Hello, I noticed, that when I try to activate my VGA head (this is a notebook: RV350 [Mobility Radeon 9600 M10]) my system freezes. I did a git bisect and got this commit as the first bad commit: f3d2ec3a5ae61215c792018320158750e7aa937c The effect I see is: The server comes up with a picture on both heads (LVDS+VGA, clone mode). The I try to switch into extended desktop via: xrandr --output VGA-0 --right-of LVDS --mode 1400x1050 This worked as expected, no I get a frozen screen. Thanks in advance Matthias
Created attachment 13639 [details] xorg.conf The xorg.conf used
Created attachment 13640 [details] Log of a X session that froze (using git version)
Please attach your xorg logs from the working version as well.
Created attachment 13641 [details] [review] possible fix Does this patch help?
Created attachment 13642 [details] Working Xorg.log
No the patch does not help
Created attachment 13643 [details] [review] Working Xorg.log got the wrong log ....
Do any of the following options help? Option "DRI" "false" Option "AGPMode" "4" Option "BusType" "PCI"
Does dualhead work if you start with VGA disconnected and then connect it after X had started?
(In reply to comment #8) > Do any of the following options help? > > Option "DRI" "false" > Option "AGPMode" "4" > Option "BusType" "PCI" Nope - still freezes and also starting X without the monitor attached, then connecting the monitor and switching it into extended desktop mode fails.
Hey, I tried again git head - no change - I just freezes. My "stable" setup is now: A checkout of c9a0cee97ca69e8fe1e1937c7670fa903214cded and cherry-picking the PAL fix (2b1fae668ddabbc72e5fc31365302ea722174df1). This brings me a working dual head setup, that I can use with my external CRT or TV-Out (well apart from bug #13872). But the upcoming render acceleration make newer commits interesting.
I checked out head again and tried some more resolutions. And got surprising results: Working: - 1152x864 - 1280x800 - 1280x960 - 1440x900 - 1280x1024 - 1680x1050 - 1600x1024 Not-working: - 1400x1050 (was working prior to the commit mentioned in the inital report) I had also another computer available and could determine, that the lock up is system wide (at least the network system goes down to). So currently I run git head at 1860x1050(LVDS) and 1280x960(VGA) -- BUT I had to disable EXA and switch back to XAA, as the refresh behaviour with EXA was unusable. With XAA every works as before (apart from the 1400x1050 resolution for VGA).
What method are you using to program the new modes? What modeline are you using for 1400x1050?
I use xrandr - for example to switch to my current mode: xrandr --output VGA-0 --mode 1280x960 --right-of LVDS The modes I mentioned were tested without the position argument. I went through the refresh rates (bottom up) and 60, 70 and 74.8 worked, 85.3 freezes hard. I'll attach the verbose output of xrandr, which contains the mode information.
Created attachment 13837 [details] xrandr --verbose output
where are all of those modes coming from? They are not in your monitor's edid. Are you adding them manually with xrandr --newmode or adding them to your config as modelines? I don't see any mention of them in your logs. Do all the of the 1400x1050 modes fail or just some? 1400x1050 (0x51) 184.0MHz +HSync +VSync h: width 1400 start 1464 end 1656 total 1960 skew 0 clock 93.9KHz v: height 1050 start 1051 end 1054 total 1100 clock 85.3Hz 1400x1050 (0x52) 155.8MHz +HSync +VSync *current h: width 1400 start 1464 end 1784 total 1912 skew 0 clock 81.5KHz v: height 1050 start 1052 end 1064 total 1090 clock 74.8Hz 1400x1050 (0x53) 151.0MHz +HSync +VSync h: width 1400 start 1464 end 1656 total 1960 skew 0 clock 77.0KHz v: height 1050 start 1051 end 1054 total 1100 clock 70.0Hz 1400x1050 (0x54) 122.0MHz +HSync +VSync h: width 1400 start 1488 end 1640 total 1880 skew 0 clock 64.9KHz v: height 1050 start 1052 end 1064 total 1082 clock 60.0Hz
Testing showed, that only the 85.3 Hz mode locks while all other modes work in 1400x1050. I have no idea where the modes come from, but I just started X and an xterm inside it - xrandr gave the same output as the attached one.
(In reply to comment #17) > Testing showed, that only the 85.3 Hz mode locks while all other modes work in > 1400x1050. I have no idea where the modes come from, but I just started X and > an xterm inside it - xrandr gave the same output as the attached one. > I wonder if your distro is shipping some xserver patch that adds the modes? Anyway, you should be able to work around this by adding a modeline and preferred mode to your xorg config to pick one of the other 1400x1050 modes.
Is this still an issue with a newer driver (6.9.0 or newer)?
Excuse me, but the Notebook taht showed this problem broke approx. five months ago and since then I use an intel based notebook. So I can't test.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.