Bug 49541 - [gma500] DPMS fails to turn off the display
Summary: [gma500] DPMS fails to turn off the display
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - Intel) (show other bugs)
Hardware: All Linux
: P3 normal
Assignee: Patrik Jakobsson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-10-25 20:35 UTC by Dainius Masiliūnas
Modified: 2016-03-23 18:28 UTC (History)
2 users (show)

See Also:
Kernel Version: 3.7.0-rc2
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Dainius Masiliūnas 2012-10-25 20:35:30 UTC
On a Fujitsu Stylistic Q550 tablet (Oak Trail), enabling DPMS results in a blanked screen, but the backlight still persists (the screen is "shining black").

I tried using both xset and vbetool, and both result in the same behaviour. Also, the same happens whether I attempt to turn off the screen while in a tty console, too. When the screen is blanked, there are no additional messages written into neither dmesg nor /var/log/messages.

Changing the backlight brightness though ACPI does work well, and the different brightness levels do show once the screen is blanked, but it does not turn off the panel completely, either.

On a related note, when attempting to put the device into sleep mode, something very similar happens - the screen stays on, but it is blanked. The device never reaches sleep, and the only way to do anything with it at that point is to force a reset.

Using Gentoo, kernel compiled directly from mainline git, X server 1.11.2.
Comment 1 Alan 2012-10-29 12:21:35 UTC
Unfortunately no code I know of documents how to do this on the Q550. In theory all the needed code is there and in oaktrail_lvds_set_power but this doesn't seem to work on Q550.
Comment 2 Dainius Masiliūnas 2012-10-29 17:29:59 UTC
Oh, that's too bad... Without this, it seems that it's also incapable of sleeping (tested all the quirks as well). And that makes it rather poor for mobile usage, unless something new is discovered about this issue...

Though I'm curious, where could the problem lie? The EFI, the hardware, something else entirely?
Comment 3 Jani Nikula 2013-04-30 10:46:25 UTC
Patrik takes care of gma500 now.
Comment 4 Patrik Jakobsson 2013-05-01 12:28:44 UTC
Hi Dainius,

Could you try the latest drm-next and see if it improves suspend/resume. pm-utils is now ignoring gma500 so we have a fair chance to do the right thing.

git://people.freedesktop.org/~airlied/linux drm-next

Thanks
Patrik
Comment 5 Dainius Masiliūnas 2013-05-01 18:32:11 UTC
Tried it, but the result is the same. If I echo either mem or freeze into /sys/power/state, the screen becomes blank, but it still shines, and it never goes into sleep.
Comment 6 Patrik Jakobsson 2013-05-06 11:32:11 UTC
Thanks for testing. Just to be sure, have you tested suspend without loading the gma500_gfx module? Are you certain the suspend issue is gma500 related?
Comment 7 Dainius Masiliūnas 2013-05-06 15:23:20 UTC
Yes. Without the module (and thus without X running), doing an echo "mem" > /sys/power/state results in the device entering sleep successfully.

Note You need to log in before you can comment on or make changes to this bug.