drm/armada: convert overlay plane to atomic state

The overlay plane support updates asynchronously to the request, but the
drm_plane_helper_update() transitional helper waits for a vblank event
before releasing the framebuffer.  Using the transitional helper would
make the call block, which would introduce a performance regression.

Convert the overlay plane update to use the atomic state structures and
methods for the plane, but implement our own legacy update method
rather than the transitional helper.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>

drivers/gpu/drm/armada/armada_crtc.c

@@ -295,19 +295,6 @@ static void armada_drm_crtc_complete_frame_work(struct armada_crtc *dcrtc,
 	spin_unlock_irqrestore(&dcrtc->irq_lock, flags);
 }
 
-static void armada_drm_crtc_complete_disable_work(struct armada_crtc *dcrtc,
-	struct armada_plane_work *work)
-{
-	unsigned long flags;
-
-	if (dcrtc->plane == work->plane)
-		dcrtc->plane = NULL;
-
-	spin_lock_irqsave(&dcrtc->irq_lock, flags);
-	armada_drm_crtc_update_regs(dcrtc, work->regs);
-	spin_unlock_irqrestore(&dcrtc->irq_lock, flags);
-}
-
 static struct armada_plane_work *
 armada_drm_crtc_alloc_plane_work(struct drm_plane *plane)
 {
@@ -1080,7 +1067,7 @@ static const struct drm_crtc_funcs armada_crtc_funcs = {
 	.disable_vblank	= armada_drm_crtc_disable_vblank,
 };
 
-static int armada_drm_plane_prepare_fb(struct drm_plane *plane,
+int armada_drm_plane_prepare_fb(struct drm_plane *plane,
 	struct drm_plane_state *state)
 {
 	DRM_DEBUG_KMS("[PLANE:%d:%s] [FB:%d]\n",
@@ -1096,7 +1083,7 @@ static int armada_drm_plane_prepare_fb(struct drm_plane *plane,
 	return 0;
 }
 
-static void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
+void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
 	struct drm_plane_state *old_state)
 {
 	DRM_DEBUG_KMS("[PLANE:%d:%s] [FB:%d]\n",
@@ -1107,7 +1094,7 @@ static void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
 		drm_framebuffer_put(old_state->fb);
 }
 
-static int armada_drm_plane_atomic_check(struct drm_plane *plane,
+int armada_drm_plane_atomic_check(struct drm_plane *plane,
 	struct drm_plane_state *state)
 {
 	if (state->fb && !WARN_ON(!state->crtc)) {
@@ -1243,66 +1230,6 @@ static const struct drm_plane_helper_funcs armada_primary_plane_helper_funcs = {
 	.atomic_disable	= armada_drm_primary_plane_atomic_disable,
 };
 
-int armada_drm_plane_disable(struct drm_plane *plane,
-			     struct drm_modeset_acquire_ctx *ctx)
-{
-	struct armada_plane *dplane = drm_to_armada_plane(plane);
-	struct armada_crtc *dcrtc;
-	struct armada_plane_work *work;
-	unsigned int idx = 0;
-	u32 sram_para1, enable_mask;
-
-	if (!plane->crtc)
-		return 0;
-
-	/*
-	 * Arrange to power down most RAMs and FIFOs if this is the primary
-	 * plane, otherwise just the YUV FIFOs for the overlay plane.
-	 */
-	if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
-		sram_para1 = CFG_PDWN256x32 | CFG_PDWN256x24 | CFG_PDWN256x8 |
-			     CFG_PDWN32x32 | CFG_PDWN64x66;
-		enable_mask = CFG_GRA_ENA;
-	} else {
-		sram_para1 = CFG_PDWN16x66 | CFG_PDWN32x66;
-		enable_mask = CFG_DMA_ENA;
-	}
-
-	dplane->state.ctrl0 &= ~enable_mask;
-
-	dcrtc = drm_to_armada_crtc(plane->crtc);
-
-	/*
-	 * Try to disable the plane and drop our ref on the framebuffer
-	 * at the next frame update. If we fail for any reason, disable
-	 * the plane immediately.
-	 */
-	work = &dplane->works[dplane->next_work];
-	work->fn = armada_drm_crtc_complete_disable_work;
-	work->cancel = armada_drm_crtc_complete_disable_work;
-	work->old_fb = plane->fb;
-
-	armada_reg_queue_mod(work->regs, idx,
-			     0, enable_mask, LCD_SPU_DMA_CTRL0);
-	armada_reg_queue_mod(work->regs, idx,
-			     sram_para1, 0, LCD_SPU_SRAM_PARA1);
-	armada_reg_queue_end(work->regs, idx);
-
-	/* Wait for any preceding work to complete, but don't wedge */
-	if (WARN_ON(!armada_drm_plane_work_wait(dplane, HZ)))
-		armada_drm_plane_work_cancel(dcrtc, dplane);
-
-	if (armada_drm_plane_work_queue(dcrtc, work)) {
-		work->fn(dcrtc, work);
-		if (work->old_fb)
-			drm_framebuffer_unreference(work->old_fb);
-	}
-
-	dplane->next_work = !dplane->next_work;
-
-	return 0;
-}
-
 static const struct drm_plane_funcs armada_primary_plane_funcs = {
 	.update_plane	= drm_plane_helper_update,
 	.disable_plane	= drm_plane_helper_disable,

drivers/gpu/drm/armada/armada_crtc.h

@@ -75,6 +75,13 @@ void armada_drm_plane_work_cancel(struct armada_crtc *dcrtc,
 void armada_drm_plane_calc_addrs(u32 *addrs, struct drm_framebuffer *fb,
 	int x, int y);
 
+int armada_drm_plane_prepare_fb(struct drm_plane *plane,
+	struct drm_plane_state *state);
+void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
+	struct drm_plane_state *old_state);
+int armada_drm_plane_atomic_check(struct drm_plane *plane,
+	struct drm_plane_state *state);
+
 struct armada_crtc {
 	struct drm_crtc		crtc;
 	const struct armada_variant *variant;

drivers/gpu/drm/armada/armada_overlay.c

@@ -7,7 +7,9 @@
  * published by the Free Software Foundation.
  */
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_plane_helper.h>
 #include "armada_crtc.h"
 #include "armada_drm.h"
 #include "armada_fb.h"
@@ -78,7 +80,7 @@ static void armada_ovl_plane_work(struct armada_crtc *dcrtc,
 	spin_unlock_irqrestore(&dcrtc->irq_lock, flags);
 }
 
-static void armada_ovl_plane_update_state(struct drm_plane_state *state,
+static unsigned int armada_ovl_plane_update_state(struct drm_plane_state *state,
 	struct armada_regs *regs)
 {
 	struct armada_ovl_plane *dplane = drm_to_armada_ovl_plane(state->plane);
@@ -180,67 +182,116 @@ static void armada_ovl_plane_update_state(struct drm_plane_state *state,
 
 	dplane->base.state.changed = idx != 0;
 
-	armada_reg_queue_end(regs, idx);
+	return idx;
 }
 
-static int
-armada_ovl_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
-	struct drm_framebuffer *fb,
-	int crtc_x, int crtc_y, unsigned crtc_w, unsigned crtc_h,
-	uint32_t src_x, uint32_t src_y, uint32_t src_w, uint32_t src_h,
-	struct drm_modeset_acquire_ctx *ctx)
+static void armada_drm_overlay_plane_atomic_update(struct drm_plane *plane,
+	struct drm_plane_state *old_state)
+{
+	struct drm_plane_state *state = plane->state;
+	struct armada_crtc *dcrtc;
+	struct armada_regs *regs;
+
+	DRM_DEBUG_KMS("[PLANE:%d:%s]\n", plane->base.id, plane->name);
+
+	if (!state->fb || WARN_ON(!state->crtc))
+		return;
+
+	DRM_DEBUG_KMS("[PLANE:%d:%s] is on [CRTC:%d:%s] with [FB:%d] visible %u->%u\n",
+		plane->base.id, plane->name,
+		state->crtc->base.id, state->crtc->name,
+		state->fb->base.id,
+		old_state->visible, state->visible);
+
+	dcrtc = drm_to_armada_crtc(state->crtc);
+	regs = dcrtc->regs + dcrtc->regs_idx;
+
+	dcrtc->regs_idx += armada_ovl_plane_update_state(state, regs);
+}
+
+static void armada_drm_overlay_plane_atomic_disable(struct drm_plane *plane,
+	struct drm_plane_state *old_state)
+{
+	struct armada_plane *dplane = drm_to_armada_plane(plane);
+	struct armada_crtc *dcrtc;
+	struct armada_regs *regs;
+	unsigned int idx = 0;
+
+	DRM_DEBUG_KMS("[PLANE:%d:%s]\n", plane->base.id, plane->name);
+
+	if (!old_state->crtc)
+		return;
+
+	DRM_DEBUG_KMS("[PLANE:%d:%s] was on [CRTC:%d:%s] with [FB:%d]\n",
+		plane->base.id, plane->name,
+		old_state->crtc->base.id, old_state->crtc->name,
+		old_state->fb->base.id);
+
+	dplane->state.ctrl0 &= ~CFG_DMA_ENA;
+
+	dcrtc = drm_to_armada_crtc(old_state->crtc);
+	regs = dcrtc->regs + dcrtc->regs_idx;
+
+	/* Disable plane and power down the YUV FIFOs */
+	armada_reg_queue_mod(regs, idx, 0, CFG_DMA_ENA, LCD_SPU_DMA_CTRL0);
+	armada_reg_queue_mod(regs, idx, CFG_PDWN16x66 | CFG_PDWN32x66, 0,
+			     LCD_SPU_SRAM_PARA1);
+
+	dcrtc->regs_idx += idx;
+
+	if (dcrtc->plane == plane)
+		dcrtc->plane = NULL;
+}
+
+static const struct drm_plane_helper_funcs armada_overlay_plane_helper_funcs = {
+	.prepare_fb	= armada_drm_plane_prepare_fb,
+	.cleanup_fb	= armada_drm_plane_cleanup_fb,
+	.atomic_check	= armada_drm_plane_atomic_check,
+	.atomic_update	= armada_drm_overlay_plane_atomic_update,
+	.atomic_disable	= armada_drm_overlay_plane_atomic_disable,
+};
+
+static int armada_overlay_commit(struct drm_plane *plane,
+	struct drm_plane_state *state)
 {
 	struct armada_ovl_plane *dplane = drm_to_armada_ovl_plane(plane);
-	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
+	const struct drm_plane_helper_funcs *plane_funcs;
+	struct armada_crtc *dcrtc = drm_to_armada_crtc(state->crtc);
 	struct armada_plane_work *work;
-	struct drm_plane_state state = {
-		.plane = plane,
-		.crtc = crtc,
-		.fb = fb,
-		.src_x = src_x,
-		.src_y = src_y,
-		.src_w = src_w,
-		.src_h = src_h,
-		.crtc_x = crtc_x,
-		.crtc_y = crtc_y,
-		.crtc_w = crtc_w,
-		.crtc_h = crtc_h,
-		.rotation = DRM_MODE_ROTATE_0,
-	};
-	struct drm_crtc_state crtc_state = {
-		.crtc = crtc,
-		.enable = crtc->enabled,
-		.mode = crtc->mode,
-	};
 	int ret;
 
-	trace_armada_ovl_plane_update(plane, crtc, fb,
-				 crtc_x, crtc_y, crtc_w, crtc_h,
-				 src_x, src_y, src_w, src_h);
-
-	ret = drm_atomic_helper_check_plane_state(&state, &crtc_state, 0,
-						  INT_MAX, true, false);
+	plane_funcs = plane->helper_private;
+	ret = plane_funcs->atomic_check(plane, state);
 	if (ret)
-		return ret;
+		goto put_state;
 
 	work = &dplane->base.works[dplane->base.next_work];
 
-	if (plane->fb != fb) {
+	if (plane->state->fb != state->fb) {
 		/*
 		 * Take a reference on the new framebuffer - we want to
 		 * hold on to it while the hardware is displaying it.
 		 */
-		drm_framebuffer_reference(fb);
+		drm_framebuffer_reference(state->fb);
 
-		work->old_fb = plane->fb;
+		work->old_fb = plane->state->fb;
 	} else {
 		work->old_fb = NULL;
 	}
 
-	armada_ovl_plane_update_state(&state, work->regs);
+	/* Point of no return */
+	swap(plane->state, state);
+
+	dcrtc->regs_idx = 0;
+	dcrtc->regs = work->regs;
 
+	plane_funcs->atomic_update(plane, state);
+
+	/* If nothing was updated, short-circuit */
 	if (!dplane->base.state.changed)
-		return 0;
+		goto put_state;
+
+	armada_reg_queue_end(dcrtc->regs, dcrtc->regs_idx);
 
 	/* Wait for pending work to complete */
 	if (armada_drm_plane_work_wait(&dplane->base, HZ / 25) == 0)
@@ -249,7 +300,7 @@ armada_ovl_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	/* Just updating the position/size? */
 	if (!dplane->base.state.vsync_update) {
 		armada_ovl_plane_work(dcrtc, work);
-		return 0;
+		goto put_state;
 	}
 
 	if (!dcrtc->plane) {
@@ -259,12 +310,48 @@ armada_ovl_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	/* Queue it for update on the next interrupt if we are enabled */
 	ret = armada_drm_plane_work_queue(dcrtc, work);
-	if (ret)
+	if (ret) {
 		DRM_ERROR("failed to queue plane work: %d\n", ret);
+		ret = 0;
+	}
 
 	dplane->base.next_work = !dplane->base.next_work;
 
-	return 0;
+put_state:
+	drm_atomic_helper_plane_destroy_state(plane, state);
+	return ret;
+}
+
+static int
+armada_ovl_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+	struct drm_framebuffer *fb,
+	int crtc_x, int crtc_y, unsigned crtc_w, unsigned crtc_h,
+	uint32_t src_x, uint32_t src_y, uint32_t src_w, uint32_t src_h,
+	struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_plane_state *state;
+
+	trace_armada_ovl_plane_update(plane, crtc, fb,
+				 crtc_x, crtc_y, crtc_w, crtc_h,
+				 src_x, src_y, src_w, src_h);
+
+	/* Construct new state for the overlay plane */
+	state = drm_atomic_helper_plane_duplicate_state(plane);
+	if (!state)
+		return -ENOMEM;
+
+	state->crtc = crtc;
+	drm_atomic_set_fb_for_plane(state, fb);
+	state->crtc_x = crtc_x;
+	state->crtc_y = crtc_y;
+	state->crtc_h = crtc_h;
+	state->crtc_w = crtc_w;
+	state->src_x = src_x;
+	state->src_y = src_y;
+	state->src_h = src_h;
+	state->src_w = src_w;
+
+	return armada_overlay_commit(plane, state);
 }
 
 static void armada_ovl_plane_destroy(struct drm_plane *plane)
@@ -355,9 +442,10 @@ static int armada_ovl_plane_set_property(struct drm_plane *plane,
 
 static const struct drm_plane_funcs armada_ovl_plane_funcs = {
 	.update_plane	= armada_ovl_plane_update,
-	.disable_plane	= armada_drm_plane_disable,
+	.disable_plane	= drm_plane_helper_disable,
 	.destroy	= armada_ovl_plane_destroy,
 	.set_property	= armada_ovl_plane_set_property,
+	.reset		= drm_atomic_helper_plane_reset,
 };
 
 static const uint32_t armada_ovl_formats[] = {
@@ -450,6 +538,9 @@ int armada_overlay_plane_create(struct drm_device *dev, unsigned long crtcs)
 	dplane->base.works[0].fn = armada_ovl_plane_work;
 	dplane->base.works[1].fn = armada_ovl_plane_work;
 
+	drm_plane_helper_add(&dplane->base.base,
+			     &armada_overlay_plane_helper_funcs);
+
 	ret = drm_universal_plane_init(dev, &dplane->base.base, crtcs,
 				       &armada_ovl_plane_funcs,
 				       armada_ovl_formats,