[PATCH v9 00/11] Add support for MT8195 SCP 2nd core

[PATCH v9 00/11] Add support for MT8195 SCP 2nd core

[Tinghan Shen]

The mediatek remoteproc driver currently only allows bringing up a 
single core SCP, e.g. MT8183. It also only bringing up the 1st 
core in SoCs with a dual-core SCP, e.g. MT8195. This series support 
to bring-up the 2nd core of the dual-core SCP.

v8 -> v9:
1. initialize l1tcm_size/l1tcm_phys at patchset 05/11 
2. rewrite patchset 06/11 to unify the flow and remove hacks

v7 -> v8:
1. update the node name of mt8192 asurada SCP rpmsg subnode
2. squash register definitions into driver patches
3. initialize local variables on the declaration at patch v8 06/11 

v6 -> v7:
1. merge the mtk_scp_cluster struct into the mtk_scp structure
   at the "Probe multi-core SCP" patch

v5 -> v6:
1. move the mtk_scp_of_regs structure from mtk_common.h to mtk_scp.c
2. rename the SCP core 0 label from 'scp' to 'scp_c0'

v4 -> v5:
1. move resource release actions to the platform driver remove operation 
2. fix dual-core watchdog handling

v3 -> v4:
1. change the representation of dual-core SCP in dts file and update SCP yaml
2. rewrite SCP driver to reflect the change of dts node
3. drop 'remove redundant call of rproc_boot for SCP' in v3 for further investigation

v2 -> v3:
1. change the representation of dual-core SCP in dts file and update SCP yaml
2. rewrite SCP driver to reflect the change of dts node
3. add SCP core 1 node to mt8195.dtsi
4. remove redundant call of rproc_boot for SCP
5. refine IPI error message

v1 -> v2:
1. update dt-binding property description
2. remove kconfig for scp dual driver
3. merge mtk_scp_dual.c and mtk_scp_subdev.c to mtk_scp.c

Tinghan Shen (11):
  dt-bindings: remoteproc: mediatek: Improve the rpmsg subnode
    definition
  arm64: dts: mediatek: Update the node name of SCP rpmsg subnode
  dt-bindings: remoteproc: mediatek: Support MT8195 dual-core SCP
  remoteproc: mediatek: Add MT8195 SCP core 1 operations
  remoteproc: mediatek: Extract remoteproc initialization flow
  remoteproc: mediatek: Probe multi-core SCP
  remoteproc: mediatek: Control SCP core 1 by rproc subdevice
  remoteproc: mediatek: Setup MT8195 SCP core 1 SRAM offset
  remoteproc: mediatek: Handle MT8195 SCP core 1 watchdog timeout
  remoteproc: mediatek: Refine ipi handler error message
  arm64: dts: mediatek: mt8195: Add SCP 2nd core

 .../bindings/remoteproc/mtk,scp.yaml          | 176 ++++++++-
 .../arm64/boot/dts/mediatek/mt8183-kukui.dtsi |   2 +-
 .../boot/dts/mediatek/mt8192-asurada.dtsi     |   2 +-
 .../boot/dts/mediatek/mt8195-cherry.dtsi      |   6 +-
 arch/arm64/boot/dts/mediatek/mt8195.dtsi      |  32 +-
 drivers/remoteproc/mtk_common.h               |  33 ++
 drivers/remoteproc/mtk_scp.c                  | 374 ++++++++++++++++--
 7 files changed, 565 insertions(+), 60 deletions(-)

-- 
2.18.0

[PATCH v9 01/11] dt-bindings: remoteproc: mediatek: Improve the rpmsg subnode definition

[Tinghan Shen]

Improve the definition of the rpmsg subnode by
assigning a distinct node name and adding the
definition source of node properties.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
---
 .../bindings/remoteproc/mtk,scp.yaml          | 31 +++++++++----------
 1 file changed, 15 insertions(+), 16 deletions(-)

diff --git a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
index 895415772d1d..271081df0e46 100644
--- a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
+++ b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
@@ -58,6 +58,18 @@ properties:
   memory-region:
     maxItems: 1
 
+  cros-ec-rpmsg:
+    $ref: /schemas/mfd/google,cros-ec.yaml
+    description:
+      This subnode represents the rpmsg device. The properties
+      of this node are defined by the individual bindings for
+      the rpmsg devices.
+
+    required:
+      - mediatek,rpmsg-name
+
+    unevaluatedProperties: false
+
 required:
   - compatible
   - reg
@@ -89,21 +101,7 @@ allOf:
         reg-names:
           maxItems: 2
 
-additionalProperties:
-  type: object
-  description:
-    Subnodes of the SCP represent rpmsg devices. The names of the devices
-    are not important. The properties of these nodes are defined by the
-    individual bindings for the rpmsg devices.
-  properties:
-    mediatek,rpmsg-name:
-      $ref: /schemas/types.yaml#/definitions/string-array
-      description:
-        Contains the name for the rpmsg device. Used to match
-        the subnode to rpmsg device announced by SCP.
-
-  required:
-    - mediatek,rpmsg-name
+additionalProperties: false
 
 examples:
   - |
@@ -118,7 +116,8 @@ examples:
         clocks = <&infracfg CLK_INFRA_SCPSYS>;
         clock-names = "main";
 
-        cros_ec {
+        cros-ec-rpmsg {
+            compatible = "google,cros-ec-rpmsg";
             mediatek,rpmsg-name = "cros-ec-rpmsg";
         };
     };
-- 
2.18.0

[PATCH v9 02/11] arm64: dts: mediatek: Update the node name of SCP rpmsg subnode

[Tinghan Shen]

Align the node name with the definition in SCP bindings.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi   | 2 +-
 arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi b/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
index fbe14b13051a..a259eb043de5 100644
--- a/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
@@ -810,7 +810,7 @@
 	pinctrl-names = "default";
 	pinctrl-0 = <&scp_pins>;
 
-	cros_ec {
+	cros-ec-rpmsg {
 		compatible = "google,cros-ec-rpmsg";
 		mediatek,rpmsg-name = "cros-ec-rpmsg";
 	};
diff --git a/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi b/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
index 9f12257ab4e7..8f14b633c1e1 100644
--- a/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
@@ -1260,7 +1260,7 @@
 	pinctrl-names = "default";
 	pinctrl-0 = <&scp_pins>;
 
-	cros-ec {
+	cros-ec-rpmsg {
 		compatible = "google,cros-ec-rpmsg";
 		mediatek,rpmsg-name = "cros-ec-rpmsg";
 	};
-- 
2.18.0

[PATCH v9 03/11] dt-bindings: remoteproc: mediatek: Support MT8195 dual-core SCP

[Tinghan Shen]

Extend the SCP binding to describe the MT8195 dual-core SCP.

Under different applications, the MT8195 SCP can be used as single-core
or dual-core. This change keeps the single-core definitions and
adds new definitions for the dual-core use case.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 .../bindings/remoteproc/mtk,scp.yaml          | 145 +++++++++++++++++-
 1 file changed, 141 insertions(+), 4 deletions(-)

diff --git a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
index 271081df0e46..09102dda4942 100644
--- a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
+++ b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
@@ -21,6 +21,7 @@ properties:
       - mediatek,mt8188-scp
       - mediatek,mt8192-scp
       - mediatek,mt8195-scp
+      - mediatek,mt8195-scp-dual
 
   reg:
     description:
@@ -31,10 +32,7 @@ properties:
 
   reg-names:
     minItems: 2
-    items:
-      - const: sram
-      - const: cfg
-      - const: l1tcm
+    maxItems: 3
 
   clocks:
     description:
@@ -70,6 +68,81 @@ properties:
 
     unevaluatedProperties: false
 
+  '#address-cells':
+    const: 1
+
+  '#size-cells':
+    const: 1
+
+  ranges:
+    description:
+      Standard ranges definition providing address translations for
+      local SCP SRAM address spaces to bus addresses.
+
+patternProperties:
+  "^scp@[a-f0-9]+$":
+    type: object
+    description:
+      The MediaTek SCP integrated to SoC might be a multi-core version.
+      The other cores are represented as child nodes of the boot core.
+      There are some integration differences for the IP like the usage of
+      address translator for translating SoC bus addresses into address space
+      for the processor.
+
+      Each SCP core has own cache memory. The SRAM and L1TCM are shared by
+      cores. The power of cache, SRAM and L1TCM power should be enabled
+      before booting SCP cores. The size of cache, SRAM, and L1TCM are varied
+      on differnt SoCs.
+
+      The SCP cores do not use an MMU, but has a set of registers to
+      control the translations between 32-bit CPU addresses into system bus
+      addresses. Cache and memory access settings are provided through a
+      Memory Protection Unit (MPU), programmable only from the SCP.
+
+    properties:
+      compatible:
+        enum:
+          - mediatek,scp-core
+
+      reg:
+        description: The base address and size of SRAM.
+        maxItems: 1
+
+      reg-names:
+        const: sram
+
+      interrupts:
+        maxItems: 1
+
+      firmware-name:
+        $ref: /schemas/types.yaml#/definitions/string
+        description:
+          If present, name (or relative path) of the file within the
+          firmware search path containing the firmware image used when
+          initializing sub cores of multi-core SCP.
+
+      memory-region:
+        maxItems: 1
+
+      cros-ec-rpmsg:
+        $ref: /schemas/mfd/google,cros-ec.yaml
+        description:
+          This subnode represents the rpmsg device. The properties
+          of this node are defined by the individual bindings for
+          the rpmsg devices.
+
+        required:
+          - mediatek,rpmsg-name
+
+        unevaluatedProperties: false
+
+    required:
+      - compatible
+      - reg
+      - reg-names
+
+    additionalProperties: false
+
 required:
   - compatible
   - reg
@@ -99,7 +172,37 @@ allOf:
         reg:
           maxItems: 2
         reg-names:
+          items:
+            - const: sram
+            - const: cfg
+  - if:
+      properties:
+        compatible:
+          enum:
+            - mediatek,mt8192-scp
+            - mediatek,mt8195-scp
+    then:
+      properties:
+        reg:
+          maxItems: 3
+        reg-names:
+          items:
+            - const: sram
+            - const: cfg
+            - const: l1tcm
+  - if:
+      properties:
+        compatible:
+          enum:
+            - mediatek,mt8195-scp-dual
+    then:
+      properties:
+        reg:
           maxItems: 2
+        reg-names:
+          items:
+            - const: cfg
+            - const: l1tcm
 
 additionalProperties: false
 
@@ -121,3 +224,37 @@ examples:
             mediatek,rpmsg-name = "cros-ec-rpmsg";
         };
     };
+
+  - |
+    scp@10500000 {
+        compatible = "mediatek,mt8195-scp-dual";
+        reg = <0x10720000 0xe0000>,
+              <0x10700000 0x8000>;
+        reg-names = "cfg", "l1tcm";
+
+        #address-cells = <1>;
+        #size-cells = <1>;
+        ranges = <0 0x10500000 0x100000>;
+
+        scp@0 {
+            compatible = "mediatek,scp-core";
+            reg = <0x0 0xa0000>;
+            reg-names = "sram";
+
+            cros-ec-rpmsg {
+                compatible = "google,cros-ec-rpmsg";
+                mediatek,rpmsg-name = "cros-ec-rpmsg";
+            };
+        };
+
+        scp@a0000 {
+            compatible = "mediatek,scp-core";
+            reg = <0xa0000 0x20000>;
+            reg-names = "sram";
+
+            cros-ec-rpmsg {
+                compatible = "google,cros-ec-rpmsg";
+                mediatek,rpmsg-name = "cros-ec-rpmsg";
+            };
+        };
+    };
-- 
2.18.0

[PATCH v9 04/11] remoteproc: mediatek: Add MT8195 SCP core 1 operations

[Tinghan Shen]

The SCP rproc driver has a set of chip dependent callbacks for
boot sequence and IRQ handling. Implement these callbacks for MT8195
SCP core 1.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_common.h |  9 ++++++
 drivers/remoteproc/mtk_scp.c    | 56 +++++++++++++++++++++++++++++++++
 2 files changed, 65 insertions(+)

diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
index ea6fa1100a00..c0905aec3b4b 100644
--- a/drivers/remoteproc/mtk_common.h
+++ b/drivers/remoteproc/mtk_common.h
@@ -47,6 +47,7 @@
 #define MT8192_SCP2SPM_IPC_CLR		0x4094
 #define MT8192_GIPC_IN_SET		0x4098
 #define MT8192_HOST_IPC_INT_BIT		BIT(0)
+#define MT8195_CORE1_HOST_IPC_INT_BIT	BIT(4)
 
 #define MT8192_CORE0_SW_RSTN_CLR	0x10000
 #define MT8192_CORE0_SW_RSTN_SET	0x10004
@@ -56,6 +57,14 @@
 
 #define MT8195_L1TCM_SRAM_PDN_RESERVED_RSI_BITS		GENMASK(7, 4)
 
+#define MT8195_CPU1_SRAM_PD			0x1084
+#define MT8195_SSHUB2APMCU_IPC_SET		0x4088
+#define MT8195_SSHUB2APMCU_IPC_CLR		0x408C
+#define MT8195_CORE1_SW_RSTN_CLR		0x20000
+#define MT8195_CORE1_SW_RSTN_SET		0x20004
+#define MT8195_CORE1_MEM_ATT_PREDEF		0x20008
+#define MT8195_CORE1_WDT_CFG			0x20034
+
 #define SCP_FW_VER_LEN			32
 #define SCP_SHARE_BUFFER_SIZE		288
 
diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index 0861b76f185f..a3b9bc158cd9 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -176,6 +176,16 @@ static void mt8192_scp_reset_deassert(struct mtk_scp *scp)
 	writel(1, scp->reg_base + MT8192_CORE0_SW_RSTN_CLR);
 }
 
+static void mt8195_scp_c1_reset_assert(struct mtk_scp *scp)
+{
+	writel(1, scp->reg_base + MT8195_CORE1_SW_RSTN_SET);
+}
+
+static void mt8195_scp_c1_reset_deassert(struct mtk_scp *scp)
+{
+	writel(1, scp->reg_base + MT8195_CORE1_SW_RSTN_CLR);
+}
+
 static void mt8183_scp_irq_handler(struct mtk_scp *scp)
 {
 	u32 scp_to_host;
@@ -212,6 +222,18 @@ static void mt8192_scp_irq_handler(struct mtk_scp *scp)
 	}
 }
 
+static void mt8195_scp_c1_irq_handler(struct mtk_scp *scp)
+{
+	u32 scp_to_host;
+
+	scp_to_host = readl(scp->reg_base + MT8195_SSHUB2APMCU_IPC_SET);
+
+	if (scp_to_host & MT8192_SCP_IPC_INT_BIT)
+		scp_ipi_handler(scp);
+
+	writel(scp_to_host, scp->reg_base + MT8195_SSHUB2APMCU_IPC_CLR);
+}
+
 static irqreturn_t scp_irq_handler(int irq, void *priv)
 {
 	struct mtk_scp *scp = priv;
@@ -453,6 +475,19 @@ static int mt8195_scp_before_load(struct mtk_scp *scp)
 	return 0;
 }
 
+static int mt8195_scp_c1_before_load(struct mtk_scp *scp)
+{
+	scp_sram_power_on(scp->reg_base + MT8195_CPU1_SRAM_PD, 0);
+
+	/* hold SCP in reset while loading FW. */
+	scp->data->scp_reset_assert(scp);
+
+	/* enable MPU for all memory regions */
+	writel(0xff, scp->reg_base + MT8195_CORE1_MEM_ATT_PREDEF);
+
+	return 0;
+}
+
 static int scp_load(struct rproc *rproc, const struct firmware *fw)
 {
 	struct mtk_scp *scp = rproc->priv;
@@ -625,6 +660,15 @@ static void mt8195_scp_stop(struct mtk_scp *scp)
 	writel(0, scp->reg_base + MT8192_CORE0_WDT_CFG);
 }
 
+static void mt8195_scp_c1_stop(struct mtk_scp *scp)
+{
+	/* Power off CPU SRAM */
+	scp_sram_power_off(scp->reg_base + MT8195_CPU1_SRAM_PD, 0);
+
+	/* Disable SCP watchdog */
+	writel(0, scp->reg_base + MT8195_CORE1_WDT_CFG);
+}
+
 static int scp_stop(struct rproc *rproc)
 {
 	struct mtk_scp *scp = (struct mtk_scp *)rproc->priv;
@@ -991,6 +1035,18 @@ static const struct mtk_scp_of_data mt8195_of_data = {
 	.host_to_scp_int_bit = MT8192_HOST_IPC_INT_BIT,
 };
 
+static const struct mtk_scp_of_data mt8195_of_data_c1 = {
+	.scp_clk_get = mt8195_scp_clk_get,
+	.scp_before_load = mt8195_scp_c1_before_load,
+	.scp_irq_handler = mt8195_scp_c1_irq_handler,
+	.scp_reset_assert = mt8195_scp_c1_reset_assert,
+	.scp_reset_deassert = mt8195_scp_c1_reset_deassert,
+	.scp_stop = mt8195_scp_c1_stop,
+	.scp_da_to_va = mt8192_scp_da_to_va,
+	.host_to_scp_reg = MT8192_GIPC_IN_SET,
+	.host_to_scp_int_bit = MT8195_CORE1_HOST_IPC_INT_BIT,
+};
+
 static const struct of_device_id mtk_scp_of_match[] = {
 	{ .compatible = "mediatek,mt8183-scp", .data = &mt8183_of_data },
 	{ .compatible = "mediatek,mt8186-scp", .data = &mt8186_of_data },
-- 
2.18.0

[PATCH v9 05/11] remoteproc: mediatek: Extract remoteproc initialization flow

[Tinghan Shen]

This is the preparation for probing multi-core SCP. The remoteproc
initialization flow is similar on cores and is reused to avoid
redundant code.

The registers of config and l1tcm are shared for multi-core
SCP. Reuse the mapped addresses for all cores.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_scp.c | 64 +++++++++++++++++++++++++-----------
 1 file changed, 45 insertions(+), 19 deletions(-)

diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index a3b9bc158cd9..32ecd1450c6f 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -23,6 +23,13 @@
 #define MAX_CODE_SIZE 0x500000
 #define SECTION_NAME_IPI_BUFFER ".ipi_buffer"
 
+struct mtk_scp_of_regs {
+	void __iomem *reg_base;
+	void __iomem *l1tcm_base;
+	size_t l1tcm_size;
+	phys_addr_t l1tcm_phys;
+};
+
 /**
  * scp_get() - get a reference to SCP.
  *
@@ -855,7 +862,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
 	}
 }
 
-static int scp_probe(struct platform_device *pdev)
+static int scp_rproc_init(struct platform_device *pdev,
+			  struct mtk_scp_of_regs *of_regs)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *np = dev->of_node;
@@ -879,6 +887,11 @@ static int scp_probe(struct platform_device *pdev)
 	scp->data = of_device_get_match_data(dev);
 	platform_set_drvdata(pdev, scp);
 
+	scp->reg_base = of_regs->reg_base;
+	scp->l1tcm_base = of_regs->l1tcm_base;
+	scp->l1tcm_size = of_regs->l1tcm_size;
+	scp->l1tcm_phys = of_regs->l1tcm_phys;
+
 	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram");
 	scp->sram_base = devm_ioremap_resource(dev, res);
 	if (IS_ERR(scp->sram_base))
@@ -888,24 +901,6 @@ static int scp_probe(struct platform_device *pdev)
 	scp->sram_size = resource_size(res);
 	scp->sram_phys = res->start;
 
-	/* l1tcm is an optional memory region */
-	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
-	scp->l1tcm_base = devm_ioremap_resource(dev, res);
-	if (IS_ERR(scp->l1tcm_base)) {
-		ret = PTR_ERR(scp->l1tcm_base);
-		if (ret != -EINVAL) {
-			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
-		}
-	} else {
-		scp->l1tcm_size = resource_size(res);
-		scp->l1tcm_phys = res->start;
-	}
-
-	scp->reg_base = devm_platform_ioremap_resource_byname(pdev, "cfg");
-	if (IS_ERR(scp->reg_base))
-		return dev_err_probe(dev, PTR_ERR(scp->reg_base),
-				     "Failed to parse and map cfg memory\n");
-
 	ret = scp->data->scp_clk_get(scp);
 	if (ret)
 		return ret;
@@ -957,6 +952,37 @@ static int scp_probe(struct platform_device *pdev)
 	return ret;
 }
 
+static int scp_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mtk_scp_of_regs scp_regs;
+	struct resource *res;
+	int ret;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg");
+	scp_regs.reg_base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(scp_regs.reg_base))
+		return dev_err_probe(dev, PTR_ERR(scp_regs.reg_base),
+				     "Failed to parse and map cfg memory\n");
+
+	/* l1tcm is an optional memory region */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
+	scp_regs.l1tcm_base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(scp_regs.l1tcm_base)) {
+		ret = PTR_ERR(scp_regs.l1tcm_base);
+		if (ret != -EINVAL)
+			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
+
+		scp_regs.l1tcm_size = 0;
+		scp_regs.l1tcm_phys = 0;
+	} else {
+		scp_regs.l1tcm_size = resource_size(res);
+		scp_regs.l1tcm_phys = res->start;
+	}
+
+	return scp_rproc_init(pdev, &scp_regs);
+}
+
 static int scp_remove(struct platform_device *pdev)
 {
 	struct mtk_scp *scp = platform_get_drvdata(pdev);
-- 
2.18.0

[PATCH v9 06/11] remoteproc: mediatek: Probe multi-core SCP

[Tinghan Shen]

The difference of single-core SCP and multi-core SCP device tree is
the presence of child device nodes described SCP cores. The SCP
driver populates the platform device and checks the child nodes
to identify whether it's a single-core SCP or a multi-core SCP.

Add the remoteproc instances for single-core SCP and multi-core SCP to
the new added SCP cluster list. When the SCP driver is removed, it
cleanup resources by walking through the cluster list.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
---
 drivers/remoteproc/mtk_common.h |   3 +
 drivers/remoteproc/mtk_scp.c    | 150 ++++++++++++++++++++++++++++----
 2 files changed, 137 insertions(+), 16 deletions(-)

diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
index c0905aec3b4b..b73b60c22ea1 100644
--- a/drivers/remoteproc/mtk_common.h
+++ b/drivers/remoteproc/mtk_common.h
@@ -128,6 +128,9 @@ struct mtk_scp {
 	size_t dram_size;
 
 	struct rproc_subdev *rpmsg_subdev;
+
+	struct list_head elem;
+	struct list_head *cluster;
 };
 
 /**
diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index 32ecd1450c6f..71882ed31e16 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -30,6 +30,8 @@ struct mtk_scp_of_regs {
 	phys_addr_t l1tcm_phys;
 };
 
+static struct list_head cluster_cores;
+
 /**
  * scp_get() - get a reference to SCP.
  *
@@ -863,7 +865,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
 }
 
 static int scp_rproc_init(struct platform_device *pdev,
-			  struct mtk_scp_of_regs *of_regs)
+			  struct mtk_scp_of_regs *of_regs,
+			  const struct mtk_scp_of_data *of_data)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *np = dev->of_node;
@@ -884,7 +887,7 @@ static int scp_rproc_init(struct platform_device *pdev,
 	scp = (struct mtk_scp *)rproc->priv;
 	scp->rproc = rproc;
 	scp->dev = dev;
-	scp->data = of_device_get_match_data(dev);
+	scp->data = of_data;
 	platform_set_drvdata(pdev, scp);
 
 	scp->reg_base = of_regs->reg_base;
@@ -934,10 +937,6 @@ static int scp_rproc_init(struct platform_device *pdev,
 		goto remove_subdev;
 	}
 
-	ret = rproc_add(rproc);
-	if (ret)
-		goto remove_subdev;
-
 	return 0;
 
 remove_subdev:
@@ -952,6 +951,109 @@ static int scp_rproc_init(struct platform_device *pdev,
 	return ret;
 }
 
+static void scp_rproc_free(struct mtk_scp *scp)
+{
+	int i;
+
+	scp_remove_rpmsg_subdev(scp);
+	scp_ipi_unregister(scp, SCP_IPI_INIT);
+	scp_unmap_memory_region(scp);
+	for (i = 0; i < SCP_IPI_MAX; i++)
+		mutex_destroy(&scp->ipi_desc[i].lock);
+	mutex_destroy(&scp->send_lock);
+}
+
+static int scp_cluster_init(struct platform_device *pdev,
+			    struct mtk_scp_of_regs *of_regs)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev_of_node(dev);
+	struct platform_device *cpdev;
+	struct device_node *child;
+	const struct mtk_scp_of_data **cluster_of_data;
+	struct list_head *cluster = &cluster_cores;
+	struct mtk_scp *scp, *temp;
+	int core_id = 0, num_cores;
+	int ret;
+
+	child = of_get_next_available_child(np, NULL);
+	if (!child)
+		return dev_err_probe(dev, -ENODEV, "No child node\n");
+
+	of_node_put(child);
+	if (of_node_name_eq(child, "cros-ec-rpmsg")) {
+		dev_dbg(dev, "single-core scp\n");
+
+		ret = scp_rproc_init(pdev, of_regs, of_device_get_match_data(dev));
+		if (ret)
+			return dev_err_probe(dev, ret, "Failed to initialize single-core scp\n");
+
+		/* the single-core drvdata is intended to be the mtk_scp object.
+		 * the mtk_scp object is used by APIs in mtk_scp_rpmsg_info.
+		 */
+		scp = platform_get_drvdata(pdev);
+		list_add_tail(&scp->elem, cluster);
+		scp->cluster = cluster;
+	} else {
+		dev_dbg(dev, "multi-core scp\n");
+
+		cluster_of_data = (const struct mtk_scp_of_data **)of_device_get_match_data(dev);
+
+		num_cores = of_get_available_child_count(np);
+
+		for_each_available_child_of_node(np, child) {
+			if (core_id >= num_cores || !cluster_of_data[core_id]) {
+				ret = -EINVAL;
+				dev_err(dev, "Not support core %d\n", core_id);
+				of_node_put(child);
+				goto init_fail;
+			}
+
+			cpdev = of_find_device_by_node(child);
+			if (!cpdev) {
+				ret = -ENODEV;
+				dev_err(dev, "Not found platform device for core %d\n", core_id);
+				of_node_put(child);
+				goto init_fail;
+			}
+
+			ret = scp_rproc_init(cpdev, of_regs, cluster_of_data[core_id]);
+			if (ret) {
+				dev_err(dev, "Failed to initialize core %d rproc\n", core_id);
+				put_device(&cpdev->dev);
+				of_node_put(child);
+				goto init_fail;
+			}
+			scp = platform_get_drvdata(cpdev);
+			list_add_tail(&scp->elem, cluster);
+			scp->cluster = cluster;
+			put_device(&cpdev->dev);
+
+			core_id++;
+		}
+	}
+
+	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
+		ret = rproc_add(scp->rproc);
+		if (ret)
+			goto add_fail;
+	}
+
+	return 0;
+
+add_fail:
+	list_for_each_entry_continue(scp, cluster, elem) {
+		rproc_del(scp->rproc);
+	}
+init_fail:
+	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
+		list_del(&scp->elem);
+		scp_rproc_free(scp);
+	}
+
+	return ret;
+}
+
 static int scp_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
@@ -980,21 +1082,30 @@ static int scp_probe(struct platform_device *pdev)
 		scp_regs.l1tcm_phys = res->start;
 	}
 
-	return scp_rproc_init(pdev, &scp_regs);
+	ret = devm_of_platform_populate(dev);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to populate platform devices\n");
+
+	INIT_LIST_HEAD(&cluster_cores);
+	platform_set_drvdata(pdev, &cluster_cores);
+
+	ret = scp_cluster_init(pdev, &scp_regs);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to initialize scp cluster\n");
+
+	return 0;
 }
 
 static int scp_remove(struct platform_device *pdev)
 {
-	struct mtk_scp *scp = platform_get_drvdata(pdev);
-	int i;
+	struct list_head *cluster = &cluster_cores;
+	struct mtk_scp *scp, *temp;
 
-	rproc_del(scp->rproc);
-	scp_remove_rpmsg_subdev(scp);
-	scp_ipi_unregister(scp, SCP_IPI_INIT);
-	scp_unmap_memory_region(scp);
-	for (i = 0; i < SCP_IPI_MAX; i++)
-		mutex_destroy(&scp->ipi_desc[i].lock);
-	mutex_destroy(&scp->send_lock);
+	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
+		list_del(&scp->elem);
+		rproc_del(scp->rproc);
+		scp_rproc_free(scp);
+	}
 
 	return 0;
 }
@@ -1073,12 +1184,19 @@ static const struct mtk_scp_of_data mt8195_of_data_c1 = {
 	.host_to_scp_int_bit = MT8195_CORE1_HOST_IPC_INT_BIT,
 };
 
+static const struct mtk_scp_of_data *mt8195_of_data_cores[] = {
+	&mt8195_of_data,
+	&mt8195_of_data_c1,
+	NULL
+};
+
 static const struct of_device_id mtk_scp_of_match[] = {
 	{ .compatible = "mediatek,mt8183-scp", .data = &mt8183_of_data },
 	{ .compatible = "mediatek,mt8186-scp", .data = &mt8186_of_data },
 	{ .compatible = "mediatek,mt8188-scp", .data = &mt8188_of_data },
 	{ .compatible = "mediatek,mt8192-scp", .data = &mt8192_of_data },
 	{ .compatible = "mediatek,mt8195-scp", .data = &mt8195_of_data },
+	{ .compatible = "mediatek,mt8195-scp-dual", .data = &mt8195_of_data_cores },
 	{},
 };
 MODULE_DEVICE_TABLE(of, mtk_scp_of_match);
-- 
2.18.0

[PATCH v9 07/11] remoteproc: mediatek: Control SCP core 1 by rproc subdevice

[Tinghan Shen]

Register SCP core 1 as a subdevice of core 0 for the boot sequence
and watchdog timeout handling. The core 1 has to boot after core 0
because the SCP clock and SRAM power is controlled by SCP core 0.
As for watchdog timeout handling, the remoteproc framework helps to
stop/start subdevices automatically when SCP driver receives watchdog
timeout event.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_common.h |  9 +++++
 drivers/remoteproc/mtk_scp.c    | 66 +++++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+)

diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
index b73b60c22ea1..f1a021ad7f66 100644
--- a/drivers/remoteproc/mtk_common.h
+++ b/drivers/remoteproc/mtk_common.h
@@ -100,6 +100,13 @@ struct mtk_scp_of_data {
 	size_t ipi_buf_offset;
 };
 
+struct mtk_scp_core_subdev {
+	struct rproc_subdev subdev;
+	struct mtk_scp *scp;
+};
+
+#define to_core_subdev(d) container_of(d, struct mtk_scp_core_subdev, subdev)
+
 struct mtk_scp {
 	struct device *dev;
 	struct rproc *rproc;
@@ -131,6 +138,8 @@ struct mtk_scp {
 
 	struct list_head elem;
 	struct list_head *cluster;
+
+	struct mtk_scp_core_subdev *core_subdev;
 };
 
 /**
diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index 71882ed31e16..ba931d550c1e 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -864,6 +864,60 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
 	}
 }
 
+static int scp_core_subdev_start(struct rproc_subdev *subdev)
+{
+	struct mtk_scp_core_subdev *core_subdev = to_core_subdev(subdev);
+	struct mtk_scp *scp = core_subdev->scp;
+
+	rproc_boot(scp->rproc);
+
+	return 0;
+}
+
+static void scp_core_subdev_stop(struct rproc_subdev *subdev, bool crashed)
+{
+	struct mtk_scp_core_subdev *core_subdev = to_core_subdev(subdev);
+	struct mtk_scp *scp = core_subdev->scp;
+
+	rproc_shutdown(scp->rproc);
+}
+
+static int scp_core_subdev_register(struct mtk_scp *scp)
+{
+	struct device *dev = scp->dev;
+	struct mtk_scp_core_subdev *core_subdev;
+	struct mtk_scp *scp_c0;
+
+	scp_c0 = list_first_entry(scp->cluster, struct mtk_scp, elem);
+	if (!scp_c0)
+		return -ENODATA;
+
+	core_subdev = devm_kzalloc(dev, sizeof(*core_subdev), GFP_KERNEL);
+	if (!core_subdev)
+		return -ENOMEM;
+
+	core_subdev->scp = scp;
+	core_subdev->subdev.start = scp_core_subdev_start;
+	core_subdev->subdev.stop = scp_core_subdev_stop;
+
+	scp->core_subdev = core_subdev;
+	rproc_add_subdev(scp_c0->rproc, &scp->core_subdev->subdev);
+
+	return 0;
+}
+
+static void scp_core_subdev_unregister(struct mtk_scp *scp)
+{
+	struct mtk_scp *scp_c0;
+
+	if (scp->core_subdev) {
+		scp_c0 = list_first_entry(scp->cluster, struct mtk_scp, elem);
+		rproc_remove_subdev(scp_c0->rproc, &scp->core_subdev->subdev);
+		devm_kfree(scp->dev, scp->core_subdev);
+		scp->core_subdev = NULL;
+	}
+}
+
 static int scp_rproc_init(struct platform_device *pdev,
 			  struct mtk_scp_of_regs *of_regs,
 			  const struct mtk_scp_of_data *of_data)
@@ -955,6 +1009,7 @@ static void scp_rproc_free(struct mtk_scp *scp)
 {
 	int i;
 
+	scp_core_subdev_unregister(scp);
 	scp_remove_rpmsg_subdev(scp);
 	scp_ipi_unregister(scp, SCP_IPI_INIT);
 	scp_unmap_memory_region(scp);
@@ -1034,6 +1089,17 @@ static int scp_cluster_init(struct platform_device *pdev,
 	}
 
 	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
+		if (!list_is_first(&scp->elem, cluster)) {
+			ret = scp_core_subdev_register(scp);
+			if (ret) {
+				dev_err_probe(scp->dev, ret, "Failed to register as subdev\n");
+				goto add_fail;
+			}
+
+			/* sub cores are booted as subdevices of core 0 */
+			scp->rproc->auto_boot = false;
+		}
+
 		ret = rproc_add(scp->rproc);
 		if (ret)
 			goto add_fail;
-- 
2.18.0

[PATCH v9 08/11] remoteproc: mediatek: Setup MT8195 SCP core 1 SRAM offset

[Tinghan Shen]

Because MT8195 SCP core 0 and core 1 both boot from head of SRAM and
have the same viewpoint of SRAM, SCP has a "core 1 SRAM offset"
configuration to control the access destination of SCP core 1 to boot
core 1 from different SRAM location.

The "core 1 SRAM offset" configuration is composed by a range
and an offset. It works like a simple memory mapped mechanism.
When SCP core 1 accesses a SRAM address located in the range,
the SCP bus adds the configured offset to the address to
shift the physical destination address on SCP SRAM. This shifting is
transparent to the software running on SCP core 1.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_common.h |  7 +++++++
 drivers/remoteproc/mtk_scp.c    | 27 +++++++++++++++++++++++++++
 2 files changed, 34 insertions(+)

diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
index f1a021ad7f66..6861543a0825 100644
--- a/drivers/remoteproc/mtk_common.h
+++ b/drivers/remoteproc/mtk_common.h
@@ -65,6 +65,13 @@
 #define MT8195_CORE1_MEM_ATT_PREDEF		0x20008
 #define MT8195_CORE1_WDT_CFG			0x20034
 
+#define MT8195_SEC_CTRL				0x85000
+#define MT8195_CORE_OFFSET_ENABLE_D		BIT(13)
+#define MT8195_CORE_OFFSET_ENABLE_I		BIT(12)
+#define MT8195_L2TCM_OFFSET_RANGE_0_LOW		0x850b0
+#define MT8195_L2TCM_OFFSET_RANGE_0_HIGH	0x850b4
+#define MT8195_L2TCM_OFFSET			0x850d0
+
 #define SCP_FW_VER_LEN			32
 #define SCP_SHARE_BUFFER_SIZE		288
 
diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index ba931d550c1e..f74ee80838e9 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -486,6 +486,9 @@ static int mt8195_scp_before_load(struct mtk_scp *scp)
 
 static int mt8195_scp_c1_before_load(struct mtk_scp *scp)
 {
+	u32 sec_ctrl;
+	struct mtk_scp *scp_c0;
+
 	scp_sram_power_on(scp->reg_base + MT8195_CPU1_SRAM_PD, 0);
 
 	/* hold SCP in reset while loading FW. */
@@ -494,6 +497,30 @@ static int mt8195_scp_c1_before_load(struct mtk_scp *scp)
 	/* enable MPU for all memory regions */
 	writel(0xff, scp->reg_base + MT8195_CORE1_MEM_ATT_PREDEF);
 
+	/*
+	 * The L2TCM_OFFSET_RANGE and L2TCM_OFFSET shift the destination address
+	 * on SRAM when SCP core 1 accesses SRAM.
+	 *
+	 * This configuration solves booting the SCP core 0 and core 1 from
+	 * different SRAM address because core 0 and core 1 both boot from
+	 * the head of SRAM by default. this must be configured before boot SCP core 1.
+	 *
+	 * The value of L2TCM_OFFSET_RANGE is from the viewpoint of SCP core 1.
+	 * When SCP core 1 issues address within the range (L2TCM_OFFSET_RANGE),
+	 * the address will be added with a fixed offset (L2TCM_OFFSET) on the bus.
+	 * The shift action is tranparent to software.
+	 */
+	writel(0, scp->reg_base + MT8195_L2TCM_OFFSET_RANGE_0_LOW);
+	writel(scp->sram_size, scp->reg_base + MT8195_L2TCM_OFFSET_RANGE_0_HIGH);
+
+	scp_c0 = list_first_entry(scp->cluster, struct mtk_scp, elem);
+	writel(scp->sram_phys - scp_c0->sram_phys, scp->reg_base + MT8195_L2TCM_OFFSET);
+
+	/* enable SRAM offset when fetching instruction and data */
+	sec_ctrl = readl(scp->reg_base + MT8195_SEC_CTRL);
+	sec_ctrl |= MT8195_CORE_OFFSET_ENABLE_I | MT8195_CORE_OFFSET_ENABLE_D;
+	writel(sec_ctrl, scp->reg_base + MT8195_SEC_CTRL);
+
 	return 0;
 }
 
-- 
2.18.0

[PATCH v9 09/11] remoteproc: mediatek: Handle MT8195 SCP core 1 watchdog timeout

[Tinghan Shen]

The MT8195 SCP core 1 watchdog timeout needs to be handled in the
SCP core 0 IRQ handler because the MT8195 SCP core 1 watchdog timeout
IRQ is wired on the same IRQ entry for core 0 watchdog timeout.
MT8195 SCP has a watchdog status register to identify the watchdog
timeout source when IRQ triggered.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_common.h |  5 +++++
 drivers/remoteproc/mtk_scp.c    | 25 ++++++++++++++++++++++++-
 2 files changed, 29 insertions(+), 1 deletion(-)

diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
index 6861543a0825..1e148cbad0fb 100644
--- a/drivers/remoteproc/mtk_common.h
+++ b/drivers/remoteproc/mtk_common.h
@@ -55,6 +55,10 @@
 #define MT8192_CORE0_WDT_IRQ		0x10030
 #define MT8192_CORE0_WDT_CFG		0x10034
 
+#define MT8195_SYS_STATUS		0x4004
+#define MT8195_CORE0_WDT		BIT(16)
+#define MT8195_CORE1_WDT		BIT(17)
+
 #define MT8195_L1TCM_SRAM_PDN_RESERVED_RSI_BITS		GENMASK(7, 4)
 
 #define MT8195_CPU1_SRAM_PD			0x1084
@@ -63,6 +67,7 @@
 #define MT8195_CORE1_SW_RSTN_CLR		0x20000
 #define MT8195_CORE1_SW_RSTN_SET		0x20004
 #define MT8195_CORE1_MEM_ATT_PREDEF		0x20008
+#define MT8195_CORE1_WDT_IRQ			0x20030
 #define MT8195_CORE1_WDT_CFG			0x20034
 
 #define MT8195_SEC_CTRL				0x85000
diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index f74ee80838e9..d08f21c1f947 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -231,6 +231,29 @@ static void mt8192_scp_irq_handler(struct mtk_scp *scp)
 	}
 }
 
+static void mt8195_scp_irq_handler(struct mtk_scp *scp)
+{
+	u32 scp_to_host;
+
+	scp_to_host = readl(scp->reg_base + MT8192_SCP2APMCU_IPC_SET);
+
+	if (scp_to_host & MT8192_SCP_IPC_INT_BIT) {
+		scp_ipi_handler(scp);
+	} else {
+		u32 reason = readl(scp->reg_base + MT8195_SYS_STATUS);
+
+		if (reason & MT8195_CORE0_WDT)
+			writel(1, scp->reg_base + MT8192_CORE0_WDT_IRQ);
+
+		if (reason & MT8195_CORE1_WDT)
+			writel(1, scp->reg_base + MT8195_CORE1_WDT_IRQ);
+
+		scp_wdt_handler(scp, reason);
+	}
+
+	writel(scp_to_host, scp->reg_base + MT8192_SCP2APMCU_IPC_CLR);
+}
+
 static void mt8195_scp_c1_irq_handler(struct mtk_scp *scp)
 {
 	u32 scp_to_host;
@@ -1256,7 +1279,7 @@ static const struct mtk_scp_of_data mt8192_of_data = {
 static const struct mtk_scp_of_data mt8195_of_data = {
 	.scp_clk_get = mt8195_scp_clk_get,
 	.scp_before_load = mt8195_scp_before_load,
-	.scp_irq_handler = mt8192_scp_irq_handler,
+	.scp_irq_handler = mt8195_scp_irq_handler,
 	.scp_reset_assert = mt8192_scp_reset_assert,
 	.scp_reset_deassert = mt8192_scp_reset_deassert,
 	.scp_stop = mt8195_scp_stop,
-- 
2.18.0

[PATCH v9 10/11] remoteproc: mediatek: Refine ipi handler error message

[Tinghan Shen]

The error message doesn't accurately reflect the cause of
the error. The error is due to a handler not being found,
not an invalid IPI ID.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 drivers/remoteproc/mtk_scp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
index d08f21c1f947..fec74631557a 100644
--- a/drivers/remoteproc/mtk_scp.c
+++ b/drivers/remoteproc/mtk_scp.c
@@ -115,7 +115,7 @@ static void scp_ipi_handler(struct mtk_scp *scp)
 	scp_ipi_lock(scp, id);
 	handler = ipi_desc[id].handler;
 	if (!handler) {
-		dev_err(scp->dev, "No such ipi id = %d\n", id);
+		dev_err(scp->dev, "No handler for ipi id = %d\n", id);
 		scp_ipi_unlock(scp, id);
 		return;
 	}
-- 
2.18.0

[PATCH v9 11/11] arm64: dts: mediatek: mt8195: Add SCP 2nd core

[Tinghan Shen]

Rewrite the MT8195 SCP device node as a cluster and
add the SCP 2nd core in it.

Since the SCP device node is changed to multi-core structure,
enable SCP cluster to enable probing SCP core 0.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
---
 .../boot/dts/mediatek/mt8195-cherry.dtsi      |  6 +++-
 arch/arm64/boot/dts/mediatek/mt8195.dtsi      | 32 ++++++++++++++-----
 2 files changed, 29 insertions(+), 9 deletions(-)

diff --git a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
index 56749cfe7c33..31415d71b6a4 100644
--- a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
@@ -933,7 +933,11 @@
 	interrupts-extended = <&pio 222 IRQ_TYPE_LEVEL_HIGH>;
 };
 
-&scp {
+&scp_cluster {
+	status = "okay";
+};
+
+&scp_c0 {
 	status = "okay";
 
 	firmware-name = "mediatek/mt8195/scp.img";
diff --git a/arch/arm64/boot/dts/mediatek/mt8195.dtsi b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
index 8fc527570791..5fe5fb32261e 100644
--- a/arch/arm64/boot/dts/mediatek/mt8195.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
@@ -826,14 +826,30 @@
 			clocks = <&infracfg_ao CLK_INFRA_AO_GCE2>;
 		};
 
-		scp: scp@10500000 {
-			compatible = "mediatek,mt8195-scp";
-			reg = <0 0x10500000 0 0x100000>,
-			      <0 0x10720000 0 0xe0000>,
-			      <0 0x10700000 0 0x8000>;
-			reg-names = "sram", "cfg", "l1tcm";
-			interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
+		scp_cluster: scp@10500000 {
+			compatible = "mediatek,mt8195-scp-dual";
+			reg = <0 0x10720000 0 0xe0000>, <0 0x10700000 0 0x8000>;
+			reg-names = "cfg", "l1tcm";
+			#address-cells = <1>;
+			#size-cells = <1>;
+			ranges = <0 0 0x10500000 0x100000>;
 			status = "disabled";
+
+			scp_c0: scp@0 {
+				compatible = "mediatek,scp-core";
+				reg = <0x0 0xa0000>;
+				reg-names = "sram";
+				interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
+				status = "disabled";
+			};
+
+			scp_c1: scp@a0000 {
+				compatible = "mediatek,scp-core";
+				reg = <0xa0000 0x20000>;
+				reg-names = "sram";
+				interrupts = <GIC_SPI 463 IRQ_TYPE_LEVEL_HIGH 0>;
+				status = "disabled";
+			};
 		};
 
 		scp_adsp: clock-controller@10720000 {
@@ -2309,7 +2325,7 @@
 				 <&iommu_vdo M4U_PORT_L19_VENC_REF_LUMA>,
 				 <&iommu_vdo M4U_PORT_L19_VENC_REF_CHROMA>;
 			interrupts = <GIC_SPI 341 IRQ_TYPE_LEVEL_HIGH 0>;
-			mediatek,scp = <&scp>;
+			mediatek,scp = <&scp_c0>;
 			clocks = <&vencsys CLK_VENC_VENC>;
 			clock-names = "venc_sel";
 			assigned-clocks = <&topckgen CLK_TOP_VENC>;
-- 
2.18.0

Re: [PATCH v9 02/11] arm64: dts: mediatek: Update the node name of SCP rpmsg subnode

[Matthias Brugger]

On 28/03/2023 04:27, Tinghan Shen wrote:
> Align the node name with the definition in SCP bindings.
> 
> Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Acked-by: Matthias Brugger <matthias.bgg@gmail.com>

> ---
>   arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi   | 2 +-
>   arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi | 2 +-
>   2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi b/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
> index fbe14b13051a..a259eb043de5 100644
> --- a/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
> +++ b/arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi
> @@ -810,7 +810,7 @@
>   	pinctrl-names = "default";
>   	pinctrl-0 = <&scp_pins>;
>   
> -	cros_ec {
> +	cros-ec-rpmsg {
>   		compatible = "google,cros-ec-rpmsg";
>   		mediatek,rpmsg-name = "cros-ec-rpmsg";
>   	};
> diff --git a/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi b/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
> index 9f12257ab4e7..8f14b633c1e1 100644
> --- a/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
> +++ b/arch/arm64/boot/dts/mediatek/mt8192-asurada.dtsi
> @@ -1260,7 +1260,7 @@
>   	pinctrl-names = "default";
>   	pinctrl-0 = <&scp_pins>;
>   
> -	cros-ec {
> +	cros-ec-rpmsg {
>   		compatible = "google,cros-ec-rpmsg";
>   		mediatek,rpmsg-name = "cros-ec-rpmsg";
>   	};

Re: [PATCH v9 11/11] arm64: dts: mediatek: mt8195: Add SCP 2nd core

[Matthias Brugger]

On 28/03/2023 04:27, Tinghan Shen wrote:
> Rewrite the MT8195 SCP device node as a cluster and
> add the SCP 2nd core in it.
> 
> Since the SCP device node is changed to multi-core structure,
> enable SCP cluster to enable probing SCP core 0.
> 
> Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

As this is a bigger change I'd prefer to take it through my tree once the driver 
and dt-bindings changes are merged. Given the fact, maybe it would make sense to 
take 2/11 through my tree as well.

Regards,
Matthias

> ---
>   .../boot/dts/mediatek/mt8195-cherry.dtsi      |  6 +++-
>   arch/arm64/boot/dts/mediatek/mt8195.dtsi      | 32 ++++++++++++++-----
>   2 files changed, 29 insertions(+), 9 deletions(-)
> 
> diff --git a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> index 56749cfe7c33..31415d71b6a4 100644
> --- a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> +++ b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> @@ -933,7 +933,11 @@
>   	interrupts-extended = <&pio 222 IRQ_TYPE_LEVEL_HIGH>;
>   };
>   
> -&scp {
> +&scp_cluster {
> +	status = "okay";
> +};
> +
> +&scp_c0 {
>   	status = "okay";
>   
>   	firmware-name = "mediatek/mt8195/scp.img";
> diff --git a/arch/arm64/boot/dts/mediatek/mt8195.dtsi b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> index 8fc527570791..5fe5fb32261e 100644
> --- a/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> +++ b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> @@ -826,14 +826,30 @@
>   			clocks = <&infracfg_ao CLK_INFRA_AO_GCE2>;
>   		};
>   
> -		scp: scp@10500000 {
> -			compatible = "mediatek,mt8195-scp";
> -			reg = <0 0x10500000 0 0x100000>,
> -			      <0 0x10720000 0 0xe0000>,
> -			      <0 0x10700000 0 0x8000>;
> -			reg-names = "sram", "cfg", "l1tcm";
> -			interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
> +		scp_cluster: scp@10500000 {
> +			compatible = "mediatek,mt8195-scp-dual";
> +			reg = <0 0x10720000 0 0xe0000>, <0 0x10700000 0 0x8000>;
> +			reg-names = "cfg", "l1tcm";
> +			#address-cells = <1>;
> +			#size-cells = <1>;
> +			ranges = <0 0 0x10500000 0x100000>;
>   			status = "disabled";
> +
> +			scp_c0: scp@0 {
> +				compatible = "mediatek,scp-core";
> +				reg = <0x0 0xa0000>;
> +				reg-names = "sram";
> +				interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
> +				status = "disabled";
> +			};
> +
> +			scp_c1: scp@a0000 {
> +				compatible = "mediatek,scp-core";
> +				reg = <0xa0000 0x20000>;
> +				reg-names = "sram";
> +				interrupts = <GIC_SPI 463 IRQ_TYPE_LEVEL_HIGH 0>;
> +				status = "disabled";
> +			};
>   		};
>   
>   		scp_adsp: clock-controller@10720000 {
> @@ -2309,7 +2325,7 @@
>   				 <&iommu_vdo M4U_PORT_L19_VENC_REF_LUMA>,
>   				 <&iommu_vdo M4U_PORT_L19_VENC_REF_CHROMA>;
>   			interrupts = <GIC_SPI 341 IRQ_TYPE_LEVEL_HIGH 0>;
> -			mediatek,scp = <&scp>;
> +			mediatek,scp = <&scp_c0>;
>   			clocks = <&vencsys CLK_VENC_VENC>;
>   			clock-names = "venc_sel";
>   			assigned-clocks = <&topckgen CLK_TOP_VENC>;

Re: [PATCH v9 05/11] remoteproc: mediatek: Extract remoteproc initialization flow

[Mathieu Poirier]

On Tue, Mar 28, 2023 at 10:27:27AM +0800, Tinghan Shen wrote:
> This is the preparation for probing multi-core SCP. The remoteproc
> initialization flow is similar on cores and is reused to avoid
> redundant code.
> 
> The registers of config and l1tcm are shared for multi-core
> SCP. Reuse the mapped addresses for all cores.
> 
> Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> ---
>  drivers/remoteproc/mtk_scp.c | 64 +++++++++++++++++++++++++-----------
>  1 file changed, 45 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
> index a3b9bc158cd9..32ecd1450c6f 100644
> --- a/drivers/remoteproc/mtk_scp.c
> +++ b/drivers/remoteproc/mtk_scp.c
> @@ -23,6 +23,13 @@
>  #define MAX_CODE_SIZE 0x500000
>  #define SECTION_NAME_IPI_BUFFER ".ipi_buffer"
>  
> +struct mtk_scp_of_regs {
> +	void __iomem *reg_base;
> +	void __iomem *l1tcm_base;
> +	size_t l1tcm_size;
> +	phys_addr_t l1tcm_phys;
> +};
> +

This should represent the cluster with a list of mtk_scp instead of @cluster_cores as
introduced in the next patch.

>  /**
>   * scp_get() - get a reference to SCP.
>   *
> @@ -855,7 +862,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
>  	}
>  }
>  
> -static int scp_probe(struct platform_device *pdev)
> +static int scp_rproc_init(struct platform_device *pdev,
> +			  struct mtk_scp_of_regs *of_regs)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> @@ -879,6 +887,11 @@ static int scp_probe(struct platform_device *pdev)
>  	scp->data = of_device_get_match_data(dev);
>  	platform_set_drvdata(pdev, scp);
>  
> +	scp->reg_base = of_regs->reg_base;
> +	scp->l1tcm_base = of_regs->l1tcm_base;
> +	scp->l1tcm_size = of_regs->l1tcm_size;
> +	scp->l1tcm_phys = of_regs->l1tcm_phys;
> +
>  	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram");
>  	scp->sram_base = devm_ioremap_resource(dev, res);
>  	if (IS_ERR(scp->sram_base))
> @@ -888,24 +901,6 @@ static int scp_probe(struct platform_device *pdev)
>  	scp->sram_size = resource_size(res);
>  	scp->sram_phys = res->start;
>  
> -	/* l1tcm is an optional memory region */
> -	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> -	scp->l1tcm_base = devm_ioremap_resource(dev, res);
> -	if (IS_ERR(scp->l1tcm_base)) {
> -		ret = PTR_ERR(scp->l1tcm_base);
> -		if (ret != -EINVAL) {
> -			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> -		}
> -	} else {

                scp->l1tcm_base = NULL;

> -		scp->l1tcm_size = resource_size(res);
> -		scp->l1tcm_phys = res->start;
> -	}
> -
> -	scp->reg_base = devm_platform_ioremap_resource_byname(pdev, "cfg");
> -	if (IS_ERR(scp->reg_base))
> -		return dev_err_probe(dev, PTR_ERR(scp->reg_base),
> -				     "Failed to parse and map cfg memory\n");
> -
>  	ret = scp->data->scp_clk_get(scp);
>  	if (ret)
>  		return ret;
> @@ -957,6 +952,37 @@ static int scp_probe(struct platform_device *pdev)
>  	return ret;
>  }
>  
> +static int scp_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct mtk_scp_of_regs scp_regs;
> +	struct resource *res;
> +	int ret;
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg");
> +	scp_regs.reg_base = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(scp_regs.reg_base))
> +		return dev_err_probe(dev, PTR_ERR(scp_regs.reg_base),
> +				     "Failed to parse and map cfg memory\n");
> +
> +	/* l1tcm is an optional memory region */
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> +	scp_regs.l1tcm_base = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(scp_regs.l1tcm_base)) {
> +		ret = PTR_ERR(scp_regs.l1tcm_base);
> +		if (ret != -EINVAL)
> +			return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> +
> +		scp_regs.l1tcm_size = 0;
> +		scp_regs.l1tcm_phys = 0;
> +	} else {
> +		scp_regs.l1tcm_size = resource_size(res);
> +		scp_regs.l1tcm_phys = res->start;
> +	}
> +
> +	return scp_rproc_init(pdev, &scp_regs);
> +}
> +
>  static int scp_remove(struct platform_device *pdev)
>  {
>  	struct mtk_scp *scp = platform_get_drvdata(pdev);
> -- 
> 2.18.0
> 

Re: [PATCH v9 06/11] remoteproc: mediatek: Probe multi-core SCP

[Mathieu Poirier]

On Tue, Mar 28, 2023 at 10:27:28AM +0800, Tinghan Shen wrote:
> The difference of single-core SCP and multi-core SCP device tree is
> the presence of child device nodes described SCP cores. The SCP
> driver populates the platform device and checks the child nodes
> to identify whether it's a single-core SCP or a multi-core SCP.
> 
> Add the remoteproc instances for single-core SCP and multi-core SCP to
> the new added SCP cluster list. When the SCP driver is removed, it
> cleanup resources by walking through the cluster list.
> 
> Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> ---
>  drivers/remoteproc/mtk_common.h |   3 +
>  drivers/remoteproc/mtk_scp.c    | 150 ++++++++++++++++++++++++++++----
>  2 files changed, 137 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/remoteproc/mtk_common.h b/drivers/remoteproc/mtk_common.h
> index c0905aec3b4b..b73b60c22ea1 100644
> --- a/drivers/remoteproc/mtk_common.h
> +++ b/drivers/remoteproc/mtk_common.h
> @@ -128,6 +128,9 @@ struct mtk_scp {
>  	size_t dram_size;
>  
>  	struct rproc_subdev *rpmsg_subdev;
> +
> +	struct list_head elem;
> +	struct list_head *cluster;
>  };
>  
>  /**
> diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
> index 32ecd1450c6f..71882ed31e16 100644
> --- a/drivers/remoteproc/mtk_scp.c
> +++ b/drivers/remoteproc/mtk_scp.c
> @@ -30,6 +30,8 @@ struct mtk_scp_of_regs {
>  	phys_addr_t l1tcm_phys;
>  };
>  
> +static struct list_head cluster_cores;
> +

See my comment in the previous patch about this.

>  /**
>   * scp_get() - get a reference to SCP.
>   *
> @@ -863,7 +865,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
>  }
>  
>  static int scp_rproc_init(struct platform_device *pdev,
> -			  struct mtk_scp_of_regs *of_regs)
> +			  struct mtk_scp_of_regs *of_regs,
> +			  const struct mtk_scp_of_data *of_data)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> @@ -884,7 +887,7 @@ static int scp_rproc_init(struct platform_device *pdev,
>  	scp = (struct mtk_scp *)rproc->priv;
>  	scp->rproc = rproc;
>  	scp->dev = dev;
> -	scp->data = of_device_get_match_data(dev);
> +	scp->data = of_data;
>  	platform_set_drvdata(pdev, scp);
>  
>  	scp->reg_base = of_regs->reg_base;
> @@ -934,10 +937,6 @@ static int scp_rproc_init(struct platform_device *pdev,
>  		goto remove_subdev;
>  	}
>  
> -	ret = rproc_add(rproc);
> -	if (ret)
> -		goto remove_subdev;
> -
>  	return 0;
>  
>  remove_subdev:
> @@ -952,6 +951,109 @@ static int scp_rproc_init(struct platform_device *pdev,
>  	return ret;
>  }
>  
> +static void scp_rproc_free(struct mtk_scp *scp)
> +{
> +	int i;
> +
> +	scp_remove_rpmsg_subdev(scp);
> +	scp_ipi_unregister(scp, SCP_IPI_INIT);
> +	scp_unmap_memory_region(scp);
> +	for (i = 0; i < SCP_IPI_MAX; i++)
> +		mutex_destroy(&scp->ipi_desc[i].lock);
> +	mutex_destroy(&scp->send_lock);
> +}
> +
> +static int scp_cluster_init(struct platform_device *pdev,
> +			    struct mtk_scp_of_regs *of_regs)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct device_node *np = dev_of_node(dev);
> +	struct platform_device *cpdev;
> +	struct device_node *child;
> +	const struct mtk_scp_of_data **cluster_of_data;
> +	struct list_head *cluster = &cluster_cores;
> +	struct mtk_scp *scp, *temp;
> +	int core_id = 0, num_cores;
> +	int ret;
> +
> +	child = of_get_next_available_child(np, NULL);
> +	if (!child)
> +		return dev_err_probe(dev, -ENODEV, "No child node\n");
> +
> +	of_node_put(child);
> +	if (of_node_name_eq(child, "cros-ec-rpmsg")) {
> +		dev_dbg(dev, "single-core scp\n");
> +
> +		ret = scp_rproc_init(pdev, of_regs, of_device_get_match_data(dev));
> +		if (ret)
> +			return dev_err_probe(dev, ret, "Failed to initialize single-core scp\n");
> +
> +		/* the single-core drvdata is intended to be the mtk_scp object.
> +		 * the mtk_scp object is used by APIs in mtk_scp_rpmsg_info.
> +		 */
> +		scp = platform_get_drvdata(pdev);
> +		list_add_tail(&scp->elem, cluster);
> +		scp->cluster = cluster;
> +	} else {
> +		dev_dbg(dev, "multi-core scp\n");
> +
> +		cluster_of_data = (const struct mtk_scp_of_data **)of_device_get_match_data(dev);
> +
> +		num_cores = of_get_available_child_count(np);
> +
> +		for_each_available_child_of_node(np, child) {
> +			if (core_id >= num_cores || !cluster_of_data[core_id]) {

The first part of the if() condition is not needed - please remove.

> +				ret = -EINVAL;
> +				dev_err(dev, "Not support core %d\n", core_id);
> +				of_node_put(child);
> +				goto init_fail;
> +			}
> +
> +			cpdev = of_find_device_by_node(child);
> +			if (!cpdev) {
> +				ret = -ENODEV;
> +				dev_err(dev, "Not found platform device for core %d\n", core_id);
> +				of_node_put(child);
> +				goto init_fail;
> +			}
> +
> +			ret = scp_rproc_init(cpdev, of_regs, cluster_of_data[core_id]);
> +			if (ret) {
> +				dev_err(dev, "Failed to initialize core %d rproc\n", core_id);
> +				put_device(&cpdev->dev);
> +				of_node_put(child);
> +				goto init_fail;
> +			}
> +			scp = platform_get_drvdata(cpdev);
> +			list_add_tail(&scp->elem, cluster);
> +			scp->cluster = cluster;
> +			put_device(&cpdev->dev);
> +
> +			core_id++;
> +		}
> +	}
> +
> +	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
> +		ret = rproc_add(scp->rproc);
> +		if (ret)
> +			goto add_fail;
> +	}
> +
> +	return 0;
> +
> +add_fail:
> +	list_for_each_entry_continue(scp, cluster, elem) {
> +		rproc_del(scp->rproc);
> +	}
> +init_fail:
> +	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
> +		list_del(&scp->elem);
> +		scp_rproc_free(scp);
> +	}
> +
> +	return ret;
> +}
> +
>  static int scp_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
> @@ -980,21 +1082,30 @@ static int scp_probe(struct platform_device *pdev)
>  		scp_regs.l1tcm_phys = res->start;
>  	}
>  
> -	return scp_rproc_init(pdev, &scp_regs);
> +	ret = devm_of_platform_populate(dev);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "Failed to populate platform devices\n");
> +
> +	INIT_LIST_HEAD(&cluster_cores);
> +	platform_set_drvdata(pdev, &cluster_cores);

Why is this needed when @cluster_cores is a global variable???

The drvdata should be the cluster I talked about in the previous patch.  Doing
this will greatly simplify this patchset.

I will stop here for this revision.

Thanks,
Mathieu

> +
> +	ret = scp_cluster_init(pdev, &scp_regs);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "Failed to initialize scp cluster\n");
> +
> +	return 0;
>  }
>  
>  static int scp_remove(struct platform_device *pdev)
>  {
> -	struct mtk_scp *scp = platform_get_drvdata(pdev);
> -	int i;
> +	struct list_head *cluster = &cluster_cores;
> +	struct mtk_scp *scp, *temp;
>  
> -	rproc_del(scp->rproc);
> -	scp_remove_rpmsg_subdev(scp);
> -	scp_ipi_unregister(scp, SCP_IPI_INIT);
> -	scp_unmap_memory_region(scp);
> -	for (i = 0; i < SCP_IPI_MAX; i++)
> -		mutex_destroy(&scp->ipi_desc[i].lock);
> -	mutex_destroy(&scp->send_lock);
> +	list_for_each_entry_safe_reverse(scp, temp, cluster, elem) {
> +		list_del(&scp->elem);
> +		rproc_del(scp->rproc);
> +		scp_rproc_free(scp);
> +	}
>  
>  	return 0;
>  }
> @@ -1073,12 +1184,19 @@ static const struct mtk_scp_of_data mt8195_of_data_c1 = {
>  	.host_to_scp_int_bit = MT8195_CORE1_HOST_IPC_INT_BIT,
>  };
>  
> +static const struct mtk_scp_of_data *mt8195_of_data_cores[] = {
> +	&mt8195_of_data,
> +	&mt8195_of_data_c1,
> +	NULL
> +};
> +
>  static const struct of_device_id mtk_scp_of_match[] = {
>  	{ .compatible = "mediatek,mt8183-scp", .data = &mt8183_of_data },
>  	{ .compatible = "mediatek,mt8186-scp", .data = &mt8186_of_data },
>  	{ .compatible = "mediatek,mt8188-scp", .data = &mt8188_of_data },
>  	{ .compatible = "mediatek,mt8192-scp", .data = &mt8192_of_data },
>  	{ .compatible = "mediatek,mt8195-scp", .data = &mt8195_of_data },
> +	{ .compatible = "mediatek,mt8195-scp-dual", .data = &mt8195_of_data_cores },
>  	{},
>  };
>  MODULE_DEVICE_TABLE(of, mtk_scp_of_match);
> -- 
> 2.18.0
> 

Re: [PATCH v9 11/11] arm64: dts: mediatek: mt8195: Add SCP 2nd core

[Mathieu Poirier]

On Fri, 31 Mar 2023 at 05:00, Matthias Brugger <matthias.bgg@gmail.com> wrote:
>
>
>
> On 28/03/2023 04:27, Tinghan Shen wrote:
> > Rewrite the MT8195 SCP device node as a cluster and
> > add the SCP 2nd core in it.
> >
> > Since the SCP device node is changed to multi-core structure,
> > enable SCP cluster to enable probing SCP core 0.
> >
> > Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> > Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
>
> As this is a bigger change I'd prefer to take it through my tree once the driver
> and dt-bindings changes are merged. Given the fact, maybe it would make sense to
> take 2/11 through my tree as well.
>

Might as well take the whole thing through your tree, which I'm
totally fine with.  But when that happens, you will also have to pick
up other potential patches that touch these files.  We'll sort it out.

> Regards,
> Matthias
>
> > ---
> >   .../boot/dts/mediatek/mt8195-cherry.dtsi      |  6 +++-
> >   arch/arm64/boot/dts/mediatek/mt8195.dtsi      | 32 ++++++++++++++-----
> >   2 files changed, 29 insertions(+), 9 deletions(-)
> >
> > diff --git a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> > index 56749cfe7c33..31415d71b6a4 100644
> > --- a/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> > +++ b/arch/arm64/boot/dts/mediatek/mt8195-cherry.dtsi
> > @@ -933,7 +933,11 @@
> >       interrupts-extended = <&pio 222 IRQ_TYPE_LEVEL_HIGH>;
> >   };
> >
> > -&scp {
> > +&scp_cluster {
> > +     status = "okay";
> > +};
> > +
> > +&scp_c0 {
> >       status = "okay";
> >
> >       firmware-name = "mediatek/mt8195/scp.img";
> > diff --git a/arch/arm64/boot/dts/mediatek/mt8195.dtsi b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> > index 8fc527570791..5fe5fb32261e 100644
> > --- a/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> > +++ b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> > @@ -826,14 +826,30 @@
> >                       clocks = <&infracfg_ao CLK_INFRA_AO_GCE2>;
> >               };
> >
> > -             scp: scp@10500000 {
> > -                     compatible = "mediatek,mt8195-scp";
> > -                     reg = <0 0x10500000 0 0x100000>,
> > -                           <0 0x10720000 0 0xe0000>,
> > -                           <0 0x10700000 0 0x8000>;
> > -                     reg-names = "sram", "cfg", "l1tcm";
> > -                     interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
> > +             scp_cluster: scp@10500000 {
> > +                     compatible = "mediatek,mt8195-scp-dual";
> > +                     reg = <0 0x10720000 0 0xe0000>, <0 0x10700000 0 0x8000>;
> > +                     reg-names = "cfg", "l1tcm";
> > +                     #address-cells = <1>;
> > +                     #size-cells = <1>;
> > +                     ranges = <0 0 0x10500000 0x100000>;
> >                       status = "disabled";
> > +
> > +                     scp_c0: scp@0 {
> > +                             compatible = "mediatek,scp-core";
> > +                             reg = <0x0 0xa0000>;
> > +                             reg-names = "sram";
> > +                             interrupts = <GIC_SPI 462 IRQ_TYPE_LEVEL_HIGH 0>;
> > +                             status = "disabled";
> > +                     };
> > +
> > +                     scp_c1: scp@a0000 {
> > +                             compatible = "mediatek,scp-core";
> > +                             reg = <0xa0000 0x20000>;
> > +                             reg-names = "sram";
> > +                             interrupts = <GIC_SPI 463 IRQ_TYPE_LEVEL_HIGH 0>;
> > +                             status = "disabled";
> > +                     };
> >               };
> >
> >               scp_adsp: clock-controller@10720000 {
> > @@ -2309,7 +2325,7 @@
> >                                <&iommu_vdo M4U_PORT_L19_VENC_REF_LUMA>,
> >                                <&iommu_vdo M4U_PORT_L19_VENC_REF_CHROMA>;
> >                       interrupts = <GIC_SPI 341 IRQ_TYPE_LEVEL_HIGH 0>;
> > -                     mediatek,scp = <&scp>;
> > +                     mediatek,scp = <&scp_c0>;
> >                       clocks = <&vencsys CLK_VENC_VENC>;
> >                       clock-names = "venc_sel";
> >                       assigned-clocks = <&topckgen CLK_TOP_VENC>;

Re: [PATCH v9 05/11] remoteproc: mediatek: Extract remoteproc initialization flow

[TingHan Shen (沈廷翰)]

Hi Mathieu,

On Fri, 2023-03-31 at 11:44 -0600, Mathieu Poirier wrote:
> External email : Please do not click links or open attachments until you have verified the sender or the content.
> 
> 
> On Tue, Mar 28, 2023 at 10:27:27AM +0800, Tinghan Shen wrote:
> > This is the preparation for probing multi-core SCP. The remoteproc
> > initialization flow is similar on cores and is reused to avoid
> > redundant code.
> > 
> > The registers of config and l1tcm are shared for multi-core
> > SCP. Reuse the mapped addresses for all cores.
> > 
> > Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> > Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> > ---
> >  drivers/remoteproc/mtk_scp.c | 64 +++++++++++++++++++++++++-----------
> >  1 file changed, 45 insertions(+), 19 deletions(-)
> > 
> > diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
> > index a3b9bc158cd9..32ecd1450c6f 100644
> > --- a/drivers/remoteproc/mtk_scp.c
> > +++ b/drivers/remoteproc/mtk_scp.c
> > @@ -23,6 +23,13 @@
> >  #define MAX_CODE_SIZE 0x500000
> >  #define SECTION_NAME_IPI_BUFFER ".ipi_buffer"
> > 
> > +struct mtk_scp_of_regs {
> > +     void __iomem *reg_base;
> > +     void __iomem *l1tcm_base;
> > +     size_t l1tcm_size;
> > +     phys_addr_t l1tcm_phys;
> > +};
> > +
> 
> This should represent the cluster with a list of mtk_scp instead of @cluster_cores as
> introduced in the next patch.

If I'm understanding you correctly, you're suggesting that @cluster_cores should be included 
as a member of this structure. Is that correct?

Best regards,
TingHan

> 
> >  /**
> >   * scp_get() - get a reference to SCP.
> >   *
> > @@ -855,7 +862,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
> >       }
> >  }
> > 
> > -static int scp_probe(struct platform_device *pdev)
> > +static int scp_rproc_init(struct platform_device *pdev,
> > +                       struct mtk_scp_of_regs *of_regs)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -879,6 +887,11 @@ static int scp_probe(struct platform_device *pdev)
> >       scp->data = of_device_get_match_data(dev);
> >       platform_set_drvdata(pdev, scp);
> > 
> > +     scp->reg_base = of_regs->reg_base;
> > +     scp->l1tcm_base = of_regs->l1tcm_base;
> > +     scp->l1tcm_size = of_regs->l1tcm_size;
> > +     scp->l1tcm_phys = of_regs->l1tcm_phys;
> > +
> >       res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram");
> >       scp->sram_base = devm_ioremap_resource(dev, res);
> >       if (IS_ERR(scp->sram_base))
> > @@ -888,24 +901,6 @@ static int scp_probe(struct platform_device *pdev)
> >       scp->sram_size = resource_size(res);
> >       scp->sram_phys = res->start;
> > 
> > -     /* l1tcm is an optional memory region */
> > -     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> > -     scp->l1tcm_base = devm_ioremap_resource(dev, res);
> > -     if (IS_ERR(scp->l1tcm_base)) {
> > -             ret = PTR_ERR(scp->l1tcm_base);
> > -             if (ret != -EINVAL) {
> > -                     return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> > -             }
> > -     } else {
> 
>                 scp->l1tcm_base = NULL;
> 
> > -             scp->l1tcm_size = resource_size(res);
> > -             scp->l1tcm_phys = res->start;
> > -     }
> > -
> > -     scp->reg_base = devm_platform_ioremap_resource_byname(pdev, "cfg");
> > -     if (IS_ERR(scp->reg_base))
> > -             return dev_err_probe(dev, PTR_ERR(scp->reg_base),
> > -                                  "Failed to parse and map cfg memory\n");
> > -
> >       ret = scp->data->scp_clk_get(scp);
> >       if (ret)
> >               return ret;
> > @@ -957,6 +952,37 @@ static int scp_probe(struct platform_device *pdev)
> >       return ret;
> >  }
> > 
> > +static int scp_probe(struct platform_device *pdev)
> > +{
> > +     struct device *dev = &pdev->dev;
> > +     struct mtk_scp_of_regs scp_regs;
> > +     struct resource *res;
> > +     int ret;
> > +
> > +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg");
> > +     scp_regs.reg_base = devm_ioremap_resource(dev, res);
> > +     if (IS_ERR(scp_regs.reg_base))
> > +             return dev_err_probe(dev, PTR_ERR(scp_regs.reg_base),
> > +                                  "Failed to parse and map cfg memory\n");
> > +
> > +     /* l1tcm is an optional memory region */
> > +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> > +     scp_regs.l1tcm_base = devm_ioremap_resource(dev, res);
> > +     if (IS_ERR(scp_regs.l1tcm_base)) {
> > +             ret = PTR_ERR(scp_regs.l1tcm_base);
> > +             if (ret != -EINVAL)
> > +                     return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> > +
> > +             scp_regs.l1tcm_size = 0;
> > +             scp_regs.l1tcm_phys = 0;
> > +     } else {
> > +             scp_regs.l1tcm_size = resource_size(res);
> > +             scp_regs.l1tcm_phys = res->start;
> > +     }
> > +
> > +     return scp_rproc_init(pdev, &scp_regs);
> > +}
> > +
> >  static int scp_remove(struct platform_device *pdev)
> >  {
> >       struct mtk_scp *scp = platform_get_drvdata(pdev);
> > --
> > 2.18.0
> > 

-- 
Best regards,
TingHan

Re: [PATCH v9 05/11] remoteproc: mediatek: Extract remoteproc initialization flow

[Mathieu Poirier]

On Wed, Apr 19, 2023 at 03:38:14AM +0000, TingHan Shen (沈廷翰) wrote:
> Hi Mathieu,
> 
> On Fri, 2023-03-31 at 11:44 -0600, Mathieu Poirier wrote:
> > External email : Please do not click links or open attachments until you have verified the sender or the content.
> > 
> > 
> > On Tue, Mar 28, 2023 at 10:27:27AM +0800, Tinghan Shen wrote:
> > > This is the preparation for probing multi-core SCP. The remoteproc
> > > initialization flow is similar on cores and is reused to avoid
> > > redundant code.
> > > 
> > > The registers of config and l1tcm are shared for multi-core
> > > SCP. Reuse the mapped addresses for all cores.
> > > 
> > > Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
> > > Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> > > ---
> > >  drivers/remoteproc/mtk_scp.c | 64 +++++++++++++++++++++++++-----------
> > >  1 file changed, 45 insertions(+), 19 deletions(-)
> > > 
> > > diff --git a/drivers/remoteproc/mtk_scp.c b/drivers/remoteproc/mtk_scp.c
> > > index a3b9bc158cd9..32ecd1450c6f 100644
> > > --- a/drivers/remoteproc/mtk_scp.c
> > > +++ b/drivers/remoteproc/mtk_scp.c
> > > @@ -23,6 +23,13 @@
> > >  #define MAX_CODE_SIZE 0x500000
> > >  #define SECTION_NAME_IPI_BUFFER ".ipi_buffer"
> > > 
> > > +struct mtk_scp_of_regs {
> > > +     void __iomem *reg_base;
> > > +     void __iomem *l1tcm_base;
> > > +     size_t l1tcm_size;
> > > +     phys_addr_t l1tcm_phys;
> > > +};
> > > +
> > 
> > This should represent the cluster with a list of mtk_scp instead of @cluster_cores as
> > introduced in the next patch.
> 
> If I'm understanding you correctly, you're suggesting that @cluster_cores should be included 
> as a member of this structure. Is that correct?

Correct.  Than this structure is allocated in probe() and added as driver data
for the platform device.  Its name should also be something like
mtk_scp_cluster or something like that.  I suggest you look at what has been
done in ti_k3_r5_remoteproc.c, your end design should be quite similar to that.
In fact you are close but a few things need to be addressed.

> 
> Best regards,
> TingHan
> 
> > 
> > >  /**
> > >   * scp_get() - get a reference to SCP.
> > >   *
> > > @@ -855,7 +862,8 @@ static void scp_remove_rpmsg_subdev(struct mtk_scp *scp)
> > >       }
> > >  }
> > > 
> > > -static int scp_probe(struct platform_device *pdev)
> > > +static int scp_rproc_init(struct platform_device *pdev,
> > > +                       struct mtk_scp_of_regs *of_regs)
> > >  {
> > >       struct device *dev = &pdev->dev;
> > >       struct device_node *np = dev->of_node;
> > > @@ -879,6 +887,11 @@ static int scp_probe(struct platform_device *pdev)
> > >       scp->data = of_device_get_match_data(dev);
> > >       platform_set_drvdata(pdev, scp);
> > > 
> > > +     scp->reg_base = of_regs->reg_base;
> > > +     scp->l1tcm_base = of_regs->l1tcm_base;
> > > +     scp->l1tcm_size = of_regs->l1tcm_size;
> > > +     scp->l1tcm_phys = of_regs->l1tcm_phys;
> > > +
> > >       res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram");
> > >       scp->sram_base = devm_ioremap_resource(dev, res);
> > >       if (IS_ERR(scp->sram_base))
> > > @@ -888,24 +901,6 @@ static int scp_probe(struct platform_device *pdev)
> > >       scp->sram_size = resource_size(res);
> > >       scp->sram_phys = res->start;
> > > 
> > > -     /* l1tcm is an optional memory region */
> > > -     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> > > -     scp->l1tcm_base = devm_ioremap_resource(dev, res);
> > > -     if (IS_ERR(scp->l1tcm_base)) {
> > > -             ret = PTR_ERR(scp->l1tcm_base);
> > > -             if (ret != -EINVAL) {
> > > -                     return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> > > -             }
> > > -     } else {
> > 
> >                 scp->l1tcm_base = NULL;
> > 
> > > -             scp->l1tcm_size = resource_size(res);
> > > -             scp->l1tcm_phys = res->start;
> > > -     }
> > > -
> > > -     scp->reg_base = devm_platform_ioremap_resource_byname(pdev, "cfg");
> > > -     if (IS_ERR(scp->reg_base))
> > > -             return dev_err_probe(dev, PTR_ERR(scp->reg_base),
> > > -                                  "Failed to parse and map cfg memory\n");
> > > -
> > >       ret = scp->data->scp_clk_get(scp);
> > >       if (ret)
> > >               return ret;
> > > @@ -957,6 +952,37 @@ static int scp_probe(struct platform_device *pdev)
> > >       return ret;
> > >  }
> > > 
> > > +static int scp_probe(struct platform_device *pdev)
> > > +{
> > > +     struct device *dev = &pdev->dev;
> > > +     struct mtk_scp_of_regs scp_regs;
> > > +     struct resource *res;
> > > +     int ret;
> > > +
> > > +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg");
> > > +     scp_regs.reg_base = devm_ioremap_resource(dev, res);
> > > +     if (IS_ERR(scp_regs.reg_base))
> > > +             return dev_err_probe(dev, PTR_ERR(scp_regs.reg_base),
> > > +                                  "Failed to parse and map cfg memory\n");
> > > +
> > > +     /* l1tcm is an optional memory region */
> > > +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
> > > +     scp_regs.l1tcm_base = devm_ioremap_resource(dev, res);
> > > +     if (IS_ERR(scp_regs.l1tcm_base)) {
> > > +             ret = PTR_ERR(scp_regs.l1tcm_base);
> > > +             if (ret != -EINVAL)
> > > +                     return dev_err_probe(dev, ret, "Failed to map l1tcm memory\n");
> > > +
> > > +             scp_regs.l1tcm_size = 0;
> > > +             scp_regs.l1tcm_phys = 0;
> > > +     } else {
> > > +             scp_regs.l1tcm_size = resource_size(res);
> > > +             scp_regs.l1tcm_phys = res->start;
> > > +     }
> > > +
> > > +     return scp_rproc_init(pdev, &scp_regs);
> > > +}
> > > +
> > >  static int scp_remove(struct platform_device *pdev)
> > >  {
> > >       struct mtk_scp *scp = platform_get_drvdata(pdev);
> > > --
> > > 2.18.0
> > > 
> 
> -- 
> Best regards,
> TingHan