DRM regression (cvs commit: src/sys/dev/drm drm-preprocess.sh...)

[Dmitry Pryanishnikov]

Hello!

On Tue, 11 Apr 2006, Eric Anholt wrote:
>>>  Update to DRM CVS as of 2006-04-09.  The most notable new feature is the
>>
>>   This commit has broken DRM on my notebook ASUS M5A (i915GM graphics). drm
>> no longer attaches; here is relevant diff of old (OK) and new (failed)
>> verbose dmesg:
>>
>>   agp0: detected 7932k stolen memory
>>   agp0: aperture size is 256M
>>   drm0: <Intel i915GM> on vgapci0
>> -info: [drm] AGP at 0xfeb80000 0MB
>> -info: [drm] Initialized i915 1.2.0 20041217
>> +error: [drm:pid0:drm_load] *ERROR* Card isn't AGP, or couldn't initialize
>>  	AGP.
>> +device_attach: drm0 attach returned 12
>>   vgapci1: <VGA-compatible display> mem 0xfea80000-0xfeafffff at device 2.1 on
>>  	pci0
>
> How about if you build the module with "make -DDRM_DEBUG=1"?

  Sure. First I did "cd /sys/modules/drm; make -DDRM_DEBUG=1", but noticed
that this didn't pass DRM_DEBUG to cc and thus made exactly the same binary.
Then I evaluated proper CFLAGS and did:

CFLAGS="-DDRM_DEBUG=1 -O2 -fno-strict-aliasing -pipe" make obj depend \
 		all install

Here is relevant part of the new verbose dmesg.boot:

vgapci0: <VGA-compatible display> port 0xec00-0xec07 mem
  0xfeb80000-0xfebfffff,0xd0000000-0xdfffffff,0xfeb40000-0xfeb7ffff irq 16 at
  device 2.0 on pci0
agp0: <Intel 82915GM (915GM GMCH) SVGA controller> on vgapci0
vgapci0: Reserved 0x80000 bytes for rid 0x10 type 3 at 0xfeb80000
vgapci0: Reserved 0x80000 bytes for rid 0x10 type 3 at 0xfeb80000
vgapci0: Reserved 0x40000 bytes for rid 0x1c type 3 at 0xfeb40000
agp0: detected 7932k stolen memory
agp0: aperture size is 256M
drm0: <Intel i915GM> on vgapci0
[drm:pid0:drm_load]
error: [drm:pid0:drm_load] *ERROR* Card isn't AGP, or couldn't initialize AGP.
[drm:pid0:drm_lastclose]
device_attach: drm0 attach returned 12
vgapci1: <VGA-compatible display> mem 0xfea80000-0xfeafffff at device 2.1 on
  pci0

So it adds just 2 messages: "[drm:pid0:drm_load]" and 
"[drm:pid0:drm_lastclose]". I'm not sure whether it's very useful, but that's
the only relevant difference.

Sincerely, Dmitry
-- 
Atlantis ISP, System Administrator
e-mail:  dmitry at atlantis.dp.ua
nic-hdl: LYNX-RIPE