Red Hat Bugzilla – Bug 41230
xfig: scaling after rotating an ellipse crash linux
Last modified: 2008-05-01 11:38:00 EDT
From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.2.16-22 i586) Description of problem: Drawing an ellipse in xfig, rotating and scaling can trigger an error in the xserver that crashes the Xserver/linux. How reproducible: Always Steps to Reproduce: 1. Draw an ellipse 2. Rotate the ellipse 3. Scale the ellipse Actual Results: On a redhat 7.1 system X/Linux hangs with only hardware reset possible, ie, severe. On a redhat 7.0 system an extra erroneous line appears - and this is erased when the window is redrawn, ie, not severe. Expected Results: The errorneous line should not be drawn. Additional info: The redhat 7.1 system I have is installed on a computer with onboard graphics card that uses the ordinary RAM as video RAM. Linux and/or the Xserver might not be able to handle that properly and xfig is just triggering an error in that connection, ie., there are both an error in the Xserver/Linux (or my bios configuration) and in xfig When I start xfig on an other computer and let the xfig display on the redhat 7.1 with onboard graphics card the error is still triggered - and the Xserver/Linux crashes - on the computer with the display - not the other computer running xfig.
i cannot reproduce this problem on my notebook and my machine. I will say it's a bug in Xserver. Could you please give more infos about your graphic card (chipset?)
I cannot reproduce this either. I need your X server logs and your XF86Config files attached uncompressed using the hyperlink below
This also happens to me, actually just starting xfig hangs my system. Looks similar to bug 47650. The system has no time to write any log files, just freezes. The problem depends on the X server, it happens with XFree86 4.0.3 (trident chipset) but not under XF86_SVGA. I also observed the problem under Mandrake 8.0 (with XFree 4).
Created attachment 34101 [details] From /var/log/
I can still reproduce the problem. My graphics card is on-board: "Build S3 Savage4 (86C370) in VT8365(VT8364)" with "Support shared memory architecture". > rpm -qa | grep X playmidi-X11-2.4-12 Xaw3d-1.5-9 XFree86-4.0.3-5 Xconfigurator-4.9.27-1 XFree86-twm-4.0.3-5 XFree86-libs-4.0.3-5 XFree86-xfs-4.0.3-5 XFree86-100dpi-fonts-4.0.3-5 XFree86-tools-4.0.3-5 XFree86-xdm-4.0.3-5 XFree86-devel-4.0.3-5 XFree86-75dpi-fonts-4.0.3-5 emacs-X11-20.7-34 Xaw3d-devel-1.5-9 I guess the problem might be connected to the shared memory. A line in the XFree log is: (WW) SAVAGE(0): Failed to set up write-combining range (0xd0000000,0x2000000)
It could help to troubleshoot if you could test something for me: Does this problem go away if you enable Option "Noaccel" in your config file, as per the XF86Config man page?
YES, the problem disappears with Option "NoAccel" in the XF86Config-4 (for xfig as well as for xdvi). BTW, even with acceleration disabled, XFree 4 still outperforms Xfree 3 on which the option accel does not work for me. I don't know if this might be related: with rh7.0, I could use Option "Accel" with XFree 3.3 (and get some acceleration) but with the XF3.3 version shipped with rh7.1, the server displays garbage when I use accel. (That's why I switched to XFree 4) In case I hadn't mentioned, my controller is a Trident Microsystems|Cyber 9525 on a Toshiba 4090 laptop.
We have released XFree86 4.1.0-15 as an update for Red Hat Linux 7.1 and 7.2 now. If you still are experiencing these problems with the server, and have upgraded to the latest version of XFree86, I would be glad to try and troubleshoot this issue further. Please update the report to let me know if this still occurs or not with 4.1.0-15. Also XFree86 4.2.0 is available in rawhide which could possibly solve the problem if 4.1.0-15 doesn't and you feel adventurous enough to experiment with rawhide. Either way, let me know how you'd like to proceed.
I just installed rh 7.2 (fresh install) and all updates. It looks like XFree86-4.1.0-15 solves my problems! I deleted all XaaNo*** options from XF86Config-4 and did not experience any hangs. Also (I did not register this as a bug), I can now suspend (and resume) my laptop without switching virtual consoles. With this version of XFree, X gets restored fine on resume.