Problems with ati driver on Tibook
John R. Dunning
jrd at jrd.org
Mon Dec 28 07:57:47 PST 2009
From: Alex Deucher <alexdeucher at gmail.com>
Date: Wed, 23 Dec 2009 09:33:05 -0500
Screen blanking turns off the crtcs and outputs and isn't directly
related to mode timing which was what the previous issue was. Macs
tend to have the backlight wired up strangely which is probably the
case for you. In some systems it's wired up to an external controller
or the levels are reversed.
Thanks again for your earlier help. I hope you'll bear with me as I
continue to try to sort this out.
It's apparent now that I have a number of issues going on
concurrently; at least some of them are due to KDE4 being, umm, a bit
raw. I'm trying to separate the KDE (and possibly X) issues from the
driver-related stuff, so I can pursue them all separately.
I'm still suspicious that I've got something in the driver, or
possibly a mismatch in how X is invoking it or something.
The backlight controls appear to be at least partially working. When
I test them manually with pbbuttonsd, they work fine. But when I tell
KDE to go to standby power on the display after a bit of idle time, it
does the thing that looks like wierd backlight problems, and
apparently changes the pixels showing (the screen image) to some
corrupted version of the desktop. It's a little hard to tell what
pixels are trying to be displayed due to the backlight.
Cycling in and out of that mode, I see the following sequence in the X
log:
in:
(II) RADEON(0): RADEONSaveScreen(2)
(II) RADEON(0): RADEONSaveScreen(0)
disable LVDS
disable LVDS
out:
(II) RADEON(0): RADEONSaveScreen(1)
enable LVDS
My surmise from observing the behaviour of the system is that
something (KDE? X?) is using multiple screens, and telling the driver
to switch them sometimes. That seems to coincide (sort of) with
somebody trying to adjust backlight power, with wierd results.
I've also tried changing the order of screen saver vs standby modes,
and had times when it would switch to some wierd screen (not the
saver) along about the time it should be going to standby. That one's
harder to reproduce; I suspect bugs in the statekeeping in KDE.
So I think the question is: what's the flow of control supposed to be,
and which bits of software are responsible for changing/managing
screens, and driving the backlight controls? If any parts of that
fall into the driver, I'd appreciate hints on debugging. Some are
likely in other areas, hints also appreciated on where to look.
Thanks again.
More information about the xorg-driver-ati
mailing list