Description of Problem: Machine hangs when doing startx on S3 ProSavage KM133 integrated video (on K7VMA motherboard). Its so badly hung that: 1.I can't switch virtual consoles. 2.Ping from a remote pc to the hung machine failed which means I just couldnot access the /var/log/XFree86.0.log file. 3.Only choice was hardware reset so that I could not save the /var/log/XFree86.0.log file at the next reboot. I have faced this problem with both X-4.0.3 and X-4.1.0 . How Reproducible: always Steps to Reproduce: 1. startx 2. 3. Actual Results: hung machine Expected Results: proper X environment Additional Information:
Set your system to boot in runlevel 3 if it is not already. It is critical that I get server log from you in order to diagnose this. What version of XFree86 are you using now? The latest is 4.0.3-22.5 Pleas attach an uncompressed log file from /var/log/XFree86.0.log and also the messages log file (do not post a large file, trim it down please). Use the link below to make file attachments.
It is already set at run-level 3 (text based login). I am using X-4.1.0. I contacted Tim of http://www.probo.com/timr/savage40.html and learnt that X-4.1.0 has their latest 1.1.7 driver. Its nearly impossible for me to get the server log as the whole system crashes when I just run "X". I can't ssh or telnet to the crashed machine from a remote machine. I made a workaround this way. I kept a remote ssh session running a "tail -f Xfree86.0.log" and then ran X. The crash occurred just after this last log message -> (II) Loading /usr/X11R6/lib/modules/drivers/savage_drv.o Right now thats the best I can do. Also lspci gives: 01:00.0 VGA compatible controller: S3 Inc.: Unknown device 8a26
When your system crashes, the XFree86.0.log should remain intact. It may get blown away somehow, but usually it is still intact. I now need the output of "lspci -v" and "lspci -vn" from you.
lspci -v output: 01:00.0 VGA compatible controller: S3 Inc.: Unknown device 8a26 (prog-if 00 [VGA]) Flags: 66Mhz, medium devsel, IRQ 5 Memory at e1000000 (32-bit, non-prefetchable) [disabled] [size=512K] Memory at d0000000 (32-bit, prefetchable) [disabled] [size=128M] Expansion ROM at <unassigned> [disabled] [size=64K] Capabilities: <available only to root> lspci -vn output: 01:00.0 Class 0300: 5333:8a26 Flags: 66Mhz, medium devsel, IRQ 5 Memory at e1000000 (32-bit, non-prefetchable) [disabled] [size=512K] Memory at d0000000 (32-bit, prefetchable) [disabled] [size=128M] Expansion ROM at <unassigned> [disabled] [size=64K] Capabilities: <available only to root>
This chipset appears to have some code in 4.1.0 but it does not appear complete. The Status document for 4.1.0 does not list it as a supported chipset (ProSavage). I have emailed the author to get details. If the chip is in fact supported, then I may add it to Xconfigurator et al. I'll update this when I hear back from Tim.
Aparently it is supposed to be supported at least partially, but there are reports of problems from a few people. Since it is not listed on the support doc, and there are known probs, I consider it a low priority problem right now. Will look into this again in the future though when I hear more from Tim.
With the latest (1.1.18) savage driver from http://www.probo.com/timr/savage40.html the lockup problem is not there anymore.
Great, closing bug, as our current driver is 1.1.20t and others have reported the problem is fixed now also. Thanks.