Bug 186321
Description
Peter Dawes
2006-03-22 21:37:51 UTC
Created attachment 126508 [details]
Output of lspci
Created attachment 126509 [details]
X logfile
Created attachment 126510 [details]
xorg.conf file
Created attachment 126511 [details]
output of lspci (properly uploaded...)
Put 'Option "nodri"' in the device section of the config, and reboot and start the X server. Note if the problem is still reproduceable, and if so, try commenting out the 'Load "dri"' line and rebooting again and noting if the issue still remains. Next, try using: Option "noaccel" in the device section. Do any of these options prevent the problem from occuring? Created attachment 126627 [details]
Log with option "nodri"
Created attachment 126628 [details]
Log from not loading DRI
Created attachment 126629 [details]
X server log with option "noaccel"
Mike - I've tried all those options, in sequence. Problem still happens - system hangs as soon as I try to log out. It looks like it's hanging as it tries to re-start the server, but I can't be sure about that. I'm attaching the logs from each try. Created attachment 126991 [details]
X server logs
See above logs, and note the "INVALID MEM ALLOCATION" and "ADVANCE_LP_RING: outring (0x2f) isn't on a QWord boundary" messages. A quick Google search for these reveals that many other distros are having exactly this problem. I'm not sure that this is the same problem, but on a Dell GX620 D 830, Intel GMA950 video, Dell 2405FPW monitor), the i810 driver worked only at very low resolution, and then it crashed whenever I logged out of Gnome, or try to kill X. The last message in /var/log/messages before the crash was always [drm] Initialized i915 1.4.0 20060119 on minor 0 The X server logs did NOT have the invalid memory allocation entry. Also, FWIW, I seem to be able to get things to almost work by using the vesa driver instead of i810. (Still problems getting 1920x1200, which looks like 1600x1200 even though xdpyinfo says it is 1920x1200. I don't think I'll report a bug about this because vesa is the wrong driver.) I am experiencing what looks to be a very similar issue, also on a Dell GX620 (see bug 192885) Can anyone reproduce this with xorg-x11-drv-i810 1.6.0 or later? (In reply to comment #14) > Can anyone reproduce this with xorg-x11-drv-i810 1.6.0 or later? Is there a package built against Core 5 anywhere? I've downloaded the source from x.org, and will try building that, but... I have here Dell Inspiron 2200 with 00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03) (PCI ID 8086:2792) and just with default configuration Section "Device" Identifier "Videocard0" Driver "i810" VendorName "Videocard vendor" BoardName "Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller" EndSection it works without problems (no Composite though). Sorry, fortgot to add the most important. My driver is xorg-x11-drv-i810-1.6.0-2 I have the identical problem to the above. My motherboard is an Intel d945gnt with integrated i810 video. I have updated the bios to 11/9 version and the problem persists. However, this time I have noted that occasionally the blow up is in the bios. My x11 log only indicated a new instruction for which there is no driver equivalent. This note is being written from my usable windows xp system. Please read my earlier comments on this same subject. The X11 files with vanilla fc6 (oct 24) are more stable then the Nov 20 release of 4 patches for x11. There was a chance that the problem was actually outside of the i9810 in Hal or in the font rendering patch that was released at end November. March 15th (ides of March). System crash when logon screen presented. My mother board is from the same family as the above reported problem. I am using the 32 bit I386 version both with and without XEN. When I removed the exclude=*xorg* from my yum.conf file, here is the status. System boots until logon screen is presented. I can follow all the messages Immediately when the logon screen being presented, it is there for a splint second and then there appears to be a bad branch to the bios, as if a address segment register was initialised to zero. I end up with a nice Intel mother board message and no keyboard or mouse. It takes a system reset to reboot to the same position where the bad branch reoccurs. Here is my configuration. Intel d945gnt mother board, 1 gig memory, ide 200g hard drive, dual core intel d930, fc6 xen or non xen version (also current to patches as of March 12th). The graphics is provided by the on-board hardware, with the default 256k shared memory). This is a two hard disk system, where each disk is bootable. Disk2 is a 250gig sata, on which fc6 also suffered with the same problem until, immediately after October 24th FC6 re-install. Immediately following that install and before updates I added the parameter exclude=*xorg* to the yum config and then installed all updates. With the xorg exclude=, system boots and operates as expected. Since then, on Fedora system 2 (the sata drive), I permitted the *.drv files to be installed. There does not appear to be any problem with the *.drv files. My initial complaint was filed for the xorg.rpm files released about November 20th. The xorg rpm files on or about that date that may contain the bug and could be one of: xorg-x11-server-Xnest-1.1.1-47.7.fc6.i386.hdr xorg-x11-server-Xorg-1.1.1-47.7.fc6.i386.hdr xorg-x11-server-sdk-1.1.1-47.7.fc6.i386.hdr xorg-x11-server-Xvfb-1.1.1-47.7.fc6.i386.hdr xorg-x11-xinit-1.0.2-15.fc6.i386.hdr I am going to use the system rescue and eliminate the logon screen (gnome) from being the cause. and will update this bug report following this exercise. As indicated, the system crashes when the logon screen is presented. I will also report on my success with "nodri" In the crash, the bios display, in color (Intel(R) 82945G Chipset family PCI Accelerated SVA 2006/... is presented. Again, my way to bypass the problem is to just remain with the xorg software that was distributed with FC6 vanilla DVD. FYI. The unofficial FC6 respin crashes at the same point. 2nd FYI, testing takes hours. so, I may just reinstall FC6 Oct 24th version, block all xorg updates, and enjoy FC6 until FC7 is out. The live CD for FC7 appears to work fine.... Here is a status report (1 hr after prev posting). System with nodri did not crash immediately, only X mouse showed on screen. Ctl-alt-bkspace gave me root prompt, from where, the startx command seemed to lock system up. I rebooted from 2nd system, renamed the xorg.conf and rebooted. The system arrived at the logon screen, showed me the bios message from the onboard graphics card, and immediately gave me the logon prompt. I am writing from the logon prompt. I will do a sync, and logoff - reboot, and come back with the hopefully encouraging news. Created attachment 150192 [details]
self explanatory. This was with vesa driver test.
Created attachment 150193 [details]
Log from last run
Created attachment 150194 [details]
self explanatory. This was with vesa driver test.
-rws--x--x 1 root root 1815124 Feb 28 16:34 Xorg Failing version
-rws--x--x 1 root root 1820676 Oct 5 00:05 /usr/bin/Xorg
I have copied 2nd file to overwrite the smaller Xorg file.
Further details soon
HORRAH HORRAH. Replacing file1 with file 2 (The fc6 vanilla release Xorg excutable solves the problem. My second system is now stable. Hip Hip Horrah. These feedback is as good as I can get it, without opening up the code to find out what is missing. (new module is smaller than older one). This new module should get onto the respin cd/dvds. This problem exists since FC5. Note the temporary solution with fc6, and absolutely no solution with fc7. Almost 8 months to fix. Has any progress been made. I cannot use XEN with a computer purchased expressly to be able to do so. With FC7 golive (May 31st) XEN and Non XEN, I have not experienced any problems. As far as I am concerned, it now works for me. Thank you again. Fedora Core 5 is no longer supported, please, could you reproduce this bug with the updated version of the currently supported distribution (Fedora Core 6, or Fedora 7, or Rawhide)? If this issue turns out to still be reproducible, please let us know in this bug report. If after a month's time we have not heard back from you, we will have to close this bug as CANTFIX/INSUFFICIENT_DATA. Setting status to NEEDINFO, and awaiting information from the reporter. Thanks in advance. Sorry, I don't have the hardware to test this available anymore. Comment #27 seems to indicate this is fixed as of Fedora 7 though. Closing per comment 27 and because we don't support the original reported distribution. |