Skip to contentSkip to conversation

drm: Don't leak fb when plane crtc coodinates are bad

PATCH In review [email protected] 4 participants 15 messages Oct 15, 2015 – Oct 16, 2015

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/5 drm: Don't leak fb when plane crtc coodinates are bad

From: Ville Syrjälä <[email protected]>

---
 drivers/gpu/drm/drm_crtc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +2 -1
index e7c8422..66233a8 100644
@@ -2342,7 +2342,8 @@ static int __setplane_internal(struct drm_plane *plane,
 	    crtc_y > INT_MAX - (int32_t) crtc_h) {
 		DRM_DEBUG_KMS("Invalid CRTC coordinates %ux%u+%d+%d\n",
 			      crtc_w, crtc_h, crtc_x, crtc_y);
-		return -ERANGE;
+		ret = -ERANGE;
+		goto out;
 	}
 
 
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#1 lore raw

PATCH 2/5 drm: Swap w/h when converting the mode to src coordidates for a rotated primary plane

From: Ville Syrjälä <[email protected]>

When converting the mode hdisplay/vdisplay to primary plane src coordinates we need to take into account the current plane rotation.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_atomic_helper.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
drivers/gpu/drm/drm_atomic_helper.c +7 -2
index 87a2a44..0c6f621 100644
@@ -1790,8 +1790,13 @@ int __drm_atomic_helper_set_config(struct drm_mode_set *set,
 	primary_state->crtc_w = set->mode->hdisplay;
 	primary_state->src_x = set->x << 16;
 	primary_state->src_y = set->y << 16;
-	primary_state->src_h = set->mode->vdisplay << 16;
-	primary_state->src_w = set->mode->hdisplay << 16;
+	if (primary_state->rotation & (BIT(DRM_ROTATE_90) | BIT(DRM_ROTATE_270))) {
+		primary_state->src_h = set->mode->hdisplay << 16;
+		primary_state->src_w = set->mode->vdisplay << 16;
+	} else {
+		primary_state->src_h = set->mode->vdisplay << 16;
+		primary_state->src_w = set->mode->hdisplay << 16;
+	}
 
 commit:
 	ret = update_output_state(state, set);
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#2 lore raw

PATCH 3/5 drm: Refactor plane src coordinate checks

In reply to [email protected]

From: Ville Syrjälä <[email protected]>

Pull the plane src coordinate checks into a separate function so that we can share them for the legacy and new stuff.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 59 +++++++++++++++++++++++-----------------------
 1 file changed, 30 insertions(+), 29 deletions(-)
drivers/gpu/drm/drm_crtc.c +30 -29
index 66233a8..227613a 100644
@@ -2286,6 +2286,32 @@ int drm_plane_check_pixel_format(const struct drm_plane *plane, u32 format)
 	return -EINVAL;
 }
 
+static int check_src_coords(uint32_t src_x, uint32_t src_y,
+			    uint32_t src_w, uint32_t src_h,
+			    const struct drm_framebuffer *fb)
+{
+	unsigned int fb_width, fb_height;
+
+	fb_width = fb->width << 16;
+	fb_height = fb->height << 16;
+
+	/* Make sure source coordinates are inside the fb. */
+	if (src_w > fb_width ||
+	    src_x > fb_width - src_w ||
+	    src_h > fb_height ||
+	    src_y > fb_height - src_h) {
+		DRM_DEBUG_KMS("Invalid source coordinates "
+			      "%u.%06ux%u.%06u+%u.%06u+%u.%06u\n",
+			      src_w >> 16, ((src_w & 0xffff) * 15625) >> 10,
+			      src_h >> 16, ((src_h & 0xffff) * 15625) >> 10,
+			      src_x >> 16, ((src_x & 0xffff) * 15625) >> 10,
+			      src_y >> 16, ((src_y & 0xffff) * 15625) >> 10);
+		return -ENOSPC;
+	}
+
+	return 0;
+}
+
 /*
  * setplane_internal - setplane handler for internal callers
  *
@@ -2305,7 +2331,6 @@ static int __setplane_internal(struct drm_plane *plane,
 			       uint32_t src_w, uint32_t src_h)
 {
 	int ret = 0;
-	unsigned int fb_width, fb_height;
 
 	/* No fb means shut it down */
 	if (!fb) {
@@ -2346,24 +2371,9 @@ static int __setplane_internal(struct drm_plane *plane,
 		goto out;
 	}
 
-
-	fb_width = fb->width << 16;
-	fb_height = fb->height << 16;
-
-	/* Make sure source coordinates are inside the fb. */
-	if (src_w > fb_width ||
-	    src_x > fb_width - src_w ||
-	    src_h > fb_height ||
-	    src_y > fb_height - src_h) {
-		DRM_DEBUG_KMS("Invalid source coordinates "
-			      "%u.%06ux%u.%06u+%u.%06u+%u.%06u\n",
-			      src_w >> 16, ((src_w & 0xffff) * 15625) >> 10,
-			      src_h >> 16, ((src_h & 0xffff) * 15625) >> 10,
-			      src_x >> 16, ((src_x & 0xffff) * 15625) >> 10,
-			      src_y >> 16, ((src_y & 0xffff) * 15625) >> 10);
-		ret = -ENOSPC;
+	ret = check_src_coords(src_x, src_y, src_w, src_h, fb);
+	if (ret)
 		goto out;
-	}
 
 	plane->old_fb = plane->fb;
 	ret = plane->funcs->update_plane(plane, crtc, fb,
@@ -2557,17 +2567,8 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 	if (crtc->invert_dimensions)
 		swap(hdisplay, vdisplay);
 
-	if (hdisplay > fb->width ||
-	    vdisplay > fb->height ||
-	    x > fb->width - hdisplay ||
-	    y > fb->height - vdisplay) {
-		DRM_DEBUG_KMS("Invalid fb size %ux%u for CRTC viewport %ux%u+%d+%d%s.\n",
-			      fb->width, fb->height, hdisplay, vdisplay, x, y,
-			      crtc->invert_dimensions ? " (inverted)" : "");
-		return -ENOSPC;
-	}
-
-	return 0;
+	return check_src_coords(x << 16, y << 16,
+				hdisplay << 16, vdisplay << 16, fb);
 }
 EXPORT_SYMBOL(drm_crtc_check_viewport);
 
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#3 lore raw

PATCH 4/5 drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

In reply to [email protected]

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#4 lore raw

Re: [PATCH 4/5] drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:01PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Can't we just remove the invert_dimensions field completely now? It's a legacy-only field, but no legacy drivers actually set it, so it winds up being completely unused.

Matt

Quoted 42 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 1 line
2.4.9
Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#5 lore raw

Re: [PATCH 4/5] drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

Quoted 11 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:01PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Can't we just remove the invert_dimensions field completely now? It's a legacy-only field, but no legacy drivers actually set it, so it winds up being completely unused.

omap sets it.

Quoted 53 lines

Matt

Quoted 42 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 1 line
2.4.9
Signature · 5 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#6 lore raw

Re: [PATCH 4/5] drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

Quoted 9 lines · Ville Syrjälä
Quoted 8 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:01PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Quoted 5 lines
Quoted 3 lines

Can't we just remove the invert_dimensions field completely now? It's a legacy-only field, but no legacy drivers actually set it, so it winds up being completely unused.

omap sets it.

Omap should be atomic now (maybe double-check with Laurent), so I think we can indeed ditch it. Follow-up series perhaps?
-Daniel

Quoted 58 lines
Quoted 54 lines

Matt

Quoted 42 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 1 line
2.4.9
Signature · 5 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
Signature · 2 lines
Ville Syrjälä
Intel OTC
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#7 lore raw

Re: [PATCH 4/5] drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

Quoted 10 lines · Daniel Vetter
Quoted 9 lines · Ville Syrjälä
Quoted 8 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:01PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

Can't we just remove the invert_dimensions field completely now? It's a legacy-only field, but no legacy drivers actually set it, so it winds up being completely unused.

omap sets it.

Omap should be atomic now (maybe double-check with Laurent), so I think we can indeed ditch it. Follow-up series perhaps?

grep ATOMIC in omap didn't turn up anything, so I assumed no.

Quoted 65 lines

-Daniel

Quoted 58 lines
Quoted 54 lines

Matt

Quoted 42 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 1 line
2.4.9
Signature · 5 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
Signature · 2 lines
Ville Syrjälä
Intel OTC
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#8 lore raw

Re: [PATCH 4/5] drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

Quoted 15 lines · Ville Syrjälä
Quoted 14 lines · Daniel Vetter
Quoted 13 lines · Ville Syrjälä
Quoted 11 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:01PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

Can't we just remove the invert_dimensions field completely now? It's a legacy-only field, but no legacy drivers actually set it, so it winds up being completely unused.

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

omap sets it.

Omap should be atomic now (maybe double-check with Laurent), so I think we can indeed ditch it. Follow-up series perhaps?

grep ATOMIC in omap didn't turn up anything, so I assumed no.

omap might not be full atomic (just like i915 isn't full atomic), but it's state-based now, so your new code below will cause the invert_dimensions to be ignored by the DRM core. And omap never uses the field internally, aside from setting it once to reflect the state values, so as far as I can see, the field is completely unused.

Matt

Quoted 70 lines
Quoted 66 lines

-Daniel

Quoted 58 lines
Quoted 54 lines

Matt

Quoted 42 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +11 -1
index 227613a..ffaa3f5 100644
@@ -2545,6 +2545,16 @@ void drm_crtc_get_hv_timing(const struct drm_display_mode *mode,
 }
 EXPORT_SYMBOL(drm_crtc_get_hv_timing);
 
+static bool invert_dimensions(const struct drm_crtc *crtc)
+{
+	if (crtc->state) {
+		return crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+							 BIT(DRM_ROTATE_270));
+	} else {
+		return crtc->invert_dimensions;
+	}
+}
+
 /**
  * drm_crtc_check_viewport - Checks that a framebuffer is big enough for the
  *     CRTC viewport
@@ -2564,7 +2574,7 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (invert_dimensions(crtc))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
Signature · 1 line
2.4.9
Signature · 5 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
Signature · 2 lines
Ville Syrjälä
Intel OTC
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 2 lines
Ville Syrjälä
Intel OTC
Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#9 lore raw

PATCH v2 4/5 drm: Check crtc viewport correctly with rotated primary plane on atomic drivers

In reply to [email protected]

From: Ville Syrjälä <[email protected]>

On atomic drivers we can dig out the primary plane rotation from the plane state instead of looking at the legacy crtc->invert_dimensions flag. The flag is not set by anyone except omapdrm, and it would be racy to set it the same way in the atomic helpers.

v2: Kill crtc->invert_dimensions totally since omap is state based
already and no one else ever used it (Matt)

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Tomi Valkeinen <[email protected]>
Cc Rob Clark <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c          | 5 +++--
 drivers/gpu/drm/omapdrm/omap_crtc.c | 3 ---
 include/drm/drm_crtc.h              | 5 -----
 3 files changed, 3 insertions(+), 10 deletions(-)
drivers/gpu/drm/drm_crtc.c +3 -2
index 227613a..a7738be 100644
@@ -677,7 +677,6 @@ int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 
 	crtc->dev = dev;
 	crtc->funcs = funcs;
-	crtc->invert_dimensions = false;
 
 	drm_modeset_lock_init(&crtc->mutex);
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
@@ -2564,7 +2563,9 @@ int drm_crtc_check_viewport(const struct drm_crtc *crtc,
 
 	drm_crtc_get_hv_timing(mode, &hdisplay, &vdisplay);
 
-	if (crtc->invert_dimensions)
+	if (crtc->state &&
+	    crtc->primary->state->rotation & (BIT(DRM_ROTATE_90) |
+					      BIT(DRM_ROTATE_270)))
 		swap(hdisplay, vdisplay);
 
 	return check_src_coords(x << 16, y << 16,
drivers/gpu/drm/omapdrm/omap_crtc.c +0 -3
index 9a4ba4f..ad09590 100644
@@ -412,9 +412,6 @@ static void omap_crtc_atomic_flush(struct drm_crtc *crtc,
 		dispc_mgr_go(omap_crtc->channel);
 		omap_irq_register(crtc->dev, &omap_crtc->vblank_irq);
 	}
-
-	crtc->invert_dimensions = !!(crtc->primary->state->rotation &
-				    (BIT(DRM_ROTATE_90) | BIT(DRM_ROTATE_270)));
 }
 
 static int omap_crtc_atomic_set_property(struct drm_crtc *crtc,
include/drm/drm_crtc.h +0 -5
index 317baf9..028e5d4 100644
@@ -407,9 +407,6 @@ struct drm_crtc_funcs {
  * @enabled: is this CRTC enabled?
  * @mode: current mode timings
  * @hwmode: mode timings as programmed to hw regs
- * @invert_dimensions: for purposes of error checking crtc vs fb sizes,
- *    invert the width/height of the crtc.  This is used if the driver
- *    is performing 90 or 270 degree rotated scanout
  * @x: x position on screen
  * @y: y position on screen
  * @funcs: CRTC control functions
@@ -458,8 +455,6 @@ struct drm_crtc {
 	 */
 	struct drm_display_mode hwmode;
 
-	bool invert_dimensions;
-
 	int x, y;
 	const struct drm_crtc_funcs *funcs;
 
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#10 lore raw

PATCH 5/5 drm: Check plane src coordinates correctly during page flip for atomic drivers

In reply to [email protected]

From: Ville Syrjälä <[email protected]>

Instead of relying on the old crtc-{x,y,mode} gunk, dig out the primary plane coordinates from the plane state when checking them against the new framebuffer during page flip.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_crtc.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +8 -1
index ffaa3f5..52feffc 100644
@@ -5193,7 +5193,14 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 		goto out;
 	}
 
-	ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, &crtc->mode, fb);
+	if (crtc->state) {
+		const struct drm_plane_state *state = crtc->primary->state;
+
+		ret = check_src_coords(state->src_x, state->src_y,
+				       state->src_w, state->src_h, fb);
+	} else {
+		ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, &crtc->mode, fb);
+	}
 	if (ret)
 		goto out;
 
Signature · 6 lines
2.4.9

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
Reviewed-by on v1 5/5
#11 lore raw

Re: [PATCH 5/5] drm: Check plane src coordinates correctly during page flip for atomic drivers

Quoted 10 lines · On Thu, Oct 15, 2015 at 08:40:02PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

Instead of relying on the old crtc-{x,y,mode} gunk, dig out the primary plane coordinates from the plane state when checking them against the new framebuffer during page flip.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>

For the series:

I also confirmed that the i-g-t test I wrote here:
http://lists.freedesktop.org/archives/intel-gfx/2015-October/077394.html
now passes with your patch series, so I believe Tvrtko's original bug report should be fixed.

Matt

Quoted 27 lines
---
 drivers/gpu/drm/drm_crtc.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_crtc.c +8 -1
index ffaa3f5..52feffc 100644
@@ -5193,7 +5193,14 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 		goto out;
 	}
 
-	ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, &crtc->mode, fb);
+	if (crtc->state) {
+		const struct drm_plane_state *state = crtc->primary->state;
+
+		ret = check_src_coords(state->src_x, state->src_y,
+				       state->src_w, state->src_h, fb);
+	} else {
+		ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, &crtc->mode, fb);
+	}
 	if (ret)
 		goto out;
 
Signature · 1 line
2.4.9
Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#12 lore raw

Re: [PATCH 5/5] drm: Check plane src coordinates correctly during page flip for atomic drivers

Quoted 12 lines · Matt Roper
Quoted 10 lines · On Thu, Oct 15, 2015 at 08:40:02PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

Instead of relying on the old crtc-{x,y,mode} gunk, dig out the primary plane coordinates from the plane state when checking them against the new framebuffer during page flip.

Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Quoted 8 lines

For the series:

I also confirmed that the i-g-t test I wrote here:

    http://lists.freedesktop.org/archives/intel-gfx/2015-October/077394.html

now passes with your patch series, so I believe Tvrtko's original bug report should be fixed.

Oh I did not realize this series is about this, perhaps because I did not see 1/5 which maybe had some more obvious clues. :)

Great, that means if we decide to merge "drm/i915: Consider plane rotation when calculating stride in skl_do_mmio_flip" and "kms_rotation_crc: Exercise page flips with 90 degree rotation" we would have working rotated legacy page flip with a test case.

Regards,

Tvrtko
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#13 lore raw

Re: [PATCH 5/5] drm: Check plane src coordinates correctly during page flip for atomic drivers

Quoted 9 lines · Tvrtko Ursulin
Quoted 7 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:02PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

Instead of relying on the old crtc-{x,y,mode} gunk, dig out the primary plane coordinates from the plane state when checking them against the new framebuffer during page flip.

Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>

For the series:

Pulled all 5 into drm-misc, thanks.

Quoted 5 lines
Quoted 4 lines

I also confirmed that the i-g-t test I wrote here:

   http://lists.freedesktop.org/archives/intel-gfx/2015-October/077394.html

now passes with your patch series, so I believe Tvrtko's original bug report should be fixed.

Quoted 7 lines

Oh I did not realize this series is about this, perhaps because I did not see 1/5 which maybe had some more obvious clues. :)

Great, that means if we decide to merge "drm/i915: Consider plane rotation when calculating stride in skl_do_mmio_flip" and "kms_rotation_crc: Exercise page flips with 90 degree rotation" we would have working rotated legacy page flip with a test case.

Matt, can you please push these igt patches and double-check that we have subtests to catch the viewport confusion here?

Thanks, Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#14 lore raw

Re: [PATCH 5/5] drm: Check plane src coordinates correctly during page flip for atomic drivers

Quoted 34 lines · Daniel Vetter
Quoted 9 lines · Tvrtko Ursulin
Quoted 7 lines · Matt Roper
Quoted 6 lines · On Thu, Oct 15, 2015 at 08:40:02PM +0300, [email protected] wrote:

From: Ville Syrjälä <[email protected]>

Instead of relying on the old crtc-{x,y,mode} gunk, dig out the primary plane coordinates from the plane state when checking them against the new framebuffer during page flip.

Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
Cc Matt Roper <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>

For the series:

Pulled all 5 into drm-misc, thanks.

Quoted 5 lines
Quoted 4 lines

I also confirmed that the i-g-t test I wrote here:

   http://lists.freedesktop.org/archives/intel-gfx/2015-October/077394.html

now passes with your patch series, so I believe Tvrtko's original bug report should be fixed.

Quoted 7 lines

Oh I did not realize this series is about this, perhaps because I did not see 1/5 which maybe had some more obvious clues. :)

Great, that means if we decide to merge "drm/i915: Consider plane rotation when calculating stride in skl_do_mmio_flip" and "kms_rotation_crc: Exercise page flips with 90 degree rotation" we would have working rotated legacy page flip with a test case.

Quoted 4 lines

Matt, can you please push these igt patches and double-check that we have subtests to catch the viewport confusion here?

Thanks, Daniel

Looks like Thomas already beat me to pushing the relevant i-g-t patches (both the one I wrote and the one Tvrtko wrote), so I think we're covered there.

Thanks.

Matt

Quoted 4 lines
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx