Hi,
I'm trying to track down a regression occuring in recent kernel : I
get the DVI output of my intel G33 chipset through a Sapphire ADD2
DVI-D SDVO adaptor card and it was working fine until
kernel 2.6.37-rc1 :
- all more recent kernels only give a black screen
- I can log through ssh and no oops nor anything suspicious is in the log
- I get an output if I uses the VGA builtin port.
I did a git bisect to track down the problem but unfortunately if got
stuck at commit 890f3359f7b84d7015104360d647ccac5f515542 that causes
an oops and doesn't allow me do do more finegrained bisection because
so far all later commits show either this oops or a black screen.
What can I do to help solve this problem ? Attached is my current git
bisect log.
Best regards,
Vincent Vanackère
On Sun, 21 Nov 2010 12:20:01 +0100, Vincent Vanackere <[email protected]> wrote:
> Hi,
>
> I'm trying to track down a regression occuring in recent kernel : I
> get the DVI output of my intel G33 chipset through a Sapphire ADD2
> DVI-D SDVO adaptor card and it was working fine until
> kernel 2.6.37-rc1 :
> - all more recent kernels only give a black screen
> - I can log through ssh and no oops nor anything suspicious is in the log
> - I get an output if I uses the VGA builtin port.
Add drm.debug=0xe to your boot command line (or module options) and attach
the dmesg. We're obviously failing to detect the output over the SDVO
card, and hopefully we might have a reason in the debug log.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
On 11/21/2010 12:49 PM, Chris Wilson wrote:
> On Sun, 21 Nov 2010 12:20:01 +0100, Vincent Vanackere<[email protected]> wrote:
>> Hi,
>>
>> I'm trying to track down a regression occuring in recent kernel : I
>> get the DVI output of my intel G33 chipset through a Sapphire ADD2
>> DVI-D SDVO adaptor card and it was working fine until
>> kernel 2.6.37-rc1 :
>> - all more recent kernels only give a black screen
>> - I can log through ssh and no oops nor anything suspicious is in the log
>> - I get an output if I uses the VGA builtin port.
>
> Add drm.debug=0xe to your boot command line (or module options) and attach
> the dmesg. We're obviously failing to detect the output over the SDVO
> card, and hopefully we might have a reason in the debug log.
> -Chris
>
I attached the dmesg from the latest git kernel (commit
b86db4744230c94e480de56f1b7f31117edbf193). Do you need the same
information with a working kernel ?
Vincent
On Sun, 21 Nov 2010 18:36:19 +0100, Vincent Vanackere <[email protected]> wrote:
> I attached the dmesg from the latest git kernel (commit
> b86db4744230c94e480de56f1b7f31117edbf193). Do you need the same
> information with a working kernel ?
It found the SDVO DVI-D device, successfully queried its modelines, and
attached it to a crtc and framebuffer... In all appearances it should be
working. Can you attach the Xorg.log and xrandr output?
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
On 11/21/2010 09:32 PM, Chris Wilson wrote:
> It found the SDVO DVI-D device, successfully queried its modelines, and
> attached it to a crtc and framebuffer... In all appearances it should be
> working. Can you attach the Xorg.log and xrandr output?
> -Chris
>
Here are the Xorg.log and xrandr output with the same kernel.
Vincent
On Sun, 21 Nov 2010 18:36:19 +0100, Vincent Vanackere <[email protected]> wrote:
> [ 8.740024] [drm:drm_crtc_helper_set_mode], [ENCODER:7:TMDS-7] set [MODE:13:1920x1200]
> [ 8.740027] [drm:intel_sdvo_debug_write], SDVOB: W: 07 01 00 00 00 (SDVO_CMD_SET_IN_OUT_MAP)
> [ 8.747826] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
> [ 8.747828] [drm:intel_sdvo_debug_write], SDVOB: W: 11 01 00 (SDVO_CMD_SET_TARGET_OUTPUT)
> [ 8.753161] [drm:intel_sdvo_debug_write], SDVOB: W: 11 01 00 (SDVO_CMD_SET_TARGET_OUTPUT)
> [ 8.758485] [drm:intel_sdvo_debug_write], SDVOB: W: 16 28 3C 80 A0 70 B0 23 40 (SDVO_CMD_SET_OUTPUT_TIMINGS_PART1)
> [ 8.771225] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
> [ 8.771227] [drm:intel_sdvo_debug_write], SDVOB: W: 10 00 (SDVO_CMD_SET_TARGET_INPUT)
> [ 8.775340] [drm:intel_sdvo_debug_write], SDVOB: W: 93 01 00 (SDVO_CMD_SET_HBUF_INDEX)
> [ 8.780688] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
This was the bug. We aborted the modeset as we tried to perform an
unsupported command. Should be fixed by:
commit da79de97d254145dcb7c08c978b1093eac15ec9c
Author: Chris Wilson <[email protected]>
Date: Mon Nov 22 11:12:46 2010 +0000
drm/i915/sdvo: Only enable HDMI encodings only if the commandset is supported
As we conflated intel_sdvo->is_hdmi with both having HDMI support on the
ADD along with having HDMI support on the monitor, we would attempt to
use HDMI encodings even if the interface did not support those commands.
Reported-by: Simon Farnsworth <[email protected]>
Signed-off-by: Chris Wilson <[email protected]>
Tested-by: Simon Farnsworth <[email protected]>
Reviewed-by: Simon Farnsworth <[email protected]>
available at
git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel.git
[drm-intel-fixes]
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
On Tue, Nov 23, 2010 at 10:08 AM, Chris Wilson <[email protected]> wrote:
> On Sun, 21 Nov 2010 18:36:19 +0100, Vincent Vanackere <[email protected]> wrote:
>> [ 8.740024] [drm:drm_crtc_helper_set_mode], [ENCODER:7:TMDS-7] set [MODE:13:1920x1200]
>> [ 8.740027] [drm:intel_sdvo_debug_write], SDVOB: W: 07 01 00 00 00 (SDVO_CMD_SET_IN_OUT_MAP)
>> [ 8.747826] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
>> [ 8.747828] [drm:intel_sdvo_debug_write], SDVOB: W: 11 01 00 (SDVO_CMD_SET_TARGET_OUTPUT)
>> [ 8.753161] [drm:intel_sdvo_debug_write], SDVOB: W: 11 01 00 (SDVO_CMD_SET_TARGET_OUTPUT)
>> [ 8.758485] [drm:intel_sdvo_debug_write], SDVOB: W: 16 28 3C 80 A0 70 B0 23 40 (SDVO_CMD_SET_OUTPUT_TIMINGS_PART1)
>> [ 8.771225] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
>> [ 8.771227] [drm:intel_sdvo_debug_write], SDVOB: W: 10 00 (SDVO_CMD_SET_TARGET_INPUT)
>> [ 8.775340] [drm:intel_sdvo_debug_write], SDVOB: W: 93 01 00 (SDVO_CMD_SET_HBUF_INDEX)
>> [ 8.780688] [drm:intel_sdvo_write_cmd], command returns response Not supported [2]
>
> This was the bug. We aborted the modeset as we tried to perform an
> unsupported command. Should be fixed by:
>
> commit da79de97d254145dcb7c08c978b1093eac15ec9c
> Author: Chris Wilson <[email protected]>
> Date: Mon Nov 22 11:12:46 2010 +0000
>
Thanks ! This patch indeed fixes my bug.
Best regards,
Vincent