(Closed) Xorg 7.1 - Nvidia drivers (all versions) re-opened

Message

AliasXZ · Post by **AliasXZ** » Mon Dec 25, 2006 1:52 am

A few months back I had the problem of my desktop becoming "unresponsive" and it appears to of happened again... - the pc appears to be running, I can SSH into it, move the mouse etc. But clicking on things does nothing, the keyboard is dead, so its an X server freeze where the mouse works but nothing else.
-
I was told this is because of Nvidia drivers, can anyone confirm this?
-
This problem happens randomly, I can run my PC for days at a time constant and not have any problems at all, while other times it happens almost straight away.
-
There is nothing in the logs, no (EE) in xorg log, no kernel problems im really stuck with this, if this is an Nvidia problem then I guess theres not a right lot I can do about it...

cheers for any thoughts

coolsnowmen · Post by **coolsnowmen** » Wed Dec 27, 2006 8:32 pm

my 2 cents:

If you were having driver issues, normally I would think you would see far worse problems than that.
In my experiance, hardware issues don't appear like this and manage to stay out of all logs (messages, xorg, kernel).
It sounds software only. You might be able to isolate whether the problem is (X or lower) or (kde/gnome) or higher by switching desktop managers ( you should be able to live with enlightenment for a couple days).
If you can ssh in, do you notice the processor hanging on anything? As in, can you correlate this behaviour with running a certain program.

AliasXZ · Post by **AliasXZ** » Wed Dec 27, 2006 9:11 pm

I am now running kernel 2.6.18-r4 and nvidia-drivers 8776, the problem has not re-occured as of yet (2 days)
-
I ran kernel 2.6.18-r5 with the same nvidia-drivers version and the problem occured twice in 1 day.
-
When i SSH'd in, there was no hanging at all (from what I could notice anyway)
-
I have played games, listenend to music, compiled, changed window managers for a while and nothing...
-
will keep you updated...

-

EDIT:

r5 kernel has same problem, using r6 and it seems to be better...

AliasXZ · Post by **AliasXZ** » Sat Jan 13, 2007 11:04 am

happened again this morning...

cmon guys, i cannot be the only person with this problem

EDIT:

I ssh'd into my pc, killed X and all was fine again so its an X problem or something...

im running Mod Xorg 7.1 - NV drivers 8667

halfgaar · Post by **halfgaar** » Sat Jan 13, 2007 11:45 am

Perhaps it's a hardware fault. Have you run memtest86+ for 24 hours? You can find it on the ultimate boot CD. The graphics card could also be the culprit. Have you over clocked it? You could try running with a different one for a while, if that's feasible.

Is your machine sufficiently cooled?

BlackEdder · Post by **BlackEdder** » Sat Jan 13, 2007 12:37 pm

I sometimes have similar problems, mostly they seem to be caused by a program (wxmaxima/wxwindows). Then if I kill the offending program from the console everything works fine again. When this happens I always retain the ability to switch to console (ctrl+alt+f1).

maxima · Post by **maxima** » Sat Jan 13, 2007 12:56 pm

Are you running composite window manager of any kind? I had exactly the same problem when I had compiz running. Now I've gone back to metacity and the problem dissappeared.

AliasXZ · Post by **AliasXZ** » Sat Jan 13, 2007 1:24 pm

halfgaar wrote:Perphaps it's a hardware fault

Ive swapped the hardware still does it, I wouldn't of thought that because it wouldn't just lock up and everything stay running if it was memory problems, if gfx card was at fault there would be some sort of display issue, but there isn't. I have also tested the gfx card in different machine for days at a time.

BlackEdder wrote:I sometimes have similar problems, mostly they seem to be caused by a program (wxmaxima/wxwindows). Then if I kill the offending program from the console everything works fine again. When this happens I always retain the ability to switch to console (ctrl+alt+f1).

I can't do that, i can only SSH into my box because the keyboard stops responding, I kill X and it restarts then all is well... I don't have any programs like that running.

maxima wrote:Are you running composite window manager of any kind? I had exactly the same problem when I had compiz running. Now I've gone back to metacity and the problem dissappeared.

No, I'm just running fluxbox with no fancy stuff on it.

however, I have been thinking, I do run a framebuffer in 1280x1024 using the vesafb with my own custom background. I think I read somewhere that with nvidia cards, if you have any sort of trouble, you should append an option in xorg.conf, but I can't, for the life in me remember or find anything, can anyone who runs a framebuffer, with an Nvidia GPU shed some light on this?
-
I will play around and report back...

EDIT:

Nvidia-drivers only become an issue if you use the official nvidia-drivers and nvidiafb, they conflict, I use vesafb so the framebuffer isn't the problem... I will try running without it though to make sure...

EDIT:

This is an ongoing issue... Found a few people with same problem. Maybe I should just run xorg 6.8? had no problems at all then.
-
Can I just downgrade xorg and xorg-server or do I have to recompile everything to do with X ?

kraix · Post by **kraix** » Sat Jan 13, 2007 2:45 pm

EDIT:
Whoops reading everything helps. Sorry. Have you tried using the official nvidia drivers?

AliasXZ · Post by **AliasXZ** » Sat Jan 13, 2007 4:40 pm

EDIT:
lol
thats about the only thing I haven't tried, suppose its worth a shot...

halfgaar · Post by **halfgaar** » Sat Jan 13, 2007 5:49 pm

Huh? I thought you were using the official drivers all the time. Above, you mentioned you were using version 8776, which sounds like the version numbering of the official driver.

rini17 · Post by **rini17** » Sat Jan 13, 2007 7:53 pm

I suspect it being more an dbus issue, it haunts me since I first installed Mandrake version that used DBUS-enabled KDE, on different hardware. It also happened few times with Debian Sarge when I installed too bleeding edge KDE backport, also on different hardware. Now it sometimes happens on Gentoo on my new machine, too. All the three machines had a NVIDIA card, and using binary drivers IIRC. In latest stable ebuilds of KDE-3.5.5 I use now it seems to be partially fixed, the system unstucks itself after a minute.

DBUS is responsible for delivering events, so if it gets stuck the it really might look like you're describing- OS itself works, but no mouse/keyboard events are being delivered to X applications.

I haven't a clue how to gather any useful information to do anything with it (like file a bug report), since it only happens here once a month or so.....

halfgaar · Post by **halfgaar** » Sat Jan 13, 2007 8:04 pm

But Dbus isn't used for mouse and keyboard events, so it seems unlikely it can freeze keyboard and mouse input.

kraix · Post by **kraix** » Sat Jan 13, 2007 11:16 pm

Huh? I thought you were using the official drivers all the time. Above, you mentioned you were using version 8776, which sounds like the version numbering of the official driver.

Well, first off he said he was using the nv drivers, I missed that also and thought he was using the official nvidia drivers also. then I looked again and saw he said, and so does his sig, that he is using the nv driver.

AliasXZ · Post by **AliasXZ** » Sun Jan 14, 2007 1:46 am

I am using the nvidia-drivers from portage.
-
for some reason I can't go above 8776, but that's another issue all together...
-
I am downgrading xorg to 6.8.2, if that doesn't work, then I guess I will put up with this as it happens randomly.
-
Thanks for all the replies

micko · Post by **micko** » Sun Jan 14, 2007 9:10 am

I've had the same problem as long as I've used nvidia-driver. Some options in Xorg.conf makes the system crash (and with crash I meand the behaviour you described) more often. Especially RenderAccel seems to have a lot to do with this. With my older nvidia-card I had to switch RenderAccel off to make my system crash less frequantly. Now with my "new" card that option has to be switched on or the crashing happens several times a day.

The problem isn't Gentoo specific. I'm having the same behaviour also in Debian. I've also heard many people using various distributions describe something similar while using nvidia-driver. I'm convinced the problem is with nvidia-driver.

Now I'm using version 9631 of the nvidia-driver on Gentoo and 8776 (if I recall correctly) in Debian and haven't had a crash in ages. Uptime with these systems have been sometimes more than 10 days. I still haven't managed to make VDR fully work on Gentoo and thus have to boot do Debian if I need some specific features of it.

gw · Post by gw » Sun Jan 14, 2007 10:11 am

I have exactly the same problem AliasXZ describes: X hangs, no more keyboard input, mouse works, ssh access to machine is possible.

I'm using nvidia-drivers v. 1.0.9746 from portage and a compositing window manager (compiz).

The problem first occured after update-worlds of mid december 06.

I get these sort of messages whenever X freezes:

Code: Select all

Jan 12 18:44:45 linux [  990.539000] NVRM: Xid (0001:00): 6, PE0000 0ec0 00d0d4dc 0000f550 00eeeeef 00fcfcfc
Jan 12 18:44:45 linux [  990.556000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 12 18:46:52 linux [ 1117.564000] NVRM: Xid (0001:00): 6, PE0000 0ec0 00ffffff 00009984 00dfe2e8 00003366
Jan 12 18:46:52 linux [ 1117.581000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 12 18:50:51 linux [ 1356.438000] NVRM: Xid (0001:00): 1, Channel 00000000 Method 00000060 Data 0101020c
Jan 12 18:50:51 linux [ 1356.454000] NVRM: Xid (0001:00): 30,  L0 -> L0
Jan 13 12:21:06 linux [ 6048.104000] NVRM: Xid (0001:00): 8, Channel 00000000
Jan 13 12:21:06 linux [ 6048.120000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 13 12:22:48 linux [ 6150.151000] NVRM: Xid (0001:00): 1, Channel 00000000 Method 00000060 Data 0101020c
Jan 13 16:30:31 linux [11322.640000] NVRM: Xid (0001:00): 6, PE0000 08c0 04ad002e 000054e4 00000000 00000000
Jan 13 16:30:31 linux [11322.656000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 13 22:42:05 linux [33620.765000] NVRM: Xid (0001:00): 1, Channel 00000000 Method 00000060 Data 0101020c
Jan 13 22:42:05 linux [33620.777000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 14 10:29:08 linux [40575.107000] NVRM: Xid (0001:00): 6, PE0000 08c0 000f040b 0000f780 008d95a4 00000000
Jan 14 10:29:08 linux [40575.124000] NVRM: Xid (0001:00): 30,  L1 -> L0
Jan 14 10:35:12 linux [40938.352000] NVRM: Xid (0001:00): 1, Channel 00000000 Method 00000060 Data 0101020c
Jan 14 10:35:12 linux [40938.364000] NVRM: Xid (0001:00): 30,  L0 -> L0

Can anyone interprete these, is it X or nvidia related?

Thanks,

gw

AliasXZ · Post by **AliasXZ** » Sun Jan 14, 2007 1:17 pm

personally i think its to do with modular, when X was monotholic I had no problems at all...
-
That's weird, my freeze doesn't display anything in logs.
-
I've had this problem for ages and now I'm fed up so, I'm installing xorg-x11-6.9.0-r3 from a fresh install as I type this...

micko · Post by **micko** » Sun Jan 14, 2007 1:33 pm

AliasXZ wrote:personally i think its to do with modular, when X was monotholic I had no problems at all...

Are you sure? If you updated nvidia-drivers about the same time you updated to modular X... Some versions of the drivers crash more often than the others. I've had also the monolithic X crash and only thing seems to affect the stability is the nvidia-drivers and it's settings in Xorg.conf. Have you had those crashes with nv-driver, or what makes you suspect it isn't caused my nvidia-drivers?

AliasXZ · Post by **AliasXZ** » Sun Jan 14, 2007 1:45 pm

I'm not saying that it definatly isn't caused by nvidia-drivers, it's just my own personal opinion from what has happened on my machine, I remember when I first updated to modular, all was fine until a couple of hours doing stuff like browsing net etc. Then it did that "lock-up" thing... I wouldn't of really minded, but it happens randomly and for no apparent reason also I noticed some old games don't render right through modular like unreal tournament for example. I have tried nearly every single version of nvidia-drivers, I can't go above 8776 though, probably because my card is a bit dated. I have also tried the official nvidia drivers off the nvidia site, those wouldn't work though, probably because again, my card is a bit dated. Is it the same on Nvidia's site where you have to use some sort of legacy drivers for older cards or are Nvidia's official ones supposed to work with every card out of the box?
-
If I have trouble in 6.9 I guess I will be forced to use modular.
-
USE="GAPING_SECURITY_HOLE" lol!

halfgaar · Post by **halfgaar** » Sun Jan 14, 2007 3:20 pm

also I noticed some old games don't render right through modular like unreal tournament for example.

Unreal tournament, 99 and 2003, runs fine here on modular X. (GeForce 6600 LE)

I have tried nearly every single version of nvidia-drivers, I can't go above 8776 though, probably because my card is a bit dated. I have also tried the official nvidia drivers off the nvidia site, those wouldn't work though, probably because again, my card is a bit dated. Is it the same on Nvidia's site where you have to use some sort of legacy drivers for older cards or are Nvidia's official ones supposed to work with every card out of the box?

Nvidia provides two sets of drivers, legacy and current. Those two are available in portage; Gentoo doesn't do anything differently than what nvidia provides. It's possible that your card can't work with the current drivers. You could try the legacy ones.

Dieter@be · Post by **Dieter@be** » Sun Jan 14, 2007 3:30 pm

Can I join the club?
I experience the same behaviour, although sometimes i can't use the mouse either, or the screen is just entirely blank.
Usually this occurs when i do something like opening a big image in the gimp, or open firefox with lot's of pages in it, so it seems like it hangs on one of those processes, but if i login via ssh the seem is very, very responsive, top showed a very low cpu usage and i couldn't find any processes to kill that might keep my system busy.

This only started occurring since recently to me, although i've been using xorg 7.* already since quite a while. nvidia-drivers and dbus on the other hand are packages that probably got updated a short while ago

AliasXZ · Post by **AliasXZ** » Sun Jan 21, 2007 6:18 pm

hmmm happened yet again after a few days on a clean install!!!
-
ok, I've changed my nvidia drivers to 9631... Apparently there are fixes so we'll see...

AliasXZ · Post by **AliasXZ** » Tue Jan 23, 2007 5:10 pm

Nope, surprisingly it hasn't fixed the problem (yes that was sarcastic)

What the hell is causing this?!

halfgaar · Post by **halfgaar** » Tue Jan 23, 2007 5:52 pm

Do you have experience with debugging with GDB? Normally, starting a program with GDB and then asking for a backtrace is easy, but I'm not sure how to start the x-server through GDB. You might want to investigate it, because it could give you a better idea of what's going on. A good backtrace is also convenient for the programmer trying to solve the bug.

(Closed) Xorg 7.1 - Nvidia drivers (all versions) re-opened

(Closed) Xorg 7.1 - Nvidia drivers (all versions) re-opened

More likely caused by DBUS