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

Newegg Daily Deals

Reply
 
Thread Tools
Old 03-13-09, 03:25 AM   #1
nevion
Registered User
 
Join Date: Dec 2004
Posts: 74
Default 180.37 xorg still bleeding memory [NVDevs, have a look]

Kde 4.2, new system, same oooold problem.

I'd really like to resolve these memory leaks. Please help me find out where the memory goes. Despite 12 gb of ram and a core i7 920, performance is dragging after 2 days. xorg restart is the only cure and time the xorg memory is freed.

Please have a look at the attached files (/proc/2951/smaps and maps, 2951=X pid).
Also, the desktop shot shows I have no apps claiming more than 120 megs of x memory (xrestop) and that xorg is the process dominating memory consumption. The heap is a gigabyte large looks like everything is referenced still according to smaps.txt. Then there's a number of sets adding up to the remaining memory that I can't account for at all.

screen shot
Attached Files
File Type: gz nvidia-bug-report.log.gz (34.9 KB, 70 views)
File Type: zip procmaps.zip (11.4 KB, 81 views)
nevion is offline   Reply With Quote
Old 03-13-09, 10:00 AM   #2
brainkilla
Registered User
 
Join Date: Mar 2008
Posts: 14
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

Excuse the off-topic nature of my question, but where can I get your .conkyrc? Thanx
brainkilla is offline   Reply With Quote
Old 03-13-09, 11:02 AM   #3
nullack
Registered User
 
Join Date: Oct 2007
Posts: 74
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

How do you know its not x itself? I saw some commits fixing leaks. What version of x arre you on?
nullack is offline   Reply With Quote
Old 03-13-09, 11:12 AM   #4
brucecadieux
Registered User
 
Join Date: Dec 2008
Posts: 78
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

nevion.

The first couple lines of your log show that you are sill using 180.22

You also have a DoS attack hitting you perhaps you should look into firewall settings
brucecadieux is offline   Reply With Quote
Old 03-13-09, 03:49 PM   #5
nevion
Registered User
 
Join Date: Dec 2004
Posts: 74
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

Look lower in my logs, I'm using 180.37 (I started upon boot with 180.22) - you can use more than one driver per boot if you know how to restart x :-).

As the logs say, xorg server 1.5.2 (opensuse package xorg-x11-server-7.4-17.3). I do not know if it is x or not but I don't see anyone complaining loudly about it. I've had such leaks for a long time too. I'm curious what makes you think I'm being Dos'd.

Also, that's not conky, it's superkaramba and a customized version of acrylicmonitor... attached.
nevion is offline   Reply With Quote
Old 03-13-09, 09:23 PM   #6
nevion
Registered User
 
Join Date: Dec 2004
Posts: 74
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

*Bump*

AaronP - please comment on this, I'm not familiar with x/lowerlevels enough to know who the culprit might be. You clue me in on what might be going on and I'll bug report the responsible parties. I need something to go on so they don't "stfu proprietary nvidia driver user!".
nevion is offline   Reply With Quote
Old 03-13-09, 11:24 PM   #7
nullack
Registered User
 
Join Date: Oct 2007
Posts: 74
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

Nevion has a look at the commits between your old version of X and revision 1.6. Note the mem leak fixes.
nullack is offline   Reply With Quote
Old 03-14-09, 12:08 AM   #8
nevion
Registered User
 
Join Date: Dec 2004
Posts: 74
Default Re: 180.37 xorg still bleeding memory [NVDevs, have a look]

I had a look at the changelog here and didn't see much concerning leaks and no one complaining about 1.5.2 leaking really but those who do seem to have a correlation with nvidia driver (reports are closed with that being the overall assumption of cause). 1.6 of xorg is a no go for me as I don't want to break the world here on my opensuse 11.1 box. And security vulnerabilities/memleaks would've been backported (if possible) by the opensuse x team which system updates would've picked up. In short don't think its an xorg problem, at least not exactly.

I have read that the resources not showing up in xrestop means they are 3d textures allocated by the driver. This means either kde/kwin/plasma (again, no intel/ati users are complaining about this problem) or that nvidia's glx stuff is wonky still. If they start using GEM for memory management it'd probably make this alot easier to find the culprit but until then, the memory has a source of being allocated by nvidia driver given that its 3d/texture memory.

Btw - my usage habbits do not change and I have no clue how I make it leak, it seems to just happen at some point, stop for a long time and then get bigger again. Firefox (w/ flash = opengl) /mplayer (vo=xv), text editing with gvim, image viewing (with gqview), konsoles, and chatting with kopete are all I really do, no fancy desktop effects to slow me down although I do have compositing enabled.
nevion 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


Similar Threads
Thread Thread Starter Forum Replies Last Post
173.14.34 -- Xorg 1.12.1 & 1.11.4, glx module fails to load mereset NVIDIA Linux 17 06-10-12 01:24 AM
Measuring card memory usage peter_ga NVIDIA Linux 0 05-24-12 08:07 PM

All times are GMT -5. The time now is 06:43 AM.


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