I was trying to upgrade my PC to get decent FlightGear performance. FlightGear (www.flightgear.org) is an open-source flight simulator. I bought 2 cards, one with a Radeon 7500QW, the other has a Radeon 9000If. Both have 128MByte memory, and are Powered by ATI, Built by Hightech Information Systems, www.hightech.com.hk. It's virtually impossible to get Built by ATI cards here. My PC is an Athlon 800 with VIA KT133 chipset. Here are my observations: XFree86 4.2.0 (RPMS from RedHat 8.0): 7500: - scenery is drawn nicely - aircraft instrument panel is not drawn correctly. According to the FlightGear mailing list this is a known problem with 4.2.0 and Radeons. - runs stably 9000: - not supported by that release. XFree86 4.2.99.3 (RPMS from RedHat 8.0.93): 7500: - scenery exhibits massive texturing errors (i.e. wrongly colored/textured triangles) - aircraft instrument panel is drawn correctly - display freezes after roughly a minute. I could most of the time log into the computer over the network, according to strace the X server called the RADEON_CP_IDLE ioctl in a loop and was getting EBUSY all the time. Killing the X server freezes the machine completely. 9000: - works nicely, no texturing problems, instrument panel drawn correctly XFree86 4.2.99.901 (RPMS from Mike Harris): 7500: - scenery still exhibits the texturing problems - the freezes are gone
You need to attach your xfree config file and log file please.
Created attachment 90083 [details] XF86Config
Created attachment 90084 [details] XFree86.0.log
I'm seeing crashes with the 9000 still as of phoebe. 128Mb card, 256Mb Athlon 550, AMD chipset system. I can duplicate this reliably by running fgfs --aircraft=ufo During running it spews r200: debugging messages which look like your mesa still has debugging printfs left in it. I hit ctrl-u ctrl-u ctrl-u (go up to about 3000 feet), then numeric 9 (throttle) a few times then swoop down and buzz the city. At that point the program hangs, dmesg is full of "radeon blah out of buffers" messages and if I kill it the entire box locks solid a few seconds later
I've seen Radeon 9000 crashes since my original posting too. A quick way to reproduce for me is fgfs --airport-id=LSZH --runway=28 and then 2 times click on the right mouse button and then move the mouse to change the view direction. Apparently I was just lucky I haven't seen a crash in the first hour... NB: I also tried to configure dual head on the R7500 card, which made the X server SIGSEGV on start.
The DRI project is aware of this problem, and hopefully will fix it before the 4.3.0 release, and get the changes into 4.3.0. In lieu of that, hopefully patches to fix it will be made available soon for me to integrate. You may wish to join the dri-devel mailing list on sourceforge to track the upstream progress, and provide them details specific to your system, etc. I'll update the report when there is new code available that is believed to address this. Thanks.
Defering for future XFree86 release.
Since this bugzilla report was filed, there have been several major updates to the X Window System, which may resolve this issue. Users who have experienced this problem are encouraged to upgrade to the latest version of Fedora Core, which can be obtained from: http://fedora.redhat.com/download If this issue turns out to still be reproduceable in the latest version of Fedora Core, please file a bug report in the X.Org bugzilla located at http://bugs.freedesktop.org in the "xorg" component. Once you've filed your bug report to X.Org, if you paste the new bug URL here, Red Hat will continue to track the issue in the centralized X.Org bug tracker, and will review any bug fixes that become available for consideration in future updates. Setting status to "CURRENTRELEASE".