[PATCH v3 00/14] Support 10-bit YUV422 and 8/10-bit YUV420 color format on DW HDMI QP

[PATCH v3 00/14] Support 10-bit YUV422 and 8/10-bit YUV420 color format on DW HDMI QP

[Cristian Ciocaltea]

This series extends the output color format support of the Rockchip DW
HDMI QP controller to cover 10-bit YUV 4:2:2 and 8/10-bit YUV 4:2:0.

Please note this has a runtime dependency on the Rockchip Samsung HDPTX
PHY driver bug fixes posted separately as [1].  While there is no build
dependency, those fixes are required to address clock rate calculation
and synchronization issues that arise when changing the color depth
(bpc) while keeping the modeline constant.

Patches 1, 7 & 14 improve VOP2 robustness on RK3588, helping recover from
exceptions and preventing random display output glitches observed when
switching modes that also change the color format, e.g. from RGB to YUV
4:2:0 and vice versa.

Patch 8 avoids an incorrect DCLK source switch for 10-bit YUV 4:2:2 by
forcing 8 bpc in the bandwidth check.

Patches 2-6 address a few vop2 related issues reported by Sashiko, while
9-11 are additional cleanups/improvements.

Patch 12 adds MEDIA_BUS_FMT_UYVY10_1X20 for 10-bit YUV 4:2:2 output,
configuring the PHY with 8 bpc.  YUV 4:2:2 always transmits two 12-bit
components per pixel regardless of color depth, so from a clock-rate
perspective it is equivalent to three 8-bit RGB components.

Patch 13 advertises YUV 4:2:0 output, now that the bus-format and VOP2
support are in place.

Tested on Radxa ROCK 5B (RK3588) and Radxa ROCK 4D (RK3576), up to
4K@60Hz YUV 4:2:0 and 4K@30Hz RGB.

[1] https://lore.kernel.org/all/20260612-hdptx-clk-fixes-v4-0-ce5e1d456cda@collabora.com/

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
Changes in v3:
- Updated patch 3 to skip any hardware programming in the callbacks
  gated by drm_atomic_helper_commit_planes() after powering down VOP2 on
  an atomic_enable() error (Sashiko)
- Ensured VBLANK events are sent immediately when atomic_enable() fails,
  hence preventing userspace processes waiting for the page flip
  completion to hang indefinitely (new patch 4, Sashiko)
- Rebased onto latest drm-misc-next
- Link to v2: https://patch.msgid.link/20260706-dw-hdmi-qp-yuv-v2-0-b91beb33dedf@collabora.com

Changes in v2:
- Collected R-b from Krzysztof on the binding patch, while also fixed
  the property ordering in the example
- Renamed vop2_clk_reset() to vop2_reset_assert_deassert() and used to
  devm_reset_control_get_optional_exclusive() in patch 6 (Philipp Zabel)
- Addressed several issues reported by Sashiko
  * Reset AXI before detaching the IOMMU domain, to close a theoretical
    window where stale or in-flight DMA transactions could fault or
    access memory untranslated after the domain is detached (patch 6)
  * Fixed resource leak on vop2_enable() error path (new patch 2)
  * Balance state on atomic_enable error paths (new patch 3)
  * Avoided division by zero when computing max_dclk (new path 4)
  * Fixed VOP2_MAX_DCLK_RATE overflow on 32-bit (new patch 5)
- Rebased onto latest drm-misc-next
- Link to v1: https://patch.msgid.link/20260617-dw-hdmi-qp-yuv-v1-0-a665cfd06d7d@collabora.com

---
Cristian Ciocaltea (14):
      dt-bindings: display: vop2: Add missing reset properties
      drm/rockchip: vop2: Fix resource leak on vop2_enable() error path
      drm/rockchip: vop2: Balance state on atomic_enable() error paths
      drm/rockchip: vop2: Send pending event when atomic_enable() fails
      drm/rockchip: vop2: Avoid division by zero when computing max_dclk
      drm/rockchip: vop2: Fix VOP2_MAX_DCLK_RATE overflow on 32-bit
      drm/rockchip: vop2: Reset AXI and DCLK to improve robustness
      drm/rockchip: vop2: Avoid DCLK source switch for 10-bit YUV422 output
      drm/rockchip: vop2: Consolidate HDMI PHY PLL clock parent switch
      drm/rockchip: vop2: Switch to enum vop_csc_format
      drm/bridge: dw-hdmi-qp: Log resolution and refresh rate in atomic_enable()
      drm/rockchip: dw_hdmi_qp: Support 10-bit YUV422 output format
      drm/rockchip: dw_hdmi_qp: Enable YUV420 output format
      arm64: dts: rockchip: Add RK3588 VOP2 resets

 .../bindings/display/rockchip/rockchip-vop2.yaml   |  46 ++++-
 arch/arm64/boot/dts/rockchip/rk3588-base.dtsi      |  12 ++
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c       |  10 +-
 drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c     |  13 +-
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c       | 215 +++++++++++++++------
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.h       |  10 +
 6 files changed, 244 insertions(+), 62 deletions(-)
---
base-commit: 671b7825dbfe9ea6e3ad3001003aeee0df48d1b5
change-id: 20260617-dw-hdmi-qp-yuv-2b0f7bb5ba81

[PATCH v3 01/14] dt-bindings: display: vop2: Add missing reset properties

[Cristian Ciocaltea]

Document the VOP2 resets corresponding to the AXI, AHB and DCLK_VP0..2
clocks, which are common to all supported SoCs, plus DCLK_VP3 which is
provided only on RK3588.

While at it, move 'iommus' in the example up into the expected position
to keep the properties in alphabetical order.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 .../bindings/display/rockchip/rockchip-vop2.yaml   | 46 +++++++++++++++++++++-
 1 file changed, 44 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/display/rockchip/rockchip-vop2.yaml b/Documentation/devicetree/bindings/display/rockchip/rockchip-vop2.yaml
index 93da1fb9adc4..f993cef0323b 100644
--- a/Documentation/devicetree/bindings/display/rockchip/rockchip-vop2.yaml
+++ b/Documentation/devicetree/bindings/display/rockchip/rockchip-vop2.yaml
@@ -82,6 +82,20 @@ properties:
       - {}
       - {}
 
+  resets:
+    minItems: 5
+    maxItems: 6
+
+  reset-names:
+    minItems: 5
+    items:
+      - const: axi
+      - const: ahb
+      - const: dclk_vp0
+      - const: dclk_vp1
+      - const: dclk_vp2
+      - const: dclk_vp3
+
   rockchip,grf:
     $ref: /schemas/types.yaml#/definitions/phandle
     description:
@@ -148,6 +162,12 @@ allOf:
         clock-names:
           maxItems: 5
 
+        resets:
+          maxItems: 5
+
+        reset-names:
+          maxItems: 5
+
         interrupts:
           maxItems: 1
 
@@ -194,6 +214,12 @@ allOf:
             - {}
             - const: pll_hdmiphy0
 
+        resets:
+          maxItems: 5
+
+        reset-names:
+          maxItems: 5
+
         interrupts:
           minItems: 4
 
@@ -246,6 +272,12 @@ allOf:
             - const: pll_hdmiphy0
             - const: pll_hdmiphy1
 
+        resets:
+          minItems: 6
+
+        reset-names:
+          minItems: 6
+
         interrupts:
           maxItems: 1
 
@@ -289,9 +321,19 @@ examples:
                               "dclk_vp0",
                               "dclk_vp1",
                               "dclk_vp2";
-                power-domains = <&power RK3568_PD_VO>;
-                rockchip,grf = <&grf>;
                 iommus = <&vop_mmu>;
+                power-domains = <&power RK3568_PD_VO>;
+                resets = <&cru SRST_A_VOP>,
+                         <&cru SRST_H_VOP>,
+                         <&cru SRST_VOP0>,
+                         <&cru SRST_VOP1>,
+                         <&cru SRST_VOP2>;
+                reset-names = "axi",
+                              "ahb",
+                              "dclk_vp0",
+                              "dclk_vp1",
+                              "dclk_vp2";
+                rockchip,grf = <&grf>;
                 vop_out: ports {
                     #address-cells = <1>;
                     #size-cells = <0>;

-- 
2.54.0

[PATCH v3 02/14] drm/rockchip: vop2: Fix resource leak on vop2_enable() error path

[Cristian Ciocaltea]

vop2_enable() returns void, so callers cannot tell whether powering up
the VOP2 succeeded.

Additionally, if rockchip_drm_dma_attach_device() fails, vop2_enable()
returns without disabling the core clocks or dropping the runtime PM
reference taken earlier, leaking both.

Make vop2_enable() return int and unwind all resources on every failure
path.  As disabling the core clocks is also needed in vop2_disable(),
factor it out into a local helper used by both.

Fixes: 604be85547ce ("drm/rockchip: Add VOP2 driver")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260623202043.8BFE21F000E9@smtp.kernel.org/
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 39 +++++++++++++++++++---------
 1 file changed, 27 insertions(+), 12 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 4cce3e336f5b..e11dc8c7f78b 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -860,6 +860,13 @@ static int vop2_core_clks_prepare_enable(struct vop2 *vop2)
 	return ret;
 }
 
+static void vop2_core_clks_disable_unprepare(struct vop2 *vop2)
+{
+	clk_disable_unprepare(vop2->pclk);
+	clk_disable_unprepare(vop2->aclk);
+	clk_disable_unprepare(vop2->hclk);
+}
+
 static void rk3588_vop2_power_domain_enable_all(struct vop2 *vop2)
 {
 	u32 pd;
@@ -871,7 +878,7 @@ static void rk3588_vop2_power_domain_enable_all(struct vop2 *vop2)
 	vop2_writel(vop2, RK3588_SYS_PD_CTRL, pd);
 }
 
-static void vop2_enable(struct vop2 *vop2)
+static int vop2_enable(struct vop2 *vop2)
 {
 	int ret;
 	u32 version;
@@ -879,25 +886,24 @@ static void vop2_enable(struct vop2 *vop2)
 	ret = pm_runtime_resume_and_get(vop2->dev);
 	if (ret < 0) {
 		drm_err(vop2->drm, "failed to get pm runtime: %d\n", ret);
-		return;
+		return ret;
 	}
 
 	ret = vop2_core_clks_prepare_enable(vop2);
-	if (ret) {
-		pm_runtime_put_sync(vop2->dev);
-		return;
-	}
+	if (ret)
+		goto err_put_pm;
 
 	ret = rockchip_drm_dma_attach_device(vop2->drm, vop2->dev);
 	if (ret) {
-		drm_err(vop2->drm, "failed to attach dma mapping, %d\n", ret);
-		return;
+		drm_err(vop2->drm, "failed to attach dma mapping: %d\n", ret);
+		goto err_disable_clks;
 	}
 
 	version = vop2_readl(vop2, RK3568_VERSION_INFO);
 	if (version != vop2->version) {
 		drm_err(vop2->drm, "Hardware version(0x%08x) mismatch\n", version);
-		return;
+		ret = -EINVAL;
+		goto err_detach_dma;
 	}
 
 	/*
@@ -932,6 +938,17 @@ static void vop2_enable(struct vop2 *vop2)
 		    VOP2_INT_BUS_ERRPR << 16 | VOP2_INT_BUS_ERRPR);
 	vop2_writel(vop2, RK3568_SYS1_INT_EN,
 		    VOP2_INT_BUS_ERRPR << 16 | VOP2_INT_BUS_ERRPR);
+
+	return 0;
+
+err_detach_dma:
+	rockchip_drm_dma_detach_device(vop2->drm, vop2->dev);
+err_disable_clks:
+	vop2_core_clks_disable_unprepare(vop2);
+err_put_pm:
+	pm_runtime_put_sync(vop2->dev);
+
+	return ret;
 }
 
 static void vop2_disable(struct vop2 *vop2)
@@ -942,9 +959,7 @@ static void vop2_disable(struct vop2 *vop2)
 
 	regcache_drop_region(vop2->map, 0, vop2_regmap_config.max_register);
 
-	clk_disable_unprepare(vop2->pclk);
-	clk_disable_unprepare(vop2->aclk);
-	clk_disable_unprepare(vop2->hclk);
+	vop2_core_clks_disable_unprepare(vop2);
 }
 
 static bool vop2_vp_dsp_lut_is_enabled(struct vop2_video_port *vp)

-- 
2.54.0

[PATCH v3 03/14] drm/rockchip: vop2: Balance state on atomic_enable() error paths

[Cristian Ciocaltea]

vop2_crtc_atomic_enable() prepares vp->dclk and, for the first CRTC,
calls vop2_enable(), incrementing the shared vop2->enable_count before
programming VOP2 registers.  Two later failure points return early with
only vop2_unlock(), leaking the dclk prepare, the enable_count
increment, and the core clocks / PM runtime reference from
vop2_enable().

Since the disable path decrements enable_count unconditionally, the
leaked increment skews the count, causing an underflow or premature
shutdown of core clocks and PM runtime while another CRTC is active.

Moreover, cleaning up on the enable error paths alone is insufficient.
A CRTC whose atomic_enable() bailed is still committed as active, so
vop2_crtc_atomic_disable() still runs its full teardown, releasing the
same dclk prepare and enable_count a second time, thus underflowing both
counts and, where vop2_enable() ran, touching VOP registers with core
clocks gated.

Fix this by reverting exactly what was set up on failure, and track the
per-VP bring-up state so atomic_disable() skips teardown when
atomic_enable() never completed.  Additionally, ensure any further
hardware programming is also skipped within the affected RPM commit
tail, i.e. in the callbacks gated by drm_atomic_helper_commit_planes():
vop2_crtc_atomic_begin(), vop2_plane_atomic_update{update|disable}() and
vop2_crtc_atomic_flush().

Fixes: 604be85547ce ("drm/rockchip: Add VOP2 driver")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260623202043.8BFE21F000E9@smtp.kernel.org/
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 68 ++++++++++++++++++++++------
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.h |  6 +++
 2 files changed, 59 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index e11dc8c7f78b..271c3cba6fb4 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -1033,6 +1033,9 @@ static void vop2_crtc_atomic_disable(struct drm_crtc *crtc,
 
 	vop2_lock(vop2);
 
+	if (!vp->enabled)
+		goto out_unlock;
+
 	old_crtc_state = drm_atomic_get_old_crtc_state(state, crtc);
 	drm_atomic_helper_disable_planes_on_crtc(old_crtc_state, false);
 
@@ -1063,11 +1066,14 @@ static void vop2_crtc_atomic_disable(struct drm_crtc *crtc,
 
 	clk_disable_unprepare(vp->dclk);
 
+	vp->enabled = false;
+
 	vop2->enable_count--;
 
 	if (!vop2->enable_count)
 		vop2_disable(vop2);
 
+out_unlock:
 	vop2_unlock(vop2);
 
 	if (crtc->state->event && !crtc->state->active) {
@@ -1197,13 +1203,21 @@ static void vop2_plane_atomic_disable(struct drm_plane *plane,
 	struct drm_plane_state *old_pstate = NULL;
 	struct vop2_win *win = to_vop2_win(plane);
 	struct vop2 *vop2 = win->vop2;
+	struct vop2_video_port *vp;
 
 	drm_dbg(vop2->drm, "%s disable\n", win->data->name);
 
 	if (state)
 		old_pstate = drm_atomic_get_old_plane_state(state, plane);
-	if (old_pstate && !old_pstate->crtc)
-		return;
+
+	if (old_pstate) {
+		if (!old_pstate->crtc)
+			return;
+
+		vp = to_vop2_video_port(old_pstate->crtc);
+		if (!vp->enabled)
+			return;
+	}
 
 	vop2_win_disable(win);
 	vop2_win_write(win, VOP2_WIN_YUV_CLIP, 0);
@@ -1298,6 +1312,9 @@ static void vop2_plane_atomic_update(struct drm_plane *plane,
 	if (WARN_ON(!crtc))
 		return;
 
+	if (!vp->enabled)
+		return;
+
 	if (!pstate->visible) {
 		vop2_plane_atomic_disable(plane, state);
 		return;
@@ -1803,8 +1820,11 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 		return;
 	}
 
-	if (!vop2->enable_count)
-		vop2_enable(vop2);
+	if (!vop2->enable_count) {
+		ret = vop2_enable(vop2);
+		if (ret)
+			goto err_unprepare_dclk;
+	}
 
 	vop2->enable_count++;
 
@@ -1831,10 +1851,8 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 		clock = vop2->ops->setup_intf_mux(vp, rkencoder->crtc_endpoint_id, polflags);
 	}
 
-	if (!clock) {
-		vop2_unlock(vop2);
-		return;
-	}
+	if (!clock)
+		goto err_put_enable_count;
 
 	if (vcstate->output_mode == ROCKCHIP_OUT_MODE_AAAA &&
 	    !(vp_data->feature & VOP2_VP_FEATURE_OUTPUT_10BIT))
@@ -1852,8 +1870,7 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 		default:
 			drm_err(vop2->drm, "Unknown DRM_MODE_CONNECTOR %d\n",
 				vcstate->output_type);
-			vop2_unlock(vop2);
-			return;
+			goto err_put_enable_count;
 		}
 	else
 		out_mode = vcstate->output_mode;
@@ -1965,6 +1982,18 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 
 	drm_crtc_vblank_on(crtc);
 
+	vp->enabled = true;
+
+	vop2_unlock(vop2);
+
+	return;
+
+err_put_enable_count:
+	if (!--vop2->enable_count)
+		vop2_disable(vop2);
+
+err_unprepare_dclk:
+	clk_disable_unprepare(vp->dclk);
 	vop2_unlock(vop2);
 }
 
@@ -2023,6 +2052,9 @@ static void vop2_crtc_atomic_begin(struct drm_crtc *crtc,
 	struct vop2_video_port *vp = to_vop2_video_port(crtc);
 	struct vop2 *vop2 = vp->vop2;
 
+	if (!vp->enabled)
+		return;
+
 	vop2->ops->setup_overlay(vp);
 }
 
@@ -2034,13 +2066,19 @@ static void vop2_crtc_atomic_flush(struct drm_crtc *crtc,
 	struct vop2_video_port *vp = to_vop2_video_port(crtc);
 	struct vop2 *vop2 = vp->vop2;
 
-	/* In case of modeset, gamma lut update already happened in atomic enable */
-	if (!drm_atomic_crtc_needs_modeset(crtc_state) && crtc_state->color_mgmt_changed)
-		vop2_crtc_atomic_try_set_gamma_locked(vop2, vp, crtc, crtc_state);
+	if (vp->enabled) {
+		/*
+		 * In case of modeset, gamma lut update already happened
+		 * in atomic enable.
+		 */
+		if (!drm_atomic_crtc_needs_modeset(crtc_state) &&
+		    crtc_state->color_mgmt_changed)
+			vop2_crtc_atomic_try_set_gamma_locked(vop2, vp, crtc, crtc_state);
 
-	vop2_post_config(crtc, false, crtc_state, old_crtc_state);
+		vop2_post_config(crtc, false, crtc_state, old_crtc_state);
 
-	vop2_cfg_done(vp);
+		vop2_cfg_done(vp);
+	}
 
 	spin_lock_irq(&crtc->dev->event_lock);
 
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
index ffcb39c130aa..b209b2d5b96d 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
@@ -243,6 +243,12 @@ struct vop2_video_port {
 
 	struct completion dsp_hold_completion;
 
+	/**
+	 * @enabled: CRTC fully brought up by atomic_enable(); allows
+	 * atomic_disable() to skip teardown after a failed enable.
+	 */
+	bool enabled;
+
 	/**
 	 * @win_mask: Bitmask of windows attached to the video port;
 	 */

-- 
2.54.0

[PATCH v3 04/14] drm/rockchip: vop2: Send pending event when atomic_enable() fails

[Cristian Ciocaltea]

vop2_crtc_atomic_flush() stashes the CRTC completion event in vp->event
and relies on the frame start interrupt (VP_INT_FS_FIELD) to deliver it
later via drm_crtc_send_vblank_event().  Since atomic_enable() runs
before atomic_flush(), bailing on an error path makes it skip
drm_crtc_vblank_on() and leave the video port disabled.  atomic_flush()
still traps the event in vp->event, but since the port never scans out,
the frame start interrupt does not fire and the event is not delivered.
Userspace waiting for the flip completion then blocks indefinitely.

Note the event cannot be flushed from atomic_disable(), as userspace
waits for the completion before committing the next state, so the
disable does not arrive and draining it there would deadlock.

Use the newly added vp->enabled flag to detect this in atomic_flush()
and send the event immediately when the port did not come up, rather
than deferring it to an interrupt that will never occur.  This also
avoids the spurious drm_crtc_vblank_get() WARN previously hit on the
failed enable path, where the vblank is not on.

Fixes: 604be85547ce ("drm/rockchip: Add VOP2 driver")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260706194935.AC7601F000E9@smtp.kernel.org/
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 271c3cba6fb4..4bb1985b8726 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -2083,8 +2083,19 @@ static void vop2_crtc_atomic_flush(struct drm_crtc *crtc,
 	spin_lock_irq(&crtc->dev->event_lock);
 
 	if (crtc->state->event) {
-		WARN_ON(drm_crtc_vblank_get(crtc));
-		vp->event = crtc->state->event;
+		/*
+		 * A failed atomic_enable() leaves the video port disabled with
+		 * no scanout, so the frame start interrupt that normally
+		 * delivers vp->event never fires. Send the event right away in
+		 * that case to avoid stalling the flip completion.
+		 */
+		if (vp->enabled) {
+			WARN_ON(drm_crtc_vblank_get(crtc));
+			vp->event = crtc->state->event;
+		} else {
+			drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		}
+
 		crtc->state->event = NULL;
 	}
 

-- 
2.54.0

[PATCH v3 05/14] drm/rockchip: vop2: Avoid division by zero when computing max_dclk

[Cristian Ciocaltea]

The DCLK reparenting logic in vop2_crtc_atomic_enable() is gated only on
the presence of an HDMI PHY PLL on the SoC.

This condition is true on any board that wires up an HDMI PHY, even when
the CRTC is currently driving a non-HDMI output such as DSI or LVDS.
Those encoders do not set vcstate->output_bpc, so it stays 0, hence the
subsequent DIV_ROUND_CLOSEST_ULL() then divides by zero.

Note this division happens before the drm_for_each_encoder_mask() loop
that would establish whether the pipe is actually driven by HDMI, so the
loop provides no protection.

Default bpc to 8 when it would otherwise be 0.  The resulting max_dclk
value is a don't-care on the non-HDMI paths that trigger this, since the
encoder loop will not select the PHY PLL as the DCLK source there.

Fixes: b0362c45c401 ("drm/rockchip: vop2: Check bpc before switching DCLK source")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260623203310.9F8C81F000E9@smtp.kernel.org/
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 4bb1985b8726..e9f5d015b645 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -1931,8 +1931,9 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 	 * to 4K@60Hz, if available, otherwise keep using the system CRU.
 	 */
 	if (vop2->pll_hdmiphy0 || vop2->pll_hdmiphy1) {
-		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8,
-							       vcstate->output_bpc);
+		unsigned int bpc = vcstate->output_bpc ?: 8;
+		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8, bpc);
+
 		if (clock <= max_dclk) {
 			drm_for_each_encoder_mask(encoder, crtc->dev, crtc_state->encoder_mask) {
 				struct rockchip_encoder *rkencoder = to_rockchip_encoder(encoder);

-- 
2.54.0

[PATCH v3 06/14] drm/rockchip: vop2: Fix VOP2_MAX_DCLK_RATE overflow on 32-bit

[Cristian Ciocaltea]

VOP2_MAX_DCLK_RATE is defined as an unsigned long and is used scaled by
the color-depth ratio when computing the maximum DCLK the HDMI PHY PLL
can source.

On architectures where unsigned long is 32-bit, VOP2_MAX_DCLK_RATE * 8
is evaluated in 32-bit arithmetic and overflows: 4800000000 wraps to
505032704 before it is promoted to the 64-bit argument of
DIV_ROUND_CLOSEST_ULL().  At 8 bpc this collapses max_dclk from the
intended 600 MHz down to roughly 63 MHz.

As a result, low-resolution modes (e.g. 640x480 or 720x480, whose pixel
clocks are well under 30 MHz) still pass the clock <= max_dclk check and
are reparented to the HDMI PHY PLL as expected, but higher-resolution
modes such as 1280x720 (~74 MHz), 1920x1080 (~148 MHz) and above are
wrongly rejected and fall back to the system CRU as the DCLK source.

Define the constant as unsigned long long so that every expression using
it is evaluated in 64-bit arithmetic, independent of the width of
unsigned long.  This future-proofs any additional users of the macro as
well.

Fixes: b0362c45c401 ("drm/rockchip: vop2: Check bpc before switching DCLK source")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260623203310.9F8C81F000E9@smtp.kernel.org/
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index e9f5d015b645..2cd04bedaa2c 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -102,7 +102,7 @@ enum vop2_afbc_format {
 	VOP2_AFBC_FMT_INVALID = -1,
 };
 
-#define VOP2_MAX_DCLK_RATE		600000000UL
+#define VOP2_MAX_DCLK_RATE		600000000ULL
 
 /*
  * bus-format types.

-- 
2.54.0

[PATCH v3 07/14] drm/rockchip: vop2: Reset AXI and DCLK to improve robustness

[Cristian Ciocaltea]

Assert the AXI reset in the CRTC disable path, and the VP DCLK reset in
the enable path.

These resets are intended to leave the hardware in a clean state for the
next use, helping recover from exceptions such as IOMMU page faults, as
well as to prevent random display output glitches, such as a blank
image, observed when switching modes that also change the color format,
e.g. from RGB to YUV420 and vice versa.

For now this seems to affect only the RK3588, hence the resets are
optional and will be provided in the device tree for this SoC only.

Co-developed-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 36 ++++++++++++++++++++++++++++
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.h |  4 ++++
 2 files changed, 40 insertions(+)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 2cd04bedaa2c..8f0f4aa05fac 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -17,6 +17,7 @@
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/regmap.h>
+#include <linux/reset.h>
 #include <linux/swab.h>
 
 #include <drm/drm.h>
@@ -867,6 +868,26 @@ static void vop2_core_clks_disable_unprepare(struct vop2 *vop2)
 	clk_disable_unprepare(vop2->hclk);
 }
 
+static void vop2_reset_assert_deassert(struct vop2 *vop2, struct reset_control *rstc)
+{
+	int ret;
+
+	if (!rstc)
+		return;
+
+	ret = reset_control_assert(rstc);
+	if (ret < 0) {
+		drm_warn(vop2->drm, "failed to assert reset: %d\n", ret);
+		return;
+	}
+
+	udelay(10);
+
+	ret = reset_control_deassert(rstc);
+	if (ret < 0)
+		drm_err(vop2->drm, "failed to deassert reset: %d\n", ret);
+}
+
 static void rk3588_vop2_power_domain_enable_all(struct vop2 *vop2)
 {
 	u32 pd;
@@ -953,6 +974,8 @@ static int vop2_enable(struct vop2 *vop2)
 
 static void vop2_disable(struct vop2 *vop2)
 {
+	vop2_reset_assert_deassert(vop2, vop2->axi_rst);
+
 	rockchip_drm_dma_detach_device(vop2->drm, vop2->dev);
 
 	pm_runtime_put_sync(vop2->dev);
@@ -1981,6 +2004,8 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 
 	vop2_crtc_atomic_try_set_gamma(vop2, vp, crtc, crtc_state);
 
+	vop2_reset_assert_deassert(vop2, vp->dclk_rst);
+
 	drm_crtc_vblank_on(crtc);
 
 	vp->enabled = true;
@@ -2596,6 +2621,12 @@ static int vop2_create_crtcs(struct vop2 *vop2)
 			return dev_err_probe(drm->dev, PTR_ERR(vp->dclk),
 					     "failed to get %s\n", dclk_name);
 
+		vp->dclk_rst = devm_reset_control_get_optional_exclusive(vop2->dev,
+									 dclk_name);
+		if (IS_ERR(vp->dclk_rst))
+			return dev_err_probe(drm->dev, PTR_ERR(vp->dclk_rst),
+					     "failed to get %s reset\n", dclk_name);
+
 		np = of_graph_get_remote_node(dev->of_node, i, -1);
 		if (!np) {
 			drm_dbg(vop2->drm, "%s: No remote for vp%d\n", __func__, i);
@@ -2955,6 +2986,11 @@ static int vop2_bind(struct device *dev, struct device *master, void *data)
 		return dev_err_probe(drm->dev, PTR_ERR(vop2->pll_hdmiphy1),
 				     "failed to get pll_hdmiphy1\n");
 
+	vop2->axi_rst = devm_reset_control_get_optional_exclusive(vop2->dev, "axi");
+	if (IS_ERR(vop2->axi_rst))
+		return dev_err_probe(drm->dev, PTR_ERR(vop2->axi_rst),
+				     "failed to get axi reset\n");
+
 	vop2->irq = platform_get_irq(pdev, 0);
 	if (vop2->irq < 0)
 		return dev_err_probe(drm->dev, vop2->irq, "cannot find irq for vop2\n");
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
index b209b2d5b96d..1c2d7d97a6c2 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h
@@ -165,6 +165,8 @@ enum vop2_win_regs {
 	VOP2_WIN_MAX_REG,
 };
 
+struct reset_control;
+
 struct vop2_regs_dump {
 	const char *name;
 	u32 base;
@@ -238,6 +240,7 @@ struct vop2_video_port {
 	struct vop2 *vop2;
 	struct clk *dclk;
 	struct clk *dclk_src;
+	struct reset_control *dclk_rst;
 	unsigned int id;
 	const struct vop2_video_port_data *data;
 
@@ -335,6 +338,7 @@ struct vop2 {
 	struct clk *pclk;
 	struct clk *pll_hdmiphy0;
 	struct clk *pll_hdmiphy1;
+	struct reset_control *axi_rst;
 
 	/* optional internal rgb encoder */
 	struct rockchip_rgb *rgb;

-- 
2.54.0

[PATCH v3 08/14] drm/rockchip: vop2: Avoid DCLK source switch for 10-bit YUV422 output

[Cristian Ciocaltea]

Currently the color depth is always factored into the DCLK source
decision for HDMI output, which can break certain modes when operating
with depths greater than 8 bpc.

When the required transmission rate exceeds the 600 MHz limit of the
HDMI PHY PLL, e.g. for 4K@60Hz 10-bit RGB output, VOP2 will normally
fall back to using the less accurate system CRU as a DCLK source,
assuming HDMI 2.1 FRL is supported by the pipeline, otherwise the mode
will be rejected.  For YUV420 output format this never happens, as it
uses half of the RGB bandwidth, hence the rate remains within the PHY
PLL limits.

On the other hand, YUV422 always transmits two 12-bit components per
clock cycle, regardless of the color depth, which from a clock-rate
perspective is equivalent to three 8-bit RGB components.  For example,
4K@60Hz 10-bit YUV422 requires the same bandwidth as 4K@60Hz 8-bit RGB,
typically 594 MHz.  However, VOP2 wrongly assumes it needs 742.5 MHz
(594 * 10 / 8) and ends up switching the DCLK source.

As a consequence, the modes requiring uncommon pixel clocks, such as
those corresponding to fractional refresh rates, will fail.  An example
is 3840x2160@59.94Hz, which would likely rely on the 593.407 MHz clock
rate unsupported by the system CRU.

Note this only affects YUV422 with color depths greater than 8 bpc; for
8-bit YUV422 the 8/bpc factor is unity and the bandwidth check is
already correct.

Prevent the incorrect switches of DCLK source to system CRU for YUV422
output format by forcing 8 bpc when checking the bandwidth.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 8f0f4aa05fac..aa81b3b01fbd 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -1954,7 +1954,15 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 	 * to 4K@60Hz, if available, otherwise keep using the system CRU.
 	 */
 	if (vop2->pll_hdmiphy0 || vop2->pll_hdmiphy1) {
-		unsigned int bpc = vcstate->output_bpc ?: 8;
+		/*
+		 * YUV422 always transmits two 12-bit components per clock
+		 * cycle, regardless of the color depth, which from a rate
+		 * perspective is equivalent to three 8-bit RGB components.
+		 * Force 8 bpc here so the bandwidth check reflects the actual
+		 * TMDS rate and avoids an unnecessary DCLK source switch.
+		 */
+		unsigned int bpc = vcstate->output_mode == ROCKCHIP_OUT_MODE_YUV422 ?
+					8 : (vcstate->output_bpc ?: 8);
 		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8, bpc);
 
 		if (clock <= max_dclk) {

-- 
2.54.0

[PATCH v3 09/14] drm/rockchip: vop2: Consolidate HDMI PHY PLL clock parent switch

[Cristian Ciocaltea]

The DCLK parent switch logic for HDMI0 and HDMI1 PHY PLLs was
duplicated, with each endpoint repeating the same clk_get_parent(),
clk_set_parent() and error handling calls.

Refactor this by first selecting the appropriate PHY PLL clock handle
based on the active HDMI endpoint, then performing the parent switch in
a single shared code path.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 36 +++++++++++-----------------
 1 file changed, 14 insertions(+), 22 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index aa81b3b01fbd..65902db7f7c0 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -1964,42 +1964,34 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
 		unsigned int bpc = vcstate->output_mode == ROCKCHIP_OUT_MODE_YUV422 ?
 					8 : (vcstate->output_bpc ?: 8);
 		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8, bpc);
+		struct clk *pll_hdmiphy = NULL;
 
 		if (clock <= max_dclk) {
 			drm_for_each_encoder_mask(encoder, crtc->dev, crtc_state->encoder_mask) {
 				struct rockchip_encoder *rkencoder = to_rockchip_encoder(encoder);
 
 				if (rkencoder->crtc_endpoint_id == ROCKCHIP_VOP2_EP_HDMI0) {
-					if (!vop2->pll_hdmiphy0)
-						break;
-
-					if (!vp->dclk_src)
-						vp->dclk_src = clk_get_parent(vp->dclk);
-
-					ret = clk_set_parent(vp->dclk, vop2->pll_hdmiphy0);
-					if (ret < 0)
-						drm_warn(vop2->drm,
-							 "Could not switch to HDMI0 PHY PLL: %d\n",
-							 ret);
+					pll_hdmiphy = vop2->pll_hdmiphy0;
 					break;
 				}
 
 				if (rkencoder->crtc_endpoint_id == ROCKCHIP_VOP2_EP_HDMI1) {
-					if (!vop2->pll_hdmiphy1)
-						break;
-
-					if (!vp->dclk_src)
-						vp->dclk_src = clk_get_parent(vp->dclk);
-
-					ret = clk_set_parent(vp->dclk, vop2->pll_hdmiphy1);
-					if (ret < 0)
-						drm_warn(vop2->drm,
-							 "Could not switch to HDMI1 PHY PLL: %d\n",
-							 ret);
+					pll_hdmiphy = vop2->pll_hdmiphy1;
 					break;
 				}
 			}
 		}
+
+		if (pll_hdmiphy) {
+			if (!vp->dclk_src)
+				vp->dclk_src = clk_get_parent(vp->dclk);
+
+			ret = clk_set_parent(vp->dclk, pll_hdmiphy);
+			if (ret < 0)
+				drm_warn(vop2->drm,
+					 "Failed to switch DCLK to HDMI PHY PLL: %d\n",
+					 ret);
+		}
 	}
 
 	clk_set_rate(vp->dclk, clock);

-- 
2.54.0

[PATCH v3 10/14] drm/rockchip: vop2: Switch to enum vop_csc_format

[Cristian Ciocaltea]

Improve code readability in vop2_setup_csc_mode() by using enum
vop_csc_format for the csc_mode variable, as well as for the return type
of the vop2_convert_csc_mode() helper, which already returns CSC_*
enumerators.

While at it, replace the nonsensical 'csc_mode = false' assignment in
the no-conversion branch with the equivalent CSC_BT601L, which carries
the same value (0) but is type-correct.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
index 65902db7f7c0..ef1e1fc2dc9e 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
@@ -648,7 +648,7 @@ static void vop2_setup_scale(struct vop2 *vop2, const struct vop2_win *win,
 	}
 }
 
-static int vop2_convert_csc_mode(int csc_mode)
+static enum vop_csc_format vop2_convert_csc_mode(int csc_mode)
 {
 	switch (csc_mode) {
 	case V4L2_COLORSPACE_SMPTE170M:
@@ -711,7 +711,7 @@ static void vop2_setup_csc_mode(struct vop2_video_port *vp,
 	int input_csc = V4L2_COLORSPACE_DEFAULT;
 	int output_csc = vcstate->color_space;
 	bool r2y_en, y2r_en;
-	int csc_mode;
+	enum vop_csc_format csc_mode;
 
 	if (is_input_yuv && !is_output_yuv) {
 		y2r_en = true;
@@ -724,7 +724,7 @@ static void vop2_setup_csc_mode(struct vop2_video_port *vp,
 	} else {
 		y2r_en = false;
 		r2y_en = false;
-		csc_mode = false;
+		csc_mode = CSC_BT601L;
 	}
 
 	vop2_win_write(win, VOP2_WIN_Y2R_EN, y2r_en);

-- 
2.54.0

[PATCH v3 11/14] drm/bridge: dw-hdmi-qp: Log resolution and refresh rate in atomic_enable()

[Cristian Ciocaltea]

The debug entry in the HDMI branch of dw_hdmi_qp_bridge_atomic_enable()
logged the mode field as the fixed 'HDMI' label, giving no information
about the actual display timing being configured.

Extend it to also include the active resolution and refresh rate,
retrieved from the CRTC mode in the incoming atomic state:

  dw_hdmi_qp_bridge_atomic_enable mode=HDMI:1920x1080@50Hz fmt=RGB rate=185625000 bpc=10

This makes the log line self-contained and directly useful when
debugging mode-setting issues, format negotiation, or TMDS rate
mismatches without having to cross-reference a separate mode dump.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
index cfd1b18c06e6..0e6142301791 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
@@ -752,6 +752,8 @@ static void dw_hdmi_qp_bridge_atomic_enable(struct drm_bridge *bridge,
 {
 	struct dw_hdmi_qp *hdmi = bridge->driver_private;
 	struct drm_connector_state *conn_state;
+	const struct drm_display_mode *mode;
+	struct drm_crtc_state *crtc_state;
 	struct drm_connector *connector;
 	unsigned int op_mode;
 
@@ -764,9 +766,15 @@ static void dw_hdmi_qp_bridge_atomic_enable(struct drm_bridge *bridge,
 		return;
 
 	if (connector->display_info.is_hdmi) {
-		dev_dbg(hdmi->dev, "%s mode=HDMI %s rate=%llu bpc=%u\n", __func__,
+		crtc_state = drm_atomic_get_new_crtc_state(state, conn_state->crtc);
+		mode = &crtc_state->mode;
+		dev_dbg(hdmi->dev,
+			"%s mode=HDMI:%ux%u@%uHz fmt=%s rate=%llu bpc=%u\n",
+			__func__, mode->hdisplay, mode->vdisplay,
+			drm_mode_vrefresh(mode),
 			drm_hdmi_connector_get_output_format_name(conn_state->hdmi.output_format),
 			conn_state->hdmi.tmds_char_rate, conn_state->hdmi.output_bpc);
+
 		op_mode = 0;
 		hdmi->tmds_char_rate = conn_state->hdmi.tmds_char_rate;
 	} else {

-- 
2.54.0

[PATCH v3 12/14] drm/rockchip: dw_hdmi_qp: Support 10-bit YUV422 output format

[Cristian Ciocaltea]

Rockchip DW HDMI QP encoder supports YUV 4:2:2 output through
ROCKCHIP_OUT_MODE_YUV422, but was limited to 8-bit depth via
MEDIA_BUS_FMT_UYVY8_1X16.  Add support for its 10-bit counterpart
MEDIA_BUS_FMT_UYVY10_1X20, which carries two 10-bit components per clock
cycle on a 20-bit wide bus.

YUV 4:2:2 always transmits two 12-bit components per pixel, regardless
of the color depth.  From a clock-rate perspective this is equivalent to
three 8-bit RGB components, so configure the HDMI PHY with 8 bpc when
YUV 4:2:2 is in use to keep its output clock aligned with the TMDS
character rate.  Otherwise the PHY PLL output would be scaled by bpc/8
for higher color depths, producing a clock rate that confuses downstream
consumers such as the VOP2 display controller.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
index 9c4cf68d79d5..78473f9a0fce 100644
--- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
+++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
@@ -215,6 +215,7 @@ dw_hdmi_qp_rockchip_encoder_atomic_check(struct drm_encoder *encoder,
 		s->output_mode = ROCKCHIP_OUT_MODE_AAAA;
 		break;
 	case MEDIA_BUS_FMT_UYVY8_1X16:
+	case MEDIA_BUS_FMT_UYVY10_1X20:
 		s->output_mode = ROCKCHIP_OUT_MODE_YUV422;
 		break;
 	case MEDIA_BUS_FMT_UYYVYY8_0_5X24:
@@ -226,7 +227,14 @@ dw_hdmi_qp_rockchip_encoder_atomic_check(struct drm_encoder *encoder,
 	}
 
 	phy_cfg.hdmi.tmds_char_rate = conn_state->hdmi.tmds_char_rate;
-	phy_cfg.hdmi.bpc = conn_state->hdmi.output_bpc;
+	/*
+	 * YUV422 always transmits two 12-bit components per clock cycle,
+	 * regardless of the color depth, which from a rate perspective is
+	 * equivalent to three 8-bit RGB components.  Force 8 bpc here to
+	 * keep the PHY PLL output aligned with the TMDS character rate.
+	 */
+	phy_cfg.hdmi.bpc = (s->output_mode == ROCKCHIP_OUT_MODE_YUV422 ?
+				8 : conn_state->hdmi.output_bpc);
 
 	ret = phy_configure(hdmi->phy, &phy_cfg);
 	if (!ret) {

-- 
2.54.0

[PATCH v3 13/14] drm/rockchip: dw_hdmi_qp: Enable YUV420 output format

[Cristian Ciocaltea]

Both RK3576 and RK3588 SoCs are capable of driving the YUV420 output
color format, and the required bus-format handling and VOP2 support are
already in place.  Advertise it via the platform supported formats so
the HDMI core can select it.

YUV420 halves the TMDS bandwidth compared to RGB/YUV444, which enables
high-resolution modes such as 4K@60Hz on links that cannot otherwise
carry the full-bandwidth signal.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
index 78473f9a0fce..1af29066a0a3 100644
--- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
+++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
@@ -611,7 +611,8 @@ static int dw_hdmi_qp_rockchip_bind(struct device *dev, struct device *master,
 
 	plat_data.supported_formats = BIT(DRM_OUTPUT_COLOR_FORMAT_RGB444) |
 				      BIT(DRM_OUTPUT_COLOR_FORMAT_YCBCR444) |
-				      BIT(DRM_OUTPUT_COLOR_FORMAT_YCBCR422);
+				      BIT(DRM_OUTPUT_COLOR_FORMAT_YCBCR422) |
+				      BIT(DRM_OUTPUT_COLOR_FORMAT_YCBCR420);
 
 	encoder = &hdmi->encoder.encoder;
 	encoder->possible_crtcs = drm_of_find_possible_crtcs(drm, dev->of_node);

-- 
2.54.0

[PATCH v3 14/14] arm64: dts: rockchip: Add RK3588 VOP2 resets

[Cristian Ciocaltea]

Add the missing reset properties to VOP2 on RK3588.

Co-developed-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
---
 arch/arm64/boot/dts/rockchip/rk3588-base.dtsi | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi b/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi
index fc1fdbfd3162..c30a6f35bc52 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi
+++ b/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi
@@ -1651,6 +1651,18 @@ vop: vop@fdd90000 {
 			      "pll_hdmiphy0";
 		iommus = <&vop_mmu>;
 		power-domains = <&power RK3588_PD_VOP>;
+		resets = <&cru SRST_A_VOP>,
+			 <&cru SRST_H_VOP>,
+			 <&cru SRST_D_VOP0>,
+			 <&cru SRST_D_VOP1>,
+			 <&cru SRST_D_VOP2>,
+			 <&cru SRST_D_VOP3>;
+		reset-names = "axi",
+			      "ahb",
+			      "dclk_vp0",
+			      "dclk_vp1",
+			      "dclk_vp2",
+			      "dclk_vp3";
 		rockchip,grf = <&sys_grf>;
 		rockchip,vop-grf = <&vop_grf>;
 		rockchip,vo1-grf = <&vo1_grf>;

-- 
2.54.0

Re: [PATCH v3 06/14] drm/rockchip: vop2: Fix VOP2_MAX_DCLK_RATE overflow on 32-bit

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The driver calls multiple sleeping functions from CRTC atomic enable, disable, and flush callbacks, violating DRM atomic context rules.
--

--- Patch [6]: [PATCH v3 06/14] drm/rockchip: vop2: Fix VOP2_MAX_DCLK_RATE overflow on 32-bit ---
commit 4edf4257240708c6864fcd134b29f53d0456d685
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: vop2: Fix VOP2_MAX_DCLK_RATE overflow on 32-bit

VOP2_MAX_DCLK_RATE is defined as an unsigned long and is used scaled by
the color-depth ratio. On 32-bit architectures, this can overflow when
multiplied. This patch changes the constant to unsigned long long to ensure
64-bit arithmetic is used.

Link: https://lore.kernel.org/all/20260623203310.9F8C81F000E9@smtp.kernel.org/

> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> index e9f5d015b6458..2cd04bedaa2c0 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> @@ -102,7 +102,7 @@ enum vop2_afbc_format {
>  	VOP2_AFBC_FMT_INVALID = -1,
>  };
>  
> -#define VOP2_MAX_DCLK_RATE		600000000UL
> +#define VOP2_MAX_DCLK_RATE		600000000ULL

[Severity: High]
This is a pre-existing issue, but I noticed some potential atomic context
violations in this file.

During non-blocking commits, the DRM core invokes CRTC callbacks like
vop2_crtc_atomic_enable(), vop2_crtc_atomic_disable(), and
vop2_crtc_atomic_flush(). These run in an atomic context where sleeping is
strictly forbidden.

However, the driver appears to invoke sleeping functions on these paths.
For example, in vop2_crtc_atomic_enable():

    vop2_lock(vop2);
    
    ...
    
    ret = clk_prepare_enable(vp->dclk);

Similarly, in vop2_crtc_atomic_disable():

    vop2_lock(vop2);

And in the flush path via vop2_crtc_atomic_try_set_gamma_locked():

    vop2_lock(vop2);

Because vop2_lock() acquires a mutex and clk_prepare_enable() can sleep,
does this violate the atomic context requirement and potentially cause system
instability or deadlocks during display state updates?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=6

Re: [PATCH v3 02/14] drm/rockchip: vop2: Fix resource leak on vop2_enable() error path

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The patch modifies `vop2_enable()` to return an error code and cleanly unwind resources on failure, but fails to update its caller `vop2_crtc_atomic_enable()` to check this return value. This converts a resource leak into a critical double-free and system crash.
--

--- Patch [2]: [PATCH v3 02/14] drm/rockchip: vop2: Fix resource leak on vop2_enable() error path ---
commit d8725e1d1e43785d7a93679c9bc6634b3ebd7fab
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: vop2: Fix resource leak on vop2_enable() error path

vop2_enable() returns void, so callers cannot tell whether powering up
the VOP2 succeeded.

Additionally, if rockchip_drm_dma_attach_device() fails, vop2_enable()
returns without disabling the core clocks or dropping the runtime PM
reference taken earlier, leaking both.

Make vop2_enable() return int and unwind all resources on every failure
path.  As disabling the core clocks is also needed in vop2_disable(),
factor it out into a local helper used by both.

Link: https://lore.kernel.org/all/20260623202043.8BFE21F000E9@smtp.kernel.org/

> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> index 4cce3e336f5b3..e11dc8c7f78ba 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c

[ ... ]

> @@ -871,7 +878,7 @@ static void rk3588_vop2_power_domain_enable_all(struct vop2 *vop2)
>  	vop2_writel(vop2, RK3588_SYS_PD_CTRL, pd);
>  }
>  
> -static void vop2_enable(struct vop2 *vop2)
> +static int vop2_enable(struct vop2 *vop2)
>  {
>  	int ret;
>  	u32 version;

[ ... ]

> @@ -932,6 +938,17 @@ static void vop2_enable(struct vop2 *vop2)
>  		    VOP2_INT_BUS_ERRPR << 16 | VOP2_INT_BUS_ERRPR);
>  	vop2_writel(vop2, RK3568_SYS1_INT_EN,
>  		    VOP2_INT_BUS_ERRPR << 16 | VOP2_INT_BUS_ERRPR);
> +
> +	return 0;
> +
> +err_detach_dma:
> +	rockchip_drm_dma_detach_device(vop2->drm, vop2->dev);
> +err_disable_clks:
> +	vop2_core_clks_disable_unprepare(vop2);
> +err_put_pm:
> +	pm_runtime_put_sync(vop2->dev);
> +
> +	return ret;
>  }

[Severity: High]
Does this code leave vop2_crtc_atomic_enable() vulnerable to a crash if
vop2_enable() fails?

The newly added error path in vop2_enable() unwinds resources (power and
clocks), invalidating the assumption that the hardware is active.

However, the caller vop2_crtc_atomic_enable() still ignores the return
value:

	if (!vop2->enable_count)
		vop2_enable(vop2);

	vop2->enable_count++;

If vop2_enable() fails (e.g., due to DMA attach failure), it disables the
core clocks and puts the PM runtime.

Because vop2_crtc_atomic_enable() ignores the error, it unconditionally
increments enable_count and proceeds to access VOP2 MMIO registers.

Can this result in a system crash (Synchronous External Abort) due to MMIO
access to unpowered/unclocked hardware, followed by a double-free of PM
runtime and clock resources on teardown?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=2

Re: [PATCH v3 03/14] drm/rockchip: vop2: Balance state on atomic_enable() error paths

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Unhandled VBLANK event on `atomic_enable` error paths causes userspace to hang indefinitely and triggers a kernel warning.
--

--- Patch [3]: [PATCH v3 03/14] drm/rockchip: vop2: Balance state on atomic_enable() error paths ---
commit 852e494693bdb0b841a64b02cd2bceb48e307231
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: vop2: Balance state on atomic_enable() error paths

This commit attempts to fix resource leaks in the error paths of
vop2_crtc_atomic_enable() by properly decrementing enable_count and
unpreparing clocks. It also tracks the per-VP bring-up state to skip 
teardown in atomic_disable() and hardware programming in atomic_flush() 
when the enable phase does not complete successfully.

Link: https://lore.kernel.org/all/20260623202043.8BFE21F000E9@smtp.kernel.org/

[ ... ]

> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> index e11dc8c7f78ba..271c3cba6fb49 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c

[ ... ]

> @@ -2034,13 +2066,19 @@ static void vop2_crtc_atomic_flush(struct drm_crtc *crtc,
>  	struct vop2_video_port *vp = to_vop2_video_port(crtc);
>  	struct vop2 *vop2 = vp->vop2;
>  
> -	/* In case of modeset, gamma lut update already happened in atomic enable */
> -	if (!drm_atomic_crtc_needs_modeset(crtc_state) && crtc_state->color_mgmt_changed)
> -		vop2_crtc_atomic_try_set_gamma_locked(vop2, vp, crtc, crtc_state);
> +	if (vp->enabled) {
> +		/*
> +		 * In case of modeset, gamma lut update already happened
> +		 * in atomic enable.
> +		 */
> +		if (!drm_atomic_crtc_needs_modeset(crtc_state) &&
> +		    crtc_state->color_mgmt_changed)
> +			vop2_crtc_atomic_try_set_gamma_locked(vop2, vp, crtc, crtc_state);
>  
> -	vop2_post_config(crtc, false, crtc_state, old_crtc_state);
> +		vop2_post_config(crtc, false, crtc_state, old_crtc_state);
>  
> -	vop2_cfg_done(vp);
> +		vop2_cfg_done(vp);
> +	}
>  
>  	spin_lock_irq(&crtc->dev->event_lock);

[Severity: High]
This is a pre-existing issue, but does this code leave userspace hanging on
atomic_enable error paths?

Since this patch uses vp->enabled to skip hardware programming on failure,
what happens to the pending vblank event? If vop2_crtc_atomic_enable() fails,
it returns before calling drm_crtc_vblank_on().

However, immediately following the spin_lock_irq() in vop2_crtc_atomic_flush(),
the code unconditionally attempts to arm the event:

	if (crtc->state->event) {
		WARN_ON(drm_crtc_vblank_get(crtc) != 0);
		drm_crtc_arm_vblank_event(crtc, crtc->state->event);
		crtc->state->event = NULL;
	}

Because vblank wasn't enabled during the failed bring-up,
drm_crtc_vblank_get() fails and triggers the WARN_ON. The event is armed but
never signaled because the disabled hardware generates no interrupts.

Should the driver consume the event directly here (e.g., using
drm_crtc_send_vblank_event) if vp->enabled is false to prevent the compositor
from blocking indefinitely?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=3

Re: [PATCH v3 05/14] drm/rockchip: vop2: Avoid division by zero when computing max_dclk

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] The expression `VOP2_MAX_DCLK_RATE * 8` silently overflows `unsigned long` on 32-bit architectures, resulting in a drastically lower `max_dclk` value.
- [High] The CRTC atomic enable callback invokes multiple sleeping operations while executing in an atomic context.
--

--- Patch [5]: [PATCH v3 05/14] drm/rockchip: vop2: Avoid division by zero when computing max_dclk ---
commit ba0b89bfb2234e515bb1a3e786beceb00c9813f0
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: vop2: Avoid division by zero when computing max_dclk

The commit modifies vop2_crtc_atomic_enable() to avoid division by zero
when vcstate->output_bpc is 0 by defaulting bpc to 8. This happens when
the CRTC drives a non-HDMI output.

Link: https://lore.kernel.org/all/20260623203310.9F8C81F000E9@smtp.kernel.org/

> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> index 4bb1985b87267..e9f5d015b6458 100644
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> @@ -1931,8 +1931,9 @@ static void vop2_crtc_atomic_enable(struct drm_crtc *crtc,
>  	 * to 4K@60Hz, if available, otherwise keep using the system CRU.
>  	 */
>  	if (vop2->pll_hdmiphy0 || vop2->pll_hdmiphy1) {
> -		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8,
> -							       vcstate->output_bpc);
> +		unsigned int bpc = vcstate->output_bpc ?: 8;
> +		unsigned long max_dclk = DIV_ROUND_CLOSEST_ULL(VOP2_MAX_DCLK_RATE * 8, bpc);

[Severity: High]
This is a pre-existing issue, but does VOP2_MAX_DCLK_RATE * 8 silently 
overflow unsigned long on 32-bit architectures?

Since VOP2_MAX_DCLK_RATE is 600000000UL, multiplying by 8 mathematically 
yields 4,800,000,000. On 32-bit architectures, this wraps around to 
505,032,704 before being passed to DIV_ROUND_CLOSEST_ULL(), which would
result in max_dclk being ~63 MHz instead of 600 MHz.

Could this prevent higher resolutions like 1080p and 4K from being 
configured properly on 32-bit kernels?

> +
>  		if (clock <= max_dclk) {
>  			drm_for_each_encoder_mask(encoder, crtc->dev, crtc_state->encoder_mask) {
>  				struct rockchip_encoder *rkencoder = to_rockchip_encoder(encoder);

[Severity: High]
This is a pre-existing issue, but does vop2_crtc_atomic_enable() execute 
sleeping operations within an atomic context?

Looking earlier in vop2_crtc_atomic_enable():

	vop2_lock(vop2);

	ret = clk_prepare_enable(vp->dclk);
	if (ret < 0) {

The function unconditionally acquires a mutex via vop2_lock() and calls 
clk_prepare_enable(), which can sleep. Since CRTC atomic enable callbacks
execute in an atomic context during non-blocking commits, could this cause
system instability or deadlocks?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=5

Re: [PATCH v3 07/14] drm/rockchip: vop2: Reset AXI and DCLK to improve robustness

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Using devm_* APIs in component bind callbacks causes memory leaks, duplicate resource registrations on unbind/rebind, and fatal bind failures. The new patch exacerbates this by requesting exclusive reset controls.
--

--- Patch [7]: [PATCH v3 07/14] drm/rockchip: vop2: Reset AXI and DCLK to improve robustness ---
commit 9b374dfb00e64eeab45f2b86b059f7b28a55a177
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: vop2: Reset AXI and DCLK to improve robustness

This patch adds assertions for the AXI reset in the CRTC disable path and the
VP DCLK reset in the enable path. These optional resets are intended to leave
the hardware in a clean state and prevent random display output glitches on
platforms like RK3588.

> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c
> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c

[ ... ]

> @@ -2596,6 +2621,12 @@ static int vop2_create_crtcs(struct vop2 *vop2)
>  			return dev_err_probe(drm->dev, PTR_ERR(vp->dclk),
>  					     "failed to get %s\n", dclk_name);
>  
> +		vp->dclk_rst = devm_reset_control_get_optional_exclusive(vop2->dev,
> +									 dclk_name);
> +		if (IS_ERR(vp->dclk_rst))
> +			return dev_err_probe(drm->dev, PTR_ERR(vp->dclk_rst),
> +					     "failed to get %s reset\n", dclk_name);
> +

[Severity: High]
This is a pre-existing issue, but will this exclusive reset control request
fail with -EBUSY on component rebind?

Because vop2_create_crtcs() is called from the component bind path in
vop2_bind() in drivers/gpu/drm/rockchip/rockchip_drm_vop2.c, any resources
allocated with devm_ functions tied to the platform device (vop2->dev) are
not released when the component is unbound. 

When the component is rebound, the previous exclusive reset control is still
held. It appears this might cause devm_reset_control_get_optional_exclusive()
to deterministically fail with -EBUSY since the prior exclusive grab was
never released.

Should this use reset_control_get_optional_exclusive() instead and be
manually released in the unbind path, or use a component-specific managed
API?

>  		np = of_graph_get_remote_node(dev->of_node, i, -1);
>  		if (!np) {
>  			drm_dbg(vop2->drm, "%s: No remote for vp%d\n", __func__, i);

[ ... ]

> @@ -2955,6 +2986,11 @@ static int vop2_bind(struct device *dev, struct device *master, void *data)
>  		return dev_err_probe(drm->dev, PTR_ERR(vop2->pll_hdmiphy1),
>  				     "failed to get pll_hdmiphy1\n");
>  
> +	vop2->axi_rst = devm_reset_control_get_optional_exclusive(vop2->dev, "axi");
> +	if (IS_ERR(vop2->axi_rst))
> +		return dev_err_probe(drm->dev, PTR_ERR(vop2->axi_rst),
> +				     "failed to get axi reset\n");
> +

[Severity: High]
This is also a pre-existing issue, but will requesting an exclusive reset
control tied to the platform device here also cause a fatal -EBUSY error
during a component rebind?

Similar to the dclk_rst allocation, vop2->axi_rst is allocated in vop2_bind()
with the devm_ API. When the DRM component master is unbound and then rebound,
vop2_bind() executes again and attempts to grab the exclusive reset control.
Since the previous devm_ allocation is tied to the persistent platform device
and wasn't released during component unbind, will this fail immediately with
-EBUSY?

>  	vop2->irq = platform_get_irq(pdev, 0);
>  	if (vop2->irq < 0)
>  		return dev_err_probe(drm->dev, vop2->irq, "cannot find irq for vop2\n");

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=7

Re: [PATCH v3 12/14] drm/rockchip: dw_hdmi_qp: Support 10-bit YUV422 output format

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The `atomic_check` callback unconditionally mutates global driver state and configures hardware PHY, breaking DRM atomic design principles and corrupting state during `TEST_ONLY` commits.
--

--- Patch [12]: [PATCH v3 12/14] drm/rockchip: dw_hdmi_qp: Support 10-bit YUV422 output format ---
commit 4acf186e6f48344c73efb358bbf5b17c6a065aa5
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: dw_hdmi_qp: Support 10-bit YUV422 output format

This commit adds support for 10-bit YUV422 output via the
MEDIA_BUS_FMT_UYVY10_1X20 format. It also forces 8 bpc for the HDMI PHY
configuration when YUV 4:2:2 is in use to keep the PHY PLL output aligned
with the TMDS character rate.

> diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> --- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> +++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
[ ... ]
> @@ -226,7 +227,14 @@ dw_hdmi_qp_rockchip_encoder_atomic_check(struct drm_encoder *encoder,
>  	}
>  
>  	phy_cfg.hdmi.tmds_char_rate = conn_state->hdmi.tmds_char_rate;
> -	phy_cfg.hdmi.bpc = conn_state->hdmi.output_bpc;
> +	/*
> +	 * YUV422 always transmits two 12-bit components per clock cycle,
> +	 * regardless of the color depth, which from a rate perspective is
> +	 * equivalent to three 8-bit RGB components.  Force 8 bpc here to
> +	 * keep the PHY PLL output aligned with the TMDS character rate.
> +	 */
> +	phy_cfg.hdmi.bpc = (s->output_mode == ROCKCHIP_OUT_MODE_YUV422 ?
> +				8 : conn_state->hdmi.output_bpc);
>  
>  	ret = phy_configure(hdmi->phy, &phy_cfg);
>  	if (!ret) {
>  		hdmi->tmds_char_rate = conn_state->hdmi.tmds_char_rate;

[Severity: High]
This is a pre-existing issue, but does dw_hdmi_qp_rockchip_encoder_atomic_check
mutate active hardware state and global driver state during TEST_ONLY commits?

Because atomic_check callbacks are executed when userspace issues a
DRM_MODE_ATOMIC_TEST_ONLY ioctl to probe display configurations, calling
phy_configure() here could actively alter the PHY hardware under an active
display.

Additionally, modifying hdmi->tmds_char_rate mutates global driver state
even if the test commit is subsequently discarded by the kernel, potentially
corrupting the active configuration. Should these side effects be deferred to
an atomic_enable or atomic_update callback instead?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=12

Re: [PATCH v3 13/14] drm/rockchip: dw_hdmi_qp: Enable YUV420 output format

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Modifying hardware or global driver state within atomic_check callback.
--

--- Patch [13]: [PATCH v3 13/14] drm/rockchip: dw_hdmi_qp: Enable YUV420 output format ---
commit 886f3880152ce2fc09e568484e921ca99dad89d4
Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>

drm/rockchip: dw_hdmi_qp: Enable YUV420 output format

Both RK3576 and RK3588 SoCs are capable of driving the YUV420 output
color format, and the required bus-format handling and VOP2 support are
already in place.  Advertise it via the platform supported formats so
the HDMI core can select it.

YUV420 halves the TMDS bandwidth compared to RGB/YUV444, which enables
high-resolution modes such as 4K@60Hz on links that cannot otherwise
carry the full-bandwidth signal.

> diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> index 78473f9a0fce3..1af29066a0a31 100644
> --- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> +++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c

[ ... ]

[Severity: High]
This isn't a bug introduced by this patch, but I noticed something concerning
in dw_hdmi_qp_rockchip_encoder_atomic_check() in this same file:

drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c:dw_hdmi_qp_rockchip_encoder_atomic_check() {
    ...
	ret = phy_configure(hdmi->phy, &phy_cfg);
	if (!ret) {
		hdmi->tmds_char_rate = conn_state->hdmi.tmds_char_rate;
		s->output_type = DRM_MODE_CONNECTOR_HDMIA;
		s->output_bpc = conn_state->hdmi.output_bpc;
		s->bus_format = ingest_fmt;
	} else {
    ...
}

Does this code accidentally configure the hardware PHY and modify global
driver state (hdmi->tmds_char_rate) during the atomic_check phase?

When userspace performs an atomic modeset with the DRM_MODE_ATOMIC_TEST_ONLY
flag, dw_hdmi_qp_rockchip_encoder_atomic_check() is invoked to validate the
state. Modifying hardware or global state here breaks the guarantee that
atomic test-only commits are free of side effects. This could cause the
hardware state to desynchronize from the DRM core's expectations.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709-dw-hdmi-qp-yuv-v3-0-a4a982a9f2e7@collabora.com?part=13