OS - out of the box and up to date FC10. HW - Asus M2A-VM motherboard with integrated Radeon X1200 Xorg freezes from time to time for no apparent reason. When it happens the mouse moves a bit, the keyboard does not respond and the image on the screen is frozen. However, the box is accessible via network. In the past it used to happen with "radeon" driver only, when I switch to radeonhd the problem disappeared. However, after some upgrade it reappeared and now happens with both radeon and radeonhd. fglrx driver were never installed on this box.
Created attachment 26736 [details] [review] Xorg.0.log
Can you try with xf86-video-ati git master or 6.12-branch?
Yeah... will probably take a day or to, to compile and reproduce the problem
git version does not compile ./configure: line 20833: syntax error near unexpected token `XINERAMA,' ./configure: line 20833: `XORG_DRIVER_CHECK_EXT(XINERAMA, xineramaproto)' libXinerama-devel.i386 is installed. Suggestions ?
(In reply to comment #4) > git version does not compile > > ./configure: line 20833: syntax error near unexpected token `XINERAMA,' > ./configure: line 20833: `XORG_DRIVER_CHECK_EXT(XINERAMA, xineramaproto)' > > libXinerama-devel.i386 is installed. > > Suggestions ? > You need xorg macros and devel packages. I'm not sure what the packages are needed on redhat off hand, probably xorg-x11-util-macros, xorg-x11-server-devel, and libdrm-devel.
xf86-video-ati-6.10.0.tar.bz2 compiles just fine, which means that all packages are installed. git version fails...
installing the xorg-x11-server-devel should fix that error.
xorg-x11-server-devel is installed.
Builds OK for me on F11 (today) and F10 (two days ago). Looking at the fedora .spec file for the xorg-x11-drv-radeonhd package, # grep BuildRequires xorg-x11-drv-radeonhd.spec BuildRequires: pkgconfig BuildRequires: pciutils-devel BuildRequires: zlib-devel BuildRequires: xorg-x11-server-sdk >= 1.1 BuildRequires: xorg-x11-server-sdk >= 1.4.99.1 BuildRequires: xorg-x11-server-sdk < 1.4.99.1 BuildRequires: libdrm-devel BuildRequires: xorg-x11-proto-devel so I certainly have that set of packages installed, if that helps.
I don't know, looks like everything is installed. Do you have a clue why 6.10 compiles just fine and git version does not ? Is it so different ?
I managed to compile it,no idea what was the problem. will try the driver now and report tomorrow.
It is hard to tell whether the git version has this problem, but it is definitely worse than the original one in one respect - it often does not wake from suspend, while the original one did quite well.
OK, it finally happened with the git version as well.
Should I switch to proprietary drivers or is there a chance for a fix ?
You can probably circumvent the issue with radeonhd by using Option "AccelMethod" "shadowfb" It looks like the issue you're seeing happens with acceleration active only. shadowfb is a bit slower (especially for video), but reasonably fast to actually use it.
Well, video is important, but apparently I have no choice - with fglrx drivers which I tried yesterday the system reboots when X starts... Hope there will be a solution soon enough, it's kinda ironic to switch from Windows to Linux for stability and have the machine freeze once a day...
AccelMethod shadowfb is VERY slow, too slow even for web browsing...
Then something is going wrong. Please attach the xorg.conf with ShadowFB, and a new Xorg.0.log.
Created attachment 27017 [details] shadowfb - xorg.conf xorg.conf with shadowfb. scrolling is painful, it takes a few seconds for page down in Firefox !
Created attachment 27018 [details] shadowfb - Xorg.0.log
Nothing in the logs that reminds me of anything. But are you using a compositing manager (desktop effects, compiz)? That won't work without 3D, of course, and shadowfb disables 3D.
No, I'm not using compiz, just plain old GNOME. Are you saying that ShadowFB should be visibly faster (for things such as scrolling in firefox) than no acceleration at all ? It is not in my case.
On scrolling shadowFB should be almost as fast as EXA. On the other hand NoAccel is close to unusable. So yes, there is a dramatic difference between shadowFB and NoAccel.
Than, I would say it looks like an additional bug. Should I create another bug report ? Is there anything I can do to help you debug this issue ? In case anybody is interested in solving it...
Looks like you're right. Unfortunately, this will be difficult to debug unless one can reproduce the issue. And I'm utterly swamped for the time being.
Does this issue occur with the preferred ati driver (xf86-vide-ati)? If so, please move this to the Driver/Radeon component. Development of radeonhd has pretty much halted and development focus is on the ati driver. Please see http://www.x.org/wiki/radeonhd If the issue does not exist in the ati driver (or if there is no response to this message), this bug will be closed as WONTFIX unless someone contributes a patch.
Closing due to lack of response. Please reopen and move to the Driver/Radeon component if this issue persists with 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.