Skip to contentSkip to conversation

Enhance EDID quirks to allow forcing a mode

PATCH Superseded Dylan Semler 3 participants 7 messages Mar 22, 2013 – Mar 25, 2013

Subsystems (2)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v3 cover Enhance EDID quirks to allow forcing a mode

Changes in this version
* Uses drm_cvt_mode() instead of drm_gtf_mode() to build modeline
* Adds bool to specify reduced blanking to edid_quirk_force_mode
* Removes preferred bit from all other modes

There is at least one monitor that doesn't report its native resolution in its EDID block. This enhancement extends the EDID quirk logic to make monitors like this "just work".

The first patch in this series sets up a new quirk list where monitors' correct width, height, refresh rate, and reduced blanking parameters are specified. When a matching monitor is attached the full mode is calculated with drm_cvt_mode() and added to the connector. The DRM_MODE_TYPE_PREFERRED bit is set on the new mode and unset from all other modes.

The first patch also defines a new quirk bit: EDID_QUIRK_FORCE_MODE. This bit needs to be set for the new quirk list discribed above to be checked.

The second patch adds the offending monitor to the quirk lists.

Dylan Semler (2):
drm: Enhance EDID quirks to explicitly set a mode
drm: Add EDID force quirk for MMT Monitor2Go HD+

 drivers/gpu/drm/drm_edid.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
Signature · 1 line
1.7.11.7
#1 lore raw

PATCH v3 1/2 drm: Enhance EDID quirks to explicitly set a mode

There is at least one monitor that doesn't report its native resolution in its EDID block. This enhancement extends the EDID quirk logic to make monitors like this "just work".

This patch sets up a new quirk list where monitors' correct width, height, refresh rate, and reduced blanking parameters are specified. When a matching monitor is attached the full mode is calculated with drm_cvt_mode() and added to the connector. The DRM_MODE_TYPE_PREFERRED bit is set on the new mode and unset from all other modes.

The patch also defines a new quirk bit: EDID_QUIRK_FORCE_MODE. This bit needs to be set for the new quirk list discribed above to be checked.

---
 drivers/gpu/drm/drm_edid.c | 78 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)
drivers/gpu/drm/drm_edid.c +78 -0
index c194f4e..38b8641 100644
@@ -68,6 +68,8 @@
 #define EDID_QUIRK_DETAILED_SYNC_PP		(1 << 6)
 /* Force reduced-blanking timings for detailed modes */
 #define EDID_QUIRK_FORCE_REDUCED_BLANKING	(1 << 7)
+/* Force specific mode for monitors that don't report correct EDIDs */
+#define EDID_QUIRK_FORCE_MODE			(1 << 8)
 
 struct detailed_mode_closure {
 	struct drm_connector *connector;
@@ -127,6 +129,16 @@ static struct edid_quirk {
 	{ "VSC", 5020, EDID_QUIRK_FORCE_REDUCED_BLANKING },
 };
 
+static struct edid_quirk_force_mode {
+	char vendor[4];
+	int product_id;
+	int hdisplay;
+	int vdisplay;
+	int vrefresh;
+	bool reduced;
+} edid_quirk_force_mode_list[] = {
+};
+
 /*
  * Autogenerated from the DMT spec.
  * This table is copied from xfree86/modes/xf86EdidModes.c.
@@ -2219,6 +2231,70 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 	return closure.modes;
 }
 
+/* Add an explicit mode based on a quirk
+ */
+static int
+do_force_quirk_modes(struct drm_connector *connector, int hdisplay,
+		     int vdisplay, int vrefresh, bool reduced)
+{
+	struct drm_display_mode *mode, *t, *cur_mode;
+	struct drm_device *dev = connector->dev;
+	int num_modes = 0;
+
+	/* sanity check display parameters */
+	if (hdisplay < 0)
+		return 0;
+	if (vdisplay < 0)
+		return 0;
+	if (vrefresh < 0)
+		return 0;
+
+	/* loop through the probed modes and clear the preferred bit */
+	list_for_each_entry_safe(cur_mode, t, &connector->probed_modes, head)
+		cur_mode->type &= ~DRM_MODE_TYPE_PREFERRED;
+
+	mode = drm_cvt_mode(dev, hdisplay, vdisplay, vrefresh, reduced, 0, 0);
+
+	if (mode) {
+		mode->type |= DRM_MODE_TYPE_PREFERRED;
+		drm_mode_probed_add(connector, mode);
+		num_modes++;
+	}
+	return num_modes;
+}
+
+/*
+ * add_force_quirk_modes - Add modes based on monitor's EDID quirks
+ * @connector: attached connector
+ * @edid: EDID block to scan
+ * @quirks: quirks to apply
+ *
+ * At least one monitor doesn't report its native resolution in its EDID block.
+ * Here we add the native mode according to this quirk
+ */
+static int
+add_force_quirk_modes(struct drm_connector *connector, struct edid *edid,
+		      u32 quirks)
+{
+	struct edid_quirk_force_mode *quirk_mode;
+	int i, num_modes = 0;
+
+	for (i = 0; i < ARRAY_SIZE(edid_quirk_force_mode_list); i++) {
+		quirk_mode = &edid_quirk_force_mode_list[i];
+
+		if (edid_vendor(edid, quirk_mode->vendor) &&
+		    (EDID_PRODUCT_ID(edid) == quirk_mode->product_id)) {
+			num_modes = do_force_quirk_modes(connector,
+					quirk_mode->hdisplay,
+					quirk_mode->vdisplay,
+					quirk_mode->vrefresh,
+					quirk_mode->reduced);
+		}
+	}
+	return num_modes;
+
+}
+
 #define HDMI_IDENTIFIER 0x000C03
 #define AUDIO_BLOCK	0x01
 #define VIDEO_BLOCK     0x02
@@ -2803,6 +2879,8 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)
 
 	if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75))
 		edid_fixup_preferred(connector, quirks);
+	if (quirks & EDID_QUIRK_FORCE_MODE)
+		num_modes += add_force_quirk_modes(connector, edid, quirks);
 
 	drm_add_display_info(edid, &connector->display_info);
 
Signature · 1 line
1.7.11.7
#2 lore raw

Re: [PATCH v3 1/2] drm: Enhance EDID quirks to explicitly set a mode

Hi, please find some review comments inline.

BR,
Jani.

Quoted 50 lines · Dylan Semler

There is at least one monitor that doesn't report its native resolution in its EDID block. This enhancement extends the EDID quirk logic to make monitors like this "just work".

This patch sets up a new quirk list where monitors' correct width, height, refresh rate, and reduced blanking parameters are specified. When a matching monitor is attached the full mode is calculated with drm_cvt_mode() and added to the connector. The DRM_MODE_TYPE_PREFERRED bit is set on the new mode and unset from all other modes.

The patch also defines a new quirk bit: EDID_QUIRK_FORCE_MODE. This bit needs to be set for the new quirk list discribed above to be checked.

---
 drivers/gpu/drm/drm_edid.c | 78 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)
drivers/gpu/drm/drm_edid.c +12 -0
index c194f4e..38b8641 100644
@@ -68,6 +68,8 @@
 #define EDID_QUIRK_DETAILED_SYNC_PP		(1 << 6)
 /* Force reduced-blanking timings for detailed modes */
 #define EDID_QUIRK_FORCE_REDUCED_BLANKING	(1 << 7)
+/* Force specific mode for monitors that don't report correct EDIDs */
+#define EDID_QUIRK_FORCE_MODE			(1 << 8)
 
 struct detailed_mode_closure {
 	struct drm_connector *connector;
@@ -127,6 +129,16 @@ static struct edid_quirk {
 	{ "VSC", 5020, EDID_QUIRK_FORCE_REDUCED_BLANKING },
 };
 
+static struct edid_quirk_force_mode {
+	char vendor[4];
+	int product_id;
+	int hdisplay;
+	int vdisplay;
+	int vrefresh;
+	bool reduced;
+} edid_quirk_force_mode_list[] = {
+};
+
 /*
  * Autogenerated from the DMT spec.
  * This table is copied from xfree86/modes/xf86EdidModes.c.

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2238

Quoted 8 lines
drivers/gpu/drm/drm_edid.c +5 -0
@@ -2219,6 +2231,70 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 	return closure.modes;
 }
 
+/* Add an explicit mode based on a quirk
+ */
+static int
+do_force_quirk_modes(struct drm_connector *connector, int hdisplay,
+		     int vdisplay, int vrefresh, bool reduced)

Nitpick: This adds one mode, not many _modes_.

Quoted 6 lines
drivers/gpu/drm/drm_edid.c +6 -0
+{
+	struct drm_display_mode *mode, *t, *cur_mode;
+	struct drm_device *dev = connector->dev;
+	int num_modes = 0;
+
+	/* sanity check display parameters */

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2254

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +10 -0
+	if (hdisplay < 0)
+		return 0;
+	if (vdisplay < 0)
+		return 0;
+	if (vrefresh < 0)
+		return 0;
+
+	/* loop through the probed modes and clear the preferred bit */
+	list_for_each_entry_safe(cur_mode, t, &connector->probed_modes, head)
+		cur_mode->type &= ~DRM_MODE_TYPE_PREFERRED;

You're not deleting entries, so list_for_each_entry() would suffice, getting rid of the temp variable.

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2256

Quoted 2 lines
drivers/gpu/drm/drm_edid.c +2 -0
+
+	mode = drm_cvt_mode(dev, hdisplay, vdisplay, vrefresh, reduced, 0, 0);

You could

	if (!mode)
		return 0;

here and get rid of the num_modes variable.

Quoted 25 lines
drivers/gpu/drm/drm_edid.c +25 -0
+
+	if (mode) {
+		mode->type |= DRM_MODE_TYPE_PREFERRED;
+		drm_mode_probed_add(connector, mode);
+		num_modes++;
+	}
+	return num_modes;
+}
+
+/*
+ * add_force_quirk_modes - Add modes based on monitor's EDID quirks
+ * @connector: attached connector
+ * @edid: EDID block to scan
+ * @quirks: quirks to apply
+ *
+ * At least one monitor doesn't report its native resolution in its EDID block.
+ * Here we add the native mode according to this quirk
+ */
+static int
+add_force_quirk_modes(struct drm_connector *connector, struct edid *edid,
+		      u32 quirks)
+{
+	struct edid_quirk_force_mode *quirk_mode;
+	int i, num_modes = 0;
+

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2291

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +10 -0
+	for (i = 0; i < ARRAY_SIZE(edid_quirk_force_mode_list); i++) {
+		quirk_mode = &edid_quirk_force_mode_list[i];
+
+		if (edid_vendor(edid, quirk_mode->vendor) &&
+		    (EDID_PRODUCT_ID(edid) == quirk_mode->product_id)) {
+			num_modes = do_force_quirk_modes(connector,
+					quirk_mode->hdisplay,
+					quirk_mode->vdisplay,
+					quirk_mode->vrefresh,
+					quirk_mode->reduced);

I was wondering why you don't bail out here. Maybe you want to be able to add more than one quirk mode? In that case you should +=, not = the num_modes.

(Note that then do_force_quirk_modes makes the last one in the array the preferred mode, clearing the DRM_MODE_TYPE_PREFERRED from the previously added quirk modes.)

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +6 -0
+		}
+	}
+	return num_modes;
+
+}
+
 #define HDMI_IDENTIFIER 0x000C03
 #define AUDIO_BLOCK	0x01
 #define VIDEO_BLOCK     0x02

drivers/gpu/drm/drm_edid.c int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid) · line 2883

Quoted 5 lines
drivers/gpu/drm/drm_edid.c +2 -0
@@ -2803,6 +2879,8 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)
 
 	if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75))
 		edid_fixup_preferred(connector, quirks);
+	if (quirks & EDID_QUIRK_FORCE_MODE)
+		num_modes += add_force_quirk_modes(connector, edid, quirks);

You don't use quirks within add_force_quirk_modes() for anything.

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +0 -0
 	drm_add_display_info(edid, &connector->display_info);
 
Signature · 6 lines
1.7.11.7

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

Re: [PATCH v3 1/2] drm: Enhance EDID quirks to explicitly set a mode

Quoted 4 lines · Jani Nikula

Hi, please find some review comments inline.

You make some good points, thanks.

Quoted 9 lines · Dylan Semler

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2238

Quoted 6 lines
drivers/gpu/drm/drm_edid.c +5 -0
+/* Add an explicit mode based on a quirk
+ */
+static int
+do_force_quirk_modes(struct drm_connector *connector, int hdisplay,
+                  int vdisplay, int vrefresh, bool reduced)

Nitpick: This adds one mode, not many _modes_.

sure, I'll fix.

Quoted 3 lines

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid,

Quoted 3 lines
drivers/gpu/drm/drm_edid.c +3 -0
+
+     /* loop through the probed modes and clear the preferred bit */
+     list_for_each_entry_safe(cur_mode, t, &connector->probed_modes,

head)

Quoted 5 lines
Quoted 1 line

+ cur_mode->type &= ~DRM_MODE_TYPE_PREFERRED;

You're not deleting entries, so list_for_each_entry() would suffice, getting rid of the temp variable.

I was going to use this but I wasn't sure if I understood all of the differences between the two. I opted for _safe because I noticed it's used in
edid_fixup_preferred() even though that routine doesn't remove entries either.
I'll fix it.

Quoted 2 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+     mode = drm_cvt_mode(dev, hdisplay, vdisplay, vrefresh, reduced,

0, 0);

Quoted 7 lines

You could

        if (!mode)
                return 0;

here and get rid of the num_modes variable.

Yes, you're right. I went with

if (mode) {
    ...
    return 1;
}
else
    return 0;

I would naively expect this to be better. Is there a reason to prefer if (!mode)?

Quoted 2 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+static int
+add_force_quirk_modes(struct drm_connector *connector, struct edid

*edid,

Quoted 20 lines
Quoted 5 lines
drivers/gpu/drm/drm_edid.c +5 -0
+                   u32 quirks)
+{
+     struct edid_quirk_force_mode *quirk_mode;
+     int i, num_modes = 0;
+

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2291

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +10 -0
+     for (i = 0; i < ARRAY_SIZE(edid_quirk_force_mode_list); i++) {
+             quirk_mode = &edid_quirk_force_mode_list[i];
+
+             if (edid_vendor(edid, quirk_mode->vendor) &&
+                 (EDID_PRODUCT_ID(edid) == quirk_mode->product_id)) {
+                     num_modes = do_force_quirk_modes(connector,
+                                     quirk_mode->hdisplay,
+                                     quirk_mode->vdisplay,
+                                     quirk_mode->vrefresh,
+                                     quirk_mode->reduced);

I was wondering why you don't bail out here. Maybe you want to be able to add more than one quirk mode? In that case you should +=, not = the num_modes.

Quoted 3 lines

(Note that then do_force_quirk_modes makes the last one in the array the preferred mode, clearing the DRM_MODE_TYPE_PREFERRED from the previously added quirk modes.)

I thought about whether multiple forced modes would ever be needed. I can't imagine it would but it's easy enough to allow for it so we might as well implement it now.

I think we'd want for it to be clear which of the forced modes will be used and
so the current behavior of making the last one in the array preferred may be the best. I will add a comment by the edid_quirk_force_mode_list describing this.

Quoted 1 line

Review int drm_add_edid_modes(struct drm_connector

Quoted 1 line
(quoted diff) +0 -0
@@ -2803,6 +2879,8 @@ int drm_add_edid_modes(struct drm_connector

*connector, struct edid *edid)

Quoted 2 lines

Review int drm_add_edid_modes(struct drm_connector · line 2879

Quoted 2 lines
(quoted diff) +0 -0
      if (quirks & (EDID_QUIRK_PREFER_LARGE_60 |

EDID_QUIRK_PREFER_LARGE_75))

Quoted 3 lines

drivers/gpu/drm/drm_edid.c int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)

Quoted 3 lines
drivers/gpu/drm/drm_edid.c +2 -0
              edid_fixup_preferred(connector, quirks);
+     if (quirks & EDID_QUIRK_FORCE_MODE)
+             num_modes += add_force_quirk_modes(connector, edid,

quirks);

Quoted 2 lines

You don't use quirks within add_force_quirk_modes() for anything.

good point. I'll remove it.

Thanks,
Dylan

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

#4 lore raw

Re: [PATCH v3 1/2] drm: Enhance EDID quirks to explicitly set a mode

Quoted 50 lines · Dylan Semler
Quoted 4 lines · Jani Nikula

Hi, please find some review comments inline.

You make some good points, thanks.

Quoted 9 lines · Dylan Semler

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2238

Quoted 6 lines
drivers/gpu/drm/drm_edid.c +5 -0
+/* Add an explicit mode based on a quirk
+ */
+static int
+do_force_quirk_modes(struct drm_connector *connector, int hdisplay,
+                  int vdisplay, int vrefresh, bool reduced)

Nitpick: This adds one mode, not many _modes_.

sure, I'll fix.

Quoted 3 lines

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid,

Quoted 3 lines
drivers/gpu/drm/drm_edid.c +3 -0
+
+     /* loop through the probed modes and clear the preferred bit */
+     list_for_each_entry_safe(cur_mode, t, &connector->probed_modes,

head)

Quoted 5 lines
Quoted 1 line

+ cur_mode->type &= ~DRM_MODE_TYPE_PREFERRED;

You're not deleting entries, so list_for_each_entry() would suffice, getting rid of the temp variable.

I was going to use this but I wasn't sure if I understood all of the differences between the two. I opted for _safe because I noticed it's used in
edid_fixup_preferred() even though that routine doesn't remove entries either.
I'll fix it.

Quoted 2 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+     mode = drm_cvt_mode(dev, hdisplay, vdisplay, vrefresh, reduced,

0, 0);

Quoted 7 lines

You could

        if (!mode)
                return 0;

here and get rid of the num_modes variable.

Quoted 8 lines

Yes, you're right. I went with

if (mode) {
    ...
    return 1;
}
else
    return 0;

Note that you should use braces in all branches if one branch requires them. See Documentation/CodingStyle.

Quoted 3 lines

I would naively expect this to be better. Is there a reason to prefer if (!mode)?

Just a personal preference of checking errors and bailing out early, and handling the happy day scenario at the top indentation level. I won't insist.

BR,
Jani.

Quoted 56 lines
Quoted 3 lines

Review

Quoted 3 lines
(quoted diff) +2 -0
+static int
+add_force_quirk_modes(struct drm_connector *connector, struct edid

*edid,

Quoted 20 lines
Quoted 5 lines
drivers/gpu/drm/drm_edid.c +5 -0
+                   u32 quirks)
+{
+     struct edid_quirk_force_mode *quirk_mode;
+     int i, num_modes = 0;
+

drivers/gpu/drm/drm_edid.c add_detailed_modes(struct drm_connector *connector, struct edid *edid, · line 2291

Quoted 10 lines
drivers/gpu/drm/drm_edid.c +10 -0
+     for (i = 0; i < ARRAY_SIZE(edid_quirk_force_mode_list); i++) {
+             quirk_mode = &edid_quirk_force_mode_list[i];
+
+             if (edid_vendor(edid, quirk_mode->vendor) &&
+                 (EDID_PRODUCT_ID(edid) == quirk_mode->product_id)) {
+                     num_modes = do_force_quirk_modes(connector,
+                                     quirk_mode->hdisplay,
+                                     quirk_mode->vdisplay,
+                                     quirk_mode->vrefresh,
+                                     quirk_mode->reduced);

I was wondering why you don't bail out here. Maybe you want to be able to add more than one quirk mode? In that case you should +=, not = the num_modes.

Quoted 3 lines

(Note that then do_force_quirk_modes makes the last one in the array the preferred mode, clearing the DRM_MODE_TYPE_PREFERRED from the previously added quirk modes.)

I thought about whether multiple forced modes would ever be needed. I can't imagine it would but it's easy enough to allow for it so we might as well implement it now.

I think we'd want for it to be clear which of the forced modes will be used and
so the current behavior of making the last one in the array preferred may be the best. I will add a comment by the edid_quirk_force_mode_list describing this.

Quoted 1 line

Review int drm_add_edid_modes(struct drm_connector

Quoted 1 line
(quoted diff) +0 -0
@@ -2803,6 +2879,8 @@ int drm_add_edid_modes(struct drm_connector

*connector, struct edid *edid)

Quoted 2 lines

Review int drm_add_edid_modes(struct drm_connector · line 2879

Quoted 2 lines
(quoted diff) +0 -0
      if (quirks & (EDID_QUIRK_PREFER_LARGE_60 |

EDID_QUIRK_PREFER_LARGE_75))

Quoted 3 lines

drivers/gpu/drm/drm_edid.c int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)

Quoted 3 lines
drivers/gpu/drm/drm_edid.c +2 -0
              edid_fixup_preferred(connector, quirks);
+     if (quirks & EDID_QUIRK_FORCE_MODE)
+             num_modes += add_force_quirk_modes(connector, edid,

quirks);

Quoted 2 lines

You don't use quirks within add_force_quirk_modes() for anything.

good point. I'll remove it.

Thanks,
Dylan

#5 lore raw

PATCH v3 2/2 drm: Add EDID force quirk for MMT Monitor2Go HD+

In reply to Dylan Semler

Set the new EDID_QUIRK_FORCE_MODE bit for the MMT Monitor2Go HD+ monitor and add it to the edid_quirk_force_mode_list.

---
 drivers/gpu/drm/drm_edid.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/drm_edid.c +5 -0
index 38b8641..2f30be4 100644
@@ -109,6 +109,9 @@ static struct edid_quirk {
 	{ "FCM", 13600, EDID_QUIRK_PREFER_LARGE_75 |
 	  EDID_QUIRK_DETAILED_IN_CM },
 
+	/* Mobile Monitor Technologies LLC, Monitor2Go HD+ */
+	{ "LLP", 0x4c54, EDID_QUIRK_FORCE_MODE},
+
 	/* LG Philips LCD LP154W01-A5 */
 	{ "LPL", 0, EDID_QUIRK_DETAILED_USE_MAXIMUM_SIZE },
 	{ "LPL", 0x2a00, EDID_QUIRK_DETAILED_USE_MAXIMUM_SIZE },
@@ -137,6 +140,8 @@ static struct edid_quirk_force_mode {
 	int vrefresh;
 	bool reduced;
 } edid_quirk_force_mode_list[] = {
+	/* Mobile Monitor Technologies LLC, Monitor2Go HD+ */
+	{ "LLP", 0x4c54, 1600, 900, 60, 1 },
 };
 
 /*
Signature · 1 line
1.7.11.7
Acked-by on the v3 series
#6 lore raw
In reply to Dylan Semler
Quoted 20 lines · Dylan Semler

Changes in this version
* Uses drm_cvt_mode() instead of drm_gtf_mode() to build modeline
* Adds bool to specify reduced blanking to edid_quirk_force_mode
* Removes preferred bit from all other modes

There is at least one monitor that doesn't report its native resolution in its EDID block. This enhancement extends the EDID quirk logic to make monitors like this "just work".

The first patch in this series sets up a new quirk list where monitors' correct width, height, refresh rate, and reduced blanking parameters are specified. When a matching monitor is attached the full mode is calculated with drm_cvt_mode() and added to the connector. The DRM_MODE_TYPE_PREFERRED bit is set on the new mode and unset from all other modes.

The first patch also defines a new quirk bit: EDID_QUIRK_FORCE_MODE. This bit needs to be set for the new quirk list discribed above to be checked.

Quoted 5 lines

The second patch adds the offending monitor to the quirk lists.

Dylan Semler (2):
drm: Enhance EDID quirks to explicitly set a mode
drm: Add EDID force quirk for MMT Monitor2Go HD+

Ah, missed this resend of your patches. Looks good to me, but I guess Dave want's to have Ajax's ack, too. Anyway for the series:

Cheers, Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch