linux-clk.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v8 0/4] MSM8998 Multimedia Clock Controller
@ 2019-11-08 23:16 Jeffrey Hugo
  2019-11-08 23:17 ` [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc Jeffrey Hugo
                   ` (3 more replies)
  0 siblings, 4 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-08 23:16 UTC (permalink / raw)
  Cc: agross, bjorn.andersson, marc.w.gonzalez, mturquette, sboyd,
	robh+dt, mark.rutland, linux-arm-msm, linux-clk, linux-kernel,
	devicetree, Jeffrey Hugo

The multimedia clock controller (mmcc) is the main clock controller for
the multimedia subsystem and is required to enable things like display and
camera.

v8:
-drop dts changes from series per Stephen's request
-fix the mislabeled mmcc example
-drop Stephen as maintainer of the mmcc binding

v7:
-port to gcc.yaml.  Drop reviewed-by for DT changes as they got completely
rewritten
-drop "clk: qcom: smd: Add XO clock for MSM8998".  Will need to find another
solution and this is not blocking right now
-convert mmcc to yaml
-drop errant clk.h include
-use blank entries in the DT when no clock is available

v6:
-drop clk_get from mmcc clock provider

v5:
-handle the case where gcc uses rpmcc for xo, but the link is not specified in dt
-have gcc select rpmcc

v4:
-fix makefile to use correct config item
-pick up tags
-fix ordering of clocks and clock-names in dt
-drop MODULE_ALIAS
-wait for xo in mmcc since that was found to be useful in some debug configs

v3:
-Rebase onto linux-next to get the final version of the clk parent rewrite
series
-Moved the bindings header to the bindings patch per Rob
-Made xo manditory for GCC to work around the lack of clk orphan probe defer
to avoid the uart console glitch

v2:
-Rebased on the "Rewrite clk parent handling" series and updated to the clk init
mechanisms introduced there.
-Marked XO clk as CLK_IGNORE_UNUSED to avoid the concern about the XO going away
"incorrectly" during late init
-Corrected the name of the XO clock to "xo"
-Dropped the fake XO clock in GCC to prevent a namespace conflict
-Fully enumerated the external clocks (DSI PLLs, etc) in the DT binding
-Cleaned up the weird newlines in the added DT node
-Added DT header file to msm8998 DT for future clients

Jeffrey Hugo (4):
  dt-bindings: clock: Document external clocks for MSM8998 gcc
  dt-bindings: clock: Convert qcom,mmcc to DT schema
  dt-bindings: clock: Add support for the MSM8998 mmcc
  clk: qcom: Add MSM8998 Multimedia Clock Controller (MMCC) driver

 .../devicetree/bindings/clock/qcom,gcc.yaml        |   47 +-
 .../devicetree/bindings/clock/qcom,mmcc.txt        |   28 -
 .../devicetree/bindings/clock/qcom,mmcc.yaml       |   95 +
 drivers/clk/qcom/Kconfig                           |    9 +
 drivers/clk/qcom/Makefile                          |    1 +
 drivers/clk/qcom/mmcc-msm8998.c                    | 2913 ++++++++++++++++++++
 include/dt-bindings/clock/qcom,mmcc-msm8998.h      |  210 ++
 7 files changed, 3261 insertions(+), 42 deletions(-)
 delete mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.txt
 create mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
 create mode 100644 drivers/clk/qcom/mmcc-msm8998.c
 create mode 100644 include/dt-bindings/clock/qcom,mmcc-msm8998.h

-- 
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.


^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-08 23:16 [PATCH v8 0/4] MSM8998 Multimedia Clock Controller Jeffrey Hugo
@ 2019-11-08 23:17 ` Jeffrey Hugo
  2019-11-12  0:44   ` Rob Herring
  2019-11-08 23:17 ` [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema Jeffrey Hugo
                   ` (2 subsequent siblings)
  3 siblings, 1 reply; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-08 23:17 UTC (permalink / raw)
  To: mturquette, sboyd, robh+dt, mark.rutland
  Cc: agross, bjorn.andersson, marc.w.gonzalez, linux-arm-msm,
	linux-clk, linux-kernel, devicetree, Jeffrey Hugo

The global clock controller on MSM8998 can consume a number of external
clocks.  Document them.

Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
---
 .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
 1 file changed, 33 insertions(+), 14 deletions(-)

diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
index e73a56f..2f3512b 100644
--- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
+++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
@@ -40,20 +40,38 @@ properties:
        - qcom,gcc-sm8150
 
   clocks:
-    minItems: 1
-    maxItems: 3
-    items:
-      - description: Board XO source
-      - description: Board active XO source
-      - description: Sleep clock source
+    oneOf:
+      #qcom,gcc-sm8150
+      #qcom,gcc-sc7180
+      - items:
+        - description: Board XO source
+        - description: Board active XO source
+        - description: Sleep clock source
+      #qcom,gcc-msm8998
+      - items:
+        - description: Board XO source
+        - description: USB 3.0 phy pipe clock
+        - description: UFS phy rx symbol clock for pipe 0
+        - description: UFS phy rx symbol clock for pipe 1
+        - description: UFS phy tx symbol clock
+        - description: PCIE phy pipe clock
 
   clock-names:
-    minItems: 1
-    maxItems: 3
-    items:
-      - const: bi_tcxo
-      - const: bi_tcxo_ao
-      - const: sleep_clk
+    oneOf:
+      #qcom,gcc-sm8150
+      #qcom,gcc-sc7180
+      - items:
+        - const: bi_tcxo
+        - const: bi_tcxo_ao
+        - const: sleep_clk
+      #qcom,gcc-msm8998
+      - items:
+        - const: xo
+        - const: usb3_pipe
+        - const: ufs_rx_symbol0
+        - const: ufs_rx_symbol1
+        - const: ufs_tx_symbol0
+        - const: pcie0_pipe
 
   '#clock-cells':
     const: 1
@@ -118,6 +136,7 @@ else:
       compatible:
         contains:
           enum:
+            - qcom,gcc-msm8998
             - qcom,gcc-sm8150
             - qcom,gcc-sc7180
   then:
@@ -179,8 +198,8 @@ examples:
     clock-controller@100000 {
       compatible = "qcom,gcc-sc7180";
       reg = <0x100000 0x1f0000>;
-      clocks = <&rpmhcc 0>, <&rpmhcc 1>;
-      clock-names = "bi_tcxo", "bi_tcxo_ao";
+      clocks = <&rpmhcc 0>, <&rpmhcc 1>, <0>;
+      clock-names = "bi_tcxo", "bi_tcxo_ao", "sleep_clk";
       #clock-cells = <1>;
       #reset-cells = <1>;
       #power-domain-cells = <1>;
-- 
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.


^ permalink raw reply related	[flat|nested] 16+ messages in thread

* [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema
  2019-11-08 23:16 [PATCH v8 0/4] MSM8998 Multimedia Clock Controller Jeffrey Hugo
  2019-11-08 23:17 ` [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc Jeffrey Hugo
@ 2019-11-08 23:17 ` Jeffrey Hugo
  2019-11-12  0:58   ` Rob Herring
  2019-11-08 23:17 ` [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc Jeffrey Hugo
  2019-11-08 23:18 ` [PATCH v8 4/4] clk: qcom: Add MSM8998 Multimedia Clock Controller (MMCC) driver Jeffrey Hugo
  3 siblings, 1 reply; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-08 23:17 UTC (permalink / raw)
  To: mturquette, sboyd, robh+dt, mark.rutland
  Cc: agross, bjorn.andersson, marc.w.gonzalez, linux-arm-msm,
	linux-clk, linux-kernel, devicetree, Jeffrey Hugo

Convert the qcom,mmcc-X clock controller binding to DT schema.

Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
---
 .../devicetree/bindings/clock/qcom,mmcc.txt        | 28 ----------
 .../devicetree/bindings/clock/qcom,mmcc.yaml       | 59 ++++++++++++++++++++++
 2 files changed, 59 insertions(+), 28 deletions(-)
 delete mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.txt
 create mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.yaml

diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt b/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
deleted file mode 100644
index 8b0f784..0000000
--- a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
+++ /dev/null
@@ -1,28 +0,0 @@
-Qualcomm Multimedia Clock & Reset Controller Binding
-----------------------------------------------------
-
-Required properties :
-- compatible : shall contain only one of the following:
-
-			"qcom,mmcc-apq8064"
-			"qcom,mmcc-apq8084"
-			"qcom,mmcc-msm8660"
-			"qcom,mmcc-msm8960"
-			"qcom,mmcc-msm8974"
-			"qcom,mmcc-msm8996"
-
-- reg : shall contain base register location and length
-- #clock-cells : shall contain 1
-- #reset-cells : shall contain 1
-
-Optional properties :
-- #power-domain-cells : shall contain 1
-
-Example:
-	clock-controller@4000000 {
-		compatible = "qcom,mmcc-msm8960";
-		reg = <0x4000000 0x1000>;
-		#clock-cells = <1>;
-		#reset-cells = <1>;
-		#power-domain-cells = <1>;
-	};
diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
new file mode 100644
index 0000000..61ed4a2
--- /dev/null
+++ b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
@@ -0,0 +1,59 @@
+# SPDX-License-Identifier: GPL-2.0-only
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/bindings/clock/qcom,mmcc.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Qualcomm Multimedia Clock & Reset Controller Binding
+
+maintainers:
+  - Jeffrey Hugo <jhugo@codeaurora.org>
+
+description: |
+  Qualcomm multimedia clock control module which supports the clocks, resets and
+  power domains.
+
+properties:
+  compatible :
+    enum:
+       - qcom,mmcc-apq8064
+       - qcom,mmcc-apq8084
+       - qcom,mmcc-msm8660
+       - qcom,mmcc-msm8960
+       - qcom,mmcc-msm8974
+       - qcom,mmcc-msm8996
+
+  '#clock-cells':
+    const: 1
+
+  '#reset-cells':
+    const: 1
+
+  '#power-domain-cells':
+    const: 1
+
+  reg:
+    maxItems: 1
+
+  protected-clocks:
+    description:
+       Protected clock specifier list as per common clock binding
+
+required:
+  - compatible
+  - reg
+  - '#clock-cells'
+  - '#reset-cells'
+  - '#power-domain-cells'
+
+examples:
+  # Example for MMCC for MSM8960:
+  - |
+    clock-controller@4000000 {
+      compatible = "qcom,mmcc-msm8960";
+      reg = <0x4000000 0x1000>;
+      #clock-cells = <1>;
+      #reset-cells = <1>;
+      #power-domain-cells = <1>;
+    };
+...
-- 
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.


^ permalink raw reply related	[flat|nested] 16+ messages in thread

* [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc
  2019-11-08 23:16 [PATCH v8 0/4] MSM8998 Multimedia Clock Controller Jeffrey Hugo
  2019-11-08 23:17 ` [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc Jeffrey Hugo
  2019-11-08 23:17 ` [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema Jeffrey Hugo
@ 2019-11-08 23:17 ` Jeffrey Hugo
  2019-11-12  0:55   ` Rob Herring
  2019-11-12  0:59   ` Rob Herring
  2019-11-08 23:18 ` [PATCH v8 4/4] clk: qcom: Add MSM8998 Multimedia Clock Controller (MMCC) driver Jeffrey Hugo
  3 siblings, 2 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-08 23:17 UTC (permalink / raw)
  To: mturquette, sboyd, robh+dt, mark.rutland
  Cc: agross, bjorn.andersson, marc.w.gonzalez, linux-arm-msm,
	linux-clk, linux-kernel, devicetree, Jeffrey Hugo

Document the multimedia clock controller found on MSM8998.

Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
---
 .../devicetree/bindings/clock/qcom,mmcc.yaml       |  36 ++++
 include/dt-bindings/clock/qcom,mmcc-msm8998.h      | 210 +++++++++++++++++++++
 2 files changed, 246 insertions(+)
 create mode 100644 include/dt-bindings/clock/qcom,mmcc-msm8998.h

diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
index 61ed4a2..520d4e0 100644
--- a/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
+++ b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
@@ -22,6 +22,31 @@ properties:
        - qcom,mmcc-msm8960
        - qcom,mmcc-msm8974
        - qcom,mmcc-msm8996
+       - qcom,mmcc-msm8998
+
+  clocks:
+    items:
+      - description: Board XO source
+      - description: Global PLL 0 clock
+      - description: DSI phy instance 0 dsi clock
+      - description: DSI phy instance 0 byte clock
+      - description: DSI phy instance 1 dsi clock
+      - description: DSI phy instance 1 byte clock
+      - description: HDMI phy PLL clock
+      - description: DisplayPort phy PLL vco clock
+      - description: DisplayPort phy PLL link clock
+
+  clock-names:
+    items:
+      - const: xo
+      - const: gpll0
+      - const: dsi0dsi
+      - const: dsi0byte
+      - const: dsi1dsi
+      - const: dsi1byte
+      - const: hdmipll
+      - const: dpvco
+      - const: dplink
 
   '#clock-cells':
     const: 1
@@ -46,6 +71,17 @@ required:
   - '#reset-cells'
   - '#power-domain-cells'
 
+if:
+  properties:
+    compatible:
+      contains:
+        const: qcom,mmcc-msm8998
+
+then:
+  required:
+    - clocks
+    - clock-names
+
 examples:
   # Example for MMCC for MSM8960:
   - |
diff --git a/include/dt-bindings/clock/qcom,mmcc-msm8998.h b/include/dt-bindings/clock/qcom,mmcc-msm8998.h
new file mode 100644
index 0000000..ecbafdb
--- /dev/null
+++ b/include/dt-bindings/clock/qcom,mmcc-msm8998.h
@@ -0,0 +1,210 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2019, The Linux Foundation. All rights reserved.
+ */
+
+#ifndef _DT_BINDINGS_CLK_MSM_MMCC_8998_H
+#define _DT_BINDINGS_CLK_MSM_MMCC_8998_H
+
+#define MMPLL0						0
+#define MMPLL0_OUT_EVEN					1
+#define MMPLL1						2
+#define MMPLL1_OUT_EVEN					3
+#define MMPLL3						4
+#define MMPLL3_OUT_EVEN					5
+#define MMPLL4						6
+#define MMPLL4_OUT_EVEN					7
+#define MMPLL5						8
+#define MMPLL5_OUT_EVEN					9
+#define MMPLL6						10
+#define MMPLL6_OUT_EVEN					11
+#define MMPLL7						12
+#define MMPLL7_OUT_EVEN					13
+#define MMPLL10						14
+#define MMPLL10_OUT_EVEN				15
+#define BYTE0_CLK_SRC					16
+#define BYTE1_CLK_SRC					17
+#define CCI_CLK_SRC					18
+#define CPP_CLK_SRC					19
+#define CSI0_CLK_SRC					20
+#define CSI1_CLK_SRC					21
+#define CSI2_CLK_SRC					22
+#define CSI3_CLK_SRC					23
+#define CSIPHY_CLK_SRC					24
+#define CSI0PHYTIMER_CLK_SRC				25
+#define CSI1PHYTIMER_CLK_SRC				26
+#define CSI2PHYTIMER_CLK_SRC				27
+#define DP_AUX_CLK_SRC					28
+#define DP_CRYPTO_CLK_SRC				29
+#define DP_LINK_CLK_SRC					30
+#define DP_PIXEL_CLK_SRC				31
+#define ESC0_CLK_SRC					32
+#define ESC1_CLK_SRC					33
+#define EXTPCLK_CLK_SRC					34
+#define FD_CORE_CLK_SRC					35
+#define HDMI_CLK_SRC					36
+#define JPEG0_CLK_SRC					37
+#define MAXI_CLK_SRC					38
+#define MCLK0_CLK_SRC					39
+#define MCLK1_CLK_SRC					40
+#define MCLK2_CLK_SRC					41
+#define MCLK3_CLK_SRC					42
+#define MDP_CLK_SRC					43
+#define VSYNC_CLK_SRC					44
+#define AHB_CLK_SRC					45
+#define AXI_CLK_SRC					46
+#define PCLK0_CLK_SRC					47
+#define PCLK1_CLK_SRC					48
+#define ROT_CLK_SRC					49
+#define VIDEO_CORE_CLK_SRC				50
+#define VIDEO_SUBCORE0_CLK_SRC				51
+#define VIDEO_SUBCORE1_CLK_SRC				52
+#define VFE0_CLK_SRC					53
+#define VFE1_CLK_SRC					54
+#define MISC_AHB_CLK					55
+#define VIDEO_CORE_CLK					56
+#define VIDEO_AHB_CLK					57
+#define VIDEO_AXI_CLK					58
+#define VIDEO_MAXI_CLK					59
+#define VIDEO_SUBCORE0_CLK				60
+#define VIDEO_SUBCORE1_CLK				61
+#define MDSS_AHB_CLK					62
+#define MDSS_HDMI_DP_AHB_CLK				63
+#define MDSS_AXI_CLK					64
+#define MDSS_PCLK0_CLK					65
+#define MDSS_PCLK1_CLK					66
+#define MDSS_MDP_CLK					67
+#define MDSS_MDP_LUT_CLK				68
+#define MDSS_EXTPCLK_CLK				69
+#define MDSS_VSYNC_CLK					70
+#define MDSS_HDMI_CLK					71
+#define MDSS_BYTE0_CLK					72
+#define MDSS_BYTE1_CLK					73
+#define MDSS_ESC0_CLK					74
+#define MDSS_ESC1_CLK					75
+#define MDSS_ROT_CLK					76
+#define MDSS_DP_LINK_CLK				77
+#define MDSS_DP_LINK_INTF_CLK				78
+#define MDSS_DP_CRYPTO_CLK				79
+#define MDSS_DP_PIXEL_CLK				80
+#define MDSS_DP_AUX_CLK					81
+#define MDSS_BYTE0_INTF_CLK				82
+#define MDSS_BYTE1_INTF_CLK				83
+#define CAMSS_CSI0PHYTIMER_CLK				84
+#define CAMSS_CSI1PHYTIMER_CLK				85
+#define CAMSS_CSI2PHYTIMER_CLK				86
+#define CAMSS_CSI0_CLK					87
+#define CAMSS_CSI0_AHB_CLK				88
+#define CAMSS_CSI0RDI_CLK				89
+#define CAMSS_CSI0PIX_CLK				90
+#define CAMSS_CSI1_CLK					91
+#define CAMSS_CSI1_AHB_CLK				92
+#define CAMSS_CSI1RDI_CLK				93
+#define CAMSS_CSI1PIX_CLK				94
+#define CAMSS_CSI2_CLK					95
+#define CAMSS_CSI2_AHB_CLK				96
+#define CAMSS_CSI2RDI_CLK				97
+#define CAMSS_CSI2PIX_CLK				98
+#define CAMSS_CSI3_CLK					99
+#define CAMSS_CSI3_AHB_CLK				100
+#define CAMSS_CSI3RDI_CLK				101
+#define CAMSS_CSI3PIX_CLK				102
+#define CAMSS_ISPIF_AHB_CLK				103
+#define CAMSS_CCI_CLK					104
+#define CAMSS_CCI_AHB_CLK				105
+#define CAMSS_MCLK0_CLK					106
+#define CAMSS_MCLK1_CLK					107
+#define CAMSS_MCLK2_CLK					108
+#define CAMSS_MCLK3_CLK					109
+#define CAMSS_TOP_AHB_CLK				110
+#define CAMSS_AHB_CLK					111
+#define CAMSS_MICRO_AHB_CLK				112
+#define CAMSS_JPEG0_CLK					113
+#define CAMSS_JPEG_AHB_CLK				114
+#define CAMSS_JPEG_AXI_CLK				115
+#define CAMSS_VFE0_AHB_CLK				116
+#define CAMSS_VFE1_AHB_CLK				117
+#define CAMSS_VFE0_CLK					118
+#define CAMSS_VFE1_CLK					119
+#define CAMSS_CPP_CLK					120
+#define CAMSS_CPP_AHB_CLK				121
+#define CAMSS_VFE_VBIF_AHB_CLK				122
+#define CAMSS_VFE_VBIF_AXI_CLK				123
+#define CAMSS_CPP_AXI_CLK				124
+#define CAMSS_CPP_VBIF_AHB_CLK				125
+#define CAMSS_CSI_VFE0_CLK				126
+#define CAMSS_CSI_VFE1_CLK				127
+#define CAMSS_VFE0_STREAM_CLK				128
+#define CAMSS_VFE1_STREAM_CLK				129
+#define CAMSS_CPHY_CSID0_CLK				130
+#define CAMSS_CPHY_CSID1_CLK				131
+#define CAMSS_CPHY_CSID2_CLK				132
+#define CAMSS_CPHY_CSID3_CLK				133
+#define CAMSS_CSIPHY0_CLK				134
+#define CAMSS_CSIPHY1_CLK				135
+#define CAMSS_CSIPHY2_CLK				136
+#define FD_CORE_CLK					137
+#define FD_CORE_UAR_CLK					138
+#define FD_AHB_CLK					139
+#define MNOC_AHB_CLK					140
+#define BIMC_SMMU_AHB_CLK				141
+#define BIMC_SMMU_AXI_CLK				142
+#define MNOC_MAXI_CLK					143
+#define VMEM_MAXI_CLK					144
+#define VMEM_AHB_CLK					145
+
+#define SPDM_BCR					0
+#define SPDM_RM_BCR					1
+#define MISC_BCR					2
+#define VIDEO_TOP_BCR					3
+#define THROTTLE_VIDEO_BCR				4
+#define MDSS_BCR					5
+#define THROTTLE_MDSS_BCR				6
+#define CAMSS_PHY0_BCR					7
+#define CAMSS_PHY1_BCR					8
+#define CAMSS_PHY2_BCR					9
+#define CAMSS_CSI0_BCR					10
+#define CAMSS_CSI0RDI_BCR				11
+#define CAMSS_CSI0PIX_BCR				12
+#define CAMSS_CSI1_BCR					13
+#define CAMSS_CSI1RDI_BCR				14
+#define CAMSS_CSI1PIX_BCR				15
+#define CAMSS_CSI2_BCR					16
+#define CAMSS_CSI2RDI_BCR				17
+#define CAMSS_CSI2PIX_BCR				18
+#define CAMSS_CSI3_BCR					19
+#define CAMSS_CSI3RDI_BCR				20
+#define CAMSS_CSI3PIX_BCR				21
+#define CAMSS_ISPIF_BCR					22
+#define CAMSS_CCI_BCR					23
+#define CAMSS_TOP_BCR					24
+#define CAMSS_AHB_BCR					25
+#define CAMSS_MICRO_BCR					26
+#define CAMSS_JPEG_BCR					27
+#define CAMSS_VFE0_BCR					28
+#define CAMSS_VFE1_BCR					29
+#define CAMSS_VFE_VBIF_BCR				30
+#define CAMSS_CPP_TOP_BCR				31
+#define CAMSS_CPP_BCR					32
+#define CAMSS_CSI_VFE0_BCR				33
+#define CAMSS_CSI_VFE1_BCR				34
+#define CAMSS_FD_BCR					35
+#define THROTTLE_CAMSS_BCR				36
+#define MNOCAHB_BCR					37
+#define MNOCAXI_BCR					38
+#define BMIC_SMMU_BCR					39
+#define MNOC_MAXI_BCR					40
+#define VMEM_BCR					41
+#define BTO_BCR						42
+
+#define VIDEO_TOP_GDSC		1
+#define VIDEO_SUBCORE0_GDSC	2
+#define VIDEO_SUBCORE1_GDSC	3
+#define MDSS_GDSC		4
+#define CAMSS_TOP_GDSC		5
+#define CAMSS_VFE0_GDSC		6
+#define CAMSS_VFE1_GDSC		7
+#define CAMSS_CPP_GDSC		8
+#define BIMC_SMMU_GDSC		9
+
+#endif
-- 
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.


^ permalink raw reply related	[flat|nested] 16+ messages in thread

* [PATCH v8 4/4] clk: qcom: Add MSM8998 Multimedia Clock Controller (MMCC) driver
  2019-11-08 23:16 [PATCH v8 0/4] MSM8998 Multimedia Clock Controller Jeffrey Hugo
                   ` (2 preceding siblings ...)
  2019-11-08 23:17 ` [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc Jeffrey Hugo
@ 2019-11-08 23:18 ` Jeffrey Hugo
  3 siblings, 0 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-08 23:18 UTC (permalink / raw)
  To: mturquette, sboyd
  Cc: agross, bjorn.andersson, marc.w.gonzalez, robh+dt, mark.rutland,
	linux-arm-msm, linux-clk, linux-kernel, devicetree, Jeffrey Hugo

Add a driver for the multimedia clock controller found on MSM8998
based devices. This should allow most multimedia device drivers
to probe and control their clocks.

Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
 drivers/clk/qcom/Kconfig        |    9 +
 drivers/clk/qcom/Makefile       |    1 +
 drivers/clk/qcom/mmcc-msm8998.c | 2913 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 2923 insertions(+)
 create mode 100644 drivers/clk/qcom/mmcc-msm8998.c

diff --git a/drivers/clk/qcom/Kconfig b/drivers/clk/qcom/Kconfig
index 3b33ef1..4b4d77b 100644
--- a/drivers/clk/qcom/Kconfig
+++ b/drivers/clk/qcom/Kconfig
@@ -229,6 +229,15 @@ config MSM_GPUCC_8998
 	  Say Y if you want to support graphics controller devices and
 	  functionality such as 3D graphics.
 
+config MSM_MMCC_8998
+	tristate "MSM8998 Multimedia Clock Controller"
+	select MSM_GCC_8998
+	select QCOM_GDSC
+	help
+	  Support for the multimedia clock controller on msm8998 devices.
+	  Say Y if you want to support multimedia devices such as display,
+	  graphics, video encode/decode, camera, etc.
+
 config QCS_GCC_404
 	tristate "QCS404 Global Clock Controller"
 	help
diff --git a/drivers/clk/qcom/Makefile b/drivers/clk/qcom/Makefile
index d899661..a4e0aa4 100644
--- a/drivers/clk/qcom/Makefile
+++ b/drivers/clk/qcom/Makefile
@@ -37,6 +37,7 @@ obj-$(CONFIG_MSM_GPUCC_8998) += gpucc-msm8998.o
 obj-$(CONFIG_MSM_MMCC_8960) += mmcc-msm8960.o
 obj-$(CONFIG_MSM_MMCC_8974) += mmcc-msm8974.o
 obj-$(CONFIG_MSM_MMCC_8996) += mmcc-msm8996.o
+obj-$(CONFIG_MSM_MMCC_8998) += mmcc-msm8998.o
 obj-$(CONFIG_QCOM_A53PLL) += a53-pll.o
 obj-$(CONFIG_QCOM_CLK_APCS_MSM8916) += apcs-msm8916.o
 obj-$(CONFIG_QCOM_CLK_RPM) += clk-rpm.o
diff --git a/drivers/clk/qcom/mmcc-msm8998.c b/drivers/clk/qcom/mmcc-msm8998.c
new file mode 100644
index 0000000..dd68983
--- /dev/null
+++ b/drivers/clk/qcom/mmcc-msm8998.c
@@ -0,0 +1,2913 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2019, The Linux Foundation. All rights reserved.
+ */
+
+#include <linux/kernel.h>
+#include <linux/bitops.h>
+#include <linux/err.h>
+#include <linux/platform_device.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/clk-provider.h>
+#include <linux/regmap.h>
+#include <linux/reset-controller.h>
+
+#include <dt-bindings/clock/qcom,mmcc-msm8998.h>
+
+#include "common.h"
+#include "clk-regmap.h"
+#include "clk-regmap-divider.h"
+#include "clk-alpha-pll.h"
+#include "clk-rcg.h"
+#include "clk-branch.h"
+#include "reset.h"
+#include "gdsc.h"
+
+enum {
+	P_XO,
+	P_GPLL0,
+	P_GPLL0_DIV,
+	P_MMPLL0_OUT_EVEN,
+	P_MMPLL1_OUT_EVEN,
+	P_MMPLL3_OUT_EVEN,
+	P_MMPLL4_OUT_EVEN,
+	P_MMPLL5_OUT_EVEN,
+	P_MMPLL6_OUT_EVEN,
+	P_MMPLL7_OUT_EVEN,
+	P_MMPLL10_OUT_EVEN,
+	P_DSI0PLL,
+	P_DSI1PLL,
+	P_DSI0PLL_BYTE,
+	P_DSI1PLL_BYTE,
+	P_HDMIPLL,
+	P_DPVCO,
+	P_DPLINK,
+	P_CORE_BI_PLL_TEST_SE,
+};
+
+static struct clk_fixed_factor gpll0_div = {
+	.mult = 1,
+	.div = 2,
+	.hw.init = &(struct clk_init_data){
+		.name = "mmss_gpll0_div",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "gpll0",
+			.name = "gpll0"
+		},
+		.num_parents = 1,
+		.ops = &clk_fixed_factor_ops,
+	},
+};
+
+static const struct clk_div_table post_div_table_fabia_even[] = {
+	{ 0x0, 1 },
+	{ 0x1, 2 },
+	{ 0x3, 4 },
+	{ 0x7, 8 },
+	{ }
+};
+
+static struct clk_alpha_pll mmpll0 = {
+	.offset = 0xc000,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr = {
+		.enable_reg = 0x1e0,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mmpll0",
+			.parent_data = &(const struct clk_parent_data){
+				.fw_name = "xo",
+				.name = "xo"
+			},
+			.num_parents = 1,
+			.ops = &clk_alpha_pll_fixed_fabia_ops,
+		},
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll0_out_even = {
+	.offset = 0xc000,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll0_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll0.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll1 = {
+	.offset = 0xc050,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr = {
+		.enable_reg = 0x1e0,
+		.enable_mask = BIT(1),
+		.hw.init = &(struct clk_init_data){
+			.name = "mmpll1",
+			.parent_data = &(const struct clk_parent_data){
+				.fw_name = "xo",
+				.name = "xo"
+			},
+			.num_parents = 1,
+			.ops = &clk_alpha_pll_fixed_fabia_ops,
+		},
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll1_out_even = {
+	.offset = 0xc050,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll1_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll1.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll3 = {
+	.offset = 0x0,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll3",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll3_out_even = {
+	.offset = 0x0,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll3_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll3.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll4 = {
+	.offset = 0x50,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll4",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll4_out_even = {
+	.offset = 0x50,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll4_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll4.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll5 = {
+	.offset = 0xa0,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll5",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll5_out_even = {
+	.offset = 0xa0,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll5_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll5.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll6 = {
+	.offset = 0xf0,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll6",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll6_out_even = {
+	.offset = 0xf0,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll6_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll6.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll7 = {
+	.offset = 0x140,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll7",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll7_out_even = {
+	.offset = 0x140,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll7_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll7.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll mmpll10 = {
+	.offset = 0x190,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll10",
+		.parent_data = &(const struct clk_parent_data){
+			.fw_name = "xo",
+			.name = "xo"
+		},
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_fixed_fabia_ops,
+	},
+};
+
+static struct clk_alpha_pll_postdiv mmpll10_out_even = {
+	.offset = 0x190,
+	.post_div_shift = 8,
+	.post_div_table = post_div_table_fabia_even,
+	.num_post_div = ARRAY_SIZE(post_div_table_fabia_even),
+	.width = 4,
+	.regs = clk_alpha_pll_regs[CLK_ALPHA_PLL_TYPE_FABIA],
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mmpll10_out_even",
+		.parent_hws = (const struct clk_hw *[]){ &mmpll10.clkr.hw },
+		.num_parents = 1,
+		.ops = &clk_alpha_pll_postdiv_fabia_ops,
+	},
+};
+
+static const struct parent_map mmss_xo_hdmi_map[] = {
+	{ P_XO, 0 },
+	{ P_HDMIPLL, 1 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_hdmi[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .fw_name = "hdmipll", .name = "hdmipll" },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_dsi0pll_dsi1pll_map[] = {
+	{ P_XO, 0 },
+	{ P_DSI0PLL, 1 },
+	{ P_DSI1PLL, 2 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_dsi0pll_dsi1pll[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .fw_name = "dsi0dsi", .name = "dsi0dsi" },
+	{ .fw_name = "dsi1dsi", .name = "dsi1dsi" },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_dsibyte_map[] = {
+	{ P_XO, 0 },
+	{ P_DSI0PLL_BYTE, 1 },
+	{ P_DSI1PLL_BYTE, 2 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_dsibyte[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .fw_name = "dsi0byte", .name = "dsi0byte" },
+	{ .fw_name = "dsi1byte", .name = "dsi1byte" },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_dp_map[] = {
+	{ P_XO, 0 },
+	{ P_DPLINK, 1 },
+	{ P_DPVCO, 2 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_dp[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .fw_name = "dplink", .name = "dplink" },
+	{ .fw_name = "dpvco", .name = "dpvco" },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_MMPLL1_OUT_EVEN, 2 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .hw = &mmpll1_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_MMPLL5_OUT_EVEN, 2 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .hw = &mmpll5_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_MMPLL3_OUT_EVEN, 3 },
+	{ P_MMPLL6_OUT_EVEN, 4 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .hw = &mmpll3_out_even.clkr.hw },
+	{ .hw = &mmpll6_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL4_OUT_EVEN, 1 },
+	{ P_MMPLL7_OUT_EVEN, 2 },
+	{ P_MMPLL10_OUT_EVEN, 3 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll4_out_even.clkr.hw },
+	{ .hw = &mmpll7_out_even.clkr.hw },
+	{ .hw = &mmpll10_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_mmpll7_mmpll10_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_MMPLL7_OUT_EVEN, 2 },
+	{ P_MMPLL10_OUT_EVEN, 3 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_mmpll7_mmpll10_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .hw = &mmpll7_out_even.clkr.hw },
+	{ .hw = &mmpll10_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static const struct parent_map mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map[] = {
+	{ P_XO, 0 },
+	{ P_MMPLL0_OUT_EVEN, 1 },
+	{ P_MMPLL4_OUT_EVEN, 2 },
+	{ P_MMPLL7_OUT_EVEN, 3 },
+	{ P_MMPLL10_OUT_EVEN, 4 },
+	{ P_GPLL0, 5 },
+	{ P_GPLL0_DIV, 6 },
+	{ P_CORE_BI_PLL_TEST_SE, 7 }
+};
+
+static const struct clk_parent_data mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div[] = {
+	{ .fw_name = "xo", .name = "xo" },
+	{ .hw = &mmpll0_out_even.clkr.hw },
+	{ .hw = &mmpll4_out_even.clkr.hw },
+	{ .hw = &mmpll7_out_even.clkr.hw },
+	{ .hw = &mmpll10_out_even.clkr.hw },
+	{ .fw_name = "gpll0", .name = "gpll0" },
+	{ .hw = &gpll0_div.hw },
+	{ .fw_name = "core_bi_pll_test_se", .name = "core_bi_pll_test_se" },
+};
+
+static struct clk_rcg2 byte0_clk_src = {
+	.cmd_rcgr = 0x2120,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsibyte_map,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "byte0_clk_src",
+		.parent_data = mmss_xo_dsibyte,
+		.num_parents = 4,
+		.ops = &clk_byte2_ops,
+		.flags = CLK_SET_RATE_PARENT,
+	},
+};
+
+static struct clk_rcg2 byte1_clk_src = {
+	.cmd_rcgr = 0x2140,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsibyte_map,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "byte1_clk_src",
+		.parent_data = mmss_xo_dsibyte,
+		.num_parents = 4,
+		.ops = &clk_byte2_ops,
+		.flags = CLK_SET_RATE_PARENT,
+	},
+};
+
+static const struct freq_tbl ftbl_cci_clk_src[] = {
+	F(37500000, P_GPLL0, 16, 0, 0),
+	F(50000000, P_GPLL0, 12, 0, 0),
+	F(100000000, P_GPLL0, 6, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 cci_clk_src = {
+	.cmd_rcgr = 0x3300,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_cci_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "cci_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_cpp_clk_src[] = {
+	F(100000000, P_GPLL0, 6, 0, 0),
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(384000000, P_MMPLL4_OUT_EVEN, 2, 0, 0),
+	F(404000000, P_MMPLL0_OUT_EVEN, 2, 0, 0),
+	F(480000000, P_MMPLL7_OUT_EVEN, 2, 0, 0),
+	F(576000000, P_MMPLL10_OUT_EVEN, 1, 0, 0),
+	F(600000000, P_GPLL0, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 cpp_clk_src = {
+	.cmd_rcgr = 0x3640,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_cpp_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "cpp_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_csi_clk_src[] = {
+	F(164571429, P_MMPLL10_OUT_EVEN, 3.5, 0, 0),
+	F(256000000, P_MMPLL4_OUT_EVEN, 3, 0, 0),
+	F(274290000, P_MMPLL7_OUT_EVEN, 3.5, 0, 0),
+	F(300000000, P_GPLL0, 2, 0, 0),
+	F(384000000, P_MMPLL4_OUT_EVEN, 2, 0, 0),
+	F(576000000, P_MMPLL10_OUT_EVEN, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 csi0_clk_src = {
+	.cmd_rcgr = 0x3090,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi0_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 csi1_clk_src = {
+	.cmd_rcgr = 0x3100,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi1_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 csi2_clk_src = {
+	.cmd_rcgr = 0x3160,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi2_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 csi3_clk_src = {
+	.cmd_rcgr = 0x31c0,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi3_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_csiphy_clk_src[] = {
+	F(164571429, P_MMPLL10_OUT_EVEN, 3.5, 0, 0),
+	F(256000000, P_MMPLL4_OUT_EVEN, 3, 0, 0),
+	F(274290000, P_MMPLL7_OUT_EVEN, 3.5, 0, 0),
+	F(300000000, P_GPLL0, 2, 0, 0),
+	F(384000000, P_MMPLL4_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 csiphy_clk_src = {
+	.cmd_rcgr = 0x3800,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csiphy_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csiphy_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_csiphytimer_clk_src[] = {
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(269333333, P_MMPLL0_OUT_EVEN, 3, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 csi0phytimer_clk_src = {
+	.cmd_rcgr = 0x3000,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csiphytimer_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi0phytimer_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 csi1phytimer_clk_src = {
+	.cmd_rcgr = 0x3030,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csiphytimer_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi1phytimer_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 csi2phytimer_clk_src = {
+	.cmd_rcgr = 0x3060,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_csiphytimer_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "csi2phytimer_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_dp_aux_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 dp_aux_clk_src = {
+	.cmd_rcgr = 0x2260,
+	.hid_width = 5,
+	.parent_map = mmss_xo_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_dp_aux_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "dp_aux_clk_src",
+		.parent_data = mmss_xo_gpll0_gpll0_div,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_dp_crypto_clk_src[] = {
+	F(101250, P_DPLINK, 1, 5, 16),
+	F(168750, P_DPLINK, 1, 5, 16),
+	F(337500, P_DPLINK, 1, 5, 16),
+	{ }
+};
+
+static struct clk_rcg2 dp_crypto_clk_src = {
+	.cmd_rcgr = 0x2220,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dp_map,
+	.freq_tbl = ftbl_dp_crypto_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "dp_crypto_clk_src",
+		.parent_data = mmss_xo_dp,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_dp_link_clk_src[] = {
+	F(162000, P_DPLINK, 2, 0, 0),
+	F(270000, P_DPLINK, 2, 0, 0),
+	F(540000, P_DPLINK, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 dp_link_clk_src = {
+	.cmd_rcgr = 0x2200,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dp_map,
+	.freq_tbl = ftbl_dp_link_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "dp_link_clk_src",
+		.parent_data = mmss_xo_dp,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_dp_pixel_clk_src[] = {
+	F(154000000, P_DPVCO, 1, 0, 0),
+	F(337500000, P_DPVCO, 2, 0, 0),
+	F(675000000, P_DPVCO, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 dp_pixel_clk_src = {
+	.cmd_rcgr = 0x2240,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dp_map,
+	.freq_tbl = ftbl_dp_pixel_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "dp_pixel_clk_src",
+		.parent_data = mmss_xo_dp,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_esc_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 esc0_clk_src = {
+	.cmd_rcgr = 0x2160,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsibyte_map,
+	.freq_tbl = ftbl_esc_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "esc0_clk_src",
+		.parent_data = mmss_xo_dsibyte,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 esc1_clk_src = {
+	.cmd_rcgr = 0x2180,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsibyte_map,
+	.freq_tbl = ftbl_esc_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "esc1_clk_src",
+		.parent_data = mmss_xo_dsibyte,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_extpclk_clk_src[] = {
+	{ .src = P_HDMIPLL },
+	{ }
+};
+
+static struct clk_rcg2 extpclk_clk_src = {
+	.cmd_rcgr = 0x2060,
+	.hid_width = 5,
+	.parent_map = mmss_xo_hdmi_map,
+	.freq_tbl = ftbl_extpclk_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "extpclk_clk_src",
+		.parent_data = mmss_xo_hdmi,
+		.num_parents = 3,
+		.ops = &clk_byte_ops,
+		.flags = CLK_SET_RATE_PARENT,
+	},
+};
+
+static const struct freq_tbl ftbl_fd_core_clk_src[] = {
+	F(100000000, P_GPLL0, 6, 0, 0),
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(404000000, P_MMPLL0_OUT_EVEN, 2, 0, 0),
+	F(480000000, P_MMPLL7_OUT_EVEN, 2, 0, 0),
+	F(576000000, P_MMPLL10_OUT_EVEN, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 fd_core_clk_src = {
+	.cmd_rcgr = 0x3b00,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_fd_core_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "fd_core_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_hdmi_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 hdmi_clk_src = {
+	.cmd_rcgr = 0x2100,
+	.hid_width = 5,
+	.parent_map = mmss_xo_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_hdmi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "hdmi_clk_src",
+		.parent_data = mmss_xo_gpll0_gpll0_div,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_jpeg0_clk_src[] = {
+	F(75000000, P_GPLL0, 8, 0, 0),
+	F(150000000, P_GPLL0, 4, 0, 0),
+	F(320000000, P_MMPLL7_OUT_EVEN, 3, 0, 0),
+	F(480000000, P_MMPLL7_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 jpeg0_clk_src = {
+	.cmd_rcgr = 0x3500,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_jpeg0_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "jpeg0_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_maxi_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	F(75000000, P_GPLL0_DIV, 4, 0, 0),
+	F(171428571, P_GPLL0, 3.5, 0, 0),
+	F(323200000, P_MMPLL0_OUT_EVEN, 2.5, 0, 0),
+	F(406000000, P_MMPLL1_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 maxi_clk_src = {
+	.cmd_rcgr = 0xf020,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_maxi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "maxi_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div,
+		.num_parents = 6,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_mclk_clk_src[] = {
+	F(4800000, P_XO, 4, 0, 0),
+	F(6000000, P_GPLL0_DIV, 10, 1, 5),
+	F(8000000, P_GPLL0_DIV, 1, 2, 75),
+	F(9600000, P_XO, 2, 0, 0),
+	F(16666667, P_GPLL0_DIV, 2, 1, 9),
+	F(19200000, P_XO, 1, 0, 0),
+	F(24000000, P_GPLL0_DIV, 1, 2, 25),
+	F(33333333, P_GPLL0_DIV, 1, 2, 9),
+	F(48000000, P_GPLL0, 1, 2, 25),
+	F(66666667, P_GPLL0, 1, 2, 9),
+	{ }
+};
+
+static struct clk_rcg2 mclk0_clk_src = {
+	.cmd_rcgr = 0x3360,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_mclk_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mclk0_clk_src",
+		.parent_data = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 mclk1_clk_src = {
+	.cmd_rcgr = 0x3390,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_mclk_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mclk1_clk_src",
+		.parent_data = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 mclk2_clk_src = {
+	.cmd_rcgr = 0x33c0,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_mclk_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mclk2_clk_src",
+		.parent_data = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 mclk3_clk_src = {
+	.cmd_rcgr = 0x33f0,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_mclk_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mclk3_clk_src",
+		.parent_data = mmss_xo_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_mdp_clk_src[] = {
+	F(85714286, P_GPLL0, 7, 0, 0),
+	F(100000000, P_GPLL0, 6, 0, 0),
+	F(150000000, P_GPLL0, 4, 0, 0),
+	F(171428571, P_GPLL0, 3.5, 0, 0),
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(275000000, P_MMPLL5_OUT_EVEN, 3, 0, 0),
+	F(300000000, P_GPLL0, 2, 0, 0),
+	F(330000000, P_MMPLL5_OUT_EVEN, 2.5, 0, 0),
+	F(412500000, P_MMPLL5_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 mdp_clk_src = {
+	.cmd_rcgr = 0x2040,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_mdp_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "mdp_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div,
+		.num_parents = 6,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_vsync_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 vsync_clk_src = {
+	.cmd_rcgr = 0x2080,
+	.hid_width = 5,
+	.parent_map = mmss_xo_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_vsync_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "vsync_clk_src",
+		.parent_data = mmss_xo_gpll0_gpll0_div,
+		.num_parents = 4,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_ahb_clk_src[] = {
+	F(19200000, P_XO, 1, 0, 0),
+	F(40000000, P_GPLL0, 15, 0, 0),
+	F(80800000, P_MMPLL0_OUT_EVEN, 10, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 ahb_clk_src = {
+	.cmd_rcgr = 0x5000,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_ahb_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "ahb_clk_src",
+		.parent_data = mmss_xo_mmpll0_gpll0_gpll0_div,
+		.num_parents = 5,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_axi_clk_src[] = {
+	F(75000000, P_GPLL0, 8, 0, 0),
+	F(171428571, P_GPLL0, 3.5, 0, 0),
+	F(240000000, P_GPLL0, 2.5, 0, 0),
+	F(323200000, P_MMPLL0_OUT_EVEN, 2.5, 0, 0),
+	F(406000000, P_MMPLL0_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+/* RO to linux */
+static struct clk_rcg2 axi_clk_src = {
+	.cmd_rcgr = 0xd000,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_axi_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "axi_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll1_gpll0_gpll0_div,
+		.num_parents = 6,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 pclk0_clk_src = {
+	.cmd_rcgr = 0x2000,
+	.mnd_width = 8,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsi0pll_dsi1pll_map,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "pclk0_clk_src",
+		.parent_data = mmss_xo_dsi0pll_dsi1pll,
+		.num_parents = 4,
+		.ops = &clk_pixel_ops,
+		.flags = CLK_SET_RATE_PARENT,
+	},
+};
+
+static struct clk_rcg2 pclk1_clk_src = {
+	.cmd_rcgr = 0x2020,
+	.mnd_width = 8,
+	.hid_width = 5,
+	.parent_map = mmss_xo_dsi0pll_dsi1pll_map,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "pclk1_clk_src",
+		.parent_data = mmss_xo_dsi0pll_dsi1pll,
+		.num_parents = 4,
+		.ops = &clk_pixel_ops,
+		.flags = CLK_SET_RATE_PARENT,
+	},
+};
+
+static const struct freq_tbl ftbl_rot_clk_src[] = {
+	F(171428571, P_GPLL0, 3.5, 0, 0),
+	F(275000000, P_MMPLL5_OUT_EVEN, 3, 0, 0),
+	F(330000000, P_MMPLL5_OUT_EVEN, 2.5, 0, 0),
+	F(412500000, P_MMPLL5_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 rot_clk_src = {
+	.cmd_rcgr = 0x21a0,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_rot_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "rot_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll5_gpll0_gpll0_div,
+		.num_parents = 6,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_video_core_clk_src[] = {
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(269330000, P_MMPLL0_OUT_EVEN, 3, 0, 0),
+	F(355200000, P_MMPLL6_OUT_EVEN, 2.5, 0, 0),
+	F(444000000, P_MMPLL6_OUT_EVEN, 2, 0, 0),
+	F(533000000, P_MMPLL3_OUT_EVEN, 2, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 video_core_clk_src = {
+	.cmd_rcgr = 0x1000,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_video_core_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "video_core_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 video_subcore0_clk_src = {
+	.cmd_rcgr = 0x1060,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_video_core_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "video_subcore0_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 video_subcore1_clk_src = {
+	.cmd_rcgr = 0x1080,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_video_core_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "video_subcore1_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll3_mmpll6_gpll0_gpll0_div,
+		.num_parents = 7,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static const struct freq_tbl ftbl_vfe_clk_src[] = {
+	F(200000000, P_GPLL0, 3, 0, 0),
+	F(300000000, P_GPLL0, 2, 0, 0),
+	F(320000000, P_MMPLL7_OUT_EVEN, 3, 0, 0),
+	F(384000000, P_MMPLL4_OUT_EVEN, 2, 0, 0),
+	F(404000000, P_MMPLL0_OUT_EVEN, 2, 0, 0),
+	F(480000000, P_MMPLL7_OUT_EVEN, 2, 0, 0),
+	F(576000000, P_MMPLL10_OUT_EVEN, 1, 0, 0),
+	F(600000000, P_GPLL0, 1, 0, 0),
+	{ }
+};
+
+static struct clk_rcg2 vfe0_clk_src = {
+	.cmd_rcgr = 0x3600,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_vfe_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "vfe0_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_rcg2 vfe1_clk_src = {
+	.cmd_rcgr = 0x3620,
+	.hid_width = 5,
+	.parent_map = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div_map,
+	.freq_tbl = ftbl_vfe_clk_src,
+	.clkr.hw.init = &(struct clk_init_data){
+		.name = "vfe1_clk_src",
+		.parent_data = mmss_xo_mmpll0_mmpll4_mmpll7_mmpll10_gpll0_gpll0_div,
+		.num_parents = 8,
+		.ops = &clk_rcg2_ops,
+	},
+};
+
+static struct clk_branch misc_ahb_clk = {
+	.halt_reg = 0x328,
+	.clkr = {
+		.enable_reg = 0x328,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "misc_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch video_core_clk = {
+	.halt_reg = 0x1028,
+	.clkr = {
+		.enable_reg = 0x1028,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_core_clk",
+			.parent_hws = (const struct clk_hw *[]){ &video_core_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch video_ahb_clk = {
+	.halt_reg = 0x1030,
+	.clkr = {
+		.enable_reg = 0x1030,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch video_axi_clk = {
+	.halt_reg = 0x1034,
+	.clkr = {
+		.enable_reg = 0x1034,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch video_maxi_clk = {
+	.halt_reg = 0x1038,
+	.clkr = {
+		.enable_reg = 0x1038,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_maxi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &maxi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch video_subcore0_clk = {
+	.halt_reg = 0x1048,
+	.clkr = {
+		.enable_reg = 0x1048,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_subcore0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &video_subcore0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch video_subcore1_clk = {
+	.halt_reg = 0x104c,
+	.clkr = {
+		.enable_reg = 0x104c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "video_subcore1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &video_subcore1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_ahb_clk = {
+	.halt_reg = 0x2308,
+	.clkr = {
+		.enable_reg = 0x2308,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_hdmi_dp_ahb_clk = {
+	.halt_reg = 0x230c,
+	.clkr = {
+		.enable_reg = 0x230c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_hdmi_dp_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_axi_clk = {
+	.halt_reg = 0x2310,
+	.clkr = {
+		.enable_reg = 0x2310,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch mdss_pclk0_clk = {
+	.halt_reg = 0x2314,
+	.clkr = {
+		.enable_reg = 0x2314,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_pclk0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &pclk0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_pclk1_clk = {
+	.halt_reg = 0x2318,
+	.clkr = {
+		.enable_reg = 0x2318,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_pclk1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &pclk1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_mdp_clk = {
+	.halt_reg = 0x231c,
+	.clkr = {
+		.enable_reg = 0x231c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_mdp_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mdp_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_mdp_lut_clk = {
+	.halt_reg = 0x2320,
+	.clkr = {
+		.enable_reg = 0x2320,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_mdp_lut_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mdp_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_extpclk_clk = {
+	.halt_reg = 0x2324,
+	.clkr = {
+		.enable_reg = 0x2324,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_extpclk_clk",
+			.parent_hws = (const struct clk_hw *[]){ &extpclk_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_vsync_clk = {
+	.halt_reg = 0x2328,
+	.clkr = {
+		.enable_reg = 0x2328,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_vsync_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vsync_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_hdmi_clk = {
+	.halt_reg = 0x2338,
+	.clkr = {
+		.enable_reg = 0x2338,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_hdmi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &hdmi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_byte0_clk = {
+	.halt_reg = 0x233c,
+	.clkr = {
+		.enable_reg = 0x233c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_byte0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &byte0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_byte1_clk = {
+	.halt_reg = 0x2340,
+	.clkr = {
+		.enable_reg = 0x2340,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_byte1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &byte1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_esc0_clk = {
+	.halt_reg = 0x2344,
+	.clkr = {
+		.enable_reg = 0x2344,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_esc0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &esc0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_esc1_clk = {
+	.halt_reg = 0x2348,
+	.clkr = {
+		.enable_reg = 0x2348,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_esc1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &esc1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_rot_clk = {
+	.halt_reg = 0x2350,
+	.clkr = {
+		.enable_reg = 0x2350,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_rot_clk",
+			.parent_hws = (const struct clk_hw *[]){ &rot_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_dp_link_clk = {
+	.halt_reg = 0x2354,
+	.clkr = {
+		.enable_reg = 0x2354,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_dp_link_clk",
+			.parent_hws = (const struct clk_hw *[]){ &dp_link_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_dp_link_intf_clk = {
+	.halt_reg = 0x2358,
+	.clkr = {
+		.enable_reg = 0x2358,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_dp_link_intf_clk",
+			.parent_hws = (const struct clk_hw *[]){ &dp_link_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_dp_crypto_clk = {
+	.halt_reg = 0x235c,
+	.clkr = {
+		.enable_reg = 0x235c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_dp_crypto_clk",
+			.parent_hws = (const struct clk_hw *[]){ &dp_crypto_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_dp_pixel_clk = {
+	.halt_reg = 0x2360,
+	.clkr = {
+		.enable_reg = 0x2360,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_dp_pixel_clk",
+			.parent_hws = (const struct clk_hw *[]){ &dp_pixel_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_dp_aux_clk = {
+	.halt_reg = 0x2364,
+	.clkr = {
+		.enable_reg = 0x2364,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_dp_aux_clk",
+			.parent_hws = (const struct clk_hw *[]){ &dp_aux_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_byte0_intf_clk = {
+	.halt_reg = 0x2374,
+	.clkr = {
+		.enable_reg = 0x2374,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_byte0_intf_clk",
+			.parent_hws = (const struct clk_hw *[]){ &byte0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mdss_byte1_intf_clk = {
+	.halt_reg = 0x2378,
+	.clkr = {
+		.enable_reg = 0x2378,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mdss_byte1_intf_clk",
+			.parent_hws = (const struct clk_hw *[]){ &byte1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi0phytimer_clk = {
+	.halt_reg = 0x3024,
+	.clkr = {
+		.enable_reg = 0x3024,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi0phytimer_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi0phytimer_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi1phytimer_clk = {
+	.halt_reg = 0x3054,
+	.clkr = {
+		.enable_reg = 0x3054,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi1phytimer_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi1phytimer_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi2phytimer_clk = {
+	.halt_reg = 0x3084,
+	.clkr = {
+		.enable_reg = 0x3084,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi2phytimer_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi2phytimer_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi0_clk = {
+	.halt_reg = 0x30b4,
+	.clkr = {
+		.enable_reg = 0x30b4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi0_ahb_clk = {
+	.halt_reg = 0x30bc,
+	.clkr = {
+		.enable_reg = 0x30bc,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi0_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi0rdi_clk = {
+	.halt_reg = 0x30d4,
+	.clkr = {
+		.enable_reg = 0x30d4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi0rdi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi0pix_clk = {
+	.halt_reg = 0x30e4,
+	.clkr = {
+		.enable_reg = 0x30e4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi0pix_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi1_clk = {
+	.halt_reg = 0x3124,
+	.clkr = {
+		.enable_reg = 0x3124,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi1_ahb_clk = {
+	.halt_reg = 0x3128,
+	.clkr = {
+		.enable_reg = 0x3128,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi1_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi1rdi_clk = {
+	.halt_reg = 0x3144,
+	.clkr = {
+		.enable_reg = 0x3144,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi1rdi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi1pix_clk = {
+	.halt_reg = 0x3154,
+	.clkr = {
+		.enable_reg = 0x3154,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi1pix_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi2_clk = {
+	.halt_reg = 0x3184,
+	.clkr = {
+		.enable_reg = 0x3184,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi2_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi2_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi2_ahb_clk = {
+	.halt_reg = 0x3188,
+	.clkr = {
+		.enable_reg = 0x3188,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi2_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi2rdi_clk = {
+	.halt_reg = 0x31a4,
+	.clkr = {
+		.enable_reg = 0x31a4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi2rdi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi2_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi2pix_clk = {
+	.halt_reg = 0x31b4,
+	.clkr = {
+		.enable_reg = 0x31b4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi2pix_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi2_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi3_clk = {
+	.halt_reg = 0x31e4,
+	.clkr = {
+		.enable_reg = 0x31e4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi3_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi3_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi3_ahb_clk = {
+	.halt_reg = 0x31e8,
+	.clkr = {
+		.enable_reg = 0x31e8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi3_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi3rdi_clk = {
+	.halt_reg = 0x3204,
+	.clkr = {
+		.enable_reg = 0x3204,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi3rdi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi3_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi3pix_clk = {
+	.halt_reg = 0x3214,
+	.clkr = {
+		.enable_reg = 0x3214,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi3pix_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csi3_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_ispif_ahb_clk = {
+	.halt_reg = 0x3224,
+	.clkr = {
+		.enable_reg = 0x3224,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_ispif_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cci_clk = {
+	.halt_reg = 0x3344,
+	.clkr = {
+		.enable_reg = 0x3344,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cci_clk",
+			.parent_hws = (const struct clk_hw *[]){ &cci_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cci_ahb_clk = {
+	.halt_reg = 0x3348,
+	.clkr = {
+		.enable_reg = 0x3348,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cci_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_mclk0_clk = {
+	.halt_reg = 0x3384,
+	.clkr = {
+		.enable_reg = 0x3384,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_mclk0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mclk0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_mclk1_clk = {
+	.halt_reg = 0x33b4,
+	.clkr = {
+		.enable_reg = 0x33b4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_mclk1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mclk1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_mclk2_clk = {
+	.halt_reg = 0x33e4,
+	.clkr = {
+		.enable_reg = 0x33e4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_mclk2_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mclk2_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_mclk3_clk = {
+	.halt_reg = 0x3414,
+	.clkr = {
+		.enable_reg = 0x3414,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_mclk3_clk",
+			.parent_hws = (const struct clk_hw *[]){ &mclk3_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_top_ahb_clk = {
+	.halt_reg = 0x3484,
+	.clkr = {
+		.enable_reg = 0x3484,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_top_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_ahb_clk = {
+	.halt_reg = 0x348c,
+	.clkr = {
+		.enable_reg = 0x348c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_micro_ahb_clk = {
+	.halt_reg = 0x3494,
+	.clkr = {
+		.enable_reg = 0x3494,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_micro_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_jpeg0_clk = {
+	.halt_reg = 0x35a8,
+	.clkr = {
+		.enable_reg = 0x35a8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_jpeg0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &jpeg0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_jpeg_ahb_clk = {
+	.halt_reg = 0x35b4,
+	.clkr = {
+		.enable_reg = 0x35b4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_jpeg_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_jpeg_axi_clk = {
+	.halt_reg = 0x35b8,
+	.clkr = {
+		.enable_reg = 0x35b8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_jpeg_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe0_ahb_clk = {
+	.halt_reg = 0x3668,
+	.clkr = {
+		.enable_reg = 0x3668,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe0_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe1_ahb_clk = {
+	.halt_reg = 0x3678,
+	.clkr = {
+		.enable_reg = 0x3678,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe1_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe0_clk = {
+	.halt_reg = 0x36a8,
+	.clkr = {
+		.enable_reg = 0x36a8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe1_clk = {
+	.halt_reg = 0x36ac,
+	.clkr = {
+		.enable_reg = 0x36ac,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cpp_clk = {
+	.halt_reg = 0x36b0,
+	.clkr = {
+		.enable_reg = 0x36b0,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cpp_clk",
+			.parent_hws = (const struct clk_hw *[]){ &cpp_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cpp_ahb_clk = {
+	.halt_reg = 0x36b4,
+	.clkr = {
+		.enable_reg = 0x36b4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cpp_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe_vbif_ahb_clk = {
+	.halt_reg = 0x36b8,
+	.clkr = {
+		.enable_reg = 0x36b8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe_vbif_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe_vbif_axi_clk = {
+	.halt_reg = 0x36bc,
+	.clkr = {
+		.enable_reg = 0x36bc,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe_vbif_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch camss_cpp_axi_clk = {
+	.halt_reg = 0x36c4,
+	.clkr = {
+		.enable_reg = 0x36c4,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cpp_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch camss_cpp_vbif_ahb_clk = {
+	.halt_reg = 0x36c8,
+	.clkr = {
+		.enable_reg = 0x36c8,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cpp_vbif_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi_vfe0_clk = {
+	.halt_reg = 0x3704,
+	.clkr = {
+		.enable_reg = 0x3704,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi_vfe0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csi_vfe1_clk = {
+	.halt_reg = 0x3714,
+	.clkr = {
+		.enable_reg = 0x3714,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csi_vfe1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe0_stream_clk = {
+	.halt_reg = 0x3720,
+	.clkr = {
+		.enable_reg = 0x3720,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe0_stream_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe0_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_vfe1_stream_clk = {
+	.halt_reg = 0x3724,
+	.clkr = {
+		.enable_reg = 0x3724,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_vfe1_stream_clk",
+			.parent_hws = (const struct clk_hw *[]){ &vfe1_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cphy_csid0_clk = {
+	.halt_reg = 0x3730,
+	.clkr = {
+		.enable_reg = 0x3730,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cphy_csid0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cphy_csid1_clk = {
+	.halt_reg = 0x3734,
+	.clkr = {
+		.enable_reg = 0x3734,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cphy_csid1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cphy_csid2_clk = {
+	.halt_reg = 0x3738,
+	.clkr = {
+		.enable_reg = 0x3738,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cphy_csid2_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_cphy_csid3_clk = {
+	.halt_reg = 0x373c,
+	.clkr = {
+		.enable_reg = 0x373c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_cphy_csid3_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csiphy0_clk = {
+	.halt_reg = 0x3740,
+	.clkr = {
+		.enable_reg = 0x3740,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csiphy0_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csiphy1_clk = {
+	.halt_reg = 0x3744,
+	.clkr = {
+		.enable_reg = 0x3744,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csiphy1_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch camss_csiphy2_clk = {
+	.halt_reg = 0x3748,
+	.clkr = {
+		.enable_reg = 0x3748,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "camss_csiphy2_clk",
+			.parent_hws = (const struct clk_hw *[]){ &csiphy_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch fd_core_clk = {
+	.halt_reg = 0x3b68,
+	.clkr = {
+		.enable_reg = 0x3b68,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "fd_core_clk",
+			.parent_hws = (const struct clk_hw *[]){ &fd_core_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch fd_core_uar_clk = {
+	.halt_reg = 0x3b6c,
+	.clkr = {
+		.enable_reg = 0x3b6c,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "fd_core_uar_clk",
+			.parent_hws = (const struct clk_hw *[]){ &fd_core_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch fd_ahb_clk = {
+	.halt_reg = 0x3b74,
+	.clkr = {
+		.enable_reg = 0x3b74,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "fd_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch mnoc_ahb_clk = {
+	.halt_reg = 0x5024,
+	.clkr = {
+		.enable_reg = 0x5024,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mnoc_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch bimc_smmu_ahb_clk = {
+	.halt_reg = 0xe004,
+	.clkr = {
+		.enable_reg = 0xe004,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "bimc_smmu_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch bimc_smmu_axi_clk = {
+	.halt_reg = 0xe008,
+	.clkr = {
+		.enable_reg = 0xe008,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "bimc_smmu_axi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &axi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+		},
+	},
+};
+
+static struct clk_branch mnoc_maxi_clk = {
+	.halt_reg = 0xf004,
+	.clkr = {
+		.enable_reg = 0xf004,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "mnoc_maxi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &maxi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch vmem_maxi_clk = {
+	.halt_reg = 0xf064,
+	.clkr = {
+		.enable_reg = 0xf064,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "vmem_maxi_clk",
+			.parent_hws = (const struct clk_hw *[]){ &maxi_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_branch vmem_ahb_clk = {
+	.halt_reg = 0xf068,
+	.clkr = {
+		.enable_reg = 0xf068,
+		.enable_mask = BIT(0),
+		.hw.init = &(struct clk_init_data){
+			.name = "vmem_ahb_clk",
+			.parent_hws = (const struct clk_hw *[]){ &ahb_clk_src.clkr.hw },
+			.num_parents = 1,
+			.ops = &clk_branch2_ops,
+			.flags = CLK_SET_RATE_PARENT,
+		},
+	},
+};
+
+static struct clk_hw *mmcc_msm8998_hws[] = {
+	&gpll0_div.hw,
+};
+
+static struct gdsc video_top_gdsc = {
+	.gdscr = 0x1024,
+	.pd = {
+		.name = "video_top",
+	},
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc video_subcore0_gdsc = {
+	.gdscr = 0x1040,
+	.pd = {
+		.name = "video_subcore0",
+	},
+	.parent = &video_top_gdsc.pd,
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc video_subcore1_gdsc = {
+	.gdscr = 0x1044,
+	.pd = {
+		.name = "video_subcore1",
+	},
+	.parent = &video_top_gdsc.pd,
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc mdss_gdsc = {
+	.gdscr = 0x2304,
+	.cxcs = (unsigned int []){ 0x2310, 0x2350, 0x231c, 0x2320 },
+	.cxc_count = 4,
+	.pd = {
+		.name = "mdss",
+	},
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc camss_top_gdsc = {
+	.gdscr = 0x34a0,
+	.cxcs = (unsigned int []){ 0x35b8, 0x36c4, 0x3704, 0x3714, 0x3494,
+				   0x35a8, 0x3868 },
+	.cxc_count = 7,
+	.pd = {
+		.name = "camss_top",
+	},
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc camss_vfe0_gdsc = {
+	.gdscr = 0x3664,
+	.pd = {
+		.name = "camss_vfe0",
+	},
+	.parent = &camss_top_gdsc.pd,
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc camss_vfe1_gdsc = {
+	.gdscr = 0x3674,
+	.pd = {
+		.name = "camss_vfe1_gdsc",
+	},
+	.parent = &camss_top_gdsc.pd,
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc camss_cpp_gdsc = {
+	.gdscr = 0x36d4,
+	.pd = {
+		.name = "camss_cpp",
+	},
+	.parent = &camss_top_gdsc.pd,
+	.pwrsts = PWRSTS_OFF_ON,
+};
+
+static struct gdsc bimc_smmu_gdsc = {
+	.gdscr = 0xe020,
+	.gds_hw_ctrl = 0xe024,
+	.pd = {
+		.name = "bimc_smmu",
+	},
+	.pwrsts = PWRSTS_OFF_ON,
+	.flags = HW_CTRL,
+};
+
+static struct clk_regmap *mmcc_msm8998_clocks[] = {
+	[MMPLL0] = &mmpll0.clkr,
+	[MMPLL0_OUT_EVEN] = &mmpll0_out_even.clkr,
+	[MMPLL1] = &mmpll1.clkr,
+	[MMPLL1_OUT_EVEN] = &mmpll1_out_even.clkr,
+	[MMPLL3] = &mmpll3.clkr,
+	[MMPLL3_OUT_EVEN] = &mmpll3_out_even.clkr,
+	[MMPLL4] = &mmpll4.clkr,
+	[MMPLL4_OUT_EVEN] = &mmpll4_out_even.clkr,
+	[MMPLL5] = &mmpll5.clkr,
+	[MMPLL5_OUT_EVEN] = &mmpll5_out_even.clkr,
+	[MMPLL6] = &mmpll6.clkr,
+	[MMPLL6_OUT_EVEN] = &mmpll6_out_even.clkr,
+	[MMPLL7] = &mmpll7.clkr,
+	[MMPLL7_OUT_EVEN] = &mmpll7_out_even.clkr,
+	[MMPLL10] = &mmpll10.clkr,
+	[MMPLL10_OUT_EVEN] = &mmpll10_out_even.clkr,
+	[BYTE0_CLK_SRC] = &byte0_clk_src.clkr,
+	[BYTE1_CLK_SRC] = &byte1_clk_src.clkr,
+	[CCI_CLK_SRC] = &cci_clk_src.clkr,
+	[CPP_CLK_SRC] = &cpp_clk_src.clkr,
+	[CSI0_CLK_SRC] = &csi0_clk_src.clkr,
+	[CSI1_CLK_SRC] = &csi1_clk_src.clkr,
+	[CSI2_CLK_SRC] = &csi2_clk_src.clkr,
+	[CSI3_CLK_SRC] = &csi3_clk_src.clkr,
+	[CSIPHY_CLK_SRC] = &csiphy_clk_src.clkr,
+	[CSI0PHYTIMER_CLK_SRC] = &csi0phytimer_clk_src.clkr,
+	[CSI1PHYTIMER_CLK_SRC] = &csi1phytimer_clk_src.clkr,
+	[CSI2PHYTIMER_CLK_SRC] = &csi2phytimer_clk_src.clkr,
+	[DP_AUX_CLK_SRC] = &dp_aux_clk_src.clkr,
+	[DP_CRYPTO_CLK_SRC] = &dp_crypto_clk_src.clkr,
+	[DP_LINK_CLK_SRC] = &dp_link_clk_src.clkr,
+	[DP_PIXEL_CLK_SRC] = &dp_pixel_clk_src.clkr,
+	[ESC0_CLK_SRC] = &esc0_clk_src.clkr,
+	[ESC1_CLK_SRC] = &esc1_clk_src.clkr,
+	[EXTPCLK_CLK_SRC] = &extpclk_clk_src.clkr,
+	[FD_CORE_CLK_SRC] = &fd_core_clk_src.clkr,
+	[HDMI_CLK_SRC] = &hdmi_clk_src.clkr,
+	[JPEG0_CLK_SRC] = &jpeg0_clk_src.clkr,
+	[MAXI_CLK_SRC] = &maxi_clk_src.clkr,
+	[MCLK0_CLK_SRC] = &mclk0_clk_src.clkr,
+	[MCLK1_CLK_SRC] = &mclk1_clk_src.clkr,
+	[MCLK2_CLK_SRC] = &mclk2_clk_src.clkr,
+	[MCLK3_CLK_SRC] = &mclk3_clk_src.clkr,
+	[MDP_CLK_SRC] = &mdp_clk_src.clkr,
+	[VSYNC_CLK_SRC] = &vsync_clk_src.clkr,
+	[AHB_CLK_SRC] = &ahb_clk_src.clkr,
+	[AXI_CLK_SRC] = &axi_clk_src.clkr,
+	[PCLK0_CLK_SRC] = &pclk0_clk_src.clkr,
+	[PCLK1_CLK_SRC] = &pclk1_clk_src.clkr,
+	[ROT_CLK_SRC] = &rot_clk_src.clkr,
+	[VIDEO_CORE_CLK_SRC] = &video_core_clk_src.clkr,
+	[VIDEO_SUBCORE0_CLK_SRC] = &video_subcore0_clk_src.clkr,
+	[VIDEO_SUBCORE1_CLK_SRC] = &video_subcore1_clk_src.clkr,
+	[VFE0_CLK_SRC] = &vfe0_clk_src.clkr,
+	[VFE1_CLK_SRC] = &vfe1_clk_src.clkr,
+	[MISC_AHB_CLK] = &misc_ahb_clk.clkr,
+	[VIDEO_CORE_CLK] = &video_core_clk.clkr,
+	[VIDEO_AHB_CLK] = &video_ahb_clk.clkr,
+	[VIDEO_AXI_CLK] = &video_axi_clk.clkr,
+	[VIDEO_MAXI_CLK] = &video_maxi_clk.clkr,
+	[VIDEO_SUBCORE0_CLK] = &video_subcore0_clk.clkr,
+	[VIDEO_SUBCORE1_CLK] = &video_subcore1_clk.clkr,
+	[MDSS_AHB_CLK] = &mdss_ahb_clk.clkr,
+	[MDSS_HDMI_DP_AHB_CLK] = &mdss_hdmi_dp_ahb_clk.clkr,
+	[MDSS_AXI_CLK] = &mdss_axi_clk.clkr,
+	[MDSS_PCLK0_CLK] = &mdss_pclk0_clk.clkr,
+	[MDSS_PCLK1_CLK] = &mdss_pclk1_clk.clkr,
+	[MDSS_MDP_CLK] = &mdss_mdp_clk.clkr,
+	[MDSS_MDP_LUT_CLK] = &mdss_mdp_lut_clk.clkr,
+	[MDSS_EXTPCLK_CLK] = &mdss_extpclk_clk.clkr,
+	[MDSS_VSYNC_CLK] = &mdss_vsync_clk.clkr,
+	[MDSS_HDMI_CLK] = &mdss_hdmi_clk.clkr,
+	[MDSS_BYTE0_CLK] = &mdss_byte0_clk.clkr,
+	[MDSS_BYTE1_CLK] = &mdss_byte1_clk.clkr,
+	[MDSS_ESC0_CLK] = &mdss_esc0_clk.clkr,
+	[MDSS_ESC1_CLK] = &mdss_esc1_clk.clkr,
+	[MDSS_ROT_CLK] = &mdss_rot_clk.clkr,
+	[MDSS_DP_LINK_CLK] = &mdss_dp_link_clk.clkr,
+	[MDSS_DP_LINK_INTF_CLK] = &mdss_dp_link_intf_clk.clkr,
+	[MDSS_DP_CRYPTO_CLK] = &mdss_dp_crypto_clk.clkr,
+	[MDSS_DP_PIXEL_CLK] = &mdss_dp_pixel_clk.clkr,
+	[MDSS_DP_AUX_CLK] = &mdss_dp_aux_clk.clkr,
+	[MDSS_BYTE0_INTF_CLK] = &mdss_byte0_intf_clk.clkr,
+	[MDSS_BYTE1_INTF_CLK] = &mdss_byte1_intf_clk.clkr,
+	[CAMSS_CSI0PHYTIMER_CLK] = &camss_csi0phytimer_clk.clkr,
+	[CAMSS_CSI1PHYTIMER_CLK] = &camss_csi1phytimer_clk.clkr,
+	[CAMSS_CSI2PHYTIMER_CLK] = &camss_csi2phytimer_clk.clkr,
+	[CAMSS_CSI0_CLK] = &camss_csi0_clk.clkr,
+	[CAMSS_CSI0_AHB_CLK] = &camss_csi0_ahb_clk.clkr,
+	[CAMSS_CSI0RDI_CLK] = &camss_csi0rdi_clk.clkr,
+	[CAMSS_CSI0PIX_CLK] = &camss_csi0pix_clk.clkr,
+	[CAMSS_CSI1_CLK] = &camss_csi1_clk.clkr,
+	[CAMSS_CSI1_AHB_CLK] = &camss_csi1_ahb_clk.clkr,
+	[CAMSS_CSI1RDI_CLK] = &camss_csi1rdi_clk.clkr,
+	[CAMSS_CSI1PIX_CLK] = &camss_csi1pix_clk.clkr,
+	[CAMSS_CSI2_CLK] = &camss_csi2_clk.clkr,
+	[CAMSS_CSI2_AHB_CLK] = &camss_csi2_ahb_clk.clkr,
+	[CAMSS_CSI2RDI_CLK] = &camss_csi2rdi_clk.clkr,
+	[CAMSS_CSI2PIX_CLK] = &camss_csi2pix_clk.clkr,
+	[CAMSS_CSI3_CLK] = &camss_csi3_clk.clkr,
+	[CAMSS_CSI3_AHB_CLK] = &camss_csi3_ahb_clk.clkr,
+	[CAMSS_CSI3RDI_CLK] = &camss_csi3rdi_clk.clkr,
+	[CAMSS_CSI3PIX_CLK] = &camss_csi3pix_clk.clkr,
+	[CAMSS_ISPIF_AHB_CLK] = &camss_ispif_ahb_clk.clkr,
+	[CAMSS_CCI_CLK] = &camss_cci_clk.clkr,
+	[CAMSS_CCI_AHB_CLK] = &camss_cci_ahb_clk.clkr,
+	[CAMSS_MCLK0_CLK] = &camss_mclk0_clk.clkr,
+	[CAMSS_MCLK1_CLK] = &camss_mclk1_clk.clkr,
+	[CAMSS_MCLK2_CLK] = &camss_mclk2_clk.clkr,
+	[CAMSS_MCLK3_CLK] = &camss_mclk3_clk.clkr,
+	[CAMSS_TOP_AHB_CLK] = &camss_top_ahb_clk.clkr,
+	[CAMSS_AHB_CLK] = &camss_ahb_clk.clkr,
+	[CAMSS_MICRO_AHB_CLK] = &camss_micro_ahb_clk.clkr,
+	[CAMSS_JPEG0_CLK] = &camss_jpeg0_clk.clkr,
+	[CAMSS_JPEG_AHB_CLK] = &camss_jpeg_ahb_clk.clkr,
+	[CAMSS_JPEG_AXI_CLK] = &camss_jpeg_axi_clk.clkr,
+	[CAMSS_VFE0_AHB_CLK] = &camss_vfe0_ahb_clk.clkr,
+	[CAMSS_VFE1_AHB_CLK] = &camss_vfe1_ahb_clk.clkr,
+	[CAMSS_VFE0_CLK] = &camss_vfe0_clk.clkr,
+	[CAMSS_VFE1_CLK] = &camss_vfe1_clk.clkr,
+	[CAMSS_CPP_CLK] = &camss_cpp_clk.clkr,
+	[CAMSS_CPP_AHB_CLK] = &camss_cpp_ahb_clk.clkr,
+	[CAMSS_VFE_VBIF_AHB_CLK] = &camss_vfe_vbif_ahb_clk.clkr,
+	[CAMSS_VFE_VBIF_AXI_CLK] = &camss_vfe_vbif_axi_clk.clkr,
+	[CAMSS_CPP_AXI_CLK] = &camss_cpp_axi_clk.clkr,
+	[CAMSS_CPP_VBIF_AHB_CLK] = &camss_cpp_vbif_ahb_clk.clkr,
+	[CAMSS_CSI_VFE0_CLK] = &camss_csi_vfe0_clk.clkr,
+	[CAMSS_CSI_VFE1_CLK] = &camss_csi_vfe1_clk.clkr,
+	[CAMSS_VFE0_STREAM_CLK] = &camss_vfe0_stream_clk.clkr,
+	[CAMSS_VFE1_STREAM_CLK] = &camss_vfe1_stream_clk.clkr,
+	[CAMSS_CPHY_CSID0_CLK] = &camss_cphy_csid0_clk.clkr,
+	[CAMSS_CPHY_CSID1_CLK] = &camss_cphy_csid1_clk.clkr,
+	[CAMSS_CPHY_CSID2_CLK] = &camss_cphy_csid2_clk.clkr,
+	[CAMSS_CPHY_CSID3_CLK] = &camss_cphy_csid3_clk.clkr,
+	[CAMSS_CSIPHY0_CLK] = &camss_csiphy0_clk.clkr,
+	[CAMSS_CSIPHY1_CLK] = &camss_csiphy1_clk.clkr,
+	[CAMSS_CSIPHY2_CLK] = &camss_csiphy2_clk.clkr,
+	[FD_CORE_CLK] = &fd_core_clk.clkr,
+	[FD_CORE_UAR_CLK] = &fd_core_uar_clk.clkr,
+	[FD_AHB_CLK] = &fd_ahb_clk.clkr,
+	[MNOC_AHB_CLK] = &mnoc_ahb_clk.clkr,
+	[BIMC_SMMU_AHB_CLK] = &bimc_smmu_ahb_clk.clkr,
+	[BIMC_SMMU_AXI_CLK] = &bimc_smmu_axi_clk.clkr,
+	[MNOC_MAXI_CLK] = &mnoc_maxi_clk.clkr,
+	[VMEM_MAXI_CLK] = &vmem_maxi_clk.clkr,
+	[VMEM_AHB_CLK] = &vmem_ahb_clk.clkr,
+};
+
+static struct gdsc *mmcc_msm8998_gdscs[] = {
+	[VIDEO_TOP_GDSC] = &video_top_gdsc,
+	[VIDEO_SUBCORE0_GDSC] = &video_subcore0_gdsc,
+	[VIDEO_SUBCORE1_GDSC] = &video_subcore1_gdsc,
+	[MDSS_GDSC] = &mdss_gdsc,
+	[CAMSS_TOP_GDSC] = &camss_top_gdsc,
+	[CAMSS_VFE0_GDSC] = &camss_vfe0_gdsc,
+	[CAMSS_VFE1_GDSC] = &camss_vfe1_gdsc,
+	[CAMSS_CPP_GDSC] = &camss_cpp_gdsc,
+	[BIMC_SMMU_GDSC] = &bimc_smmu_gdsc,
+};
+
+static const struct qcom_reset_map mmcc_msm8998_resets[] = {
+	[SPDM_BCR] = { 0x200 },
+	[SPDM_RM_BCR] = { 0x300 },
+	[MISC_BCR] = { 0x320 },
+	[VIDEO_TOP_BCR] = { 0x1020 },
+	[THROTTLE_VIDEO_BCR] = { 0x1180 },
+	[MDSS_BCR] = { 0x2300 },
+	[THROTTLE_MDSS_BCR] = { 0x2460 },
+	[CAMSS_PHY0_BCR] = { 0x3020 },
+	[CAMSS_PHY1_BCR] = { 0x3050 },
+	[CAMSS_PHY2_BCR] = { 0x3080 },
+	[CAMSS_CSI0_BCR] = { 0x30b0 },
+	[CAMSS_CSI0RDI_BCR] = { 0x30d0 },
+	[CAMSS_CSI0PIX_BCR] = { 0x30e0 },
+	[CAMSS_CSI1_BCR] = { 0x3120 },
+	[CAMSS_CSI1RDI_BCR] = { 0x3140 },
+	[CAMSS_CSI1PIX_BCR] = { 0x3150 },
+	[CAMSS_CSI2_BCR] = { 0x3180 },
+	[CAMSS_CSI2RDI_BCR] = { 0x31a0 },
+	[CAMSS_CSI2PIX_BCR] = { 0x31b0 },
+	[CAMSS_CSI3_BCR] = { 0x31e0 },
+	[CAMSS_CSI3RDI_BCR] = { 0x3200 },
+	[CAMSS_CSI3PIX_BCR] = { 0x3210 },
+	[CAMSS_ISPIF_BCR] = { 0x3220 },
+	[CAMSS_CCI_BCR] = { 0x3340 },
+	[CAMSS_TOP_BCR] = { 0x3480 },
+	[CAMSS_AHB_BCR] = { 0x3488 },
+	[CAMSS_MICRO_BCR] = { 0x3490 },
+	[CAMSS_JPEG_BCR] = { 0x35a0 },
+	[CAMSS_VFE0_BCR] = { 0x3660 },
+	[CAMSS_VFE1_BCR] = { 0x3670 },
+	[CAMSS_VFE_VBIF_BCR] = { 0x36a0 },
+	[CAMSS_CPP_TOP_BCR] = { 0x36c0 },
+	[CAMSS_CPP_BCR] = { 0x36d0 },
+	[CAMSS_CSI_VFE0_BCR] = { 0x3700 },
+	[CAMSS_CSI_VFE1_BCR] = { 0x3710 },
+	[CAMSS_FD_BCR] = { 0x3b60 },
+	[THROTTLE_CAMSS_BCR] = { 0x3c30 },
+	[MNOCAHB_BCR] = { 0x5020 },
+	[MNOCAXI_BCR] = { 0xd020 },
+	[BMIC_SMMU_BCR] = { 0xe000 },
+	[MNOC_MAXI_BCR] = { 0xf000 },
+	[VMEM_BCR] = { 0xf060 },
+	[BTO_BCR] = { 0x10004 },
+};
+
+static const struct regmap_config mmcc_msm8998_regmap_config = {
+	.reg_bits	= 32,
+	.reg_stride	= 4,
+	.val_bits	= 32,
+	.max_register	= 0x10004,
+	.fast_io	= true,
+};
+
+static const struct qcom_cc_desc mmcc_msm8998_desc = {
+	.config = &mmcc_msm8998_regmap_config,
+	.clks = mmcc_msm8998_clocks,
+	.num_clks = ARRAY_SIZE(mmcc_msm8998_clocks),
+	.resets = mmcc_msm8998_resets,
+	.num_resets = ARRAY_SIZE(mmcc_msm8998_resets),
+	.gdscs = mmcc_msm8998_gdscs,
+	.num_gdscs = ARRAY_SIZE(mmcc_msm8998_gdscs),
+	.clk_hws = mmcc_msm8998_hws,
+	.num_clk_hws = ARRAY_SIZE(mmcc_msm8998_hws),
+};
+
+static const struct of_device_id mmcc_msm8998_match_table[] = {
+	{ .compatible = "qcom,mmcc-msm8998" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, mmcc_msm8998_match_table);
+
+static int mmcc_msm8998_probe(struct platform_device *pdev)
+{
+	struct regmap *regmap;
+
+	regmap = qcom_cc_map(pdev, &mmcc_msm8998_desc);
+	if (IS_ERR(regmap))
+		return PTR_ERR(regmap);
+
+	return qcom_cc_really_probe(pdev, &mmcc_msm8998_desc, regmap);
+}
+
+static struct platform_driver mmcc_msm8998_driver = {
+	.probe		= mmcc_msm8998_probe,
+	.driver		= {
+		.name	= "mmcc-msm8998",
+		.of_match_table = mmcc_msm8998_match_table,
+	},
+};
+module_platform_driver(mmcc_msm8998_driver);
+
+MODULE_DESCRIPTION("QCOM MMCC MSM8998 Driver");
+MODULE_LICENSE("GPL v2");
-- 
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.


^ permalink raw reply related	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-08 23:17 ` [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc Jeffrey Hugo
@ 2019-11-12  0:44   ` Rob Herring
  2019-11-12 16:25     ` Jeffrey Hugo
  0 siblings, 1 reply; 16+ messages in thread
From: Rob Herring @ 2019-11-12  0:44 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: mturquette, sboyd, mark.rutland, agross, bjorn.andersson,
	marc.w.gonzalez, linux-arm-msm, linux-clk, linux-kernel,
	devicetree

On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
> The global clock controller on MSM8998 can consume a number of external
> clocks.  Document them.
> 
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---
>  .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
>  1 file changed, 33 insertions(+), 14 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> index e73a56f..2f3512b 100644
> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> @@ -40,20 +40,38 @@ properties:
>         - qcom,gcc-sm8150
>  
>    clocks:
> -    minItems: 1

1 or 2 clocks are no longer allowed?

> -    maxItems: 3
> -    items:
> -      - description: Board XO source
> -      - description: Board active XO source
> -      - description: Sleep clock source
> +    oneOf:
> +      #qcom,gcc-sm8150
> +      #qcom,gcc-sc7180

Typically, this would be an if/then schema, but I'm okay with leaving it 
like this. Depends whether you want to check the clocks match the 
compatible.

> +      - items:
> +        - description: Board XO source
> +        - description: Board active XO source
> +        - description: Sleep clock source
> +      #qcom,gcc-msm8998
> +      - items:
> +        - description: Board XO source
> +        - description: USB 3.0 phy pipe clock
> +        - description: UFS phy rx symbol clock for pipe 0
> +        - description: UFS phy rx symbol clock for pipe 1
> +        - description: UFS phy tx symbol clock
> +        - description: PCIE phy pipe clock
>  
>    clock-names:
> -    minItems: 1
> -    maxItems: 3
> -    items:
> -      - const: bi_tcxo
> -      - const: bi_tcxo_ao
> -      - const: sleep_clk
> +    oneOf:
> +      #qcom,gcc-sm8150
> +      #qcom,gcc-sc7180
> +      - items:
> +        - const: bi_tcxo
> +        - const: bi_tcxo_ao
> +        - const: sleep_clk
> +      #qcom,gcc-msm8998
> +      - items:
> +        - const: xo
> +        - const: usb3_pipe
> +        - const: ufs_rx_symbol0
> +        - const: ufs_rx_symbol1
> +        - const: ufs_tx_symbol0
> +        - const: pcie0_pipe
>  
>    '#clock-cells':
>      const: 1
> @@ -118,6 +136,7 @@ else:
>        compatible:
>          contains:
>            enum:
> +            - qcom,gcc-msm8998
>              - qcom,gcc-sm8150
>              - qcom,gcc-sc7180
>    then:
> @@ -179,8 +198,8 @@ examples:
>      clock-controller@100000 {
>        compatible = "qcom,gcc-sc7180";
>        reg = <0x100000 0x1f0000>;
> -      clocks = <&rpmhcc 0>, <&rpmhcc 1>;
> -      clock-names = "bi_tcxo", "bi_tcxo_ao";
> +      clocks = <&rpmhcc 0>, <&rpmhcc 1>, <0>;
> +      clock-names = "bi_tcxo", "bi_tcxo_ao", "sleep_clk";

The patch subject says 8998, but this is changing sc7180. 

>        #clock-cells = <1>;
>        #reset-cells = <1>;
>        #power-domain-cells = <1>;
> -- 
> Qualcomm Technologies, Inc. is a member of the
> Code Aurora Forum, a Linux Foundation Collaborative Project.
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc
  2019-11-08 23:17 ` [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc Jeffrey Hugo
@ 2019-11-12  0:55   ` Rob Herring
  2019-11-12 16:07     ` Jeffrey Hugo
  2019-11-12  0:59   ` Rob Herring
  1 sibling, 1 reply; 16+ messages in thread
From: Rob Herring @ 2019-11-12  0:55 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: mturquette, sboyd, robh+dt, mark.rutland, agross,
	bjorn.andersson, marc.w.gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree, Jeffrey Hugo

On Fri,  8 Nov 2019 16:17:48 -0700, Jeffrey Hugo wrote:
> Document the multimedia clock controller found on MSM8998.
> 
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---
>  .../devicetree/bindings/clock/qcom,mmcc.yaml       |  36 ++++
>  include/dt-bindings/clock/qcom,mmcc-msm8998.h      | 210 +++++++++++++++++++++
>  2 files changed, 246 insertions(+)
>  create mode 100644 include/dt-bindings/clock/qcom,mmcc-msm8998.h
> 

Please add Acked-by/Reviewed-by tags when posting new versions. However,
there's no need to repost patches *only* to add the tags. The upstream
maintainer will do that for acks received on the version they apply.

If a tag was not added on purpose, please state why and what changed.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema
  2019-11-08 23:17 ` [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema Jeffrey Hugo
@ 2019-11-12  0:58   ` Rob Herring
  2019-11-12 16:11     ` Jeffrey Hugo
  0 siblings, 1 reply; 16+ messages in thread
From: Rob Herring @ 2019-11-12  0:58 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: mturquette, sboyd, mark.rutland, agross, bjorn.andersson,
	marc.w.gonzalez, linux-arm-msm, linux-clk, linux-kernel,
	devicetree

On Fri, Nov 08, 2019 at 04:17:33PM -0700, Jeffrey Hugo wrote:
> Convert the qcom,mmcc-X clock controller binding to DT schema.
> 
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---
>  .../devicetree/bindings/clock/qcom,mmcc.txt        | 28 ----------
>  .../devicetree/bindings/clock/qcom,mmcc.yaml       | 59 ++++++++++++++++++++++
>  2 files changed, 59 insertions(+), 28 deletions(-)
>  delete mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.txt
>  create mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
> 
> diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt b/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
> deleted file mode 100644
> index 8b0f784..0000000
> --- a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
> +++ /dev/null
> @@ -1,28 +0,0 @@
> -Qualcomm Multimedia Clock & Reset Controller Binding
> -----------------------------------------------------
> -
> -Required properties :
> -- compatible : shall contain only one of the following:
> -
> -			"qcom,mmcc-apq8064"
> -			"qcom,mmcc-apq8084"
> -			"qcom,mmcc-msm8660"
> -			"qcom,mmcc-msm8960"
> -			"qcom,mmcc-msm8974"
> -			"qcom,mmcc-msm8996"
> -
> -- reg : shall contain base register location and length
> -- #clock-cells : shall contain 1
> -- #reset-cells : shall contain 1
> -
> -Optional properties :
> -- #power-domain-cells : shall contain 1
> -
> -Example:
> -	clock-controller@4000000 {
> -		compatible = "qcom,mmcc-msm8960";
> -		reg = <0x4000000 0x1000>;
> -		#clock-cells = <1>;
> -		#reset-cells = <1>;
> -		#power-domain-cells = <1>;
> -	};
> diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
> new file mode 100644
> index 0000000..61ed4a2
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
> @@ -0,0 +1,59 @@
> +# SPDX-License-Identifier: GPL-2.0-only
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/bindings/clock/qcom,mmcc.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Qualcomm Multimedia Clock & Reset Controller Binding
> +
> +maintainers:
> +  - Jeffrey Hugo <jhugo@codeaurora.org>
> +
> +description: |
> +  Qualcomm multimedia clock control module which supports the clocks, resets and
> +  power domains.
> +
> +properties:
> +  compatible :
> +    enum:
> +       - qcom,mmcc-apq8064
> +       - qcom,mmcc-apq8084
> +       - qcom,mmcc-msm8660
> +       - qcom,mmcc-msm8960
> +       - qcom,mmcc-msm8974
> +       - qcom,mmcc-msm8996
> +
> +  '#clock-cells':
> +    const: 1
> +
> +  '#reset-cells':
> +    const: 1
> +
> +  '#power-domain-cells':
> +    const: 1
> +
> +  reg:
> +    maxItems: 1
> +
> +  protected-clocks:
> +    description:
> +       Protected clock specifier list as per common clock binding

Wasn't documented before. Okay to add here, but mention it in the commit 
msg.

> +
> +required:
> +  - compatible
> +  - reg
> +  - '#clock-cells'
> +  - '#reset-cells'
> +  - '#power-domain-cells'
> +
> +examples:
> +  # Example for MMCC for MSM8960:
> +  - |
> +    clock-controller@4000000 {
> +      compatible = "qcom,mmcc-msm8960";
> +      reg = <0x4000000 0x1000>;
> +      #clock-cells = <1>;
> +      #reset-cells = <1>;
> +      #power-domain-cells = <1>;
> +    };
> +...
> -- 
> Qualcomm Technologies, Inc. is a member of the
> Code Aurora Forum, a Linux Foundation Collaborative Project.
> 

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc
  2019-11-08 23:17 ` [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc Jeffrey Hugo
  2019-11-12  0:55   ` Rob Herring
@ 2019-11-12  0:59   ` Rob Herring
  1 sibling, 0 replies; 16+ messages in thread
From: Rob Herring @ 2019-11-12  0:59 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: mturquette, sboyd, robh+dt, mark.rutland, agross,
	bjorn.andersson, marc.w.gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree, Jeffrey Hugo

On Fri,  8 Nov 2019 16:17:48 -0700, Jeffrey Hugo wrote:
> Document the multimedia clock controller found on MSM8998.
> 
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---
>  .../devicetree/bindings/clock/qcom,mmcc.yaml       |  36 ++++
>  include/dt-bindings/clock/qcom,mmcc-msm8998.h      | 210 +++++++++++++++++++++
>  2 files changed, 246 insertions(+)
>  create mode 100644 include/dt-bindings/clock/qcom,mmcc-msm8998.h
> 

Reviewed-by: Rob Herring <robh@kernel.org>

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc
  2019-11-12  0:55   ` Rob Herring
@ 2019-11-12 16:07     ` Jeffrey Hugo
  0 siblings, 0 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-12 16:07 UTC (permalink / raw)
  To: Rob Herring
  Cc: mturquette, sboyd, robh+dt, mark.rutland, agross,
	bjorn.andersson, marc.w.gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree

On 11/11/2019 5:55 PM, Rob Herring wrote:
> On Fri,  8 Nov 2019 16:17:48 -0700, Jeffrey Hugo wrote:
>> Document the multimedia clock controller found on MSM8998.
>>
>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>> ---
>>   .../devicetree/bindings/clock/qcom,mmcc.yaml       |  36 ++++
>>   include/dt-bindings/clock/qcom,mmcc-msm8998.h      | 210 +++++++++++++++++++++
>>   2 files changed, 246 insertions(+)
>>   create mode 100644 include/dt-bindings/clock/qcom,mmcc-msm8998.h
>>
> 
> Please add Acked-by/Reviewed-by tags when posting new versions. However,
> there's no need to repost patches *only* to add the tags. The upstream
> maintainer will do that for acks received on the version they apply.
> 
> If a tag was not added on purpose, please state why and what changed.
> 

Ya know, normally you are right about this, as I did something stupid, 
but in this case I did mention in the cover letter that tags were 
dropped because the bindings were re-written in yaml (ie no one had 
reviewed the new code).

-- 
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema
  2019-11-12  0:58   ` Rob Herring
@ 2019-11-12 16:11     ` Jeffrey Hugo
  0 siblings, 0 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-12 16:11 UTC (permalink / raw)
  To: Rob Herring
  Cc: mturquette, sboyd, mark.rutland, agross, bjorn.andersson,
	marc.w.gonzalez, linux-arm-msm, linux-clk, linux-kernel,
	devicetree

On 11/11/2019 5:58 PM, Rob Herring wrote:
> On Fri, Nov 08, 2019 at 04:17:33PM -0700, Jeffrey Hugo wrote:
>> Convert the qcom,mmcc-X clock controller binding to DT schema.
>>
>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>> ---
>>   .../devicetree/bindings/clock/qcom,mmcc.txt        | 28 ----------
>>   .../devicetree/bindings/clock/qcom,mmcc.yaml       | 59 ++++++++++++++++++++++
>>   2 files changed, 59 insertions(+), 28 deletions(-)
>>   delete mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.txt
>>   create mode 100644 Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
>>
>> diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt b/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
>> deleted file mode 100644
>> index 8b0f784..0000000
>> --- a/Documentation/devicetree/bindings/clock/qcom,mmcc.txt
>> +++ /dev/null
>> @@ -1,28 +0,0 @@
>> -Qualcomm Multimedia Clock & Reset Controller Binding
>> -----------------------------------------------------
>> -
>> -Required properties :
>> -- compatible : shall contain only one of the following:
>> -
>> -			"qcom,mmcc-apq8064"
>> -			"qcom,mmcc-apq8084"
>> -			"qcom,mmcc-msm8660"
>> -			"qcom,mmcc-msm8960"
>> -			"qcom,mmcc-msm8974"
>> -			"qcom,mmcc-msm8996"
>> -
>> -- reg : shall contain base register location and length
>> -- #clock-cells : shall contain 1
>> -- #reset-cells : shall contain 1
>> -
>> -Optional properties :
>> -- #power-domain-cells : shall contain 1
>> -
>> -Example:
>> -	clock-controller@4000000 {
>> -		compatible = "qcom,mmcc-msm8960";
>> -		reg = <0x4000000 0x1000>;
>> -		#clock-cells = <1>;
>> -		#reset-cells = <1>;
>> -		#power-domain-cells = <1>;
>> -	};
>> diff --git a/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
>> new file mode 100644
>> index 0000000..61ed4a2
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/clock/qcom,mmcc.yaml
>> @@ -0,0 +1,59 @@
>> +# SPDX-License-Identifier: GPL-2.0-only
>> +%YAML 1.2
>> +---
>> +$id: http://devicetree.org/schemas/bindings/clock/qcom,mmcc.yaml#
>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>> +
>> +title: Qualcomm Multimedia Clock & Reset Controller Binding
>> +
>> +maintainers:
>> +  - Jeffrey Hugo <jhugo@codeaurora.org>
>> +
>> +description: |
>> +  Qualcomm multimedia clock control module which supports the clocks, resets and
>> +  power domains.
>> +
>> +properties:
>> +  compatible :
>> +    enum:
>> +       - qcom,mmcc-apq8064
>> +       - qcom,mmcc-apq8084
>> +       - qcom,mmcc-msm8660
>> +       - qcom,mmcc-msm8960
>> +       - qcom,mmcc-msm8974
>> +       - qcom,mmcc-msm8996
>> +
>> +  '#clock-cells':
>> +    const: 1
>> +
>> +  '#reset-cells':
>> +    const: 1
>> +
>> +  '#power-domain-cells':
>> +    const: 1
>> +
>> +  reg:
>> +    maxItems: 1
>> +
>> +  protected-clocks:
>> +    description:
>> +       Protected clock specifier list as per common clock binding
> 
> Wasn't documented before. Okay to add here, but mention it in the commit
> msg.

Its a generic property that applies to all providers.  In the non-yaml 
bindings, its was documented in the clock-binding.txt, thus implicitly 
documented for all the bindings.  That doesn't work in yaml, so I'm 
making it explicit here (following the GCC bindings you already 
approved).  Nothing changes.

> 
>> +
>> +required:
>> +  - compatible
>> +  - reg
>> +  - '#clock-cells'
>> +  - '#reset-cells'
>> +  - '#power-domain-cells'
>> +
>> +examples:
>> +  # Example for MMCC for MSM8960:
>> +  - |
>> +    clock-controller@4000000 {
>> +      compatible = "qcom,mmcc-msm8960";
>> +      reg = <0x4000000 0x1000>;
>> +      #clock-cells = <1>;
>> +      #reset-cells = <1>;
>> +      #power-domain-cells = <1>;
>> +    };
>> +...
>> -- 
>> Qualcomm Technologies, Inc. is a member of the
>> Code Aurora Forum, a Linux Foundation Collaborative Project.
>>


-- 
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-12  0:44   ` Rob Herring
@ 2019-11-12 16:25     ` Jeffrey Hugo
  2019-11-12 18:37       ` Rob Herring
  0 siblings, 1 reply; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-12 16:25 UTC (permalink / raw)
  To: Rob Herring
  Cc: mturquette, sboyd, mark.rutland, agross, bjorn.andersson,
	marc.w.gonzalez, linux-arm-msm, linux-clk, linux-kernel,
	devicetree

On 11/11/2019 5:44 PM, Rob Herring wrote:
> On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
>> The global clock controller on MSM8998 can consume a number of external
>> clocks.  Document them.
>>
>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>> ---
>>   .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
>>   1 file changed, 33 insertions(+), 14 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>> index e73a56f..2f3512b 100644
>> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>> @@ -40,20 +40,38 @@ properties:
>>          - qcom,gcc-sm8150
>>   
>>     clocks:
>> -    minItems: 1
> 
> 1 or 2 clocks are no longer allowed?

Correct.

The primary reason is that Stephen indicated in previous discussions 
that if the hardware exists, it should be indicated in DT, regardless if 
the driver uses it.  In the 7180 and 8150 case, the hardware exists, so 
these should not be optional.

The secondary reason is I found that the schema was broken anyways.  In 
the way it was written, if you implemented sleep, you could not skip 
xo_ao, however there is a dts that did exactly that.

The third reason was that I couldn't find a way to write valid yaml to 
preserve the original meaning.  when you have an "items" as a subnode of 
"oneOf", you no longer have control over the minItems/maxItems, so all 3 
became required anyways.  I find it disappointing that the "version" of 
Yaml used for DT bindings is not documented, so after several hours of 
trial and error, I just gave up since I found this to work (failed cases 
just gave me an error with no indication of what was wrong, not even a 
line number).

> 
>> -    maxItems: 3
>> -    items:
>> -      - description: Board XO source
>> -      - description: Board active XO source
>> -      - description: Sleep clock source
>> +    oneOf:
>> +      #qcom,gcc-sm8150
>> +      #qcom,gcc-sc7180
> 
> Typically, this would be an if/then schema, but I'm okay with leaving it
> like this. Depends whether you want to check the clocks match the
> compatible.

Is there an example somewhere?  The only thing I found was 
example-schema.yaml which seemed to suggest this way.

> 
>> +      - items:
>> +        - description: Board XO source
>> +        - description: Board active XO source
>> +        - description: Sleep clock source
>> +      #qcom,gcc-msm8998
>> +      - items:
>> +        - description: Board XO source
>> +        - description: USB 3.0 phy pipe clock
>> +        - description: UFS phy rx symbol clock for pipe 0
>> +        - description: UFS phy rx symbol clock for pipe 1
>> +        - description: UFS phy tx symbol clock
>> +        - description: PCIE phy pipe clock
>>   
>>     clock-names:
>> -    minItems: 1
>> -    maxItems: 3
>> -    items:
>> -      - const: bi_tcxo
>> -      - const: bi_tcxo_ao
>> -      - const: sleep_clk
>> +    oneOf:
>> +      #qcom,gcc-sm8150
>> +      #qcom,gcc-sc7180
>> +      - items:
>> +        - const: bi_tcxo
>> +        - const: bi_tcxo_ao
>> +        - const: sleep_clk
>> +      #qcom,gcc-msm8998
>> +      - items:
>> +        - const: xo
>> +        - const: usb3_pipe
>> +        - const: ufs_rx_symbol0
>> +        - const: ufs_rx_symbol1
>> +        - const: ufs_tx_symbol0
>> +        - const: pcie0_pipe
>>   
>>     '#clock-cells':
>>       const: 1
>> @@ -118,6 +136,7 @@ else:
>>         compatible:
>>           contains:
>>             enum:
>> +            - qcom,gcc-msm8998
>>               - qcom,gcc-sm8150
>>               - qcom,gcc-sc7180
>>     then:
>> @@ -179,8 +198,8 @@ examples:
>>       clock-controller@100000 {
>>         compatible = "qcom,gcc-sc7180";
>>         reg = <0x100000 0x1f0000>;
>> -      clocks = <&rpmhcc 0>, <&rpmhcc 1>;
>> -      clock-names = "bi_tcxo", "bi_tcxo_ao";
>> +      clocks = <&rpmhcc 0>, <&rpmhcc 1>, <0>;
>> +      clock-names = "bi_tcxo", "bi_tcxo_ao", "sleep_clk";
> 
> The patch subject says 8998, but this is changing sc7180.

I'm fixing up the example so that it no longer fails checks.  See the 
above comment.

> 
>>         #clock-cells = <1>;
>>         #reset-cells = <1>;
>>         #power-domain-cells = <1>;
>> -- 
>> Qualcomm Technologies, Inc. is a member of the
>> Code Aurora Forum, a Linux Foundation Collaborative Project.
>>


-- 
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-12 16:25     ` Jeffrey Hugo
@ 2019-11-12 18:37       ` Rob Herring
  2019-11-12 19:38         ` Jeffrey Hugo
  0 siblings, 1 reply; 16+ messages in thread
From: Rob Herring @ 2019-11-12 18:37 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: Michael Turquette, Stephen Boyd, Mark Rutland, Andy Gross,
	Bjorn Andersson, Marc Gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree

On Tue, Nov 12, 2019 at 10:25 AM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
>
> On 11/11/2019 5:44 PM, Rob Herring wrote:
> > On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
> >> The global clock controller on MSM8998 can consume a number of external
> >> clocks.  Document them.
> >>
> >> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> >> ---
> >>   .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
> >>   1 file changed, 33 insertions(+), 14 deletions(-)
> >>
> >> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >> index e73a56f..2f3512b 100644
> >> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >> @@ -40,20 +40,38 @@ properties:
> >>          - qcom,gcc-sm8150
> >>
> >>     clocks:
> >> -    minItems: 1
> >
> > 1 or 2 clocks are no longer allowed?
>
> Correct.
>
> The primary reason is that Stephen indicated in previous discussions
> that if the hardware exists, it should be indicated in DT, regardless if
> the driver uses it.  In the 7180 and 8150 case, the hardware exists, so
> these should not be optional.

Agreed. The commit message should mention this though.

>
> The secondary reason is I found that the schema was broken anyways.  In
> the way it was written, if you implemented sleep, you could not skip
> xo_ao, however there is a dts that did exactly that.

If a dts can be updated in a compatible way, we should do that rather
than carry inconsistencies into the schema.

> The third reason was that I couldn't find a way to write valid yaml to
> preserve the original meaning.  when you have an "items" as a subnode of
> "oneOf", you no longer have control over the minItems/maxItems, so all 3
> became required anyways.

That would be a bug. You're saying something like this doesn't work?:

oneOf:
  - minItems: 1
    maxItems: 3
    items:
      - const: a
      - const: b
      - const: c

>  I find it disappointing that the "version" of
> Yaml used for DT bindings is not documented,

Not sure which part you mean? json-schema is the vocabulary which has
a spec. The meta-schema then constrains what the json-schema structure
should look like. That's still evolving a bit as I try to improve it
based on mistakes people make. Then there's the intermediate .dt.yaml
format used internally. That's supposed to stay internal and may go
away when/if we integrate the validation into dtc.

> so after several hours of
> trial and error, I just gave up since I found this to work (failed cases
> just gave me an error with no indication of what was wrong, not even a
> line number).

Schema failures or dts failures? It is possible to get line numbers
for either, but that makes validation much slower. In the latter case,
the line numbers aren't too useful either given they are for the
.dt.yaml file and not the .dts source file (dtc integration would
solve that). Adding '-n' to dt-doc-validate or dt-validate will turn
them on though.

Yes, error messages need work. I have some idea how to improve them,
but haven't had time to implement. Too many binding reviews... You can
get more detail with '-v' option. It's *way* more verbose, but not
necessarily more useful.

Rob

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-12 18:37       ` Rob Herring
@ 2019-11-12 19:38         ` Jeffrey Hugo
  2019-11-12 21:18           ` Rob Herring
  0 siblings, 1 reply; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-12 19:38 UTC (permalink / raw)
  To: Rob Herring
  Cc: Michael Turquette, Stephen Boyd, Mark Rutland, Andy Gross,
	Bjorn Andersson, Marc Gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree

On 11/12/2019 11:37 AM, Rob Herring wrote:
> On Tue, Nov 12, 2019 at 10:25 AM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
>>
>> On 11/11/2019 5:44 PM, Rob Herring wrote:
>>> On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
>>>> The global clock controller on MSM8998 can consume a number of external
>>>> clocks.  Document them.
>>>>
>>>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>>>> ---
>>>>    .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
>>>>    1 file changed, 33 insertions(+), 14 deletions(-)
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>> index e73a56f..2f3512b 100644
>>>> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>> @@ -40,20 +40,38 @@ properties:
>>>>           - qcom,gcc-sm8150
>>>>
>>>>      clocks:
>>>> -    minItems: 1
>>>
>>> 1 or 2 clocks are no longer allowed?
>>
>> Correct.
>>
>> The primary reason is that Stephen indicated in previous discussions
>> that if the hardware exists, it should be indicated in DT, regardless if
>> the driver uses it.  In the 7180 and 8150 case, the hardware exists, so
>> these should not be optional.
> 
> Agreed. The commit message should mention this though.

Fair enough, will do.

> 
>>
>> The secondary reason is I found that the schema was broken anyways.  In
>> the way it was written, if you implemented sleep, you could not skip
>> xo_ao, however there is a dts that did exactly that.
> 
> If a dts can be updated in a compatible way, we should do that rather
> than carry inconsistencies into the schema.
> 
>> The third reason was that I couldn't find a way to write valid yaml to
>> preserve the original meaning.  when you have an "items" as a subnode of
>> "oneOf", you no longer have control over the minItems/maxItems, so all 3
>> became required anyways.
> 
> That would be a bug. You're saying something like this doesn't work?:
> 
> oneOf:
>    - minItems: 1
>      maxItems: 3
>      items:
>        - const: a
>        - const: b
>        - const: c

Yes.  That specifically won't work.  "items" would need to have the dash 
preceding it, otherwise it won't compile if you have more than one.  But 
ignoring that, yes, when it compiled, and I saw the output from the 
check failing (after adding verbose mode), min and max for the items 
list would be 3, and the check would fail.

> 
>>   I find it disappointing that the "version" of
>> Yaml used for DT bindings is not documented,
> 
> Not sure which part you mean? json-schema is the vocabulary which has
> a spec. The meta-schema then constrains what the json-schema structure
> should look like. That's still evolving a bit as I try to improve it
> based on mistakes people make. Then there's the intermediate .dt.yaml
> format used internally. That's supposed to stay internal and may go
> away when/if we integrate the validation into dtc.

So, this is probably off-topic, but hopefully you'll find this useful.

I'm probably in the minority, but I really haven't used json-schema nor 
yaml before.  I have experience with other "schema" languages, so I 
figured I could pick what I need from the documentation.

The only documentation I see is writing-schema.md and example-schema.yaml

To me, writing-schema.md is insufficient.  Its better than nothing, so 
I'm still glad it exists, but I don't have any confidence I can really 
write a binding yaml from scratch based on it.  It does a good thing by 
telling you what are important properties of a binding, so based on that 
you can kind of start to understand how existing bindings actually work. 
  Its great in telling you how to run the validation checks (the Running 
checks) section.  The dependencies section is awesome from my 
perspective - most projects seem to assume you just know what their 
dependencies are, and its painful to try to figure them out when you get 
cryptic errors during make.

Where it really fails is that I get no sense of the language.  As a 
minimum a lexigraphic specification that would allow me to write a 
compiler (I've done this before).  Then I would understand what are the 
keywords, and where they are valid.  I wouldn't understand what they 
mean, but at-least I can look at some implemented examples and 
extrapolate from there.

Have you by chance ever looked at the ACPI spec?  Maybe not the best 
example, but its the one that comes to my mind first.  ACPI has ACPI 
Source Language (ASL).  Its an interpreted hardware description language 
that doesn't match yaml, but I think the ACPI spec does a reasonable job 
of describing it.  You have a lexographic definition which seems to be 
really helpful to ACPICA in implementing the intrepreter.  It lists all 
of the valid operators, types, etc.  It provides detailed references of 
each keyword - how they are used, what they do, etc.  Its not the 
greatest at "how to write ASL 101" or "these are common problems that 
people face, and how they can be solved", but atleast with what there 
is, I could read every keyword that seems to be possibly related to what 
I want to do, and hazard a guess if it would work for my problem.

Perhaps that is outside the scope of the writing-schema.md document, 
that is fair.  However, I argue that the document does not provide 
sufficient references.  The document provides a reference to the 
json-schema spec, but the spec is kinda useless (atleast I feel that it 
is).  "minItems" is not defined anywhere in the spec.  What does it 
mean?  How can I use it?  Specific to minItems/maxItems, I'll I've 
gathered about it is from example-schema.yaml which indicates its a way 
to identify mandatory and optional values for a property, but it doesn't 
describe the fact that order matters, and you cannot mix/match things - 
IE it looks like you need atleast min items, and at most max items, but 
even if you have enough items to satisfy min, there cannot be gaps (you 
can't pick items 1, 5, 10 from the list).  I only found that out from 
running the validation checks with trial/error.

There is no reference to the yaml spec, despite the document stating 
that the bindings are written in yaml.

However, having found the yaml spec, its really not much better than the 
json-schema spec, and it doesn't line up because as the document states, 
the bindings are not really written in yaml - its a subset of yaml where 
a ton of the boilerplate "code" is skipped.

What is boilerplate that is skipped?  IMO, if you are not strictly 
adhering to yaml, then you need to clearly document your own derivative 
language so that someone like me whom is being introduced to all of this 
for the first time can start to figure out some of it.  It would be 
helpful to look at other yaml examples, and understand what is 
considered to be boilerplate so I can translate that to a DT binding.

I understand, the majority of the above is complaints and demands which 
is really not fair to you, since you are spending what I presume to be 
your "non-dayjob" time to make the community better.  However, I don't 
really know how to contribute to make the documentation better.  I don't 
understand enough.  As far as this topic is concerned, I'm a dumb monkey 
banging on a keyboard hoping to get close enough to Shakespeare to pass 
mustard by accident, and maybe learn something along the way so that 
next time, I might have an idea of how to do something of what I need.

Hopefully you've made it this far - that ended up being a lot more text 
that I thought it would be.  I really hope this is useful feedback to 
you, but let me know if I am still not clear on something.  I will try 
my best to clarify more.  If you feel like I can contribute somehow, 
just let me know.

> 
>> so after several hours of
>> trial and error, I just gave up since I found this to work (failed cases
>> just gave me an error with no indication of what was wrong, not even a
>> line number).
> 
> Schema failures or dts failures? It is possible to get line numbers
> for either, but that makes validation much slower. In the latter case,
> the line numbers aren't too useful either given they are for the
> .dt.yaml file and not the .dts source file (dtc integration would
> solve that). Adding '-n' to dt-doc-validate or dt-validate will turn
> them on though.

Schema compilation failures.  I don't recall the exact error message, 
but it was something like "no valid schema found, continuing". 
Essentially running "dt_binding_check".  I tried with -v but wasn't 
getting much more in this case.  I didn't try -n.

> 
> Yes, error messages need work. I have some idea how to improve them,
> but haven't had time to implement. Too many binding reviews... You can
> get more detail with '-v' option. It's *way* more verbose, but not
> necessarily more useful.
> 
> Rob
> 


-- 
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-12 19:38         ` Jeffrey Hugo
@ 2019-11-12 21:18           ` Rob Herring
  2019-11-12 22:03             ` Jeffrey Hugo
  0 siblings, 1 reply; 16+ messages in thread
From: Rob Herring @ 2019-11-12 21:18 UTC (permalink / raw)
  To: Jeffrey Hugo
  Cc: Michael Turquette, Stephen Boyd, Mark Rutland, Andy Gross,
	Bjorn Andersson, Marc Gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree

On Tue, Nov 12, 2019 at 1:38 PM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
>
> On 11/12/2019 11:37 AM, Rob Herring wrote:
> > On Tue, Nov 12, 2019 at 10:25 AM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
> >>
> >> On 11/11/2019 5:44 PM, Rob Herring wrote:
> >>> On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
> >>>> The global clock controller on MSM8998 can consume a number of external
> >>>> clocks.  Document them.
> >>>>
> >>>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> >>>> ---
> >>>>    .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
> >>>>    1 file changed, 33 insertions(+), 14 deletions(-)
> >>>>
> >>>> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >>>> index e73a56f..2f3512b 100644
> >>>> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >>>> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
> >>>> @@ -40,20 +40,38 @@ properties:
> >>>>           - qcom,gcc-sm8150
> >>>>
> >>>>      clocks:
> >>>> -    minItems: 1
> >>>
> >>> 1 or 2 clocks are no longer allowed?
> >>
> >> Correct.
> >>
> >> The primary reason is that Stephen indicated in previous discussions
> >> that if the hardware exists, it should be indicated in DT, regardless if
> >> the driver uses it.  In the 7180 and 8150 case, the hardware exists, so
> >> these should not be optional.
> >
> > Agreed. The commit message should mention this though.
>
> Fair enough, will do.
>
> >
> >>
> >> The secondary reason is I found that the schema was broken anyways.  In
> >> the way it was written, if you implemented sleep, you could not skip
> >> xo_ao, however there is a dts that did exactly that.
> >
> > If a dts can be updated in a compatible way, we should do that rather
> > than carry inconsistencies into the schema.
> >
> >> The third reason was that I couldn't find a way to write valid yaml to
> >> preserve the original meaning.  when you have an "items" as a subnode of
> >> "oneOf", you no longer have control over the minItems/maxItems, so all 3
> >> became required anyways.
> >
> > That would be a bug. You're saying something like this doesn't work?:
> >
> > oneOf:
> >    - minItems: 1
> >      maxItems: 3
> >      items:
> >        - const: a
> >        - const: b
> >        - const: c
>
> Yes.  That specifically won't work.  "items" would need to have the dash
> preceding it, otherwise it won't compile if you have more than one.  But
> ignoring that, yes, when it compiled, and I saw the output from the
> check failing (after adding verbose mode), min and max for the items
> list would be 3, and the check would fail.

A '-' before items would make oneOf have 2 separate schemas. That
would pass with any values for 1-3 items except it would fail for 3
items with [a, b, c] because 2 oneOf clauses pass.

> >>   I find it disappointing that the "version" of
> >> Yaml used for DT bindings is not documented,
> >
> > Not sure which part you mean? json-schema is the vocabulary which has
> > a spec. The meta-schema then constrains what the json-schema structure
> > should look like. That's still evolving a bit as I try to improve it
> > based on mistakes people make. Then there's the intermediate .dt.yaml
> > format used internally. That's supposed to stay internal and may go
> > away when/if we integrate the validation into dtc.
>
> So, this is probably off-topic, but hopefully you'll find this useful.

I'm interested in knowing the pain points.

> I'm probably in the minority, but I really haven't used json-schema nor
> yaml before.  I have experience with other "schema" languages, so I
> figured I could pick what I need from the documentation.

Well, json-schema was new to me before this. There's definitely some
things I really don't love about it, but it's better than trying to
define our own language. It's generally been able to handle some of
the more complex cases.

> The only documentation I see is writing-schema.md and example-schema.yaml
>
> To me, writing-schema.md is insufficient.  Its better than nothing, so
> I'm still glad it exists, but I don't have any confidence I can really
> write a binding yaml from scratch based on it.  It does a good thing by
> telling you what are important properties of a binding, so based on that
> you can kind of start to understand how existing bindings actually work.
>   Its great in telling you how to run the validation checks (the Running
> checks) section.  The dependencies section is awesome from my
> perspective - most projects seem to assume you just know what their
> dependencies are, and its painful to try to figure them out when you get
> cryptic errors during make.
>
> Where it really fails is that I get no sense of the language.  As a
> minimum a lexigraphic specification that would allow me to write a
> compiler (I've done this before).  Then I would understand what are the
> keywords, and where they are valid.  I wouldn't understand what they
> mean, but at-least I can look at some implemented examples and
> extrapolate from there.
>
> Have you by chance ever looked at the ACPI spec?  Maybe not the best
> example, but its the one that comes to my mind first.  ACPI has ACPI
> Source Language (ASL).  Its an interpreted hardware description language
> that doesn't match yaml, but I think the ACPI spec does a reasonable job
> of describing it.  You have a lexographic definition which seems to be
> really helpful to ACPICA in implementing the intrepreter.  It lists all
> of the valid operators, types, etc.  It provides detailed references of
> each keyword - how they are used, what they do, etc.  Its not the
> greatest at "how to write ASL 101" or "these are common problems that
> people face, and how they can be solved", but atleast with what there
> is, I could read every keyword that seems to be possibly related to what
> I want to do, and hazard a guess if it would work for my problem.

I have not read the ACPI spec.

> Perhaps that is outside the scope of the writing-schema.md document,
> that is fair.  However, I argue that the document does not provide
> sufficient references.  The document provides a reference to the
> json-schema spec, but the spec is kinda useless (atleast I feel that it
> is).  "minItems" is not defined anywhere in the spec.  What does it
> mean?  How can I use it?  Specific to minItems/maxItems, I'll I've
> gathered about it is from example-schema.yaml which indicates its a way
> to identify mandatory and optional values for a property, but it doesn't
> describe the fact that order matters, and you cannot mix/match things -
> IE it looks like you need atleast min items, and at most max items, but
> even if you have enough items to satisfy min, there cannot be gaps (you
> can't pick items 1, 5, 10 from the list).  I only found that out from
> running the validation checks with trial/error.

I think you looked at the 'Core' spec rather than the 'Validation' spec:
http://json-schema.org/draft/2019-09/json-schema-validation.html

Though that has moved on to a newer version and we're still on draft7
which is here:
https://tools.ietf.org/html/draft-handrews-json-schema-validation-01

I guess a direct link to this with 'Details on json-schema keywords is
here' would be helpful.

minItems/maxItems is the one area we deviate from json-schema
defaults. That's what the 'Property Schema' section calls out.

Order matters for DT too, so that aspect matches up well with
json-schema. That's been a common issue in dts files, so schema
starting to enforce that will be good for new bindings, but somewhat
painful for existing ones.

> There is no reference to the yaml spec, despite the document stating
> that the bindings are written in yaml.
>
> However, having found the yaml spec, its really not much better than the
> json-schema spec, and it doesn't line up because as the document states,
> the bindings are not really written in yaml - its a subset of yaml where
> a ton of the boilerplate "code" is skipped.

Yeah, there's a lot to YAML that no one uses and I too find the spec
pretty useless (hence why no reference). Like most other uses I've
encountered, we're using a JSON compatible subset which is just lists
and dicts of key/value pairs. The main thing folks need to know and
trip up on are: indentation is important (including no tabs) and pay
attention to '-' or lack of.

> What is boilerplate that is skipped?  IMO, if you are not strictly
> adhering to yaml, then you need to clearly document your own derivative
> language so that someone like me whom is being introduced to all of this
> for the first time can start to figure out some of it.  It would be
> helpful to look at other yaml examples, and understand what is
> considered to be boilerplate so I can translate that to a DT binding.

We're not skipping any boilerplate. We're not using advanced features
like tags or anchors. You can use any YAML parser including online
ones to read the files.

> I understand, the majority of the above is complaints and demands which
> is really not fair to you, since you are spending what I presume to be
> your "non-dayjob" time to make the community better.

It's my day job or part of it, just not enough hours in the day...

> However, I don't
> really know how to contribute to make the documentation better.  I don't
> understand enough.  As far as this topic is concerned, I'm a dumb monkey
> banging on a keyboard hoping to get close enough to Shakespeare to pass
> mustard by accident, and maybe learn something along the way so that
> next time, I might have an idea of how to do something of what I need.

The challenge is providing enough information to write bindings
without being json-schema experts. My hope is really to build up
enough examples and make the meta-schema good enough to keep folks
within the lines. Maybe that's a flawed approach, but even getting
folks to follow writing-schema.rst and run 'make dt_binding_check' has
been a challenge.

> Hopefully you've made it this far - that ended up being a lot more text
> that I thought it would be.  I really hope this is useful feedback to
> you, but let me know if I am still not clear on something.  I will try
> my best to clarify more.  If you feel like I can contribute somehow,
> just let me know.
>
> >
> >> so after several hours of
> >> trial and error, I just gave up since I found this to work (failed cases
> >> just gave me an error with no indication of what was wrong, not even a
> >> line number).
> >
> > Schema failures or dts failures? It is possible to get line numbers
> > for either, but that makes validation much slower. In the latter case,
> > the line numbers aren't too useful either given they are for the
> > .dt.yaml file and not the .dts source file (dtc integration would
> > solve that). Adding '-n' to dt-doc-validate or dt-validate will turn
> > them on though.
>
> Schema compilation failures.  I don't recall the exact error message,
> but it was something like "no valid schema found, continuing".
> Essentially running "dt_binding_check".  I tried with -v but wasn't
> getting much more in this case.  I didn't try -n.

That's before we even validate the schema, so something has gone wrong
pretty early. You may get farther with 'make -k'. I'll have to look
into it. The schemas are actually built twice. They are all built into
processed-schema.yaml. That's supposed to skip any with errors and is
what's used to validate dts files. If that's failing for some reason,
then it's going to be pretty vague. The dt_binding_check rule also
fully validates each binding schema and builds and validates the
examples. It should print more detailed errors (though still sometimes
vague).

Rob

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc
  2019-11-12 21:18           ` Rob Herring
@ 2019-11-12 22:03             ` Jeffrey Hugo
  0 siblings, 0 replies; 16+ messages in thread
From: Jeffrey Hugo @ 2019-11-12 22:03 UTC (permalink / raw)
  To: Rob Herring
  Cc: Michael Turquette, Stephen Boyd, Mark Rutland, Andy Gross,
	Bjorn Andersson, Marc Gonzalez, linux-arm-msm, linux-clk,
	linux-kernel, devicetree

On 11/12/2019 2:18 PM, Rob Herring wrote:
> On Tue, Nov 12, 2019 at 1:38 PM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
>>
>> On 11/12/2019 11:37 AM, Rob Herring wrote:
>>> On Tue, Nov 12, 2019 at 10:25 AM Jeffrey Hugo <jhugo@codeaurora.org> wrote:
>>>>
>>>> On 11/11/2019 5:44 PM, Rob Herring wrote:
>>>>> On Fri, Nov 08, 2019 at 04:17:16PM -0700, Jeffrey Hugo wrote:
>>>>>> The global clock controller on MSM8998 can consume a number of external
>>>>>> clocks.  Document them.
>>>>>>
>>>>>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>>>>>> ---
>>>>>>     .../devicetree/bindings/clock/qcom,gcc.yaml        | 47 +++++++++++++++-------
>>>>>>     1 file changed, 33 insertions(+), 14 deletions(-)
>>>>>>
>>>>>> diff --git a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>>>> index e73a56f..2f3512b 100644
>>>>>> --- a/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>>>> +++ b/Documentation/devicetree/bindings/clock/qcom,gcc.yaml
>>>>>> @@ -40,20 +40,38 @@ properties:
>>>>>>            - qcom,gcc-sm8150
>>>>>>
>>>>>>       clocks:
>>>>>> -    minItems: 1
>>>>>
>>>>> 1 or 2 clocks are no longer allowed?
>>>>
>>>> Correct.
>>>>
>>>> The primary reason is that Stephen indicated in previous discussions
>>>> that if the hardware exists, it should be indicated in DT, regardless if
>>>> the driver uses it.  In the 7180 and 8150 case, the hardware exists, so
>>>> these should not be optional.
>>>
>>> Agreed. The commit message should mention this though.
>>
>> Fair enough, will do.
>>
>>>
>>>>
>>>> The secondary reason is I found that the schema was broken anyways.  In
>>>> the way it was written, if you implemented sleep, you could not skip
>>>> xo_ao, however there is a dts that did exactly that.
>>>
>>> If a dts can be updated in a compatible way, we should do that rather
>>> than carry inconsistencies into the schema.
>>>
>>>> The third reason was that I couldn't find a way to write valid yaml to
>>>> preserve the original meaning.  when you have an "items" as a subnode of
>>>> "oneOf", you no longer have control over the minItems/maxItems, so all 3
>>>> became required anyways.
>>>
>>> That would be a bug. You're saying something like this doesn't work?:
>>>
>>> oneOf:
>>>     - minItems: 1
>>>       maxItems: 3
>>>       items:
>>>         - const: a
>>>         - const: b
>>>         - const: c
>>
>> Yes.  That specifically won't work.  "items" would need to have the dash
>> preceding it, otherwise it won't compile if you have more than one.  But
>> ignoring that, yes, when it compiled, and I saw the output from the
>> check failing (after adding verbose mode), min and max for the items
>> list would be 3, and the check would fail.
> 
> A '-' before items would make oneOf have 2 separate schemas. That
> would pass with any values for 1-3 items except it would fail for 3
> items with [a, b, c] because 2 oneOf clauses pass.

What I was trying to do was something like:

oneOf:
     -minItems: 1
     -maxItems: 3
     -items:
       -const: a
       -const: b
       -const: c
     -items:
       -const: x
       -const: y
       -const: z

Where you have to have either [x, y, z] xor a set from [a, b, c].  One 
of the two items lists, where min/max is applied to the first one.  "-" 
on both of the items is needed since you can't seem to have the same tag 
more than one at the same scope level.

Probably this was a flawed idea from the start.

> 
>>>>    I find it disappointing that the "version" of
>>>> Yaml used for DT bindings is not documented,
>>>
>>> Not sure which part you mean? json-schema is the vocabulary which has
>>> a spec. The meta-schema then constrains what the json-schema structure
>>> should look like. That's still evolving a bit as I try to improve it
>>> based on mistakes people make. Then there's the intermediate .dt.yaml
>>> format used internally. That's supposed to stay internal and may go
>>> away when/if we integrate the validation into dtc.
>>
>> So, this is probably off-topic, but hopefully you'll find this useful.
> 
> I'm interested in knowing the pain points.
> 
>> I'm probably in the minority, but I really haven't used json-schema nor
>> yaml before.  I have experience with other "schema" languages, so I
>> figured I could pick what I need from the documentation.
> 
> Well, json-schema was new to me before this. There's definitely some
> things I really don't love about it, but it's better than trying to
> define our own language. It's generally been able to handle some of
> the more complex cases.
> 
>> The only documentation I see is writing-schema.md and example-schema.yaml
>>
>> To me, writing-schema.md is insufficient.  Its better than nothing, so
>> I'm still glad it exists, but I don't have any confidence I can really
>> write a binding yaml from scratch based on it.  It does a good thing by
>> telling you what are important properties of a binding, so based on that
>> you can kind of start to understand how existing bindings actually work.
>>    Its great in telling you how to run the validation checks (the Running
>> checks) section.  The dependencies section is awesome from my
>> perspective - most projects seem to assume you just know what their
>> dependencies are, and its painful to try to figure them out when you get
>> cryptic errors during make.
>>
>> Where it really fails is that I get no sense of the language.  As a
>> minimum a lexigraphic specification that would allow me to write a
>> compiler (I've done this before).  Then I would understand what are the
>> keywords, and where they are valid.  I wouldn't understand what they
>> mean, but at-least I can look at some implemented examples and
>> extrapolate from there.
>>
>> Have you by chance ever looked at the ACPI spec?  Maybe not the best
>> example, but its the one that comes to my mind first.  ACPI has ACPI
>> Source Language (ASL).  Its an interpreted hardware description language
>> that doesn't match yaml, but I think the ACPI spec does a reasonable job
>> of describing it.  You have a lexographic definition which seems to be
>> really helpful to ACPICA in implementing the intrepreter.  It lists all
>> of the valid operators, types, etc.  It provides detailed references of
>> each keyword - how they are used, what they do, etc.  Its not the
>> greatest at "how to write ASL 101" or "these are common problems that
>> people face, and how they can be solved", but atleast with what there
>> is, I could read every keyword that seems to be possibly related to what
>> I want to do, and hazard a guess if it would work for my problem.
> 
> I have not read the ACPI spec.
> 
>> Perhaps that is outside the scope of the writing-schema.md document,
>> that is fair.  However, I argue that the document does not provide
>> sufficient references.  The document provides a reference to the
>> json-schema spec, but the spec is kinda useless (atleast I feel that it
>> is).  "minItems" is not defined anywhere in the spec.  What does it
>> mean?  How can I use it?  Specific to minItems/maxItems, I'll I've
>> gathered about it is from example-schema.yaml which indicates its a way
>> to identify mandatory and optional values for a property, but it doesn't
>> describe the fact that order matters, and you cannot mix/match things -
>> IE it looks like you need atleast min items, and at most max items, but
>> even if you have enough items to satisfy min, there cannot be gaps (you
>> can't pick items 1, 5, 10 from the list).  I only found that out from
>> running the validation checks with trial/error.
> 
> I think you looked at the 'Core' spec rather than the 'Validation' spec:
> http://json-schema.org/draft/2019-09/json-schema-validation.html
> 
> Though that has moved on to a newer version and we're still on draft7
> which is here:
> https://tools.ietf.org/html/draft-handrews-json-schema-validation-01

Yes, that looks completely different than what I read.  Thanks for the 
direct link.  I'm going to go read it.

> 
> I guess a direct link to this with 'Details on json-schema keywords is
> here' would be helpful.

Yes please.  Or atleast a "Hey, there are actually two specs, 'core' and 
'validation'. The 'validation' one is the relevant one.  Hopefully that 
clarifies any confusion"

> 
> minItems/maxItems is the one area we deviate from json-schema
> defaults. That's what the 'Property Schema' section calls out.
> 
> Order matters for DT too, so that aspect matches up well with
> json-schema. That's been a common issue in dts files, so schema
> starting to enforce that will be good for new bindings, but somewhat
> painful for existing ones.

You are right, order does matter in DT.  I think I've gotten used to 
just having -names, and assuming if a, b, c, and d are all listed as 
optional, that means you could have a and c.  However that kind of 
breaks the index mapping, so if you have c, you really need a and b as 
well.  I was attempting to apply that concept to schema, and it wasn't 
working.  I suspect that concept shouldn't be valid normally.

> 
>> There is no reference to the yaml spec, despite the document stating
>> that the bindings are written in yaml.
>>
>> However, having found the yaml spec, its really not much better than the
>> json-schema spec, and it doesn't line up because as the document states,
>> the bindings are not really written in yaml - its a subset of yaml where
>> a ton of the boilerplate "code" is skipped.
> 
> Yeah, there's a lot to YAML that no one uses and I too find the spec
> pretty useless (hence why no reference). Like most other uses I've
> encountered, we're using a JSON compatible subset which is just lists
> and dicts of key/value pairs. The main thing folks need to know and
> trip up on are: indentation is important (including no tabs) and pay
> attention to '-' or lack of.
> 
>> What is boilerplate that is skipped?  IMO, if you are not strictly
>> adhering to yaml, then you need to clearly document your own derivative
>> language so that someone like me whom is being introduced to all of this
>> for the first time can start to figure out some of it.  It would be
>> helpful to look at other yaml examples, and understand what is
>> considered to be boilerplate so I can translate that to a DT binding.
> 
> We're not skipping any boilerplate. We're not using advanced features
> like tags or anchors. You can use any YAML parser including online
> ones to read the files.

Ok, so I feel like I've misunderstood this except from writing-schema.md:

"The Devicetree schemas don't exactly match the YAML encoded DT data 
produced by dtc. They are simplified to make them more compact and avoid 
a bunch of boilerplate."

I thought this meant the bindings were simplified to be more readable, 
by skipping boilerplate text.  What does it actually mean?

> 
>> I understand, the majority of the above is complaints and demands which
>> is really not fair to you, since you are spending what I presume to be
>> your "non-dayjob" time to make the community better.
> 
> It's my day job or part of it, just not enough hours in the day...
> 
>> However, I don't
>> really know how to contribute to make the documentation better.  I don't
>> understand enough.  As far as this topic is concerned, I'm a dumb monkey
>> banging on a keyboard hoping to get close enough to Shakespeare to pass
>> mustard by accident, and maybe learn something along the way so that
>> next time, I might have an idea of how to do something of what I need.
> 
> The challenge is providing enough information to write bindings
> without being json-schema experts. My hope is really to build up
> enough examples and make the meta-schema good enough to keep folks
> within the lines. Maybe that's a flawed approach, but even getting
> folks to follow writing-schema.rst and run 'make dt_binding_check' has
> been a challenge.
> 
>> Hopefully you've made it this far - that ended up being a lot more text
>> that I thought it would be.  I really hope this is useful feedback to
>> you, but let me know if I am still not clear on something.  I will try
>> my best to clarify more.  If you feel like I can contribute somehow,
>> just let me know.
>>
>>>
>>>> so after several hours of
>>>> trial and error, I just gave up since I found this to work (failed cases
>>>> just gave me an error with no indication of what was wrong, not even a
>>>> line number).
>>>
>>> Schema failures or dts failures? It is possible to get line numbers
>>> for either, but that makes validation much slower. In the latter case,
>>> the line numbers aren't too useful either given they are for the
>>> .dt.yaml file and not the .dts source file (dtc integration would
>>> solve that). Adding '-n' to dt-doc-validate or dt-validate will turn
>>> them on though.
>>
>> Schema compilation failures.  I don't recall the exact error message,
>> but it was something like "no valid schema found, continuing".
>> Essentially running "dt_binding_check".  I tried with -v but wasn't
>> getting much more in this case.  I didn't try -n.
> 
> That's before we even validate the schema, so something has gone wrong
> pretty early. You may get farther with 'make -k'. I'll have to look
> into it. The schemas are actually built twice. They are all built into
> processed-schema.yaml. That's supposed to skip any with errors and is
> what's used to validate dts files. If that's failing for some reason,
> then it's going to be pretty vague. The dt_binding_check rule also
> fully validates each binding schema and builds and validates the
> examples. It should print more detailed errors (though still sometimes
> vague).
> 
> Rob
> 


-- 
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.

^ permalink raw reply	[flat|nested] 16+ messages in thread

end of thread, other threads:[~2019-11-12 22:03 UTC | newest]

Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-08 23:16 [PATCH v8 0/4] MSM8998 Multimedia Clock Controller Jeffrey Hugo
2019-11-08 23:17 ` [PATCH v8 1/4] dt-bindings: clock: Document external clocks for MSM8998 gcc Jeffrey Hugo
2019-11-12  0:44   ` Rob Herring
2019-11-12 16:25     ` Jeffrey Hugo
2019-11-12 18:37       ` Rob Herring
2019-11-12 19:38         ` Jeffrey Hugo
2019-11-12 21:18           ` Rob Herring
2019-11-12 22:03             ` Jeffrey Hugo
2019-11-08 23:17 ` [PATCH v8 2/4] dt-bindings: clock: Convert qcom,mmcc to DT schema Jeffrey Hugo
2019-11-12  0:58   ` Rob Herring
2019-11-12 16:11     ` Jeffrey Hugo
2019-11-08 23:17 ` [PATCH v8 3/4] dt-bindings: clock: Add support for the MSM8998 mmcc Jeffrey Hugo
2019-11-12  0:55   ` Rob Herring
2019-11-12 16:07     ` Jeffrey Hugo
2019-11-12  0:59   ` Rob Herring
2019-11-08 23:18 ` [PATCH v8 4/4] clk: qcom: Add MSM8998 Multimedia Clock Controller (MMCC) driver Jeffrey Hugo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).