Hello,
This series is to fix an issue that surfaced after damage clipping was
enabled for the virtio-gpu by commit 01f05940a9a7 ("drm/virtio: Enable
fb damage clips property for the primary plane").
After that change, flickering artifacts was reported to be present with
both weston and wlroots wayland compositors when running in a virtual
machine. The cause was identified by Sima Vetter, who pointed out that
virtio-gpu does per-buffer uploads and for this reason it needs to do
a buffer damage handling, instead of frame damage handling.
Their suggestion was to extend the damage helpers to cover that case
and given that there's isn't a buffer damage accumulation algorithm
(e.g: buffer age), just do a full plane update if the framebuffer that
is attached to a plane changed since the last plane update (page-flip).
It is a v4 that addresses issues pointed out by Sima Vetter in v3:
https://lists.freedesktop.org/archives/dri-devel/2023-November/431409.html
Patch #1 adds a ignore_damage_clips field to struct drm_plane_state to be
set by drivers that want the damage helpers to ignore the damage clips.
Patch #2 fixes the virtio-gpu damage handling logic by asking the damage
helper to ignore the damage clips if the framebuffer attached to a plane
has changed since the last page-flip.
Patch #3 does the same but for the vmwgfx driver that also needs to handle
buffer damage and should have the same issue (although I haven't tested it
due not having a VMWare setup).
Patch #4 adds to the KMS damage tracking kernel-doc some paragraphs about
damage tracking types and references to links that explain frame damage vs
buffer damage.
Finally patch #5 adds an item to the DRM todo, about the need to implement
some buffer damage accumulation algorithm instead of just doing full plane
updates in this case.
Because commit 01f05940a9a7 landed in v6.4, the first 2 patches are marked
as Fixes and Cc stable.
I've tested this on a VM with weston, was able to reproduce the issue
reported and the patches did fix the problem.
Best regards,
Javier
Changes in v4:
- Refer in ignore_damage_clips kernel-doc to "Damage Tracking Properties"
KMS documentation section (Sima Vetter).
- Add another paragraph to "Damage Tracking Properties" section to mention
the fields that drivers with per-buffer upload target should check to set
drm_plane_state.ignore_damage_clips (Sima Vetter).
- Reference the &drm_plane_state.ignore_damage_clips and the damage helpers
in the buffer damage TODO entry (Sima Vetter).
Changes in v3:
- Fix typo in the kernel-doc (Simon Ser).
- Add a paragraph explaining what the problem in the kernel is and
make it clear that the refeference documents are related to how
user-space handles this case (Thomas Zimmermann).
Changes in v2:
- Add a struct drm_plane_state .ignore_damage_clips to set in the plane's
.atomic_check, instead of having different helpers (Thomas Zimmermann).
- Set struct drm_plane_state .ignore_damage_clips in virtio-gpu plane's
.atomic_check instead of using a different helpers (Thomas Zimmermann).
- Set struct drm_plane_state .ignore_damage_clips in vmwgfx plane's
.atomic_check instead of using a different helpers (Thomas Zimmermann).
Javier Martinez Canillas (5):
drm: Allow drivers to indicate the damage helpers to ignore damage
clips
drm/virtio: Disable damage clipping if FB changed since last page-flip
drm/vmwgfx: Disable damage clipping if FB changed since last page-flip
drm/plane: Extend damage tracking kernel-doc
drm/todo: Add entry about implementing buffer age for damage tracking
Documentation/gpu/drm-kms.rst | 2 ++
Documentation/gpu/todo.rst | 23 ++++++++++++++++++++
drivers/gpu/drm/drm_damage_helper.c | 3 ++-
drivers/gpu/drm/drm_plane.c | 30 ++++++++++++++++++++++++++
drivers/gpu/drm/virtio/virtgpu_plane.c | 10 +++++++++
drivers/gpu/drm/vmwgfx/vmwgfx_kms.c | 11 ++++++++++
include/drm/drm_plane.h | 10 +++++++++
7 files changed, 88 insertions(+), 1 deletion(-)
--
2.41.0
The driver does per-buffer uploads and needs to force a full plane update
if the plane's attached framebuffer has change since the last page-flip.
Suggested-by: Sima Vetter <[email protected]>
Signed-off-by: Javier Martinez Canillas <[email protected]>
Reviewed-by: Thomas Zimmermann <[email protected]>
Reviewed-by: Zack Rusin <[email protected]>
Acked-by: Sima Vetter <[email protected]>
---
(no changes since v2)
Changes in v2:
- Set struct drm_plane_state .ignore_damage_clips in vmwgfx plane's
.atomic_check instead of using a different helpers (Thomas Zimmermann).
drivers/gpu/drm/vmwgfx/vmwgfx_kms.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 818b7f109f53..f9364bf222e3 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -837,10 +837,21 @@ int vmw_du_primary_plane_atomic_check(struct drm_plane *plane,
{
struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
plane);
+ struct drm_plane_state *old_state = drm_atomic_get_old_plane_state(state,
+ plane);
struct drm_crtc_state *crtc_state = NULL;
struct drm_framebuffer *new_fb = new_state->fb;
+ struct drm_framebuffer *old_fb = old_state->fb;
int ret;
+ /*
+ * Ignore damage clips if the framebuffer attached to the plane's state
+ * has changed since the last plane update (page-flip). In this case, a
+ * full plane update should happen because uploads are done per-buffer.
+ */
+ if (old_fb != new_fb)
+ new_state->ignore_damage_clips = true;
+
if (new_state->crtc)
crtc_state = drm_atomic_get_new_crtc_state(state,
new_state->crtc);
--
2.41.0
The driver does per-buffer uploads and needs to force a full plane update
if the plane's attached framebuffer has change since the last page-flip.
Fixes: 01f05940a9a7 ("drm/virtio: Enable fb damage clips property for the primary plane")
Cc: <[email protected]> # v6.4+
Reported-by: nerdopolis <[email protected]>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218115
Suggested-by: Sima Vetter <[email protected]>
Signed-off-by: Javier Martinez Canillas <[email protected]>
Reviewed-by: Thomas Zimmermann <[email protected]>
Reviewed-by: Zack Rusin <[email protected]>
Acked-by: Sima Vetter <[email protected]>
---
(no changes since v2)
Changes in v2:
- Set struct drm_plane_state .ignore_damage_clips in virtio-gpu plane's
.atomic_check instead of using a different helpers (Thomas Zimmermann).
drivers/gpu/drm/virtio/virtgpu_plane.c | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/drivers/gpu/drm/virtio/virtgpu_plane.c b/drivers/gpu/drm/virtio/virtgpu_plane.c
index a2e045f3a000..a1ef657eba07 100644
--- a/drivers/gpu/drm/virtio/virtgpu_plane.c
+++ b/drivers/gpu/drm/virtio/virtgpu_plane.c
@@ -79,6 +79,8 @@ static int virtio_gpu_plane_atomic_check(struct drm_plane *plane,
{
struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
plane);
+ struct drm_plane_state *old_plane_state = drm_atomic_get_old_plane_state(state,
+ plane);
bool is_cursor = plane->type == DRM_PLANE_TYPE_CURSOR;
struct drm_crtc_state *crtc_state;
int ret;
@@ -86,6 +88,14 @@ static int virtio_gpu_plane_atomic_check(struct drm_plane *plane,
if (!new_plane_state->fb || WARN_ON(!new_plane_state->crtc))
return 0;
+ /*
+ * Ignore damage clips if the framebuffer attached to the plane's state
+ * has changed since the last plane update (page-flip). In this case, a
+ * full plane update should happen because uploads are done per-buffer.
+ */
+ if (old_plane_state->fb != new_plane_state->fb)
+ new_plane_state->ignore_damage_clips = true;
+
crtc_state = drm_atomic_get_crtc_state(state,
new_plane_state->crtc);
if (IS_ERR(crtc_state))
--
2.41.0
The "Damage Tracking Properties" section in the documentation doesn't have
info about the two type of damage handling: frame damage vs buffer damage.
Add it to the section and mention that helpers only support frame damage,
and how drivers handling buffer damage can indicate that the damage clips
should be ignored.
Also add references to further documentation about the two damage types.
Suggested-by: Simon Ser <[email protected]>
Signed-off-by: Javier Martinez Canillas <[email protected]>
Reviewed-by: Simon Ser <[email protected]>
Reviewed-by: Thomas Zimmermann <[email protected]>
Reviewed-by: Zack Rusin <[email protected]>
Acked-by: Sima Vetter <[email protected]>
---
Changes in v4:
- Add another paragraph to "Damage Tracking Properties" section to mention
the fields that drivers with per-buffer upload target should check to set
drm_plane_state.ignore_damage_clips (Sima Vetter).
Changes in v3:
- Fix typo in the kernel-doc (Simon Ser).
- Add a paragraph explaining what the problem in the kernel is and
make it clear that the refeference documents are related to how
user-space handles this case (Thomas Zimmermann).
drivers/gpu/drm/drm_plane.c | 30 ++++++++++++++++++++++++++++++
1 file changed, 30 insertions(+)
diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c
index 24e7998d1731..662e0ba2707a 100644
--- a/drivers/gpu/drm/drm_plane.c
+++ b/drivers/gpu/drm/drm_plane.c
@@ -1442,6 +1442,36 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
* Drivers implementing damage can use drm_atomic_helper_damage_iter_init() and
* drm_atomic_helper_damage_iter_next() helper iterator function to get damage
* rectangles clipped to &drm_plane_state.src.
+ *
+ * Note that there are two types of damage handling: frame damage and buffer
+ * damage, the type of damage handling implemented depends on a driver's upload
+ * target. Drivers implementing a per-plane or per-CRTC upload target need to
+ * handle frame damage, while drivers implementing a per-buffer upload target
+ * need to handle buffer damage.
+ *
+ * The existing damage helpers only support the frame damage type, there is no
+ * buffer age support or similar damage accumulation algorithm implemented yet.
+ *
+ * Only drivers handling frame damage can use the mentioned damage helpers to
+ * iterate over the damaged regions. Drivers that handle buffer damage, must set
+ * &drm_plane_state.ignore_damage_clips for drm_atomic_helper_damage_iter_init()
+ * to know that damage clips should be ignored and return &drm_plane_state.src
+ * as the damage rectangle, to force a full plane update.
+ *
+ * Drivers with a per-buffer upload target could compare the &drm_plane_state.fb
+ * of the old and new plane states to determine if the framebuffer attached to a
+ * plane has changed or not since the last plane update. If &drm_plane_state.fb
+ * has changed, then &drm_plane_state.ignore_damage_clips must be set to true.
+ *
+ * That is because drivers with a per-plane upload target, expect the backing
+ * storage buffer to not change for a given plane. If the upload buffer changes
+ * between page flips, the new upload buffer has to be updated as a whole. This
+ * can be improved in the future if support for frame damage is added to the DRM
+ * damage helpers, similarly to how user-space already handle this case as it is
+ * explained in the following documents:
+ *
+ * https://registry.khronos.org/EGL/extensions/KHR/EGL_KHR_swap_buffers_with_damage.txt
+ * https://emersion.fr/blog/2019/intro-to-damage-tracking/
*/
/**
--
2.41.0
Currently, only damage tracking for frame damage is supported. If a driver
needs to do buffer damage (e.g: the framebuffer attached to plane's state
has changed since the last page-flip), the damage helpers just fallback to
a full plane update.
Add en entry in the TODO about implementing buffer age or any other damage
accumulation algorithm for buffer damage handling.
Suggested-by: Simon Ser <[email protected]>
Signed-off-by: Javier Martinez Canillas <[email protected]>
Reviewed-by: Simon Ser <[email protected]>
Reviewed-by: Thomas Zimmermann <[email protected]>
Reviewed-by: Zack Rusin <[email protected]>
Acked-by: Sima Vetter <[email protected]>
---
Changes in v4:
- Reference the &drm_plane_state.ignore_damage_clips and the damage helpers
in the buffer damage TODO entry (Sima Vetter).
Documentation/gpu/todo.rst | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst
index b62c7fa0c2bc..503d57c75215 100644
--- a/Documentation/gpu/todo.rst
+++ b/Documentation/gpu/todo.rst
@@ -782,6 +782,29 @@ Contact: Hans de Goede
Level: Advanced
+Buffer age or other damage accumulation algorithm for buffer damage
+===================================================================
+
+Drivers that do per-buffer uploads, need a buffer damage handling (rather than
+frame damage like drivers that do per-plane or per-CRTC uploads), but there is
+no support to get the buffer age or any other damage accumulation algorithm.
+
+For this reason, the damage helpers just fallback to a full plane update if the
+framebuffer attached to a plane has changed since the last page-flip. Drivers
+set &drm_plane_state.ignore_damage_clips to true as indication to
+drm_atomic_helper_damage_iter_init() and drm_atomic_helper_damage_iter_next()
+helpers that the damage clips should be ignored.
+
+This should be improved to get damage tracking properly working on drivers that
+do per-buffer uploads.
+
+More information about damage tracking and references to learning materials can
+be found in :ref:`damage_tracking_properties`.
+
+Contact: Javier Martinez Canillas <[email protected]>
+
+Level: Advanced
+
Outside DRM
===========
--
2.41.0
It allows drivers to set a struct drm_plane_state .ignore_damage_clips in
their plane's .atomic_check callback, as an indication to damage helpers
such as drm_atomic_helper_damage_iter_init() that the damage clips should
be ignored.
To be used by drivers that do per-buffer (e.g: virtio-gpu) uploads (rather
than per-plane uploads), since these type of drivers need to handle buffer
damages instead of frame damages.
That way, these drivers could force a full plane update if the framebuffer
attached to a plane's state has changed since the last update (page-flip).
Fixes: 01f05940a9a7 ("drm/virtio: Enable fb damage clips property for the primary plane")
Cc: <[email protected]> # v6.4+
Reported-by: nerdopolis <[email protected]>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218115
Suggested-by: Thomas Zimmermann <[email protected]>
Signed-off-by: Javier Martinez Canillas <[email protected]>
Reviewed-by: Thomas Zimmermann <[email protected]>
Reviewed-by: Zack Rusin <[email protected]>
Acked-by: Sima Vetter <[email protected]>
---
Changes in v4:
- Refer in ignore_damage_clips kernel-doc to "Damage Tracking Properties"
KMS documentation section (Sima Vetter).
Changes in v2:
- Add a struct drm_plane_state .ignore_damage_clips to set in the plane's
.atomic_check, instead of having different helpers (Thomas Zimmermann).
Documentation/gpu/drm-kms.rst | 2 ++
drivers/gpu/drm/drm_damage_helper.c | 3 ++-
include/drm/drm_plane.h | 10 ++++++++++
3 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/Documentation/gpu/drm-kms.rst b/Documentation/gpu/drm-kms.rst
index 270d320407c7..a98a7e04e86f 100644
--- a/Documentation/gpu/drm-kms.rst
+++ b/Documentation/gpu/drm-kms.rst
@@ -548,6 +548,8 @@ Plane Composition Properties
.. kernel-doc:: drivers/gpu/drm/drm_blend.c
:doc: overview
+.. _damage_tracking_properties:
+
Damage Tracking Properties
--------------------------
diff --git a/drivers/gpu/drm/drm_damage_helper.c b/drivers/gpu/drm/drm_damage_helper.c
index d8b2955e88fd..afb02aae707b 100644
--- a/drivers/gpu/drm/drm_damage_helper.c
+++ b/drivers/gpu/drm/drm_damage_helper.c
@@ -241,7 +241,8 @@ drm_atomic_helper_damage_iter_init(struct drm_atomic_helper_damage_iter *iter,
iter->plane_src.x2 = (src.x2 >> 16) + !!(src.x2 & 0xFFFF);
iter->plane_src.y2 = (src.y2 >> 16) + !!(src.y2 & 0xFFFF);
- if (!iter->clips || !drm_rect_equals(&state->src, &old_state->src)) {
+ if (!iter->clips || state->ignore_damage_clips ||
+ !drm_rect_equals(&state->src, &old_state->src)) {
iter->clips = NULL;
iter->num_clips = 0;
iter->full_update = true;
diff --git a/include/drm/drm_plane.h b/include/drm/drm_plane.h
index 79d62856defb..fef775200a81 100644
--- a/include/drm/drm_plane.h
+++ b/include/drm/drm_plane.h
@@ -190,6 +190,16 @@ struct drm_plane_state {
*/
struct drm_property_blob *fb_damage_clips;
+ /**
+ * @ignore_damage_clips:
+ *
+ * Set by drivers to indicate the drm_atomic_helper_damage_iter_init()
+ * helper that the @fb_damage_clips blob property should be ignored.
+ *
+ * See :ref:`damage_tracking_properties` for more information.
+ */
+ bool ignore_damage_clips;
+
/**
* @src:
*
--
2.41.0
Javier Martinez Canillas <[email protected]> writes:
> Hello,
>
> This series is to fix an issue that surfaced after damage clipping was
> enabled for the virtio-gpu by commit 01f05940a9a7 ("drm/virtio: Enable
> fb damage clips property for the primary plane").
>
> After that change, flickering artifacts was reported to be present with
> both weston and wlroots wayland compositors when running in a virtual
> machine. The cause was identified by Sima Vetter, who pointed out that
> virtio-gpu does per-buffer uploads and for this reason it needs to do
> a buffer damage handling, instead of frame damage handling.
>
> Their suggestion was to extend the damage helpers to cover that case
> and given that there's isn't a buffer damage accumulation algorithm
> (e.g: buffer age), just do a full plane update if the framebuffer that
> is attached to a plane changed since the last plane update (page-flip).
>
> It is a v4 that addresses issues pointed out by Sima Vetter in v3:
>
> https://lists.freedesktop.org/archives/dri-devel/2023-November/431409.html
>
> Patch #1 adds a ignore_damage_clips field to struct drm_plane_state to be
> set by drivers that want the damage helpers to ignore the damage clips.
>
> Patch #2 fixes the virtio-gpu damage handling logic by asking the damage
> helper to ignore the damage clips if the framebuffer attached to a plane
> has changed since the last page-flip.
>
> Patch #3 does the same but for the vmwgfx driver that also needs to handle
> buffer damage and should have the same issue (although I haven't tested it
> due not having a VMWare setup).
>
> Patch #4 adds to the KMS damage tracking kernel-doc some paragraphs about
> damage tracking types and references to links that explain frame damage vs
> buffer damage.
>
> Finally patch #5 adds an item to the DRM todo, about the need to implement
> some buffer damage accumulation algorithm instead of just doing full plane
> updates in this case.
>
> Because commit 01f05940a9a7 landed in v6.4, the first 2 patches are marked
> as Fixes and Cc stable.
>
> I've tested this on a VM with weston, was able to reproduce the issue
> reported and the patches did fix the problem.
>
> Best regards,
> Javier
>
> Changes in v4:
> - Refer in ignore_damage_clips kernel-doc to "Damage Tracking Properties"
> KMS documentation section (Sima Vetter).
> - Add another paragraph to "Damage Tracking Properties" section to mention
> the fields that drivers with per-buffer upload target should check to set
> drm_plane_state.ignore_damage_clips (Sima Vetter).
> - Reference the &drm_plane_state.ignore_damage_clips and the damage helpers
> in the buffer damage TODO entry (Sima Vetter).
>
> Changes in v3:
> - Fix typo in the kernel-doc (Simon Ser).
> - Add a paragraph explaining what the problem in the kernel is and
> make it clear that the refeference documents are related to how
> user-space handles this case (Thomas Zimmermann).
>
> Changes in v2:
> - Add a struct drm_plane_state .ignore_damage_clips to set in the plane's
> .atomic_check, instead of having different helpers (Thomas Zimmermann).
> - Set struct drm_plane_state .ignore_damage_clips in virtio-gpu plane's
> .atomic_check instead of using a different helpers (Thomas Zimmermann).
> - Set struct drm_plane_state .ignore_damage_clips in vmwgfx plane's
> .atomic_check instead of using a different helpers (Thomas Zimmermann).
>
> Javier Martinez Canillas (5):
> drm: Allow drivers to indicate the damage helpers to ignore damage
> clips
> drm/virtio: Disable damage clipping if FB changed since last page-flip
> drm/vmwgfx: Disable damage clipping if FB changed since last page-flip
> drm/plane: Extend damage tracking kernel-doc
> drm/todo: Add entry about implementing buffer age for damage tracking
>
Pushed to drm-misc (drm-misc-next). Thanks!
--
Best regards,
Javier Martinez Canillas
Core Platforms
Red Hat