Hello. This morning I updated to xf86-video-ati 6.7.196 after having used xf86-video-ati 6.7.195 for a month without any problem. In the few hours since the upgrade I had a strange problem: for two times while I was normally using xorg the screen blinked, the status led of my monitor went from green (normal) to yellow (changing mode) and then to green again while the normal output was replaced by a grey pattern (picture taken with digital camera attached in the next comments). The mouse keeps working, the grey pattern reacted to changes (ie. activating compiz effects with the keyboard shortcuts), switching to console and then back to X fixed the problem in every occasion. The second time I actually checked the refresh rate with the OSD menu: it changed from the normal "H 68.76 KHz V 85.22 Hz" to "H 67.22 KHz V 83.11 Hz". A screenshot taken with the PrintScreen key showed the correct image. I'm using an ATI Radeon X550 (RV370 PCI-E), more info in xorg.conf and Xorg.0.log attached.
Created attachment 12522 [details] Xorg.0.log
Created attachment 12523 [details] xorg.conf
Created attachment 12524 [details] Grey pattern.
Created attachment 12525 [details] xrandr-output.txt I reverted to xf86-video-ati 6.7.185 and noticed another change which could or could not be significant: xf86-video-ati 6.7.195: xrandr -q Screen 0: minimum 320 x 200, current 1024 x 768, maximum 2048 x 1200 VGA-0 connected 1024x768+0+0 (normal left inverted right x axis y axis) 310mm x 230mm 1024x768 85.0*+ 84.9 75.1 70.1 60.0 1280x1024 59.9 800x600 84.9 72.2 75.0 60.3 56.2 640x480 75.0 72.8 60.0 720x400 70.1 DVI-0 disconnected (normal left inverted right x axis y axis) S-video disconnected (normal left inverted right x axis y axis) xf86-video-ati 6.7.196: xrandr -q Screen 0: minimum 320 x 200, current 1024 x 768, maximum 2048 x 1200 VGA-0 connected 1024x768+0+0 (normal left inverted right x axis y axis) 310mm x 230mm 1024x768 85.0*+ 84.9 75.1 70.1 60.0 1280x1024 59.9 800x600 84.9 72.2 75.0 60.3 56.2 640x480 75.0 72.8 60.0 720x400 70.1 DVI-0 unknown connection 1024x768+0+0 (normal left inverted right x axis y axis) 0mm x 0mm 1024x768 60.0* 800x600 60.3 640x480 59.9 S-video disconnected (normal left inverted right x axis y axis) I'm not using the DVI output of the card, there's nothing connected. Attached the --verbose output.
(In reply to comment #4) > Created an attachment (id=12525) [details] > xrandr-output.txt > > I reverted to xf86-video-ati 6.7.185 and noticed another change which could or > could not be significant: We default the tv dac to status unknown if there is no monitor attached as we cannot reliably do load detection on it yet. You can turn of the output by default by disabling it in your config (http://www.intellinuxgraphics.com/dualhead.html) or issuing the following randr command: xrandr --output DVI-0 --off does turning that output off fix the problem?
*** Bug 13229 has been marked as a duplicate of this bug. ***
Created attachment 12561 [details] new xorg.conf (In reply to comment #5) > > We default the tv dac to status unknown if there is no monitor attached as we > cannot reliably do load detection on it yet. You can turn of the output by > default by disabling it in your config > (http://www.intellinuxgraphics.com/dualhead.html) or issuing the following > randr command: > xrandr --output DVI-0 --off > does turning that output off fix the problem? > I ran a couple of hours without problems after having disabled the DVI output with xrandr, but I'm not sure if it really fixed the problem or if it was just luck, because I can't reliably trigger it. When I tried to modify the xorg.conf file to disable the DVI output X couldn't even start: when I added the Option "Ignore" in any "Monitor" section (DVI, VGA, S-video) it failed reporting "Screen(s) found, but none have a usable configuration." just after having processed that part of the file. Maybe there's something wrong in my new xorg.conf, attached; I'd be grateful if someone could point me the error.
(In reply to comment #7) > When I tried to modify the xorg.conf file to disable the DVI output X couldn't > even start: when I added the Option "Ignore" in any "Monitor" section (DVI, > VGA, S-video) it failed reporting "Screen(s) found, but none have a usable > configuration." just after having processed that part of the file. > Maybe there's something wrong in my new xorg.conf, attached; I'd be grateful if > someone could point me the error. > If you actually want to ignore the output, you need to set it to true: Option "Ignore" "true" of you just want to default the output to the off state: Option "Enable" "false"
(In reply to comment #8) > (In reply to comment #7) > > > When I tried to modify the xorg.conf file to disable the DVI output X couldn't > > even start: when I added the Option "Ignore" in any "Monitor" section (DVI, > > VGA, S-video) it failed reporting "Screen(s) found, but none have a usable > > configuration." just after having processed that part of the file. > > Maybe there's something wrong in my new xorg.conf, attached; I'd be grateful if > > someone could point me the error. > > > > If you actually want to ignore the output, you need to set it to true: > Option "Ignore" "true" > > of you just want to default the output to the off state: > Option "Enable" "false" > That's the problem: when I add Option "Ignore" "True" to a "Monitor" section, DVI or TV-out, X doesn't start anymore. By the way, I had again my problem this morning after 30s from boot, while I was starting firefox, before I could launch xrandr in a console. I recompiled the driver after having reverted the relevant commit (5db3afaa1fdb69d382ac769ef40191a4b964d28e) and another one as well (bug 13244) and so far everything's good, but I've been using the new driver for just a couple of hours.
(In reply to comment #9) > That's the problem: when I add Option "Ignore" "True" to a "Monitor" section, > DVI or TV-out, X doesn't start anymore. Can you attach the log in the failed case?
Created attachment 12573 [details] Xorg.0.log (In reply to comment #10) > (In reply to comment #9) > > > That's the problem: when I add Option "Ignore" "True" to a "Monitor" section, > > DVI or TV-out, X doesn't start anymore. > > > Can you attach the log in the failed case? > Here it is.
Ok, I've been using 6.7.196 with commit 5db3afaa1fdb69d382ac769ef40191a4b964d28e reverted for a few days and the bug didn't happened anymore, so it was most probably related to that change. I don't know exactly why it was introduced but if it breaks the standard one-monitor-connected-to-the-vga-port setup it doesn't seem too much viable.
(In reply to comment #12) > Ok, I've been using 6.7.196 with commit > 5db3afaa1fdb69d382ac769ef40191a4b964d28e reverted for a few days and the bug > didn't happened anymore, so it was most probably related to that change. > I don't know exactly why it was introduced but if it breaks the standard > one-monitor-connected-to-the-vga-port setup it doesn't seem too much viable. > Can you try again with ati git master?
Using git master now, no problems in some ~2½ hours, but I'm noticing a certain "loss of smoothness" of the compiz effects, specially the ones who involve all the screen (cube rotations, etc.). It could be just a sensation, however. I'm going to report if the problem happens again or in a couple days.
(In reply to comment #14) > Using git master now, no problems in some ~2½ hours, but I'm noticing a > certain "loss of smoothness" of the compiz effects, specially the ones who > involve all the screen (cube rotations, etc.). It could be just a sensation, > however. > > I'm going to report if the problem happens again or in a couple days. > I spoke too soon, the problem just came back.
Can you try and use git-bisect to isolate the change that introduced the problem?
(In reply to comment #16) > Can you try and use git-bisect to isolate the change that introduced the > problem? > Well, as of today the responsible change seems to be commit http://gitweb.freedesktop.org/?p=xorg/driver/xf86-video-ati.git;a=commit;h=5db3afaa1fdb69d382ac769ef40191a4b964d28e Without that change everything works, with that change I experience this mode changes. I can bisect, but since the problem is apparently random (or at least I can't figure out what triggers it) and it can happen in 30s-6h from starting X this could take a _long_ time.
Okay, I misread comment #15 to apply to that as well.
Do you still have this problem with ati git master?
(In reply to comment #19) > Do you still have this problem with ati git master? > Sorry for being late, I had a rough week. I've been using a git snapshot (last included commit: 3cfbcf4cafbdfdb33411d16e51fb1f77cd0f93dd) for the last couple days and my problem did not present itself again.
Ok, I've been using daily git snapshots for the last 5 days and the problem's gone. I'm closing the bug (I hope it's ok if I do it myself).
Bug 13229 was marked as a duplicate of this one, and still affects ati master. The problem arises because VGA output is now enabled by default even if no external monitor is attached, and for some reason for me it has a resolution of 1152x768 (not 1024x768). So, it's as if I'm using an external monitor that's slightly bigger than the laptop screen, and the mouse carries on past the right-hand edge of the display. With metacity, windows still maximise to the correct size (but Fitts' law is broken), and with simpler window managers (e.g. evilwm) the windows maximise to the size of the virtual screen. I should note that "xrandr --output VGA-0 --off" does fix the issue for me until restarting X, but I expect quite a few people will notice this bug and won't want to edit xorg.conf. My logs and config files are at http://bugs.debian.org/458415 .
The problem is you are using an XPRESS card. We are not able to reliably do DDC or load detection on the VGA port right now so there is no way to know whether or not a monitor is connected or not. I hope to sort out the DDC stuff at some point, but until then, we need to leave the status as unknown or some people's cards won't come up at all. In the interim, you can add a monitor section linked to the VGA output and set it to be disabled by default. See this page for more info: http://www.intellinuxgraphics.com/dualhead.html
Closing but for inactivity, please reopen and update informations if you still have this issue with recent xf86-video-ati.
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.