when you say sounds, what type of sounds? Doom sounds or just rubbish?
You should be stracing the doom.x86 process. In this example that would
be 13237. Furthermore, you need to get an strace before the crash.
strace -o /tmp/doomtrace -f <doom3_executabl_name>
Now, if you've soft rebooted and the screen came up garbled, then some
process is not talking nicely to the card and the problem is
persistent.
On Wed, 2005-11-16 at 17:21 +0100, Keith Vassallo wrote:
> I launched Doom III and waited for it to freeze, then I ssh'ed into the
> system and ran top.
>
> The top most process was X (with some 99.9% CPU activity). Sometimes
> doom.x86 (the Doom III process, of course) would rise to the top for a
> few moments, but then go down again. The relevant lines from top:
>
> When X was on top:
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 11515 root 25 0 305m 32m 8824 R 99.8 2.2 0:45.56 X
> 1 root 16 0 2548 536 452 S 0.0 0.0 0:01.15 init
> 2 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0
>
> When doom.x86 was on top:
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 13237 keithvas 23 0 802m 715m 8320 R 97.3 47.7 2:04.93 doom.x86
> 11515 root 25 0 305m 32m 8824 R 2.4 2.2 0:36.50 X
> 13235 keithvas 16 0 10424 1276 928 R 0.2 0.1 0:00.38 top
>
> My load average just kept rising, and rising. At a point it was 20.86,
> 8.14, 3.16 (whatever that means).
>
> Before the game freezes, the top most process is doom.x86 (with around
> 85% CPU usage) and X comes in at third or fourth place.
>
> Whilst the game is frozen, I occasionally here sound from my speakers.
> Other times I hear nothing. When I'm hearing sounds, I'm able to type in
> commands via SSH. When I'm not hearing sounds, nothing appears and the
> system grinds to a halt. This suggests that (perhaps) the system tries
> to recover but then is frozen again.
>
> When I run strace on X, I keep getting the following:
> rt_sigreturn(0xe) = 36984
> --- SIGALRM (Alarm clock) @ 0 (0) ---
> rt_sigreturn(0xe) = 36984
>
> Doom III spawns several processes, here's what they say:
> [ Process PID=13250 runs in 32 bit mode. ]
> select(19, [18], NULL, NULL, NULL
>
> [ Process PID=13251 runs in 32 bit mode. ]
> getppid() = 13250
> poll([{fd=15, events=POLLIN}], 1, 2000) = 0
>
> [ Process PID=13253 runs in 32 bit mode. ]
> gettimeofday({2811813141633945, 586448462527099936}, NULL) = 0
> nanosleep({38654705664000000, 586448610090393576}, NULL) = 0
> gettimeofday({2868712868371353, 586448630030824480}, NULL) = 0
>
> I tried to reboot the PC. I did kill -9 (the X pid) and kill -9 (the
> doom III pid), but when I looked at the machine's monitor it just turned
> off (i.e. no output from the video card). I then tried to reboot via
> SSH. The system looked like it tried to reboot, but then a garbled
> screen appeared and nothing more happened.
>
> The second time I tried this, kill -9 (the X pid) stopped sound coming
> from my speakers, but the image just froze (i.e. frozen, not turned
> off). I then killed doom, but the machine remained frozen. After this,
> the machine completely froze (i.e. no commands, even from SSH).
>
> Keith
>
> Andrew Cilia wrote:
> > Keith,
> > just a point on something you wrote in an earlier email. If you can
> > ssh in, then the system has not crashed as such. It would seem that your
> > video subsystem is stuck. Given that you can get in via ssh, is there
> > anything that is hogging the CPU that you can see via top. If so, get an
> > strace on it as this will help in debugging.
> >
> > Another thing. Are you running in SMP mode?
> >
> > subsystem is
> > On Wed, 2005-11-16 at 07:26 +0100, Keith Vassallo wrote:
> >
> >> When I bought the card, I was using kernel 2.6.12, I re-compiled that,
> >> and it still didn't work. I also compiled 2.6.13 - still didn't work.
> >>
> >> glxgears works. Also, some games like GTA: Vice City work with no
> >> problems. Also, If I set the quality of graphics in Doom III to medium,
> >> the game works.
> >>
> >> It seems like only very intensive use makes the box die. It could be
> >> something to do with the bug Andrew mentioned earlier (the bug is only a
> >> problem with very intensive use). It can't be a hardware problem,
> >> Windows worked.
> >>
> >> Keith
> >>
> >> Jean Azzopardi wrote:
> >>
> >>> I will not pretend to be an expert on this, but have you tried
> >>> recompiling the
> >>> kernel?
> >>> Also. do simpler 3D apps work? Such as glxgears? Try running glxgears in
> >>> a
> >>> terminal...
> >>>
> >>> On Tuesday 15 November 2005 6:30 pm, Keith Vassallo wrote:
> >>>
> >>>
> >>>> The messages I get from dmesg after start-up show no errors - just that
> >>>> AGPART is being loaded, finds a graphics card and everything seems
> >>>> normal.
> >>>>
> >>>> I had already checked out the thread you pointed me to. The thread
> >>>> provides a patch to fix the problem, but I'm supposed to already have
> >>>> that patch - Gentoo have included it in their ebuild (of which I have
> >>>> the latest, 7676).
> >>>>
> >>>> I tried to figure out what they're talking about when they mention
> >>>> global_flush_tlb(), and found this interesting post:
> >>>> http://marc.theaimsgroup.com/?l=linux-kernel&m=112928307319954&w=2
> >>>>
> >>>> It seems, as is also said in the nvidia post, that this is being worked
> >>>> on in kernel 2.6.14. Unfortunately, the thread I found mentions delays -
> >>>> not total freezes - being caused by this "bug". The patch is against
> >>>> 2.6.14-rc4. I don't totally understand kernel terminology - but if the
> >>>> patch is against rc4, does that mean it will be included in rc5?
> >>>> Whichever it's included in, the latest (testing) kernel available on
> >>>> Portage is 2.6.14-r2.
> >>>>
> >>>> Seems like this will be a waiting game, unless anyone else has
> >>>> suggestions.
> >>>>
> >>>> Keith
> >>>>
> >>>> Andrew Cilia wrote:
> >>>>
> >>>>
> >>>>> Do you get any messages after agpart module kicks in during startup? Did
> >>>>> you try some other forums besides gentoo? For example, I found this:
> >>>>>
> >>>>> http://www.nvnews.net/vbulletin/showthread.php?t=57990
> >>>>>
> >>>>> Cheers
> >>>>>
> >>>>>
> >>>>> On Tue, 2005-11-15 at 16:01 +0100, Keith Vassallo wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>>> Hey Guys,
> >>>>>>
> >>>>>> I've recently upgraded to an XFX GeForce 6800 GT from my previous card,
> >>>>>> a GeForce FX 5200. Since having done so, I'm having problems playing
> >>>>>>
> >>>>>>
> >>> games.
> >>>
> >>>
> >>>>>> When starting Doom3 in "Ultra high" or "high" quality mode, the game
> >>>>>> locks up either seconds, or minutes, after the game begins. The whole
> >>>>>> system freezes, neither CTRL+ALT+BACKSPACE nor anything else works.
> >>>>>> Starting doom in "medium" quality mode seems to stop the problem from
> >>>>>> happening, although I haven't played the game for longer then 30mins.
> >>>>>> When starting Half Life 2, the game freezes a few seconds after the
> >>>>>> menu
> >>>>>> is displayed. With Counterstrike: Source, the game loads a map, then
> >>>>>> sends me back to the desktop.
> >>>>>>
> >>>>>> Here's some information you may need:
> >>>>>>
> >>>>>> Gentoo running amd64 on AMD Athlon 64 3000+ (Socket 939)
> >>>>>> XFX GeForce 6800 GT (AGP)
> >>>>>> 1.5GB DDR RAM
> >>>>>>
> >>>>>> cat /proc/driver/nvidia/agp/status:
> >>>>>>
> >>>>>> Status: Enabled
> >>>>>> Driver: AGPGART
> >>>>>> AGP Rate: 8x
> >>>>>> Fast Writes: Disabled
> >>>>>> SBA: Enabled
> >>>>>>
> >>>>>> Kernel: 2.6.13-gentoo-r3
> >>>>>> nvidia-kernel: 1.0.7676
> >>>>>> nvidia-glx: 1.0.7676-r2
> >>>>>>
> >>>>>> I've also used nvidia-settings to check the card temperature. 15mins
> >>>>>> after boot, nvidia-settings reports:
> >>>>>>
> >>>>>> Core Temperature: 44C
> >>>>>> Ambient Temperature: 36C
> >>>>>>
> >>>>>> I've searched through the Gentoo forums for similar problems, and this
> >>>>>> has been reported one or two times, none of these people found a
> >>>>>> solution (or posted about it). I looked on my motherboard for capacitor
> >>>>>> decay (as described in another post) and haven't found any. I also
> >>>>>> don't
> >>>>>> have X Composite extensions enabled.
> >>>>>>
> >>>>>> Whenever the PC crashes, I can see the following in /var/log/messages
> >>>>>> (I
> >>>>>> have to SSH to do this, the machine is too frozen to launch a terminal
> >>>>>> locally)
> >>>>>>
> >>>>>> Nov 3 22:05:06 silver NVRM: Xid: 25, L0 -> L0
> >>>>>> Nov 3 22:05:06 silver NVRM: Xid: 6, PE0000 1f08 00000000 00000000
> >>>>>> 00f1efeb 00000000
> >>>>>> Nov 3 22:05:09 silver NVRM: Xid: 8, Channel 00000020
> >>>>>> Nov 3 22:05:17 silver NVRM: Xid: 8, Channel 00000020
> >>>>>>
> >>>>>> etc...
> >>>>>>
> >>>>>> I've installed Windows on this machine and a few games just to test -
> >>>>>> everything worked fine, so it can't be a hardware problem.
> >>>>>>
> >>>>>> Any help would be greatly appreciated.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>> _______________________________________________
> >>>> MLUG-list mailing list
> >>>> [email protected]
> >>>> http://mailserv.megabyte.net/mailman/listinfo/mlug-list
> >>>>
> >>>>
> >>>>
> >>> _______________________________________________
> >>> MLUG-list mailing list
> >>> [email protected]
> >>> http://mailserv.megabyte.net/mailman/listinfo/mlug-list
> >>>
> >>>
> >> _______________________________________________
> >> MLUG-list mailing list
> >> [email protected]
> >> http://mailserv.megabyte.net/mailman/listinfo/mlug-list
> >>
>
> _______________________________________________
> MLUG-list mailing list
> [email protected]
> http://mailserv.megabyte.net/mailman/listinfo/mlug-list
--
Andrew Cilia B.Sc. RHCE MCSE ICSE
Senior Software Consultant
Philip Toledo Limited
Computer & Communications Solutions
Notabile Road, Mriehel BKR01, Malta
Tel: (+356) 2144 5566
Mobile: (+356) 9943 0588
Fax: (+356) 2148 4316
Website : www.ptl.com.mt
The information contained in this email is confidential and may be privileged.
It is intended for the addressee only, if you are not the intended recipient
please notify the sender and delete the email immediately. The contents of
this e-mail must not be disclosed or copied without the senders consent. We
cannot accept any responsibility for viruses. Any views expressed in this
message are those of the individual sender, except where the sender
specifically states them to be the view of Philip Toledo Limited
