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

Newegg Daily Deals

Reply
 
Thread Tools
Old 07-02-09, 09:40 AM   #1
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

I am getting a graphics corruption ever since the 18x.x drivers. I tried 185.18.14 today and it is still there. This didn't appear in the 177.x drivers.

This is how the xine control window should look like and what I get in 17x.x:



This is what I get in 18x.x:




This only applies to 32 bit binaries running in a 64 bit OS. But as I said, it worked fine until 18x.x. I can take the same literal xine install and same machine, just upgrade the NVidia drivers from 17x.x to 18x.x and it breaks. I continues to work for the 64 bit xine.

The problem appears both when running the 32 bit binary inside and outside a chroot.

Is it possible that one of the 32 bit libraries installed by the NVidia drivers is screwed up?

%%

In case this can't be solved:

I was happy with 17x.x but given kernel 2.6.28.7 it won't locate the kernel sources anymore (even if I give all the environment variables, e.g. SYSSRC). Is there a patch available to compile 177 on that kernel?

%%

Maybe I can try installing just the 32 bit libs from the 17x.x and have the 18x.x drivers in the kernel and in the 64 bit userland?

Any other creative solutions?

Thank you.
Attached Files
File Type: gz nvidia-bug-report.log.gz (30.8 KB, 85 views)
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 07-02-09, 09:47 AM   #2
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

Hmmm, doesn't look like the NVidia userland can possibly be responsible for this:

ldd /usr/local32/bin/xine:
linux-gate.so.1 => (0xffffe000)
libXv.so.1 => /emul/ia32-linux/usr/lib/libXv.so.1 (0xf7f7d000)
libpthread.so.0 => /lib32/libpthread.so.0 (0xf7f66000)
libSM.so.6 => /emul/ia32-linux/usr/lib/libSM.so.6 (0xf7f5e000)
libICE.so.6 => /emul/ia32-linux/usr/lib/libICE.so.6 (0xf7f46000)
libXinerama.so.1 => /usr/local32/lib/libXinerama.so.1 (0xf7f43000)
libXxf86vm.so.1 => /emul/ia32-linux/usr/lib/libXxf86vm.so.1 (0xf7f3e000)
libpng12.so.0 => /emul/ia32-linux/usr/lib/libpng12.so.0 (0xf7f1b000)
libxine.so.1 => /usr/local32/lib/libxine.so.1 (0xf7ed2000)
libz.so.1 => /usr/lib32/libz.so.1 (0xf7ebc000)
librt.so.1 => /lib32/librt.so.1 (0xf7eb3000)
libresolv.so.2 => /lib32/libresolv.so.2 (0xf7ea0000)
libnsl.so.1 => /lib32/libnsl.so.1 (0xf7e88000)
libdl.so.2 => /lib32/libdl.so.2 (0xf7e84000)
libXext.so.6 => /emul/ia32-linux/usr/lib/libXext.so.6 (0xf7e76000)
libXtst.so.6 => /emul/ia32-linux/usr/lib/libXtst.so.6 (0xf7e70000)
libX11.so.6 => /emul/ia32-linux/usr/lib/libX11.so.6 (0xf7d81000)
libm.so.6 => /lib32/libm.so.6 (0xf7d5d000)
libc.so.6 => /lib32/libc.so.6 (0xf7c0b000)
/lib/ld-linux.so.2 (0xf7fac000)
libXau.so.6 => /usr/lib32/libXau.so.6 (0xf7c08000)
libxcb-xlib.so.0 => /usr/lib32/libxcb-xlib.so.0 (0xf7c05000)
libxcb.so.1 => /usr/lib32/libxcb.so.1 (0xf7bed000)
libXdmcp.so.6 => /usr/lib32/libXdmcp.so.6 (0xf7be8000)
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 07-02-09, 02:14 PM   #3
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

Looks like this kernel and NVidia combo has more problems:

Code:
Jul  2 10:26:28 wavehh kernel: BUG: using smp_processor_id() in preemptible [00000000] code: Xorg/7507
Jul  2 10:26:28 wavehh kernel: caller is paravirt_get_lazy_mode+0x21/0x40
Jul  2 10:26:28 wavehh kernel: Pid: 7507, comm: Xorg Tainted: P           2.6.28.7-perfctr #3
Jul  2 10:26:28 wavehh kernel: Call Trace:
Jul  2 10:26:28 wavehh kernel:  [<ffffffff80454ebb>] debug_smp_processor_id+0xdb/0xe0
Jul  2 10:26:28 wavehh kernel:  [<ffffffff8022d841>] paravirt_get_lazy_mode+0x21/0x40
Jul  2 10:26:28 wavehh kernel:  [<ffffffff80234f04>] change_page_attr_set_clr+0x124/0x3a0
Jul  2 10:26:28 wavehh kernel:  [<ffffffff802353dd>] _set_memory_uc+0x1d/0x30
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa08eb0a8>] nv_vm_malloc_pages+0x208/0x350 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa08e7c8b>] nv_alloc_pages+0x21b/0x490 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa07bc1b1>] _nv011901rm+0x3f/0x54 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa080c261>] _nv005266rm+0x44/0x49 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa0776d68>] _nv008160rm+0x98/0x16d [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa0776ea5>] _nv008161rm+0x4c/0x1c0 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa0792f28>] _nv003884rm+0x182/0x21a [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa079323d>] _nv003894rm+0x27d/0x68d [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa0793103>] _nv003894rm+0x143/0x68d [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa078f951>] _nv007107rm+0x77/0xbd [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa078f92a>] _nv007107rm+0x50/0xbd [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa05a41f5>] _nv003716rm+0x145/0x5b4 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa0814581>] rm_ioctl+0x2f/0x67 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa08e877a>] nv_kern_ioctl+0x1ca/0x470 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffffa08e8a6c>] nv_kern_unlocked_ioctl+0x1c/0x30 [nvidia]
Jul  2 10:26:28 wavehh kernel:  [<ffffffff802d19af>] vfs_ioctl+0x2f/0xb0
Jul  2 10:26:28 wavehh kernel:  [<ffffffff802d1aae>] do_vfs_ioctl+0x7e/0x480
Jul  2 10:26:28 wavehh kernel:  [<ffffffff802d1f51>] sys_ioctl+0xa1/0xb0
Jul  2 10:26:28 wavehh kernel:  [<ffffffff802134ca>] system_call_fastpath+0x16/0x1b
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 07-06-09, 03:10 PM   #4
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

I hacked up the 177.82 drivers to compile with kernel 2.6.30.1.

I didn't test it yet but just as a heads-up for others who have problems with the 18x.xx NVidia drivers, I have a diff for you.
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 07-09-09, 01:50 PM   #5
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

Bump.

I should probably try the xine mailing list.

Is it really the case that nobody has ever seen this kind of corruption before?

Am I the only one using 32 bit apps like that in a 64 bit NVidia X11 server?
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 07-25-09, 03:45 PM   #6
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

190.x has the same problem, except that even less pixels are visible.

I noticed that the working 64 bit Xine has the top window manager decoration (not not the sides or bottom). Maybe the kind of window that is being created here has to do with the problem?
__________________

My Unix benchmark results
uOpt is offline   Reply With Quote
Old 09-22-09, 09:11 PM   #7
AaronP
NVIDIA Corporation
 
AaronP's Avatar
 
Join Date: Mar 2005
Posts: 2,487
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

I just built a 32-bit copy of Ubuntu's version of Xine (xine-lib 1.1.16.3 and xine-ui-0.99.5+cvs20070914) and it works fine. There must be something else wrong. What else are you running? Does the problem go away if you use a kernel without paravirtualization?
AaronP is offline   Reply With Quote
Old 09-23-09, 09:36 PM   #8
uOpt
FreeBSD cheering section
 
uOpt's Avatar
 
Join Date: Nov 2004
Location: Boston, MA, USA
Posts: 609
Default Re: What would cause this kind of graphics corruption in 18x.x (not in 17x.x)

Did you run this in a 32 bit chroot or inside the regular environment? I noticed that a chroot 32-bit xine works for me, too. It's the one outside that is affected (a /usr/local32 tree).

I went forward to 185.18.36, which still broke it with serverflage to have SHM pixmaps off and back to 177.x fixed it again.

To make more sure there is no userland change from the NVidia drivers involved I md5ed all shared libs that are used by my xine according to ldd. There was no change between the 185 and 177 installation (my xine is not bound to libGL).
__________________

My Unix benchmark results
uOpt 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 11:10 AM.


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