Dell Latitude C600 backlight remains on in S3 suspend
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
HAL |
Fix Released
|
Medium
|
|||
hal-info (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
Older Dell Latitude C600 upgraded from 8.10 to 9.04, LCD backlight now remains on in S3 suspend. Specifically, suspend/resume overall works as expected, the panel is shut off, but the backlight remains lit (even with the lid closed). Reproducible every time, and a clear regression from 8.10 and earlier. Normal DPMS functionality and 'xset dpms force off' work as expected, though in either case, the backlight and panel will be powered back on during the suspend process, thus the only available workaround for this is to avoid suspend and power the machine off.
The device identifies itself as 'ATI Technologies Inc Rage Mobility M3 AGP 2x (rev 02)', Xorg uses the R128 driver for this chip. There are no unusual messages in dmesg or the Xorg log and no other functionality seems to be affected. Oddly enough, the backlight will actually shut off momentarily on a resume, and is then powered back on just prior to the mode switch back to X.
tags: | added: patch-forwarded-upstream |
Changed in hal: | |
importance: | Unknown → Medium |
status: | Unknown → Fix Released |
tags: |
added: patch-accepted-upstream removed: patch-forwarded-upstream |
Changed in hal: | |
importance: | Medium → Unknown |
Changed in hal: | |
importance: | Unknown → Medium |
Discovered that correct behavior occurs when using 'pm-suspend --quirk- dpms-suspend' , and also that 'radeontool light off' has no effect. Not sure whether that did under 8.10 or earlier, but the only listed quirk in /usr/share/ hal/fdi/ information/ 10freedesktop/ 20-video- quirk-pm- dell.fdi for the C600 is 'radeon_off' (which has not changed since prior releases). Adding 'dpms_suspend' restores correct backlight behavior, patch attached. Also verified that removing the 'radeon_off' quirk has no noticable effect.
Still unsure what caused the regression, although my current suspicion is that something (kernel?) broke whatever magic is behind the 'radeon_off' quirk. Will post findings to upstream mailing list and see what happens.