[PATCH v2 0/2] Add support for Renesas RZ/G3L LVDS encoder

[PATCH v2 0/2] Add support for Renesas RZ/G3L LVDS encoder

[Biju]

From: Biju Das <biju.das.jz@bp.renesas.com>

Add support for the RZ/G3L LVDS encoder driver. It operates in single-link
mode with 4 lanes (Data) + 1 lane (Clock) and supports pixel clock rates
from 25 to 87 MHz. The LVDS module cannot be used at the same time as
MIPI-DSI. However, LVDS and the DSI interface share a peripheral clock and
the MIPI_DSI_PRESET_N reset signal. Also, the MIPI_DSI_CMN_RSTB and
MIPI_DSI_ARESET_N reset signals must be asserted before using the LVDS
module.

v2->v3:
 * Dropped patch#1 as it is accepted.
 * Replace drm_atomic_state with drm_atomic_commit in
   rzg3l_lvds_atomic_{en,dis}able().
 * Drop local variable ret and dev_err() messages in
   rzg3l_lvds_atomic_enable(); use WARN_ON() instead to
   capture unexpected failures since atomic_enable should not fail.
 * Drop local variable next_bridge from rzg3l_lvds_probe().
v1->v2:
 * Collected the tags for binding patches.
 * Dropped unused function rzg3l_lvds_is_connected() and removed the 
   corresponding header file rzg3l_lvds.h
 * Dropped next_bridge from struct rzg3l_lvds instead using bridge's
   next_bridge.
 * Replaced pm_runtime_resume_and_get()->pm_runtime_get_sync() as
   atomic_enable doesn't fail and for each enable there always will be an
   atomic_disable() call.
 * Started using DEFINE_RUNTIME_DEV_PM_OPS for PM callback.
 * Replaced rzg3l_lvds_parse_dt() with devm_drm_of_get_bridge() in probe()
 * Started using reset_control_bulk_*() in rzg3l_lvds_pm_runtime_{suspend,
   resume}().

Biju Das (2):
  dt-bindings: display: bridge: Document Renesas RZ/G3L LVDS encoder
  drm: renesas: rz-du: Add support for RZ/G3L LVDS encoder

 .../bridge/renesas,r9a08g046-lvds.yaml        | 128 ++++++++
 drivers/gpu/drm/renesas/rz-du/Kconfig         |  13 +
 drivers/gpu/drm/renesas/rz-du/Makefile        |   1 +
 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c    | 277 ++++++++++++++++++
 .../gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h   |  26 ++
 5 files changed, 445 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/bridge/renesas,r9a08g046-lvds.yaml
 create mode 100644 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c
 create mode 100644 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h

-- 
2.43.0

[PATCH v2 1/2] dt-bindings: display: bridge: Document Renesas RZ/G3L LVDS encoder

[Biju]

From: Biju Das <biju.das.jz@bp.renesas.com>

Document the LVDS encoder IP found on the RZ/G3L SoC. It supports
single-link mode. LVDS and the DSI interface share a peripheral clock and
the MIPI_DSI_PRESET_N reset signal. However, the LVDS module cannot be
used at the same time as MIPI-DSI.

Signed-off-by: Tommaso Merciai <tommaso.merciai.xr@bp.renesas.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
---
v2->v3:
 * No change.
v1->v2:
 * Collected tag.
---
 .../bridge/renesas,r9a08g046-lvds.yaml        | 128 ++++++++++++++++++
 1 file changed, 128 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/display/bridge/renesas,r9a08g046-lvds.yaml

diff --git a/Documentation/devicetree/bindings/display/bridge/renesas,r9a08g046-lvds.yaml b/Documentation/devicetree/bindings/display/bridge/renesas,r9a08g046-lvds.yaml
new file mode 100644
index 000000000000..b1f6d020ae7b
--- /dev/null
+++ b/Documentation/devicetree/bindings/display/bridge/renesas,r9a08g046-lvds.yaml
@@ -0,0 +1,128 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/bridge/renesas,r9a08g046-lvds.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/G3L LVDS Encoder
+
+maintainers:
+  - Biju Das <biju.das.jz@bp.renesas.com>
+  - Tommaso Merciai <tommaso.merciai.xr@bp.renesas.com>
+
+description: |
+  This binding describe the LVDS encoder embedded in the Renesas RZ/G3L
+  SoC. The encoder can operate in LVDS Single-link mode with 4 lanes
+  (Data) + 1 lane (Clock).
+
+properties:
+  compatible:
+    const: renesas,r9a08g046-lvds
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    items:
+      - description: Peripheral clock
+      - description: PHY clock
+      - description: Dot clock
+
+  clock-names:
+    items:
+      - const: pclk
+      - const: phyclk
+      - const: dotclk
+
+  resets:
+    items:
+      - description: LVDS_RESET_N
+      - description: MIPI_DSI_PRESET_N
+      - description: MIPI_DSI_CMN_RSTB
+      - description: MIPI_DSI_ARESET_N
+
+  reset-names:
+    items:
+      - const: lvdrst
+      - const: prst
+      - const: rst
+      - const: arst
+
+  power-domains:
+    maxItems: 1
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Input channel, directly connected to the Display Unit.
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: |
+          Output channel, directly connected to the LVDS panel or bridge.
+
+    required:
+      - port@0
+      - port@1
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+  - resets
+  - reset-names
+  - power-domains
+  - ports
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/renesas,r9a08g046-cpg.h>
+
+    lvds-cmn@108a0000 {
+        compatible = "renesas,r9a08g046-lvds-cmn",
+                     "simple-mfd", "syscon";
+        reg = <0x108a0000 0x10000>;
+        #address-cells = <1>;
+        #size-cells = <1>;
+
+        lvds0: lvds@10 {
+            compatible = "renesas,r9a08g046-lvds";
+            reg = <0x10 0x8>;
+            clocks = <&cpg CPG_MOD R9A08G046_MIPI_DSI_PCLK>,
+                     <&cpg CPG_MOD R9A08G046_LVDS_PLLCLK>,
+                     <&cpg CPG_MOD R9A08G046_LVDS_CLK_DOT0>;
+            clock-names = "pclk", "phyclk", "dotclk";
+            resets = <&cpg R9A08G046_LVDS_RESET_N>,
+                     <&cpg R9A08G046_MIPI_DSI_PRESET_N>,
+                     <&cpg R9A08G046_MIPI_DSI_CMN_RSTB>,
+                     <&cpg R9A08G046_MIPI_DSI_ARESET_N>;
+            reset-names = "lvdrst", "prst", "rst", "arst";
+            power-domains = <&cpg>;
+
+            ports {
+                #address-cells = <1>;
+                #size-cells = <0>;
+
+                port@0 {
+                    reg = <0>;
+                    lvds0_in: endpoint {
+                        remote-endpoint = <&du_out_lvds0>;
+                    };
+                };
+
+                port@1 {
+                    reg = <1>;
+                    lvds0_out: endpoint {
+                        remote-endpoint = <&panel_in>;
+                    };
+                };
+            };
+        };
+    };
+...
-- 
2.43.0

[PATCH v2 2/2] drm: renesas: rz-du: Add support for RZ/G3L LVDS encoder

[Biju]

From: Biju Das <biju.das.jz@bp.renesas.com>

Add support for the RZ/G3L LVDS encoder driver. It operates in single-link
mode with 4 lanes (Data) + 1 lane (Clock) and supports pixel clock rates
from 25 to 87 MHz. The LVDS module cannot be used at the same time as
MIPI-DSI. However, LVDS and the DSI interface share a peripheral clock and
the MIPI_DSI_PRESET_N reset signal. Also, the MIPI_DSI_CMN_RSTB and
MIPI_DSI_ARESET_N reset signals must be asserted before using the LVDS
module.

Signed-off-by: Tommaso Merciai <tommaso.merciai.xr@bp.renesas.com>
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
---
v2->v3:
 * Replace drm_atomic_state with drm_atomic_commit in
   rzg3l_lvds_atomic_{en,dis}able().
 * Drop local variable ret and dev_err() messages in
   rzg3l_lvds_atomic_enable(); use WARN_ON() instead to
   capture unexpected failures since atomic_enable should not fail.
 * Drop local variable next_bridge from rzg3l_lvds_probe().
v1->v2:
 * Dropped unused function rzg3l_lvds_is_connected() and removed the 
   corresponding header file rzg3l_lvds.h
 * Dropped next_bridge from struct rzg3l_lvds instead using bridge's
   next_bridge.
 * Replaced pm_runtime_resume_and_get()->pm_runtime_get_sync() as
   atomic_enable doesn't fail and for each enable there always will be an
   atomic_disable() call.
 * Started using DEFINE_RUNTIME_DEV_PM_OPS for PM callback.
 * Replaced rzg3l_lvds_parse_dt() with devm_drm_of_get_bridge() in probe()
 * Started using reset_control_bulk_*() in rzg3l_lvds_pm_runtime_{suspend,
   resume}()
---
 drivers/gpu/drm/renesas/rz-du/Kconfig         |  13 +
 drivers/gpu/drm/renesas/rz-du/Makefile        |   1 +
 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c    | 277 ++++++++++++++++++
 .../gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h   |  26 ++
 4 files changed, 317 insertions(+)
 create mode 100644 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c
 create mode 100644 drivers/gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h

diff --git a/drivers/gpu/drm/renesas/rz-du/Kconfig b/drivers/gpu/drm/renesas/rz-du/Kconfig
index 7f2ef7137ae5..cbfc7b6bccb8 100644
--- a/drivers/gpu/drm/renesas/rz-du/Kconfig
+++ b/drivers/gpu/drm/renesas/rz-du/Kconfig
@@ -26,3 +26,16 @@ config DRM_RZG2L_MIPI_DSI
 	def_tristate DRM_RZG2L_DU
 	depends on DRM_RZG2L_USE_MIPI_DSI
 	select DRM_MIPI_DSI
+
+config DRM_RZG3L_USE_LVDS
+	bool "RZ/G3L DU LVDS Encoder Support"
+	depends on DRM_BRIDGE && OF
+	default DRM_RZG2L_DU
+	help
+	  Enable support for the RZ/G3L Display Unit embedded LVDS encoders.
+
+config DRM_RZG3L_LVDS
+	def_tristate DRM_RZG2L_DU
+	depends on DRM_RZG3L_USE_LVDS
+	select DRM_KMS_HELPER
+	select DRM_PANEL
diff --git a/drivers/gpu/drm/renesas/rz-du/Makefile b/drivers/gpu/drm/renesas/rz-du/Makefile
index 2987900ea6b6..46decb7ac4f1 100644
--- a/drivers/gpu/drm/renesas/rz-du/Makefile
+++ b/drivers/gpu/drm/renesas/rz-du/Makefile
@@ -8,3 +8,4 @@ rzg2l-du-drm-$(CONFIG_VIDEO_RENESAS_VSP1)	+= rzg2l_du_vsp.o
 obj-$(CONFIG_DRM_RZG2L_DU)		+= rzg2l-du-drm.o
 
 obj-$(CONFIG_DRM_RZG2L_MIPI_DSI)	+= rzg2l_mipi_dsi.o
+obj-$(CONFIG_DRM_RZG3L_LVDS)		+= rzg3l_lvds.o
diff --git a/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c b/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c
new file mode 100644
index 000000000000..a51c3e5a2efe
--- /dev/null
+++ b/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds.c
@@ -0,0 +1,277 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * RZ/G3L LVDS Encoder Driver
+ *
+ * Copyright (C) 2026 Renesas Electronics Corporation
+ */
+
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/media-bus-format.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_graph.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+#include <linux/reset.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_bridge.h>
+#include <drm/drm_of.h>
+#include <drm/drm_panel.h>
+#include <drm/drm_probe_helper.h>
+
+#include "rzg3l_lvds_regs.h"
+
+enum rzg3l_lvds_mode {
+	RZG3L_LVDS_MODE_JEIDA = 0,
+	RZG3L_LVDS_MODE_JEIDA_MIRROR = 1,
+	RZG3L_LVDS_MODE_MODE2 = 2,
+	RZG3L_LVDS_MODE_MODE2_MIRROR = 3,
+	RZG3L_LVDS_MODE_VESA = 4,
+	RZG3L_LVDS_MODE_VESA_MIRROR = 5,
+	RZG3L_LVDS_MODE_MODE6 = 6,
+	RZG3L_LVDS_MODE_MODE6_MIRROR = 7,
+};
+
+struct rzg3l_lvds {
+	struct device *dev;
+	struct reset_control *prstc;
+	struct reset_control *lvd_rstc;
+	struct regmap *regmap;
+	struct drm_bridge bridge;
+};
+
+#define bridge_to_rzg3l_lvds(b) \
+	container_of(b, struct rzg3l_lvds, bridge)
+
+/* -----------------------------------------------------------------------------
+ * Bridge
+ */
+
+static void rzg3l_lvds_atomic_enable(struct drm_bridge *bridge,
+				     struct drm_atomic_commit *state)
+{
+	struct rzg3l_lvds *lvds = bridge_to_rzg3l_lvds(bridge);
+	const struct drm_bridge_state *bridge_state;
+	u32 fmt;
+
+	/* Get the LVDS format from the bridge state. */
+	bridge_state = drm_atomic_get_new_bridge_state(state, bridge);
+	if (WARN_ON(!bridge_state))
+		return;
+
+	switch (bridge_state->output_bus_cfg.format) {
+	case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA:
+		fmt = RZG3L_LVDS_MODE_JEIDA;
+		break;
+	case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG:
+		fmt = RZG3L_LVDS_MODE_VESA;
+		break;
+	default:
+		fmt = RZG3L_LVDS_MODE_VESA;
+		dev_warn(lvds->dev, "Unsupported bus fmt 0x%04x\n",
+			 bridge_state->output_bus_cfg.format);
+		break;
+	}
+
+	if (WARN_ON(pm_runtime_get_sync(lvds->dev) < 0))
+		return;
+
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_EN_BGR, LVDS_0_PHY_CH_EN_BGR);
+	fsleep(20);
+
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_EN_LDO, LVDS_0_PHY_CH_EN_LDO);
+	fsleep(10);
+
+	regmap_write(lvds->regmap, LVDS_CMN, LVDS_CMN_RST_PHY0_SEL);
+	regmap_update_bits(lvds->regmap, LVDS_0_CTL_OFFSET,
+			   LVDS_0_CTL_FMT_SEL_MSK,
+			   FIELD_PREP(LVDS_0_CTL_FMT_SEL_MSK, fmt));
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_IO_EN_MSK, LVDS_0_PHY_CH_IO_EN);
+	regmap_write(lvds->regmap, LVDS_CMN,
+		     LVDS_CMN_RST_PHY0_SEL | LVDS_CMN_PHY_RESET);
+	fsleep(100);
+}
+
+static void rzg3l_lvds_atomic_disable(struct drm_bridge *bridge,
+				      struct drm_atomic_commit *state)
+{
+	struct rzg3l_lvds *lvds = bridge_to_rzg3l_lvds(bridge);
+
+	regmap_update_bits(lvds->regmap, LVDS_CMN, LVDS_CMN_PHY_RESET, 0);
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_IO_EN_MSK, 0);
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_EN_LDO, 0);
+	regmap_update_bits(lvds->regmap, LVDS_0_PHY_OFFSET,
+			   LVDS_0_PHY_CH_EN_BGR, 0);
+
+	pm_runtime_put(lvds->dev);
+}
+
+static int rzg3l_lvds_attach(struct drm_bridge *bridge,
+			     struct drm_encoder *encoder,
+			     enum drm_bridge_attach_flags flags)
+{
+	struct rzg3l_lvds *lvds = bridge_to_rzg3l_lvds(bridge);
+
+	if (!lvds->bridge.next_bridge)
+		return 0;
+
+	return drm_bridge_attach(encoder, lvds->bridge.next_bridge, bridge, flags);
+}
+
+static enum drm_mode_status
+rzg3l_lvds_bridge_mode_valid(struct drm_bridge *bridge,
+			     const struct drm_display_info *info,
+			     const struct drm_display_mode *mode)
+{
+	if (mode->clock > 87000)
+		return MODE_CLOCK_HIGH;
+
+	if (mode->clock < 25000)
+		return MODE_CLOCK_LOW;
+
+	return MODE_OK;
+}
+
+static const struct drm_bridge_funcs rzg3l_lvds_bridge_ops = {
+	.attach = rzg3l_lvds_attach,
+	.atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
+	.atomic_reset = drm_atomic_helper_bridge_reset,
+	.atomic_enable = rzg3l_lvds_atomic_enable,
+	.atomic_disable = rzg3l_lvds_atomic_disable,
+	.mode_valid = rzg3l_lvds_bridge_mode_valid,
+};
+
+/* -----------------------------------------------------------------------------
+ * Power Management
+ */
+
+static int rzg3l_lvds_pm_runtime_suspend(struct device *dev)
+{
+	struct rzg3l_lvds *lvds = dev_get_drvdata(dev);
+	struct reset_control_bulk_data resets[] = {
+		{ .rstc = lvds->lvd_rstc },
+		{ .rstc = lvds->prstc },
+	};
+
+	return reset_control_bulk_assert(ARRAY_SIZE(resets), resets);
+}
+
+static int rzg3l_lvds_pm_runtime_resume(struct device *dev)
+{
+	struct rzg3l_lvds *lvds = dev_get_drvdata(dev);
+	struct reset_control_bulk_data resets[] = {
+		{ .rstc = lvds->lvd_rstc },
+		{ .rstc = lvds->prstc },
+	};
+
+	return reset_control_bulk_deassert(ARRAY_SIZE(resets), resets);
+}
+
+static DEFINE_RUNTIME_DEV_PM_OPS(rzg3l_lvds_pm_ops,
+				 rzg3l_lvds_pm_runtime_suspend,
+				 rzg3l_lvds_pm_runtime_resume, NULL);
+
+/* -----------------------------------------------------------------------------
+ * Probe & Remove
+ */
+
+static int rzg3l_lvds_probe(struct platform_device *pdev)
+{
+	struct reset_control *rstc, *arstc;
+	struct device *dev = &pdev->dev;
+	struct rzg3l_lvds *lvds;
+	int ret;
+
+	lvds = devm_drm_bridge_alloc(dev, struct rzg3l_lvds, bridge,
+				     &rzg3l_lvds_bridge_ops);
+	if (IS_ERR(lvds))
+		return PTR_ERR(lvds);
+
+	lvds->dev = dev;
+	lvds->bridge.of_node = pdev->dev.of_node;
+
+	lvds->regmap = syscon_node_to_regmap(dev->of_node->parent);
+	if (IS_ERR(lvds->regmap))
+		return PTR_ERR(lvds->regmap);
+
+	rstc = devm_reset_control_get_optional_exclusive(dev, "rst");
+	if (IS_ERR(rstc))
+		return dev_err_probe(dev, PTR_ERR(rstc), "failed to get rst\n");
+
+	arstc = devm_reset_control_get_optional_exclusive(dev, "arst");
+	if (IS_ERR(arstc))
+		return dev_err_probe(dev, PTR_ERR(arstc),
+				     "failed to get arst\n");
+
+	lvds->prstc = devm_reset_control_get_shared(dev, "prst");
+	if (IS_ERR(lvds->prstc))
+		return dev_err_probe(dev, PTR_ERR(lvds->prstc),
+				     "failed to get prst\n");
+
+	lvds->lvd_rstc = devm_reset_control_get_shared(dev, "lvdrst");
+	if (IS_ERR(lvds->lvd_rstc))
+		return dev_err_probe(dev, PTR_ERR(lvds->lvd_rstc),
+				     "failed to get core reset\n");
+
+	platform_set_drvdata(pdev, lvds);
+	ret = devm_pm_runtime_enable(dev);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to enable Runtime PM\n");
+
+	lvds->bridge.next_bridge = devm_drm_of_get_bridge(dev, dev->of_node, 1, 0);
+	if (IS_ERR(lvds->bridge.next_bridge))
+		return dev_err_probe(dev, PTR_ERR(lvds->bridge.next_bridge),
+				     "failed to get next bridge\n");
+
+	ret = reset_control_assert(rstc);
+	if (ret < 0)
+		return ret;
+
+	ret = reset_control_assert(arstc);
+	if (ret < 0)
+		return ret;
+
+	ret = devm_drm_bridge_add(dev, &lvds->bridge);
+	if (ret)
+		return dev_err_probe(dev, ret,
+				     "Failed to register drm bridge\n");
+
+	return ret;
+}
+
+static const struct of_device_id rzg3l_lvds_of_table[] = {
+	{ .compatible = "renesas,r9a08g046-lvds" },
+	{ /* sentinel */ }
+};
+
+MODULE_DEVICE_TABLE(of, rzg3l_lvds_of_table);
+
+static struct platform_driver rzg3l_lvds_platform_driver = {
+	.probe		= rzg3l_lvds_probe,
+	.driver		= {
+		.name	= "rzg3l-lvds",
+		.pm	= pm_ptr(&rzg3l_lvds_pm_ops),
+		.of_match_table = rzg3l_lvds_of_table,
+	},
+};
+
+module_platform_driver(rzg3l_lvds_platform_driver);
+
+MODULE_AUTHOR("Biju Das <biju.das.jz@bp.renesas.com>");
+MODULE_AUTHOR("Tommaso Merciai <tommaso.merciai.xr@bp.renesas.com>");
+MODULE_DESCRIPTION("Renesas RZ/G3L LVDS Encoder Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h b/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h
new file mode 100644
index 000000000000..281b7648f168
--- /dev/null
+++ b/drivers/gpu/drm/renesas/rz-du/rzg3l_lvds_regs.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * RZ/G3L LVDS Interface Registers Definitions
+ *
+ * Copyright (C) 2026 Renesas Electronics Corporation
+ *
+ */
+
+#ifndef __RZG3L_LVDS_REGS_H__
+#define __RZG3L_LVDS_REGS_H__
+
+#define LVDS_CMN			0x00
+#define LVDS_CMN_RST_PHY0_SEL		(1 << 24)
+#define LVDS_CMN_RST_PHY0_SEL_CH0	(1 << 24)
+#define LVDS_CMN_PHY_RESET		(1 << 0)
+
+#define LVDS_0_PHY_OFFSET		0x10
+#define LVDS_0_PHY_CH_IO_EN_MSK		(0x1f)
+#define LVDS_0_PHY_CH_IO_EN		(LVDS_0_PHY_CH_IO_EN_MSK << 0)
+#define LVDS_0_PHY_CH_EN_BGR		BIT(8)
+#define LVDS_0_PHY_CH_EN_LDO		BIT(9)
+
+#define LVDS_0_CTL_OFFSET		0x14
+#define LVDS_0_CTL_FMT_SEL_MSK		GENMASK(23, 20)
+
+#endif /* __RZG3L_LVDS_REGS_H__ */
-- 
2.43.0

Claude review: Add support for Renesas RZ/G3L LVDS encoder

[Claude Code Review Bot]

Overall Series Review

Subject: Add support for Renesas RZ/G3L LVDS encoder
Author: Biju <biju.das.au@gmail.com>
Patches: 3
Reviewed: 2026-05-25T16:56:33.740823

---

This is a 2-patch series adding support for the Renesas RZ/G3L LVDS encoder: a DT binding document (patch 1) and the driver itself (patch 2). The series is well-structured and follows patterns established by the existing `rzg2l_mipi_dsi.c` sibling driver. The code is generally clean and the binding document is complete.

However, there are several issues ranging from minor to moderate:

1. **Unused includes** in the driver (`of_device.h`, `of_graph.h`, several others).
2. **`rst` and `arst` reset controls are acquired but never stored** — they're asserted in probe and then lost, meaning they can never be deasserted on remove or error paths.
3. **`pm_runtime_get_sync` error path leaks the PM reference count** — on failure, `pm_runtime_put` is never called.
4. **Register header style inconsistency** — mixes `BIT()` macro with raw `(1 << n)` shifts.
5. **Cover letter says "v2->v3" but subject says "[PATCH v2]"** — version numbering is inconsistent.

The DT binding (patch 1) looks correct and already has an Acked-by from Conor Dooley.

---

---
Generated by Claude Code Patch Reviewer

Claude review: dt-bindings: display: bridge: Document Renesas RZ/G3L LVDS encoder

[Claude Code Review Bot]

Patch Review

This patch adds the YAML DT binding for `renesas,r9a08g046-lvds`. It is well-written and already carries an Acked-by.

**Minor nits:**

1. **Description typo**: Line 803 says `"This binding describe"` — should be `"This binding describes"`.

2. **Reset ordering**: The binding defines resets as `lvdrst, prst, rst, arst`, which is fine, but the driver acquires `rst` and `arst` as `_optional_exclusive` while `prst` and `lvdrst` are `_shared`. It would be helpful to document in the binding description which resets are shared with the DSI module, since the cover letter text explains this constraint but the binding doesn't.

3. **Example node naming**: The example uses `lvds-cmn@108a0000` as the parent syscon MFD node and `lvds@10` as the child. This looks correct for a syscon child node with relative register offsets.

Overall this patch looks good. No blocking issues.

---

---
Generated by Claude Code Patch Reviewer

Claude review: drm: renesas: rz-du: Add support for RZ/G3L LVDS encoder

[Claude Code Review Bot]

Patch Review

**Issues:**

**(1) `rst` and `arst` reset controls are fire-and-forget (medium severity)**

In `rzg3l_lvds_probe()`, the `rstc` and `arstc` resets are acquired, asserted, and then the local variables go out of scope. They are never stored in `struct rzg3l_lvds`, so they can never be deasserted:

```c
struct reset_control *rstc, *arstc;
...
rstc = devm_reset_control_get_optional_exclusive(dev, "rst");
...
arstc = devm_reset_control_get_optional_exclusive(dev, "arst");
...
ret = reset_control_assert(rstc);
...
ret = reset_control_assert(arstc);
```

Per the cover letter, these resets "must be asserted before using the LVDS module" and should stay asserted. If this is intentional (assert once, never deassert), it works because `devm_*` will release the reset control on driver remove, but a comment would clarify the intent. Additionally, using `_exclusive` reset controls that are only asserted and never deasserted is unusual — if they truly must stay asserted for the lifetime of the device, consider whether `_shared` would be more appropriate if the DSI driver also needs to manipulate them.

**(2) `pm_runtime_get_sync` error path leaks PM refcount (medium severity)**

In `rzg3l_lvds_atomic_enable()`:

```c
if (WARN_ON(pm_runtime_get_sync(lvds->dev) < 0))
    return;
```

`pm_runtime_get_sync()` increments the reference count even on failure. On error, you must call `pm_runtime_put_noidle()` or `pm_runtime_put()` to avoid a PM refcount leak. The matching `pm_runtime_put()` in `rzg3l_lvds_atomic_disable()` will never be called if enable bailed out. Consider:
```c
if (WARN_ON(pm_runtime_get_sync(lvds->dev) < 0)) {
    pm_runtime_put_noidle(lvds->dev);
    return;
}
```
Or alternatively, use `pm_runtime_resume_and_get()` which doesn't increment the refcount on failure.

**(3) Unused includes (minor)**

Several headers appear unnecessary:
- `linux/of_device.h` — nothing from this header is used (no `of_device_get_match_data()` etc.)
- `linux/of_graph.h` — no `of_graph_*` functions are called
- `linux/bitfield.h` — wait, `FIELD_PREP` is used, so this one is needed
- `linux/mfd/syscon.h` — `syscon_node_to_regmap()` is used, so needed
- `linux/regmap.h` — `regmap_*` functions used, needed
- `linux/media-bus-format.h` — `MEDIA_BUS_FMT_*` used, needed
- `drm/drm_panel.h` — no panel functions called directly, but `DRM_PANEL` is selected in Kconfig. This include is not needed in this source file.
- `drm/drm_probe_helper.h` — no probe helper functions called.

**(4) Register header mixes `BIT()` and raw shifts (minor style)**

In `rzg3l_lvds_regs.h`:

```c
#define LVDS_CMN_RST_PHY0_SEL		(1 << 24)
#define LVDS_CMN_RST_PHY0_SEL_CH0	(1 << 24)
#define LVDS_CMN_PHY_RESET		(1 << 0)
```

These should use `BIT()` for consistency with the rest of the file which uses `BIT(8)`, `BIT(9)`, `GENMASK()`:

```c
#define LVDS_0_PHY_CH_EN_BGR		BIT(8)
#define LVDS_0_PHY_CH_EN_LDO		BIT(9)
```

Also, `LVDS_CMN_RST_PHY0_SEL` and `LVDS_CMN_RST_PHY0_SEL_CH0` are defined to the same value `(1 << 24)` — one of them appears redundant.

**(5) `LVDS_0_PHY_CH_IO_EN` is redundant (minor)**

```c
#define LVDS_0_PHY_CH_IO_EN_MSK		(0x1f)
#define LVDS_0_PHY_CH_IO_EN		(LVDS_0_PHY_CH_IO_EN_MSK << 0)
```

`LVDS_0_PHY_CH_IO_EN` is just `LVDS_0_PHY_CH_IO_EN_MSK << 0`, which is the same value. Having both is confusing. This could be simplified, or at least the `<< 0` should be removed.

**(6) `select DRM_PANEL` in Kconfig may be unnecessary (minor)**

The driver doesn't use any panel API directly — it uses `devm_drm_of_get_bridge()` which handles panel wrapping via `drm_panel_bridge`. The `DRM_PANEL` symbol is typically pulled in transitively by the bridge infrastructure. Check whether `DRM_PANEL` is really needed here.

**(7) `select DRM_KMS_HELPER` — check necessity (minor)**

Similar to above, `DRM_KMS_HELPER` is often pulled in by `DRM_BRIDGE`. Verify it's needed as a direct dependency.

**(8) The `return ret;` at the end of probe could be `return 0;` (nitpick)**

```c
ret = devm_drm_bridge_add(dev, &lvds->bridge);
if (ret)
    return dev_err_probe(dev, ret,
                 "Failed to register drm bridge\n");

return ret;
```

At this point `ret` is always 0 (the error case already returned). Using `return 0;` is clearer.

**(9) Version numbering confusion (meta)**

The cover letter changelog says "v2->v3" but the subject lines say `[PATCH v2 0/2]`. This appears to actually be v3 based on the changelog content.

**(10) Bridge attach returns 0 when there's no next_bridge (needs verification)**

```c
static int rzg3l_lvds_attach(...)
{
    struct rzg3l_lvds *lvds = bridge_to_rzg3l_lvds(bridge);

    if (!lvds->bridge.next_bridge)
        return 0;

    return drm_bridge_attach(encoder, lvds->bridge.next_bridge, bridge, flags);
}
```

The `next_bridge` is set from `devm_drm_of_get_bridge()` in probe, which returns an error if not found. So `next_bridge` should always be non-NULL if probe succeeded. This NULL check is harmless defensive code but shouldn't be reachable. Not a bug, but the check is dead code given the current probe logic.

---
Generated by Claude Code Patch Reviewer