Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
doesn't have an HWCFG block but a separate register set.
The missing registers are added in the a6xx.xml.h file that would
require a subsequent sync and the non-existent hwcfg is handled
in a6xx_set_hwcg().
The A750 GPU info are added under the adreno_is_a750() macro and
the ADRENO_7XX_GEN3 family id.
This adds:
- the GMU and SMMU bindings
- DRM driver changes
- DT nodes
Dependencies: None
Tested using Mesa's !26934 Merge Request [0] on the SM8650-QRD
and with kmscube & vkcube to test basic rendering.
[0] https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26934
Signed-off-by: Neil Armstrong <[email protected]>
---
Neil Armstrong (5):
dt-bindings: display/msm/gmu: Document Adreno 750 GMU
dt-bindings: arm-smmu: Document SM8650 GPU SMMU
drm: msm: add support for A750 GPU
arm64: dts: qcom: sm8650: add GPU nodes
arm64: dts: qcom: sm8650-qrd: enable GPU
.../devicetree/bindings/display/msm/gmu.yaml | 1 +
.../devicetree/bindings/iommu/arm,smmu.yaml | 7 +-
arch/arm64/boot/dts/qcom/sm8650-qrd.dts | 8 +
arch/arm64/boot/dts/qcom/sm8650.dtsi | 169 +++++++++++++++++++++
drivers/gpu/drm/msm/adreno/a6xx.xml.h | 8 +
drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 2 +
drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 29 +++-
drivers/gpu/drm/msm/adreno/adreno_device.c | 14 ++
drivers/gpu/drm/msm/adreno/adreno_gpu.h | 9 +-
9 files changed, 241 insertions(+), 6 deletions(-)
---
base-commit: 84baf172e2fa30d6d6d0fb8ed076b47e836b74f1
change-id: 20240208-topic-sm8650-gpu-489d5e2c2b17
Best regards,
--
Neil Armstrong <[email protected]>
Document the GPU SMMU found on the SM8650 platform.
Signed-off-by: Neil Armstrong <[email protected]>
---
Documentation/devicetree/bindings/iommu/arm,smmu.yaml | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/Documentation/devicetree/bindings/iommu/arm,smmu.yaml b/Documentation/devicetree/bindings/iommu/arm,smmu.yaml
index a4042ae24770..3ad5c850f3bf 100644
--- a/Documentation/devicetree/bindings/iommu/arm,smmu.yaml
+++ b/Documentation/devicetree/bindings/iommu/arm,smmu.yaml
@@ -93,6 +93,7 @@ properties:
- qcom,sm8350-smmu-500
- qcom,sm8450-smmu-500
- qcom,sm8550-smmu-500
+ - qcom,sm8650-smmu-500
- const: qcom,adreno-smmu
- const: qcom,smmu-500
- const: arm,mmu-500
@@ -508,7 +509,10 @@ allOf:
- if:
properties:
compatible:
- const: qcom,sm8550-smmu-500
+ contains:
+ enum:
+ - qcom,sm8550-smmu-500
+ - qcom,sm8650-smmu-500
then:
properties:
clock-names:
@@ -544,7 +548,6 @@ allOf:
- qcom,sdx65-smmu-500
- qcom,sm6350-smmu-500
- qcom,sm6375-smmu-500
- - qcom,sm8650-smmu-500
- qcom,x1e80100-smmu-500
then:
properties:
--
2.34.1
Add support for the A750 GPU found on the SM8650 platform
Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
doesn't have an HWCFG block but a separate register set.
The missing registers are added in the a6xx.xml.h file that would
require a subsequent sync and the non-existent hwcfg is handled
in a6xx_set_hwcg().
The A750 GPU info are added under the adreno_is_a750() macro and
the ADRENO_7XX_GEN3 family id.
Signed-off-by: Neil Armstrong <[email protected]>
---
drivers/gpu/drm/msm/adreno/a6xx.xml.h | 8 ++++++++
drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 2 ++
drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 29 ++++++++++++++++++++++++++---
drivers/gpu/drm/msm/adreno/adreno_device.c | 14 ++++++++++++++
drivers/gpu/drm/msm/adreno/adreno_gpu.h | 9 ++++++++-
5 files changed, 58 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/msm/adreno/a6xx.xml.h b/drivers/gpu/drm/msm/adreno/a6xx.xml.h
index 863b5e3b0e67..c4db4e0c0819 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a6xx.xml.h
@@ -1725,6 +1725,8 @@ static inline uint32_t REG_A6XX_RBBM_PERFCTR_RBBM_SEL(uint32_t i0) { return 0x00
#define REG_A6XX_RBBM_BLOCK_SW_RESET_CMD2 0x00000046
+#define REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL 0x000000ad
+
#define REG_A6XX_RBBM_CLOCK_CNTL 0x000000ae
#define REG_A6XX_RBBM_CLOCK_CNTL_SP0 0x000000b0
@@ -1939,12 +1941,18 @@ static inline uint32_t REG_A6XX_RBBM_PERFCTR_RBBM_SEL(uint32_t i0) { return 0x00
#define REG_A6XX_RBBM_CLOCK_HYST_HLSQ 0x0000011d
+#define REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD 0x0000011e
+
+#define REG_A7XX_RBBM_CGC_P2S_TRIG_CMD 0x0000011f
+
#define REG_A6XX_RBBM_CLOCK_CNTL_TEX_FCHE 0x00000120
#define REG_A6XX_RBBM_CLOCK_DELAY_TEX_FCHE 0x00000121
#define REG_A6XX_RBBM_CLOCK_HYST_TEX_FCHE 0x00000122
+#define REG_A7XX_RBBM_CGC_P2S_STATUS 0x00000122
+
#define REG_A7XX_RBBM_CLOCK_HYST2_VFD 0x0000012f
#define REG_A6XX_RBBM_LPAC_GBIF_CLIENT_QOS_CNTL 0x000005ff
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
index 8c4900444b2c..325881d8ff08 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
@@ -842,6 +842,8 @@ static int a6xx_gmu_fw_start(struct a6xx_gmu *gmu, unsigned int state)
*/
if (adreno_is_a740(adreno_gpu))
chipid_min = 2;
+ else if (adreno_is_a750(adreno_gpu))
+ chipid_min = 9;
else
return -EINVAL;
diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
index c0bc924cd302..472991584053 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
@@ -958,10 +958,11 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
const struct adreno_reglist *reg;
+ bool skip_programming = !(adreno_gpu->info->hwcg || adreno_is_a7xx(adreno_gpu));
unsigned int i;
u32 val, clock_cntl_on, cgc_mode;
- if (!adreno_gpu->info->hwcg)
+ if (skip_programming)
return;
if (adreno_is_a630(adreno_gpu))
@@ -982,6 +983,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
state ? 0x5555 : 0);
}
+ if (!adreno_gpu->info->hwcg) {
+ gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
+ gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
+
+ if (state) {
+ gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
+
+ if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
+ val & BIT(0), 1, 10)) {
+ dev_err(&gpu->pdev->dev, "RBBM_CGC_P2S_STATUS TXDONE Poll failed\n");
+ return;
+ }
+
+ gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 0);
+ }
+
+ return;
+ }
+
val = gpu_read(gpu, REG_A6XX_RBBM_CLOCK_CNTL);
/* Don't re-program the registers if they are already correct */
@@ -1239,7 +1259,9 @@ static void a6xx_set_cp_protect(struct msm_gpu *gpu)
count = ARRAY_SIZE(a660_protect);
count_max = 48;
BUILD_BUG_ON(ARRAY_SIZE(a660_protect) > 48);
- } else if (adreno_is_a730(adreno_gpu) || adreno_is_a740(adreno_gpu)) {
+ } else if (adreno_is_a730(adreno_gpu) ||
+ adreno_is_a740(adreno_gpu) ||
+ adreno_is_a750(adreno_gpu)) {
regs = a730_protect;
count = ARRAY_SIZE(a730_protect);
count_max = 48;
@@ -2880,7 +2902,8 @@ struct msm_gpu *a6xx_gpu_init(struct drm_device *dev)
/* gpu->info only gets assigned in adreno_gpu_init() */
is_a7xx = config->info->family == ADRENO_7XX_GEN1 ||
- config->info->family == ADRENO_7XX_GEN2;
+ config->info->family == ADRENO_7XX_GEN2 ||
+ config->info->family == ADRENO_7XX_GEN3;
a6xx_llc_slices_init(pdev, a6xx_gpu, is_a7xx);
diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c
index 2ce7d7b1690d..e2582c91d7e7 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_device.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_device.c
@@ -522,6 +522,20 @@ static const struct adreno_info gpulist[] = {
.zapfw = "a740_zap.mdt",
.hwcg = a740_hwcg,
.address_space_size = SZ_16G,
+ }, {
+ .chip_ids = ADRENO_CHIP_IDS(0x43051401), /* "C520v2" */
+ .family = ADRENO_7XX_GEN3,
+ .fw = {
+ [ADRENO_FW_SQE] = "gen70900_sqe.fw",
+ [ADRENO_FW_GMU] = "gmu_gen70900.bin",
+ },
+ .gmem = 3 * SZ_1M,
+ .inactive_period = DRM_MSM_INACTIVE_PERIOD,
+ .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT |
+ ADRENO_QUIRK_HAS_HW_APRIV,
+ .init = a6xx_gpu_init,
+ .zapfw = "gen70900_zap.mbn",
+ .address_space_size = SZ_16G,
},
};
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
index bc14df96feb0..744fa18067f8 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
@@ -48,6 +48,7 @@ enum adreno_family {
ADRENO_6XX_GEN4, /* a660 family */
ADRENO_7XX_GEN1, /* a730 family */
ADRENO_7XX_GEN2, /* a740 family */
+ ADRENO_7XX_GEN3, /* a750 family */
};
#define ADRENO_QUIRK_TWO_PASS_USE_WFI BIT(0)
@@ -423,12 +424,18 @@ static inline int adreno_is_a740(struct adreno_gpu *gpu)
return gpu->info->chip_ids[0] == 0x43050a01;
}
+static inline int adreno_is_a750(struct adreno_gpu *gpu)
+{
+ return gpu->info->chip_ids[0] == 0x43051401;
+}
+
/* Placeholder to make future diffs smaller */
static inline int adreno_is_a740_family(struct adreno_gpu *gpu)
{
if (WARN_ON_ONCE(!gpu->info))
return false;
- return gpu->info->family == ADRENO_7XX_GEN2;
+ return gpu->info->family == ADRENO_7XX_GEN2 ||
+ gpu->info->family == ADRENO_7XX_GEN3;
}
static inline int adreno_is_a7xx(struct adreno_gpu *gpu)
--
2.34.1
Add GPU nodes for the SM8650 platform.
Signed-off-by: Neil Armstrong <[email protected]>
---
arch/arm64/boot/dts/qcom/sm8650.dtsi | 169 +++++++++++++++++++++++++++++++++++
1 file changed, 169 insertions(+)
diff --git a/arch/arm64/boot/dts/qcom/sm8650.dtsi b/arch/arm64/boot/dts/qcom/sm8650.dtsi
index c455ca4e6475..f6f9e603fb2f 100644
--- a/arch/arm64/boot/dts/qcom/sm8650.dtsi
+++ b/arch/arm64/boot/dts/qcom/sm8650.dtsi
@@ -2582,6 +2582,131 @@ tcsr: clock-controller@1fc0000 {
#reset-cells = <1>;
};
+ gpu: gpu@3d00000 {
+ compatible = "qcom,adreno-43051401", "qcom,adreno";
+ reg = <0x0 0x03d00000 0x0 0x40000>,
+ <0x0 0x03d9e000 0x0 0x1000>,
+ <0x0 0x03d61000 0x0 0x800>;
+ reg-names = "kgsl_3d0_reg_memory",
+ "cx_mem",
+ "cx_dbgc";
+
+ interrupts = <GIC_SPI 300 IRQ_TYPE_LEVEL_HIGH>;
+
+ iommus = <&adreno_smmu 0 0x0>,
+ <&adreno_smmu 1 0x0>;
+
+ operating-points-v2 = <&gpu_opp_table>;
+
+ interconnects = <&gem_noc MASTER_GFX3D 0 &mc_virt SLAVE_EBI1 0>;
+ interconnect-names = "gfx-mem";
+
+ qcom,gmu = <&gmu>;
+
+ status = "disabled";
+
+ zap-shader {
+ memory-region = <&gpu_micro_code_mem>;
+ };
+
+ /* Speedbin needs more work on A740+, keep only lower freqs */
+ gpu_opp_table: opp-table {
+ compatible = "operating-points-v2";
+
+ opp-680000000 {
+ opp-hz = /bits/ 64 <680000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_SVS_L1>;
+ };
+
+ opp-629000000 {
+ opp-hz = /bits/ 64 <629000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_SVS_L0>;
+ };
+
+ opp-578000000 {
+ opp-hz = /bits/ 64 <578000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_SVS>;
+ };
+
+ opp-500000000 {
+ opp-hz = /bits/ 64 <500000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS_L1>;
+ };
+
+ opp-422000000 {
+ opp-hz = /bits/ 64 <422000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS>;
+ };
+
+ opp-366000000 {
+ opp-hz = /bits/ 64 <366000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS_D0>;
+ };
+
+ opp-310000000 {
+ opp-hz = /bits/ 64 <310000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS_D1>;
+ };
+
+ opp-231000000 {
+ opp-hz = /bits/ 64 <231000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS_D2>;
+ };
+ };
+ };
+
+ gmu: gmu@3d6a000 {
+ compatible = "qcom,adreno-gmu-750.1", "qcom,adreno-gmu";
+ reg = <0x0 0x03d6a000 0x0 0x35000>,
+ <0x0 0x03d50000 0x0 0x10000>,
+ <0x0 0x0b280000 0x0 0x10000>;
+ reg-names = "gmu", "rscc", "gmu_pdc";
+
+ interrupts = <GIC_SPI 304 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 305 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "hfi", "gmu";
+
+ clocks = <&gpucc GPU_CC_AHB_CLK>,
+ <&gpucc GPU_CC_CX_GMU_CLK>,
+ <&gpucc GPU_CC_CXO_CLK>,
+ <&gcc GCC_DDRSS_GPU_AXI_CLK>,
+ <&gcc GCC_GPU_MEMNOC_GFX_CLK>,
+ <&gpucc GPU_CC_HUB_CX_INT_CLK>,
+ <&gpucc GPU_CC_DEMET_CLK>;
+ clock-names = "ahb",
+ "gmu",
+ "cxo",
+ "axi",
+ "memnoc",
+ "hub",
+ "demet";
+
+ power-domains = <&gpucc GPU_CX_GDSC>,
+ <&gpucc GPU_GX_GDSC>;
+ power-domain-names = "cx",
+ "gx";
+
+ iommus = <&adreno_smmu 5 0x0>;
+
+ qcom,qmp = <&aoss_qmp>;
+
+ operating-points-v2 = <&gmu_opp_table>;
+
+ gmu_opp_table: opp-table {
+ compatible = "operating-points-v2";
+
+ opp-625000000 {
+ opp-hz = /bits/ 64 <625000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_SVS>;
+ };
+
+ opp-260000000 {
+ opp-hz = /bits/ 64 <260000000>;
+ opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS>;
+ };
+ };
+ };
+
gpucc: clock-controller@3d90000 {
compatible = "qcom,sm8650-gpucc";
reg = <0 0x03d90000 0 0xa000>;
@@ -2595,6 +2720,50 @@ gpucc: clock-controller@3d90000 {
#power-domain-cells = <1>;
};
+ adreno_smmu: iommu@3da0000 {
+ compatible = "qcom,sm8650-smmu-500", "qcom,adreno-smmu",
+ "qcom,smmu-500", "arm,mmu-500";
+ reg = <0x0 0x03da0000 0x0 0x40000>;
+ #iommu-cells = <2>;
+ #global-interrupts = <1>;
+ interrupts = <GIC_SPI 673 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 677 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 678 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 679 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 680 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 681 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 682 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 683 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 684 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 685 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 686 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 687 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 422 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 476 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 574 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 575 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 576 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 577 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 659 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 661 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 664 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 665 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 666 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 668 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 669 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 699 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&gpucc GPU_CC_HLOS1_VOTE_GPU_SMMU_CLK>,
+ <&gcc GCC_GPU_MEMNOC_GFX_CLK>,
+ <&gcc GCC_GPU_SNOC_DVM_GFX_CLK>,
+ <&gpucc GPU_CC_AHB_CLK>;
+ clock-names = "hlos",
+ "bus",
+ "iface",
+ "ahb";
+ power-domains = <&gpucc GPU_CX_GDSC>;
+ dma-coherent;
+ };
+
ipa: ipa@3f40000 {
compatible = "qcom,sm8650-ipa", "qcom,sm8550-ipa";
--
2.34.1
Document the Adreno 750 GMU found on the SM8650 platform.
Signed-off-by: Neil Armstrong <[email protected]>
---
Documentation/devicetree/bindings/display/msm/gmu.yaml | 1 +
1 file changed, 1 insertion(+)
diff --git a/Documentation/devicetree/bindings/display/msm/gmu.yaml b/Documentation/devicetree/bindings/display/msm/gmu.yaml
index 4e1c25b42908..b3837368a260 100644
--- a/Documentation/devicetree/bindings/display/msm/gmu.yaml
+++ b/Documentation/devicetree/bindings/display/msm/gmu.yaml
@@ -224,6 +224,7 @@ allOf:
enum:
- qcom,adreno-gmu-730.1
- qcom,adreno-gmu-740.1
+ - qcom,adreno-gmu-750.1
then:
properties:
reg:
--
2.34.1
Add path of the GPU firmware for the SM8650-QRD board
Signed-off-by: Neil Armstrong <[email protected]>
---
arch/arm64/boot/dts/qcom/sm8650-qrd.dts | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
index 8515498553bf..7151f3dc67c4 100644
--- a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
+++ b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
@@ -525,6 +525,14 @@ &ipa {
status = "okay";
};
+&gpu {
+ status = "okay";
+
+ zap-shader {
+ firmware-name = "qcom/sm8650/gen70900_zap.mbn";
+ };
+};
+
&mdss {
status = "okay";
};
--
2.34.1
On 12.02.2024 11:37, Neil Armstrong wrote:
> Document the Adreno 750 GMU found on the SM8650 platform.
>
> Signed-off-by: Neil Armstrong <[email protected]>
> ---
Reviewed-by: Konrad Dybcio <[email protected]>
Konrad
On 12.02.2024 11:37, Neil Armstrong wrote:
> Add support for the A750 GPU found on the SM8650 platform
>
> Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
> doesn't have an HWCFG block but a separate register set.
>
> The missing registers are added in the a6xx.xml.h file that would
> require a subsequent sync and the non-existent hwcfg is handled
> in a6xx_set_hwcg().
These should also be submitted to mesa to make sure the next header sync
doesn't wipe them
[...]
> --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> @@ -958,10 +958,11 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
> struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
> const struct adreno_reglist *reg;
> + bool skip_programming = !(adreno_gpu->info->hwcg || adreno_is_a7xx(adreno_gpu));
is_a750?
> unsigned int i;
> u32 val, clock_cntl_on, cgc_mode;
>
> - if (!adreno_gpu->info->hwcg)
> + if (skip_programming)
> return;
>
> if (adreno_is_a630(adreno_gpu))
> @@ -982,6 +983,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
> state ? 0x5555 : 0);
> }
>
> + if (!adreno_gpu->info->hwcg) {
I don't think this block of code is reachable now, no?
Maybe remove the skip_programming and if_a750 here?
> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
> +
> + if (state) {
> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
> +
> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
> + val & BIT(0), 1, 10)) {
We should define that bit name (the err suggests it's
REG_A7XX_RBBM_GCC_P2S_STATUS_TXDONE or so)
[...]
> +static inline int adreno_is_a750(struct adreno_gpu *gpu)
> +{
> + return gpu->info->chip_ids[0] == 0x43051401;
> +}
> +
> /* Placeholder to make future diffs smaller */
Please also remove this comment now that it's invalid
Konrad
On 12.02.2024 11:37, Neil Armstrong wrote:
> Add GPU nodes for the SM8650 platform.
>
> Signed-off-by: Neil Armstrong <[email protected]>
> ---
> arch/arm64/boot/dts/qcom/sm8650.dtsi | 169 +++++++++++++++++++++++++++++++++++
> 1 file changed, 169 insertions(+)
>
> diff --git a/arch/arm64/boot/dts/qcom/sm8650.dtsi b/arch/arm64/boot/dts/qcom/sm8650.dtsi
> index c455ca4e6475..f6f9e603fb2f 100644
> --- a/arch/arm64/boot/dts/qcom/sm8650.dtsi
> +++ b/arch/arm64/boot/dts/qcom/sm8650.dtsi
> @@ -2582,6 +2582,131 @@ tcsr: clock-controller@1fc0000 {
> #reset-cells = <1>;
> };
>
> + gpu: gpu@3d00000 {
> + compatible = "qcom,adreno-43051401", "qcom,adreno";
> + reg = <0x0 0x03d00000 0x0 0x40000>,
> + <0x0 0x03d9e000 0x0 0x1000>,
> + <0x0 0x03d61000 0x0 0x800>;
> + reg-names = "kgsl_3d0_reg_memory",
> + "cx_mem",
> + "cx_dbgc";
> +
> + interrupts = <GIC_SPI 300 IRQ_TYPE_LEVEL_HIGH>;
> +
> + iommus = <&adreno_smmu 0 0x0>,
> + <&adreno_smmu 1 0x0>;
> +
> + operating-points-v2 = <&gpu_opp_table>;
> +
> + interconnects = <&gem_noc MASTER_GFX3D 0 &mc_virt SLAVE_EBI1 0>;
> + interconnect-names = "gfx-mem";
ICC should be unnecessary
Konrad
On 12.02.2024 11:37, Neil Armstrong wrote:
> Add path of the GPU firmware for the SM8650-QRD board
>
> Signed-off-by: Neil Armstrong <[email protected]>
> ---
Reviewed-by: Konrad Dybcio <[email protected]>
Konrad
On Mon, 12 Feb 2024 at 12:37, Neil Armstrong <[email protected]> wrote:
>
> Add path of the GPU firmware for the SM8650-QRD board
>
> Signed-off-by: Neil Armstrong <[email protected]>
> ---
> arch/arm64/boot/dts/qcom/sm8650-qrd.dts | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
> index 8515498553bf..7151f3dc67c4 100644
> --- a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
> +++ b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
> @@ -525,6 +525,14 @@ &ipa {
> status = "okay";
> };
>
> +&gpu {
> + status = "okay";
> +
> + zap-shader {
> + firmware-name = "qcom/sm8650/gen70900_zap.mbn";
But why? For sm8450 / sm8550 we had "a730_zap.mbn" and "a740_zap.mbn"
> + };
> +};
> +
> &mdss {
> status = "okay";
> };
>
> --
> 2.34.1
>
--
With best wishes
Dmitry
Hi,
On 12/02/2024 14:32, Dmitry Baryshkov wrote:
> On Mon, 12 Feb 2024 at 12:37, Neil Armstrong <[email protected]> wrote:
>>
>> Add path of the GPU firmware for the SM8650-QRD board
>>
>> Signed-off-by: Neil Armstrong <[email protected]>
>> ---
>> arch/arm64/boot/dts/qcom/sm8650-qrd.dts | 8 ++++++++
>> 1 file changed, 8 insertions(+)
>>
>> diff --git a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
>> index 8515498553bf..7151f3dc67c4 100644
>> --- a/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
>> +++ b/arch/arm64/boot/dts/qcom/sm8650-qrd.dts
>> @@ -525,6 +525,14 @@ &ipa {
>> status = "okay";
>> };
>>
>> +&gpu {
>> + status = "okay";
>> +
>> + zap-shader {
>> + firmware-name = "qcom/sm8650/gen70900_zap.mbn";
>
> But why? For sm8450 / sm8550 we had "a730_zap.mbn" and "a740_zap.mbn"
No idea why, this is the vendor's fw name, it uses the chipid (2 for sm8550, 9 for sm8650) instead of the commercial name like the seq file (gen70900_sqe.fw).
Neil
>
>> + };
>> +};
>> +
>> &mdss {
>> status = "okay";
>> };
>>
>> --
>> 2.34.1
>>
>
>
On 12/02/2024 11:50, Konrad Dybcio wrote:
> On 12.02.2024 11:37, Neil Armstrong wrote:
>> Add GPU nodes for the SM8650 platform.
>>
>> Signed-off-by: Neil Armstrong <[email protected]>
>> ---
>> arch/arm64/boot/dts/qcom/sm8650.dtsi | 169 +++++++++++++++++++++++++++++++++++
>> 1 file changed, 169 insertions(+)
>>
>> diff --git a/arch/arm64/boot/dts/qcom/sm8650.dtsi b/arch/arm64/boot/dts/qcom/sm8650.dtsi
>> index c455ca4e6475..f6f9e603fb2f 100644
>> --- a/arch/arm64/boot/dts/qcom/sm8650.dtsi
>> +++ b/arch/arm64/boot/dts/qcom/sm8650.dtsi
>> @@ -2582,6 +2582,131 @@ tcsr: clock-controller@1fc0000 {
>> #reset-cells = <1>;
>> };
>>
>> + gpu: gpu@3d00000 {
>> + compatible = "qcom,adreno-43051401", "qcom,adreno";
>> + reg = <0x0 0x03d00000 0x0 0x40000>,
>> + <0x0 0x03d9e000 0x0 0x1000>,
>> + <0x0 0x03d61000 0x0 0x800>;
>> + reg-names = "kgsl_3d0_reg_memory",
>> + "cx_mem",
>> + "cx_dbgc";
>> +
>> + interrupts = <GIC_SPI 300 IRQ_TYPE_LEVEL_HIGH>;
>> +
>> + iommus = <&adreno_smmu 0 0x0>,
>> + <&adreno_smmu 1 0x0>;
>> +
>> + operating-points-v2 = <&gpu_opp_table>;
>> +
>> + interconnects = <&gem_noc MASTER_GFX3D 0 &mc_virt SLAVE_EBI1 0>;
>> + interconnect-names = "gfx-mem";
>
> ICC should be unnecessary
Ack I'll remove
Neil
>
> Konrad
On 12/02/2024 11:46, Konrad Dybcio wrote:
> On 12.02.2024 11:37, Neil Armstrong wrote:
>> Add support for the A750 GPU found on the SM8650 platform
>>
>> Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
>> doesn't have an HWCFG block but a separate register set.
>>
>> The missing registers are added in the a6xx.xml.h file that would
>> require a subsequent sync and the non-existent hwcfg is handled
>> in a6xx_set_hwcg().
>
> These should also be submitted to mesa to make sure the next header sync
> doesn't wipe them
Ack submitting them right now: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27576
>
> [...]
>
>> --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>> @@ -958,10 +958,11 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
>> struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
>> const struct adreno_reglist *reg;
>> + bool skip_programming = !(adreno_gpu->info->hwcg || adreno_is_a7xx(adreno_gpu));
>
> is_a750?
OK right, I was thinking of the next gpu which will probably also miss an hwcfg
>
>> unsigned int i;
>> u32 val, clock_cntl_on, cgc_mode;
>>
>> - if (!adreno_gpu->info->hwcg)
>> + if (skip_programming)
>> return;
>>
>> if (adreno_is_a630(adreno_gpu))
>> @@ -982,6 +983,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>> state ? 0x5555 : 0);
>> }
>>
>> + if (!adreno_gpu->info->hwcg) {
>
> I don't think this block of code is reachable now, no?
It is because we didn't skip when adreno_is_a7xx(adreno_gpu)
>
> Maybe remove the skip_programming and if_a750 here?
This would require:
>> - if (!adreno_gpu->info->hwcg || )
>> + if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
and:
>> + if (adreno_is_a750(adreno_gpu)) {
But if the next gpu also doesn't have an hwcfg, we will need to use
the current design...
I just tried with:
====================><===============================
@@ -961,7 +961,7 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
unsigned int i;
u32 val, clock_cntl_on, cgc_mode;
- if (!adreno_gpu->info->hwcg)
+ if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
return;
if (adreno_is_a630(adreno_gpu))
@@ -982,6 +982,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
state ? 0x5555 : 0);
}
+ if (adreno_is_a750(adreno_gpu)) {
+ gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
+ gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
+
+ if (state) {
+ gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
+
+ if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
+ val & A7XX_RBBM_CGC_P2S_STATUS_TXDONE, 1, 10)) {
+ dev_err(&gpu->pdev->dev, "RBBM_CGC_P2S_STATUS TXDONE Poll failed\n");
+ return;
+ }
+
+ gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 0);
+ }
+
+ return;
+ }
+
val = gpu_read(gpu, REG_A6XX_RBBM_CLOCK_CNTL);
/* Don't re-program the registers if they are already correct */
====================><===============================
And it works fine, does it work it for you ?
>
>> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
>> +
>> + if (state) {
>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
>> +
>> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
>> + val & BIT(0), 1, 10)) {
>
> We should define that bit name (the err suggests it's
> REG_A7XX_RBBM_GCC_P2S_STATUS_TXDONE or so)
>
> [...]
>
>> +static inline int adreno_is_a750(struct adreno_gpu *gpu)
>> +{
>> + return gpu->info->chip_ids[0] == 0x43051401;
>> +}
>> +
>> /* Placeholder to make future diffs smaller */
>
> Please also remove this comment now that it's invalid
Ack
>
> Konrad
Thanks,
Neil
On 12.02.2024 15:45, Neil Armstrong wrote:
> On 12/02/2024 11:46, Konrad Dybcio wrote:
>> On 12.02.2024 11:37, Neil Armstrong wrote:
>>> Add support for the A750 GPU found on the SM8650 platform
>>>
>>> Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
>>> doesn't have an HWCFG block but a separate register set.
>>>
>>> The missing registers are added in the a6xx.xml.h file that would
>>> require a subsequent sync and the non-existent hwcfg is handled
>>> in a6xx_set_hwcg().
>>
>> These should also be submitted to mesa to make sure the next header sync
>> doesn't wipe them
>
> Ack submitting them right now: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27576
Thanks
>
>>
>> [...]
>>
>>> --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>>> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>>> @@ -958,10 +958,11 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>>> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
>>> struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
>>> const struct adreno_reglist *reg;
>>> + bool skip_programming = !(adreno_gpu->info->hwcg || adreno_is_a7xx(adreno_gpu));
>>
>> is_a750?
>
> OK right, I was thinking of the next gpu which will probably also miss an hwcfg
>
>>
>>> unsigned int i;
>>> u32 val, clock_cntl_on, cgc_mode;
>>> - if (!adreno_gpu->info->hwcg)
>>> + if (skip_programming)
>>> return;
>>> if (adreno_is_a630(adreno_gpu))
>>> @@ -982,6 +983,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>>> state ? 0x5555 : 0);
>>> }
>>> + if (!adreno_gpu->info->hwcg) {
>>
>> I don't think this block of code is reachable now, no?
>
> It is because we didn't skip when adreno_is_a7xx(adreno_gpu)
Ahh I misread the brackets within the assignment
>
>>
>> Maybe remove the skip_programming and if_a750 here?
> This would require:
>>> - if (!adreno_gpu->info->hwcg || )
>>> + if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
>
> and:
>
>>> + if (adreno_is_a750(adreno_gpu)) {
>
> But if the next gpu also doesn't have an hwcfg, we will need to use
> the current design...
>
> I just tried with:
> ====================><===============================
> @@ -961,7 +961,7 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
> unsigned int i;
> u32 val, clock_cntl_on, cgc_mode;
>
> - if (!adreno_gpu->info->hwcg)
> + if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
> return;
>
> if (adreno_is_a630(adreno_gpu))
> @@ -982,6 +982,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
> state ? 0x5555 : 0);
> }
>
> + if (adreno_is_a750(adreno_gpu)) {
> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
> +
> + if (state) {
> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
> +
> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
> + val & A7XX_RBBM_CGC_P2S_STATUS_TXDONE, 1, 10)) {
> + dev_err(&gpu->pdev->dev, "RBBM_CGC_P2S_STATUS TXDONE Poll failed\n");
> + return;
> + }
> +
> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 0);
> + }
> +
> + return;
> + }
> +
> val = gpu_read(gpu, REG_A6XX_RBBM_CLOCK_CNTL);
>
> /* Don't re-program the registers if they are already correct */
> ====================><===============================
>
> And it works fine, does it work it for you ?
Let's keep it as-is in the original submission, as I've mentioned, I had
misread the code
Konrad
>
>>
>>> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
>>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
>>> +
>>> + if (state) {
>>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
>>> +
>>> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
>>> + val & BIT(0), 1, 10)) {
>>
>> We should define that bit name (the err suggests it's
>> REG_A7XX_RBBM_GCC_P2S_STATUS_TXDONE or so)
>>
>> [...]
>>
>>> +static inline int adreno_is_a750(struct adreno_gpu *gpu)
>>> +{
>>> + return gpu->info->chip_ids[0] == 0x43051401;
>>> +}
>>> +
>>> /* Placeholder to make future diffs smaller */
>>
>> Please also remove this comment now that it's invalid
>
> Ack
>
>>
>> Konrad
>
> Thanks,
> Neil
>
On 14/02/2024 22:43, Konrad Dybcio wrote:
> On 12.02.2024 15:45, Neil Armstrong wrote:
>> On 12/02/2024 11:46, Konrad Dybcio wrote:
>>> On 12.02.2024 11:37, Neil Armstrong wrote:
>>>> Add support for the A750 GPU found on the SM8650 platform
>>>>
>>>> Unlike the the very close A740 GPU on the SM8550 SoC, the A750 GPU
>>>> doesn't have an HWCFG block but a separate register set.
>>>>
>>>> The missing registers are added in the a6xx.xml.h file that would
>>>> require a subsequent sync and the non-existent hwcfg is handled
>>>> in a6xx_set_hwcg().
>>>
>>> These should also be submitted to mesa to make sure the next header sync
>>> doesn't wipe them
>>
>> Ack submitting them right now: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27576
>
> Thanks
>
>>
>>>
>>> [...]
>>>
>>>> --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>>>> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
>>>> @@ -958,10 +958,11 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>>>> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
>>>> struct a6xx_gmu *gmu = &a6xx_gpu->gmu;
>>>> const struct adreno_reglist *reg;
>>>> + bool skip_programming = !(adreno_gpu->info->hwcg || adreno_is_a7xx(adreno_gpu));
>>>
>>> is_a750?
>>
>> OK right, I was thinking of the next gpu which will probably also miss an hwcfg
>>
>>>
>>>> unsigned int i;
>>>> u32 val, clock_cntl_on, cgc_mode;
>>>> - if (!adreno_gpu->info->hwcg)
>>>> + if (skip_programming)
>>>> return;
>>>> if (adreno_is_a630(adreno_gpu))
>>>> @@ -982,6 +983,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>>>> state ? 0x5555 : 0);
>>>> }
>>>> + if (!adreno_gpu->info->hwcg) {
>>>
>>> I don't think this block of code is reachable now, no?
>>
>> It is because we didn't skip when adreno_is_a7xx(adreno_gpu)
>
> Ahh I misread the brackets within the assignment
>
>>
>>>
>>> Maybe remove the skip_programming and if_a750 here?
>> This would require:
>>>> - if (!adreno_gpu->info->hwcg || )
>>>> + if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
>>
>> and:
>>
>>>> + if (adreno_is_a750(adreno_gpu)) {
>>
>> But if the next gpu also doesn't have an hwcfg, we will need to use
>> the current design...
>>
>> I just tried with:
>> ====================><===============================
>> @@ -961,7 +961,7 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>> unsigned int i;
>> u32 val, clock_cntl_on, cgc_mode;
>>
>> - if (!adreno_gpu->info->hwcg)
>> + if (!(adreno_gpu->info->hwcg || adreno_is_a750(adreno_gpu)))
>> return;
>>
>> if (adreno_is_a630(adreno_gpu))
>> @@ -982,6 +982,25 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state)
>> state ? 0x5555 : 0);
>> }
>>
>> + if (adreno_is_a750(adreno_gpu)) {
>> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
>> +
>> + if (state) {
>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
>> +
>> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
>> + val & A7XX_RBBM_CGC_P2S_STATUS_TXDONE, 1, 10)) {
>> + dev_err(&gpu->pdev->dev, "RBBM_CGC_P2S_STATUS TXDONE Poll failed\n");
>> + return;
>> + }
>> +
>> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 0);
>> + }
>> +
>> + return;
>> + }
>> +
>> val = gpu_read(gpu, REG_A6XX_RBBM_CLOCK_CNTL);
>>
>> /* Don't re-program the registers if they are already correct */
>> ====================><===============================
>>
>> And it works fine, does it work it for you ?
>
> Let's keep it as-is in the original submission, as I've mentioned, I had
> misread the code
Ack thanks
Neil
>
> Konrad
>
>>
>>>
>>>> + gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1);
>>>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_GLOBAL_LOAD_CMD, state ? 1 : 0);
>>>> +
>>>> + if (state) {
>>>> + gpu_write(gpu, REG_A7XX_RBBM_CGC_P2S_TRIG_CMD, 1);
>>>> +
>>>> + if (gpu_poll_timeout(gpu, REG_A7XX_RBBM_CGC_P2S_STATUS, val,
>>>> + val & BIT(0), 1, 10)) {
>>>
>>> We should define that bit name (the err suggests it's
>>> REG_A7XX_RBBM_GCC_P2S_STATUS_TXDONE or so)
>>>
>>> [...]
>>>
>>>> +static inline int adreno_is_a750(struct adreno_gpu *gpu)
>>>> +{
>>>> + return gpu->info->chip_ids[0] == 0x43051401;
>>>> +}
>>>> +
>>>> /* Placeholder to make future diffs smaller */
>>>
>>> Please also remove this comment now that it's invalid
>>
>> Ack
>>
>>>
>>> Konrad
>>
>> Thanks,
>> Neil
>>