Hi, I have an Asus p5ld2-vm with the i945g integrated video cards. I have pentium d 820 cpu and a gentoo compiled in 64Bit. I've tried various version of xorg-x11 (from 6.8.2 with a patch to 7.0 rc2) and all do the same result: my video cards is recognized correctly and the driver works well, but when I close the X session the computer freeze and I must do ad hard reset. I've do my test with the 2.6.14 kernel from gentoo and the i915 kernel driver was load correctly. With the vesa driver there's no problem. What's the problem? Thanks ENx
Created attachment 3839 [details] My xorg log This is the log of the xorg server when it crashes
Created attachment 3840 [details] My config file This is my config file using the i810 driver
Can you VT switch or does that lockup too ? If so, try adding Option "VBERestore" to your Device section.
(In reply to comment #3) > Can you VT switch or does that lockup too ? > > If so, try adding > > Option "VBERestore" > > to your Device section. yes I can VT switch without problem. Instead, when I close the X session the screen remain blank with a not blinkng cursor. ENx
O.k. the next thing to do is check to see if you can ssh in to your machine from a network and get a copy of the logfile when it crashed.
You might also want to look at the output of 'dmesg' too.
(In reply to comment #6) > You might also want to look at the output of 'dmesg' too. The problem is that the machine will freeze completely, and I cannot connect to it in any way (no ssh ftp etc...). The log I posted, is the one I get when I hard reboot after the crash. ENx
Have you tried commenting out the Load "dri" line in your config file ?
(In reply to comment #8) > Have you tried commenting out the > > Load "dri" > > line in your config file ? Yes, I've tried with and without the kernel module compiled, but, when I tried with a kernel that doesn't have the i915 module at the close of the x server appear this sentence: "It is not supported in any way. bugs may be filled in the bugzilla ay http://bugs.freedesktop.org". When I tried with a kernel that has the i915 module it simply freeze the pc. May be a kernel related problem? I will try with a vanilla kernel and othe kernel and in the end, with another linux distro. ENx
I've tried with a 2.6.15_rc1 vanilla kernel without any success. The result is the same: a blank screen with a not blinkng corsor.
(In reply to comment #9) > (In reply to comment #8) > > Have you tried commenting out the > > > > Load "dri" > > > > line in your config file ? > > Yes, I've tried with and without the kernel module compiled, but, when I tried with a kernel that > doesn't have the i915 module at the close of the x server appear this sentence: > "It is not supported in any way. bugs may be filled in the bugzilla ay > http://bugs.freedesktop.org". > When I tried with a kernel that has the i915 module it simply freeze the pc. > May be a kernel related problem? I will try with a vanilla kernel and othe kernel and in the end, > with another linux distro. > I wasn't talking about the kernel driver here, I was talking about the X config file xorg.conf. Try commenting out the line that says... Load "dri" and see if that helps.
I've commented out that line but none have changed! The problem can be that I've compiled the x server in 64bits mode? ENx
You might want to try again with X.Org 6.9/7.0.
(In reply to comment #13) > You might want to try again with X.Org 6.9/7.0. I've tried X.Org 7.0 on my gentoo, but the result is the same. In console appear this message when i stop the X.Org server: Unable to handle kernel paging request at ffff810027def810 RIP: <ffffffff8011e969> {dump_pagetable+81} Modules linked in: 915 drm [...] [...] The pc stop and I can only do an hard reset. I've used the 1.4.1.3 version of i810 video driver installed with the portage system (it is marked unstable only for x86 arch, but none for amd64 arch). It is possible that the i810 driver is not stable for x86_64 architecture? ENx
You need to attach more of that kernel output so we can see where things are going wrong.
I think I had the same problem. I have a DELL GX620 machine which has a i945G chipset. Both SLES 9 SP3 x86_64 and Ubuntu 5.10 x86_64's X with the i810 driver will cause kernel panic when I terminates X, either by logging off GNOME/KDE, shutting down/reboot system or killing X manually. Even Ubuntu Dapper Drake x86_64's xorg 7.0 has this problem. If I use VESA driver all these distributions' X will work with no problem. The only known (by me) distribution which seems have fixed this problem is RHEL 4 U2 x86_64. It's i810 X driver works fine on this machine. Maybe we can find some clue in it's patches. I will take a snapshot of the kernel panic screen and post here later.
I have installed gentoo linux in 32 bit mode (pentium4) and all is working fine. There's no problem with the i810 driver! I think it's a specific problem of 64 bit version of the driver. ENx
This is a snapshot of the kernel panic. I'm sorry it's not very clear. http://www.flickr.com/photo_zoom.gne?id=112755726&size=o
Can you press SHIFT+PGUP before taking the shot to get the top most part of the crash ???
I tried but Shift+Pageup never work at this time. I'm seeking a serial port cable to capture the output on another machine, but for the time I have only this kind of screenshots :-( Is it possible to tune the kernel to print only the first screen of panic output? (In reply to comment #19) > Can you press SHIFT+PGUP before taking the shot to get the top most part of the > crash ???
Finally I got a serial port cable and captured the kernel panic output! I don't know if I can use a paste bin so I just attached it here. The first one is by manually kill the X process when running an AMD64 generic kenrel. The second one is by reboot the system when running an AMD64 Xeon kernel, which is my default kernel.
Created attachment 5009 [details] Generic AMD64 kernel panic output (by manually killing X server)
Created attachment 5010 [details] AMD 64 Xeon kernel panic output (by rebooting the system)
I have seen this bug in the same was as Ming Lei. Intel Celeron with AMD64 extensions, running on Abit P5GV-MX
I have been able to work around this bug by setting Execute Disable Function to ENABLED in BIOS (Advanced -> CPU Configuration). X now terminates and restarts without issue.
This looks like a kernel bug in the agp module. The kernel is crashing in the agp_free_memory() function. Have any of you reported that to the kernel folks ?
I have exactly the same problem. This VGA is based in the 82865G chipset, and is the builtin VGA of an Asrock 775i65GV motherboard, with a Celeron 2,9 Ghz processor.
Maybe you can fix it in exactly the same way by changing the option in the BIOS ?
That is correct. Enabling "disable execute" fixed the crashes, except when running Xorg under Xen. Then, the problems appear again, and more serious. No need to kill X or switching to text mode for crashing. Adding agp=off, however, made my system stable, even under xen. This fact suggests (but does not demonstrate) a kernel bug in the agp code. On the other hand I congratulate the original reporter for being able to obtain an Oops.
If not, you might want to check with the manufacturer if they have an updated BIOS.
whoops, wrong bug. ignore that.
Hi, Gustavo Boiko pointed me to this bug. I've spent some time debugging a very similar bug recently (probably this one), and it was a kernel bug fixed by this patch: http://lkml.org/lkml/2006/10/18/68 Could you check if this patch solves the problem for you?
The bug priority was upgraded (P2->high) with the bugzilla configuration change. I'm Changing the priority back to the normal one. Sorry for the spam.
I am experiencing this bug on the RedHat Enterprise Linux 5 final release. My machine is a Toshiba Satellite A105-S4334 with a Core 2 T5500 and *no* ability to turn "Execute Disable" on in the BIOS (tried to get Toshiba to patch the BIOS, they refuse). Anyway, I can confirm that the 1 line kernel patch referred to above completely resolved the X crash on exit for me. I believe FC6 has moved up to kernel 2.6.20 so the patch should be there for updated FC6 boxes. I have an open bug with RedHat (#232748) requesting that the 1 line fix be added to the RHEL kernel also. Jay
Thanks. Closing this as it's NOTOURBUG.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.