All of lore.kernel.org
 help / color / mirror / Atom feed
From: Maxime Ripard <mripard@kernel.org>
To: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
	 Thomas Zimmermann <tzimmermann@suse.de>,
	David Airlie <airlied@gmail.com>,
	 Simona Vetter <simona@ffwll.ch>,
	Louis Chauvet <louis.chauvet@bootlin.com>,
	 Haneen Mohammed <hamohammed.sa@gmail.com>,
	 Melissa Wen <melissa.srw@gmail.com>
Cc: "Daniel Stone" <daniels@collabora.com>,
	"Harry Wentland" <harry.wentland@amd.com>,
	"Jocelyn Falempe" <jfalempe@redhat.com>,
	"Jonas Ådahl" <jadahl@redhat.com>,
	"Michel Dänzer" <mdaenzer@redhat.com>,
	"Pekka Paalanen" <pekka.paalanen@collabora.com>,
	"Sebastian Wick" <sebastian.wick@redhat.com>,
	"Simon Ser" <contact@emersion.fr>,
	"Victoria Brekenfeld" <victoria@system76.com>,
	"Xaver Hugl" <xaver.hugl@kde.org>,
	dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	"Maxime Ripard" <mripard@kernel.org>
Subject: [PATCH v7 14/16] drm/atomic-uapi: Add DRM_MODE_ATOMIC_RESET flag
Date: Fri, 30 Oct 2026 11:27:46 +0100	[thread overview]
Message-ID: <20261030-drm-reset-state-flag-v7-14-4e55eaeae683@kernel.org> (raw)
In-Reply-To: <20261030-drm-reset-state-flag-v7-0-4e55eaeae683@kernel.org>

From: Sebastian Wick <sebastian.wick@redhat.com>

Userspace currently has no atomic way to reset all KMS object states
to their defaults. To bring a display pipeline to a known state, a
compositor must explicitly set every property on every object, which
requires tracking which properties exist and what their defaults are.

Introduce DRM_MODE_ATOMIC_RESET (0x0800) which, when passed to the
atomic ioctl, fills the commit with default states for all KMS
objects before applying the properties supplied in the request.
Properties not explicitly included in the commit remain at their
defaults (CRTCs inactive, planes disabled, connectors unbound, and
so on).

The flag cannot be combined with DRM_MODE_PAGE_FLIP_ASYNC, since a
full state reset is incompatible with an async flip.

Since the commit contains every CRTC of the device or lease, including
the ones that are disabled before and after it, no page flip event is
attached to those: they can't report a flip, and drm_atomic_crtc_check()
would reject the whole commit. Such a CRTC doesn't count towards the
CRTCs that can signal either, so a commit that leaves them all disabled
while asking for an event is still rejected instead of waiting for an
event that would never come.

Signed-off-by: Sebastian Wick <sebastian.wick@redhat.com>
Co-developed-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 drivers/gpu/drm/drm_atomic.c      |  3 +++
 drivers/gpu/drm/drm_atomic_uapi.c | 44 ++++++++++++++++++++++++++++++++++++---
 drivers/gpu/drm/drm_ioctl.c       |  4 ++++
 include/uapi/drm/drm.h            | 10 +++++++++
 include/uapi/drm/drm_mode.h       | 17 ++++++++++++++-
 5 files changed, 74 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 6bc0d6339f14..16530ce62455 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -1614,4 +1614,7 @@ EXPORT_SYMBOL(drm_atomic_get_new_bridge_state);
  * drm_atomic_helper_colorop_create_state() and do not have a per-driver hook.
  *
+ * This is used to report the %DRM_CAP_ATOMIC_RESET capability to userspace,
+ * and as a precondition in drm_atomic_commit_fill_with_defaults().
+ *
  * Returns:
  * True if all objects implement atomic_create_state, false otherwise.
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 1eebcf6f5e05..77d0713c7e2d 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -1441,8 +1441,27 @@ static int prepare_signaling(struct drm_device *dev,
 	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
 		s32 __user *fence_ptr;
+		bool page_flip_event;
 
 		fence_ptr = get_out_fence_for_crtc(crtc_state->state, crtc);
 
-		if (arg->flags & DRM_MODE_PAGE_FLIP_EVENT || fence_ptr) {
+		page_flip_event = arg->flags & DRM_MODE_PAGE_FLIP_EVENT;
+
+		/*
+		 * A commit with the DRM_MODE_ATOMIC_RESET flag contains
+		 * every CRTC of the device or lease, even the ones that are
+		 * disabled before and after the commit. Those can't report
+		 * a page flip, and drm_atomic_crtc_check() would reject the
+		 * whole commit, so don't attach any event to them.
+		 */
+		if (page_flip_event && !fence_ptr &&
+		    (arg->flags & DRM_MODE_ATOMIC_RESET) &&
+		    !crtc_state->active) {
+			struct drm_crtc_state *old_crtc_state;
+
+			old_crtc_state = drm_atomic_get_old_crtc_state(state, crtc);
+			page_flip_event = old_crtc_state->active;
+		}
+
+		if (page_flip_event || fence_ptr) {
 			struct drm_pending_vblank_event *e;
 
@@ -1454,5 +1473,5 @@ static int prepare_signaling(struct drm_device *dev,
 		}
 
-		if (arg->flags & DRM_MODE_PAGE_FLIP_EVENT) {
+		if (page_flip_event) {
 			struct drm_pending_vblank_event *e = crtc_state->event;
 
@@ -1495,5 +1514,6 @@ static int prepare_signaling(struct drm_device *dev,
 		}
 
-		c++;
+		if (page_flip_event)
+			c++;
 	}
 
@@ -1674,4 +1694,11 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
 	}
 
+	if ((arg->flags & DRM_MODE_ATOMIC_RESET) &&
+			(arg->flags & DRM_MODE_PAGE_FLIP_ASYNC)) {
+		drm_dbg_atomic(dev,
+			       "commit failed: reset cannot be combined with async flip\n");
+		return -EINVAL;
+	}
+
 	state = drm_atomic_commit_alloc(dev);
 	if (!state)
@@ -1689,4 +1716,15 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
 	num_fences = 0;
 
+	if (arg->flags & DRM_MODE_ATOMIC_RESET) {
+		if (!drm_atomic_implements_create_state(dev)) {
+			ret = -EOPNOTSUPP;
+			goto out;
+		}
+
+		ret = drm_atomic_commit_fill_with_defaults(file_priv, state);
+		if (ret)
+			goto out;
+	}
+
 	for (i = 0; i < arg->count_objs; i++) {
 		uint32_t obj_id, count_props;
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index 0dbf04d4aa9e..f666a13da6bf 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -306,4 +306,8 @@ static int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_
 			     dev->mode_config.async_page_flip;
 		break;
+	case DRM_CAP_ATOMIC_RESET:
+		req->value = drm_core_check_feature(dev, DRIVER_ATOMIC) &&
+			     drm_atomic_implements_create_state(dev);
+		break;
 	default:
 		return -EINVAL;
diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h
index bc7ef7684099..b6e2f2edd122 100644
--- a/include/uapi/drm/drm.h
+++ b/include/uapi/drm/drm.h
@@ -796,4 +796,14 @@ struct drm_gem_change_handle {
 #define DRM_CAP_ATOMIC_ASYNC_PAGE_FLIP	0x15
 
+/**
+ * DRM_CAP_ATOMIC_RESET
+ *
+ * If set to 1, the driver supports the &DRM_MODE_ATOMIC_RESET flag in
+ * &DRM_IOCTL_MODE_ATOMIC commits. When supported, userspace can pass that
+ * flag to reset all KMS object states to their defaults before applying
+ * property changes.
+ */
+#define DRM_CAP_ATOMIC_RESET	0x16
+
 /* DRM_IOCTL_GET_CAP ioctl argument type */
 struct drm_get_cap {
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index ed1a660a3dfd..1ab5c0fb3804 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -1329,4 +1329,18 @@ struct drm_mode_destroy_dumb {
  */
 #define DRM_MODE_ATOMIC_ALLOW_MODESET 0x0400
+/**
+ * DRM_MODE_ATOMIC_RESET
+ *
+ * Reset all KMS object states (CRTCs, planes, connectors, color operations)
+ * to their default values before applying the properties in this commit.
+ * Properties not explicitly included in the commit will remain at their
+ * defaults (CRTCs inactive, planes disabled, connectors unbound, etc.).
+ *
+ * CRTCs that are disabled before and after the commit won't send any
+ * &DRM_MODE_PAGE_FLIP_EVENT.
+ *
+ * This flag cannot be combined with &DRM_MODE_PAGE_FLIP_ASYNC.
+ */
+#define DRM_MODE_ATOMIC_RESET 0x0800
 
 /**
@@ -1341,5 +1355,6 @@ struct drm_mode_destroy_dumb {
 		DRM_MODE_ATOMIC_TEST_ONLY |\
 		DRM_MODE_ATOMIC_NONBLOCK |\
-		DRM_MODE_ATOMIC_ALLOW_MODESET)
+		DRM_MODE_ATOMIC_ALLOW_MODESET |\
+		DRM_MODE_ATOMIC_RESET)
 
 struct drm_mode_atomic {

-- 
2.55.0


  parent reply	other threads:[~2026-10-30 10:28 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-30 10:27 [PATCH v7 00/16] drm: Add DRM_MODE_ATOMIC_RESET flag Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 01/16] drm/amd/display: Fix planes atomic_create_state implementation Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 02/16] drm/solomon: ssd130x: Drop the init-only assumption in ssd130x_crtc_create_state() Maxime Ripard
2026-10-30 10:39   ` Javier Martinez Canillas
2026-10-30 10:27 ` [PATCH v7 03/16] drm/atomic: Switch to krealloc_array() in drm_atomic_get_private_obj_state() Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 04/16] drm/atomic: Use __GFP_ZERO instead of explicit memset " Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 05/16] drm/atomic: Use __GFP_ZERO instead of explicit memset in drm_atomic_get_connector_state() Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 06/16] drm/atomic: Create function to insert CRTC state into a commit Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 07/16] drm/atomic: Create function to insert plane " Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 08/16] drm/atomic: Create function to insert colorop " Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 09/16] drm/atomic: Create function to insert private obj " Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 10/16] drm/atomic: Create function to insert connector " Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 11/16] drm/atomic: Add drm_atomic_implements_create_state() helper Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 12/16] drm/atomic: Allow filling a commit with pristine object states Maxime Ripard
2026-10-30 10:53   ` sashiko-bot
2026-10-30 10:27 ` [PATCH v7 13/16] drm/tests: atomic: Add tests for drm_atomic_commit_fill_with_defaults() Maxime Ripard
2026-10-30 10:39   ` sashiko-bot
2026-10-30 10:27 ` Maxime Ripard [this message]
2026-10-30 10:54   ` [PATCH v7 14/16] drm/atomic-uapi: Add DRM_MODE_ATOMIC_RESET flag sashiko-bot
2026-10-30 10:27 ` [PATCH v7 15/16] drm/vkms: Switch container_of helpers to container_of_const Maxime Ripard
2026-10-30 10:27 ` [PATCH v7 16/16] drm/vkms: Add driver-specific plane property for testing Maxime Ripard
2026-10-30 10:40   ` sashiko-bot

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=20261030-drm-reset-state-flag-v7-14-4e55eaeae683@kernel.org \
    --to=mripard@kernel.org \
    --cc=airlied@gmail.com \
    --cc=contact@emersion.fr \
    --cc=daniels@collabora.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=hamohammed.sa@gmail.com \
    --cc=harry.wentland@amd.com \
    --cc=jadahl@redhat.com \
    --cc=jfalempe@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=louis.chauvet@bootlin.com \
    --cc=maarten.lankhorst@linux.intel.com \
    --cc=mdaenzer@redhat.com \
    --cc=melissa.srw@gmail.com \
    --cc=pekka.paalanen@collabora.com \
    --cc=sebastian.wick@redhat.com \
    --cc=simona@ffwll.ch \
    --cc=tzimmermann@suse.de \
    --cc=victoria@system76.com \
    --cc=xaver.hugl@kde.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.