[REGRESSION] 3.4-rc* radeon: No DVI-A output after commit e00e8b5e760cb

[REGRESSION] 3.4-rc* radeon: No DVI-A output after commit e00e8b5e760cb

[Takashi Iwai]

Hi,

I noticed that one machine here gives only the blank output with
3.4-rc's.  The bisection lead me to affecting commit:

    commit e00e8b5e760cbbe9067daeae5454d67c44c8d035
    Author: Alex Deucher <alexander.deucher@amd.com>
    Date:   Fri Mar 16 12:22:09 2012 -0400

    drm/radeon/kms: fix analog load detection on DVI-I connectors
    
Reverting this commit helped, it goes back to normal.

On this system, the VGA monitor is connected to DVI over a VGA-DVI
connector (and even VGA-switcher).  So, yeah, it's a strange setup.
But a regression is a regression.

I'm willing to test any patch.


thanks,

Takashi

Re: [REGRESSION] 3.4-rc* radeon: No DVI-A output after commit e00e8b5e760cb

[Takashi Iwai]

At Wed, 18 Apr 2012 14:39:54 +0200,
Takashi Iwai wrote:
> 
> Hi,
> 
> I noticed that one machine here gives only the blank output with
> 3.4-rc's.  The bisection lead me to affecting commit:
> 
>     commit e00e8b5e760cbbe9067daeae5454d67c44c8d035
>     Author: Alex Deucher <alexander.deucher@amd.com>
>     Date:   Fri Mar 16 12:22:09 2012 -0400
> 
>     drm/radeon/kms: fix analog load detection on DVI-I connectors
>     
> Reverting this commit helped, it goes back to normal.
> 
> On this system, the VGA monitor is connected to DVI over a VGA-DVI
> connector (and even VGA-switcher).  So, yeah, it's a strange setup.
> But a regression is a regression.

Looking at the commit again, it appears like a simple logic failure.
It should be "&&" instead of "||".


Takashi

[PATCH] drm/radeon/kms: fix the regression of DVI connector check

[Takashi Iwai]

The check of the encoder type in the commit [e00e8b5e: drm/radeon/kms:
fix analog load detection on DVI-I connectors] is obviously wrong, and
it's the culprit of the regression on my workstation with DVI-analog
connection resulting in the blank output.

Fixed the typo now.

Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
 drivers/gpu/drm/radeon/radeon_connectors.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index bd05156..aa8268d 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -970,7 +970,7 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 
 			encoder = obj_to_encoder(obj);
 
-			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC ||
+			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC &&
 			    encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
 				continue;
 
-- 
1.7.9.2

Re: [PATCH] drm/radeon/kms: fix the regression of DVI connector check

[Alex Deucher]

On Wed, Apr 18, 2012 at 9:21 AM, Takashi Iwai <tiwai@suse.de> wrote:
> The check of the encoder type in the commit [e00e8b5e: drm/radeon/kms:
> fix analog load detection on DVI-I connectors] is obviously wrong, and
> it's the culprit of the regression on my workstation with DVI-analog
> connection resulting in the blank output.
>
> Fixed the typo now.
>
> Cc: <stable@vger.kernel.org>
> Signed-off-by: Takashi Iwai <tiwai@suse.de>

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

> ---
>  drivers/gpu/drm/radeon/radeon_connectors.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
> index bd05156..aa8268d 100644
> --- a/drivers/gpu/drm/radeon/radeon_connectors.c
> +++ b/drivers/gpu/drm/radeon/radeon_connectors.c
> @@ -970,7 +970,7 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
>
>                        encoder = obj_to_encoder(obj);
>
> -                       if (encoder->encoder_type != DRM_MODE_ENCODER_DAC ||
> +                       if (encoder->encoder_type != DRM_MODE_ENCODER_DAC &&
>                            encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
>                                continue;
>
> --
> 1.7.9.2
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH] drm/radeon: fix logic inversion in load detect fix

[alexdeucher]

From: Alex Deucher <alexander.deucher@amd.com>

Reported-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/radeon/radeon_connectors.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 910ea6d..4bbb6d4 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -977,7 +977,7 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 
 			encoder = obj_to_encoder(obj);
 
-			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC ||
+			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC &&
 			    encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
 				continue;
 
-- 
1.7.7.5

Re: [REGRESSION] 3.4-rc* radeon: No DVI-A output after commit e00e8b5e760cb

[Alex Deucher]

On Wed, Apr 18, 2012 at 8:39 AM, Takashi Iwai <tiwai@suse.de> wrote:
> Hi,
>
> I noticed that one machine here gives only the blank output with
> 3.4-rc's.  The bisection lead me to affecting commit:
>
>    commit e00e8b5e760cbbe9067daeae5454d67c44c8d035
>    Author: Alex Deucher <alexander.deucher@amd.com>
>    Date:   Fri Mar 16 12:22:09 2012 -0400
>
>    drm/radeon/kms: fix analog load detection on DVI-I connectors
>
> Reverting this commit helped, it goes back to normal.
>
> On this system, the VGA monitor is connected to DVI over a VGA-DVI
> connector (and even VGA-switcher).  So, yeah, it's a strange setup.
> But a regression is a regression.
>

I don't think it's a regression per se.  Prior to that patch, the
analog vga was using the wrong load detect function (it's using the
one for DP bridge chips) which returned connector_status_unknown if it
was called on an invalid connector.  The patch just sets the behaviour
back to the original behaviour prior to support for DP bridge chips
being added (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=d629a3ceb4fc1ab5aab737b964100d114aba1173).
 The problem is the driver is not able to detect the monitor on the
other side of the KVM (a common problem with KVMs).  The real problem
is that some desktop environments choose to interpret
connector_status_unknown as connected.  It's not really viable to
always return connector_status_unknown as then all ports would be
considered connected when in truth they would not be.

Alex

> I'm willing to test any patch.
>
>
> thanks,
>
> Takashi
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel