 100%, but the thinkpad_acpi continues to
> > > report the previous value.  I should mention that I'm running BIOS
> > > 2.12 and the "video" module is loaded on my system.  (Is loading this
> > > module a mistake?)
> > 
> > THAT might also be Lenovo doing a botched up job in the DSDT. If this is
> > true, it simply means you can say bye-bye to brightness support on
> > thinkpad-acpi for the x61/t61 and later, there is only so much breakage I
> > will tolerate before I decide they do NOT care about it and will keep
> > breaking things further and further, and just drop the support.
> 
> I think this is actually an issue with the -intel X driver, which pokes 
> the chipset backlight registers directly and overrides the platform 
> legacy backlight controller.
> 
> Still, I suspect that the best approach to handling the backlight stuff 
> on new Thinkpads is just to leave it up to the video extension. Just 
> check if _BCL, _BCM and _BQC are implemented and then refuse to create 
> the backlight?

I will check for that.  AND (_BCL, _BCM, _BCQ, _BQC) or OR (_BCL, _BCM, _BCQ,
_BQC) ?  (BQC is there because Micosoft are a bunch of irresponsible **** and got it
wrong in some reference document).

--

-- 

Re: Brightness not quite tracked on X60s, thinkpad-acpi-0.15-20070815_v2.6.22.2.patch

On Mon, Oct 01, 2007 at 08:59:53AM -0300, Henrique de Moraes Holschuh wrote:

> I will check for that.  AND (_BCL, _BCM, _BCQ, _BQC) or OR (_BCL, _BCM, _BCQ,
> _BQC) ?  (BQC is there because Micosoft are a bunch of irresponsible **** and got it
> wrong in some reference document).

The ACPI video driver does AND, so I think you want

_BCL AND _BCM AND (_BCQ OR _BQC)

to check that the video driver will bind and create the backlight class 
device, and then refuse to do so in the Thinkpad driver.
--

-- 
Matthew Garrett | mjg59@...

-------------------------------------------------------------------------
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/

Thinkpad R61 8918-5QG LED control

Hello

I wonder if the LED control should work on this model. LED 1
is the orange part of the battery and LED 7 is sleep (moon).
All the other leds don't change when i try to set or unset them.

I'm using the latest 0.16-20070925 thinkpad/ibm acpi.

Controlling the wlan led would be nice, because it's always off
here unlike the bluetooth led which switches nicly when the
device gets enabled or disabled.

Christian

-------------------------------------------------------------------------
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/

thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

I have released version 0.17-20071002 of thinkpad-acpi through the
sourceforge.net release system.

Patches are available for 2.6.20, 2.6.21, 2.6.22 and 2.6.23-rc at:
http://sourceforge.net/project/showfiles.php?group_id=117042&package_id=230205

git users can get it directly from tags in:
git://repo.or.cz/linux-2.6/linux-acpi-2.6/ibm-acpi-2.6.git

This release needs testing, and it needs testing FAST.  It has two important
features that could make it for 2.6.24, if I get reports of how well they
work in less than a week.

* Support for 16-level brightness for Lenovo thinkpads
  (X61, T61 *needs* this.  X60 likely does so, as well.  T60 appears not to,
   but I welcome tests, it should auto-detect).

  Note that this may or may not fix issues with the brightness *hot keys*,
  but it *really* is supposed to fix all issues with the sysfs/procfs
  backlight/brightness interface on the ThinkPads with 16 levels.

* Revamped NVRAM polling support for hot keys.  I had to redo a lot of the
  code, so that it plugs nicely into hotkey_mask.

  I have received so far *NO* reports from Thinkpads that really benefit
  from NVRAM polling, such as the T23 and T30 and earlier.

--

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond

Re: thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

Henrique de Moraes Holschuh wrote:
> This release needs testing, and it needs testing FAST.  It has two important
> features that could make it for 2.6.24, if I get reports of how well they
> work in less than a week.
> 
> * Support for 16-level brightness for Lenovo thinkpads
>   (X61, T61 *needs* this.  X60 likely does so, as well.  T60 appears not to,
>    but I welcome tests, it should auto-detect).

This support works fine on my T60.  But I noticed a small problem: I tried

  echo 15 > /sys/class/backlight/thinkpad_screen/brightness

which generated EINVAL, as expected.  But afterwards all values
generated EINVAL, even reasonable ones such as 0 through 7.  EINVAL was
also generated when trying to set the value through procfs, either by
up/down or level N.  Prior to trying 15, the 0-7 values didn't signal
EINVAL errors.

The kern.log is filled with messages like

  Oct  3 00:21:50 ravna kernel: set_level status: 0

which originate in "drivers/acpi/video.c".  My interpretation is that
some variable in ACPI got set to a bad value and is now stuck there.

-------------------------------------------------------------------------
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/

Re: thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

On Wed, 03 Oct 2007, Chris Hanson wrote:
> Henrique de Moraes Holschuh wrote:
> > * Support for 16-level brightness for Lenovo thinkpads
> >   (X61, T61 *needs* this.  X60 likely does so, as well.  T60 appears not to,
> >    but I welcome tests, it should auto-detect).
> 
> This support works fine on my T60.  But I noticed a small problem: I tried

Thanks for testing, and for the report!

>   echo 15 > /sys/class/backlight/thinkpad_screen/brightness
> 
> which generated EINVAL, as expected.  But afterwards all values
> generated EINVAL, even reasonable ones such as 0 through 7.  EINVAL was
> also generated when trying to set the value through procfs, either by
> up/down or level N.  Prior to trying 15, the 0-7 values didn't signal
> EINVAL errors.

Can't reproduce this on 2.6.22 and a T43.  Given that the driver returns
EINVAL well before it even tries to poke the hardware, it should not make a
difference which thinkpad we try it on.  So, that leaves us with
thinkpad-acpi bugs, or a backlight class bug.  Which kernel version?

Also, just to be sure, cat
/sys/class/backlight/thinkpad_screen/max_brightness returns 7 in your T60,
correct?

> The kern.log is filled with messages like
> 
>   Oct  3 00:21:50 ravna kernel: set_level status: 0

Re: thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

Henrique de Moraes Holschuh wrote:
> Can't reproduce this on 2.6.22 and a T43.  Given that the driver returns
> EINVAL well before it even tries to poke the hardware, it should not make a
> difference which thinkpad we try it on.  So, that leaves us with
> thinkpad-acpi bugs, or a backlight class bug.  Which kernel version?

2.6.22.9 + numerous patches: whoopie patches; deferrable timers; round
jiffies; ALPM; HRT.

> Also, just to be sure, cat
> /sys/class/backlight/thinkpad_screen/max_brightness returns 7 in your T60,
> correct?

Correct.

> Should be impossible to happen as long as max_brightness is not 15 (not only
> because of the backlight class, but also because it means thinkpad-acpi is
> supposed to refuse to write such stuff to the hardware).  But a backlight
> variable being stuck is not at all difficult to imagine.  Let's try to track
> it down...

OK, what else do you need?

-------------------------------------------------------------------------
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/

Re: thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

On Wed, 03 Oct 2007, Chris Hanson wrote:
> > because of the backlight class, but also because it means thinkpad-acpi is
> > supposed to refuse to write such stuff to the hardware).  But a backlight

But it was not refusing to *read* it.  Try the attached patch.  Does it
solve the issue?  If it does, it means the T60 has one extra bit of data in
NVRAM that might be interesting to find out.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

commit 2207713b918c5144d3b355150bf8cd53b48b991d
Author: Henrique de Moraes Holschuh <hmh@...>
Date:   Wed Oct 3 09:04:46 2007 -0300

    ACPI: thinkpad-acpi: fix issues in brightness 16-levels support

    Fix some issues with the new 16-level brightness support.

    Signed-off-by: Henrique de Moraes Holschuh <hmh@...>

diff --git a/drivers/misc/thinkpad_acpi.c b/drivers/misc/thinkpad_acpi.c
index b0ef877..45956f5 100644
--- a/drivers/misc/thinkpad_acpi.c
+++ b/drivers/misc/thinkpad_acpi.c
 <at>  <at>  -3584,6 +3584,13  <at>  <at>  static int __init brightness_check_levels(void)

Re: [ltp] thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

On Wed, 03 Oct 2007, Jan Gutter wrote:
> On Wed, 2007-10-03 at 08:50 -0300, Henrique de Moraes Holschuh wrote:
> > Yes.  Please send me the output of dmidecode (edit out the UUID and serial
> > numbers) and of acpidump.  I need what is in the SSDTs and DSDT.
> 
> Attaching...

Please try the attached patch on top of 0.17-20071002.  Does it fix the
issue?

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

commit 1b81ea7ba1dc07398e951cfdeacaf6cc6ad716eb
Author: Henrique de Moraes Holschuh <hmh@...>
Date:   Wed Oct 3 10:03:32 2007 -0300

    ACPI: thinkpad-acpi: fix issues in brightness 16-levels support (v2)

    Fix some issues with the new 16-level brightness support.

    Signed-off-by: Henrique de Moraes Holschuh <hmh@...>

diff --git a/drivers/misc/thinkpad_acpi.c b/drivers/misc/thinkpad_acpi.c
index b0ef877..a4884aa 100644
--- a/drivers/misc/thinkpad_acpi.c