Bug 16515 - [bisected] Radeon rv280 can't boot on kernel 2.6.35.
[bisected] Radeon rv280 can't boot on kernel 2.6.35.
Status: RESOLVED OBSOLETE
Product: Drivers
Classification: Unclassified
Component: Video(AGP)
x86-64 Linux
: P1 normal
Assigned To: Dave Airlie
:
Depends on:
Blocks: 16055
  Show dependency treegraph
 
Reported: 2010-08-04 16:10 UTC by Albert Gall
Modified: 2012-08-09 15:29 UTC (History)
6 users (show)

See Also:
Kernel Version: 2.6.35
Tree: Mainline
Regression: Yes


Attachments
git bisect 2.6.35 bad 2.6.34 god (3.28 KB, text/x-log)
2010-08-04 16:10 UTC, Albert Gall
Details
lspci -vv (1.77 KB, text/plain)
2010-08-04 16:14 UTC, Albert Gall
Details
dmesg (48.68 KB, text/plain)
2010-08-04 16:15 UTC, Albert Gall
Details
radeontool regmatch '*' > working.dump on kernel 2.6.35 after git revert (19.91 KB, application/octet-stream)
2010-08-04 18:52 UTC, Albert Gall
Details

Description Albert Gall 2010-08-04 16:10:46 UTC
Created attachment 27350 [details]
git bisect 2.6.35 bad 2.6.34 god

Radeon agp rv280 can't boot on kernel 2.6.35 using dvi port, on vga port works as
well, on kernel 2.6.34 work fine all ports.
I try to revert this patch and post result here to night.
Comment 1 Albert Gall 2010-08-04 16:14:32 UTC
Created attachment 27351 [details]
lspci -vv
Comment 2 Albert Gall 2010-08-04 16:15:23 UTC
Created attachment 27352 [details]
dmesg
Comment 3 Albert Gall 2010-08-04 18:52:53 UTC
Created attachment 27353 [details]
radeontool regmatch '*' > working.dump on kernel 2.6.35 after git revert
Comment 4 Alex Deucher 2010-08-04 19:48:00 UTC
Can you attach a a broken dump as well (2.6.35 without the revert)?  Also, please only provide dumps with module option modeset=1.
Comment 5 Albert Gall 2010-08-04 20:15:55 UTC
(In reply to comment #4)
> Can you attach a a broken dump as well (2.6.35 without the revert)?  Also,
> please only provide dumps with module option modeset=1.

No, on 2.6.35 without the revert always get a kernel panic or hard reboot and console not shows anything.
Comment 6 Albert Gall 2010-08-04 20:23:48 UTC
The first dump is good and have module option modeset=1 by default.
Comment 7 Rafael J. Wysocki 2010-08-04 23:05:23 UTC
Handled-By : Alex Deucher <alexdeucher@gmail.com>
Comment 8 Alex Deucher 2010-08-16 16:02:24 UTC
Does booting with radeon.agpmode=-1 help?  This might be related to bug 16581 or bug 16574.  Can you try the patches attached to those bugs?
Comment 9 Albert Gall 2010-08-17 18:50:46 UTC
(In reply to comment #8)
> Does booting with radeon.agpmode=-1 help?  This might be related to bug 16581
> or bug 16574.  Can you try the patches attached to those bugs?


The option radeon.agpmode=-1 does not work and the patches do not solve the problem, the result is always the same as you know.
Comment 10 Rafael J. Wysocki 2010-09-01 22:17:24 UTC
On Wednesday, September 01, 2010, Johan Kullstam wrote:
> "Rafael J. Wysocki" <rjw@sisk.pl> writes:
> 
> > This message has been generated automatically as a part of a report
> > of regressions introduced between 2.6.34 and 2.6.35.
> >
> > The following bug entry is on the current list of known regressions
> > introduced between 2.6.34 and 2.6.35.  Please verify if it still should
> > be listed and let the tracking team know (either way).
> >
> >
> > Bug-Entry	: http://bugzilla.kernel.org/show_bug.cgi?id=16515
> > Subject		: Radeon rv280 can't boot on kernel 2.6.35.
> > Submitter	: Albert Gall <ss3vdr@gmail.com>
> > Date		: 2010-08-04 16:10 (26 days old)
> > Handled-By	: Alex Deucher <alexdeucher@gmail.com>
> 
> This bug was introduced with 2.6.35-rc1 and continues through
> 2.6.36-rc3.  It is still a bug.  I cannot boot with modeset=1 on DVI
> port (not sure about VGA port status).
Comment 11 Florian Mickler 2010-09-08 12:30:50 UTC
First-Bad-Commit: https://bugzilla.kernel.org/attachment.cgi?id=27350


6b8b1786a8c29ce6e32298b93ac8d4a18a2b11c4 is the first bad commit
commit 6b8b1786a8c29ce6e32298b93ac8d4a18a2b11c4
Author: Jerome Glisse <jglisse@redhat.com>
Date:   Wed Apr 7 10:21:31 2010 +0000

    drm/radeon/kms: enable use of unmappable VRAM V2
    
    This patch enable the use of unmappable VRAM thanks to
    previous TTM infrastructure change.
    
    V2 update after io_mem_reserve/io_mem_free callback balancing
    
    Signed-off-by: Jerome Glisse <jglisse@redhat.com>
    Signed-off-by: Dave Airlie <airlied@redhat.com>
Comment 12 Florian Mickler 2010-09-30 05:49:34 UTC
Ping?
Comment 13 Alex Deucher 2010-09-30 14:37:38 UTC
How much vram does your card actually have (not what's reported in dmesg)?  Is it 128 MB or 256 MB?  Also in comment 1, are you saying that vga works fine but dvi does not or that neither port works?
Comment 14 Albert Gall 2010-09-30 17:23:14 UTC
(In reply to comment #13)
> How much vram does your card actually have (not what's reported in dmesg)?  Is
> it 128 MB or 256 MB?  Also in comment 1, are you saying that vga works fine but
> dvi does not or that neither port works?

I think my video card have 128MB of video ram, but I'm not really sure.
Xorg.log show on UMS: 

[    45.948] (--) RADEON(0): Mapped VideoRAM: 131072 kByte (128 bit DDR SDRAM)

and lspci:

01:00.0

Region 0: Memory at e8000000 (32-bit, prefetchable) [size=128M]
Region 1: I/O ports at c000 [size=256]
Region 2: Memory at ff4f0000 (32-bit, non-prefetchable) [size=64K]
Expansion ROM at ff4c0000 [disabled] [size=128K]

01:00.1

Region 0: Memory at e0000000 (32-bit, prefetchable) [size=128M]
Region 1: Memory at ff4e0000 (32-bit, non-prefetchable) [size=64K]

VGA to VGA work fine here but DVI to HDMI no.
Comment 15 Albert Gall 2010-09-30 17:38:56 UTC
This is the card:

http://www.amazon.co.uk/Club-3D-Radeon-9250-Graphics/dp/B000IOM5TW

This shows 256MB DDR.
Comment 16 Albert Gall 2011-01-04 21:23:34 UTC
I test git kernel 2.6.37-rc8 and bug persist.
Comment 18 Albert Gall 2012-03-14 22:32:52 UTC
(In reply to comment #17)
> Can you try these two patches:
> http://lists.freedesktop.org/archives/dri-devel/2012-March/020070.html
> http://lists.freedesktop.org/archives/dri-devel/2012-March/020071.html

Now I have new hardware and my old pc is in pieces when I have some time I'll try.

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