nV News Forums

 
 

nV News Forums (http://www.nvnews.net/vbulletin/index.php)
-   NVIDIA Linux (http://www.nvnews.net/vbulletin/forumdisplay.php?f=14)
-   -   [BUG] Ioremap error with nvidia in 2.6.38-rc6 (http://www.nvnews.net/vbulletin/showthread.php?t=160186)

rprabhu 03-02-11 01:26 AM

[BUG] Ioremap error with nvidia in 2.6.38-rc6
 
When testing 2.6.38-rc6, while starting xorg I noticed following


===========================================
Code:

Feb 27 21:53:12 localhost kernel: [  622.957621] ioremap error for 0xbbffe000-0xbbfff000, requested 0x10, got 0x0
Feb 27 21:53:12 localhost kernel: [  622.957626] ioremap error for 0xbbffe000-0xbbfff000, requested 0x10, got 0x0
Feb 27 21:53:12 localhost kernel: [  622.957631] ioremap error for 0xbbffe000-0xbbfff000, requested 0x10, got 0x0
Feb 27 21:53:12 localhost kernel: [  622.957636] ioremap error for 0xbbffe000-0xbbfff000, requested 0x10, got 0x0

=======================================

Going by the kernel source, I noticed that request and response are differing over the cacheability of the memory region.

Sifting through the kernel git, I noticed

http://git.kernel.org/?p=linux/kerne...69e069dc2e0b8d

So, this looks like the culprit.

For Intel 9 code, they have made following change,

http://git.kernel.org/?p=linux/kerne...c62f65e0851fd3

Now, looking at the nvidia module source code -- os-interface.c I observed call to os_map_kernel_space being made. The bug may be in calling with wrong mode for opregion (the term used in above commits)

Edit:
Is changing NV_IOREMAP_NOCACHE(vaddr, start, size_bytes) to NV_IOREMAP(vaddr, start, size_bytes) for NV_MEMORY_DEFAULT going to fix this ?

oyvind 03-02-11 07:26 AM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
I'm seeing these errors as well, using nvidia 270.29 + 2.6.38-rc6-based kernel (built for Lucid x86-64 by Ubuntu kernel PPA).

[ 10.560094] ioremap error for 0xbf67c000-0xbf67d000, requested 0x10, got 0x0
[ 10.560127] ioremap error for 0xbf76d000-0xbf76e000, requested 0x10, got 0x0
[ 10.560130] ioremap error for 0xbf76d000-0xbf76e000, requested 0x10, got 0x0
[ 10.560133] ioremap error for 0xbf76d000-0xbf76e000, requested 0x10, got 0x0
[ 10.560136] ioremap error for 0xbf76d000-0xbf76e000, requested 0x10, got 0x0
[ 10.560139] ioremap error for 0xbf766000-0xbf767000, requested 0x10, got 0x0
[ 10.560142] ioremap error for 0xbf74d000-0xbf74e000, requested 0x10, got 0x0

gilboa 03-13-11 04:15 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
Seeing the same, 260.19.36, 2.6.38rc6.
Funny thing though, I'm seeing this on an AMD machine, but the same driver/kernel combo works just fine on a Xeon workstation.

rprabhu,

Have you had any luck finding which io*remap call fails?

- Gilboa

gilboa 03-13-11 04:20 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
FWIW:

$ cat /var/log/messages | grep ioremap
Mar 13 09:21:13 gilboa-wx-dev kernel: [ 93.732328] ioremap error for 0xbfcf3000-0xbfcf4000, requested 0x10, got 0x0
Mar 13 22:53:26 gilboa-wx-dev kernel: [48826.547871] ioremap error for 0xbfcf3000-0xbfcf4000, requested 0x10, got 0x0
$ cat /proc/iomem | grep bfcf3000
bfcf3000-bfcfffff : ACPI Tables

rprabhu 03-14-11 10:27 AM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
Quote:

Originally Posted by gilboa (Post 2404388)
Seeing the same, 260.19.36, 2.6.38rc6.
Funny thing though, I'm seeing this on an AMD machine, but the same driver/kernel combo works just fine on a Xeon workstation.

rprabhu,

Have you had any luck finding which io*remap call fails?

- Gilboa

It is not related to Intel in anyway. I just quoted the Intel commit to show similarity. You can look at os_map_kernel_space function in os-interface.c to see the callpoints. I think it may require a fix in the closed part of the driver. Also, related to your latter comment, yes, it is related to ACPI (check the other commit in my comment which caused this).

rprabhu 03-17-11 01:50 AM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
Has this bug been fixed ? It seems to be crashing the system last time I used (which was rc8). This is important since .38 has been released and distro packagers will be blocked on this.

zander 03-17-11 10:31 AM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
A bug has been filed and we'll address the problem in a future release. It shouldn't cause any crashes - the mapping request's memory type is modified to match the kernel's mappings' memory type.

rprabhu 03-17-11 02:24 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
Quote:

Originally Posted by zander (Post 2406309)
A bug has been filed and we'll address the problem in a future release. It shouldn't cause any crashes - the mapping request's memory type is modified to match the kernel's mappings' memory type.

Good to hear that, thanks. I had crash only once with it (of course i switched back to 37 after crash). Another place (and the reason why I presumed the crash was due to it) it caused a nuisance of sorts was when using OSSv4 for sound, it was causing a lot of crackling noise in the speaker while playing music, with ALSA I have had no such issues though (but crash happened when using ALSA).

rprabhu 03-19-11 02:26 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
1 Attachment(s)
Edit: Correction, The crash was not related to nvidia, but I have seen those Xid messages in my dmesg nevertheless.


My laptop crashed again (hard crash, no sysrq possible, so most probably graphics related).

This may be related

Code:

Mar 19 22:34:35 localhost kernel: [39218.242039] NVRM: Xid (0000:02:00): 56, CMDre 00000003 000000c0 b0ef122d 00000007 00000000
Mar 19 22:34:35 localhost kernel: [39218.242054] NVRM: Xid (0000:02:00): 56, CMDre 00000004 000000c0 b0ef122d 00000007 00000000

Above message was in my kernel log. That looks to be related to nvidia.

I am also attaching a bug report generated from nvidia-bug-report.

gilboa 03-21-11 05:50 AM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
Thus far, no crashes on my side - only very slow performance. (DRI is enabled)

- Gilboa

Sky777 05-15-11 02:48 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
1 Attachment(s)
with 2.6.38 and 270.41.06 I also have the

Quote:

[ 31.774550] ioremap error for 0xffb0000-0xffb1000, requested 0x10, got 0x0
may be due to this my usb wireless keyboard doesn't work

danix 05-20-11 04:23 PM

Re: [BUG] Ioremap error with nvidia in 2.6.38-rc6
 
We fixed some issues that we believe to be related to this in 270.41.19 and 275.09.


All times are GMT -5. The time now is 06:36 AM.

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