public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Hans de Goede <hdegoede@redhat.com>
To: "Maarten Lankhorst" <maarten.lankhorst@linux.intel.com>,
	"Jani Nikula" <jani.nikula@linux.intel.com>,
	"Joonas Lahtinen" <joonas.lahtinen@linux.intel.com>,
	"Rodrigo Vivi" <rodrigo.vivi@intel.com>,
	"Ville Syrjälä" <ville.syrjala@linux.intel.com>
Cc: intel-gfx <intel-gfx@lists.freedesktop.org>,
	dri-devel@lists.freedesktop.org
Subject: [PATCH] drm/i915: Program LUT before intel_color_commit() if LUT was not previously set
Date: Sun, 20 Oct 2019 20:19:33 +0200	[thread overview]
Message-ID: <20191020181933.54829-1-hdegoede@redhat.com> (raw)

Since commit 051a6d8d3ca0 ("drm/i915: Move LUT programming to happen after
vblank waits"), I am seeing an ugly colored flash of the first few display
lines on 2 Cherry Trail devices when the gamma table gets set for the first
time. A blue flash on a GPD win and a yellow flash on an Asus T100HA.

The problem is that since this change, the LUT is programmed after the
write *and latching* of the double-buffered register which causes the LUT
to be used starting at the next frame. This means that the old LUT is still
used for the first couple of lines of the display. If no LUT was in use
before then the LUT registers may contain bogus values. This leads to
messed up colors until the new LUT values are written. At least on CHT DSI
panels this causes messed up colors on the first few lines.

This commit fixes this by adding a load_lut_before_commit boolean,
modifying intel_begin_crtc_commit to load the luts earlier if this is set,
and setting this from intel_color_check when a LUT table was not in use
before (and thus may contain bogus values), or when the table size
changes.

Fixes: 051a6d8d3ca0 ("drm/i915: Move LUT programming to happen after vblank waits")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/gpu/drm/i915/display/intel_color.c    | 26 +++++++++++++++++++
 drivers/gpu/drm/i915/display/intel_display.c  |  7 +++++
 .../drm/i915/display/intel_display_types.h    |  3 +++
 3 files changed, 36 insertions(+)

diff --git a/drivers/gpu/drm/i915/display/intel_color.c b/drivers/gpu/drm/i915/display/intel_color.c
index 71a0201437a9..0da6dcc5bebd 100644
--- a/drivers/gpu/drm/i915/display/intel_color.c
+++ b/drivers/gpu/drm/i915/display/intel_color.c
@@ -1052,6 +1052,32 @@ intel_color_add_affected_planes(struct intel_crtc_state *new_crtc_state)
 		new_crtc_state->update_planes |= BIT(plane->id);
 	}
 
+	/*
+	 * Normally we load the LUTs after vblank / after the double-buffer
+	 * registers written by commit have been latched, this avoids a
+	 * gamma change mid-way the screen. This does mean that the first
+	 * few lines of the display will (sometimes) still use the old
+	 * table. This is fine when changing an existing LUT, but if this
+	 * is the first time the LUT gets loaded, then the hw may contain
+	 * random values, causing the first lines to have funky colors.
+	 *
+	 * So if were enabling a LUT for the first time or changing the table
+	 * size, then we must do this before the commit to avoid corrupting
+	 * the first lines of the display.
+	 */
+	if (!old_crtc_state->base.gamma_lut && new_crtc_state->base.gamma_lut)
+		new_crtc_state->load_lut_before_commit = true;
+	else if (!old_crtc_state->base.degamma_lut &&
+		 new_crtc_state->base.degamma_lut)
+		new_crtc_state->load_lut_before_commit = true;
+	else if (old_crtc_state->base.gamma_lut &&
+		 new_crtc_state->base.gamma_lut &&
+		 lut_is_legacy(old_crtc_state->base.gamma_lut) !=
+			lut_is_legacy(new_crtc_state->base.gamma_lut))
+		new_crtc_state->load_lut_before_commit = true;
+	else
+		new_crtc_state->load_lut_before_commit = false;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
index aa54bb22796d..21442b0dd134 100644
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -14033,6 +14033,7 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
 		if (new_crtc_state->base.active &&
 		    !needs_modeset(new_crtc_state) &&
+		    !new_crtc_state->load_lut_before_commit &&
 		    (new_crtc_state->base.color_mgmt_changed ||
 		     new_crtc_state->update_pipe))
 			intel_color_load_luts(new_crtc_state);
@@ -14529,6 +14530,12 @@ static void intel_begin_crtc_commit(struct intel_atomic_state *state,
 		intel_atomic_get_new_crtc_state(state, crtc);
 	bool modeset = needs_modeset(new_crtc_state);
 
+	if (!modeset &&
+	    new_crtc_state->load_lut_before_commit &&
+	    (new_crtc_state->base.color_mgmt_changed ||
+	     new_crtc_state->update_pipe))
+		intel_color_load_luts(new_crtc_state);
+
 	/* Perform vblank evasion around commit operation */
 	intel_pipe_update_start(new_crtc_state);
 
diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
index 449abaea619f..bbdeb3be64e6 100644
--- a/drivers/gpu/drm/i915/display/intel_display_types.h
+++ b/drivers/gpu/drm/i915/display/intel_display_types.h
@@ -973,6 +973,9 @@ struct intel_crtc_state {
 	/* enable pipe csc? */
 	bool csc_enable;
 
+	/* load luts before color settings commit */
+	bool load_lut_before_commit;
+
 	/* Display Stream compression state */
 	struct {
 		bool compression_enable;
-- 
2.23.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

             reply	other threads:[~2019-10-20 18:19 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-20 18:19 Hans de Goede [this message]
2019-10-20 18:34 ` ✗ Fi.CI.BUILD: failure for drm/i915: Program LUT before intel_color_commit() if LUT was not previously set Patchwork
2019-10-21  7:45 ` [PATCH] " Hans de Goede
2019-10-21 14:39 ` Ville Syrjälä
2019-10-25 19:23   ` Hans de Goede
2019-10-25 19:23     ` [Intel-gfx] " Hans de Goede
2019-10-25 19:45     ` Ville Syrjälä
2019-10-25 19:45       ` [Intel-gfx] " Ville Syrjälä
2019-10-28 18:56       ` Hans de Goede
2019-10-28 18:56         ` [Intel-gfx] " Hans de Goede

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191020181933.54829-1-hdegoede@redhat.com \
    --to=hdegoede@redhat.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jani.nikula@linux.intel.com \
    --cc=joonas.lahtinen@linux.intel.com \
    --cc=maarten.lankhorst@linux.intel.com \
    --cc=rodrigo.vivi@intel.com \
    --cc=ville.syrjala@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox