http://bugzilla.novell.com/show_bug.cgi?id=578578 http://bugzilla.novell.com/show_bug.cgi?id=578578#c Jeff Mahoney <jeffm@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|kernel-maintainers@forge.pr |trenn@novell.com |ovo.novell.com | -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=578578 http://bugzilla.novell.com/show_bug.cgi?id=578578#c2 Matthias Hanisch <mjh@kabelmail.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|mjh@kabelmail.de | --- Comment #2 from Matthias Hanisch <mjh@kabelmail.de> 2010-08-05 22:30:21 UTC --- Meanwhile I upgraded to 11.3. It shows the same behavior. The brightness control keys (Fn-LEFT/RIGHT) cause a value increment/decrement in /sys/class/backlight/acpi_video0/brightness. But, the it has no effect on the (physical) screen brightness. Current workaround: monitor the value above and set the brightness directly with setpci -s 00:02.0 F4.B=$value where $value is between 00 and FF, based on the brightness values above from 0 to 9. Graphics card: Intel GMA 4500 X Server: Hmmm... rpm -qf /usr/bin/Xorg gives xorg-x11-server-7.5_1.8.0-9.4.x86_64 what else do you want to know? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=578578 http://bugzilla.novell.com/show_bug.cgi?id=578578#c3 Thomas Renninger <trenn@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P4 - Low Status|NEW |ASSIGNED CC| |trenn@novell.com Severity|Normal |Minor --- Comment #3 from Thomas Renninger <trenn@novell.com> 2010-08-06 15:40:46 UTC --- Yep, it was already discussed that a PCI config based backlight interface should be added to the Intel graphics kernel driver as fall back if no other (video.ko or one of the native laptop drivers) cannot serve it. But nobody did the work yet. If you see an update somewhere mainline, I'd be glad to backport it. I'll ping here if I see news on a list, please do the same. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=578578 http://bugzilla.novell.com/show_bug.cgi?id=578578#c4 --- Comment #4 from Matthias Hanisch <mjh@kabelmail.de> 2010-08-06 21:25:11 UTC --- Yes, will do. The solution/workaround I am currently using, is described in detail at http://www.opensuse-forum.de/bildschirmhelligkeit-einstellen-extensa-5635z-d... Just to complete the current status. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c6 Michael Chang <mchang@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jlee@novell.com --- Comment #6 from Michael Chang <mchang@novell.com> 2011-08-24 05:53:16 UTC --- Hi Matthias, Yes. This is related with "Intel backlight combination mode", that is the backlight brightness level can be changed via LPBC (that is the PCI configuration register your workaround uses) OR i915's BLC_PWM_CTL register. As far as I know, we have to access both registers at the same time to guarantee a reasonable dynamic backlight range (The LPBC is acts a multiplier of 255 with BLC_PWM_CTL value). This combination mode is used on older Intel graphics and yours is fall into this category (GMA4500 is used in Montevina platform, we ever came across similar issue on this hardware). For modern Intel graphics, the LPBC register was obsoleted. I also CC Joey Lee as he is maintainer of ACER acpi/wmi driver. This might also be an ACPI issue as these function keys would trigger SCI events (Q event) to execute corresponding ACPI method (Q method) an change the backlight (via the opregion or whatever means) in it. First, it would be good if you could help to provide dmesg while your are changing the backlight, whith following kernel command-line parameter. drm.debug=0x02 I would suggest a step, for example ..like $ dmesg -c $ xrandr --output LVDS1 --set BACKLIGHT 6 $ xrandr --output LVDS1 --set BACKLIGHT 9 $ dmesg > /tmp/xrandr-backlight.log Thanks. :) Thanks. :) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c8 --- Comment #8 from Michael Chang <mchang@novell.com> 2011-08-25 03:12:23 UTC --- Thanks. To me that nothing printed is reasonable as this implies that drm driver didn't perform any real brightness setting in respond of attempting to set it. This is weird but possible, there's breakage of something in the middle of the process in prior of driver was asked to set a value. :) I think we can rule out combination mode as it didn't take place and focus on the ACPI issue. From your log that's Intel opRegion bios relies on in order to communicate with vga driver (a memory region when you write some value, an interrupt is trigger automatically to notice vga driver to fetch the value and do something (like change brightness) accordingly). Would you please help to dump the ACPI for debugging? $ acpidump > acpidump.hex And attach the acpidump.hex .. Thanks a lot. :) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c10 Joey Lee <jlee@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |mjh@kabelmail.de --- Comment #10 from Joey Lee <jlee@suse.com> 2011-08-31 04:16:17 UTC --- Could you please try to add acpi_osi="Linux" kernel parameter to /boot/grub/menu.lst? like the following: kernel /boot/vmlinuz-2.6.32.29-0.3-pae root=/dev/disk/by-id/ata-WDC_WD1200BEVS-08UST0_WD-WXC608060890-part2 resume=/dev/disk/by-id/ata-WDC_WD1200BEVS-08UST0_WD-WXC608060890-part1 splash=silent showopts vga=0x317 acpi_osi="Linux" <==== add here Then, reboot your machine try to echo a number to acpi video brightness node: echo 4 > /sys/class/backlight/acpi_video0/brightness Looks it works or not. If yes, that means your machine have incomplete _BCM function support by EC firmware. Does your machine's brightness control works fine before ? or ALWAYS fail ? What is your original OS when you bought in this machine? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c12 --- Comment #12 from Joey Lee <jlee@suse.com> 2011-09-14 10:19:25 UTC --- (In reply to comment #11)

First, sorry for the long delay. Yes, the option acpi_osi="Linux" helps in both kernels I am using. This is 2.6.37.6-0.5-desktop and 2.6.39-32-desktop!

In detail: - the "echo" command above changed the brightness - the shortcut keys for brightness change work - the power management changed the brightness when I disconnected from power

Great!

I am quite sure, that I tried also with ACPI options (don't remember which one) when I faced that problem first (around 2009/2010), but at that time nothing helped.

To your questions: Brightness control never worked on that laptop. But I directly installed openSUSE on it. The laptop was shipped without a "real" OS. Linpus Linux was installed by default, which I deleted immediately after one test startup.

Thank you very much for that information! If this is the suggested workaround, this would be absolutely ok for me. Everything works as expected until now. I'll leave this option enabled and report if it causes any side effects.

If you need additional info for more investigations or to improve the ACPI driver in handling poor firmware, do not hesitate to ask.

Again, thank you for your support.

On some Acer machine, their BIOS ship OS with acpi_osi="Linux", but it's not a good idea because nobody tested the "Linux path" in BIOS, even Acer's ODM also didn't do the job. That's why kernel acpi team choice suggest avoid to use acpi_osi=Linux. If the osi Linux parameter works fine to you, like you said, we should test the machine after use this parameter, e.g. suspend to RAM or suspend to Disk. If everything works fine, then that's good! But, we will not to add this parameter to kernel for the default. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c14 --- Comment #14 from Michael Chang <mchang@suse.com> 2011-09-14 10:47:03 UTC --- (In reply to comment #13)

Again, this is a machine that have EC firmware problem to handle the brightness level. If the EC firmware works fine, then don't need add acpi_osi=Linux:

To verify it we might need to install Windows (better be Windows 7) on it and see if brightness works. But It's not worth for doing it IMO. :) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c16 Thomas Renninger <trenn@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WONTFIX --- Comment #16 from Thomas Renninger <trenn@novell.com> 2011-09-15 10:12:46 UTC --- Hm, on kernel bugzilla, there is the additional "resolved documented" state for bugs which would be nice for this one as well. If I understood Joey Lee correctly there is a Windows brightness driver which more and more OEMs are using to which we are not 100% compatible. As there fortunately is an easy workaround, I resolve the bug. If there is a more general problem with brightness interfaces (on this one without acpi_osi=Linux on others this might not work), this sounds as it should get discussed on the acpi or x86_platform mailing list or in a bugzilla.kernel.org bug with the goal to get things work mainline. I could imagine there are similar bugs there where you could attach your acpidump output as another example and offer testing if you want to help. Maybe it helps that someone finds what's missing on Linux and can get fixed/tested if the kernel guy does not have such HW. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=578578 https://bugzilla.novell.com/show_bug.cgi?id=578578#c18 --- Comment #18 from Joey Lee <jlee@suse.com> 2011-09-15 12:15:51 UTC --- (In reply to comment #14)

(In reply to comment #13)

...

Again, this is a machine that have EC firmware problem to handle the brightness level. If the EC firmware works fine, then don't need add acpi_osi=Linux:

To verify it we might need to install Windows (better be Windows 7) on it and see if brightness works. But It's not worth for doing it IMO. :)

Install Windows might not be work, because we need Acer's windows 7 preload image, not just install a standard windows 7. At least download the Windows 7 video driver from acer website and install it. Then compare the brightness control function before acer's driver installed and after. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.