Bug 4978 - Failed to allocate mem resource for AGP bridge
Failed to allocate mem resource for AGP bridge
Status: REJECTED INVALID
Product: Drivers
Classification: Unclassified
Component: PCI
i386 Linux
: P2 normal
Assigned To: Greg Kroah-Hartman
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2005-08-01 08:41 UTC by Ortwin Glück
Modified: 2005-09-07 13:38 UTC (History)
0 users

See Also:
Kernel Version: 2.6.13-rc2, 2.6.13-rc3, 2.6.13-rc4, 2.6.13-rc5, 2.6.13-rc6
Tree: Mainline
Regression: ---


Attachments
dmesg (15.07 KB, text/plain)
2005-08-01 08:41 UTC, Ortwin Glück
Details
lspci (1.25 KB, text/plain)
2005-08-01 08:42 UTC, Ortwin Glück
Details
dmesg-2.6.13-rc5 (12.81 KB, text/plain)
2005-08-04 15:24 UTC, Ortwin Glück
Details

Description Ortwin Glück 2005-08-01 08:41:21 UTC
I see the line
PCI: Failed to allocate mem resource #6:10000@f0000000 for 0000:01:00.0
on boot. The nvidia binary driver does not work then.
The message does not appear with 2.6.12.2 and the nvidia driver works.

I attach dmesg and lspci output.
Comment 1 Ortwin Glück 2005-08-01 08:41:58 UTC
Created attachment 5453 [details]
dmesg
Comment 2 Ortwin Glück 2005-08-01 08:42:26 UTC
Created attachment 5454 [details]
lspci
Comment 3 Greg Kroah-Hartman 2005-08-04 13:11:52 UTC
Can you test 2.6.13-rc5 and see if this is fixed there?
Comment 4 Ortwin Glück 2005-08-04 15:21:19 UTC
Same effect with 2.6.13-rc5
Comment 5 Ortwin Glück 2005-08-04 15:24:14 UTC
Created attachment 5508 [details]
dmesg-2.6.13-rc5
Comment 6 Ortwin Glück 2005-08-15 12:02:29 UTC
Okay, I tried all 2.6.13-rc releases. All but rc1 show the mem resource failure.
So the problem came in between rc1 and rc2.
The nVidia driver does not work with rc1 either, so I suppose this has nothing
(or not much) to do with each other. I'll try and isolate the patch that causes
the problem.
Comment 7 Ortwin Glück 2005-08-15 12:58:13 UTC
Backing out this patch
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;h=70bcd53451f68873a9bfaa14a07ecf6142ea94f9;hp=87325263cd4fbcdeb29994e69e96550b39981621;hb=299de0343c7d18448a69c635378342e9214b14af;f=arch/i386/pci/common.c
removes the warning. Maybe Ivan has a clue what is going on here.

nVidia now complains it can not find the device...
Comment 8 Greg Kroah-Hartman 2005-08-18 00:26:23 UTC
There's nothing we can do if the binary nvidia driver breaks, sorry.  Please
contact them.
Comment 9 Ortwin Glück 2005-08-18 01:08:43 UTC
Greg,

I don't care about the nV driver. The original problem is the warning issued by
the kernel. I want to know if this warning is serious and points to a hardware
support problem that needs to be resolved. Or is it rather an informational
message that does not cause any harm. Please clarify.

Ortwin
Comment 10 Greg Kroah-Hartman 2005-08-18 01:39:47 UTC
I'd like to say it's just informational, and can be safely ignored, but I'm not
really sure.  Some machines need that patch and don't work without it.  Others
(like yours and bug #4965) seem to not like it at all.

But, as you are dealing with the binary nvidia driver, there's really nothing I 
can do to help out, sorry.  Try asking nvidia if this is an issue or not.
Comment 11 Ortwin Glück 2005-09-07 13:38:47 UTC
Never mind, the nvidia driver issue is something else (EXPORT_SYMBOL_GPL in
sysfs). See http://bugs.gentoo.org/show_bug.cgi?id=104369

Note You need to log in before you can comment on or make changes to this bug.