On Saturday 04 February 2012 20:19:16 Joey Lee wrote:
Add Cc. to dri mail
Hi Pali,
於 五，2012-02-03 於 16:24 +0100，Pali Rohár 提到：
On Friday 20 January 2012 17:55:57 Pali Rohár wrote:
On Friday 20 January 2012 11:28:59 joeyli wrote:
於 五，2012-01-20 於 11:12 +0800，joeyli 提到：
Hi Pali,
Sorry for I am late reply you.
於 二，2012-01-17 於 19:10 +0100，Pali Rohár 提到：
On Wednesday 21 December 2011 12:45:07 Pali Rohár wrote: > Hello, > > I tried to boot with all 3 strings in acpi_os_name, but > nothing was > changed. I'm attaching dmesg outputs, one from BIOS > mode, one from > UEFI mode. Both are with "Microsoft Windows NT".
Did you looked at my logs?
I checked your dmesg log, found the acpi_os_name kernel parameter is :
[ 0.000000] Command line: BOOT_IMAGE=/vmlinuz-3.2.0-4-generic root=UUID=4ec6272a-8551-40f3-a1b1-3e10984f3e69 ro pcie_aspm=force acpi.debug_level=0x2 acpi.debug_layer=0xFFFFFFFF "acpi_os_name=Microsoft Windows NT" splash vt.handoff=7
We still need by-pass the os name check in DSDT when test function key,> > > > please help to feed: acpi_os_name="Windows 2009"
And, looks like the acpi debug level not enough, please kindly change acpi> > > > debug parameter to: acpi.debug_level=0xF acpi.debug_layer=0xFFFFFFFF log_buf_len=5M
summary: acpi.debug_level=0xF acpi.debug_layer=0xFFFFFFFF log_buf_len=5M acpi_os_name="Windows 2009"
Forgot remind, please remember press brightness function key a couple of times before you dump the dmesg and messages log.
Thanks a lot! Joey Lee
Hello,
there was no acpi log, so I recompiled ubuntu kernel with CONFIG_ACPI_DEBUG=y and CONFIG_ACPI_DEBUG_FUNC_TRACE. Then I started kernel with your params.
Now I'm attaching very long debug output. I belive it will be now usefull.
Pressing brightness keys did not show anything in log.
After writing 0 to /sys/class/backlight/acpi_video0/brightness in log appear: [ 57.675070] [ACPI Debug] Integer 0x000000000000000B ...
And after writing 10: [ 99.865208] [ACPI Debug] Integer 0x0000000000000048 [ 99.865295] evmisc-0120 [4294967289] ev_queue_notify_reques: Dispatching Notify on [DGFX] Node ffff880136246c80 Value 0xD0 (**Device Specific**) [ 99.865350] video-1474 [4294967289] video_bus_notify> > : Unsupported event [0xd0]
Do you need more logs? Or is this enought?
Yes, as you point out, this is a doubt for video bus received 0xD0 event but nobody take care it.
Traced dsdt of EliteBook 8460p from you, the _BCM like this:
Method (_BCM, 1, Serialized) /* ATI _BCM, perlog, run this _BCM */ { Store (_SB.BCM (Arg0), Local0) /* set next level, normally return 0x1 if XP sp2 or later */ If (Local0) /* if XP sp2 or later */ { Store (BRID, Local1) If (LEqual (SBRV (), 0x00)) /* normally SBRV return 1, will not emit SMI */ { _SB.SSMI (0xEA74, 0x04, Local1, 0x00, 0x00) }
Signal (\_SB.BEVT) /* emit BEVT event, HKFR(HotkeyFunctionResponse) waiting it, but why? */ } }
_BCM call SBRV to setup brightness value to variable ABRI:
Method (SBRV, 0, Serialized) /* call by ATI _BCM */ { Store (_SB.SBRC (), ABRI) /* SBRC() return the brightness value, why store to ABRI? only used in PEGP.DGFX.AFN2 */ Or (PSBR, 0x80, PSBR) Notify (^, 0xD0) /* notify method's parent: PEGP.DGFX by 0xD0 */ Return (0x01) /* normally return 1 */ }The PEGP.DGFX acpi device was binding to acpi/video driver, the above ASL code emit a 0xD0 bus event to video.c but cann't process it. Even we add a new bus event in video.c and generate a acpi event, there still need another acpi driver should take care it.
I thought this acpi event might need take care by radeon drm, but I am not good for radeon, need more help.
Per your acpi debug log, the brightness value was changed normally when you access _BCM:
83133 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.674669] ACPI: Execute Method [_SB_.PCI0.PEGP.DGFX.LCD_._BCM] (Node ffff8 801362472a8) # start test _BCM manually 83134 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.674736] exregion-0199 [01] ex_system_memory_space: System-Memory (width 8 ) R/W 0 Address=00000000BF7ACE1C 83135 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.674748] exregion-0199 [02] ex_system_memory_space: System-Memory (width 8 ) R/W 1 Address=00000000BF7ACE1C ... 83179 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.675070] [ACPI Debug] Integer 0x000000000000000B # brightness value is 11 83180 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.675090] evmisc-0120 [4294967289] ev_queue_notify_reques: Dispatching No tify on [DGFX] Node ffff880136246c80 Value 0xD0 (**Device Specific**) 83181 Jan 20 17:17:51 Pali-EliteBook kernel: [ 57.675099] video-1474 [4294967288] video_bus_notify : Unsupported event [0xd0] ... 83197 Jan 20 17:18:33 Pali-EliteBook kernel: [ 99.863593] ACPI: Execute Method [_SB_.PCI0.PEGP.DGFX.LCD_._BCM] (Node ffff8 801362472a8) 83198 Jan 20 17:18:33 Pali-EliteBook kernel: [ 99.863817] exregion-0199 [01] ex_system_memory_space: System-Memory (width 8 ) R/W 0 Address=00000000BF7ACE1C 83243 Jan 20 17:18:33 Pali-EliteBook kernel: [ 99.865208] [ACPI Debug] Integer 0x0000000000000048 # brightness value is 72 83244 Jan 20 17:18:33 Pali-EliteBook kernel: [ 99.865295] evmisc-0120 [4294967289] ev_queue_notify_reques: Dispatching Notify on [DGFX] Node ffff880136246c80 Value 0xD0 (**Device Specific**) 83245 Jan 20 17:18:33 Pali-EliteBook kernel: [ 99.865350] video-1474 [4294967289] video_bus_notify : Unsupported event [0xd0]
The brightness values are 11(0x0B) and 72(0x48), that means the BCM method return a good value and set to ABRI, ABRI waiting the guy who take the 0xD0 event to read then change brightness.
Another doubt is the latest statement in _BCM, it emit a BEVT event:
Signal (_SB.BEVT)
Only HKFR(HotkeyFunctionResponse) method is waiting this event, it should related to how the HP implement brightness function key on Windows through wmi.
Of course this issue really close related to video driver, even more, we might need to know hp wmi for how to implement on Windows.
Unfortunately, sorry for I don't have any solution to you, now, I will continue to trace and find any support from other experts.
Thanks a lot! Joey Lee
Hi!
now I found that ALS button which enable/disable ambient light sensor working with linux and ALS can also decrease/increase brightness.
ALS can be enabled or disabled via WMI (acpi) on linux too. I found function which enable ALS in linux kernel. See source code of function set_als in hp-wmi.c: http://tomoyo.sourceforge.jp/cgi- bin/lxr/source/drivers/platform/x86/hp-wmi.c#L443
This function can enable ALS which can decrease brightness (but only to specific one level) via WMI. And WMI is ACPI extension, so maybe it is really possible to change brightness via acpi without DRI support.
But I do not WMI, ACPI and DSDT code. Can you try to decode what is called in acpi when that set_als function in hp-wmi is called? I think here in acpi can be some magic which can manipulate with display brightness. At least ACPI must call somewhat to change brightness...
Note: Enabling ALS on my notebook change brightness to some specific level and disabling ALS will revert brightness back. From userspace ALS can be enabled/disabled via this sysfs entry: /sys/devices/platform/hp-wmi/als
於 二，2012-03-20 於 23:03 +0100，Pali Rohár 提到：
Another doubt is the latest statement in _BCM, it emit a BEVT event:
Signal (_SB.BEVT)
Only HKFR(HotkeyFunctionResponse) method is waiting this event, it should related to how the HP implement brightness function key on Windows through wmi.
Of course this issue really close related to video driver, even more, we might need to know hp wmi for how to implement on Windows.
Unfortunately, sorry for I don't have any solution to you, now, I will continue to trace and find any support from other experts.
Thanks a lot! Joey Lee
Hi!
now I found that ALS button which enable/disable ambient light sensor working with linux and ALS can also decrease/increase brightness.
ALS can be enabled or disabled via WMI (acpi) on linux too. I found function which enable ALS in linux kernel. See source code of function set_als in hp-wmi.c: http://tomoyo.sourceforge.jp/cgi- bin/lxr/source/drivers/platform/x86/hp-wmi.c#L443
This function can enable ALS which can decrease brightness (but only to specific one level) via WMI. And WMI is ACPI extension, so maybe it is really possible to change brightness via acpi without DRI support.
But I do not WMI, ACPI and DSDT code. Can you try to decode what is called in acpi when that set_als function in hp-wmi is called? I think here in acpi can be some magic which can manipulate with display brightness. At least ACPI must call somewhat to change brightness...
Note: Enabling ALS on my notebook change brightness to some specific level and disabling ALS will revert brightness back. From userspace ALS can be enabled/disabled via this sysfs entry: /sys/devices/platform/hp-wmi/als
-- Pali Rohár pali.rohar@gmail.com
Traced the your dsdt from 8460p.
Let's see "set ALS" first, when set ALS enable, BIOS didn't _eat_ any input value for brightness level:
Method (WHCM, 2, NotSerialized) { CreateDWordField (Arg1, 0x00, SNIN) /* signature */ CreateDWordField (Arg1, 0x04, COMD) /* command = write ? 0x2 : 0x1 */ CreateDWordField (Arg1, 0x08, CMTP) /* commandtype */ CreateDWordField (Arg1, 0x0C, DASI) Store ("HandleWMICommand Enter", Debug) ... If (LEqual (COMD, 0x02)) /* command write */ { Store ("write BIOS command", Debug) .... If (LEqual (CMTP, 0x03)) /* write HPWMI_ALS_QUERY 0x3 */ { Store (^WSAL (DDWD), Local2) /* call ^WSAL and feed DDWD */ Store (0x00, RTCD) }
Method (WSAL, 1, NotSerialized) /* write HPWMI_ALS_QUERY 0x3 */ { If (LEqual (PRDT, 0x03)) { Return (Package (0x02) { 0x04, 0x00 }) }
_SB.SSMI (0xEA75, 0x02, 0x03, 0x574D4953, 0x00) /* emit SMI, BIOS code take job but didn't feed Arg0 to BIOS */ Return (WFDA ()) }
Everything handle by BIOS function after emit SMI event, and it didn't give BIOS any input value of brightness. So, didn't find any way to change brightness through set ALS.
Then, I found another interesting command for 'write brightness', hp-wmi doesn't have implementation for this command:
If (LEqual (CMTP, 0x06)) /* set brightness? */ { Store ("write Brightness", Debug) Store (^SBBC (DDWD), Local2) /* call DDWD */ Store (Local2, Debug) Store (0x00, RTCD) }
Method (SBBC, 1, NotSerialized) /* write brighenss? didn't do anything? */ { Return (Package (0x02) /* just only return a package */ { 0x04, 0x00 }) }
Unfortunately, the 0x06 command also doesn't do anything for change brightness on your machine.
Sorry! Didn't find any way can change brightness on your machine unless we have drm driver support ATI acpi events.
Thanks a lot! Joey Lee
Hello, I have some good news. Radeon patches from this post http://lists.freedesktop.org/archives/dri-devel/2012-July/025535.html export brightness file /sys/class/backlight/radeon_bl/brightness And finally with these patches I'm able to change brightness on my EliteBook.
On Sat, Jul 28, 2012 at 4:47 PM, Pali Rohár pali.rohar@gmail.com wrote:
Hello, I have some good news. Radeon patches from this post http://lists.freedesktop.org/archives/dri-devel/2012-July/025535.html export brightness file /sys/class/backlight/radeon_bl/brightness And finally with these patches I'm able to change brightness on my EliteBook.
Wow, it's great news :) I was reading yesterday info about ACPI patches from AMD on Phronix and thinking will it help or no, and now I have an answer :)
Regards Artur
於 一，2012-07-30 於 10:17 +0200，Artur Flinta 提到：
On Sat, Jul 28, 2012 at 4:47 PM, Pali Rohár pali.rohar@gmail.com wrote:
Hello, I have some good news. Radeon patches from this post http://lists.freedesktop.org/archives/dri-devel/2012-July/025535.html export brightness file /sys/class/backlight/radeon_bl/brightness And finally with these patches I'm able to change brightness on my EliteBook.
Wow, it's great news :) I was reading yesterday info about ACPI patches from AMD on Phronix and thinking will it help or no, and now I have an answer :)
Regards Artur
Yes, thanks for Alex's 3 patches, those patches also works to me on my machine for backlight control.
Thanks Joey LEe
On Sat, Jul 28, 2012 at 4:47 PM, Pali Rohár pali.rohar@gmail.com wrote:
Hello, I have some good news. Radeon patches from this post http://lists.freedesktop.org/archives/dri-devel/2012-July/025535.html export brightness file /sys/class/backlight/radeon_bl/brightness And finally with these patches I'm able to change brightness on my EliteBook.
Nice, in your older mail I read:
The PEGP.DGFX acpi device was binding to acpi/video driver, the above ASL code emit a 0xD0 bus event to video.c but cann't process it. Even we add a new bus event in video.c and generate a acpi event, there still need another acpi driver should take care it.
That event (0xd0) is probably a custom event announced through ATIF. I'm working on the code to deal with it, the latest patches are here:
http://lists.freedesktop.org/archives/dri-devel/2012-July/025646.html
You first need this one with the ACPI header: http://lists.freedesktop.org/archives/dri-devel/2012-July/025517.html
and the other 3 that you've already tested. With my patches on top you should be able to change the brightness using the hotkeys.
I'd be grateful if you could test them :-)
Luca
dri-devel@lists.freedesktop.org
-
Artur Flinta -
joeyli -
Luca Tettamanti -
Pali Rohár