Hide Forgot
Created attachment 1152159 [details] Xorg.0.log and coredump portions of journal Steps to Reproduce: 1.dnf update previously working Rawhide/Plasma installation 2.reboot 3.login on vtty 4.startx, or startx /usr/bin/icewm Actual results: 1.radeon driver unloads 2.segfault 3.coredump Expected results: 1.radeon driver stays loaded 2.Plasma or IceWM session Additional info: P4 CPU, 2G RAM F23 & F24 both work as expected. 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV200 [Radeon 7500/7500 LE] (prog-if 00 [VGA controller]) Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Dell Optiplex GX260 Flags: bus master, stepping, 66MHz, medium devsel, latency 64, IRQ 16 Memory at f0000000 (32-bit, prefetchable) [size=128M] I/O ports at de00 [size=256] Memory at fe9f0000 (32-bit, non-prefetchable) [size=64K] Expansion ROM at fea00000 [disabled] [size=128K] Capabilities: [58] AGP version 2.0 Capabilities: [50] Power Management version 2 Kernel driver in use: radeon Kernel modules: radeon
Created attachment 1153590 [details] Xorg.0.log, dmesg tail, journalctl -b from i686 rv200 host m7ncd Same problem on another F25 machine with which F24 and Tumbleweed work as expected.
Created attachment 1153592 [details] Xorg.0.log and hwinfo --gfxcard from P4 i686 rv380 host gx28b I'm unable to tell whether or not this might be related. Here, with xorg-x11-drv-ati-7.6.1-3.20160215gitd41fccc.fc24.i686 installed, X simply refuses to run, with no evidence I can locate of any segfault. Using neither xorg.conf* nor xrandr to affect X startup, first error messages are (WW) glamor requires at least 128 instructions (64 reported) (EE) modeset(0): Failed to initialize glamor at ScreenInit() time. which I've been seeing a lot of lately on various hardware in both Rawhide, (IIRC) F24, various openSUSEes and Stretch, trying intentionally to get built-in modeset driver used instead of the otherwise default intel or radeon driver.
This is due to X server changes to drop DRI1. xorg-x11-drv-ati needs a fix to not check DRI1 symbols before loading.
This needs a better solution, but for now I've moved the needed symbol to the DRI2 code.