dt-bindings: drm/msm: Update bindings for MDP5
PATCH 1/2 dt-bindings: drm/msm: Update bindings for MDP5
MDP5 has a different compatible string (which causes checkpatch warnings when we try to add MDP5 device nodes). It also has different list of clocks.
Update the bindings for MDP5.
--- Documentation/devicetree/bindings/drm/msm/mdp.txt | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-)
Documentation/devicetree/bindings/drm/msm/mdp.txt +16 -7
@@ -3,18 +3,27 @@ Qualcomm adreno/snapdragon display controller Required properties: - compatible: * "qcom,mdp" - mdp4 + * "qcom,mdss_mdp" - mdp5 - reg: Physical base address and length of the controller's registers. - interrupts: The interrupt signal from the display controller. - connectors: array of phandles for output device(s) - clocks: device clocks See ../clocks/clock-bindings.txt for details. -- clock-names: the following clocks are required: - * "core_clk" - * "iface_clk" - * "lut_clk" - * "src_clk" - * "hdmi_clk" - * "mpd_clk" +- clock-names: the following clocks are required. + For MDP4: + * "core_clk" + * "iface_clk" + * "lut_clk" + * "src_clk" + * "hdmi_clk" + * "mdp_clk" + For MDP5: + * "bus_clk" + * "iface_clk" + * "core_clk_src" + * "core_clk" + * "lut_clk" (some MDP5 versions may not need this) + * "vsync_clk" Optional properties: - gpus: phandle for gpu device
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
PATCH 2/2 arm64: dtsi: Add device nodes for GPU and MDSS on MSM8916
Add device nodes for gpu, mdp and dsi blocks.
The gpu and dsi nodes are missing the 'power-domains' property needed for configuring GDSC. This needs to be added when its available.
--- arch/arm64/boot/dts/qcom/msm8916.dtsi | 105 ++++++++++++++++++++++++++++++++++ 1 file changed, 105 insertions(+)
arch/arm64/boot/dts/qcom/msm8916.dtsi +105 -0
@@ -391,6 +391,111 @@ interrupt-controller; #interrupt-cells = <4>; }; + + gpu: qcom,adreno-3xx@01c00000 { + compatible = "qcom,adreno-3xx"; + #stream-id-cells = <16>; + reg = <0x01c00000 0x20000>; + reg-names = "kgsl_3d0_reg_memory"; + interrupts = <GIC_SPI 33 IRQ_TYPE_NONE>; + interrupt-names = "kgsl_3d0_irq"; + clocks = + <&gcc GCC_OXILI_GFX3D_CLK>, + <&gcc GCC_OXILI_AHB_CLK>, + <&gcc GCC_OXILI_GMEM_CLK>, + <&gcc GCC_BIMC_GFX_CLK>, + <&gcc GCC_BIMC_GPU_CLK>, + <&gcc GFX3D_CLK_SRC>; + clock-names = + "core_clk", + "iface_clk", + "mem_clk", + "mem_iface_clk", + "alt_mem_iface_clk", + "gfx3d_clk_src"; + + qcom,chipid = <0x03000600>; + qcom,gpu-pwrlevels { + compatible = "qcom,gpu-pwrlevels"; + qcom,gpu-pwrlevel@0 { + qcom,gpu-freq = <400000000>; + }; + qcom,gpu-pwrlevel@1 { + qcom,gpu-freq = <19200000>; + }; + }; + }; + + mdss_mdp: qcom,mdss_mdp@1a00000 { + compatible = "qcom,mdss_mdp"; + reg = <0x1a00000 0x90000>, + <0x1ac8000 0x3000>; + reg-names = "mdp_phys", + "vbif_phys"; + interrupts = <GIC_SPI 72 IRQ_TYPE_NONE>; + + interrupt-controller; + #interrupt-cells = <1>; + + connectors = <&mdss_dsi0>; + gpus = <&gpu>; + + clocks = <&gcc GCC_MDSS_AHB_CLK>, + <&gcc GCC_MDSS_AXI_CLK>, + <&gcc MDP_CLK_SRC>, + <&gcc GCC_MDSS_MDP_CLK>, + <&gcc GCC_MDSS_VSYNC_CLK>; + clock-names = "iface_clk", + "bus_clk", + "core_clk_src", + "core_clk", + "vsync_clk"; + }; + + mdss_dsi0: qcom,mdss_dsi@1a98000 { + compatible = "qcom,mdss-dsi-ctrl"; + qcom,dsi-host-index = <0>; + #address-cells = <1>; + #size-cells = <0>; + reg = <0x1a98000 0x25c>; + reg-names = "dsi_ctrl"; + interrupt-parent = <&mdss_mdp>; + interrupts = <4 IRQ_TYPE_NONE>; + + clocks = <&gcc GCC_MDSS_MDP_CLK>, + <&gcc GCC_MDSS_AHB_CLK>, + <&gcc GCC_MDSS_AXI_CLK>, + <&gcc GCC_MDSS_BYTE0_CLK>, + <&gcc GCC_MDSS_PCLK0_CLK>, + <&gcc GCC_MDSS_ESC0_CLK>, + <&gcc BYTE0_CLK_SRC>, + <&gcc PCLK0_CLK_SRC>; + clock-names = "mdp_core_clk", + "iface_clk", + "bus_clk", + "byte_clk", + "pixel_clk", + "core_clk", + "byte_clk_src", + "pixel_clk_src"; + qcom,dsi-phy = <&mdss_dsi_phy0>; + status = "disabled"; + }; + + mdss_dsi_phy0: qcom,mdss_dsi_phy@1a98300 { + compatible = "qcom,dsi-phy-28nm-lp"; + qcom,dsi-phy-index = <0>; + reg = <0x1a98300 0xd4>, + <0x1a98500 0x280>, + <0x1a98780 0x30>; + reg-names = "dsi_pll", + "dsi_phy", + "dsi_phy_regulator"; + + clocks = <&gcc GCC_MDSS_AHB_CLK>; + clock-names = "iface_clk"; + status = "disabled"; + }; }; };
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
Quoted 16 lines · Archit Taneja
MDP5 has a different compatible string (which causes checkpatch warnings when we try to add MDP5 device nodes). It also has different list of clocks.
Update the bindings for MDP5.
Signed-off-by Archit Taneja <[email protected]>--- Documentation/devicetree/bindings/drm/msm/mdp.txt | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-)Documentation/devicetree/bindings/drm/msm/mdp.txt +0
-0
Documentation/devicetree/bindings/drm/msm/mdp.txt Qualcomm adreno/snapdragon display controller · line 6
Quoted 4 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt +1
-0@@ -3,18 +3,27 @@ Qualcomm adreno/snapdragon display controller Required properties: - compatible: * "qcom,mdp" - mdp4 + * "qcom,mdss_mdp" - mdp5
Wouldn't it be better to name these "qcom,mdp4" and "qcom,mdp5" so that we don't have to rely on some magic detection mechanism to support future versions?
Regards,
Bjorn
Quoted 12 lines · Bjorn Andersson
Quoted 11 lines · Archit Taneja
MDP5 has a different compatible string (which causes checkpatch warnings when we try to add MDP5 device nodes). It also has different list of clocks.
Update the bindings for MDP5.
Signed-off-by Archit Taneja <[email protected]>--- Documentation/devicetree/bindings/drm/msm/mdp.txt | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-)
Quoted 13 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt Qualcomm adreno/snapdragon display controller · line 6
Quoted 9 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt +1
-0@@ -3,18 +3,27 @@ Qualcomm adreno/snapdragon display controller Required properties: - compatible: * "qcom,mdp" - mdp4 + * "qcom,mdss_mdp" - mdp5Wouldn't it be better to name these "qcom,mdp4" and "qcom,mdp5" so that we don't have to rely on some magic detection mechanism to support future versions?
My guess is that compatibility string "mdss_mdp" came from the downstream kernels. Chips having MDP4 were supported on downstream kernels that preceded DT.
Since we don't have any MDP DT upstream, I think we could go with the string names you suggested. It might break things for people who might try to use an old downstream dts file with a new kernel. But that doesn't sound like the worst thing.
Archit
Quoted 4 lines
Regards,
Bjorn
Signature · 5 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 35 lines · Archit Taneja
Quoted 16 lines · Bjorn Andersson
Quoted 13 lines · Archit Taneja
MDP5 has a different compatible string (which causes checkpatch warnings when we try to add MDP5 device nodes). It also has different list of clocks.
Update the bindings for MDP5.
Signed-off-by Archit Taneja <[email protected]>--- Documentation/devicetree/bindings/drm/msm/mdp.txt | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-)Quoted 15 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt Qualcomm adreno/snapdragon display controller · line 6
Quoted 10 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt +1
-0@@ -3,18 +3,27 @@ Qualcomm adreno/snapdragon display controller Required properties: - compatible: * "qcom,mdp" - mdp4 + * "qcom,mdss_mdp" - mdp5Wouldn't it be better to name these "qcom,mdp4" and "qcom,mdp5" so that we don't have to rely on some magic detection mechanism to support future versions?
Quoted 8 lines
My guess is that compatibility string "
mdss_mdp" came from the downstream kernels. Chips having MDP4 were supported on downstream kernels that preceded DT.Since we don't have any MDP DT upstream, I think we could go with the string names you suggested. It might break things for people who might try to use an old downstream dts file with a new kernel. But that doesn't sound like the worst thing.
it would break things for downstream 3.10/mdp5 dt files.. but what I'd suggest is to make the code look for both qcom,mdp5 and qcom,mdss_mdp but only document the former
BR,
-R
Quoted 10 lines
Archit
Quoted 4 lines
Regards,
Bjorn--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
Quoted 40 lines · Rob Clark
Quoted 39 lines · Archit Taneja
Quoted 16 lines · Bjorn Andersson
Quoted 13 lines · Archit Taneja
MDP5 has a different compatible string (which causes checkpatch warnings when we try to add MDP5 device nodes). It also has different list of clocks.
Update the bindings for MDP5.
Signed-off-by Archit Taneja <[email protected]>--- Documentation/devicetree/bindings/drm/msm/mdp.txt | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-)Quoted 15 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt Qualcomm adreno/snapdragon display controller · line 6
Quoted 10 lines
Documentation/devicetree/bindings/drm/msm/mdp.txt +1
-0@@ -3,18 +3,27 @@ Qualcomm adreno/snapdragon display controller Required properties: - compatible: * "qcom,mdp" - mdp4 + * "qcom,mdss_mdp" - mdp5Wouldn't it be better to name these "qcom,mdp4" and "qcom,mdp5" so that we don't have to rely on some magic detection mechanism to support future versions?
My guess is that compatibility string "
mdss_mdp" came from the downstream kernels. Chips having MDP4 were supported on downstream kernels that preceded DT.
Quoted 8 lines
Quoted 4 lines
Since we don't have any MDP DT upstream, I think we could go with the string names you suggested. It might break things for people who might try to use an old downstream dts file with a new kernel. But that doesn't sound like the worst thing.
it would break things for downstream 3.10/mdp5 dt files.. but what I'd suggest is to make the code look for both qcom,mdp5 and qcom,mdss_mdp but only document the former
That's a good idea. I'll update the patches to do that.
Archit
Quoted 14 lines
BR,
-RQuoted 10 lines
Archit
Quoted 4 lines
Regards,
Bjorn--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project