Skip to contentSkip to conversation

drm/i915: set the AVI VIC of the HDMI mode

PATCH In review Paulo Zanoni 4 participants 11 messages Nov 21, 2012 – Nov 23, 2012

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: set the AVI VIC of the HDMI mode

From: Paulo Zanoni <[email protected]>

We currently set "0" as the VIC value of the AVI InfoFrames. According to the specs this should be fine and work for every mode, so to my point of view we can't consider the current behavior as a bug. The problem is that we recently received a bug report (Kernel bug #50371) from a user that has an AV receiver that gives a black screen for any mode with VIC set to 0.

So in order to make at least some modes work for him, this patch sets the correct VIC number when sending AVI InfoFrames. We add a generic drm function to calculate the VIC number and then call it from the i915 driver.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=50371

Cc Thierry Reding <[email protected]>
---
 drivers/gpu/drm/drm_edid.c        |   19 +++++++++++++++++++
 drivers/gpu/drm/drm_modes.c       |    3 ++-
 drivers/gpu/drm/i915/intel_hdmi.c |    2 ++
 include/drm/drm_crtc.h            |    4 +++-
 4 files changed, 26 insertions(+), 2 deletions(-)

Patch applies on top of Daniel's drm-intel-next-queued. I'm not sure who exactly is going to merge this (Dave or Daniel).

drivers/gpu/drm/drm_edid.c +19 -0
index fadcd44..c57fc46 100644
@@ -2062,3 +2062,22 @@ int drm_add_modes_noedid(struct drm_connector *connector,
 	return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
+
+/**
+ * drm_mode_vic - return the CEA-861 VIC of a given mode
+ * @mode: mode
+ *
+ * RETURNS:
+ * The VIC number, 0 in case it's not a CEA-861 mode.
+ */
+uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)
+{
+	uint8_t i;
+
+	for (i = 0; i < drm_num_cea_modes; i++)
+		if (drm_mode_equal(mode, &edid_cea_modes[i]))
+			return i + 1;
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_mode_cea_vic);
drivers/gpu/drm/drm_modes.c +2 -1
index 59450f3..9ef6750 100644
@@ -768,7 +768,8 @@ EXPORT_SYMBOL(drm_mode_duplicate);
  * RETURNS:
  * True if the modes are equal, false otherwise.
  */
-bool drm_mode_equal(struct drm_display_mode *mode1, struct drm_display_mode *mode2)
+bool drm_mode_equal(const struct drm_display_mode *mode1,
+		    const struct drm_display_mode *mode2)
 {
 	/* do clock check convert to PICOS so fb modes get matched
 	 * the same */
drivers/gpu/drm/i915/intel_hdmi.c +2 -0
index 1dcfd5b..eaf70d6 100644
@@ -340,6 +340,8 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		avi_if.body.avi.YQ_CN_PR |= DIP_AVI_PR_2;
 
+	avi_if.body.avi.VIC = drm_mode_cea_vic(adjusted_mode);
+
 	intel_set_infoframe(encoder, &avi_if);
 }
 
include/drm/drm_crtc.h +3 -1
index 3fa18b7..245220e 100644
@@ -892,7 +892,8 @@ extern void drm_mode_config_init(struct drm_device *dev);
 extern void drm_mode_config_reset(struct drm_device *dev);
 extern void drm_mode_config_cleanup(struct drm_device *dev);
 extern void drm_mode_set_name(struct drm_display_mode *mode);
-extern bool drm_mode_equal(struct drm_display_mode *mode1, struct drm_display_mode *mode2);
+extern bool drm_mode_equal(const struct drm_display_mode *mode1,
+			   const struct drm_display_mode *mode2);
 extern int drm_mode_width(struct drm_display_mode *mode);
 extern int drm_mode_height(struct drm_display_mode *mode);
 
@@ -1053,6 +1054,7 @@ extern struct drm_display_mode *drm_gtf_mode_complex(struct drm_device *dev,
 				int GTF_2C, int GTF_K, int GTF_2J);
 extern int drm_add_modes_noedid(struct drm_connector *connector,
 				int hdisplay, int vdisplay);
+extern uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode);
 
 extern int drm_edid_header_is_valid(const u8 *raw_edid);
 extern bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid);
Signature · 1 line
1.7.10.4
#1 lore raw
Quoted 44 lines · Paulo Zanoni

From: Paulo Zanoni <[email protected]>

We currently set "0" as the VIC value of the AVI InfoFrames. According to the specs this should be fine and work for every mode, so to my point of view we can't consider the current behavior as a bug. The problem is that we recently received a bug report (Kernel bug #50371) from a user that has an AV receiver that gives a black screen for any mode with VIC set to 0.

So in order to make at least some modes work for him, this patch sets the correct VIC number when sending AVI InfoFrames. We add a generic drm function to calculate the VIC number and then call it from the i915 driver.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=50371

Cc Thierry Reding <[email protected]>
---
 drivers/gpu/drm/drm_edid.c        |   19 +++++++++++++++++++
 drivers/gpu/drm/drm_modes.c       |    3 ++-
 drivers/gpu/drm/i915/intel_hdmi.c |    2 ++
 include/drm/drm_crtc.h            |    4 +++-
 4 files changed, 26 insertions(+), 2 deletions(-)

Patch applies on top of Daniel's drm-intel-next-queued. I'm not sure who exactly is going to merge this (Dave or Daniel).

drivers/gpu/drm/drm_edid.c +9 -0
index fadcd44..c57fc46 100644
@@ -2062,3 +2062,22 @@ int drm_add_modes_noedid(struct drm_connector *connector,
 	return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
+
+/**
+ * drm_mode_vic - return the CEA-861 VIC of a given mode
+ * @mode: mode
+ *
+ * RETURNS:
+ * The VIC number, 0 in case it's not a CEA-861 mode.
+ */
+uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2083

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +10 -0
+{
+	uint8_t i;
+
+	for (i = 0; i < drm_num_cea_modes; i++)
+		if (drm_mode_equal(mode, &edid_cea_modes[i]))
+			return i + 1;
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_mode_cea_vic);

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Thierry

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

#2 lore raw
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

I think the best way is to pick the infoframe implementation you best like from one of these, move it into the helper, then improve it until you're happy. And then convert over 1-2 other drivers. At least that's been my approach for the recent dp helper refactoring.

Cheers, Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#3 lore raw
Quoted 11 lines · Daniel Vetter
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

Quoted 4 lines

I think the best way is to pick the infoframe implementation you best like from one of these, move it into the helper, then improve it until you're happy. And then convert over 1-2 other drivers. At least that's been my approach for the recent dp helper refactoring.

And if you could use that opportunity to integrate the kerneldoc for edid/eld and the new infoframe code into the drm docbook, that would be rather awesome ;-)
-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#4 lore raw
Quoted 12 lines · Daniel Vetter
Quoted 11 lines · Daniel Vetter
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

Quoted 8 lines
Quoted 4 lines

I think the best way is to pick the infoframe implementation you best like from one of these, move it into the helper, then improve it until you're happy. And then convert over 1-2 other drivers. At least that's been my approach for the recent dp helper refactoring.

And if you could use that opportunity to integrate the kerneldoc for edid/eld and the new infoframe code into the drm docbook, that would be rather awesome ;-)

I suck at documentation =), but I'll see what I can do.

Thierry

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

#5 lore raw
In reply to Daniel Vetter
Quoted 10 lines · Daniel Vetter
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

I'll have to rely on somebody else to do the testing since I don't have an HDMI capable hardware except Tegra to run this on. But yes, I had planned to convert Tegra and at least one other driver for reference. But I guess while at it I could just as well convert all of them.

Thierry

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

#6 lore raw

2012/11/22 Thierry Reding <[email protected]>:

Quoted 12 lines · Daniel Vetter
Quoted 10 lines
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

Quoted 4 lines

I'll have to rely on somebody else to do the testing since I don't have an HDMI capable hardware except Tegra to run this on. But yes, I had planned to convert Tegra and at least one other driver for reference. But I guess while at it I could just as well convert all of them.

I'll take a look at radeon (I'll try to convert it to the new functions) over weekend.

Signature · 5 lines
Rafał
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#7 lore raw
Quoted 13 lines · Rafał Miłecki

2012/11/22 Thierry Reding <[email protected]>:

Quoted 12 lines · Daniel Vetter
Quoted 10 lines
Quoted 2 lines · Thierry Reding

Oh great, so I copied that table for nothing. Thanks for Cc'ing, I can reuse that in the HDMI infoframe series.

Wrt the infoframe series, I think it'd be awesome if you could convert i915 and radeon (iirc the existing drivers with the "best" avi infoframe support) over to the new code. This gives some nice validation, both by testing on actual hw and that the interface is sane, since it'll be used by 2-3 different drivers then.

Quoted 7 lines
Quoted 4 lines

I'll have to rely on somebody else to do the testing since I don't have an HDMI capable hardware except Tegra to run this on. But yes, I had planned to convert Tegra and at least one other driver for reference. But I guess while at it I could just as well convert all of them.

I'll take a look at radeon (I'll try to convert it to the new functions) over weekend.

Okay, great! I think for radeon things should be the easiest since it doesn't currently fill in anything but the colorspace field. Judging by the bug report that Paulo mentioned this will probably not be enough for some hardware, but should be enough according to the specification. Having more data in the AVI infoframe shouldn't hurt, though.

Thierry

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#8 lore raw
In reply to Paulo Zanoni
Quoted 32 lines · Paulo Zanoni

From: Paulo Zanoni <[email protected]>

We currently set "0" as the VIC value of the AVI InfoFrames. According to the specs this should be fine and work for every mode, so to my point of view we can't consider the current behavior as a bug. The problem is that we recently received a bug report (Kernel bug #50371) from a user that has an AV receiver that gives a black screen for any mode with VIC set to 0.

So in order to make at least some modes work for him, this patch sets the correct VIC number when sending AVI InfoFrames. We add a generic drm function to calculate the VIC number and then call it from the i915 driver.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=50371

Cc Thierry Reding <[email protected]>
---
 drivers/gpu/drm/drm_edid.c        |   19 +++++++++++++++++++
 drivers/gpu/drm/drm_modes.c       |    3 ++-
 drivers/gpu/drm/i915/intel_hdmi.c |    2 ++
 include/drm/drm_crtc.h            |    4 +++-
 4 files changed, 26 insertions(+), 2 deletions(-)

Patch applies on top of Daniel's drm-intel-next-queued. I'm not sure who exactly is going to merge this (Dave or Daniel).

drivers/gpu/drm/drm_edid.c +0 -0
index fadcd44..c57fc46 100644

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2067

Quoted 6 lines
drivers/gpu/drm/drm_edid.c +3 -0
@@ -2062,3 +2062,22 @@ int drm_add_modes_noedid(struct drm_connector *connector,
 	return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
+
+/**
+ * drm_mode_vic - return the CEA-861 VIC of a given mode

The name in the comment here doesn't match the actual function name.

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2073

Quoted 6 lines
drivers/gpu/drm/drm_edid.c +6 -0
+ * @mode: mode
+ *
+ * RETURNS:
+ * The VIC number, 0 in case it's not a CEA-861 mode.
+ */
+uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)

I understand the reason for returning an uint8_t here, but ...

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2077

Quoted 4 lines
drivers/gpu/drm/drm_edid.c +4 -0
+{
+	uint8_t i;
+
+	for (i = 0; i < drm_num_cea_modes; i++)

... maybe unsigned int would be better for the iteration variable here. Looking at drm_edid_modes.h, drm_num_cea_modes is actually signed, which isn't necessary to store an array size, so maybe that should be changed as well.

Quoted 11 lines
drivers/gpu/drm/drm_edid.c +6 -0
+		if (drm_mode_equal(mode, &edid_cea_modes[i]))
+			return i + 1;
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_mode_cea_vic);
drivers/gpu/drm/drm_modes.c +0 -0
index 59450f3..9ef6750 100644

drivers/gpu/drm/drm_modes.c EXPORT_SYMBOL(drm_mode_duplicate); · line 772

Quoted 6 lines
drivers/gpu/drm/drm_modes.c +2 -1
@@ -768,7 +768,8 @@ EXPORT_SYMBOL(drm_mode_duplicate);
  * RETURNS:
  * True if the modes are equal, false otherwise.
  */
-bool drm_mode_equal(struct drm_display_mode *mode1, struct drm_display_mode *mode2)
+bool drm_mode_equal(const struct drm_display_mode *mode1,
+		    const struct drm_display_mode *mode2)

I think this change warrants a separate commit.

Thierry

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

#9 lore raw

Hi

2012/11/22 Thierry Reding <[email protected]>:

Quoted 28 lines · Paulo Zanoni
Quoted 27 lines

From: Paulo Zanoni <[email protected]>

We currently set "0" as the VIC value of the AVI InfoFrames. According to the specs this should be fine and work for every mode, so to my point of view we can't consider the current behavior as a bug. The problem is that we recently received a bug report (Kernel bug #50371) from a user that has an AV receiver that gives a black screen for any mode with VIC set to 0.

So in order to make at least some modes work for him, this patch sets the correct VIC number when sending AVI InfoFrames. We add a generic drm function to calculate the VIC number and then call it from the i915 driver.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=50371

Cc Thierry Reding <[email protected]>
---
 drivers/gpu/drm/drm_edid.c        |   19 +++++++++++++++++++
 drivers/gpu/drm/drm_modes.c       |    3 ++-
 drivers/gpu/drm/i915/intel_hdmi.c |    2 ++
 include/drm/drm_crtc.h            |    4 +++-
 4 files changed, 26 insertions(+), 2 deletions(-)

Patch applies on top of Daniel's drm-intel-next-queued. I'm not sure who exactly is going to merge this (Dave or Daniel).

Quoted 13 lines

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2067

Quoted 11 lines
drivers/gpu/drm/drm_edid.c +3 -0
index fadcd44..c57fc46 100644
@@ -2062,3 +2062,22 @@ int drm_add_modes_noedid(struct drm_connector *connector,
      return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
+
+/**
+ * drm_mode_vic - return the CEA-861 VIC of a given mode

The name in the comment here doesn't match the actual function name.

Will fix.

Quoted 10 lines

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2073

Quoted 7 lines
drivers/gpu/drm/drm_edid.c +6 -0
+ * @mode: mode
+ *
+ * RETURNS:
+ * The VIC number, 0 in case it's not a CEA-861 mode.
+ */
+uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)

I understand the reason for returning an uint8_t here, but ...

Quoted 9 lines

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2077

Quoted 4 lines
drivers/gpu/drm/drm_edid.c +4 -0
+{
+     uint8_t i;
+
+     for (i = 0; i < drm_num_cea_modes; i++)

... maybe unsigned int would be better for the iteration variable here. Looking at drm_edid_modes.h, drm_num_cea_modes is actually signed, which isn't necessary to store an array size, so maybe that should be changed as well.

I used uint8_t because "i" is the thing we return. I don't think there's a perfect solution to this problem and I don't really think it matters too much.

Quoted 12 lines
Quoted 12 lines
drivers/gpu/drm/drm_edid.c +6 -0
+             if (drm_mode_equal(mode, &edid_cea_modes[i]))
+                     return i + 1;
+
+     return 0;
+}
+EXPORT_SYMBOL(drm_mode_cea_vic);
drivers/gpu/drm/drm_modes.c +0 -0
index 59450f3..9ef6750 100644
@@ -768,7 +768,8 @@ EXPORT_SYMBOL(drm_mode_duplicate);
Quoted 8 lines

drivers/gpu/drm/drm_modes.c EXPORT_SYMBOL(drm_mode_duplicate); · line 772

Quoted 6 lines
drivers/gpu/drm/drm_modes.c +2 -1
  * RETURNS:
  * True if the modes are equal, false otherwise.
  */
-bool drm_mode_equal(struct drm_display_mode *mode1, struct drm_display_mode *mode2)
+bool drm_mode_equal(const struct drm_display_mode *mode1,
+                 const struct drm_display_mode *mode2)

I think this change warrants a separate commit.

I just noticed Dave's tree already has these things as const. So V2 will be based on Dave's drm-next tree.

Quoted 2 lines

Thierry

Signature · 1 line
Paulo Zanoni
#10 lore raw
Quoted 4 lines · Paulo Zanoni

Hi

2012/11/22 Thierry Reding <[email protected]>:

Quoted 1 line · Paulo Zanoni

[...]

Quoted 19 lines
Quoted 9 lines
Quoted 6 lines

+ * @mode: mode
+ *
+ * RETURNS:
+ * The VIC number, 0 in case it's not a CEA-861 mode.
+ */
+uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)

I understand the reason for returning an uint8_t here, but ...

Quoted 9 lines

drivers/gpu/drm/drm_edid.c int drm_add_modes_noedid(struct drm_connector *connector, · line 2077

Quoted 4 lines
drivers/gpu/drm/drm_edid.c +4 -0
+{
+     uint8_t i;
+
+     for (i = 0; i < drm_num_cea_modes; i++)

... maybe unsigned int would be better for the iteration variable here. Looking at drm_edid_modes.h, drm_num_cea_modes is actually signed, which isn't necessary to store an array size, so maybe that should be changed as well.

Quoted 3 lines

I used uint8_t because "i" is the thing we return. I don't think there's a perfect solution to this problem and I don't really think it matters too much.

Fair enough, uint8_t should work fine.

Thierry

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel