If colorop TYPE requires a data blob, userspace have to set a blob
whenever enables this colorop, i.e. when setting this colorop bypass
property to false. Clarify this in the documentation and reject the
atomic commit if colorop TYPE requires a DATA but set bypass to false
and no blob submitted.

Fixes: e5719e7f1900 ("drm/colorop: Add 3x4 CTM type")
Reviewed-by: Harry Wentland <[email protected]>
Signed-off-by: Melissa Wen <[email protected]>
---
 drivers/gpu/drm/drm_atomic.c  | 22 ++++++++++++++++++++--
 drivers/gpu/drm/drm_colorop.c |  5 ++++-
 include/drm/drm_colorop.h     |  6 ++++--
 3 files changed, 28 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 36c9550c2749..d8b56fa47d45 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -990,8 +990,13 @@ drm_atomic_add_pipeline_colorops(struct drm_atomic_commit 
*state,
  * drm_atomic_colorop_check - check new colorop state
  * @new_colorop_state: new colorop state to check
  *
- * Ensure that the colorop in @new_colorop_state belongs to an active color
- * pipeline, i.e. it's in the chain of colorops set to the color_pipeline
+ * Check that a colorop whose TYPE requires a data blob has one when it's
+ * enabled, i.e. userspace can't clear (or never set) the DATA property while
+ * taking the colorop out of bypass, since drivers would have nothing to
+ * program.
+ *
+ * Also ensure that the colorop in @new_colorop_state belongs to an active
+ * color pipeline, i.e. it's in the chain of colorops set to the color_pipeline
  * property of current, old or new plane state.
  *
  * Userspace is allowed to finalize colorop's settings in the same commit that
@@ -1015,6 +1020,19 @@ static int drm_atomic_colorop_check(const struct 
drm_colorop_state *new_colorop_
        if (state->duplicated)
                return 0;
 
+       /*
+        * Reject if colorop TYPE requires a DATA but set bypass to false and
+        * no blob submitted
+        */
+       if (new_colorop_state->colorop->data_property &&
+           !new_colorop_state->bypass && !new_colorop_state->data) {
+               drm_dbg_atomic(new_colorop_state->colorop->dev,
+                              "[COLOROP:%d:%d] enabled without a DATA blob\n",
+                              new_colorop_state->colorop->base.id,
+                              new_colorop_state->colorop->type);
+               return -EINVAL;
+       }
+
        /* Not a plane colorop */
        if (!plane)
                return 0;
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 95da03d6c65c..842fd7dc31e6 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -55,7 +55,10 @@
  * enumerated 1D curves, 1D LUTs, 3D LUTs, matrices, etc. See the
  * &drm_colorop_type documentation for information on each type.
  *
- * If a colorop advertises the BYPASS property it can be bypassed.
+ * If a colorop advertises the BYPASS property it can be bypassed. If a
+ * colorop advertises DATA and BYPASS properties, BYPASS can only be set to
+ * false if DATA contains a blob. An atomic commit that leaves such a
+ * colorop enabled without a blob will fail, returning -EINVAL.
  *
  * Information about colorop and color pipeline design decisions can be
  * found at rfc/color_pipeline.rst, but note that this document will
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 11a88e9477f1..952c30b06d83 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -283,8 +283,10 @@ struct drm_colorop_state {
        /**
         * @data:
         *
-        * Data blob for any TYPE that requires such a blob. The
-        * interpretation of the blob is TYPE-specific.
+        * Data blob for any TYPE that requires such a blob. The interpretation
+        * of the blob is TYPE-specific. Whenever @bypass is false, @data must
+        * be provided; otherwise, the atomic check will reject an active
+        * colorop without a blob.
         *
         * See the &drm_colorop_type documentation for how blob is laid
         * out.
-- 
2.53.0

Reply via email to