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

Newegg Daily Deals

Reply
 
Thread Tools
Old 02-17-12, 11:45 AM   #1
stuartlevy
Registered User
 
Join Date: May 2003
Location: flat, steamy Champaign, IL
Posts: 36
Default Yay! 295.20 fixes strange Shake libc-related(?) bug

Wanted to report that you've recently (295.20) *fixed* a strange low-level problem (I/O? libc?), in case this helps others.

We're using an old 32bit Linux version of Shake (4.10.0606) on our 64bit machines. It worked fine until we installed nVidia 280.*/285.*/290.* drivers, and then began misbehaving. Its graphics worked normally, but FileOut image-writing failed bizarrely for a variety of image formats (at least TIFF/JPEG/PNG), writing either 0-length files, or (for PNG) 40+-byte files with a PNG header and no data.

No errors reported, no filesystems full. Reverting to older nVidia driver (e.g. 190.42) makes Shake work fine, but is only possible with distributions with old Xorg ABIs.

The failure happens reliably in a variety of environments: Fedora 9 x64, Linux Mint x64 (based on Ubuntu Oneiric), CentOS 6.2 x64 -- 32-bit-glibc versions from 2.8 to 2.13 -- with at least nVidia drivers 280.13, 285.05.09, and 290.10.

strace'ing didn't turn up anything suspicious. I suspect some low-level interference with libc (perhaps a colliding global variable?), but don't know.

Delighted to find that, with 295.20 nVidia driver, Shake again saves files normally. Haven't surveyed systematically to tell the oldest driver version where the problem appeared. Can't guess how this relates to any of the Changelogged changes.
stuartlevy is offline   Reply With Quote
Old 02-17-12, 03:52 PM   #2
stuartlevy
Registered User
 
Join Date: May 2003
Location: flat, steamy Champaign, IL
Posts: 36
Default Re: Yay! 295.20 fixes strange Shake libc-related(?) bug

Followup: on checking more versions, it appears
260.19.21 and before don't have the bug described above
280.13, 285.05.09, and 290.10 do have this bug
295.09 has the bug fixed already
295.20 still has it fixed.

Perhaps this is one of the "two bugs that caused sporadic application crashes in some multi-threaded OpenGL applications" mention in the 295.09 Changelog. For us, no crashes, but reliably bad data.
stuartlevy 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 02:44 PM.


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