Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 06-19-10, 11:53 AM   #1
nusch
Registered User
 
Join Date: Feb 2009
Posts: 7
Default 100% CPU usage on geforce 9600M GS

I have problems with stability of this driver since purchase laptop with nvidia. Tried various drivers, through long time always the newest beta from your ftp, now sticked to those from ubuntu repo, currently
195.36.24-0ubuntu1~10.04. Today again Xorg hanged due to nvidia driver, losing few hours of my work. Will this driver be ever in 50% as stable as Windows version? Although it happens once per few days I must save everything every 10minutes, can't left computer overnight or use clock alarm from KDE because I never sure if it will not hang kernel even doing nothing.

Here is a log from today's hang - kernel responsive, one core eaten by Xorg, KDE freezed not responding to keyboard, mouse except cursor moving, logged via ssh and attached gdb to Xorg process:
Code:
(gdb) backtrace full
#0  0x00007f2f0680813d in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#1  0x00007f2f0680997d in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#2  0x00007f2f06866c89 in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#3  0x00007f2f06ad3393 in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#4  0x00007f2f06ad1a8a in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#5  0x00007f2f06ad359c in ?? () from /usr/lib/xorg/extra-modules/nvidia_drv.so
No symbol table info available.
#6  0x00000000004da5fc in damagePutImage (pDrawable=0x3d1fc00, pGC=0x6b82810,
    depth=<value optimized out>, x=0, y=0, w=131969032, h=12, leftPad=0,
    format=2, pImage=0x7ddb008 "\240\240\240")
    at ../../../miext/damage/damage.c:905
        pGCPriv = 0x73648f0
        oldFuncs = 0x7c9120
#7  0x000000000042e8e9 in ProcPutImage (client=0x2bb9e90)
    at ../../dix/dispatch.c:2067
        pGC = 0x6b82810
        pDraw = 0x3d1fc00
        length = <value optimized out>
#8  0x0000000000430c3c in Dispatch () at ../../dix/dispatch.c:439
        result = <value optimized out>
        client = 0x2bb9e90
        nready = 0
        start_tick = 1252580
#9  0x00000000004261aa in main (argc=8, argv=0x7ddbe8, envp=<value optimized out>)
    at ../../dix/main.c:285
        i = 1
        alwaysCheckForInput = {0, 1}
Where is the problem ? Funds, lack of debug or developers ?
Why you advising adding different options to kernel which limit functionality or interfere HPET what only decrease probability of deadlock not eliminating it, instead of debug and fix the issue?
nusch is offline   Reply With Quote
Old 06-19-10, 03:06 PM   #2
AaronP
NVIDIA Corporation
 
AaronP's Avatar
 
Join Date: Mar 2005
Posts: 2,487
Default Re: 100% CPU usage on geforce 9600M GS

Quote:
Originally Posted by nusch View Post
Code:
(gdb) backtrace full
[...]
#6  0x00000000004da5fc in damagePutImage (pDrawable=0x3d1fc00, pGC=0x6b82810,
    depth=<value optimized out>, x=0, y=0, w=131969032, h=12, leftPad=0,
    format=2, pImage=0x7ddb008 "\240\240\240")
    at ../../../miext/damage/damage.c:905
        pGCPriv = 0x73648f0
        oldFuncs = 0x7c9120
#7  0x000000000042e8e9 in ProcPutImage (client=0x2bb9e90)
    at ../../dix/dispatch.c:2067
        pGC = 0x6b82810
        pDraw = 0x3d1fc00
        length = <value optimized out>
#8  0x0000000000430c3c in Dispatch () at ../../dix/dispatch.c:439
        result = <value optimized out>
        client = 0x2bb9e90
        nready = 0
        start_tick = 1252580
#9  0x00000000004261aa in main (argc=8, argv=0x7ddbe8, envp=<value optimized out>)
    at ../../dix/main.c:285
        i = 1
        alwaysCheckForInput = {0, 1}
How long did you wait? This doesn't list the depth, but even at just depth 8, this is an upload of 1.47 GB to the graphics card. It's possible that the optimization level of the X server makes that parameter unreliable... If you use GDB's "finish" command, does ProcPutImage ever complete?

Also, please see the forum sticky post about how to generate a bug report log file.
AaronP is offline   Reply With Quote
Old 06-20-10, 07:33 AM   #3
nusch
Registered User
 
Join Date: Feb 2009
Posts: 7
Default Re: 100% CPU usage on geforce 9600M GS

Attached nvidia-bug-report, but it doesn't contain informations about bug itself because I've run it while ago and rebooted computer since the crash. I didn't try finish command and I don't know if it locks exactly at that point, but for sure some subset of calls is executed in a loop. I'm experiencing often such hangs that screen is frozen - sometimes with cursor sometimes without, and after minute return to responsible state. This time I waited about 20 minutes, then forced kdm restart.

Is the 1.47 GB normal data transfer between CPU and GPU? It's qute big for me and my GPU doesn't have such ammount of RAM. During the crash I had desktop effects in KDE enabled although didn't have any graphic intensive programs enabled: Chromium with ~20 tabs, firefox with 5 tabs, VirtualBox, IM, and Amarok. I'm only guesing but this crash could be causes by simultaneous screen updates in different regions - e.g Amarok OSD saying about new song, IM about new message and some internal KDE notification. While not using one of them, I'm experiencing fewer crashes.
Also long playing some flash powered web content (especially fullscreen) causes flickering or hangs(not adobe plugin but whole X or kernel)
Attached Files
File Type: gz nvidia-bug-report.log.gz (57.6 KB, 60 views)
nusch is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 09:55 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.