Wavering Arrandale output

PATCH Gustavo F. Padovan 2 participants 4 messages May 13, 2011 – Jun 24, 2011

Subsystems (2)

DRM drivers Intel i915

lore

Share & tools

lore

Gustavo F. Padovan

author

2011-05-13 20:44 UTC #0 lore raw

Hi Chris,

Bug #28306 is going to complete his first birthday at the end of this month, it is nagging around for a long time and I have many interest in fix this.

So I tested your patch in duplicated bug #36599 and it doesn't work, but I managed to add some printks(see diff) there to help debug.

[    1.033866] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.033868] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.033869] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 1
[    1.034120] [drm:intel_crtc_mode_set] *ERROR* is_lvds 1
[    1.034121] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
(null)
[    1.191465] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.191466] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.191468] [drm:intel_crtc_mode_set] *ERROR* encoder->type 1
[    1.191469] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 0
[    1.191720] [drm:intel_crtc_mode_set] *ERROR* is_lvds 0
[    1.191721] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder

(null)

It seems your patch has no effect once has_edp_encoder is null. Can you please look into this? I can provide any test you may want, just send me the patches. ;)

Regards,

[0] https://bugs.freedesktop.org/show_bug.cgi?id=28306
[1] https://bugs.freedesktop.org/show_bug.cgi?id=36599

drivers/gpu/drm/i915/intel_display.c +15 -2

index c38a001..a3c53cb 100644
@@ -4744,26 +4744,37 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
        if (HAS_PCH_SPLIT(dev)) {
                bool use_ssc = intel_panel_use_ssc(dev_priv);
 
+               DRM_ERROR("1. use_ssc %d", use_ssc);
+
                /* Are any non-SSC outputs enabled? */
                list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
                        if (encoder->base.crtc == NULL)
                                continue;
 
-                       if (encoder->type == INTEL_OUTPUT_ANALOG)
+                       DRM_ERROR("encoder->type %d", encoder->type);
+                       if (encoder->type == INTEL_OUTPUT_ANALOG) {
                                use_ssc = false;
+                               break;
+                       }
                }
 
+               DRM_ERROR("2. use_ssc %d", use_ssc);
+
                temp = I915_READ(PCH_DREF_CONTROL);
                /* Always enable nonspread source */
                temp &= ~DREF_NONSPREAD_SOURCE_MASK;
                temp |= DREF_NONSPREAD_SOURCE_ENABLE;
                temp &= ~DREF_SSC_SOURCE_MASK;
-               temp |= DREF_SSC_SOURCE_ENABLE;
+               if (use_ssc)
+                       temp |= DREF_SSC_SOURCE_ENABLE;
                I915_WRITE(PCH_DREF_CONTROL, temp);
 
                POSTING_READ(PCH_DREF_CONTROL);
                udelay(200);
 
+               DRM_ERROR("is_lvds %d", is_lvds);
+               DRM_ERROR("has_edp_encoder %p", has_edp_encoder);
+
                if (has_edp_encoder) {
                        if (use_ssc) {
                                temp |= DREF_SSC1_ENABLE;

@@ -4775,6 +4786,8 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
                        temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
 
                        /* Enable CPU source on CPU attached eDP */
+                       DRM_ERROR("!intel_encoder_is_pch_edp(&has_edp_encoder->base) %d",
+                               !intel_encoder_is_pch_edp(&has_edp_encoder->base));
                        if (!intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
                                if (use_ssc)

Signature · 2 lines

Gustavo F. Padovan
http://profusion.mobi

Chris Wilson

2011-05-14 07:55 UTC #1 lore raw

Quoted 20 lines · Gustavo F. Padovan

Hi Chris,

Bug #28306 is going to complete his first birthday at the end of this month, it is nagging around for a long time and I have many interest in fix this.

So I tested your patch in duplicated bug #36599 and it doesn't work, but I managed to add some printks(see diff) there to help debug.

[    1.033866] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.033868] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.033869] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 1
[    1.034120] [drm:intel_crtc_mode_set] *ERROR* is_lvds 1
[    1.034121] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
(null)
[    1.191465] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.191466] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.191468] [drm:intel_crtc_mode_set] *ERROR* encoder->type 1
[    1.191469] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 0
[    1.191720] [drm:intel_crtc_mode_set] *ERROR* is_lvds 0
[    1.191721] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder

Quoted 5 lines

(null)
It seems your patch has no effect once has_edp_encoder is null. Can you please look into this? I can provide any test you may want, just send me the patches. ;)

With no eDP on the system, we wouldn't expect to apply any of the eDP settings.

Even after your fix to the patch, nothing? Can you try just setting use_ssc=0 (and keep your fix).

In 633f2ea26665d37bb3c8ae30799aa14988622653 (drm/i915: Disable SSC for outputs other than LVDS or DP) I tried to implement another aspect of source-clock control, that is the patch that appeared to work if you only had VGA connected. So the next step would be to combine the two patches. -Chris

Signature · 1 line

Chris Wilson, Intel Open Source Technology Centre

Gustavo F. Padovan

author

2011-05-14 19:31 UTC #2 lore raw

Quoted 27 lines · Chris Wilson

Quoted 20 lines · Gustavo F. Padovan
Hi Chris,
Bug #28306 is going to complete his first birthday at the end of this month, it is nagging around for a long time and I have many interest in fix this.
So I tested your patch in duplicated bug #36599 and it doesn't work, but I managed to add some printks(see diff) there to help debug.
[    1.033866] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.033868] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.033869] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 1
[    1.034120] [drm:intel_crtc_mode_set] *ERROR* is_lvds 1
[    1.034121] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
(null)
[    1.191465] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.191466] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.191468] [drm:intel_crtc_mode_set] *ERROR* encoder->type 1
[    1.191469] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 0
[    1.191720] [drm:intel_crtc_mode_set] *ERROR* is_lvds 0
[    1.191721] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
Quoted 5 lines
(null)
It seems your patch has no effect once has_edp_encoder is null. Can you please look into this? I can provide any test you may want, just send me the patches. ;)

Quoted 5 lines

With no eDP on the system, we wouldn't expect to apply any of the eDP settings.
Even after your fix to the patch, nothing? Can you try just setting use_ssc=0 (and keep your fix).

Nothing.

Signature · 2 lines

Gustavo F. Padovan
http://profusion.mobi

Gustavo F. Padovan

author

2011-06-24 20:38 UTC #3 lore raw

In reply to Chris Wilson

Quoted 30 lines · Chris Wilson

Quoted 20 lines · Gustavo F. Padovan
Hi Chris,
Bug #28306 is going to complete his first birthday at the end of this month, it is nagging around for a long time and I have many interest in fix this.
So I tested your patch in duplicated bug #36599 and it doesn't work, but I managed to add some printks(see diff) there to help debug.
[    1.033866] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.033868] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.033869] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 1
[    1.034120] [drm:intel_crtc_mode_set] *ERROR* is_lvds 1
[    1.034121] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
(null)
[    1.191465] [drm:intel_crtc_mode_set] *ERROR* 1. use_ssc 1
[    1.191466] [drm:intel_crtc_mode_set] *ERROR* encoder->type 4
[    1.191468] [drm:intel_crtc_mode_set] *ERROR* encoder->type 1
[    1.191469] [drm:intel_crtc_mode_set] *ERROR* 2. use_ssc 0
[    1.191720] [drm:intel_crtc_mode_set] *ERROR* is_lvds 0
[    1.191721] [drm:intel_crtc_mode_set] *ERROR* has_edp_encoder
Quoted 5 lines
(null)
It seems your patch has no effect once has_edp_encoder is null. Can you please look into this? I can provide any test you may want, just send me the patches. ;)
With no eDP on the system, we wouldn't expect to apply any of the eDP settings.

Quoted 7 lines

Even after your fix to the patch, nothing? Can you try just setting use_ssc=0 (and keep your fix).
In 633f2ea26665d37bb3c8ae30799aa14988622653 (drm/i915: Disable SSC for outputs other than LVDS or DP) I tried to implement another aspect of source-clock control, that is the patch that appeared to work if you only had VGA connected. So the next step would be to combine the two patches.

No fixes for this yet. There are many devices out there with this bug. Can someone please take a look on this? Just send me patches, I can test them.

Gustavo