[PATCH 090/207] drm/amdgpu: enable JPEG3.0 for Sienna_Cichlid

[PATCH 090/207] drm/amdgpu: enable JPEG3.0 for Sienna_Cichlid

[Alex Deucher]

From: Leo Liu <leo.liu@amd.com>

By adding JPEG HW block to Sienna_Cichlid

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 81032b0d8512..320d4db471cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -57,6 +57,7 @@
 #include "vcn_v2_0.h"
 #include "jpeg_v2_0.h"
 #include "vcn_v3_0.h"
+#include "jpeg_v3_0.h"
 #include "dce_virtual.h"
 #include "mes_v10_1.h"
 #include "mxgpu_nv.h"
@@ -499,6 +500,7 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
 		amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block);
 		amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
+		amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block);
 		if (adev->enable_mes)
 			amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block);
 		break;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 091/207] drm/amdgpu/gfx10: add gc golden setting for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add gc golden setting for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 55 ++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 69458d5ccb09..750f2ff9ed9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -62,6 +62,10 @@
 
 #define mmCGTT_GS_NGG_CLK_CTRL	0x5087
 #define mmCGTT_GS_NGG_CLK_CTRL_BASE_IDX	1
+#define mmCGTT_SPI_RA0_CLK_CTRL 0x507a
+#define mmCGTT_SPI_RA0_CLK_CTRL_BASE_IDX 1
+#define mmCGTT_SPI_RA1_CLK_CTRL 0x507b
+#define mmCGTT_SPI_RA1_CLK_CTRL_BASE_IDX 1
 
 #define GB_ADDR_CONFIG__NUM_PKRS__SHIFT                                                                       0x8
 #define GB_ADDR_CONFIG__NUM_PKRS_MASK                                                                         0x00000700L
@@ -92,6 +96,8 @@
 #define CP_RB_DOORBELL_RANGE_LOWER__DOORBELL_RANGE_LOWER_Sienna_Cichlid__SHIFT	0x2
 #define CP_RB_DOORBELL_RANGE_UPPER__DOORBELL_RANGE_UPPER_Sienna_Cichlid_MASK	0x00000FFCL
 #define PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC_Sienna_Cichlid_MASK	0x00300000L
+#define mmGCR_GENERAL_CNTL_Sienna_Cichlid			0x1580
+#define mmGCR_GENERAL_CNTL_Sienna_Cichlid_BASE_IDX	0
 
 MODULE_FIRMWARE("amdgpu/navi10_ce.bin");
 MODULE_FIRMWARE("amdgpu/navi10_pfp.bin");
@@ -3050,6 +3056,47 @@ static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_2_nv12[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xFFFFFFFF, 0xe0000000)
 };
 
+static const struct soc15_reg_golden golden_settings_gc_10_3[] =
+{
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_PS_CLK_CTRL, 0xff7f0fff, 0x78000100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0xff7f0fff, 0x30000100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA1_CLK_CTRL, 0xff7f0fff, 0x7e000100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCPF_GCR_CNTL, 0x0007ffff, 0x0000c000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000200),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_EXCEPTION_CONTROL, 0x7fff0f1f, 0x00b80000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Sienna_Cichlid, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x10f80988),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf17fffff, 0x01200007),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0xffffffff, 0x00070104),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_CONFIG, 0xe07df47f, 0x00180070),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER0_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER1_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER10_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER11_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER12_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER13_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER14_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER15_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER2_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER3_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER4_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER5_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER6_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER7_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER8_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER9_SELECT, 0xf0f001ff, 0x00000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00000000)
+};
+
+static const struct soc15_reg_golden golden_settings_gc_10_3_sienna_cichlid[] =
+{
+	/* Pending on emulation bring up */
+};
+
 #define DEFAULT_SH_MEM_CONFIG \
 	((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
 	 (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
@@ -3230,6 +3277,14 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev)
 						golden_settings_gc_rlc_spm_10_1_2_nv12,
 						(const u32)ARRAY_SIZE(golden_settings_gc_rlc_spm_10_1_2_nv12));
 		break;
+	case CHIP_SIENNA_CICHLID:
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_10_3,
+						(const u32)ARRAY_SIZE(golden_settings_gc_10_3));
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_10_3_sienna_cichlid,
+						(const u32)ARRAY_SIZE(golden_settings_gc_10_3_sienna_cichlid));
+		break;
 	default:
 		break;
 	}
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 092/207] drm/amdgpu: force pa_sc_tile_steering_override to 0 for gfx10.3

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

pa_sc_tile_steering_override is only programmable for
gfx10.0/10.1/10.2

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 21 ++++++++-------------
 1 file changed, 8 insertions(+), 13 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 750f2ff9ed9f..66d7ae456768 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -95,7 +95,6 @@
 #define CP_RB_DOORBELL_RANGE_LOWER__DOORBELL_RANGE_LOWER_Sienna_Cichlid_MASK	0x00000FFCL
 #define CP_RB_DOORBELL_RANGE_LOWER__DOORBELL_RANGE_LOWER_Sienna_Cichlid__SHIFT	0x2
 #define CP_RB_DOORBELL_RANGE_UPPER__DOORBELL_RANGE_UPPER_Sienna_Cichlid_MASK	0x00000FFCL
-#define PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC_Sienna_Cichlid_MASK	0x00300000L
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid			0x1580
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid_BASE_IDX	0
 
@@ -4468,6 +4467,11 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade
 	uint32_t num_packer_per_sc;
 	uint32_t pa_sc_tile_steering_override;
 
+	/* for ASICs that integrates GFX v10.3
+	 * pa_sc_tile_steering_override should be set to 0 */
+	if (adev->asic_type == CHIP_SIENNA_CICHLID)
+		return 0;
+
 	/* init num_sc */
 	num_sc = adev->gfx.config.max_shader_engines * adev->gfx.config.max_sh_per_se *
 			adev->gfx.config.num_sc_per_sh;
@@ -4485,18 +4489,9 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade
 	pa_sc_tile_steering_override |=
 		(order_base_2(num_rb_per_sc) << PA_SC_TILE_STEERING_OVERRIDE__NUM_RB_PER_SC__SHIFT) &
 		PA_SC_TILE_STEERING_OVERRIDE__NUM_RB_PER_SC_MASK;
-	switch (adev->asic_type) {
-	case CHIP_SIENNA_CICHLID:
-		pa_sc_tile_steering_override |=
-			(order_base_2(num_packer_per_sc) << PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC__SHIFT) &
-			PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC_Sienna_Cichlid_MASK;
-		break;
-	default:
-		pa_sc_tile_steering_override |=
-			(order_base_2(num_packer_per_sc) << PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC__SHIFT) &
-			PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC_MASK;
-		break;
-	}
+	pa_sc_tile_steering_override |=
+		(order_base_2(num_packer_per_sc) << PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC__SHIFT) &
+		PA_SC_TILE_STEERING_OVERRIDE__NUM_PACKER_PER_SC_MASK;
 
 	return pa_sc_tile_steering_override;
 }
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 093/207] drm/amdgpu: add cp firmware backdoor loading triger

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Triger CP ucode addr and data to backdoor load CP firmware.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 37 ++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 66d7ae456768..b62c755dff21 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -98,6 +98,19 @@
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid			0x1580
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid_BASE_IDX	0
 
+#define mmCP_HYP_PFP_UCODE_ADDR			0x5814
+#define mmCP_HYP_PFP_UCODE_ADDR_BASE_IDX	1
+#define mmCP_HYP_PFP_UCODE_DATA			0x5815
+#define mmCP_HYP_PFP_UCODE_DATA_BASE_IDX	1
+#define mmCP_HYP_CE_UCODE_ADDR			0x5818
+#define mmCP_HYP_CE_UCODE_ADDR_BASE_IDX		1
+#define mmCP_HYP_CE_UCODE_DATA			0x5819
+#define mmCP_HYP_CE_UCODE_DATA_BASE_IDX		1
+#define mmCP_HYP_ME_UCODE_ADDR			0x5816
+#define mmCP_HYP_ME_UCODE_ADDR_BASE_IDX		1
+#define mmCP_HYP_ME_UCODE_DATA			0x5817
+#define mmCP_HYP_ME_UCODE_DATA_BASE_IDX		1
+
 MODULE_FIRMWARE("amdgpu/navi10_ce.bin");
 MODULE_FIRMWARE("amdgpu/navi10_pfp.bin");
 MODULE_FIRMWARE("amdgpu/navi10_me.bin");
@@ -5393,6 +5406,14 @@ static int gfx_v10_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCP_PFP_IC_BASE_HI,
 		upper_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
 
+	WREG32_SOC15(GC, 0, mmCP_HYP_PFP_UCODE_ADDR, 0);
+
+	for (i = 0; i < pfp_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, mmCP_HYP_PFP_UCODE_DATA,
+			     le32_to_cpup(fw_data + pfp_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, mmCP_HYP_PFP_UCODE_ADDR, adev->gfx.pfp_fw_version);
+
 	return 0;
 }
 
@@ -5462,6 +5483,14 @@ static int gfx_v10_0_cp_gfx_load_ce_microcode(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCP_CE_IC_BASE_HI,
 		upper_32_bits(adev->gfx.ce.ce_fw_gpu_addr));
 
+	WREG32_SOC15(GC, 0, mmCP_HYP_CE_UCODE_ADDR, 0);
+
+	for (i = 0; i < ce_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, mmCP_HYP_CE_UCODE_DATA,
+			     le32_to_cpup(fw_data + ce_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, mmCP_HYP_CE_UCODE_ADDR, adev->gfx.ce_fw_version);
+
 	return 0;
 }
 
@@ -5531,6 +5560,14 @@ static int gfx_v10_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCP_ME_IC_BASE_HI,
 		upper_32_bits(adev->gfx.me.me_fw_gpu_addr));
 
+	WREG32_SOC15(GC, 0, mmCP_HYP_ME_UCODE_ADDR, 0);
+
+	for (i = 0; i < me_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, mmCP_HYP_ME_UCODE_DATA,
+			     le32_to_cpup(fw_data + me_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, mmCP_HYP_ME_UCODE_ADDR, adev->gfx.me_fw_version);
+
 	return 0;
 }
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 094/207] drm/amdgpu: disable gfxoff for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Temporary disable gfxoff for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index b62c755dff21..940b4563312b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3531,6 +3531,9 @@ static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev)
 		if (!gfx_v10_0_navi10_gfxoff_should_enable(adev))
 			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 		break;
+	case CHIP_SIENNA_CICHLID:
+		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
+		break;
 	default:
 		break;
 	}
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 095/207] drm/amdgpu: add vram_info v2_5 in atomfirmware header

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

vram_info v2_5 was introduced to support sienna_cichlid

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/include/atomfirmware.h | 124 +++++++++++++++++++++
 1 file changed, 124 insertions(+)

diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h
index b36ea8340afa..2c3c73f1e6ee 100644
--- a/drivers/gpu/drm/amd/include/atomfirmware.h
+++ b/drivers/gpu/drm/amd/include/atomfirmware.h
@@ -2313,6 +2313,130 @@ struct atom_vram_info_header_v2_4 {
   struct   atom_vram_module_v10  vram_module[16];        // just for allocation, real number of blocks is in ucNumOfVRAMModule;
 };
 
+struct atom_vram_module_v11 {
+	// Design Specific Values
+	uint32_t  memory_size;                   // Total memory size in unit of MB for CONFIG_MEMSIZE zeros
+	uint32_t  channel_enable;                // bit vector, each bit indicate specific channel enable or not
+	uint16_t  mem_voltage;                   // mem_voltage
+	uint16_t  vram_module_size;              // Size of atom_vram_module_v9
+	uint8_t   ext_memory_id;                 // Current memory module ID
+	uint8_t   memory_type;                   // enum of atom_dgpu_vram_type
+	uint8_t   channel_num;                   // Number of mem. channels supported in this module
+	uint8_t   channel_width;                 // CHANNEL_16BIT/CHANNEL_32BIT/CHANNEL_64BIT
+	uint8_t   density;                       // _8Mx32, _16Mx32, _16Mx16, _32Mx16
+	uint8_t   tunningset_id;                 // MC phy registers set per.
+	uint16_t  reserved[4];                   // reserved
+	uint8_t   vender_rev_id;                 // [7:4] Revision, [3:0] Vendor code
+	uint8_t   refreshrate;			 // [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms)
+	uint8_t   vram_flags;			 // bit0= bankgroup enable
+	uint8_t   vram_rsd2;			 // reserved
+	uint16_t  gddr6_mr10;                    // gddr6 mode register10 value
+	uint16_t  gddr6_mr0;                     // gddr6 mode register0 value
+	uint16_t  gddr6_mr1;                     // gddr6 mode register1 value
+	uint16_t  gddr6_mr2;                     // gddr6 mode register2 value
+	uint16_t  gddr6_mr4;                     // gddr6 mode register4 value
+	uint16_t  gddr6_mr7;                     // gddr6 mode register7 value
+	uint16_t  gddr6_mr8;                     // gddr6 mode register8 value
+	char    dram_pnstring[40];               // part number end with '0'.
+};
+
+struct atom_gddr6_ac_timing_v2_5 {
+	uint32_t  u32umc_id_access;
+	uint8_t  RL;
+	uint8_t  WL;
+	uint8_t  tRAS;
+	uint8_t  tRC;
+
+	uint16_t  tREFI;
+	uint8_t  tRFC;
+	uint8_t  tRFCpb;
+
+	uint8_t  tRREFD;
+	uint8_t  tRCDRD;
+	uint8_t  tRCDWR;
+	uint8_t  tRP;
+
+	uint8_t  tRRDS;
+	uint8_t  tRRDL;
+	uint8_t  tWR;
+	uint8_t  tWTRS;
+
+	uint8_t  tWTRL;
+	uint8_t  tFAW;
+	uint8_t  tCCDS;
+	uint8_t  tCCDL;
+
+	uint8_t  tCRCRL;
+	uint8_t  tCRCWL;
+	uint8_t  tCKE;
+	uint8_t  tCKSRE;
+
+	uint8_t  tCKSRX;
+	uint8_t  tRTPS;
+	uint8_t  tRTPL;
+	uint8_t  tMRD;
+
+	uint8_t  tMOD;
+	uint8_t  tXS;
+	uint8_t  tXHP;
+	uint8_t  tXSMRS;
+
+	uint32_t  tXSH;
+
+	uint8_t  tPD;
+	uint8_t  tXP;
+	uint8_t  tCPDED;
+	uint8_t  tACTPDE;
+
+	uint8_t  tPREPDE;
+	uint8_t  tREFPDE;
+	uint8_t  tMRSPDEN;
+	uint8_t  tRDSRE;
+
+	uint8_t  tWRSRE;
+	uint8_t  tPPD;
+	uint8_t  tCCDMW;
+	uint8_t  tWTRTR;
+
+	uint8_t  tLTLTR;
+	uint8_t  tREFTR;
+	uint8_t  VNDR;
+	uint8_t  reserved[9];
+};
+
+struct atom_gddr6_bit_byte_remap {
+	uint32_t dphy_byteremap;    //mmUMC_DPHY_ByteRemap
+	uint32_t dphy_bitremap0;    //mmUMC_DPHY_BitRemap0
+	uint32_t dphy_bitremap1;    //mmUMC_DPHY_BitRemap1
+	uint32_t dphy_bitremap2;    //mmUMC_DPHY_BitRemap2
+	uint32_t aphy_bitremap0;    //mmUMC_APHY_BitRemap0
+	uint32_t aphy_bitremap1;    //mmUMC_APHY_BitRemap1
+	uint32_t phy_dram;          //mmUMC_PHY_DRAM
+};
+
+struct atom_gddr6_dram_data_remap {
+	uint32_t table_size;
+	uint8_t phyintf_ck_inverted[8];     //UMC_PHY_PHYINTF_CNTL.INV_CK
+	struct atom_gddr6_bit_byte_remap bit_byte_remap[16];
+};
+
+struct atom_vram_info_header_v2_5 {
+	struct   atom_common_table_header table_header;
+	uint16_t mem_adjust_tbloffset;                         // offset of atom_umc_init_reg_block structure for memory vendor specific UMC adjust settings
+	uint16_t gddr6_ac_timing_offset;                     // offset of atom_gddr6_ac_timing_v2_5 structure for memory clock specific UMC settings
+	uint16_t mc_adjust_pertile_tbloffset;                  // offset of atom_umc_init_reg_block structure for Per Byte Offset Preset Settings
+	uint16_t mc_phyinit_tbloffset;                         // offset of atom_umc_init_reg_block structure for MC phy init set
+	uint16_t dram_data_remap_tbloffset;                    // offset of atom_gddr6_dram_data_remap array to indicate DRAM data lane to GPU mapping
+	uint16_t reserved;                                     // offset of reserved
+	uint16_t post_ucode_init_offset;                       // offset of atom_umc_init_reg_block structure for MC phy init after MC uCode complete umc init
+	uint16_t strobe_mode_patch_tbloffset;                  // offset of atom_umc_init_reg_block structure for Strobe Mode memory clock specific UMC settings
+	uint8_t  vram_module_num;                              // indicate number of VRAM module
+	uint8_t  umcip_min_ver;
+	uint8_t  umcip_max_ver;
+	uint8_t  mc_phy_tile_num;                              // indicate the MCD tile number which use in DramDataRemapTbl and usMcAdjustPerTileTblOffset
+	struct   atom_vram_module_v11  vram_module[16];        // just for allocation, real number of blocks is in ucNumOfVRAMModule;
+};
+
 /* 
   ***************************************************************************
     Data Table voltageobject_info  structure
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 096/207] drm/amdgpu: support query vram info for sienna_cichlid

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

support query vram_module v11 and vram_info v2_5
for sienna_cichlid

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c  | 22 +++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 58f9d8c3a17a..659d845dbebe 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -120,11 +120,13 @@ union umc_info {
 union vram_info {
 	struct atom_vram_info_header_v2_3 v23;
 	struct atom_vram_info_header_v2_4 v24;
+	struct atom_vram_info_header_v2_5 v25;
 };
 
 union vram_module {
 	struct atom_vram_module_v9 v9;
 	struct atom_vram_module_v10 v10;
+	struct atom_vram_module_v11 v11;
 };
 
 static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
@@ -260,6 +262,26 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 				if (vram_vendor)
 					*vram_vendor = mem_vendor;
 				break;
+			case 5:
+				if (module_id > vram_info->v25.vram_module_num)
+					module_id = 0;
+				vram_module = (union vram_module *)vram_info->v25.vram_module;
+				while (i < module_id) {
+					vram_module = (union vram_module *)
+						((u8 *)vram_module + vram_module->v11.vram_module_size);
+					i++;
+				}
+				mem_type = vram_module->v11.memory_type;
+				if (vram_type)
+					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+				mem_channel_number = vram_module->v11.channel_num;
+				mem_channel_width = vram_module->v11.channel_width;
+				if (vram_width)
+					*vram_width = mem_channel_number * (1 << mem_channel_width);
+				mem_vendor = (vram_module->v11.vender_rev_id) & 0xF;
+				if (vram_vendor)
+					*vram_vendor = mem_vendor;
+				break;
 			default:
 				return -EINVAL;
 			}
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 097/207] drm/amdgpu: drop gfx_v10_0_tiling_mode_table_init

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

tiling mode table is not used anymore for gfx10

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 8 --------
 1 file changed, 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 940b4563312b..ce6ba3ba9223 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -4401,12 +4401,6 @@ static int gfx_v10_0_sw_fini(void *handle)
 	return 0;
 }
 
-
-static void gfx_v10_0_tiling_mode_table_init(struct amdgpu_device *adev)
-{
-	/* TODO */
-}
-
 static void gfx_v10_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
 				   u32 sh_num, u32 instance)
 {
@@ -4656,8 +4650,6 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev)
 
 	WREG32_FIELD15(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
 
-	gfx_v10_0_tiling_mode_table_init(adev);
-
 	gfx_v10_0_setup_rb(adev);
 	gfx_v10_0_get_cu_info(adev, &adev->gfx.cu_info);
 	gfx_v10_0_get_tcc_info(adev);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 098/207] drm/amdgpu: only send one sdma firmware for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

As all four sdma firmware are same, PSP only receive one SDMA fw.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 4ed542b69377..2fd40d0d7a0e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -1683,6 +1683,15 @@ static int psp_np_fw_load(struct psp_context *psp)
 		if (fw_load_skip_check(psp, ucode))
 			continue;
 
+		if (psp->autoload_supported &&
+		    adev->asic_type == CHIP_SIENNA_CICHLID &&
+		    (ucode->ucode_id == AMDGPU_UCODE_ID_SDMA1 ||
+		     ucode->ucode_id == AMDGPU_UCODE_ID_SDMA2 ||
+		     ucode->ucode_id == AMDGPU_UCODE_ID_SDMA3))
+			/* PSP only receive one SDMA fw for sienna_cichlid,
+			 * as all four sdma fw are same */
+			continue;
+
 		psp_print_fw_hdr(psp, ucode);
 
 		ret = psp_execute_np_fw_load(psp, ucode);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 099/207] drm/amdgpu: add firmware_info v3_4 structure for Sienna_Cichlid

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

firmware_info v3_4 strucure will be used by kernel driver
to query various parameters set by VBIOS for Sienna_Cichlid

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/include/atomfirmware.h | 36 ++++++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h
index 2c3c73f1e6ee..f7e98855cac7 100644
--- a/drivers/gpu/drm/amd/include/atomfirmware.h
+++ b/drivers/gpu/drm/amd/include/atomfirmware.h
@@ -562,6 +562,42 @@ struct atom_firmware_info_v3_3
   uint32_t reserved2[2];
 };
 
+struct atom_firmware_info_v3_4 {
+	struct atom_common_table_header table_header;
+	uint32_t firmware_revision;
+	uint32_t bootup_sclk_in10khz;
+	uint32_t bootup_mclk_in10khz;
+	uint32_t firmware_capability;             // enum atombios_firmware_capability
+	uint32_t main_call_parser_entry;          /* direct address of main parser call in VBIOS binary. */
+	uint32_t bios_scratch_reg_startaddr;      // 1st bios scratch register dword address
+	uint16_t bootup_vddc_mv;
+	uint16_t bootup_vddci_mv;
+	uint16_t bootup_mvddc_mv;
+	uint16_t bootup_vddgfx_mv;
+	uint8_t  mem_module_id;
+	uint8_t  coolingsolution_id;              /*0: Air cooling; 1: Liquid cooling ... */
+	uint8_t  reserved1[2];
+	uint32_t mc_baseaddr_high;
+	uint32_t mc_baseaddr_low;
+	uint8_t  board_i2c_feature_id;            // enum of atom_board_i2c_feature_id_def
+	uint8_t  board_i2c_feature_gpio_id;       // i2c id find in gpio_lut data table gpio_id
+	uint8_t  board_i2c_feature_slave_addr;
+	uint8_t  reserved3;
+	uint16_t bootup_mvddq_mv;
+	uint16_t bootup_mvpp_mv;
+	uint32_t zfbstartaddrin16mb;
+	uint32_t pplib_pptable_id;                // if pplib_pptable_id!=0, pplib get powerplay table inside driver instead of from VBIOS
+	uint32_t mvdd_ratio;                      // mvdd_raio = (real mvdd in power rail)*1000/(mvdd_output_from_svi2)
+	uint16_t hw_bootup_vddgfx_mv;             // hw default vddgfx voltage level decide by board strap
+	uint16_t hw_bootup_vddc_mv;               // hw default vddc voltage level decide by board strap
+	uint16_t hw_bootup_mvddc_mv;              // hw default mvddc voltage level decide by board strap
+	uint16_t hw_bootup_vddci_mv;              // hw default vddci voltage level decide by board strap
+	uint32_t maco_pwrlimit_mw;                // bomaco mode power limit in unit of m-watt
+	uint32_t usb_pwrlimit_mw;                 // power limit when USB is enable in unit of m-watt
+	uint32_t fw_reserved_size_in_kb;          // VBIOS reserved extra fw size in unit of kb.
+	uint32_t reserved[5];
+};
+
 /* 
   ***************************************************************************
     Data Table lcd_info  structure
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 100/207] drm/amdgpu: add atomfirmware helper funciton to query reserved fb size

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

fw_reserved_size_in_kb is introduced for driver to query
the TMR region reserved by PSP BL in Sienna_Cichlid and onwards

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c  | 38 +++++++++++++++++++
 .../gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h  |  1 +
 2 files changed, 39 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 659d845dbebe..a9adccfda4c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -325,6 +325,9 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev)
 
 union firmware_info {
 	struct atom_firmware_info_v3_1 v31;
+	struct atom_firmware_info_v3_2 v32;
+	struct atom_firmware_info_v3_3 v33;
+	struct atom_firmware_info_v3_4 v34;
 };
 
 /*
@@ -590,3 +593,38 @@ int amdgpu_atomfirmware_get_mem_train_info(struct amdgpu_device *adev)
 	adev->fw_vram_usage.mem_train_support = true;
 	return 0;
 }
+
+int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
+{
+	struct atom_context *ctx = adev->mode_info.atom_context;
+	union firmware_info *firmware_info;
+	int index;
+	u16 data_offset, size;
+	u8 frev, crev;
+	int fw_reserved_fb_size;
+
+	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+			firmwareinfo);
+
+	if (!amdgpu_atom_parse_data_header(ctx, index, &size,
+				&frev, &crev, &data_offset))
+		/* fail to parse data_header */
+		return 0;
+
+	firmware_info = (union firmware_info *)(ctx->bios + data_offset);
+
+	if (frev !=3)
+		return -EINVAL;
+
+	switch (crev) {
+	case 4:
+		fw_reserved_fb_size =
+			(firmware_info->v34.fw_reserved_size_in_kb << 10);
+		break;
+	default:
+		fw_reserved_fb_size = 0;
+		break;
+	}
+
+	return fw_reserved_fb_size;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
index 434fe2fa0089..3a5ed339903e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
@@ -36,5 +36,6 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
+int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev);
 
 #endif
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 101/207] drm/amdgpu: switch to query reserved fb size from vbios (v3)

[Alex Deucher]

From: Hawking Zhang <Hawking.Zhang@amd.com>

For Sienna_Cichlid, query fw_reserved_fb_size from vbios directly.
For navi1x, fall back to default 64K TMR size.
For pre-navi, no need to reserve tmr region in top LFB.

v2: fix TMR define (Alex)
v3: partially revert size change

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 21 ++++++++++++++++-----
 1 file changed, 16 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 93d057f92dcd..b5e050b82172 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -58,6 +58,7 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_sdma.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_atomfirmware.h"
 #include "bif/bif_4_1_d.h"
 
 #define AMDGPU_TTM_VRAM_MAX_DW_READ	(size_t)128
@@ -1961,13 +1962,23 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	 * reserve TMR memory at the top of VRAM which holds
 	 * IP Discovery data and is protected by PSP.
 	 */
+	adev->discovery_tmr_size =
+		amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
+	if (!adev->discovery_tmr_size &&
+	    adev->asic_type >= CHIP_NAVI10 &&
+	    amdgpu_discovery) {
+		/* if fw_reserved_fb_size is 0 from vbios,
+		 * then fallback to the default tmr_size */
+		adev->discovery_tmr_size = DISCOVERY_TMR_SIZE;
+	}
+
 	if (adev->discovery_tmr_size > 0) {
 		r = amdgpu_bo_create_kernel_at(adev,
-			adev->gmc.real_vram_size - adev->discovery_tmr_size,
-			adev->discovery_tmr_size,
-			AMDGPU_GEM_DOMAIN_VRAM,
-			&adev->discovery_memory,
-			NULL);
+					adev->gmc.real_vram_size - adev->discovery_tmr_size,
+					adev->discovery_tmr_size,
+					AMDGPU_GEM_DOMAIN_VRAM,
+					&adev->discovery_memory,
+					NULL);
 		if (r)
 			return r;
 	}
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 102/207] drm/amdgpu: open GFX clock gating for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Open GFX MGCG, CGCG and 3DCG for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 320d4db471cb..5d1c2eba3412 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -714,7 +714,10 @@ static int nv_common_early_init(void *handle)
 		adev->external_rev_id = adev->rev_id + 0xa;
 		break;
 	case CHIP_SIENNA_CICHLID:
-		adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG |
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_VCN_MGCG |
 			AMD_CG_SUPPORT_JPEG_MGCG;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_JPEG;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 103/207] drm/amdgpu: change the offset for VCN FW cache window

[Alex Deucher]

From: Leo Liu <leo.liu@amd.com>

The signed header is added

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 011edbdd4e55..371c70a1e611 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -356,11 +356,8 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
 		WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
 			upper_32_bits(adev->vcn.inst[inst].gpu_addr));
 		offset = size;
-		/* No signed header for now from firmware
 		WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_OFFSET0,
 			AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
-		*/
-		WREG32_SOC15(UVD, inst, mmUVD_VCPU_CACHE_OFFSET0, 0);
 	}
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_SIZE0, size);
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 104/207] drm/amdgpu: fix the PSP front door loading VCN firmware

[Alex Deucher]

From: Leo Liu <leo.liu@amd.com>

for the second instance with correct index

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 371c70a1e611..dddae2b8f0f9 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -345,9 +345,9 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
 	/* cache window 0: fw */
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
-			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo));
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
 		WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
-			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi));
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
 		WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_OFFSET0, 0);
 		offset = 0;
 	} else {
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 105/207] drm/amdgpu: set the LMI ctrl and reset earlier

[Alex Deucher]

From: Leo Liu <leo.liu@amd.com>

So the LMI register will be programmed properly

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index dddae2b8f0f9..517484292303 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -712,6 +712,15 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
 		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0,
 			~UVD_MASTINT_EN__VCPU_EN_MASK);
 
+		/* enable LMI MC and UMC channels */
+		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
+			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+		tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
+		tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+		tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+		WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
+
 		/* setup mmUVD_LMI_CTRL */
 		tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL);
 		WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp |
@@ -752,15 +761,6 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
 		WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG,
 			adev->gfx.config.gb_addr_config);
 
-		/* enable LMI MC and UMC channels */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
-			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
-
-		tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
-		tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
-		tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
-
 		/* unblock VCPU register access */
 		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0,
 			~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 106/207] drm/amdgpu: update golden setting for gfx10.3

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Update gfx golden setting for gfx10.3.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index ce6ba3ba9223..752032eba6ec 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3082,6 +3082,8 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x10f80988),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf17fffff, 0x01200007),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_TIMEOUT_COUNTER, 0xffffffff, 0x00000800),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE_2, 0xffffffbf, 0x00000820),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0xffffffff, 0x00070104),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_CONFIG, 0xe07df47f, 0x00180070),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER0_SELECT, 0xf0f001ff, 0x00000000),
@@ -3101,7 +3103,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER8_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER9_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00000000)
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00800000)
 };
 
 static const struct soc15_reg_golden golden_settings_gc_10_3_sienna_cichlid[] =
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 107/207] drm/amd/amdgpu: fix the HDP LS/DS/SD programming

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN
have to be set for SRAM LS/DS/SD

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 5d1c2eba3412..081fde347dd4 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -889,6 +889,16 @@ static void nv_update_hdp_mem_power_gating(struct amdgpu_device *adev,
 						 RC_MEM_POWER_DS_EN, enable);
 	}
 
+	/* confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN have to
+	 * be set for SRAM LS/DS/SD */
+	if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
+							AMD_CG_SUPPORT_HDP_SD)) {
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+						IPH_MEM_POWER_CTRL_EN, 1);
+		hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+						RC_MEM_POWER_CTRL_EN, 1);
+	}
+
 	WREG32_SOC15(HDP, 0, mmHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
 
 	/* restore IPH & RC clock override after clock/power mode changing */
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 108/207] drm/amd/amdgpu: add HDP mgcg and ls support

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

add HDP mgcg and ls support and verified

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 081fde347dd4..f497fbcc6f5c 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -718,7 +718,9 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_GFX_CGCG |
 			AMD_CG_SUPPORT_GFX_3D_CGCG |
 			AMD_CG_SUPPORT_VCN_MGCG |
-			AMD_CG_SUPPORT_JPEG_MGCG;
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_HDP_MGCG |
+			AMD_CG_SUPPORT_HDP_LS;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_JPEG;
 		adev->external_rev_id = adev->rev_id + 0x28;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 109/207] drm/amd/amdgpu: add IH cg support

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

IH cg verified

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index f497fbcc6f5c..9559eaf25591 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -720,7 +720,8 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_VCN_MGCG |
 			AMD_CG_SUPPORT_JPEG_MGCG |
 			AMD_CG_SUPPORT_HDP_MGCG |
-			AMD_CG_SUPPORT_HDP_LS;
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_IH_CG;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_JPEG;
 		adev->external_rev_id = adev->rev_id + 0x28;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 110/207] drm/amd/amdgpu: add athub ls support

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

athub ls is bounded with hdp ls,verified.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 9559eaf25591..451557544b85 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -721,7 +721,8 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_JPEG_MGCG |
 			AMD_CG_SUPPORT_HDP_MGCG |
 			AMD_CG_SUPPORT_HDP_LS |
-			AMD_CG_SUPPORT_IH_CG;
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_MC_LS;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_JPEG;
 		adev->external_rev_id = adev->rev_id + 0x28;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 111/207] drm/amdgpu: Enable Multi Media Hub (MMHUB) Clock Gating for sienna_cichlid.

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable mmhub clockgating.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 451557544b85..89c3727b7a8e 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -717,6 +717,7 @@ static int nv_common_early_init(void *handle)
 		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
 			AMD_CG_SUPPORT_GFX_CGCG |
 			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_MC_MGCG |
 			AMD_CG_SUPPORT_VCN_MGCG |
 			AMD_CG_SUPPORT_JPEG_MGCG |
 			AMD_CG_SUPPORT_HDP_MGCG |
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 112/207] drm/amdgpu: fix SDMA hdp flush engine conflict

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Each of HDP flush engine should be used by one ring, correct allocate of
hdp flush engine to SDMA ring.
Correct me value of each SDMA ring, as it was cleared when init microcode.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index 52206050adb9..f072cef28b60 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -391,10 +391,7 @@ static void sdma_v5_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	u32 ref_and_mask = 0;
 	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
 
-	if (ring->me == 0)
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0;
-	else
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma1;
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
 			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
@@ -1224,6 +1221,7 @@ static int sdma_v5_2_sw_init(void *handle)
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
 		ring->use_doorbell = true;
+		ring->me = i;
 
 		DRM_INFO("use_doorbell being set to: [%s]\n",
 				ring->use_doorbell?"true":"false");
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 113/207] drm/amdgpu: enable 3D pipe 1 on Sienna_Cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Only disable 3D pipe 1 on navi1x, enable 3D pipe 1 on Sienna_Cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 752032eba6ec..6c52363d5662 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -55,6 +55,7 @@
  * 2. Async ring
  */
 #define GFX10_NUM_GFX_RINGS_NV1X	1
+#define GFX10_NUM_GFX_RINGS_Sienna_Cichlid	2
 #define GFX10_MEC_HPD_SIZE	2048
 
 #define F32_CE_PROGRAM_RAM_SIZE		65536
@@ -7057,7 +7058,18 @@ static int gfx_v10_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_NV1X;
+	switch (adev->asic_type) {
+	case CHIP_NAVI10:
+	case CHIP_NAVI14:
+	case CHIP_NAVI12:
+		adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_NV1X;
+		break;
+	case CHIP_SIENNA_CICHLID:
+		adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_Sienna_Cichlid;
+		break;
+	default:
+		break;
+	}
 
 	adev->gfx.num_compute_rings = AMDGPU_MAX_COMPUTE_RINGS;
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 114/207] sound/pci/hda: add sienna_cichlid audio asic id for sienna_cichlid up

[Alex Deucher]

From: Hersen Wu <hersenxs.wu@amd.com>

dp/hdmi ati hda is not shown in audio settings

Signed-off-by: Hersen Wu <hersenxs.wu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 sound/pci/hda/hda_intel.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
index 92a042e34d3e..4188bc4bd4e2 100644
--- a/sound/pci/hda/hda_intel.c
+++ b/sound/pci/hda/hda_intel.c
@@ -2643,6 +2643,9 @@ static const struct pci_device_id azx_ids[] = {
 	{ PCI_DEVICE(0x1002, 0xab38),
 	  .driver_data = AZX_DRIVER_ATIHDMI_NS | AZX_DCAPS_PRESET_ATI_HDMI_NS |
 	  AZX_DCAPS_PM_RUNTIME },
+	{ PCI_DEVICE(0x1002, 0xab28),
+	  .driver_data = AZX_DRIVER_ATIHDMI_NS | AZX_DCAPS_PRESET_ATI_HDMI_NS |
+	  AZX_DCAPS_PM_RUNTIME },
 	/* VIA VT8251/VT8237A */
 	{ PCI_DEVICE(0x1106, 0x3288), .driver_data = AZX_DRIVER_VIA },
 	/* VIA GFX VT7122/VX900 */
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 115/207] drm/amdgpu/dc: Add missing Sienna_Cichlid chip id

[Alex Deucher]

From: "Jerry (Fangzhi) Zuo" <Jerry.Zuo@amd.com>

Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 3b6588e15c6b..15de344438d2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2782,6 +2782,9 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
 	case CHIP_NAVI14:
 	case CHIP_NAVI12:
 	case CHIP_RENOIR:
+#endif
+#if defined(CONFIG_DRM_AMD_DC_DCN3_0)
+	case CHIP_SIENNA_CICHLID:
 #endif
 		return amdgpu_dc != 0;
 #endif
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 116/207] drm/amdkfd: Support Sienna_Cichlid KFD v4

[Alex Deucher]

From: Yong Zhao <Yong.Zhao@amd.com>

v4: drop get_tile_config, comment out other callbacks

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/Makefile           |  13 +-
 .../drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c  | 834 ++++++++++++++++++
 drivers/gpu/drm/amd/amdkfd/kfd_crat.c         |   1 +
 drivers/gpu/drm/amd/amdkfd/kfd_device.c       |  21 +
 .../drm/amd/amdkfd/kfd_device_queue_manager.c |   1 +
 drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c  |   1 +
 .../gpu/drm/amd/amdkfd/kfd_packet_manager.c   |   1 +
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c     |   1 +
 8 files changed, 867 insertions(+), 6 deletions(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c

diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 179070b12cde..bfcfb034aed5 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -176,12 +176,13 @@ AMDKFD_PATH := ../amdkfd
 include $(FULL_AMD_PATH)/amdkfd/Makefile
 amdgpu-y += $(AMDKFD_FILES)
 amdgpu-y += \
-	 amdgpu_amdkfd_fence.o \
-	 amdgpu_amdkfd_gpuvm.o \
-	 amdgpu_amdkfd_gfx_v8.o \
-	 amdgpu_amdkfd_gfx_v9.o \
-	 amdgpu_amdkfd_arcturus.o \
-	 amdgpu_amdkfd_gfx_v10.o
+	amdgpu_amdkfd_fence.o \
+	amdgpu_amdkfd_gpuvm.o \
+	amdgpu_amdkfd_gfx_v8.o \
+	amdgpu_amdkfd_gfx_v9.o \
+	amdgpu_amdkfd_arcturus.o \
+	amdgpu_amdkfd_gfx_v10.o \
+	amdgpu_amdkfd_gfx_v10_3.o
 
 ifneq ($(CONFIG_DRM_AMDGPU_CIK),)
 amdgpu-y += amdgpu_amdkfd_gfx_v7.o
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
new file mode 100644
index 000000000000..7e59e473a190
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
@@ -0,0 +1,834 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include <linux/mmu_context.h>
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "gc/gc_10_3_0_offset.h"
+#include "gc/gc_10_3_0_sh_mask.h"
+#include "navi10_enum.h"
+#include "oss/osssys_5_0_0_offset.h"
+#include "oss/osssys_5_0_0_sh_mask.h"
+#include "soc15_common.h"
+#include "v10_structs.h"
+#include "nv.h"
+#include "nvd.h"
+#include "gfxhub_v2_1.h"
+
+enum hqd_dequeue_request_type {
+	NO_ACTION = 0,
+	DRAIN_PIPE,
+	RESET_WAVES,
+	SAVE_WAVES
+};
+
+static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
+{
+	return (struct amdgpu_device *)kgd;
+}
+
+static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	mutex_lock(&adev->srbm_mutex);
+	nv_grbm_select(adev, mec, pipe, queue, vmid);
+}
+
+static void unlock_srbm(struct kgd_dev *kgd)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(kgd, mec, pipe, queue_id, 0);
+}
+
+static uint64_t get_queue_mask(struct amdgpu_device *adev,
+			       uint32_t pipe_id, uint32_t queue_id)
+{
+	unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe +
+			queue_id;
+
+	return 1ull << bit;
+}
+
+static void release_queue(struct kgd_dev *kgd)
+{
+	unlock_srbm(kgd);
+}
+
+static void program_sh_mem_settings_v10_3(struct kgd_dev *kgd, uint32_t vmid,
+					uint32_t sh_mem_config,
+					uint32_t sh_mem_ape1_base,
+					uint32_t sh_mem_ape1_limit,
+					uint32_t sh_mem_bases)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	lock_srbm(kgd, 0, 0, 0, vmid);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
+	/* APE1 no longer exists on GFX9 */
+
+	unlock_srbm(kgd);
+}
+
+/* ATC is defeatured on Sienna_Cichlid */
+static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid,
+					unsigned int vmid)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT;
+
+	/* Mapping vmid to pasid also for IH block */
+	pr_debug("mapping vmid %d -> pasid %d in IH block for GFX client\n",
+			vmid, pasid);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid, value);
+
+	return 0;
+}
+
+static int init_interrupts_v10_3(struct kgd_dev *kgd, uint32_t pipe_id)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t mec;
+	uint32_t pipe;
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(kgd, mec, pipe, 0, 0);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmCPC_INT_CNTL),
+		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
+		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
+
+	unlock_srbm(kgd);
+
+	return 0;
+}
+
+static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
+				unsigned int engine_id,
+				unsigned int queue_id)
+{
+	uint32_t sdma_engine_reg_base = 0;
+	uint32_t sdma_rlc_reg_offset;
+
+	switch (engine_id) {
+	default:
+		dev_warn(adev->dev,
+			 "Invalid sdma engine id (%d), using engine id 0\n",
+			 engine_id);
+		/* fall through */
+	case 0:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA0, 0,
+				mmSDMA0_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL;
+		break;
+	case 1:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA1, 0,
+				mmSDMA1_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL;
+		break;
+	case 2:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA2, 0,
+				mmSDMA2_RLC0_RB_CNTL) - mmSDMA2_RLC0_RB_CNTL;
+		break;
+	case 3:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA3, 0,
+				mmSDMA3_RLC0_RB_CNTL) - mmSDMA2_RLC0_RB_CNTL;
+		break;
+	}
+
+	sdma_rlc_reg_offset = sdma_engine_reg_base
+		+ queue_id * (mmSDMA0_RLC1_RB_CNTL - mmSDMA0_RLC0_RB_CNTL);
+
+	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
+			queue_id, sdma_rlc_reg_offset);
+
+	return sdma_rlc_reg_offset;
+}
+
+static inline struct v10_compute_mqd *get_mqd(void *mqd)
+{
+	return (struct v10_compute_mqd *)mqd;
+}
+
+static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v10_sdma_mqd *)mqd;
+}
+
+static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr,
+			uint32_t wptr_shift, uint32_t wptr_mask,
+			struct mm_struct *mm)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	struct v10_compute_mqd *m;
+	uint32_t *mqd_hqd;
+	uint32_t reg, hqd_base, data;
+
+	m = get_mqd(mqd);
+
+	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
+	acquire_queue(kgd, pipe_id, queue_id);
+
+	/* HIQ is set during driver init period with vmid set to 0*/
+	if (m->cp_hqd_vmid == 0) {
+		uint32_t value, mec, pipe;
+
+		mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+		pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+		pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
+			mec, pipe, queue_id);
+		value = RREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CP_SCHEDULERS));
+		value = REG_SET_FIELD(value, RLC_CP_SCHEDULERS, scheduler1,
+			((mec << 5) | (pipe << 3) | queue_id | 0x80));
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CP_SCHEDULERS), value);
+	}
+
+	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
+	mqd_hqd = &m->cp_mqd_base_addr_lo;
+	hqd_base = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
+
+	for (reg = hqd_base;
+	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
+		WREG32(reg, mqd_hqd[reg - hqd_base]);
+
+
+	/* Activate doorbell logic before triggering WPTR poll. */
+	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
+			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data);
+
+	if (wptr) {
+		/* Don't read wptr with get_user because the user
+		 * context may not be accessible (if this function
+		 * runs in a work queue). Instead trigger a one-shot
+		 * polling read from memory in the CP. This assumes
+		 * that wptr is GPU-accessible in the queue's VMID via
+		 * ATC or SVM. WPTR==RPTR before starting the poll so
+		 * the CP starts fetching new commands from the right
+		 * place.
+		 *
+		 * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit
+		 * tricky. Assume that the queue didn't overflow. The
+		 * number of valid bits in the 32-bit RPTR depends on
+		 * the queue size. The remaining bits are taken from
+		 * the saved 64-bit WPTR. If the WPTR wrapped, add the
+		 * queue size.
+		 */
+		uint32_t queue_size =
+			2 << REG_GET_FIELD(m->cp_hqd_pq_control,
+					   CP_HQD_PQ_CONTROL, QUEUE_SIZE);
+		uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1);
+
+		if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr)
+			guessed_wptr += queue_size;
+		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
+		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
+
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+		       lower_32_bits(guessed_wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+		       upper_32_bits(guessed_wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+		       lower_32_bits((uint64_t)wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+		       upper_32_bits((uint64_t)wptr));
+		pr_debug("%s setting CP_PQ_WPTR_POLL_CNTL1 to %x\n", __func__,
+			 (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1),
+		       (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+	}
+
+	/* Start the EOP fetcher */
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
+	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
+			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
+
+	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
+
+	release_queue(kgd);
+
+	return 0;
+}
+
+static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
+			    uint32_t pipe_id, uint32_t queue_id,
+			    uint32_t doorbell_off)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct v10_compute_mqd *m;
+	uint32_t mec, pipe;
+	int r;
+
+	m = get_mqd(mqd);
+
+	acquire_queue(kgd, pipe_id, queue_id);
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
+		 mec, pipe, queue_id);
+
+	spin_lock(&adev->gfx.kiq.ring_lock);
+	r = amdgpu_ring_alloc(kiq_ring, 7);
+	if (r) {
+		pr_err("Failed to alloc KIQ (%d).\n", r);
+		goto out_unlock;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(m->cp_hqd_vmid) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(queue_id) |
+			  PACKET3_MAP_QUEUES_PIPE(pipe) |
+			  PACKET3_MAP_QUEUES_ME((mec - 1)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(1) | /* engine_sel: hiq */
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_MAP_QUEUES_DOORBELL_OFFSET(doorbell_off));
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_hi);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_hi);
+	amdgpu_ring_commit(kiq_ring);
+
+out_unlock:
+	spin_unlock(&adev->gfx.kiq.ring_lock);
+	release_queue(kgd);
+
+	return r;
+}
+
+static int hqd_dump_v10_3(struct kgd_dev *kgd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t i = 0, reg;
+#define HQD_N_REGS 56
+#define DUMP_REG(addr) do {				\
+		if (WARN_ON_ONCE(i >= HQD_N_REGS))	\
+			break;				\
+		(*dump)[i][0] = (addr) << 2;		\
+		(*dump)[i++][1] = RREG32(addr);		\
+	} while (0)
+
+	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	acquire_queue(kgd, pipe_id, queue_id);
+
+	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
+	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
+		DUMP_REG(reg);
+
+	release_queue(kgd);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd,
+			     uint32_t __user *wptr, struct mm_struct *mm)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	unsigned long end_jiffies;
+	uint32_t data;
+	uint64_t data64;
+	uint64_t __user *wptr64 = (uint64_t __user *)wptr;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
+		m->sdmax_rlcx_rb_cntl & (~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK));
+
+	end_jiffies = msecs_to_jiffies(2000) + jiffies;
+	while (true) {
+		data = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
+		if (data & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL_OFFSET,
+	       m->sdmax_rlcx_doorbell_offset);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA0_RLC0_DOORBELL,
+			     ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, data);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR,
+				m->sdmax_rlcx_rb_rptr);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI,
+				m->sdmax_rlcx_rb_rptr_hi);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 1);
+	if (read_user_wptr(mm, wptr64, data64)) {
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
+		       lower_32_bits(data64));
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
+		       upper_32_bits(data64));
+	} else {
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
+		       m->sdmax_rlcx_rb_rptr);
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
+		       m->sdmax_rlcx_rb_rptr_hi);
+	}
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 0);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE, m->sdmax_rlcx_rb_base);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE_HI,
+			m->sdmax_rlcx_rb_base_hi);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_LO,
+			m->sdmax_rlcx_rb_rptr_addr_lo);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_HI,
+			m->sdmax_rlcx_rb_rptr_addr_hi);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA0_RLC0_RB_CNTL,
+			     RB_ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, data);
+
+	return 0;
+}
+
+static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd,
+			     uint32_t engine_id, uint32_t queue_id,
+			     uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
+			engine_id, queue_id);
+	uint32_t i = 0, reg;
+#undef HQD_N_REGS
+#define HQD_N_REGS (19+6+7+10)
+
+	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	for (reg = mmSDMA0_RLC0_RB_CNTL; reg <= mmSDMA0_RLC0_DOORBELL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_STATUS; reg <= mmSDMA0_RLC0_CSA_ADDR_HI; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_IB_SUB_REMAIN;
+	     reg <= mmSDMA0_RLC0_MINOR_PTR_UPDATE; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_MIDCMD_DATA0;
+	     reg <= mmSDMA0_RLC0_MIDCMD_CNTL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address,
+				uint32_t pipe_id, uint32_t queue_id)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t act;
+	bool retval = false;
+	uint32_t low, high;
+
+	acquire_queue(kgd, pipe_id, queue_id);
+	act = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
+	if (act) {
+		low = lower_32_bits(queue_address >> 8);
+		high = upper_32_bits(queue_address >> 8);
+
+		if (low == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE)) &&
+		   high == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE_HI)))
+			retval = true;
+	}
+	release_queue(kgd);
+	return retval;
+}
+
+static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t sdma_rlc_rb_cntl;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);
+
+	if (sdma_rlc_rb_cntl & SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK)
+		return true;
+
+	return false;
+}
+
+static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+				enum kfd_preempt_type reset_type,
+				unsigned int utimeout, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	enum hqd_dequeue_request_type type;
+	unsigned long end_jiffies;
+	uint32_t temp;
+	struct v10_compute_mqd *m = get_mqd(mqd);
+
+	acquire_queue(kgd, pipe_id, queue_id);
+
+	if (m->cp_hqd_vmid == 0)
+		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+
+	switch (reset_type) {
+	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
+		type = DRAIN_PIPE;
+		break;
+	case KFD_PREEMPT_TYPE_WAVEFRONT_RESET:
+		type = RESET_WAVES;
+		break;
+	default:
+		type = DRAIN_PIPE;
+		break;
+	}
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
+
+	end_jiffies = (utimeout * HZ / 1000) + jiffies;
+	while (true) {
+		temp = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
+		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("cp queue pipe %d queue %d preemption failed\n",
+					pipe_id, queue_id);
+			release_queue(kgd);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	release_queue(kgd);
+	return 0;
+}
+
+static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+				unsigned int utimeout)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t temp;
+	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);
+	temp = temp & ~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK;
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, temp);
+
+	while (true) {
+		temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
+		if (temp & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, 0);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
+		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL) |
+		SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK);
+
+	m->sdmax_rlcx_rb_rptr = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR);
+	m->sdmax_rlcx_rb_rptr_hi =
+		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI);
+
+	return 0;
+}
+
+
+static int address_watch_disable_v10_3(struct kgd_dev *kgd)
+{
+	return 0;
+}
+
+static int address_watch_execute_v10_3(struct kgd_dev *kgd,
+					unsigned int watch_point_id,
+					uint32_t cntl_val,
+					uint32_t addr_hi,
+					uint32_t addr_lo)
+{
+	return 0;
+}
+
+static int wave_control_execute_v10_3(struct kgd_dev *kgd,
+					uint32_t gfx_index_val,
+					uint32_t sq_cmd)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX), gfx_index_val);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd);
+
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		INSTANCE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SA_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SE_BROADCAST_WRITES, 1);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX), data);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+static uint32_t address_watch_get_offset_v10_3(struct kgd_dev *kgd,
+					unsigned int watch_point_id,
+					unsigned int reg_offset)
+{
+	return 0;
+}
+
+static void set_vm_context_page_table_base_v10_3(struct kgd_dev *kgd, uint32_t vmid,
+		uint64_t page_table_base)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	/* SDMA is on gfxhub as well for Navi1* series */
+	gfxhub_v2_1_setup_vm_pt_regs(adev, vmid, page_table_base);
+}
+
+#if 0
+uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd,
+				uint32_t trap_debug_wave_launch_mode,
+				uint32_t vmid)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t data = 0;
+	uint32_t orig_wave_cntl_value;
+	uint32_t orig_stall_vmid;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	orig_wave_cntl_value = RREG32(SOC15_REG_OFFSET(GC,
+				0,
+				mmSPI_GDBG_WAVE_CNTL));
+	orig_stall_vmid = REG_GET_FIELD(orig_wave_cntl_value,
+			SPI_GDBG_WAVE_CNTL,
+			STALL_VMID);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	data = 0;
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), orig_stall_vmid);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd,
+						uint32_t trap_override,
+						uint32_t trap_mask)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	data = 0;
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK,
+			EXCP_EN, trap_mask);
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK,
+			REPLACE, trap_override);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t data = 0;
+	bool is_stall_mode;
+	bool is_mode_set;
+
+	is_stall_mode = (wave_launch_mode == 4);
+	is_mode_set = (wave_launch_mode != 0 && wave_launch_mode != 4);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+			VMID_MASK, is_mode_set ? 1 << vmid : 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+			MODE, is_mode_set ? wave_launch_mode : 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data);
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL,
+			STALL_VMID, is_stall_mode ? 1 << vmid : 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL,
+			STALL_RA, is_stall_mode ? 1 : 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+/* kgd_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values
+ * The values read are:
+ *	ib_offload_wait_time     -- Wait Count for Indirect Buffer Offloads.
+ *	atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads.
+ *	wrm_offload_wait_time    -- Wait Count for WAIT_REG_MEM Offloads.
+ *	gws_wait_time            -- Wait Count for Global Wave Syncs.
+ *	que_sleep_wait_time      -- Wait Count for Dequeue Retry.
+ *	sch_wave_wait_time       -- Wait Count for Scheduling Wave Message.
+ *	sem_rearm_wait_time      -- Wait Count for Semaphore re-arm.
+ *	deq_retry_wait_time      -- Wait Count for Global Wave Syncs.
+ */
+void get_iq_wait_times_v10_3(struct kgd_dev *kgd,
+					uint32_t *wait_times)
+
+{
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	*wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2));
+}
+
+void build_grace_period_packet_info_v10_3(struct kgd_dev *kgd,
+						uint32_t wait_times,
+						uint32_t grace_period,
+						uint32_t *reg_offset,
+						uint32_t *reg_data)
+{
+	*reg_data = wait_times;
+
+	*reg_data = REG_SET_FIELD(*reg_data,
+			CP_IQ_WAIT_TIME2,
+			SCH_WAVE,
+			grace_period);
+
+	*reg_offset = mmCP_IQ_WAIT_TIME2;
+}
+#endif
+
+const struct kfd2kgd_calls gfx_v10_3_kfd2kgd = {
+	.program_sh_mem_settings = program_sh_mem_settings_v10_3,
+	.set_pasid_vmid_mapping = set_pasid_vmid_mapping_v10_3,
+	.init_interrupts = init_interrupts_v10_3,
+	.hqd_load = hqd_load_v10_3,
+	.hiq_mqd_load = hiq_mqd_load_v10_3,
+	.hqd_sdma_load = hqd_sdma_load_v10_3,
+	.hqd_dump = hqd_dump_v10_3,
+	.hqd_sdma_dump = hqd_sdma_dump_v10_3,
+	.hqd_is_occupied = hqd_is_occupied_v10_3,
+	.hqd_sdma_is_occupied = hqd_sdma_is_occupied_v10_3,
+	.hqd_destroy = hqd_destroy_v10_3,
+	.hqd_sdma_destroy = hqd_sdma_destroy_v10_3,
+	.address_watch_disable = address_watch_disable_v10_3,
+	.address_watch_execute = address_watch_execute_v10_3,
+	.wave_control_execute = wave_control_execute_v10_3,
+	.address_watch_get_offset = address_watch_get_offset_v10_3,
+	.get_atc_vmid_pasid_mapping_info = NULL,
+	.set_vm_context_page_table_base = set_vm_context_page_table_base_v10_3,
+	.get_hive_id = amdgpu_amdkfd_get_hive_id,
+#if 0
+	.enable_debug_trap = enable_debug_trap_v10_3,
+	.disable_debug_trap = disable_debug_trap_v10_3,
+	.set_wave_launch_trap_override = set_wave_launch_trap_override_v10_3,
+	.set_wave_launch_mode = set_wave_launch_mode_v10_3,
+	.get_iq_wait_times = get_iq_wait_times_v10_3,
+	.build_grace_period_packet_info = build_grace_period_packet_info_v10_3,
+#endif
+};
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index 1009a3b8dcc2..9deadfd8f929 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -678,6 +678,7 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
 	case CHIP_NAVI10:
 	case CHIP_NAVI12:
 	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
 		pcache_info = navi10_cache_info;
 		num_of_cache_types = ARRAY_SIZE(navi10_cache_info);
 		break;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 2c030c2b5b8d..25a2703621a7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -46,6 +46,7 @@ extern const struct kfd2kgd_calls gfx_v8_kfd2kgd;
 extern const struct kfd2kgd_calls gfx_v9_kfd2kgd;
 extern const struct kfd2kgd_calls arcturus_kfd2kgd;
 extern const struct kfd2kgd_calls gfx_v10_kfd2kgd;
+extern const struct kfd2kgd_calls gfx_v10_3_kfd2kgd;
 
 static const struct kfd2kgd_calls *kfd2kgd_funcs[] = {
 #ifdef KFD_SUPPORT_IOMMU_V2
@@ -72,6 +73,7 @@ static const struct kfd2kgd_calls *kfd2kgd_funcs[] = {
 	[CHIP_NAVI10] = &gfx_v10_kfd2kgd,
 	[CHIP_NAVI12] = &gfx_v10_kfd2kgd,
 	[CHIP_NAVI14] = &gfx_v10_kfd2kgd,
+	[CHIP_SIENNA_CICHLID] = &gfx_v10_3_kfd2kgd,
 };
 
 #ifdef KFD_SUPPORT_IOMMU_V2
@@ -458,6 +460,24 @@ static const struct kfd_device_info navi14_device_info = {
 	.num_sdma_queues_per_engine = 8,
 };
 
+static const struct kfd_device_info sienna_cichlid_device_info = {
+	.asic_family = CHIP_SIENNA_CICHLID,
+	.asic_name = "sienna_cichlid",
+	.max_pasid_bits = 16,
+	.max_no_of_hqd  = 24,
+	.doorbell_size  = 8,
+	.ih_ring_entry_size = 8 * sizeof(uint32_t),
+	.event_interrupt_class = &event_interrupt_class_v9,
+	.num_of_watch_points = 4,
+	.mqd_size_aligned = MQD_SIZE_ALIGNED,
+	.needs_iommu_device = false,
+	.supports_cwsr = true,
+	.needs_pci_atomics = false,
+	.num_sdma_engines = 4,
+	.num_xgmi_sdma_engines = 0,
+	.num_sdma_queues_per_engine = 8,
+};
+
 /* For each entry, [0] is regular and [1] is virtualisation device. */
 static const struct kfd_device_info *kfd_supported_devices[][2] = {
 #ifdef KFD_SUPPORT_IOMMU_V2
@@ -480,6 +500,7 @@ static const struct kfd_device_info *kfd_supported_devices[][2] = {
 	[CHIP_NAVI10] = {&navi10_device_info, NULL},
 	[CHIP_NAVI12] = {&navi12_device_info, &navi12_device_info},
 	[CHIP_NAVI14] = {&navi14_device_info, NULL},
+	[CHIP_SIENNA_CICHLID] = {&sienna_cichlid_device_info, NULL},
 };
 
 static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 6293017bd5bf..fcc927c2f887 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -1943,6 +1943,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 	case CHIP_NAVI10:
 	case CHIP_NAVI12:
 	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
 		device_queue_manager_init_v10_navi10(&dqm->asic_ops);
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
index 78714f9a8b11..b4674cf73132 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
@@ -415,6 +415,7 @@ int kfd_init_apertures(struct kfd_process *process)
 			case CHIP_NAVI10:
 			case CHIP_NAVI12:
 			case CHIP_NAVI14:
+			case CHIP_SIENNA_CICHLID:
 				kfd_init_apertures_v9(pdd, id);
 				break;
 			default:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
index 685ca82d42fe..89d7f08d749f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
@@ -245,6 +245,7 @@ int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
 	case CHIP_NAVI10:
 	case CHIP_NAVI12:
 	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
 		pm->pmf = &kfd_v9_pm_funcs;
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index d5e2585d6f34..4025efc619c7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -1363,6 +1363,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	case CHIP_NAVI10:
 	case CHIP_NAVI12:
 	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
 		dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 <<
 			HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) &
 			HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 117/207] drm/amdkfd: Add Sienna_Cichlid trap handler support

[Alex Deucher]

From: Jay Cornwall <jay.cornwall@amd.com>

- Replace SQC stores with TCP stores
- Synchronize with MSG_SAVEWAVE via lgkmcnt
- HW_REG_IB_STS is now read-only

Signed-off-by: Jay Cornwall <jay.cornwall@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../gpu/drm/amd/amdkfd/cwsr_trap_handler.h    | 844 +++++++++++++-----
 .../amd/amdkfd/cwsr_trap_handler_gfx10.asm    | 154 +++-
 drivers/gpu/drm/amd/amdkfd/kfd_device.c       |   4 +
 3 files changed, 731 insertions(+), 271 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
index d3400da6ab64..0aeff2cc3efe 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
@@ -679,8 +679,8 @@ static const uint32_t cwsr_trap_gfx9_hex[] = {
 	0xbf810000, 0x00000000,
 };
 
-static const uint32_t cwsr_trap_gfx10_hex[] = {
-	0xbf820001, 0xbf8201c1,
+static const uint32_t cwsr_trap_nv1x_hex[] = {
+	0xbf820001, 0xbf8201bc,
 	0xb0804004, 0xb978f802,
 	0x8a788678, 0xb971f803,
 	0x876eff71, 0x00000400,
@@ -736,7 +736,22 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x90768176, 0x88777677,
 	0xbefb037c, 0xbefa0380,
 	0xb97302dc, 0x8f739973,
-	0x8873737f, 0xb97a2a05,
+	0x8873737f, 0xbefe03c1,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0xbeff0380, 0xbf820002,
+	0xbeff03c1, 0xbf82000b,
+	0xbef603ff, 0x01000000,
+	0xe0704000, 0x7a5d0000,
+	0xe0704080, 0x7a5d0100,
+	0xe0704100, 0x7a5d0200,
+	0xe0704180, 0x7a5d0300,
+	0xbf82000a, 0xbef603ff,
+	0x01000000, 0xe0704000,
+	0x7a5d0000, 0xe0704100,
+	0x7a5d0100, 0xe0704200,
+	0x7a5d0200, 0xe0704300,
+	0x7a5d0300, 0xb97a2a05,
 	0x807a817a, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0x8f7a897a,
@@ -779,9 +794,7 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xb97bf815, 0xbefe037c,
 	0xbefc037a, 0xf4611efa,
 	0xf8000000, 0x807a847a,
-	0xbefc037e, 0x8776ff7f,
-	0x04000000, 0xbeef0380,
-	0x886f6f76, 0xb97a2a05,
+	0xbefc037e, 0xb97a2a05,
 	0x807a817a, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0x8f7a897a,
@@ -809,190 +822,105 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xf469013a, 0xfa000010,
 	0xf469023a, 0xfa000020,
 	0x8074b074, 0x82758075,
-	0xbef40372, 0xbefa0380,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0xbeff0380,
-	0xbf820002, 0xbeff03c1,
-	0xbf82000b, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x7a5d0000, 0xe0704080,
-	0x7a5d0100, 0xe0704100,
-	0x7a5d0200, 0xe0704180,
-	0x7a5d0300, 0xbf82000a,
-	0xbef603ff, 0x01000000,
-	0xe0704000, 0x7a5d0000,
-	0xe0704100, 0x7a5d0100,
-	0xe0704200, 0x7a5d0200,
-	0xe0704300, 0x7a5d0300,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0xbeff0380,
-	0xbf820001, 0xbeff03c1,
-	0xb9714306, 0x8771c171,
-	0xbf840046, 0xbf8a0000,
-	0x8776ff6f, 0x04000000,
-	0xbf840042, 0x8f718671,
-	0x8f718271, 0xbef60371,
-	0xb97a2a05, 0x807a817a,
+	0xbef40372, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0x8f7a897a, 0xbf820001,
-	0x8f7a8a7a, 0xb9761e06,
-	0x8f768a76, 0x807a767a,
-	0x807aff7a, 0x00000200,
-	0x807aff7a, 0x00000080,
-	0xbef603ff, 0x01000000,
-	0xd7650000, 0x000100c1,
-	0xd7660000, 0x000200c1,
-	0x16000084, 0x907c9973,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb9714306,
+	0x8771c171, 0xbf840046,
+	0xbf8a0000, 0x8776ff73,
+	0x04000000, 0xbf840042,
+	0x8f718671, 0x8f718271,
+	0xbef60371, 0xb97a2a05,
+	0x807a817a, 0x907c9973,
 	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850012,
-	0xbe8303ff, 0x00000080,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0xb9761e06, 0x8f768a76,
+	0x807a767a, 0x807aff7a,
+	0x00000200, 0x807aff7a,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0xd7650000,
+	0x000100c1, 0xd7660000,
+	0x000200c1, 0x16000084,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850012, 0xbe8303ff,
+	0x00000080, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x7a5d0100,
-	0x807c037c, 0x807a037a,
-	0xd5250000, 0x0001ff00,
-	0x00000080, 0xbf0a717c,
-	0xbf85fff4, 0xbf820011,
-	0xbe8303ff, 0x00000100,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x7a5d0100, 0x807c037c,
+	0x807a037a, 0xd5250000,
+	0x0001ff00, 0x00000080,
+	0xbf0a717c, 0xbf85fff4,
+	0xbf820011, 0xbe8303ff,
+	0x00000100, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x7a5d0100,
-	0x807c037c, 0x807a037a,
-	0xd5250000, 0x0001ff00,
-	0x00000100, 0xbf0a717c,
-	0xbf85fff4, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850004,
-	0xbefa03ff, 0x00000200,
-	0xbeff0380, 0xbf820003,
-	0xbefa03ff, 0x00000400,
-	0xbeff03c1, 0xb9712a05,
-	0x80718171, 0x8f718271,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850017,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x7a5d0100, 0x807c037c,
+	0x807a037a, 0xd5250000,
+	0x0001ff00, 0x00000100,
+	0xbf0a717c, 0xbf85fff4,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850004, 0xbefa03ff,
+	0x00000200, 0xbeff0380,
+	0xbf820003, 0xbefa03ff,
+	0x00000400, 0xbeff03c1,
+	0xb9712a05, 0x80718171,
+	0x8f718271, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850017, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a717c, 0xbf840037,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x7a5d0000,
+	0xe0704080, 0x7a5d0100,
+	0xe0704100, 0x7a5d0200,
+	0xe0704180, 0x7a5d0300,
+	0x807c847c, 0x807aff7a,
+	0x00000200, 0xbf0a717c,
+	0xbf85ffef, 0xbf820025,
 	0xbef603ff, 0x01000000,
 	0xbefc0384, 0xbf0a717c,
-	0xbf840037, 0x7e008700,
+	0xbf840020, 0x7e008700,
 	0x7e028701, 0x7e048702,
 	0x7e068703, 0xe0704000,
-	0x7a5d0000, 0xe0704080,
-	0x7a5d0100, 0xe0704100,
-	0x7a5d0200, 0xe0704180,
+	0x7a5d0000, 0xe0704100,
+	0x7a5d0100, 0xe0704200,
+	0x7a5d0200, 0xe0704300,
 	0x7a5d0300, 0x807c847c,
-	0x807aff7a, 0x00000200,
+	0x807aff7a, 0x00000400,
 	0xbf0a717c, 0xbf85ffef,
-	0xbf820025, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a717c, 0xbf840020,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
+	0xb9711e06, 0x8771c171,
+	0xbf84000c, 0x8f718371,
+	0x80717c71, 0xbefe03c1,
+	0xbeff0380, 0x7e008700,
 	0xe0704000, 0x7a5d0000,
-	0xe0704100, 0x7a5d0100,
-	0xe0704200, 0x7a5d0200,
-	0xe0704300, 0x7a5d0300,
-	0x807c847c, 0x807aff7a,
-	0x00000400, 0xbf0a717c,
-	0xbf85ffef, 0xb9711e06,
-	0x8771c171, 0xbf84000c,
-	0x8f718371, 0x80717c71,
-	0xbefe03c1, 0xbeff0380,
-	0x7e008700, 0xe0704000,
-	0x7a5d0000, 0x807c817c,
-	0x807aff7a, 0x00000080,
-	0xbf0a717c, 0xbf85fff8,
-	0xbf820142, 0xbef4037e,
-	0x8775ff7f, 0x0000ffff,
-	0x8875ff75, 0x00040000,
-	0xbef60380, 0xbef703ff,
-	0x10807fac, 0x8772ff7f,
-	0x08000000, 0x90728372,
-	0x88777277, 0x8772ff7f,
-	0x70000000, 0x90728172,
-	0x88777277, 0xb97302dc,
-	0x8f739973, 0x8873737f,
-	0x8772ff7f, 0x04000000,
-	0xbf840036, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb96f4306,
-	0x876fc16f, 0xbf84002b,
-	0x8f6f866f, 0x8f6f826f,
-	0xbef6036f, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0x8078ff78,
-	0x00000080, 0xbef603ff,
-	0x01000000, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850009,
-	0xe0310000, 0x781d0000,
-	0x807cff7c, 0x00000080,
-	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff8,
-	0xbf820008, 0xe0310000,
-	0x781d0000, 0x807cff7c,
-	0x00000100, 0x8078ff78,
-	0x00000100, 0xbf0a6f7c,
-	0xbf85fff8, 0xbef80380,
+	0x807c817c, 0x807aff7a,
+	0x00000080, 0xbf0a717c,
+	0xbf85fff8, 0xbf820142,
+	0xbef4037e, 0x8775ff7f,
+	0x0000ffff, 0x8875ff75,
+	0x00040000, 0xbef60380,
+	0xbef703ff, 0x10807fac,
+	0x8772ff7f, 0x08000000,
+	0x90728372, 0x88777277,
+	0x8772ff7f, 0x70000000,
+	0x90728172, 0x88777277,
+	0xb97302dc, 0x8f739973,
+	0x8873737f, 0x8772ff7f,
+	0x04000000, 0xbf840036,
 	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
-	0xb96f2a05, 0x806f816f,
-	0x8f6f826f, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850021, 0xbef603ff,
-	0x01000000, 0xbef20378,
-	0x8078ff78, 0x00000200,
-	0xbefc0384, 0xe0304000,
-	0x785d0000, 0xe0304080,
-	0x785d0100, 0xe0304100,
-	0x785d0200, 0xe0304180,
-	0x785d0300, 0xbf8c3f70,
-	0x7e008500, 0x7e028501,
-	0x7e048502, 0x7e068503,
-	0x807c847c, 0x8078ff78,
-	0x00000200, 0xbf0a6f7c,
-	0xbf85ffee, 0xe0304000,
-	0x725d0000, 0xe0304080,
-	0x725d0100, 0xe0304100,
-	0x725d0200, 0xe0304180,
-	0x725d0300, 0xbf820032,
-	0xbef603ff, 0x01000000,
-	0xbef20378, 0x8078ff78,
-	0x00000400, 0xbefc0384,
-	0xe0304000, 0x785d0000,
-	0xe0304100, 0x785d0100,
-	0xe0304200, 0x785d0200,
-	0xe0304300, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
-	0x7e028501, 0x7e048502,
-	0x7e068503, 0x807c847c,
-	0x8078ff78, 0x00000400,
-	0xbf0a6f7c, 0xbf85ffee,
-	0xb96f1e06, 0x876fc16f,
-	0xbf84000e, 0x8f6f836f,
-	0x806f7c6f, 0xbefe03c1,
-	0xbeff0380, 0xe0304000,
-	0x785d0000, 0xbf8c3f70,
-	0x7e008500, 0x807c817c,
-	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff7,
-	0xbeff03c1, 0xe0304000,
-	0x725d0000, 0xe0304100,
-	0x725d0100, 0xe0304200,
-	0x725d0200, 0xe0304300,
-	0x725d0300, 0xbf8c3f70,
+	0xb96f4306, 0x876fc16f,
+	0xbf84002b, 0x8f6f866f,
+	0x8f6f826f, 0xbef6036f,
 	0xb9782a05, 0x80788178,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
@@ -1000,77 +928,147 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x8f788a78, 0xb9721e06,
 	0x8f728a72, 0x80787278,
 	0x8078ff78, 0x00000200,
-	0x80f8ff78, 0x00000050,
+	0x8078ff78, 0x00000080,
 	0xbef603ff, 0x01000000,
-	0xbefc03ff, 0x0000006c,
-	0x80f89078, 0xf429003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc847c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0x80f8a078, 0xf42d003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc887c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0x80f8c078, 0xf431003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc907c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0xbe883108, 0xbe8a310a,
-	0xbe8c310c, 0xbe8e310e,
-	0xbf06807c, 0xbf84fff0,
-	0xb9782a05, 0x80788178,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850009, 0xe0310000,
+	0x781d0000, 0x807cff7c,
+	0x00000080, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff8, 0xbf820008,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000100,
+	0x8078ff78, 0x00000100,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbef80380, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb9721e06,
-	0x8f728a72, 0x80787278,
-	0x8078ff78, 0x00000200,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb96f2a05,
+	0x806f816f, 0x8f6f826f,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850021,
 	0xbef603ff, 0x01000000,
-	0xf4211bfa, 0xf0000000,
-	0x80788478, 0xf4211b3a,
+	0xbef20378, 0x8078ff78,
+	0x00000200, 0xbefc0384,
+	0xe0304000, 0x785d0000,
+	0xe0304080, 0x785d0100,
+	0xe0304100, 0x785d0200,
+	0xe0304180, 0x785d0300,
+	0xbf8c3f70, 0x7e008500,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
+	0x8078ff78, 0x00000200,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xe0304000, 0x725d0000,
+	0xe0304080, 0x725d0100,
+	0xe0304100, 0x725d0200,
+	0xe0304180, 0x725d0300,
+	0xbf820032, 0xbef603ff,
+	0x01000000, 0xbef20378,
+	0x8078ff78, 0x00000400,
+	0xbefc0384, 0xe0304000,
+	0x785d0000, 0xe0304100,
+	0x785d0100, 0xe0304200,
+	0x785d0200, 0xe0304300,
+	0x785d0300, 0xbf8c3f70,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807c847c, 0x8078ff78,
+	0x00000400, 0xbf0a6f7c,
+	0xbf85ffee, 0xb96f1e06,
+	0x876fc16f, 0xbf84000e,
+	0x8f6f836f, 0x806f7c6f,
+	0xbefe03c1, 0xbeff0380,
+	0xe0304000, 0x785d0000,
+	0xbf8c3f70, 0x7e008500,
+	0x807c817c, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff7, 0xbeff03c1,
+	0xe0304000, 0x725d0000,
+	0xe0304100, 0x725d0100,
+	0xe0304200, 0x725d0200,
+	0xe0304300, 0x725d0300,
+	0xbf8c3f70, 0xb9782a05,
+	0x80788178, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb9721e06, 0x8f728a72,
+	0x80787278, 0x8078ff78,
+	0x00000200, 0x80f8ff78,
+	0x00000050, 0xbef603ff,
+	0x01000000, 0xbefc03ff,
+	0x0000006c, 0x80f89078,
+	0xf429003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc847c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0x80f8a078,
+	0xf42d003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc887c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0x80f8c078,
+	0xf431003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc907c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0xbe883108,
+	0xbe8a310a, 0xbe8c310c,
+	0xbe8e310e, 0xbf06807c,
+	0xbf84fff0, 0xb9782a05,
+	0x80788178, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb9721e06, 0x8f728a72,
+	0x80787278, 0x8078ff78,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0xf4211bfa,
 	0xf0000000, 0x80788478,
-	0xf4211b7a, 0xf0000000,
-	0x80788478, 0xf4211eba,
+	0xf4211b3a, 0xf0000000,
+	0x80788478, 0xf4211b7a,
 	0xf0000000, 0x80788478,
-	0xf4211efa, 0xf0000000,
-	0x80788478, 0xf4211c3a,
+	0xf4211eba, 0xf0000000,
+	0x80788478, 0xf4211efa,
 	0xf0000000, 0x80788478,
-	0xf4211c7a, 0xf0000000,
-	0x80788478, 0xf4211e7a,
+	0xf4211c3a, 0xf0000000,
+	0x80788478, 0xf4211c7a,
 	0xf0000000, 0x80788478,
-	0xf4211cfa, 0xf0000000,
-	0x80788478, 0xf4211bba,
+	0xf4211e7a, 0xf0000000,
+	0x80788478, 0xf4211cfa,
 	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef815, 0xbef2036d,
-	0x876dff72, 0x0000ffff,
-	0xbefc036f, 0xbefe037a,
-	0xbeff037b, 0x876f71ff,
-	0x000003ff, 0xb9ef4803,
-	0xb9f9f816, 0x876f71ff,
-	0xfffff800, 0x906f8b6f,
-	0xb9efa2c3, 0xb9f3f801,
-	0x876fff72, 0xfc000000,
-	0x906f9a6f, 0x8f6f906f,
-	0xbef30380, 0x88736f73,
-	0x876fff72, 0x02000000,
-	0x906f996f, 0x8f6f8f6f,
+	0xb9eef814, 0xf4211bba,
+	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef815,
+	0xbef2036d, 0x876dff72,
+	0x0000ffff, 0xbefc036f,
+	0xbefe037a, 0xbeff037b,
+	0x876f71ff, 0x000003ff,
+	0xb9ef4803, 0xb9f9f816,
+	0x876f71ff, 0xfffff800,
+	0x906f8b6f, 0xb9efa2c3,
+	0xb9f3f801, 0x876fff72,
+	0xfc000000, 0x906f9a6f,
+	0x8f6f906f, 0xbef30380,
 	0x88736f73, 0x876fff72,
-	0x01000000, 0x906f986f,
-	0x8f6f996f, 0x88736f73,
-	0x876fff70, 0x00800000,
-	0x906f976f, 0xb9f3f807,
-	0x87fe7e7e, 0x87ea6a6a,
-	0xb9f0f802, 0xbf8a0000,
-	0xbe80226c, 0xbf810000,
+	0x02000000, 0x906f996f,
+	0x8f6f8f6f, 0x88736f73,
+	0x876fff72, 0x01000000,
+	0x906f986f, 0x8f6f996f,
+	0x88736f73, 0x876fff70,
+	0x00800000, 0x906f976f,
+	0xb9f3f807, 0x87fe7e7e,
+	0x87ea6a6a, 0xb9f0f802,
+	0xbf8a0000, 0xbe80226c,
+	0xbf810000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
-	0xbf9f0000, 0x00000000,
 };
+
 static const uint32_t cwsr_trap_arcturus_hex[] = {
 	0xbf820001, 0xbf8202c4,
 	0xb8f8f802, 0x89788678,
@@ -1560,3 +1558,389 @@ static const uint32_t cwsr_trap_arcturus_hex[] = {
 	0xbf8a0000, 0x95806f6c,
 	0xbf810000, 0x00000000,
 };
+
+static const uint32_t cwsr_trap_gfx10_hex[] = {
+	0xbf820001, 0xbf8201c7,
+	0xb0804004, 0xb978f802,
+	0x8a788678, 0xb971f803,
+	0x876eff71, 0x00000400,
+	0xbf85001d, 0x876eff71,
+	0x00000100, 0xbf840002,
+	0x8878ff78, 0x00002000,
+	0xb970f812, 0xb971f813,
+	0x8ff08870, 0xf4051bb8,
+	0xfa000000, 0xbf8cc07f,
+	0xf4051c38, 0xfa000008,
+	0xbf8cc07f, 0x87ee6e6e,
+	0xbf840001, 0xbe80206e,
+	0xb971f803, 0x8771ff71,
+	0x000001ff, 0xbf850002,
+	0x806c846c, 0x826d806d,
+	0x876dff6d, 0x0000ffff,
+	0x87fe7e7e, 0x87ea6a6a,
+	0xb9f8f802, 0xbe80226c,
+	0xb971f803, 0x8771ff71,
+	0x00000100, 0xbf840006,
+	0xbef60380, 0xb9f60203,
+	0x876dff6d, 0x0000ffff,
+	0x80ec886c, 0x82ed806d,
+	0xbef60380, 0xb9f60283,
+	0xb972f816, 0xbeee037e,
+	0xbeef037f, 0xbefe0480,
+	0xbf900004, 0xbf8cc07f,
+	0xbef4037e, 0x8775ff7f,
+	0x0000ffff, 0x8875ff75,
+	0x00040000, 0xbef60380,
+	0xbef703ff, 0x10807fac,
+	0x8776ff7f, 0x08000000,
+	0x90768376, 0x88777677,
+	0x8776ff7f, 0x70000000,
+	0x90768176, 0x88777677,
+	0xbefb037c, 0xbefa0380,
+	0xb97302dc, 0x8f739973,
+	0x8873737f, 0xbefe03c1,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0xbeff0380, 0xbf820002,
+	0xbeff03c1, 0xbf82000b,
+	0xbef603ff, 0x01000000,
+	0xe0704000, 0x7a5d0000,
+	0xe0704080, 0x7a5d0100,
+	0xe0704100, 0x7a5d0200,
+	0xe0704180, 0x7a5d0300,
+	0xbf82000a, 0xbef603ff,
+	0x01000000, 0xe0704000,
+	0x7a5d0000, 0xe0704100,
+	0x7a5d0100, 0xe0704200,
+	0x7a5d0200, 0xe0704300,
+	0x7a5d0300, 0xb97a2a05,
+	0x807a817a, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0xb9761e06, 0x8f768a76,
+	0x807a767a, 0x807aff7a,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0xbefc0380, 0xd7610002,
+	0x0000f87b, 0x807c817c,
+	0xd7610002, 0x0000f86c,
+	0x807c817c, 0xd7610002,
+	0x0000f86d, 0x807c817c,
+	0xd7610002, 0x0000f86e,
+	0x807c817c, 0xd7610002,
+	0x0000f86f, 0x807c817c,
+	0xd7610002, 0x0000f878,
+	0x807c817c, 0xb971f803,
+	0xd7610002, 0x0000f871,
+	0x807c817c, 0xd7610002,
+	0x0000f872, 0x807c817c,
+	0xb97bf801, 0xd7610002,
+	0x0000f87b, 0x807c817c,
+	0xb97bf814, 0xd7610002,
+	0x0000f87b, 0x807c817c,
+	0xb97bf815, 0xd7610002,
+	0x0000f87b, 0x807c817c,
+	0xbeff0380, 0xe0704000,
+	0x7a5d0200, 0xb97a2a05,
+	0x807a817a, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0xb9761e06, 0x8f768a76,
+	0x807a767a, 0xbef603ff,
+	0x01000000, 0xbef90380,
+	0xbefc0380, 0xbf800000,
+	0xbe802f00, 0xbe822f02,
+	0xbe842f04, 0xbe862f06,
+	0xbe882f08, 0xbe8a2f0a,
+	0xbe8c2f0c, 0xbe8e2f0e,
+	0xd7610002, 0x0000f200,
+	0x80798179, 0xd7610002,
+	0x0000f201, 0x80798179,
+	0xd7610002, 0x0000f202,
+	0x80798179, 0xd7610002,
+	0x0000f203, 0x80798179,
+	0xd7610002, 0x0000f204,
+	0x80798179, 0xd7610002,
+	0x0000f205, 0x80798179,
+	0xd7610002, 0x0000f206,
+	0x80798179, 0xd7610002,
+	0x0000f207, 0x80798179,
+	0xd7610002, 0x0000f208,
+	0x80798179, 0xd7610002,
+	0x0000f209, 0x80798179,
+	0xd7610002, 0x0000f20a,
+	0x80798179, 0xd7610002,
+	0x0000f20b, 0x80798179,
+	0xd7610002, 0x0000f20c,
+	0x80798179, 0xd7610002,
+	0x0000f20d, 0x80798179,
+	0xd7610002, 0x0000f20e,
+	0x80798179, 0xd7610002,
+	0x0000f20f, 0x80798179,
+	0xbf06a079, 0xbf840006,
+	0xe0704000, 0x7a5d0200,
+	0x807aff7a, 0x00000080,
+	0xbef90380, 0x7e040280,
+	0x807c907c, 0xbf0aff7c,
+	0x00000060, 0xbf85ffbc,
+	0xbe802f00, 0xbe822f02,
+	0xbe842f04, 0xbe862f06,
+	0xbe882f08, 0xbe8a2f0a,
+	0xd7610002, 0x0000f200,
+	0x80798179, 0xd7610002,
+	0x0000f201, 0x80798179,
+	0xd7610002, 0x0000f202,
+	0x80798179, 0xd7610002,
+	0x0000f203, 0x80798179,
+	0xd7610002, 0x0000f204,
+	0x80798179, 0xd7610002,
+	0x0000f205, 0x80798179,
+	0xd7610002, 0x0000f206,
+	0x80798179, 0xd7610002,
+	0x0000f207, 0x80798179,
+	0xd7610002, 0x0000f208,
+	0x80798179, 0xd7610002,
+	0x0000f209, 0x80798179,
+	0xd7610002, 0x0000f20a,
+	0x80798179, 0xd7610002,
+	0x0000f20b, 0x80798179,
+	0xe0704000, 0x7a5d0200,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820001, 0xbeff03c1,
+	0xb9714306, 0x8771c171,
+	0xbf840046, 0xbf8a0000,
+	0x8776ff73, 0x04000000,
+	0xbf840042, 0x8f718671,
+	0x8f718271, 0xbef60371,
+	0xb97a2a05, 0x807a817a,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0x8f7a897a, 0xbf820001,
+	0x8f7a8a7a, 0xb9761e06,
+	0x8f768a76, 0x807a767a,
+	0x807aff7a, 0x00000200,
+	0x807aff7a, 0x00000080,
+	0xbef603ff, 0x01000000,
+	0xd7650000, 0x000100c1,
+	0xd7660000, 0x000200c1,
+	0x16000084, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850012,
+	0xbe8303ff, 0x00000080,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x7a5d0100,
+	0x807c037c, 0x807a037a,
+	0xd5250000, 0x0001ff00,
+	0x00000080, 0xbf0a717c,
+	0xbf85fff4, 0xbf820011,
+	0xbe8303ff, 0x00000100,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x7a5d0100,
+	0x807c037c, 0x807a037a,
+	0xd5250000, 0x0001ff00,
+	0x00000100, 0xbf0a717c,
+	0xbf85fff4, 0xbefe03c1,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850004,
+	0xbefa03ff, 0x00000200,
+	0xbeff0380, 0xbf820003,
+	0xbefa03ff, 0x00000400,
+	0xbeff03c1, 0xb9712a05,
+	0x80718171, 0x8f718271,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850017,
+	0xbef603ff, 0x01000000,
+	0xbefc0384, 0xbf0a717c,
+	0xbf840037, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xe0704000,
+	0x7a5d0000, 0xe0704080,
+	0x7a5d0100, 0xe0704100,
+	0x7a5d0200, 0xe0704180,
+	0x7a5d0300, 0x807c847c,
+	0x807aff7a, 0x00000200,
+	0xbf0a717c, 0xbf85ffef,
+	0xbf820025, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a717c, 0xbf840020,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x7a5d0000,
+	0xe0704100, 0x7a5d0100,
+	0xe0704200, 0x7a5d0200,
+	0xe0704300, 0x7a5d0300,
+	0x807c847c, 0x807aff7a,
+	0x00000400, 0xbf0a717c,
+	0xbf85ffef, 0xb9711e06,
+	0x8771c171, 0xbf84000c,
+	0x8f718371, 0x80717c71,
+	0xbefe03c1, 0xbeff0380,
+	0x7e008700, 0xe0704000,
+	0x7a5d0000, 0x807c817c,
+	0x807aff7a, 0x00000080,
+	0xbf0a717c, 0xbf85fff8,
+	0xbf82012e, 0xbef4037e,
+	0x8775ff7f, 0x0000ffff,
+	0x8875ff75, 0x00040000,
+	0xbef60380, 0xbef703ff,
+	0x10807fac, 0x8772ff7f,
+	0x08000000, 0x90728372,
+	0x88777277, 0x8772ff7f,
+	0x70000000, 0x90728172,
+	0x88777277, 0xb97302dc,
+	0x8f739973, 0x8873737f,
+	0x8772ff7f, 0x04000000,
+	0xbf840036, 0xbefe03c1,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb96f4306,
+	0x876fc16f, 0xbf84002b,
+	0x8f6f866f, 0x8f6f826f,
+	0xbef6036f, 0xb9782a05,
+	0x80788178, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb9721e06, 0x8f728a72,
+	0x80787278, 0x8078ff78,
+	0x00000200, 0x8078ff78,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850009,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000080,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbf820008, 0xe0310000,
+	0x781d0000, 0x807cff7c,
+	0x00000100, 0x8078ff78,
+	0x00000100, 0xbf0a6f7c,
+	0xbf85fff8, 0xbef80380,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820001, 0xbeff03c1,
+	0xb96f2a05, 0x806f816f,
+	0x8f6f826f, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850021, 0xbef603ff,
+	0x01000000, 0xbef20378,
+	0x8078ff78, 0x00000200,
+	0xbefc0384, 0xe0304000,
+	0x785d0000, 0xe0304080,
+	0x785d0100, 0xe0304100,
+	0x785d0200, 0xe0304180,
+	0x785d0300, 0xbf8c3f70,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807c847c, 0x8078ff78,
+	0x00000200, 0xbf0a6f7c,
+	0xbf85ffee, 0xe0304000,
+	0x725d0000, 0xe0304080,
+	0x725d0100, 0xe0304100,
+	0x725d0200, 0xe0304180,
+	0x725d0300, 0xbf820032,
+	0xbef603ff, 0x01000000,
+	0xbef20378, 0x8078ff78,
+	0x00000400, 0xbefc0384,
+	0xe0304000, 0x785d0000,
+	0xe0304100, 0x785d0100,
+	0xe0304200, 0x785d0200,
+	0xe0304300, 0x785d0300,
+	0xbf8c3f70, 0x7e008500,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xb96f1e06, 0x876fc16f,
+	0xbf84000e, 0x8f6f836f,
+	0x806f7c6f, 0xbefe03c1,
+	0xbeff0380, 0xe0304000,
+	0x785d0000, 0xbf8c3f70,
+	0x7e008500, 0x807c817c,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff7,
+	0xbeff03c1, 0xe0304000,
+	0x725d0000, 0xe0304100,
+	0x725d0100, 0xe0304200,
+	0x725d0200, 0xe0304300,
+	0x725d0300, 0xbf8c3f70,
+	0xb9782a05, 0x80788178,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb9721e06,
+	0x8f728a72, 0x80787278,
+	0x8078ff78, 0x00000200,
+	0x80f8ff78, 0x00000050,
+	0xbef603ff, 0x01000000,
+	0xbefc03ff, 0x0000006c,
+	0x80f89078, 0xf429003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc847c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0x80f8a078, 0xf42d003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc887c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0x80f8c078, 0xf431003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0xbe883108, 0xbe8a310a,
+	0xbe8c310c, 0xbe8e310e,
+	0xbf06807c, 0xbf84fff0,
+	0xb9782a05, 0x80788178,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb9721e06,
+	0x8f728a72, 0x80787278,
+	0x8078ff78, 0x00000200,
+	0xbef603ff, 0x01000000,
+	0xf4211bfa, 0xf0000000,
+	0x80788478, 0xf4211b3a,
+	0xf0000000, 0x80788478,
+	0xf4211b7a, 0xf0000000,
+	0x80788478, 0xf4211eba,
+	0xf0000000, 0x80788478,
+	0xf4211efa, 0xf0000000,
+	0x80788478, 0xf4211c3a,
+	0xf0000000, 0x80788478,
+	0xf4211c7a, 0xf0000000,
+	0x80788478, 0xf4211e7a,
+	0xf0000000, 0x80788478,
+	0xf4211cfa, 0xf0000000,
+	0x80788478, 0xf4211bba,
+	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef814,
+	0xf4211bba, 0xf0000000,
+	0x80788478, 0xbf8cc07f,
+	0xb9eef815, 0xbef2036d,
+	0x876dff72, 0x0000ffff,
+	0xbefc036f, 0xbefe037a,
+	0xbeff037b, 0x876f71ff,
+	0x000003ff, 0xb9ef4803,
+	0xb9f9f816, 0x876f71ff,
+	0xfffff800, 0x906f8b6f,
+	0xb9efa2c3, 0xb9f3f801,
+	0x87fe7e7e, 0x87ea6a6a,
+	0xb9f0f802, 0xbf8a0000,
+	0xbe80226c, 0xbf810000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0x00000000,
+};
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
index 4433bda2ce25..ce70bba8ac06 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
@@ -20,6 +20,17 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+/* To compile this assembly code:
+ *
+ * Navi1x:
+ *   PROJECT=gfx10 ./sp3 ASIC_TARGET_NAVI1X=1 cwsr_trap_handler_gfx10.asm -hex tmp.hex
+ *
+ * Others:
+ *   PROJECT=gfx10 ./sp3 ASIC_TARGET_NAVI1X=0 cwsr_trap_handler_gfx10.asm -hex tmp.hex
+ */
+
+var NO_SQC_STORE				= !ASIC_TARGET_NAVI1X
+
 var SQ_WAVE_STATUS_INST_ATC_SHIFT		= 23
 var SQ_WAVE_STATUS_INST_ATC_MASK		= 0x00800000
 var SQ_WAVE_STATUS_SPI_PRIO_MASK		= 0x00000006
@@ -170,6 +181,8 @@ L_SKIP_RESTORE:
 	s_or_b32	s_save_status, s_save_status, SQ_WAVE_STATUS_HALT_MASK
 
 L_FETCH_2ND_TRAP:
+
+if ASIC_TARGET_NAVI1X
 	// Preserve and clear scalar XNACK state before issuing scalar loads.
 	// Save IB_STS.REPLAY_W64H[25], RCNT[21:16], FIRST_REPLAY[15] into
 	// unused space ttmp11[31:24].
@@ -183,6 +196,7 @@ L_FETCH_2ND_TRAP:
 	s_or_b32	ttmp11, ttmp11, ttmp3
 	s_andn2_b32	ttmp2, ttmp2, (SQ_WAVE_IB_STS_REPLAY_W64H_MASK | SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK)
 	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
+end
 
 	// Read second-level TBA/TMA from first-level TMA and jump if available.
 	// ttmp[2:5] and ttmp12 can be used (others hold SPI-initialized debug data)
@@ -207,6 +221,7 @@ L_NO_NEXT_TRAP:
 L_EXCP_CASE:
 	s_and_b32	ttmp1, ttmp1, 0xFFFF
 
+if ASIC_TARGET_NAVI1X
 	// Restore SQ_WAVE_IB_STS.
 	s_lshr_b32	ttmp2, ttmp11, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
 	s_and_b32	ttmp3, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
@@ -214,6 +229,7 @@ L_EXCP_CASE:
 	s_and_b32	ttmp2, ttmp2, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
 	s_or_b32	ttmp2, ttmp2, ttmp3
 	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
+end
 
 	// Restore SQ_WAVE_STATUS.
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
@@ -240,6 +256,8 @@ L_NO_PC_REWIND:
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_SAVECTX_SHIFT, 1), s_save_tmp	//clear saveCtx bit
 
 	s_getreg_b32	s_save_xnack_mask, hwreg(HW_REG_SHADER_XNACK_MASK)
+
+if ASIC_TARGET_NAVI1X
 	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_RCNT_SHIFT, SQ_WAVE_IB_STS_RCNT_SIZE)
 	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_RCNT_SHIFT
 	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
@@ -253,6 +271,7 @@ L_NO_PC_REWIND:
 	s_and_b32	s_save_tmp, s_save_tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG
 
 	s_setreg_b32	hwreg(HW_REG_IB_STS), s_save_tmp
+end
 
 	/* inform SPI the readiness and wait for SPI's go signal */
 	s_mov_b32	s_save_exec_lo, exec_lo					//save EXEC and use EXEC for the go signal from SPI
@@ -261,12 +280,16 @@ L_NO_PC_REWIND:
 
 	s_sendmsg	sendmsg(MSG_SAVEWAVE)					//send SPI a message and wait for SPI's write to EXEC
 
+if ASIC_TARGET_NAVI1X
 L_SLEEP:
 	// sleep 1 (64clk) is not enough for 8 waves per SIMD, which will cause
 	// SQ hang, since the 7,8th wave could not get arbit to exec inst, while
 	// other waves are stuck into the sleep-loop and waiting for wrexec!=0
 	s_sleep		0x2
 	s_cbranch_execz	L_SLEEP
+else
+	s_waitcnt	lgkmcnt(0)
+end
 
 	/* setup Resource Contants */
 	s_mov_b32	s_save_buf_rsrc0, s_save_spi_init_lo			//base_addr_lo
@@ -289,6 +312,38 @@ L_SLEEP:
 	s_lshl_b32	s_wave_size, s_wave_size, S_WAVE_SIZE
 	s_or_b32	s_wave_size, s_save_spi_init_hi, s_wave_size		//share s_wave_size with exec_hi, it's at bit25
 
+	/* save first 4 VGPRs, needed for SGPR save */
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_SAVE_4VGPR_EXEC_HI
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_SAVE_4VGPR_WAVE32
+L_ENABLE_SAVE_4VGPR_EXEC_HI:
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+	s_branch	L_SAVE_4VGPR_WAVE64
+L_SAVE_4VGPR_WAVE32:
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR Allocated in 4-GPR granularity
+
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*3
+	s_branch	L_SAVE_HWREG
+
+L_SAVE_4VGPR_WAVE64:
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR Allocated in 4-GPR granularity
+
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*3
+
 	/* save HW registers */
 
 L_SAVE_HWREG:
@@ -300,6 +355,13 @@ L_SAVE_HWREG:
 
 	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
 
+if NO_SQC_STORE
+	v_mov_b32	v0, 0x0							//Offset[31:0] from buffer resource
+	v_mov_b32	v1, 0x0							//Offset[63:32] from buffer resource
+	v_mov_b32	v2, 0x0							//Set of SGPRs for TCP store
+	s_mov_b32	m0, 0x0							//Next lane of v2 to write to
+end
+
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_pc_lo, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_pc_hi, s_save_buf_rsrc0, s_save_mem_offset)
@@ -320,10 +382,11 @@ L_SAVE_HWREG:
 	s_getreg_b32	s_save_m0, hwreg(HW_REG_SHADER_FLAT_SCRATCH_HI)
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 
-	/* the first wave in the threadgroup */
-	s_and_b32	s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
-	s_mov_b32	s_save_exec_hi, 0x0
-	s_or_b32	s_save_exec_hi, s_save_tmp, s_save_exec_hi		// save first wave bit to s_save_exec_hi.bits[26]
+if NO_SQC_STORE
+	// Write HWREG/SGPRs with 32 VGPR lanes, wave32 is common case.
+	s_mov_b32	exec_hi, 0x0
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+end
 
 	/* save SGPRs */
 	// Save SGPR before LDS save, then the s0 to s4 can be used during LDS save...
@@ -334,10 +397,14 @@ L_SAVE_HWREG:
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, s_save_tmp
 	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
 
+if NO_SQC_STORE
+	s_mov_b32	ttmp13, 0x0						//next VGPR lane to copy SGPR into
+else
 	// backup s_save_buf_rsrc0,1 to s_save_pc_lo/hi, since write_16sgpr_to_mem function will change the rsrc0
 	s_mov_b32	s_save_xnack_mask, s_save_buf_rsrc0
 	s_add_u32	s_save_buf_rsrc0, s_save_buf_rsrc0, s_save_mem_offset
 	s_addc_u32	s_save_buf_rsrc1, s_save_buf_rsrc1, 0
+end
 
 	s_mov_b32	m0, 0x0							//SGPR initial index value =0
 	s_nop		0x0							//Manually inserted wait states
@@ -353,6 +420,18 @@ L_SAVE_SGPR_LOOP:
 	s_movrels_b64	s14, s14						//s14 = s[14+m0], s15 = s[15+m0]
 
 	write_16sgpr_to_mem(s0, s_save_buf_rsrc0, s_save_mem_offset)
+
+if NO_SQC_STORE
+	s_cmp_eq_u32	ttmp13, 0x20						//have 32 VGPR lanes filled?
+	s_cbranch_scc0	L_SAVE_SGPR_SKIP_TCP_STORE
+
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, 0x80
+	s_mov_b32	ttmp13, 0x0
+	v_mov_b32	v2, 0x0
+L_SAVE_SGPR_SKIP_TCP_STORE:
+end
+
 	s_add_u32	m0, m0, 16						//next sgpr index
 	s_cmp_lt_u32	m0, 96							//scc = (m0 < first 96 SGPR) ? 1 : 0
 	s_cbranch_scc1	L_SAVE_SGPR_LOOP					//first 96 SGPR save is complete?
@@ -366,43 +445,12 @@ L_SAVE_SGPR_LOOP:
 	s_movrels_b64	s10, s10						//s10 = s[10+m0], s11 = s[11+m0]
 	write_12sgpr_to_mem(s0, s_save_buf_rsrc0, s_save_mem_offset)
 
+if NO_SQC_STORE
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+else
 	// restore s_save_buf_rsrc0,1
 	s_mov_b32	s_save_buf_rsrc0, s_save_xnack_mask
-
-	/* save first 4 VGPR, then LDS save could use   */
-	// each wave will alloc 4 vgprs at least...
-
-	s_mov_b32	s_save_mem_offset, 0
- 	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
-	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
-	s_and_b32	m0, m0, 1
-	s_cmp_eq_u32	m0, 1
-	s_cbranch_scc1	L_ENABLE_SAVE_4VGPR_EXEC_HI
-	s_mov_b32	exec_hi, 0x00000000
-	s_branch	L_SAVE_4VGPR_WAVE32
-L_ENABLE_SAVE_4VGPR_EXEC_HI:
-	s_mov_b32	exec_hi, 0xFFFFFFFF
-	s_branch	L_SAVE_4VGPR_WAVE64
-L_SAVE_4VGPR_WAVE32:
-	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
-
-	// VGPR Allocated in 4-GPR granularity
-
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*3
-	s_branch	L_SAVE_LDS
-
-L_SAVE_4VGPR_WAVE64:
-	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
-
-	// VGPR Allocated in 4-GPR granularity
-
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*3
+end
 
 	/* save LDS */
 
@@ -423,7 +471,7 @@ L_SAVE_LDS_NORMAL:
 	s_cbranch_scc0	L_SAVE_LDS_DONE						//no lds used? jump to L_SAVE_DONE
 
 	s_barrier								//LDS is used? wait for other waves in the same TG
-	s_and_b32	s_save_tmp, s_save_exec_hi, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
+	s_and_b32	s_save_tmp, s_wave_size, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
 	s_cbranch_scc0	L_SAVE_LDS_DONE
 
 	// first wave do LDS save;
@@ -856,6 +904,8 @@ L_RESTORE_HWREG:
 	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE), s_restore_m0
 	s_setreg_b32	hwreg(HW_REG_MODE), s_restore_mode
+
+if ASIC_TARGET_NAVI1X
 	s_and_b32	s_restore_m0, s_restore_tmp, S_SAVE_PC_HI_RCNT_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_RCNT_SHIFT
 	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_RCNT_SHIFT
@@ -873,6 +923,7 @@ L_RESTORE_HWREG:
 	s_and_b32	s_restore_m0, s_restore_status, SQ_WAVE_STATUS_INST_ATC_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_STATUS_INST_ATC_SHIFT
 	s_setreg_b32 	hwreg(HW_REG_IB_STS), s_restore_mode
+end
 
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
@@ -887,15 +938,28 @@ L_END_PGM:
 end
 
 function write_hwreg_to_mem(s, s_rsrc, s_mem_offset)
+if NO_SQC_STORE
+	// Copy into VGPR for later TCP store.
+	v_writelane_b32	v2, s, m0
+	s_add_u32	m0, m0, 0x1
+else
 	s_mov_b32	exec_lo, m0
 	s_mov_b32	m0, s_mem_offset
 	s_buffer_store_dword	s, s_rsrc, m0 glc:1
 	s_add_u32	s_mem_offset, s_mem_offset, 4
 	s_mov_b32	m0, exec_lo
 end
+end
 
 
 function write_16sgpr_to_mem(s, s_rsrc, s_mem_offset)
+if NO_SQC_STORE
+	// Copy into VGPR for later TCP store.
+	for var sgpr_idx = 0; sgpr_idx < 16; sgpr_idx ++
+		v_writelane_b32	v2, s[sgpr_idx], ttmp13
+		s_add_u32	ttmp13, ttmp13, 0x1
+	end
+else
 	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
 	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
 	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
@@ -903,15 +967,23 @@ function write_16sgpr_to_mem(s, s_rsrc, s_mem_offset)
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*16
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
 end
+end
 
 function write_12sgpr_to_mem(s, s_rsrc, s_mem_offset)
+if NO_SQC_STORE
+	// Copy into VGPR for later TCP store.
+	for var sgpr_idx = 0; sgpr_idx < 12; sgpr_idx ++
+		v_writelane_b32	v2, s[sgpr_idx], ttmp13
+		s_add_u32	ttmp13, ttmp13, 0x1
+	end
+else
 	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
 	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
 	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*12
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
 end
-
+end
 
 function read_hwreg_from_mem(s, s_rsrc, s_mem_offset)
 	s_buffer_load_dword	s, s_rsrc, s_mem_offset glc:1
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 25a2703621a7..c36a352776a7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -580,6 +580,10 @@ static void kfd_cwsr_init(struct kfd_dev *kfd)
 			BUILD_BUG_ON(sizeof(cwsr_trap_gfx9_hex) > PAGE_SIZE);
 			kfd->cwsr_isa = cwsr_trap_gfx9_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx9_hex);
+		} else if (kfd->device_info->asic_family < CHIP_SIENNA_CICHLID) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_nv1x_hex) > PAGE_SIZE);
+			kfd->cwsr_isa = cwsr_trap_nv1x_hex;
+			kfd->cwsr_isa_size = sizeof(cwsr_trap_nv1x_hex);
 		} else {
 			BUILD_BUG_ON(sizeof(cwsr_trap_gfx10_hex) > PAGE_SIZE);
 			kfd->cwsr_isa = cwsr_trap_gfx10_hex;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 118/207] drm/amdkfd: Support newer assemblers in gfx10 trap handler

[Alex Deucher]

From: Jay Cornwall <jay.cornwall@amd.com>

The contents of macros are parsed by the assembler before conditions
have been tested. This causes assembly errors when using IP-specific
instructions in the IP-unified trap handler.

Add a preprocessing step to filter IP-specific code.

Also guard a Navi1x-specific instruction (no effect on Sienna_Cichlid).

Signed-off-by: Jay Cornwall <jay.cornwall@amd.com>
Reviewed-by: Yong Zhao <Yong.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../gpu/drm/amd/amdkfd/cwsr_trap_handler.h    | 597 +++++++++---------
 .../amd/amdkfd/cwsr_trap_handler_gfx10.asm    |  78 ++-
 2 files changed, 341 insertions(+), 334 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
index 0aeff2cc3efe..7e9e2d606682 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
@@ -1560,7 +1560,7 @@ static const uint32_t cwsr_trap_arcturus_hex[] = {
 };
 
 static const uint32_t cwsr_trap_gfx10_hex[] = {
-	0xbf820001, 0xbf8201c7,
+	0xbf820001, 0xbf8201c6,
 	0xb0804004, 0xb978f802,
 	0x8a788678, 0xb971f803,
 	0x876eff71, 0x00000400,
@@ -1585,297 +1585,227 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x876dff6d, 0x0000ffff,
 	0x80ec886c, 0x82ed806d,
 	0xbef60380, 0xb9f60283,
-	0xb972f816, 0xbeee037e,
-	0xbeef037f, 0xbefe0480,
-	0xbf900004, 0xbf8cc07f,
-	0xbef4037e, 0x8775ff7f,
-	0x0000ffff, 0x8875ff75,
-	0x00040000, 0xbef60380,
-	0xbef703ff, 0x10807fac,
-	0x8776ff7f, 0x08000000,
-	0x90768376, 0x88777677,
-	0x8776ff7f, 0x70000000,
-	0x90768176, 0x88777677,
-	0xbefb037c, 0xbefa0380,
-	0xb97302dc, 0x8f739973,
-	0x8873737f, 0xbefe03c1,
+	0xbeee037e, 0xbeef037f,
+	0xbefe0480, 0xbf900004,
+	0xbf8cc07f, 0xbef4037e,
+	0x8775ff7f, 0x0000ffff,
+	0x8875ff75, 0x00040000,
+	0xbef60380, 0xbef703ff,
+	0x10807fac, 0x8776ff7f,
+	0x08000000, 0x90768376,
+	0x88777677, 0x8776ff7f,
+	0x70000000, 0x90768176,
+	0x88777677, 0xbefb037c,
+	0xbefa0380, 0xb97302dc,
+	0x8f739973, 0x8873737f,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820002, 0xbeff03c1,
+	0xbf82000b, 0xbef603ff,
+	0x01000000, 0xe0704000,
+	0x7a5d0000, 0xe0704080,
+	0x7a5d0100, 0xe0704100,
+	0x7a5d0200, 0xe0704180,
+	0x7a5d0300, 0xbf82000a,
+	0xbef603ff, 0x01000000,
+	0xe0704000, 0x7a5d0000,
+	0xe0704100, 0x7a5d0100,
+	0xe0704200, 0x7a5d0200,
+	0xe0704300, 0x7a5d0300,
+	0xb97a2a05, 0x807a817a,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820002,
-	0xbeff03c1, 0xbf82000b,
+	0x8f7a897a, 0xbf820001,
+	0x8f7a8a7a, 0xb9761e06,
+	0x8f768a76, 0x807a767a,
+	0x807aff7a, 0x00000200,
 	0xbef603ff, 0x01000000,
-	0xe0704000, 0x7a5d0000,
-	0xe0704080, 0x7a5d0100,
-	0xe0704100, 0x7a5d0200,
-	0xe0704180, 0x7a5d0300,
-	0xbf82000a, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x7a5d0000, 0xe0704100,
-	0x7a5d0100, 0xe0704200,
-	0x7a5d0200, 0xe0704300,
-	0x7a5d0300, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0x807aff7a,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0x7e000280,
-	0x7e020280, 0x7e040280,
-	0xbefc0380, 0xd7610002,
-	0x0000f87b, 0x807c817c,
-	0xd7610002, 0x0000f86c,
+	0x7e000280, 0x7e020280,
+	0x7e040280, 0xbefc0380,
+	0xd7610002, 0x0000f87b,
 	0x807c817c, 0xd7610002,
-	0x0000f86d, 0x807c817c,
-	0xd7610002, 0x0000f86e,
+	0x0000f86c, 0x807c817c,
+	0xd7610002, 0x0000f86d,
 	0x807c817c, 0xd7610002,
-	0x0000f86f, 0x807c817c,
-	0xd7610002, 0x0000f878,
-	0x807c817c, 0xb971f803,
-	0xd7610002, 0x0000f871,
+	0x0000f86e, 0x807c817c,
+	0xd7610002, 0x0000f86f,
 	0x807c817c, 0xd7610002,
-	0x0000f872, 0x807c817c,
-	0xb97bf801, 0xd7610002,
-	0x0000f87b, 0x807c817c,
-	0xb97bf814, 0xd7610002,
-	0x0000f87b, 0x807c817c,
-	0xb97bf815, 0xd7610002,
-	0x0000f87b, 0x807c817c,
-	0xbeff0380, 0xe0704000,
-	0x7a5d0200, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0xbef603ff,
-	0x01000000, 0xbef90380,
-	0xbefc0380, 0xbf800000,
-	0xbe802f00, 0xbe822f02,
-	0xbe842f04, 0xbe862f06,
-	0xbe882f08, 0xbe8a2f0a,
-	0xbe8c2f0c, 0xbe8e2f0e,
-	0xd7610002, 0x0000f200,
+	0x0000f878, 0x807c817c,
+	0xb971f803, 0xd7610002,
+	0x0000f871, 0x807c817c,
+	0xd7610002, 0x0000f872,
+	0x807c817c, 0xb97bf801,
+	0xd7610002, 0x0000f87b,
+	0x807c817c, 0xb97bf814,
+	0xd7610002, 0x0000f87b,
+	0x807c817c, 0xb97bf815,
+	0xd7610002, 0x0000f87b,
+	0x807c817c, 0xbeff0380,
+	0xe0704000, 0x7a5d0200,
+	0xb97a2a05, 0x807a817a,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0x8f7a897a, 0xbf820001,
+	0x8f7a8a7a, 0xb9761e06,
+	0x8f768a76, 0x807a767a,
+	0xbef603ff, 0x01000000,
+	0xbef90380, 0xbefc0380,
+	0xbf800000, 0xbe802f00,
+	0xbe822f02, 0xbe842f04,
+	0xbe862f06, 0xbe882f08,
+	0xbe8a2f0a, 0xbe8c2f0c,
+	0xbe8e2f0e, 0xd7610002,
+	0x0000f200, 0x80798179,
+	0xd7610002, 0x0000f201,
 	0x80798179, 0xd7610002,
-	0x0000f201, 0x80798179,
-	0xd7610002, 0x0000f202,
+	0x0000f202, 0x80798179,
+	0xd7610002, 0x0000f203,
 	0x80798179, 0xd7610002,
-	0x0000f203, 0x80798179,
-	0xd7610002, 0x0000f204,
+	0x0000f204, 0x80798179,
+	0xd7610002, 0x0000f205,
 	0x80798179, 0xd7610002,
-	0x0000f205, 0x80798179,
-	0xd7610002, 0x0000f206,
+	0x0000f206, 0x80798179,
+	0xd7610002, 0x0000f207,
 	0x80798179, 0xd7610002,
-	0x0000f207, 0x80798179,
-	0xd7610002, 0x0000f208,
+	0x0000f208, 0x80798179,
+	0xd7610002, 0x0000f209,
 	0x80798179, 0xd7610002,
-	0x0000f209, 0x80798179,
-	0xd7610002, 0x0000f20a,
+	0x0000f20a, 0x80798179,
+	0xd7610002, 0x0000f20b,
 	0x80798179, 0xd7610002,
-	0x0000f20b, 0x80798179,
-	0xd7610002, 0x0000f20c,
+	0x0000f20c, 0x80798179,
+	0xd7610002, 0x0000f20d,
 	0x80798179, 0xd7610002,
-	0x0000f20d, 0x80798179,
-	0xd7610002, 0x0000f20e,
+	0x0000f20e, 0x80798179,
+	0xd7610002, 0x0000f20f,
+	0x80798179, 0xbf06a079,
+	0xbf840006, 0xe0704000,
+	0x7a5d0200, 0x807aff7a,
+	0x00000080, 0xbef90380,
+	0x7e040280, 0x807c907c,
+	0xbf0aff7c, 0x00000060,
+	0xbf85ffbc, 0xbe802f00,
+	0xbe822f02, 0xbe842f04,
+	0xbe862f06, 0xbe882f08,
+	0xbe8a2f0a, 0xd7610002,
+	0x0000f200, 0x80798179,
+	0xd7610002, 0x0000f201,
 	0x80798179, 0xd7610002,
-	0x0000f20f, 0x80798179,
-	0xbf06a079, 0xbf840006,
-	0xe0704000, 0x7a5d0200,
-	0x807aff7a, 0x00000080,
-	0xbef90380, 0x7e040280,
-	0x807c907c, 0xbf0aff7c,
-	0x00000060, 0xbf85ffbc,
-	0xbe802f00, 0xbe822f02,
-	0xbe842f04, 0xbe862f06,
-	0xbe882f08, 0xbe8a2f0a,
-	0xd7610002, 0x0000f200,
-	0x80798179, 0xd7610002,
-	0x0000f201, 0x80798179,
-	0xd7610002, 0x0000f202,
+	0x0000f202, 0x80798179,
+	0xd7610002, 0x0000f203,
 	0x80798179, 0xd7610002,
-	0x0000f203, 0x80798179,
-	0xd7610002, 0x0000f204,
+	0x0000f204, 0x80798179,
+	0xd7610002, 0x0000f205,
 	0x80798179, 0xd7610002,
-	0x0000f205, 0x80798179,
-	0xd7610002, 0x0000f206,
+	0x0000f206, 0x80798179,
+	0xd7610002, 0x0000f207,
 	0x80798179, 0xd7610002,
-	0x0000f207, 0x80798179,
-	0xd7610002, 0x0000f208,
+	0x0000f208, 0x80798179,
+	0xd7610002, 0x0000f209,
 	0x80798179, 0xd7610002,
-	0x0000f209, 0x80798179,
-	0xd7610002, 0x0000f20a,
-	0x80798179, 0xd7610002,
-	0x0000f20b, 0x80798179,
-	0xe0704000, 0x7a5d0200,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0xbeff0380,
-	0xbf820001, 0xbeff03c1,
-	0xb9714306, 0x8771c171,
-	0xbf840046, 0xbf8a0000,
-	0x8776ff73, 0x04000000,
-	0xbf840042, 0x8f718671,
-	0x8f718271, 0xbef60371,
-	0xb97a2a05, 0x807a817a,
+	0x0000f20a, 0x80798179,
+	0xd7610002, 0x0000f20b,
+	0x80798179, 0xe0704000,
+	0x7a5d0200, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0x8f7a897a, 0xbf820001,
-	0x8f7a8a7a, 0xb9761e06,
-	0x8f768a76, 0x807a767a,
-	0x807aff7a, 0x00000200,
-	0x807aff7a, 0x00000080,
-	0xbef603ff, 0x01000000,
-	0xd7650000, 0x000100c1,
-	0xd7660000, 0x000200c1,
-	0x16000084, 0x907c9973,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb9714306,
+	0x8771c171, 0xbf840046,
+	0xbf8a0000, 0x8776ff73,
+	0x04000000, 0xbf840042,
+	0x8f718671, 0x8f718271,
+	0xbef60371, 0xb97a2a05,
+	0x807a817a, 0x907c9973,
 	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850012,
-	0xbe8303ff, 0x00000080,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0xb9761e06, 0x8f768a76,
+	0x807a767a, 0x807aff7a,
+	0x00000200, 0x807aff7a,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0xd7650000,
+	0x000100c1, 0xd7660000,
+	0x000200c1, 0x16000084,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850012, 0xbe8303ff,
+	0x00000080, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x7a5d0100,
-	0x807c037c, 0x807a037a,
-	0xd5250000, 0x0001ff00,
-	0x00000080, 0xbf0a717c,
-	0xbf85fff4, 0xbf820011,
-	0xbe8303ff, 0x00000100,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x7a5d0100, 0x807c037c,
+	0x807a037a, 0xd5250000,
+	0x0001ff00, 0x00000080,
+	0xbf0a717c, 0xbf85fff4,
+	0xbf820011, 0xbe8303ff,
+	0x00000100, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x7a5d0100,
-	0x807c037c, 0x807a037a,
-	0xd5250000, 0x0001ff00,
-	0x00000100, 0xbf0a717c,
-	0xbf85fff4, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850004,
-	0xbefa03ff, 0x00000200,
-	0xbeff0380, 0xbf820003,
-	0xbefa03ff, 0x00000400,
-	0xbeff03c1, 0xb9712a05,
-	0x80718171, 0x8f718271,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850017,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x7a5d0100, 0x807c037c,
+	0x807a037a, 0xd5250000,
+	0x0001ff00, 0x00000100,
+	0xbf0a717c, 0xbf85fff4,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850004, 0xbefa03ff,
+	0x00000200, 0xbeff0380,
+	0xbf820003, 0xbefa03ff,
+	0x00000400, 0xbeff03c1,
+	0xb9712a05, 0x80718171,
+	0x8f718271, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850017, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a717c, 0xbf840037,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x7a5d0000,
+	0xe0704080, 0x7a5d0100,
+	0xe0704100, 0x7a5d0200,
+	0xe0704180, 0x7a5d0300,
+	0x807c847c, 0x807aff7a,
+	0x00000200, 0xbf0a717c,
+	0xbf85ffef, 0xbf820025,
 	0xbef603ff, 0x01000000,
 	0xbefc0384, 0xbf0a717c,
-	0xbf840037, 0x7e008700,
+	0xbf840020, 0x7e008700,
 	0x7e028701, 0x7e048702,
 	0x7e068703, 0xe0704000,
-	0x7a5d0000, 0xe0704080,
-	0x7a5d0100, 0xe0704100,
-	0x7a5d0200, 0xe0704180,
+	0x7a5d0000, 0xe0704100,
+	0x7a5d0100, 0xe0704200,
+	0x7a5d0200, 0xe0704300,
 	0x7a5d0300, 0x807c847c,
-	0x807aff7a, 0x00000200,
+	0x807aff7a, 0x00000400,
 	0xbf0a717c, 0xbf85ffef,
-	0xbf820025, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a717c, 0xbf840020,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
+	0xb9711e06, 0x8771c171,
+	0xbf84000c, 0x8f718371,
+	0x80717c71, 0xbefe03c1,
+	0xbeff0380, 0x7e008700,
 	0xe0704000, 0x7a5d0000,
-	0xe0704100, 0x7a5d0100,
-	0xe0704200, 0x7a5d0200,
-	0xe0704300, 0x7a5d0300,
-	0x807c847c, 0x807aff7a,
-	0x00000400, 0xbf0a717c,
-	0xbf85ffef, 0xb9711e06,
-	0x8771c171, 0xbf84000c,
-	0x8f718371, 0x80717c71,
-	0xbefe03c1, 0xbeff0380,
-	0x7e008700, 0xe0704000,
-	0x7a5d0000, 0x807c817c,
-	0x807aff7a, 0x00000080,
-	0xbf0a717c, 0xbf85fff8,
-	0xbf82012e, 0xbef4037e,
-	0x8775ff7f, 0x0000ffff,
-	0x8875ff75, 0x00040000,
-	0xbef60380, 0xbef703ff,
-	0x10807fac, 0x8772ff7f,
-	0x08000000, 0x90728372,
-	0x88777277, 0x8772ff7f,
-	0x70000000, 0x90728172,
-	0x88777277, 0xb97302dc,
-	0x8f739973, 0x8873737f,
-	0x8772ff7f, 0x04000000,
-	0xbf840036, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb96f4306,
-	0x876fc16f, 0xbf84002b,
-	0x8f6f866f, 0x8f6f826f,
-	0xbef6036f, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0x8078ff78,
-	0x00000080, 0xbef603ff,
-	0x01000000, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850009,
-	0xe0310000, 0x781d0000,
-	0x807cff7c, 0x00000080,
-	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff8,
-	0xbf820008, 0xe0310000,
-	0x781d0000, 0x807cff7c,
-	0x00000100, 0x8078ff78,
-	0x00000100, 0xbf0a6f7c,
-	0xbf85fff8, 0xbef80380,
+	0x807c817c, 0x807aff7a,
+	0x00000080, 0xbf0a717c,
+	0xbf85fff8, 0xbf82012d,
+	0xbef4037e, 0x8775ff7f,
+	0x0000ffff, 0x8875ff75,
+	0x00040000, 0xbef60380,
+	0xbef703ff, 0x10807fac,
+	0x8772ff7f, 0x08000000,
+	0x90728372, 0x88777277,
+	0x8772ff7f, 0x70000000,
+	0x90728172, 0x88777277,
+	0xb97302dc, 0x8f739973,
+	0x8873737f, 0x8772ff7f,
+	0x04000000, 0xbf840036,
 	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
-	0xb96f2a05, 0x806f816f,
-	0x8f6f826f, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850021, 0xbef603ff,
-	0x01000000, 0xbef20378,
-	0x8078ff78, 0x00000200,
-	0xbefc0384, 0xe0304000,
-	0x785d0000, 0xe0304080,
-	0x785d0100, 0xe0304100,
-	0x785d0200, 0xe0304180,
-	0x785d0300, 0xbf8c3f70,
-	0x7e008500, 0x7e028501,
-	0x7e048502, 0x7e068503,
-	0x807c847c, 0x8078ff78,
-	0x00000200, 0xbf0a6f7c,
-	0xbf85ffee, 0xe0304000,
-	0x725d0000, 0xe0304080,
-	0x725d0100, 0xe0304100,
-	0x725d0200, 0xe0304180,
-	0x725d0300, 0xbf820032,
-	0xbef603ff, 0x01000000,
-	0xbef20378, 0x8078ff78,
-	0x00000400, 0xbefc0384,
-	0xe0304000, 0x785d0000,
-	0xe0304100, 0x785d0100,
-	0xe0304200, 0x785d0200,
-	0xe0304300, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
-	0x7e028501, 0x7e048502,
-	0x7e068503, 0x807c847c,
-	0x8078ff78, 0x00000400,
-	0xbf0a6f7c, 0xbf85ffee,
-	0xb96f1e06, 0x876fc16f,
-	0xbf84000e, 0x8f6f836f,
-	0x806f7c6f, 0xbefe03c1,
-	0xbeff0380, 0xe0304000,
-	0x785d0000, 0xbf8c3f70,
-	0x7e008500, 0x807c817c,
-	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff7,
-	0xbeff03c1, 0xe0304000,
-	0x725d0000, 0xe0304100,
-	0x725d0100, 0xe0304200,
-	0x725d0200, 0xe0304300,
-	0x725d0300, 0xbf8c3f70,
+	0xb96f4306, 0x876fc16f,
+	0xbf84002b, 0x8f6f866f,
+	0x8f6f826f, 0xbef6036f,
 	0xb9782a05, 0x80788178,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
@@ -1883,58 +1813,127 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x8f788a78, 0xb9721e06,
 	0x8f728a72, 0x80787278,
 	0x8078ff78, 0x00000200,
-	0x80f8ff78, 0x00000050,
+	0x8078ff78, 0x00000080,
 	0xbef603ff, 0x01000000,
-	0xbefc03ff, 0x0000006c,
-	0x80f89078, 0xf429003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc847c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0x80f8a078, 0xf42d003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc887c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0x80f8c078, 0xf431003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc907c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0xbe883108, 0xbe8a310a,
-	0xbe8c310c, 0xbe8e310e,
-	0xbf06807c, 0xbf84fff0,
-	0xb9782a05, 0x80788178,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850009, 0xe0310000,
+	0x781d0000, 0x807cff7c,
+	0x00000080, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff8, 0xbf820008,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000100,
+	0x8078ff78, 0x00000100,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbef80380, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb9721e06,
-	0x8f728a72, 0x80787278,
-	0x8078ff78, 0x00000200,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb96f2a05,
+	0x806f816f, 0x8f6f826f,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850021,
 	0xbef603ff, 0x01000000,
-	0xf4211bfa, 0xf0000000,
-	0x80788478, 0xf4211b3a,
+	0xbef20378, 0x8078ff78,
+	0x00000200, 0xbefc0384,
+	0xe0304000, 0x785d0000,
+	0xe0304080, 0x785d0100,
+	0xe0304100, 0x785d0200,
+	0xe0304180, 0x785d0300,
+	0xbf8c3f70, 0x7e008500,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
+	0x8078ff78, 0x00000200,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xe0304000, 0x725d0000,
+	0xe0304080, 0x725d0100,
+	0xe0304100, 0x725d0200,
+	0xe0304180, 0x725d0300,
+	0xbf820032, 0xbef603ff,
+	0x01000000, 0xbef20378,
+	0x8078ff78, 0x00000400,
+	0xbefc0384, 0xe0304000,
+	0x785d0000, 0xe0304100,
+	0x785d0100, 0xe0304200,
+	0x785d0200, 0xe0304300,
+	0x785d0300, 0xbf8c3f70,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807c847c, 0x8078ff78,
+	0x00000400, 0xbf0a6f7c,
+	0xbf85ffee, 0xb96f1e06,
+	0x876fc16f, 0xbf84000e,
+	0x8f6f836f, 0x806f7c6f,
+	0xbefe03c1, 0xbeff0380,
+	0xe0304000, 0x785d0000,
+	0xbf8c3f70, 0x7e008500,
+	0x807c817c, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff7, 0xbeff03c1,
+	0xe0304000, 0x725d0000,
+	0xe0304100, 0x725d0100,
+	0xe0304200, 0x725d0200,
+	0xe0304300, 0x725d0300,
+	0xbf8c3f70, 0xb9782a05,
+	0x80788178, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb9721e06, 0x8f728a72,
+	0x80787278, 0x8078ff78,
+	0x00000200, 0x80f8ff78,
+	0x00000050, 0xbef603ff,
+	0x01000000, 0xbefc03ff,
+	0x0000006c, 0x80f89078,
+	0xf429003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc847c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0x80f8a078,
+	0xf42d003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc887c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0x80f8c078,
+	0xf431003a, 0xf0000000,
+	0xbf8cc07f, 0x80fc907c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0xbe883108,
+	0xbe8a310a, 0xbe8c310c,
+	0xbe8e310e, 0xbf06807c,
+	0xbf84fff0, 0xb9782a05,
+	0x80788178, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb9721e06, 0x8f728a72,
+	0x80787278, 0x8078ff78,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0xf4211bfa,
 	0xf0000000, 0x80788478,
-	0xf4211b7a, 0xf0000000,
-	0x80788478, 0xf4211eba,
+	0xf4211b3a, 0xf0000000,
+	0x80788478, 0xf4211b7a,
 	0xf0000000, 0x80788478,
-	0xf4211efa, 0xf0000000,
-	0x80788478, 0xf4211c3a,
+	0xf4211eba, 0xf0000000,
+	0x80788478, 0xf4211efa,
 	0xf0000000, 0x80788478,
-	0xf4211c7a, 0xf0000000,
-	0x80788478, 0xf4211e7a,
+	0xf4211c3a, 0xf0000000,
+	0x80788478, 0xf4211c7a,
 	0xf0000000, 0x80788478,
-	0xf4211cfa, 0xf0000000,
-	0x80788478, 0xf4211bba,
+	0xf4211e7a, 0xf0000000,
+	0x80788478, 0xf4211cfa,
 	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef815, 0xbef2036d,
-	0x876dff72, 0x0000ffff,
-	0xbefc036f, 0xbefe037a,
-	0xbeff037b, 0x876f71ff,
-	0x000003ff, 0xb9ef4803,
-	0xb9f9f816, 0x876f71ff,
+	0xb9eef814, 0xf4211bba,
+	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef815,
+	0xbef2036d, 0x876dff72,
+	0x0000ffff, 0xbefc036f,
+	0xbefe037a, 0xbeff037b,
+	0x876f71ff, 0x000003ff,
+	0xb9ef4803, 0x876f71ff,
 	0xfffff800, 0x906f8b6f,
 	0xb9efa2c3, 0xb9f3f801,
 	0x87fe7e7e, 0x87ea6a6a,
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
index ce70bba8ac06..05dd29035772 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
@@ -23,13 +23,15 @@
 /* To compile this assembly code:
  *
  * Navi1x:
- *   PROJECT=gfx10 ./sp3 ASIC_TARGET_NAVI1X=1 cwsr_trap_handler_gfx10.asm -hex tmp.hex
+ *   cpp -DASIC_TARGET_NAVI1X=1 cwsr_trap_handler_gfx10.asm -P -o nv1x.sp3
+ *   sp3-nv1x nv1x.sp3 -hex nv1x.hex
  *
  * Others:
- *   PROJECT=gfx10 ./sp3 ASIC_TARGET_NAVI1X=0 cwsr_trap_handler_gfx10.asm -hex tmp.hex
+ *   cpp -DASIC_TARGET_NAVI1X=0 cwsr_trap_handler_gfx10.asm -P -o gfx10.sp3
+ *   sp3-gfx10 gfx10.sp3 -hex gfx10.hex
  */
 
-var NO_SQC_STORE				= !ASIC_TARGET_NAVI1X
+#define NO_SQC_STORE !ASIC_TARGET_NAVI1X
 
 var SQ_WAVE_STATUS_INST_ATC_SHIFT		= 23
 var SQ_WAVE_STATUS_INST_ATC_MASK		= 0x00800000
@@ -182,7 +184,7 @@ L_SKIP_RESTORE:
 
 L_FETCH_2ND_TRAP:
 
-if ASIC_TARGET_NAVI1X
+#if ASIC_TARGET_NAVI1X
 	// Preserve and clear scalar XNACK state before issuing scalar loads.
 	// Save IB_STS.REPLAY_W64H[25], RCNT[21:16], FIRST_REPLAY[15] into
 	// unused space ttmp11[31:24].
@@ -196,7 +198,7 @@ if ASIC_TARGET_NAVI1X
 	s_or_b32	ttmp11, ttmp11, ttmp3
 	s_andn2_b32	ttmp2, ttmp2, (SQ_WAVE_IB_STS_REPLAY_W64H_MASK | SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK)
 	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
-end
+#endif
 
 	// Read second-level TBA/TMA from first-level TMA and jump if available.
 	// ttmp[2:5] and ttmp12 can be used (others hold SPI-initialized debug data)
@@ -221,7 +223,7 @@ L_NO_NEXT_TRAP:
 L_EXCP_CASE:
 	s_and_b32	ttmp1, ttmp1, 0xFFFF
 
-if ASIC_TARGET_NAVI1X
+#if ASIC_TARGET_NAVI1X
 	// Restore SQ_WAVE_IB_STS.
 	s_lshr_b32	ttmp2, ttmp11, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
 	s_and_b32	ttmp3, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
@@ -229,7 +231,7 @@ if ASIC_TARGET_NAVI1X
 	s_and_b32	ttmp2, ttmp2, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
 	s_or_b32	ttmp2, ttmp2, ttmp3
 	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
-end
+#endif
 
 	// Restore SQ_WAVE_STATUS.
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
@@ -255,9 +257,9 @@ L_NO_PC_REWIND:
 	s_mov_b32	s_save_tmp, 0
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_SAVECTX_SHIFT, 1), s_save_tmp	//clear saveCtx bit
 
+#if ASIC_TARGET_NAVI1X
 	s_getreg_b32	s_save_xnack_mask, hwreg(HW_REG_SHADER_XNACK_MASK)
 
-if ASIC_TARGET_NAVI1X
 	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_RCNT_SHIFT, SQ_WAVE_IB_STS_RCNT_SIZE)
 	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_RCNT_SHIFT
 	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
@@ -271,7 +273,7 @@ if ASIC_TARGET_NAVI1X
 	s_and_b32	s_save_tmp, s_save_tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG
 
 	s_setreg_b32	hwreg(HW_REG_IB_STS), s_save_tmp
-end
+#endif
 
 	/* inform SPI the readiness and wait for SPI's go signal */
 	s_mov_b32	s_save_exec_lo, exec_lo					//save EXEC and use EXEC for the go signal from SPI
@@ -280,16 +282,16 @@ end
 
 	s_sendmsg	sendmsg(MSG_SAVEWAVE)					//send SPI a message and wait for SPI's write to EXEC
 
-if ASIC_TARGET_NAVI1X
+#if ASIC_TARGET_NAVI1X
 L_SLEEP:
 	// sleep 1 (64clk) is not enough for 8 waves per SIMD, which will cause
 	// SQ hang, since the 7,8th wave could not get arbit to exec inst, while
 	// other waves are stuck into the sleep-loop and waiting for wrexec!=0
 	s_sleep		0x2
 	s_cbranch_execz	L_SLEEP
-else
+#else
 	s_waitcnt	lgkmcnt(0)
-end
+#endif
 
 	/* setup Resource Contants */
 	s_mov_b32	s_save_buf_rsrc0, s_save_spi_init_lo			//base_addr_lo
@@ -355,12 +357,12 @@ L_SAVE_HWREG:
 
 	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
 
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	v_mov_b32	v0, 0x0							//Offset[31:0] from buffer resource
 	v_mov_b32	v1, 0x0							//Offset[63:32] from buffer resource
 	v_mov_b32	v2, 0x0							//Set of SGPRs for TCP store
 	s_mov_b32	m0, 0x0							//Next lane of v2 to write to
-end
+#endif
 
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_pc_lo, s_save_buf_rsrc0, s_save_mem_offset)
@@ -371,6 +373,8 @@ end
 
 	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
 	write_hwreg_to_mem(s_save_trapsts, s_save_buf_rsrc0, s_save_mem_offset)
+
+	// Not used on Sienna_Cichlid but keep layout same for debugger.
 	write_hwreg_to_mem(s_save_xnack_mask, s_save_buf_rsrc0, s_save_mem_offset)
 
 	s_getreg_b32	s_save_m0, hwreg(HW_REG_MODE)
@@ -382,11 +386,11 @@ end
 	s_getreg_b32	s_save_m0, hwreg(HW_REG_SHADER_FLAT_SCRATCH_HI)
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	// Write HWREG/SGPRs with 32 VGPR lanes, wave32 is common case.
 	s_mov_b32	exec_hi, 0x0
 	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-end
+#endif
 
 	/* save SGPRs */
 	// Save SGPR before LDS save, then the s0 to s4 can be used during LDS save...
@@ -397,14 +401,14 @@ end
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, s_save_tmp
 	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
 
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	s_mov_b32	ttmp13, 0x0						//next VGPR lane to copy SGPR into
-else
+#else
 	// backup s_save_buf_rsrc0,1 to s_save_pc_lo/hi, since write_16sgpr_to_mem function will change the rsrc0
 	s_mov_b32	s_save_xnack_mask, s_save_buf_rsrc0
 	s_add_u32	s_save_buf_rsrc0, s_save_buf_rsrc0, s_save_mem_offset
 	s_addc_u32	s_save_buf_rsrc1, s_save_buf_rsrc1, 0
-end
+#endif
 
 	s_mov_b32	m0, 0x0							//SGPR initial index value =0
 	s_nop		0x0							//Manually inserted wait states
@@ -421,7 +425,7 @@ L_SAVE_SGPR_LOOP:
 
 	write_16sgpr_to_mem(s0, s_save_buf_rsrc0, s_save_mem_offset)
 
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	s_cmp_eq_u32	ttmp13, 0x20						//have 32 VGPR lanes filled?
 	s_cbranch_scc0	L_SAVE_SGPR_SKIP_TCP_STORE
 
@@ -430,7 +434,7 @@ if NO_SQC_STORE
 	s_mov_b32	ttmp13, 0x0
 	v_mov_b32	v2, 0x0
 L_SAVE_SGPR_SKIP_TCP_STORE:
-end
+#endif
 
 	s_add_u32	m0, m0, 16						//next sgpr index
 	s_cmp_lt_u32	m0, 96							//scc = (m0 < first 96 SGPR) ? 1 : 0
@@ -445,12 +449,12 @@ end
 	s_movrels_b64	s10, s10						//s10 = s[10+m0], s11 = s[11+m0]
 	write_12sgpr_to_mem(s0, s_save_buf_rsrc0, s_save_mem_offset)
 
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-else
+#else
 	// restore s_save_buf_rsrc0,1
 	s_mov_b32	s_save_buf_rsrc0, s_save_xnack_mask
-end
+#endif
 
 	/* save LDS */
 
@@ -899,13 +903,17 @@ L_RESTORE_HWREG:
 
 	s_and_b32	s_restore_m0, SQ_WAVE_TRAPSTS_PRE_SAVECTX_MASK, s_restore_trapsts
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_PRE_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_PRE_SAVECTX_SIZE), s_restore_m0
+
+#if ASIC_TARGET_NAVI1X
 	s_setreg_b32	hwreg(HW_REG_SHADER_XNACK_MASK), s_restore_xnack_mask
+#endif
+
 	s_and_b32	s_restore_m0, SQ_WAVE_TRAPSTS_POST_SAVECTX_MASK, s_restore_trapsts
 	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE), s_restore_m0
 	s_setreg_b32	hwreg(HW_REG_MODE), s_restore_mode
 
-if ASIC_TARGET_NAVI1X
+#if ASIC_TARGET_NAVI1X
 	s_and_b32	s_restore_m0, s_restore_tmp, S_SAVE_PC_HI_RCNT_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_RCNT_SHIFT
 	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_RCNT_SHIFT
@@ -923,7 +931,7 @@ if ASIC_TARGET_NAVI1X
 	s_and_b32	s_restore_m0, s_restore_status, SQ_WAVE_STATUS_INST_ATC_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_STATUS_INST_ATC_SHIFT
 	s_setreg_b32 	hwreg(HW_REG_IB_STS), s_restore_mode
-end
+#endif
 
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
@@ -938,51 +946,51 @@ L_END_PGM:
 end
 
 function write_hwreg_to_mem(s, s_rsrc, s_mem_offset)
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	// Copy into VGPR for later TCP store.
 	v_writelane_b32	v2, s, m0
 	s_add_u32	m0, m0, 0x1
-else
+#else
 	s_mov_b32	exec_lo, m0
 	s_mov_b32	m0, s_mem_offset
 	s_buffer_store_dword	s, s_rsrc, m0 glc:1
 	s_add_u32	s_mem_offset, s_mem_offset, 4
 	s_mov_b32	m0, exec_lo
-end
+#endif
 end
 
 
 function write_16sgpr_to_mem(s, s_rsrc, s_mem_offset)
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	// Copy into VGPR for later TCP store.
 	for var sgpr_idx = 0; sgpr_idx < 16; sgpr_idx ++
 		v_writelane_b32	v2, s[sgpr_idx], ttmp13
 		s_add_u32	ttmp13, ttmp13, 0x1
 	end
-else
+#else
 	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
 	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
 	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
 	s_buffer_store_dwordx4	s[12], s_rsrc, 48 glc:1
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*16
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
-end
+#endif
 end
 
 function write_12sgpr_to_mem(s, s_rsrc, s_mem_offset)
-if NO_SQC_STORE
+#if NO_SQC_STORE
 	// Copy into VGPR for later TCP store.
 	for var sgpr_idx = 0; sgpr_idx < 12; sgpr_idx ++
 		v_writelane_b32	v2, s[sgpr_idx], ttmp13
 		s_add_u32	ttmp13, ttmp13, 0x1
 	end
-else
+#else
 	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
 	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
 	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*12
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
-end
+#endif
 end
 
 function read_hwreg_from_mem(s, s_rsrc, s_mem_offset)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 119/207] drm/amdkfd: Support debugger in Navi1x trap handler

[Alex Deucher]

From: Jay Cornwall <jay.cornwall@amd.com>

- Preserve scalar GPRs ttmp[4:11] and ttmp13
- Add single step exception during context save workaround
- Remove incorrect PC adjustment during context save

Signed-off-by: Jay Cornwall <jay.cornwall@amd.com>
Reviewed-by: Yong Zhao <Yong.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../gpu/drm/amd/amdkfd/cwsr_trap_handler.h    | 1368 +++++++++--------
 .../amd/amdkfd/cwsr_trap_handler_gfx10.asm    |  141 +-
 2 files changed, 789 insertions(+), 720 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
index 7e9e2d606682..577d901fdb63 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
@@ -680,389 +680,404 @@ static const uint32_t cwsr_trap_gfx9_hex[] = {
 };
 
 static const uint32_t cwsr_trap_nv1x_hex[] = {
-	0xbf820001, 0xbf8201bc,
+	0xbf820001, 0xbf8201cd,
 	0xb0804004, 0xb978f802,
-	0x8a788678, 0xb971f803,
-	0x876eff71, 0x00000400,
-	0xbf850033, 0x876eff71,
-	0x00000100, 0xbf840002,
-	0x8878ff78, 0x00002000,
-	0x8a77ff77, 0xff000000,
-	0xb96ef807, 0x876fff6e,
-	0x02000000, 0x8f6f866f,
-	0x88776f77, 0x876fff6e,
-	0x003f8000, 0x8f6f896f,
-	0x88776f77, 0x8a6eff6e,
-	0x023f8000, 0xb9eef807,
-	0xb97af812, 0xb97bf813,
-	0x8ffa887a, 0xf4051bbd,
-	0xfa000000, 0xbf8cc07f,
-	0xf4051ebd, 0xfa000008,
-	0xbf8cc07f, 0x87ee6e6e,
-	0xbf840001, 0xbe80206e,
-	0xb971f803, 0x8771ff71,
-	0x000001ff, 0xbf850002,
-	0x806c846c, 0x826d806d,
-	0x876dff6d, 0x0000ffff,
-	0x906e8977, 0x876fff6e,
-	0x003f8000, 0x906e8677,
-	0x876eff6e, 0x02000000,
-	0x886e6f6e, 0xb9eef807,
-	0x87fe7e7e, 0x87ea6a6a,
-	0xb9f8f802, 0xbe80226c,
-	0xb971f803, 0x8771ff71,
-	0x00000100, 0xbf840006,
-	0xbef60380, 0xb9f60203,
-	0x876dff6d, 0x0000ffff,
-	0x80ec886c, 0x82ed806d,
-	0xbef60380, 0xb9f60283,
-	0xb972f816, 0xb9762c07,
-	0x8f769a76, 0x886d766d,
-	0xb97603c7, 0x8f769976,
-	0x886d766d, 0xb9760647,
-	0x8f769876, 0x886d766d,
-	0xb976f807, 0x8776ff76,
-	0x00007fff, 0xb9f6f807,
+	0x8a788678, 0xb96ef801,
+	0x876eff6e, 0x00000800,
+	0xbf840003, 0x876eff78,
+	0x00002000, 0xbf840009,
+	0xb97bf803, 0x876eff7b,
+	0x00000400, 0xbf850033,
+	0x876eff7b, 0x00000100,
+	0xbf840002, 0x8878ff78,
+	0x00002000, 0x8a77ff77,
+	0xff000000, 0xb96ef807,
+	0x876fff6e, 0x02000000,
+	0x8f6f866f, 0x88776f77,
+	0x876fff6e, 0x003f8000,
+	0x8f6f896f, 0x88776f77,
+	0x8a6eff6e, 0x023f8000,
+	0xb9eef807, 0xb97af812,
+	0xb97bf813, 0x8ffa887a,
+	0xf4051bbd, 0xfa000000,
+	0xbf8cc07f, 0xf4051ebd,
+	0xfa000008, 0xbf8cc07f,
+	0x87ee6e6e, 0xbf840001,
+	0xbe80206e, 0xb97bf803,
+	0x877bff7b, 0x000001ff,
+	0xbf850002, 0x806c846c,
+	0x826d806d, 0x876dff6d,
+	0x0000ffff, 0x906e8977,
+	0x876fff6e, 0x003f8000,
+	0x906e8677, 0x876eff6e,
+	0x02000000, 0x886e6f6e,
+	0xb9eef807, 0x87fe7e7e,
+	0x87ea6a6a, 0xb9f8f802,
+	0xbe80226c, 0x876dff6d,
+	0x0000ffff, 0xbefa0380,
+	0xb9fa0283, 0xb97a2c07,
+	0x8f7a9a7a, 0x886d7a6d,
+	0xb97a03c7, 0x8f7a997a,
+	0x886d7a6d, 0xb97a0647,
+	0x8f7a987a, 0x886d7a6d,
+	0xb97af807, 0x877aff7a,
+	0x00007fff, 0xb9faf807,
 	0xbeee037e, 0xbeef037f,
 	0xbefe0480, 0xbf900004,
 	0xbf8e0002, 0xbf88fffe,
+	0xb97b02dc, 0x8f7b997b,
+	0x887b7b7f, 0xb97a2a05,
+	0x807a817a, 0xbf0d997b,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0x877bff7f, 0x0000ffff,
+	0x807aff7a, 0x00000200,
+	0x807a7e7a, 0x827b807b,
+	0xf4491c3d, 0xfa000050,
+	0xf4491d3d, 0xfa000060,
+	0xf4411e7d, 0xfa000074,
 	0xbef4037e, 0x8775ff7f,
 	0x0000ffff, 0x8875ff75,
 	0x00040000, 0xbef60380,
 	0xbef703ff, 0x10807fac,
-	0x8776ff7f, 0x08000000,
-	0x90768376, 0x88777677,
-	0x8776ff7f, 0x70000000,
-	0x90768176, 0x88777677,
-	0xbefb037c, 0xbefa0380,
+	0x877aff7f, 0x08000000,
+	0x907a837a, 0x88777a77,
+	0x877aff7f, 0x70000000,
+	0x907a817a, 0x88777a77,
+	0xbef1037c, 0xbef00380,
 	0xb97302dc, 0x8f739973,
-	0x8873737f, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820002,
-	0xbeff03c1, 0xbf82000b,
-	0xbef603ff, 0x01000000,
-	0xe0704000, 0x7a5d0000,
-	0xe0704080, 0x7a5d0100,
-	0xe0704100, 0x7a5d0200,
-	0xe0704180, 0x7a5d0300,
-	0xbf82000a, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x7a5d0000, 0xe0704100,
-	0x7a5d0100, 0xe0704200,
-	0x7a5d0200, 0xe0704300,
-	0x7a5d0300, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
+	0x8873737f, 0xb97bf816,
+	0xba80f816, 0x00000000,
+	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0x807aff7a,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0xbefe037c,
-	0xbefc037a, 0xf4611efa,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xbefe037c,
-	0xbefc037a, 0xf4611b3a,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xbefe037c,
-	0xbefc037a, 0xf4611b7a,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xbefe037c,
-	0xbefc037a, 0xf4611bba,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xbefe037c,
-	0xbefc037a, 0xf4611bfa,
-	0xf8000000, 0x807a847a,
+	0xbf850002, 0xbeff0380,
+	0xbf820002, 0xbeff03c1,
+	0xbf82000b, 0xbef603ff,
+	0x01000000, 0xe0704000,
+	0x705d0000, 0xe0704080,
+	0x705d0100, 0xe0704100,
+	0x705d0200, 0xe0704180,
+	0x705d0300, 0xbf82000a,
+	0xbef603ff, 0x01000000,
+	0xe0704000, 0x705d0000,
+	0xe0704100, 0x705d0100,
+	0xe0704200, 0x705d0200,
+	0xe0704300, 0x705d0300,
+	0xb9702a05, 0x80708170,
+	0xbf0d9973, 0xbf850002,
+	0x8f708970, 0xbf820001,
+	0x8f708a70, 0xb97a1e06,
+	0x8f7a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0xbef603ff, 0x01000000,
+	0xbefe037c, 0xbefc0370,
+	0xf4611c7a, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xbefe037c, 0xbefc0370,
+	0xf4611b3a, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xbefe037c, 0xbefc0370,
+	0xf4611b7a, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xbefe037c, 0xbefc0370,
+	0xf4611bba, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xbefe037c, 0xbefc0370,
+	0xf4611bfa, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xbefe037c, 0xbefc0370,
+	0xf4611e3a, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xb97af803, 0xbefe037c,
+	0xbefc0370, 0xf4611eba,
+	0xf8000000, 0x80708470,
 	0xbefc037e, 0xbefe037c,
-	0xbefc037a, 0xf4611e3a,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xb971f803,
-	0xbefe037c, 0xbefc037a,
+	0xbefc0370, 0xf4611efa,
+	0xf8000000, 0x80708470,
+	0xbefc037e, 0xb971f801,
+	0xbefe037c, 0xbefc0370,
 	0xf4611c7a, 0xf8000000,
-	0x807a847a, 0xbefc037e,
-	0xbefe037c, 0xbefc037a,
-	0xf4611cba, 0xf8000000,
-	0x807a847a, 0xbefc037e,
-	0xb97bf801, 0xbefe037c,
-	0xbefc037a, 0xf4611efa,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xb97bf814,
-	0xbefe037c, 0xbefc037a,
-	0xf4611efa, 0xf8000000,
-	0x807a847a, 0xbefc037e,
-	0xb97bf815, 0xbefe037c,
-	0xbefc037a, 0xf4611efa,
-	0xf8000000, 0x807a847a,
-	0xbefc037e, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0xbef603ff,
-	0x01000000, 0xbef20374,
-	0x80747a74, 0x82758075,
-	0xbefc0380, 0xbf800000,
-	0xbe802f00, 0xbe822f02,
-	0xbe842f04, 0xbe862f06,
-	0xbe882f08, 0xbe8a2f0a,
-	0xbe8c2f0c, 0xbe8e2f0e,
-	0xf469003a, 0xfa000000,
-	0xf469013a, 0xfa000010,
-	0xf469023a, 0xfa000020,
-	0xf469033a, 0xfa000030,
-	0x8074c074, 0x82758075,
-	0x807c907c, 0xbf0aff7c,
-	0x00000060, 0xbf85ffea,
-	0xbe802f00, 0xbe822f02,
-	0xbe842f04, 0xbe862f06,
-	0xbe882f08, 0xbe8a2f0a,
-	0xf469003a, 0xfa000000,
-	0xf469013a, 0xfa000010,
-	0xf469023a, 0xfa000020,
-	0x8074b074, 0x82758075,
-	0xbef40372, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb9714306,
-	0x8771c171, 0xbf840046,
-	0xbf8a0000, 0x8776ff73,
-	0x04000000, 0xbf840042,
-	0x8f718671, 0x8f718271,
-	0xbef60371, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0x807aff7a,
-	0x00000200, 0x807aff7a,
-	0x00000080, 0xbef603ff,
-	0x01000000, 0xd7650000,
-	0x000100c1, 0xd7660000,
-	0x000200c1, 0x16000084,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbefc0380,
-	0xbf850012, 0xbe8303ff,
-	0x00000080, 0xbf800000,
-	0xbf800000, 0xbf800000,
-	0xd8d80000, 0x01000000,
-	0xbf8c0000, 0xe0704000,
-	0x7a5d0100, 0x807c037c,
-	0x807a037a, 0xd5250000,
-	0x0001ff00, 0x00000080,
-	0xbf0a717c, 0xbf85fff4,
-	0xbf820011, 0xbe8303ff,
-	0x00000100, 0xbf800000,
-	0xbf800000, 0xbf800000,
-	0xd8d80000, 0x01000000,
-	0xbf8c0000, 0xe0704000,
-	0x7a5d0100, 0x807c037c,
-	0x807a037a, 0xd5250000,
-	0x0001ff00, 0x00000100,
-	0xbf0a717c, 0xbf85fff4,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850004, 0xbefa03ff,
-	0x00000200, 0xbeff0380,
-	0xbf820003, 0xbefa03ff,
-	0x00000400, 0xbeff03c1,
-	0xb9712a05, 0x80718171,
-	0x8f718271, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850017, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a717c, 0xbf840037,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
-	0xe0704000, 0x7a5d0000,
-	0xe0704080, 0x7a5d0100,
-	0xe0704100, 0x7a5d0200,
-	0xe0704180, 0x7a5d0300,
-	0x807c847c, 0x807aff7a,
-	0x00000200, 0xbf0a717c,
-	0xbf85ffef, 0xbf820025,
+	0x80708470, 0xbefc037e,
+	0xb971f814, 0xbefe037c,
+	0xbefc0370, 0xf4611c7a,
+	0xf8000000, 0x80708470,
+	0xbefc037e, 0xb971f815,
+	0xbefe037c, 0xbefc0370,
+	0xf4611c7a, 0xf8000000,
+	0x80708470, 0xbefc037e,
+	0xb9702a05, 0x80708170,
+	0xbf0d9973, 0xbf850002,
+	0x8f708970, 0xbf820001,
+	0x8f708a70, 0xb97a1e06,
+	0x8f7a8a7a, 0x80707a70,
 	0xbef603ff, 0x01000000,
-	0xbefc0384, 0xbf0a717c,
-	0xbf840020, 0x7e008700,
-	0x7e028701, 0x7e048702,
-	0x7e068703, 0xe0704000,
-	0x7a5d0000, 0xe0704100,
-	0x7a5d0100, 0xe0704200,
-	0x7a5d0200, 0xe0704300,
-	0x7a5d0300, 0x807c847c,
-	0x807aff7a, 0x00000400,
-	0xbf0a717c, 0xbf85ffef,
-	0xb9711e06, 0x8771c171,
-	0xbf84000c, 0x8f718371,
-	0x80717c71, 0xbefe03c1,
-	0xbeff0380, 0x7e008700,
-	0xe0704000, 0x7a5d0000,
-	0x807c817c, 0x807aff7a,
-	0x00000080, 0xbf0a717c,
-	0xbf85fff8, 0xbf820142,
-	0xbef4037e, 0x8775ff7f,
-	0x0000ffff, 0x8875ff75,
-	0x00040000, 0xbef60380,
-	0xbef703ff, 0x10807fac,
-	0x8772ff7f, 0x08000000,
-	0x90728372, 0x88777277,
-	0x8772ff7f, 0x70000000,
-	0x90728172, 0x88777277,
-	0xb97302dc, 0x8f739973,
-	0x8873737f, 0x8772ff7f,
-	0x04000000, 0xbf840036,
+	0xbefb0374, 0x80747074,
+	0x82758075, 0xbefc0380,
+	0xbf800000, 0xbe802f00,
+	0xbe822f02, 0xbe842f04,
+	0xbe862f06, 0xbe882f08,
+	0xbe8a2f0a, 0xbe8c2f0c,
+	0xbe8e2f0e, 0xf469003a,
+	0xfa000000, 0xf469013a,
+	0xfa000010, 0xf469023a,
+	0xfa000020, 0xf469033a,
+	0xfa000030, 0x8074c074,
+	0x82758075, 0x807c907c,
+	0xbf0aff7c, 0x00000060,
+	0xbf85ffea, 0xbe802f00,
+	0xbe822f02, 0xbe842f04,
+	0xbe862f06, 0xbe882f08,
+	0xbe8a2f0a, 0xf469003a,
+	0xfa000000, 0xf469013a,
+	0xfa000010, 0xf469023a,
+	0xfa000020, 0x8074b074,
+	0x82758075, 0xbef4037b,
 	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
-	0xb96f4306, 0x876fc16f,
-	0xbf84002b, 0x8f6f866f,
-	0x8f6f826f, 0xbef6036f,
-	0xb9782a05, 0x80788178,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb9721e06,
-	0x8f728a72, 0x80787278,
-	0x8078ff78, 0x00000200,
-	0x8078ff78, 0x00000080,
+	0xb97b4306, 0x877bc17b,
+	0xbf840044, 0xbf8a0000,
+	0x877aff73, 0x04000000,
+	0xbf840040, 0x8f7b867b,
+	0x8f7b827b, 0xbef6037b,
+	0xb9702a05, 0x80708170,
+	0xbf0d9973, 0xbf850002,
+	0x8f708970, 0xbf820001,
+	0x8f708a70, 0xb97a1e06,
+	0x8f7a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000080,
 	0xbef603ff, 0x01000000,
+	0xd7650000, 0x000100c1,
+	0xd7660000, 0x000200c1,
+	0x16000084, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850012,
+	0xbe8303ff, 0x00000080,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x705d0100,
+	0x807c037c, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000080, 0xbf0a7b7c,
+	0xbf85fff4, 0xbf820011,
+	0xbe8303ff, 0x00000100,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x705d0100,
+	0x807c037c, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000100, 0xbf0a7b7c,
+	0xbf85fff4, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbefc0380,
-	0xbf850009, 0xe0310000,
-	0x781d0000, 0x807cff7c,
-	0x00000080, 0x8078ff78,
-	0x00000080, 0xbf0a6f7c,
-	0xbf85fff8, 0xbf820008,
-	0xe0310000, 0x781d0000,
-	0x807cff7c, 0x00000100,
-	0x8078ff78, 0x00000100,
-	0xbf0a6f7c, 0xbf85fff8,
-	0xbef80380, 0xbefe03c1,
+	0xbf06817c, 0xbf850004,
+	0xbef003ff, 0x00000200,
+	0xbeff0380, 0xbf820003,
+	0xbef003ff, 0x00000400,
+	0xbeff03c1, 0xb97b2a05,
+	0x807b817b, 0x8f7b827b,
 	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850017,
+	0xbef603ff, 0x01000000,
+	0xbefc0384, 0xbf0a7b7c,
+	0xbf840037, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xe0704000,
+	0x705d0000, 0xe0704080,
+	0x705d0100, 0xe0704100,
+	0x705d0200, 0xe0704180,
+	0x705d0300, 0x807c847c,
+	0x8070ff70, 0x00000200,
+	0xbf0a7b7c, 0xbf85ffef,
+	0xbf820025, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a7b7c, 0xbf840020,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x705d0000,
+	0xe0704100, 0x705d0100,
+	0xe0704200, 0x705d0200,
+	0xe0704300, 0x705d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xb97b1e06,
+	0x877bc17b, 0xbf84000c,
+	0x8f7b837b, 0x807b7c7b,
+	0xbefe03c1, 0xbeff0380,
+	0x7e008700, 0xe0704000,
+	0x705d0000, 0x807c817c,
+	0x8070ff70, 0x00000080,
+	0xbf0a7b7c, 0xbf85fff8,
+	0xbf82014f, 0xbef4037e,
+	0x8775ff7f, 0x0000ffff,
+	0x8875ff75, 0x00040000,
+	0xbef60380, 0xbef703ff,
+	0x10807fac, 0x876eff7f,
+	0x08000000, 0x906e836e,
+	0x88776e77, 0x876eff7f,
+	0x70000000, 0x906e816e,
+	0x88776e77, 0xb97202dc,
+	0x8f729972, 0x8872727f,
+	0x876eff7f, 0x04000000,
+	0xbf840034, 0xbefe03c1,
+	0x907c9972, 0x877c817c,
 	0xbf06817c, 0xbf850002,
 	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb96f2a05,
-	0x806f816f, 0x8f6f826f,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850021,
-	0xbef603ff, 0x01000000,
-	0xbef20378, 0x8078ff78,
-	0x00000200, 0xbefc0384,
-	0xe0304000, 0x785d0000,
-	0xe0304080, 0x785d0100,
-	0xe0304100, 0x785d0200,
-	0xe0304180, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
-	0x7e028501, 0x7e048502,
-	0x7e068503, 0x807c847c,
+	0xbeff03c1, 0xb96f4306,
+	0x876fc16f, 0xbf840029,
+	0x8f6f866f, 0x8f6f826f,
+	0xbef6036f, 0xb9782a05,
+	0x80788178, 0xbf0d9972,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb96e1e06, 0x8f6e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x8078ff78,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850009,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000080,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbf820008, 0xe0310000,
+	0x781d0000, 0x807cff7c,
+	0x00000100, 0x8078ff78,
+	0x00000100, 0xbf0a6f7c,
+	0xbf85fff8, 0xbef80380,
+	0xbefe03c1, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820001, 0xbeff03c1,
+	0xb96f2a05, 0x806f816f,
+	0x8f6f826f, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbf850021, 0xbef603ff,
+	0x01000000, 0xbeee0378,
 	0x8078ff78, 0x00000200,
-	0xbf0a6f7c, 0xbf85ffee,
-	0xe0304000, 0x725d0000,
-	0xe0304080, 0x725d0100,
-	0xe0304100, 0x725d0200,
-	0xe0304180, 0x725d0300,
-	0xbf820032, 0xbef603ff,
-	0x01000000, 0xbef20378,
-	0x8078ff78, 0x00000400,
 	0xbefc0384, 0xe0304000,
-	0x785d0000, 0xe0304100,
-	0x785d0100, 0xe0304200,
-	0x785d0200, 0xe0304300,
+	0x785d0000, 0xe0304080,
+	0x785d0100, 0xe0304100,
+	0x785d0200, 0xe0304180,
 	0x785d0300, 0xbf8c3f70,
 	0x7e008500, 0x7e028501,
 	0x7e048502, 0x7e068503,
 	0x807c847c, 0x8078ff78,
-	0x00000400, 0xbf0a6f7c,
-	0xbf85ffee, 0xb96f1e06,
-	0x876fc16f, 0xbf84000e,
-	0x8f6f836f, 0x806f7c6f,
-	0xbefe03c1, 0xbeff0380,
+	0x00000200, 0xbf0a6f7c,
+	0xbf85ffee, 0xe0304000,
+	0x6e5d0000, 0xe0304080,
+	0x6e5d0100, 0xe0304100,
+	0x6e5d0200, 0xe0304180,
+	0x6e5d0300, 0xbf820032,
+	0xbef603ff, 0x01000000,
+	0xbeee0378, 0x8078ff78,
+	0x00000400, 0xbefc0384,
 	0xe0304000, 0x785d0000,
+	0xe0304100, 0x785d0100,
+	0xe0304200, 0x785d0200,
+	0xe0304300, 0x785d0300,
 	0xbf8c3f70, 0x7e008500,
-	0x807c817c, 0x8078ff78,
-	0x00000080, 0xbf0a6f7c,
-	0xbf85fff7, 0xbeff03c1,
-	0xe0304000, 0x725d0000,
-	0xe0304100, 0x725d0100,
-	0xe0304200, 0x725d0200,
-	0xe0304300, 0x725d0300,
-	0xbf8c3f70, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0x80f8ff78,
-	0x00000050, 0xbef603ff,
-	0x01000000, 0xbefc03ff,
-	0x0000006c, 0x80f89078,
-	0xf429003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc847c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0x80f8a078,
-	0xf42d003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc887c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0xbe843104,
-	0xbe863106, 0x80f8c078,
-	0xf431003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc907c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0xbe843104,
-	0xbe863106, 0xbe883108,
-	0xbe8a310a, 0xbe8c310c,
-	0xbe8e310e, 0xbf06807c,
-	0xbf84fff0, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0xf4211bfa,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xb96f1e06, 0x876fc16f,
+	0xbf84000e, 0x8f6f836f,
+	0x806f7c6f, 0xbefe03c1,
+	0xbeff0380, 0xe0304000,
+	0x785d0000, 0xbf8c3f70,
+	0x7e008500, 0x807c817c,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff7,
+	0xbeff03c1, 0xe0304000,
+	0x6e5d0000, 0xe0304100,
+	0x6e5d0100, 0xe0304200,
+	0x6e5d0200, 0xe0304300,
+	0x6e5d0300, 0xbf8c3f70,
+	0xb9782a05, 0x80788178,
+	0xbf0d9972, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb96e1e06,
+	0x8f6e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0x80f8ff78, 0x00000050,
+	0xbef603ff, 0x01000000,
+	0xbefc03ff, 0x0000006c,
+	0x80f89078, 0xf429003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc847c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0x80f8a078, 0xf42d003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc887c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0x80f8c078, 0xf431003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0xbe883108, 0xbe8a310a,
+	0xbe8c310c, 0xbe8e310e,
+	0xbf06807c, 0xbf84fff0,
+	0xb9782a05, 0x80788178,
+	0xbf0d9972, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb96e1e06,
+	0x8f6e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0xbef603ff, 0x01000000,
+	0xf4211bfa, 0xf0000000,
+	0x80788478, 0xf4211b3a,
 	0xf0000000, 0x80788478,
-	0xf4211b3a, 0xf0000000,
-	0x80788478, 0xf4211b7a,
+	0xf4211b7a, 0xf0000000,
+	0x80788478, 0xf4211c3a,
 	0xf0000000, 0x80788478,
-	0xf4211eba, 0xf0000000,
-	0x80788478, 0xf4211efa,
+	0xf4211c7a, 0xf0000000,
+	0x80788478, 0xf4211eba,
 	0xf0000000, 0x80788478,
-	0xf4211c3a, 0xf0000000,
-	0x80788478, 0xf4211c7a,
+	0xf4211efa, 0xf0000000,
+	0x80788478, 0xf4211e7a,
 	0xf0000000, 0x80788478,
-	0xf4211e7a, 0xf0000000,
-	0x80788478, 0xf4211cfa,
+	0xf4211cfa, 0xf0000000,
+	0x80788478, 0xf4211bba,
 	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef814, 0xf4211bba,
-	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef815,
-	0xbef2036d, 0x876dff72,
-	0x0000ffff, 0xbefc036f,
-	0xbefe037a, 0xbeff037b,
-	0x876f71ff, 0x000003ff,
+	0xb9eef815, 0xbefc036f,
+	0xbefe0370, 0xbeff0371,
+	0x876f7bff, 0x000003ff,
 	0xb9ef4803, 0xb9f9f816,
-	0x876f71ff, 0xfffff800,
+	0x876f7bff, 0xfffff800,
 	0x906f8b6f, 0xb9efa2c3,
-	0xb9f3f801, 0x876fff72,
+	0xb9f3f801, 0xb96e2a05,
+	0x806e816e, 0xbf0d9972,
+	0xbf850002, 0x8f6e896e,
+	0xbf820001, 0x8f6e8a6e,
+	0x806eff6e, 0x00000200,
+	0x806e746e, 0x826f8075,
+	0x876fff6f, 0x0000ffff,
+	0xf4091c37, 0xfa000050,
+	0xf4091d37, 0xfa000060,
+	0xf4011e77, 0xfa000074,
+	0xbf8cc07f, 0x876fff6d,
 	0xfc000000, 0x906f9a6f,
-	0x8f6f906f, 0xbef30380,
-	0x88736f73, 0x876fff72,
+	0x8f6f906f, 0xbeee0380,
+	0x886e6f6e, 0x876fff6d,
 	0x02000000, 0x906f996f,
-	0x8f6f8f6f, 0x88736f73,
-	0x876fff72, 0x01000000,
+	0x8f6f8f6f, 0x886e6f6e,
+	0x876fff6d, 0x01000000,
 	0x906f986f, 0x8f6f996f,
-	0x88736f73, 0x876fff70,
+	0x886e6f6e, 0x876fff7a,
 	0x00800000, 0x906f976f,
-	0xb9f3f807, 0x87fe7e7e,
-	0x87ea6a6a, 0xb9f0f802,
+	0xb9eef807, 0x876dff6d,
+	0x0000ffff, 0x87fe7e7e,
+	0x87ea6a6a, 0xb9faf802,
 	0xbf8a0000, 0xbe80226c,
 	0xbf810000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
@@ -1560,384 +1575,395 @@ static const uint32_t cwsr_trap_arcturus_hex[] = {
 };
 
 static const uint32_t cwsr_trap_gfx10_hex[] = {
-	0xbf820001, 0xbf8201c6,
+	0xbf820001, 0xbf8201cf,
 	0xb0804004, 0xb978f802,
-	0x8a788678, 0xb971f803,
-	0x876eff71, 0x00000400,
-	0xbf85001d, 0x876eff71,
-	0x00000100, 0xbf840002,
-	0x8878ff78, 0x00002000,
-	0xb970f812, 0xb971f813,
-	0x8ff08870, 0xf4051bb8,
-	0xfa000000, 0xbf8cc07f,
-	0xf4051c38, 0xfa000008,
-	0xbf8cc07f, 0x87ee6e6e,
-	0xbf840001, 0xbe80206e,
-	0xb971f803, 0x8771ff71,
-	0x000001ff, 0xbf850002,
-	0x806c846c, 0x826d806d,
-	0x876dff6d, 0x0000ffff,
-	0x87fe7e7e, 0x87ea6a6a,
-	0xb9f8f802, 0xbe80226c,
-	0xb971f803, 0x8771ff71,
-	0x00000100, 0xbf840006,
-	0xbef60380, 0xb9f60203,
-	0x876dff6d, 0x0000ffff,
-	0x80ec886c, 0x82ed806d,
-	0xbef60380, 0xb9f60283,
-	0xbeee037e, 0xbeef037f,
-	0xbefe0480, 0xbf900004,
-	0xbf8cc07f, 0xbef4037e,
-	0x8775ff7f, 0x0000ffff,
-	0x8875ff75, 0x00040000,
-	0xbef60380, 0xbef703ff,
-	0x10807fac, 0x8776ff7f,
-	0x08000000, 0x90768376,
-	0x88777677, 0x8776ff7f,
-	0x70000000, 0x90768176,
-	0x88777677, 0xbefb037c,
-	0xbefa0380, 0xb97302dc,
-	0x8f739973, 0x8873737f,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0xbeff0380,
-	0xbf820002, 0xbeff03c1,
-	0xbf82000b, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x7a5d0000, 0xe0704080,
-	0x7a5d0100, 0xe0704100,
-	0x7a5d0200, 0xe0704180,
-	0x7a5d0300, 0xbf82000a,
-	0xbef603ff, 0x01000000,
-	0xe0704000, 0x7a5d0000,
-	0xe0704100, 0x7a5d0100,
-	0xe0704200, 0x7a5d0200,
-	0xe0704300, 0x7a5d0300,
-	0xb97a2a05, 0x807a817a,
+	0x8a788678, 0xb96ef801,
+	0x876eff6e, 0x00000800,
+	0xbf840003, 0x876eff78,
+	0x00002000, 0xbf840009,
+	0xb97bf803, 0x876eff7b,
+	0x00000400, 0xbf85001d,
+	0x876eff7b, 0x00000100,
+	0xbf840002, 0x8878ff78,
+	0x00002000, 0xb97af812,
+	0xb97bf813, 0x8ffa887a,
+	0xf4051bbd, 0xfa000000,
+	0xbf8cc07f, 0xf4051ebd,
+	0xfa000008, 0xbf8cc07f,
+	0x87ee6e6e, 0xbf840001,
+	0xbe80206e, 0xb97bf803,
+	0x877bff7b, 0x000001ff,
+	0xbf850002, 0x806c846c,
+	0x826d806d, 0x876dff6d,
+	0x0000ffff, 0x87fe7e7e,
+	0x87ea6a6a, 0xb9f8f802,
+	0xbe80226c, 0x876dff6d,
+	0x0000ffff, 0xbefa0380,
+	0xb9fa0283, 0xbeee037e,
+	0xbeef037f, 0xbefe0480,
+	0xbf900004, 0xbf8cc07f,
+	0xb97b02dc, 0x8f7b997b,
+	0x887b7b7f, 0xb97a2a05,
+	0x807a817a, 0xbf0d997b,
+	0xbf850002, 0x8f7a897a,
+	0xbf820001, 0x8f7a8a7a,
+	0x877bff7f, 0x0000ffff,
+	0x807aff7a, 0x00000200,
+	0x807a7e7a, 0x827b807b,
+	0xbef4037e, 0x8775ff7f,
+	0x0000ffff, 0x8875ff75,
+	0x00040000, 0xbef60380,
+	0xbef703ff, 0x10807fac,
+	0x877aff7f, 0x08000000,
+	0x907a837a, 0x88777a77,
+	0x877aff7f, 0x70000000,
+	0x907a817a, 0x88777a77,
+	0xbef1037c, 0xbef00380,
+	0xb97302dc, 0x8f739973,
+	0x8873737f, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
 	0xbf06817c, 0xbf850002,
-	0x8f7a897a, 0xbf820001,
-	0x8f7a8a7a, 0xb9761e06,
-	0x8f768a76, 0x807a767a,
-	0x807aff7a, 0x00000200,
+	0xbeff0380, 0xbf820002,
+	0xbeff03c1, 0xbf82000b,
 	0xbef603ff, 0x01000000,
-	0x7e000280, 0x7e020280,
-	0x7e040280, 0xbefc0380,
-	0xd7610002, 0x0000f87b,
+	0xe0704000, 0x705d0000,
+	0xe0704080, 0x705d0100,
+	0xe0704100, 0x705d0200,
+	0xe0704180, 0x705d0300,
+	0xbf82000a, 0xbef603ff,
+	0x01000000, 0xe0704000,
+	0x705d0000, 0xe0704100,
+	0x705d0100, 0xe0704200,
+	0x705d0200, 0xe0704300,
+	0x705d0300, 0xb9702a05,
+	0x80708170, 0xbf0d9973,
+	0xbf850002, 0x8f708970,
+	0xbf820001, 0x8f708a70,
+	0xb97a1e06, 0x8f7a8a7a,
+	0x80707a70, 0x8070ff70,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0xbefc0380, 0xd7610002,
+	0x0000f871, 0x807c817c,
+	0xd7610002, 0x0000f86c,
 	0x807c817c, 0xd7610002,
-	0x0000f86c, 0x807c817c,
-	0xd7610002, 0x0000f86d,
+	0x0000f86d, 0x807c817c,
+	0xd7610002, 0x0000f86e,
 	0x807c817c, 0xd7610002,
-	0x0000f86e, 0x807c817c,
-	0xd7610002, 0x0000f86f,
+	0x0000f86f, 0x807c817c,
+	0xd7610002, 0x0000f878,
+	0x807c817c, 0xb97af803,
+	0xd7610002, 0x0000f87a,
 	0x807c817c, 0xd7610002,
-	0x0000f878, 0x807c817c,
-	0xb971f803, 0xd7610002,
+	0x0000f87b, 0x807c817c,
+	0xb971f801, 0xd7610002,
 	0x0000f871, 0x807c817c,
-	0xd7610002, 0x0000f872,
-	0x807c817c, 0xb97bf801,
-	0xd7610002, 0x0000f87b,
-	0x807c817c, 0xb97bf814,
-	0xd7610002, 0x0000f87b,
-	0x807c817c, 0xb97bf815,
-	0xd7610002, 0x0000f87b,
-	0x807c817c, 0xbeff0380,
-	0xe0704000, 0x7a5d0200,
-	0xb97a2a05, 0x807a817a,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0x8f7a897a, 0xbf820001,
-	0x8f7a8a7a, 0xb9761e06,
-	0x8f768a76, 0x807a767a,
-	0xbef603ff, 0x01000000,
-	0xbef90380, 0xbefc0380,
-	0xbf800000, 0xbe802f00,
-	0xbe822f02, 0xbe842f04,
-	0xbe862f06, 0xbe882f08,
-	0xbe8a2f0a, 0xbe8c2f0c,
-	0xbe8e2f0e, 0xd7610002,
-	0x0000f200, 0x80798179,
-	0xd7610002, 0x0000f201,
+	0xb971f814, 0xd7610002,
+	0x0000f871, 0x807c817c,
+	0xb971f815, 0xd7610002,
+	0x0000f871, 0x807c817c,
+	0xbeff0380, 0xe0704000,
+	0x705d0200, 0xb9702a05,
+	0x80708170, 0xbf0d9973,
+	0xbf850002, 0x8f708970,
+	0xbf820001, 0x8f708a70,
+	0xb97a1e06, 0x8f7a8a7a,
+	0x80707a70, 0xbef603ff,
+	0x01000000, 0xbef90380,
+	0xbefc0380, 0xbf800000,
+	0xbe802f00, 0xbe822f02,
+	0xbe842f04, 0xbe862f06,
+	0xbe882f08, 0xbe8a2f0a,
+	0xbe8c2f0c, 0xbe8e2f0e,
+	0xd7610002, 0x0000f200,
 	0x80798179, 0xd7610002,
-	0x0000f202, 0x80798179,
-	0xd7610002, 0x0000f203,
+	0x0000f201, 0x80798179,
+	0xd7610002, 0x0000f202,
 	0x80798179, 0xd7610002,
-	0x0000f204, 0x80798179,
-	0xd7610002, 0x0000f205,
+	0x0000f203, 0x80798179,
+	0xd7610002, 0x0000f204,
 	0x80798179, 0xd7610002,
-	0x0000f206, 0x80798179,
-	0xd7610002, 0x0000f207,
+	0x0000f205, 0x80798179,
+	0xd7610002, 0x0000f206,
 	0x80798179, 0xd7610002,
-	0x0000f208, 0x80798179,
-	0xd7610002, 0x0000f209,
+	0x0000f207, 0x80798179,
+	0xd7610002, 0x0000f208,
 	0x80798179, 0xd7610002,
-	0x0000f20a, 0x80798179,
-	0xd7610002, 0x0000f20b,
+	0x0000f209, 0x80798179,
+	0xd7610002, 0x0000f20a,
 	0x80798179, 0xd7610002,
-	0x0000f20c, 0x80798179,
-	0xd7610002, 0x0000f20d,
+	0x0000f20b, 0x80798179,
+	0xd7610002, 0x0000f20c,
 	0x80798179, 0xd7610002,
-	0x0000f20e, 0x80798179,
-	0xd7610002, 0x0000f20f,
-	0x80798179, 0xbf06a079,
-	0xbf840006, 0xe0704000,
-	0x7a5d0200, 0x807aff7a,
-	0x00000080, 0xbef90380,
-	0x7e040280, 0x807c907c,
-	0xbf0aff7c, 0x00000060,
-	0xbf85ffbc, 0xbe802f00,
-	0xbe822f02, 0xbe842f04,
-	0xbe862f06, 0xbe882f08,
-	0xbe8a2f0a, 0xd7610002,
-	0x0000f200, 0x80798179,
-	0xd7610002, 0x0000f201,
+	0x0000f20d, 0x80798179,
+	0xd7610002, 0x0000f20e,
 	0x80798179, 0xd7610002,
-	0x0000f202, 0x80798179,
-	0xd7610002, 0x0000f203,
+	0x0000f20f, 0x80798179,
+	0xbf06a079, 0xbf840006,
+	0xe0704000, 0x705d0200,
+	0x8070ff70, 0x00000080,
+	0xbef90380, 0x7e040280,
+	0x807c907c, 0xbf0aff7c,
+	0x00000060, 0xbf85ffbc,
+	0xbe802f00, 0xbe822f02,
+	0xbe842f04, 0xbe862f06,
+	0xbe882f08, 0xbe8a2f0a,
+	0xd7610002, 0x0000f200,
 	0x80798179, 0xd7610002,
-	0x0000f204, 0x80798179,
-	0xd7610002, 0x0000f205,
+	0x0000f201, 0x80798179,
+	0xd7610002, 0x0000f202,
 	0x80798179, 0xd7610002,
-	0x0000f206, 0x80798179,
-	0xd7610002, 0x0000f207,
+	0x0000f203, 0x80798179,
+	0xd7610002, 0x0000f204,
 	0x80798179, 0xd7610002,
-	0x0000f208, 0x80798179,
-	0xd7610002, 0x0000f209,
+	0x0000f205, 0x80798179,
+	0xd7610002, 0x0000f206,
 	0x80798179, 0xd7610002,
-	0x0000f20a, 0x80798179,
-	0xd7610002, 0x0000f20b,
-	0x80798179, 0xe0704000,
-	0x7a5d0200, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb9714306,
-	0x8771c171, 0xbf840046,
-	0xbf8a0000, 0x8776ff73,
-	0x04000000, 0xbf840042,
-	0x8f718671, 0x8f718271,
-	0xbef60371, 0xb97a2a05,
-	0x807a817a, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f7a897a,
-	0xbf820001, 0x8f7a8a7a,
-	0xb9761e06, 0x8f768a76,
-	0x807a767a, 0x807aff7a,
-	0x00000200, 0x807aff7a,
-	0x00000080, 0xbef603ff,
-	0x01000000, 0xd7650000,
-	0x000100c1, 0xd7660000,
-	0x000200c1, 0x16000084,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbefc0380,
-	0xbf850012, 0xbe8303ff,
-	0x00000080, 0xbf800000,
-	0xbf800000, 0xbf800000,
-	0xd8d80000, 0x01000000,
-	0xbf8c0000, 0xe0704000,
-	0x7a5d0100, 0x807c037c,
-	0x807a037a, 0xd5250000,
-	0x0001ff00, 0x00000080,
-	0xbf0a717c, 0xbf85fff4,
-	0xbf820011, 0xbe8303ff,
-	0x00000100, 0xbf800000,
-	0xbf800000, 0xbf800000,
-	0xd8d80000, 0x01000000,
-	0xbf8c0000, 0xe0704000,
-	0x7a5d0100, 0x807c037c,
-	0x807a037a, 0xd5250000,
-	0x0001ff00, 0x00000100,
-	0xbf0a717c, 0xbf85fff4,
-	0xbefe03c1, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850004, 0xbefa03ff,
-	0x00000200, 0xbeff0380,
-	0xbf820003, 0xbefa03ff,
-	0x00000400, 0xbeff03c1,
-	0xb9712a05, 0x80718171,
-	0x8f718271, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850017, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a717c, 0xbf840037,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
-	0xe0704000, 0x7a5d0000,
-	0xe0704080, 0x7a5d0100,
-	0xe0704100, 0x7a5d0200,
-	0xe0704180, 0x7a5d0300,
-	0x807c847c, 0x807aff7a,
-	0x00000200, 0xbf0a717c,
-	0xbf85ffef, 0xbf820025,
-	0xbef603ff, 0x01000000,
-	0xbefc0384, 0xbf0a717c,
-	0xbf840020, 0x7e008700,
-	0x7e028701, 0x7e048702,
-	0x7e068703, 0xe0704000,
-	0x7a5d0000, 0xe0704100,
-	0x7a5d0100, 0xe0704200,
-	0x7a5d0200, 0xe0704300,
-	0x7a5d0300, 0x807c847c,
-	0x807aff7a, 0x00000400,
-	0xbf0a717c, 0xbf85ffef,
-	0xb9711e06, 0x8771c171,
-	0xbf84000c, 0x8f718371,
-	0x80717c71, 0xbefe03c1,
-	0xbeff0380, 0x7e008700,
-	0xe0704000, 0x7a5d0000,
-	0x807c817c, 0x807aff7a,
-	0x00000080, 0xbf0a717c,
-	0xbf85fff8, 0xbf82012d,
-	0xbef4037e, 0x8775ff7f,
-	0x0000ffff, 0x8875ff75,
-	0x00040000, 0xbef60380,
-	0xbef703ff, 0x10807fac,
-	0x8772ff7f, 0x08000000,
-	0x90728372, 0x88777277,
-	0x8772ff7f, 0x70000000,
-	0x90728172, 0x88777277,
-	0xb97302dc, 0x8f739973,
-	0x8873737f, 0x8772ff7f,
-	0x04000000, 0xbf840036,
+	0x0000f207, 0x80798179,
+	0xd7610002, 0x0000f208,
+	0x80798179, 0xd7610002,
+	0x0000f209, 0x80798179,
+	0xd7610002, 0x0000f20a,
+	0x80798179, 0xd7610002,
+	0x0000f20b, 0x80798179,
+	0xe0704000, 0x705d0200,
 	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
-	0xb96f4306, 0x876fc16f,
-	0xbf84002b, 0x8f6f866f,
-	0x8f6f826f, 0xbef6036f,
-	0xb9782a05, 0x80788178,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb9721e06,
-	0x8f728a72, 0x80787278,
-	0x8078ff78, 0x00000200,
-	0x8078ff78, 0x00000080,
+	0xb97b4306, 0x877bc17b,
+	0xbf840044, 0xbf8a0000,
+	0x877aff73, 0x04000000,
+	0xbf840040, 0x8f7b867b,
+	0x8f7b827b, 0xbef6037b,
+	0xb9702a05, 0x80708170,
+	0xbf0d9973, 0xbf850002,
+	0x8f708970, 0xbf820001,
+	0x8f708a70, 0xb97a1e06,
+	0x8f7a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000080,
 	0xbef603ff, 0x01000000,
+	0xd7650000, 0x000100c1,
+	0xd7660000, 0x000200c1,
+	0x16000084, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850012,
+	0xbe8303ff, 0x00000080,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x705d0100,
+	0x807c037c, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000080, 0xbf0a7b7c,
+	0xbf85fff4, 0xbf820011,
+	0xbe8303ff, 0x00000100,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8c0000,
+	0xe0704000, 0x705d0100,
+	0x807c037c, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000100, 0xbf0a7b7c,
+	0xbf85fff4, 0xbefe03c1,
 	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbefc0380,
-	0xbf850009, 0xe0310000,
-	0x781d0000, 0x807cff7c,
-	0x00000080, 0x8078ff78,
-	0x00000080, 0xbf0a6f7c,
-	0xbf85fff8, 0xbf820008,
-	0xe0310000, 0x781d0000,
-	0x807cff7c, 0x00000100,
-	0x8078ff78, 0x00000100,
-	0xbf0a6f7c, 0xbf85fff8,
-	0xbef80380, 0xbefe03c1,
+	0xbf06817c, 0xbf850004,
+	0xbef003ff, 0x00000200,
+	0xbeff0380, 0xbf820003,
+	0xbef003ff, 0x00000400,
+	0xbeff03c1, 0xb97b2a05,
+	0x807b817b, 0x8f7b827b,
 	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbf850017,
+	0xbef603ff, 0x01000000,
+	0xbefc0384, 0xbf0a7b7c,
+	0xbf840037, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xe0704000,
+	0x705d0000, 0xe0704080,
+	0x705d0100, 0xe0704100,
+	0x705d0200, 0xe0704180,
+	0x705d0300, 0x807c847c,
+	0x8070ff70, 0x00000200,
+	0xbf0a7b7c, 0xbf85ffef,
+	0xbf820025, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a7b7c, 0xbf840020,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x705d0000,
+	0xe0704100, 0x705d0100,
+	0xe0704200, 0x705d0200,
+	0xe0704300, 0x705d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xb97b1e06,
+	0x877bc17b, 0xbf84000c,
+	0x8f7b837b, 0x807b7c7b,
+	0xbefe03c1, 0xbeff0380,
+	0x7e008700, 0xe0704000,
+	0x705d0000, 0x807c817c,
+	0x8070ff70, 0x00000080,
+	0xbf0a7b7c, 0xbf85fff8,
+	0xbf82013a, 0xbef4037e,
+	0x8775ff7f, 0x0000ffff,
+	0x8875ff75, 0x00040000,
+	0xbef60380, 0xbef703ff,
+	0x10807fac, 0x876eff7f,
+	0x08000000, 0x906e836e,
+	0x88776e77, 0x876eff7f,
+	0x70000000, 0x906e816e,
+	0x88776e77, 0xb97202dc,
+	0x8f729972, 0x8872727f,
+	0x876eff7f, 0x04000000,
+	0xbf840034, 0xbefe03c1,
+	0x907c9972, 0x877c817c,
 	0xbf06817c, 0xbf850002,
 	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb96f2a05,
-	0x806f816f, 0x8f6f826f,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850021,
-	0xbef603ff, 0x01000000,
-	0xbef20378, 0x8078ff78,
-	0x00000200, 0xbefc0384,
-	0xe0304000, 0x785d0000,
-	0xe0304080, 0x785d0100,
-	0xe0304100, 0x785d0200,
-	0xe0304180, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
-	0x7e028501, 0x7e048502,
-	0x7e068503, 0x807c847c,
+	0xbeff03c1, 0xb96f4306,
+	0x876fc16f, 0xbf840029,
+	0x8f6f866f, 0x8f6f826f,
+	0xbef6036f, 0xb9782a05,
+	0x80788178, 0xbf0d9972,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb96e1e06, 0x8f6e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x8078ff78,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbefc0380, 0xbf850009,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000080,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbf820008, 0xe0310000,
+	0x781d0000, 0x807cff7c,
+	0x00000100, 0x8078ff78,
+	0x00000100, 0xbf0a6f7c,
+	0xbf85fff8, 0xbef80380,
+	0xbefe03c1, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820001, 0xbeff03c1,
+	0xb96f2a05, 0x806f816f,
+	0x8f6f826f, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbf850021, 0xbef603ff,
+	0x01000000, 0xbeee0378,
 	0x8078ff78, 0x00000200,
-	0xbf0a6f7c, 0xbf85ffee,
-	0xe0304000, 0x725d0000,
-	0xe0304080, 0x725d0100,
-	0xe0304100, 0x725d0200,
-	0xe0304180, 0x725d0300,
-	0xbf820032, 0xbef603ff,
-	0x01000000, 0xbef20378,
-	0x8078ff78, 0x00000400,
 	0xbefc0384, 0xe0304000,
-	0x785d0000, 0xe0304100,
-	0x785d0100, 0xe0304200,
-	0x785d0200, 0xe0304300,
+	0x785d0000, 0xe0304080,
+	0x785d0100, 0xe0304100,
+	0x785d0200, 0xe0304180,
 	0x785d0300, 0xbf8c3f70,
 	0x7e008500, 0x7e028501,
 	0x7e048502, 0x7e068503,
 	0x807c847c, 0x8078ff78,
-	0x00000400, 0xbf0a6f7c,
-	0xbf85ffee, 0xb96f1e06,
-	0x876fc16f, 0xbf84000e,
-	0x8f6f836f, 0x806f7c6f,
-	0xbefe03c1, 0xbeff0380,
+	0x00000200, 0xbf0a6f7c,
+	0xbf85ffee, 0xe0304000,
+	0x6e5d0000, 0xe0304080,
+	0x6e5d0100, 0xe0304100,
+	0x6e5d0200, 0xe0304180,
+	0x6e5d0300, 0xbf820032,
+	0xbef603ff, 0x01000000,
+	0xbeee0378, 0x8078ff78,
+	0x00000400, 0xbefc0384,
 	0xe0304000, 0x785d0000,
+	0xe0304100, 0x785d0100,
+	0xe0304200, 0x785d0200,
+	0xe0304300, 0x785d0300,
 	0xbf8c3f70, 0x7e008500,
-	0x807c817c, 0x8078ff78,
-	0x00000080, 0xbf0a6f7c,
-	0xbf85fff7, 0xbeff03c1,
-	0xe0304000, 0x725d0000,
-	0xe0304100, 0x725d0100,
-	0xe0304200, 0x725d0200,
-	0xe0304300, 0x725d0300,
-	0xbf8c3f70, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0x80f8ff78,
-	0x00000050, 0xbef603ff,
-	0x01000000, 0xbefc03ff,
-	0x0000006c, 0x80f89078,
-	0xf429003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc847c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0x80f8a078,
-	0xf42d003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc887c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0xbe843104,
-	0xbe863106, 0x80f8c078,
-	0xf431003a, 0xf0000000,
-	0xbf8cc07f, 0x80fc907c,
-	0xbf800000, 0xbe803100,
-	0xbe823102, 0xbe843104,
-	0xbe863106, 0xbe883108,
-	0xbe8a310a, 0xbe8c310c,
-	0xbe8e310e, 0xbf06807c,
-	0xbf84fff0, 0xb9782a05,
-	0x80788178, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb9721e06, 0x8f728a72,
-	0x80787278, 0x8078ff78,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0xf4211bfa,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xb96f1e06, 0x876fc16f,
+	0xbf84000e, 0x8f6f836f,
+	0x806f7c6f, 0xbefe03c1,
+	0xbeff0380, 0xe0304000,
+	0x785d0000, 0xbf8c3f70,
+	0x7e008500, 0x807c817c,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7c, 0xbf85fff7,
+	0xbeff03c1, 0xe0304000,
+	0x6e5d0000, 0xe0304100,
+	0x6e5d0100, 0xe0304200,
+	0x6e5d0200, 0xe0304300,
+	0x6e5d0300, 0xbf8c3f70,
+	0xb9782a05, 0x80788178,
+	0xbf0d9972, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb96e1e06,
+	0x8f6e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0x80f8ff78, 0x00000050,
+	0xbef603ff, 0x01000000,
+	0xbefc03ff, 0x0000006c,
+	0x80f89078, 0xf429003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc847c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0x80f8a078, 0xf42d003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc887c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0x80f8c078, 0xf431003a,
+	0xf0000000, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe803100, 0xbe823102,
+	0xbe843104, 0xbe863106,
+	0xbe883108, 0xbe8a310a,
+	0xbe8c310c, 0xbe8e310e,
+	0xbf06807c, 0xbf84fff0,
+	0xb9782a05, 0x80788178,
+	0xbf0d9972, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb96e1e06,
+	0x8f6e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0xbef603ff, 0x01000000,
+	0xf4211bfa, 0xf0000000,
+	0x80788478, 0xf4211b3a,
 	0xf0000000, 0x80788478,
-	0xf4211b3a, 0xf0000000,
-	0x80788478, 0xf4211b7a,
+	0xf4211b7a, 0xf0000000,
+	0x80788478, 0xf4211c3a,
 	0xf0000000, 0x80788478,
-	0xf4211eba, 0xf0000000,
-	0x80788478, 0xf4211efa,
+	0xf4211c7a, 0xf0000000,
+	0x80788478, 0xf4211eba,
 	0xf0000000, 0x80788478,
-	0xf4211c3a, 0xf0000000,
-	0x80788478, 0xf4211c7a,
+	0xf4211efa, 0xf0000000,
+	0x80788478, 0xf4211e7a,
 	0xf0000000, 0x80788478,
-	0xf4211e7a, 0xf0000000,
-	0x80788478, 0xf4211cfa,
+	0xf4211cfa, 0xf0000000,
+	0x80788478, 0xf4211bba,
 	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef814, 0xf4211bba,
-	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef815,
-	0xbef2036d, 0x876dff72,
-	0x0000ffff, 0xbefc036f,
-	0xbefe037a, 0xbeff037b,
-	0x876f71ff, 0x000003ff,
-	0xb9ef4803, 0x876f71ff,
+	0xb9eef815, 0xbefc036f,
+	0xbefe0370, 0xbeff0371,
+	0x876f7bff, 0x000003ff,
+	0xb9ef4803, 0x876f7bff,
 	0xfffff800, 0x906f8b6f,
 	0xb9efa2c3, 0xb9f3f801,
+	0xb96e2a05, 0x806e816e,
+	0xbf0d9972, 0xbf850002,
+	0x8f6e896e, 0xbf820001,
+	0x8f6e8a6e, 0x806eff6e,
+	0x00000200, 0x806e746e,
+	0x826f8075, 0x876fff6f,
+	0x0000ffff, 0xf4091c37,
+	0xfa000050, 0xf4091d37,
+	0xfa000060, 0xf4011e77,
+	0xfa000074, 0xbf8cc07f,
+	0x876dff6d, 0x0000ffff,
 	0x87fe7e7e, 0x87ea6a6a,
-	0xb9f0f802, 0xbf8a0000,
+	0xb9faf802, 0xbf8a0000,
 	0xbe80226c, 0xbf810000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
index 05dd29035772..5b220f2a7501 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
@@ -33,6 +33,8 @@
 
 #define NO_SQC_STORE !ASIC_TARGET_NAVI1X
 
+var SINGLE_STEP_MISSED_WORKAROUND		= 1	//workaround for lost MODE.DEBUG_EN exception when SAVECTX raised
+
 var SQ_WAVE_STATUS_INST_ATC_SHIFT		= 23
 var SQ_WAVE_STATUS_INST_ATC_MASK		= 0x00800000
 var SQ_WAVE_STATUS_SPI_PRIO_MASK		= 0x00000006
@@ -72,6 +74,8 @@ var SQ_WAVE_IB_STS_RCNT_SIZE			= 6
 var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK	= 0x003F8000
 var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG	= 0x00007FFF
 
+var SQ_WAVE_MODE_DEBUG_EN_MASK			= 0x800
+
 var SQ_BUF_RSRC_WORD1_ATC_SHIFT			= 24
 var SQ_BUF_RSRC_WORD3_MTYPE_SHIFT		= 27
 
@@ -109,17 +113,19 @@ var s_save_pc_hi				= ttmp1
 var s_save_exec_lo				= ttmp2
 var s_save_exec_hi				= ttmp3
 var s_save_status				= ttmp12
-var s_save_trapsts				= ttmp5
-var s_save_xnack_mask				= ttmp6
+var s_save_trapsts				= ttmp15
+var s_save_xnack_mask				= s_save_trapsts
 var s_wave_size					= ttmp7
 var s_save_buf_rsrc0				= ttmp8
 var s_save_buf_rsrc1				= ttmp9
 var s_save_buf_rsrc2				= ttmp10
 var s_save_buf_rsrc3				= ttmp11
-var s_save_mem_offset				= ttmp14
+var s_save_mem_offset				= ttmp4
 var s_save_alloc_size				= s_save_trapsts
-var s_save_tmp					= s_save_buf_rsrc2
-var s_save_m0					= ttmp15
+var s_save_tmp					= ttmp14
+var s_save_m0					= ttmp5
+var s_save_ttmps_lo				= s_save_tmp
+var s_save_ttmps_hi				= s_save_trapsts
 
 var S_RESTORE_BUF_RSRC_WORD1_STRIDE		= S_SAVE_BUF_RSRC_WORD1_STRIDE
 var S_RESTORE_BUF_RSRC_WORD3_MISC		= S_SAVE_BUF_RSRC_WORD3_MISC
@@ -141,23 +147,25 @@ var s_restore_spi_init_lo			= exec_lo
 var s_restore_spi_init_hi			= exec_hi
 var s_restore_mem_offset			= ttmp12
 var s_restore_alloc_size			= ttmp3
-var s_restore_tmp				= ttmp6
+var s_restore_tmp				= ttmp2
 var s_restore_mem_offset_save			= s_restore_tmp
 var s_restore_m0				= s_restore_alloc_size
 var s_restore_mode				= ttmp7
-var s_restore_flat_scratch			= ttmp2
+var s_restore_flat_scratch			= s_restore_tmp
 var s_restore_pc_lo				= ttmp0
 var s_restore_pc_hi				= ttmp1
-var s_restore_exec_lo				= ttmp14
-var s_restore_exec_hi				= ttmp15
-var s_restore_status				= ttmp4
-var s_restore_trapsts				= ttmp5
+var s_restore_exec_lo				= ttmp4
+var s_restore_exec_hi				= ttmp5
+var s_restore_status				= ttmp14
+var s_restore_trapsts				= ttmp15
 var s_restore_xnack_mask			= ttmp13
 var s_restore_buf_rsrc0				= ttmp8
 var s_restore_buf_rsrc1				= ttmp9
 var s_restore_buf_rsrc2				= ttmp10
 var s_restore_buf_rsrc3				= ttmp11
-var s_restore_size				= ttmp7
+var s_restore_size				= ttmp6
+var s_restore_ttmps_lo				= s_restore_tmp
+var s_restore_ttmps_hi				= s_restore_alloc_size
 
 shader main
 	asic(DEFAULT)
@@ -172,6 +180,24 @@ L_JUMP_TO_RESTORE:
 L_SKIP_RESTORE:
 	s_getreg_b32	s_save_status, hwreg(HW_REG_STATUS)			//save STATUS since we will change SCC
 	s_andn2_b32	s_save_status, s_save_status, SQ_WAVE_STATUS_SPI_PRIO_MASK
+
+if SINGLE_STEP_MISSED_WORKAROUND
+	// No single step exceptions if MODE.DEBUG_EN=0.
+	s_getreg_b32    ttmp2, hwreg(HW_REG_MODE)
+	s_and_b32       ttmp2, ttmp2, SQ_WAVE_MODE_DEBUG_EN_MASK
+	s_cbranch_scc0  L_NO_SINGLE_STEP_WORKAROUND
+
+	// Second-level trap already handled exception if STATUS.HALT=1.
+	s_and_b32       ttmp2, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+
+	// Prioritize single step exception over context save.
+	// Second-level trap will halt wave and RFE, re-entering for SAVECTX.
+	s_cbranch_scc0  L_FETCH_2ND_TRAP
+
+L_NO_SINGLE_STEP_WORKAROUND:
+end
+
+
 	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
 	s_and_b32	ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_SAVECTX_MASK	//check whether this is for save
 	s_cbranch_scc1	L_SAVE
@@ -241,25 +267,11 @@ L_EXCP_CASE:
 	s_rfe_b64	[ttmp0, ttmp1]
 
 L_SAVE:
-	//check whether there is mem_viol
-	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
-	s_and_b32	s_save_trapsts, s_save_trapsts, SQ_WAVE_TRAPSTS_MEM_VIOL_MASK
-	s_cbranch_scc0	L_NO_PC_REWIND
-
-	//if so, need rewind PC assuming GDS operation gets NACKed
-	s_mov_b32	s_save_tmp, 0
-	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_MEM_VIOL_SHIFT, 1), s_save_tmp	//clear mem_viol bit
 	s_and_b32	s_save_pc_hi, s_save_pc_hi, 0x0000ffff			//pc[47:32]
-	s_sub_u32	s_save_pc_lo, s_save_pc_lo, 8				//pc[31:0]-8
-	s_subb_u32	s_save_pc_hi, s_save_pc_hi, 0x0
-
-L_NO_PC_REWIND:
 	s_mov_b32	s_save_tmp, 0
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_SAVECTX_SHIFT, 1), s_save_tmp	//clear saveCtx bit
 
 #if ASIC_TARGET_NAVI1X
-	s_getreg_b32	s_save_xnack_mask, hwreg(HW_REG_SHADER_XNACK_MASK)
-
 	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_RCNT_SHIFT, SQ_WAVE_IB_STS_RCNT_SIZE)
 	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_RCNT_SHIFT
 	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
@@ -293,6 +305,21 @@ L_SLEEP:
 	s_waitcnt	lgkmcnt(0)
 #endif
 
+	// Save trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
+	// ttmp SR memory offset : size(VGPR)+size(SGPR)+0x40
+	get_wave_size(s_save_ttmps_hi)
+	get_vgpr_size_bytes(s_save_ttmps_lo, s_save_ttmps_hi)
+	s_and_b32	s_save_ttmps_hi, s_save_spi_init_hi, 0xFFFF
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, get_sgpr_size_bytes()
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, s_save_spi_init_lo
+	s_addc_u32	s_save_ttmps_hi, s_save_ttmps_hi, 0x0
+
+#if ASIC_TARGET_NAVI1X
+	s_store_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_save_ttmps_lo, s_save_ttmps_hi], 0x50 glc:1
+	s_store_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_save_ttmps_lo, s_save_ttmps_hi], 0x60 glc:1
+	s_store_dword   ttmp13, [s_save_ttmps_lo, s_save_ttmps_hi], 0x74 glc:1
+#endif
+
 	/* setup Resource Contants */
 	s_mov_b32	s_save_buf_rsrc0, s_save_spi_init_lo			//base_addr_lo
 	s_and_b32	s_save_buf_rsrc1, s_save_spi_init_hi, 0x0000FFFF	//base_addr_hi
@@ -310,9 +337,13 @@ L_SLEEP:
 
 	/* global mem offset */
 	s_mov_b32	s_save_mem_offset, 0x0
-	s_getreg_b32	s_wave_size, hwreg(HW_REG_IB_STS2,SQ_WAVE_IB_STS2_WAVE64_SHIFT,SQ_WAVE_IB_STS2_WAVE64_SIZE)
-	s_lshl_b32	s_wave_size, s_wave_size, S_WAVE_SIZE
-	s_or_b32	s_wave_size, s_save_spi_init_hi, s_wave_size		//share s_wave_size with exec_hi, it's at bit25
+	get_wave_size(s_wave_size)
+
+#if ASIC_TARGET_NAVI1X
+	// Save and clear vector XNACK state late to free up SGPRs.
+	s_getreg_b32	s_save_xnack_mask, hwreg(HW_REG_SHADER_XNACK_MASK)
+	s_setreg_imm32_b32	hwreg(HW_REG_SHADER_XNACK_MASK), 0x0
+#endif
 
 	/* save first 4 VGPRs, needed for SGPR save */
 	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
@@ -371,8 +402,8 @@ L_SAVE_HWREG:
 	write_hwreg_to_mem(s_save_exec_hi, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_status, s_save_buf_rsrc0, s_save_mem_offset)
 
-	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
-	write_hwreg_to_mem(s_save_trapsts, s_save_buf_rsrc0, s_save_mem_offset)
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_TRAPSTS)
+	write_hwreg_to_mem(s_save_tmp, s_save_buf_rsrc0, s_save_mem_offset)
 
 	// Not used on Sienna_Cichlid but keep layout same for debugger.
 	write_hwreg_to_mem(s_save_xnack_mask, s_save_buf_rsrc0, s_save_mem_offset)
@@ -650,9 +681,7 @@ L_RESTORE:
 	s_lshr_b32	s_restore_tmp, s_restore_tmp, (S_RESTORE_SPI_INIT_MTYPE_SHIFT-SQ_BUF_RSRC_WORD3_MTYPE_SHIFT)
 	s_or_b32	s_restore_buf_rsrc3, s_restore_buf_rsrc3, s_restore_tmp	//or MTYPE
 	//determine it is wave32 or wave64
-	s_getreg_b32	s_restore_size, hwreg(HW_REG_IB_STS2,SQ_WAVE_IB_STS2_WAVE64_SHIFT,SQ_WAVE_IB_STS2_WAVE64_SIZE)
-	s_lshl_b32	s_restore_size, s_restore_size, S_WAVE_SIZE
-	s_or_b32	s_restore_size, s_restore_spi_init_hi, s_restore_size
+	get_wave_size(s_restore_size)
 
 	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_FIRST_WAVE_MASK
 	s_cbranch_scc0	L_RESTORE_VGPR
@@ -686,7 +715,7 @@ L_RESTORE_LDS_NORMAL:
 
 	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
 
-	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_lshr_b32	m0, s_restore_size, S_WAVE_SIZE
 	s_and_b32	m0, m0, 1
 	s_cmp_eq_u32	m0, 1
 	s_mov_b32	m0, 0x0
@@ -894,9 +923,6 @@ L_RESTORE_HWREG:
 
 	s_setreg_b32	hwreg(HW_REG_SHADER_FLAT_SCRATCH_HI), s_restore_flat_scratch
 
-	s_mov_b32	s_restore_tmp, s_restore_pc_hi
-	s_and_b32	s_restore_pc_hi, s_restore_tmp, 0x0000ffff		//pc[47:32] //Do it here in order not to affect STATUS
-
 	s_mov_b32	m0, s_restore_m0
 	s_mov_b32	exec_lo, s_restore_exec_lo
 	s_mov_b32	exec_hi, s_restore_exec_hi
@@ -913,26 +939,39 @@ L_RESTORE_HWREG:
 	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE), s_restore_m0
 	s_setreg_b32	hwreg(HW_REG_MODE), s_restore_mode
 
+	// Restore trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
+	// ttmp SR memory offset : size(VGPR)+size(SGPR)+0x40
+	get_vgpr_size_bytes(s_restore_ttmps_lo, s_restore_size)
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, get_sgpr_size_bytes()
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, s_restore_buf_rsrc0
+	s_addc_u32	s_restore_ttmps_hi, s_restore_buf_rsrc1, 0x0
+	s_and_b32	s_restore_ttmps_hi, s_restore_ttmps_hi, 0xFFFF
+	s_load_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x50 glc:1
+	s_load_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x60 glc:1
+	s_load_dword	ttmp13, [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x74 glc:1
+	s_waitcnt	lgkmcnt(0)
+
 #if ASIC_TARGET_NAVI1X
-	s_and_b32	s_restore_m0, s_restore_tmp, S_SAVE_PC_HI_RCNT_MASK
+	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_RCNT_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_RCNT_SHIFT
 	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_RCNT_SHIFT
-	s_mov_b32	s_restore_mode, 0x0
-	s_or_b32	s_restore_mode, s_restore_mode, s_restore_m0
-	s_and_b32	s_restore_m0, s_restore_tmp, S_SAVE_PC_HI_FIRST_REPLAY_MASK
+	s_mov_b32	s_restore_tmp, 0x0
+	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
+	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_FIRST_REPLAY_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
 	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT
-	s_or_b32	s_restore_mode, s_restore_mode, s_restore_m0
-	s_and_b32	s_restore_m0, s_restore_tmp, S_SAVE_PC_HI_REPLAY_W64H_MASK
+	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
+	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_REPLAY_W64H_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_REPLAY_W64H_SHIFT
 	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT
-	s_or_b32	s_restore_mode, s_restore_mode, s_restore_m0
+	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
 
 	s_and_b32	s_restore_m0, s_restore_status, SQ_WAVE_STATUS_INST_ATC_MASK
 	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_STATUS_INST_ATC_SHIFT
-	s_setreg_b32 	hwreg(HW_REG_IB_STS), s_restore_mode
+	s_setreg_b32 	hwreg(HW_REG_IB_STS), s_restore_tmp
 #endif
 
+	s_and_b32	s_restore_pc_hi, s_restore_pc_hi, 0x0000ffff		//pc[47:32] //Do it here in order not to affect STATUS
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
 	s_setreg_b32	hwreg(HW_REG_STATUS), s_restore_status			// SCC is included, which is changed by previous salu
@@ -1022,9 +1061,7 @@ end
 function get_vgpr_size_bytes(s_vgpr_size_byte, s_size)
 	s_getreg_b32	s_vgpr_size_byte, hwreg(HW_REG_GPR_ALLOC,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE)
 	s_add_u32	s_vgpr_size_byte, s_vgpr_size_byte, 1
-	s_lshr_b32	m0, s_size, S_WAVE_SIZE
-	s_and_b32	m0, m0, 1
-	s_cmp_eq_u32	m0, 1
+	s_bitcmp1_b32	s_size, S_WAVE_SIZE
 	s_cbranch_scc1	L_ENABLE_SHIFT_W64
 	s_lshl_b32	s_vgpr_size_byte, s_vgpr_size_byte, (2+7)		//Number of VGPRs = (vgpr_size + 1) * 4 * 32 * 4   (non-zero value)
 	s_branch	L_SHIFT_DONE
@@ -1045,3 +1082,9 @@ end
 function get_hwreg_size_bytes
 	return 128
 end
+
+function get_wave_size(s_reg)
+	s_getreg_b32	s_reg, hwreg(HW_REG_IB_STS2,SQ_WAVE_IB_STS2_WAVE64_SHIFT,SQ_WAVE_IB_STS2_WAVE64_SIZE)
+	s_lshl_b32	s_reg, s_reg, S_WAVE_SIZE
+	s_or_b32	s_reg, s_save_spi_init_hi, s_reg			//share with exec_hi, it's at bit25
+end
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 120/207] drm/amdkfd: sienna_cichlid virtual function support

[Alex Deucher]

From: shaoyunl <shaoyun.liu@amd.com>

amdkfd add support for sienna_cichlid virtual function

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Yong Zhao <Yong.Zhao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdkfd/kfd_device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index c36a352776a7..22348cebaf36 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -500,7 +500,7 @@ static const struct kfd_device_info *kfd_supported_devices[][2] = {
 	[CHIP_NAVI10] = {&navi10_device_info, NULL},
 	[CHIP_NAVI12] = {&navi12_device_info, &navi12_device_info},
 	[CHIP_NAVI14] = {&navi14_device_info, NULL},
-	[CHIP_SIENNA_CICHLID] = {&sienna_cichlid_device_info, NULL},
+	[CHIP_SIENNA_CICHLID] = {&sienna_cichlid_device_info, &sienna_cichlid_device_info},
 };
 
 static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 121/207] drm/amd/powerplay: enable VR0HOT for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable the feature of Voltage Regulator (VR) Hot for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 5e2b01a565b3..6824cc804083 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -299,6 +299,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_FW_DSTATE_BIT)
 				| FEATURE_MASK(FEATURE_DF_CSTATE_BIT)
 				| FEATURE_MASK(FEATURE_GFX_SS_BIT)
+				| FEATURE_MASK(FEATURE_VR0HOT_BIT)
 				| FEATURE_MASK(FEATURE_PPT_BIT)
 				| FEATURE_MASK(FEATURE_TDC_BIT)
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 122/207] drm/amd/powerplay: enable FCLK DS for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable the feature of FCLK Deep Sleep for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 6824cc804083..2d844569ed8d 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -296,6 +296,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_DPM_FCLK_BIT)
 				| FEATURE_MASK(FEATURE_DS_SOCCLK_BIT)
 				| FEATURE_MASK(FEATURE_DS_DCEFCLK_BIT)
+				| FEATURE_MASK(FEATURE_DS_FCLK_BIT)
 				| FEATURE_MASK(FEATURE_FW_DSTATE_BIT)
 				| FEATURE_MASK(FEATURE_DF_CSTATE_BIT)
 				| FEATURE_MASK(FEATURE_GFX_SS_BIT)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 123/207] drm/amdgpu: fix typo for vcn3/jpeg3 idle check

[Alex Deucher]

From: James Zhu <James.Zhu@amd.com>

fix typo for vcn3/jpeg3 idle check

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c | 2 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
index 2fc778cad91a..9dbc616b38cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
@@ -479,7 +479,7 @@ static int jpeg_v3_0_set_clockgating_state(void *handle,
 	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
 
 	if (enable) {
-		if (jpeg_v3_0_is_idle(handle))
+		if (!jpeg_v3_0_is_idle(handle))
 			return -EBUSY;
 		jpeg_v3_0_enable_clock_gating(adev);
 	} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 517484292303..340cb344a35d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -1174,7 +1174,7 @@ static int vcn_v3_0_set_clockgating_state(void *handle,
 			continue;
 
 		if (enable) {
-			if (RREG32_SOC15(VCN, i, mmUVD_STATUS) == UVD_STATUS__IDLE)
+			if (RREG32_SOC15(VCN, i, mmUVD_STATUS) != UVD_STATUS__IDLE)
 				return -EBUSY;
 			vcn_v3_0_enable_clock_gating(adev, i);
 		} else {
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 124/207] drm/amd/powerplay: enable MM DPM PG for sienna_cichlid (v2)

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable VCN dpm set for sienna_cichlid.
Enable JPEG dpm set for sienna_cichlid.

v2: squash in BACO fix (Kenneth)

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/inc/smu_types.h |  1 +
 .../drm/amd/powerplay/sienna_cichlid_ppt.c    | 51 +++++++++++++++++--
 2 files changed, 48 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
index 1ef791f36888..3efadf2cff9c 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
@@ -247,6 +247,7 @@ enum smu_clk_type {
        __SMU_DUMMY_MAP(FW_DSTATE),                     	\
        __SMU_DUMMY_MAP(BACO),                          	\
        __SMU_DUMMY_MAP(VCN_PG),                        	\
+       __SMU_DUMMY_MAP(MM_DPM_PG),                     	\
        __SMU_DUMMY_MAP(JPEG_PG),                       	\
        __SMU_DUMMY_MAP(USB_PG),                        	\
        __SMU_DUMMY_MAP(RSMU_SMN_CG),                   	\
diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 2d844569ed8d..80a099877a68 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -141,6 +141,7 @@ static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_feature_mask_map[SMU_FEAT
 	FEA_MAP(FW_DSTATE),
 	FEA_MAP(GFXOFF),
 	FEA_MAP(BACO),
+	FEA_MAP(MM_DPM_PG),
 	FEA_MAP(RSMU_SMN_CG),
 	FEA_MAP(PPT),
 	FEA_MAP(TDC),
@@ -327,6 +328,10 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	if (adev->pm.pp_feature & PP_SCLK_DEEP_SLEEP_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DS_GFXCLK_BIT);
 
+	if (smu->adev->pg_flags & AMD_PG_SUPPORT_VCN ||
+	    smu->adev->pg_flags & AMD_PG_SUPPORT_JPEG)
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_MM_DPM_PG_BIT);
+
 	return 0;
 }
 
@@ -485,15 +490,21 @@ static int sienna_cichlid_dpm_set_uvd_enable(struct smu_context *smu, bool enabl
 
 	if (enable) {
 		/* vcn dpm on is a prerequisite for vcn power gate messages */
-		if (smu_feature_is_enabled(smu, SMU_FEATURE_VCN_PG_BIT)) {
-			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpVcn, 1, NULL);
+		if (smu_feature_is_enabled(smu, SMU_FEATURE_MM_DPM_PG_BIT)) {
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpVcn, 0, NULL);
+			if (ret)
+				return ret;
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpVcn, 0x10000, NULL);
 			if (ret)
 				return ret;
 		}
 		power_gate->vcn_gated = false;
 	} else {
-		if (smu_feature_is_enabled(smu, SMU_FEATURE_VCN_PG_BIT)) {
-			ret = smu_send_smc_msg(smu, SMU_MSG_PowerDownVcn, NULL);
+		if (smu_feature_is_enabled(smu, SMU_FEATURE_MM_DPM_PG_BIT)) {
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownVcn, 0, NULL);
+			if (ret)
+				return ret;
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownVcn, 0x10000, NULL);
 			if (ret)
 				return ret;
 		}
@@ -503,6 +514,37 @@ static int sienna_cichlid_dpm_set_uvd_enable(struct smu_context *smu, bool enabl
 	return ret;
 }
 
+static int sienna_cichlid_dpm_set_jpeg_enable(struct smu_context *smu, bool enable)
+{
+	struct smu_power_context *smu_power = &smu->smu_power;
+	struct smu_power_gate *power_gate = &smu_power->power_gate;
+	int ret = 0;
+
+	if (enable) {
+		if (smu_feature_is_enabled(smu, SMU_FEATURE_MM_DPM_PG_BIT)) {
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0, NULL);
+			if (ret)
+				return ret;
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0x10000, NULL);
+			if (ret)
+				return ret;
+		}
+		power_gate->jpeg_gated = false;
+	} else {
+		if (smu_feature_is_enabled(smu, SMU_FEATURE_MM_DPM_PG_BIT)) {
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0, NULL);
+			if (ret)
+				return ret;
+			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0x10000, NULL);
+			if (ret)
+				return ret;
+		}
+		power_gate->jpeg_gated = true;
+	}
+
+	return ret;
+}
+
 static int sienna_cichlid_get_current_clk_freq_by_table(struct smu_context *smu,
 				       enum smu_clk_type clk_type,
 				       uint32_t *value)
@@ -2214,6 +2256,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.get_allowed_feature_mask = sienna_cichlid_get_allowed_feature_mask,
 	.set_default_dpm_table = sienna_cichlid_set_default_dpm_table,
 	.dpm_set_uvd_enable = sienna_cichlid_dpm_set_uvd_enable,
+	.dpm_set_jpeg_enable = sienna_cichlid_dpm_set_jpeg_enable,
 	.get_current_clk_freq_by_table = sienna_cichlid_get_current_clk_freq_by_table,
 	.print_clk_levels = sienna_cichlid_print_clk_levels,
 	.force_clk_levels = sienna_cichlid_force_clk_levels,
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 125/207] drm/amdgpu: Sienna_Cichlid don't enable SMU for SRIOV

[Alex Deucher]

From: shaoyunl <shaoyun.liu@amd.com>

SMU firmware already been loaded from host, don't enable it for now.
May need to re-work it if we want to enable the SMU for guest in the future.

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 89c3727b7a8e..f4d4d91ebee2 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -493,7 +493,7 @@ int nv_set_ip_blocks(struct amdgpu_device *adev)
 		if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP))
 			amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP &&
-		    is_support_sw_smu(adev))
+		    is_support_sw_smu(adev) && !amdgpu_sriov_vf(adev))
 			amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
 		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
 			amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 126/207] drm/amd/powerplay: enable BACO for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable BACO for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 80a099877a68..e0067921a3e9 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -103,6 +103,7 @@ static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_message_map[SMU_MSG_MAX_C
 	MSG_MAP(PowerUpJpeg,			PPSMC_MSG_PowerUpJpeg),
 	MSG_MAP(PowerDownJpeg,			PPSMC_MSG_PowerDownJpeg),
 	MSG_MAP(BacoAudioD3PME,			PPSMC_MSG_BacoAudioD3PME),
+	MSG_MAP(ArmD3,				PPSMC_MSG_ArmD3),
 };
 
 static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_clk_map[SMU_CLK_COUNT] = {
@@ -304,6 +305,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_VR0HOT_BIT)
 				| FEATURE_MASK(FEATURE_PPT_BIT)
 				| FEATURE_MASK(FEATURE_TDC_BIT)
+				| FEATURE_MASK(FEATURE_BACO_BIT)
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
 				| FEATURE_MASK(FEATURE_THERMAL_BIT);
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 127/207] drm/amd/powerplay: enable APCC DFLL for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable APCC DFLL for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index e0067921a3e9..2db4b3fb0cf9 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -162,6 +162,7 @@ static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_feature_mask_map[SMU_FEAT
 	FEA_MAP(TEMP_DEPENDENT_VMIN),
 	FEA_MAP(MMHUB_PG),
 	FEA_MAP(ATHUB_PG),
+	FEA_MAP(APCC_DFLL),
 };
 
 static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_table_map[SMU_TABLE_COUNT] = {
@@ -306,6 +307,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_PPT_BIT)
 				| FEATURE_MASK(FEATURE_TDC_BIT)
 				| FEATURE_MASK(FEATURE_BACO_BIT)
+				| FEATURE_MASK(FEATURE_APCC_DFLL_BIT)
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
 				| FEATURE_MASK(FEATURE_THERMAL_BIT);
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 128/207] drm/amd/powerplay: add function to get power limit for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add function to get pptable power limit for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 2db4b3fb0cf9..2843e1be4289 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -1501,6 +1501,12 @@ static int sienna_cichlid_display_disable_memory_clock_switch(struct smu_context
 	return ret;
 }
 
+static uint32_t sienna_cichlid_get_pptable_power_limit(struct smu_context *smu)
+{
+	PPTable_t *pptable = smu->smu_table.driver_pptable;
+	return pptable->SocketPowerLimitAc[PPT_THROTTLER_PPT0];
+}
+
 static int sienna_cichlid_get_power_limit(struct smu_context *smu,
 				     uint32_t *limit,
 				     bool cap)
@@ -2335,6 +2341,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.get_dpm_ultimate_freq = smu_v11_0_get_dpm_ultimate_freq,
 	.set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range,
 	.override_pcie_parameters = smu_v11_0_override_pcie_parameters,
+	.get_pptable_power_limit = sienna_cichlid_get_pptable_power_limit,
 };
 
 void sienna_cichlid_set_ppt_funcs(struct smu_context *smu)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 129/207] drm/amdgpu/mes: update mes fw api

[Alex Deucher]

From: Le Ma <le.ma@amd.com>

Update mes_api_def.h to match the latest mes fw.

v2: clean up coding style based on kernel standards:
  - fix indentation and alignment
  - break long lines
  - put the opening brace last on the line
  - remove unnecessary blank line and space
  - replace uint(32|64) with standard uint(32|64)_t

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/mes_api_def.h | 644 ++++++++++++-----------
 1 file changed, 341 insertions(+), 303 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/mes_api_def.h b/drivers/gpu/drm/amd/amdgpu/mes_api_def.h
index da040845de00..3f4fca5fd1da 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_api_def.h
+++ b/drivers/gpu/drm/amd/amdgpu/mes_api_def.h
@@ -26,95 +26,88 @@
 
 #pragma pack(push, 4)
 
-typedef uint32_t uint32;
-typedef uint64_t uint64;
-
 #define MES_API_VERSION 1
 
-//Driver submits one API(cmd) as a single Frame and this command size is same for all API
-//to ease the debugging and parsing of ring buffer.
-enum {API_FRAME_SIZE_IN_DWORDS = 64};
+/* Driver submits one API(cmd) as a single Frame and this command size is same
+ * for all API to ease the debugging and parsing of ring buffer.
+ */
+enum { API_FRAME_SIZE_IN_DWORDS = 64 };
 
-//To avoid command in scheduler context to be overwritten whenenver mutilple interrupts come in, 
-//this creates another queue
-enum {API_NUMBER_OF_COMMAND_MAX   = 32};
+/* To avoid command in scheduler context to be overwritten whenenver mutilple
+ * interrupts come in, this creates another queue.
+ */
+enum { API_NUMBER_OF_COMMAND_MAX = 32 };
 
-enum MES_API_TYPE
-{
-    MES_API_TYPE_SCHEDULER = 1,
-    MES_API_TYPE_MAX
+enum MES_API_TYPE {
+	MES_API_TYPE_SCHEDULER = 1,
+	MES_API_TYPE_MAX
 };
 
-enum MES_SCH_API_OPCODE
-{
-    MES_SCH_API_SET_HW_RSRC                  = 0,
-    MES_SCH_API_SET_SCHEDULING_CONFIG        = 1, //agreegated db, quantums, etc
-    MES_SCH_API_ADD_QUEUE                    = 2,
-    MES_SCH_API_REMOVE_QUEUE                 = 3,
-    MES_SCH_API_PERFORM_YIELD                = 4,
-    MES_SCH_API_SET_GANG_PRIORITY_LEVEL      = 5, //For windows GANG = Context
-    MES_SCH_API_SUSPEND                      = 6,
-    MES_SCH_API_RESUME                       = 7,
-    MES_SCH_API_RESET                        = 8,
-    MES_SCH_API_SET_LOG_BUFFER               = 9,
-    MES_SCH_API_CHANGE_GANG_PRORITY          = 10,
-    MES_SCH_API_QUERY_SCHEDULER_STATUS       = 11,
-    MES_SCH_API_PROGRAM_GDS                  = 12,
-    MES_SCH_API_MAX                          = 0xFF
+enum MES_SCH_API_OPCODE {
+	MES_SCH_API_SET_HW_RSRC			= 0,
+	MES_SCH_API_SET_SCHEDULING_CONFIG	= 1, /* agreegated db, quantums, etc */
+	MES_SCH_API_ADD_QUEUE			= 2,
+	MES_SCH_API_REMOVE_QUEUE		= 3,
+	MES_SCH_API_PERFORM_YIELD		= 4,
+	MES_SCH_API_SET_GANG_PRIORITY_LEVEL	= 5,
+	MES_SCH_API_SUSPEND			= 6,
+	MES_SCH_API_RESUME			= 7,
+	MES_SCH_API_RESET			= 8,
+	MES_SCH_API_SET_LOG_BUFFER		= 9,
+	MES_SCH_API_CHANGE_GANG_PRORITY		= 10,
+	MES_SCH_API_QUERY_SCHEDULER_STATUS	= 11,
+	MES_SCH_API_PROGRAM_GDS			= 12,
+	MES_SCH_API_SET_DEBUG_VMID		= 13,
+	MES_SCH_API_MISC			= 14,
+	MES_SCH_API_MAX				= 0xFF
 };
 
-union MES_API_HEADER
-{
-    struct
-    {
-        uint32 type	: 4; // 0 - Invalid; 1 - Scheduling; 2 - TBD
-        uint32 opcode   : 8;
-        uint32 dwsize   : 8; //including header
-        uint32 reserved : 12;
-    };
-
-    uint32 u32All;
+union MES_API_HEADER {
+	struct {
+		uint32_t type		: 4; /* 0 - Invalid; 1 - Scheduling; 2 - TBD */
+		uint32_t opcode		: 8;
+		uint32_t dwsize		: 8; /* including header */
+		uint32_t reserved	: 12;
+	};
+
+	uint32_t	u32All;
 };
 
-enum MES_AMD_PRIORITY_LEVEL
-{
-    AMD_PRIORITY_LEVEL_LOW       = 0,
-    AMD_PRIORITY_LEVEL_NORMAL    = 1,
-    AMD_PRIORITY_LEVEL_MEDIUM    = 2,
-    AMD_PRIORITY_LEVEL_HIGH      = 3,
-    AMD_PRIORITY_LEVEL_REALTIME  = 4,
-    AMD_PRIORITY_NUM_LEVELS
+enum MES_AMD_PRIORITY_LEVEL {
+	AMD_PRIORITY_LEVEL_LOW		= 0,
+	AMD_PRIORITY_LEVEL_NORMAL	= 1,
+	AMD_PRIORITY_LEVEL_MEDIUM	= 2,
+	AMD_PRIORITY_LEVEL_HIGH		= 3,
+	AMD_PRIORITY_LEVEL_REALTIME	= 4,
+	AMD_PRIORITY_NUM_LEVELS
 };
 
-enum MES_QUEUE_TYPE
-{
-    MES_QUEUE_TYPE_GFX,
-    MES_QUEUE_TYPE_COMPUTE,
-    MES_QUEUE_TYPE_SDMA,
-    MES_QUEUE_TYPE_MAX,
+enum MES_QUEUE_TYPE {
+	MES_QUEUE_TYPE_GFX,
+	MES_QUEUE_TYPE_COMPUTE,
+	MES_QUEUE_TYPE_SDMA,
+	MES_QUEUE_TYPE_MAX,
 };
 
-struct MES_API_STATUS
-{
-    uint64 api_completion_fence_addr;
-    uint64 api_completion_fence_value;
+struct MES_API_STATUS {
+	uint64_t	api_completion_fence_addr;
+	uint64_t	api_completion_fence_value;
 };
 
 enum { MAX_COMPUTE_PIPES = 8 };
-enum { MAX_GFX_PIPES	 = 2 };
-enum { MAX_SDMA_PIPES    = 2 };
+enum { MAX_GFX_PIPES = 2 };
+enum { MAX_SDMA_PIPES = 2 };
 
-enum { MAX_COMPUTE_HQD_PER_PIPE     = 8 };
-enum { MAX_GFX_HQD_PER_PIPE         = 8 };
-enum { MAX_SDMA_HQD_PER_PIPE        = 10 };
+enum { MAX_COMPUTE_HQD_PER_PIPE = 8 };
+enum { MAX_GFX_HQD_PER_PIPE = 8 };
+enum { MAX_SDMA_HQD_PER_PIPE = 10 };
 
 enum { MAX_QUEUES_IN_A_GANG = 8 };
 
-enum VM_HUB_TYPE
-{
-    VM_HUB_TYPE_GC = 0,
-    VM_HUB_TYPE_MM = 1,
-    VM_HUB_TYPE_MAX,
+enum VM_HUB_TYPE {
+	VM_HUB_TYPE_GC = 0,
+	VM_HUB_TYPE_MM = 1,
+	VM_HUB_TYPE_MAX,
 };
 
 enum { VMID_INVALID = 0xffff };
@@ -122,283 +115,328 @@ enum { VMID_INVALID = 0xffff };
 enum { MAX_VMID_GCHUB = 16 };
 enum { MAX_VMID_MMHUB = 16 };
 
-enum MES_LOG_OPERATION
-{
-    MES_LOG_OPERATION_CONTEXT_STATE_CHANGE = 0
+enum MES_LOG_OPERATION {
+	MES_LOG_OPERATION_CONTEXT_STATE_CHANGE = 0
 };
 
-enum MES_LOG_CONTEXT_STATE
-{
-    MES_LOG_CONTEXT_STATE_IDLE = 0,
-    MES_LOG_CONTEXT_STATE_RUNNING = 1,
-    MES_LOG_CONTEXT_STATE_READY = 2,
-    MES_LOG_CONTEXT_STATE_READY_STANDBY = 3,
+enum MES_LOG_CONTEXT_STATE {
+	MES_LOG_CONTEXT_STATE_IDLE		= 0,
+	MES_LOG_CONTEXT_STATE_RUNNING		= 1,
+	MES_LOG_CONTEXT_STATE_READY		= 2,
+	MES_LOG_CONTEXT_STATE_READY_STANDBY	= 3,
 };
 
-struct MES_LOG_CONTEXT_STATE_CHANGE
-{
-    void*                   h_context;
-    enum MES_LOG_CONTEXT_STATE   new_context_state;
+struct MES_LOG_CONTEXT_STATE_CHANGE {
+	void				*h_context;
+	enum MES_LOG_CONTEXT_STATE	new_context_state;
 };
 
-struct MES_LOG_ENTRY_HEADER
-{
-    uint32 first_free_entry_index;
-    uint32 wraparound_count;
-    uint64 number_of_entries;
-    uint64 reserved[2];
+struct MES_LOG_ENTRY_HEADER {
+	uint32_t	first_free_entry_index;
+	uint32_t	wraparound_count;
+	uint64_t	number_of_entries;
+	uint64_t	reserved[2];
 };
 
-struct MES_LOG_ENTRY_DATA
-{
-    uint64 gpu_time_stamp;
-    uint32 operation_type; //operation_type is of MES_LOG_OPERATION type
-    uint32 reserved_operation_type_bits;
-    union
-    {
-        struct MES_LOG_CONTEXT_STATE_CHANGE  context_state_change;
-        uint64                        reserved_operation_data[2];
-    };
+struct MES_LOG_ENTRY_DATA {
+	uint64_t	gpu_time_stamp;
+	uint32_t	operation_type; /* operation_type is of MES_LOG_OPERATION type */
+	uint32_t	reserved_operation_type_bits;
+	union {
+		struct MES_LOG_CONTEXT_STATE_CHANGE	context_state_change;
+		uint64_t				reserved_operation_data[2];
+	};
 };
 
-struct MES_LOG_BUFFER
-{
-    struct MES_LOG_ENTRY_HEADER header;
-    struct MES_LOG_ENTRY_DATA   entries[1];
+struct MES_LOG_BUFFER {
+	struct MES_LOG_ENTRY_HEADER	header;
+	struct MES_LOG_ENTRY_DATA	entries[1];
 };
 
-union MESAPI_SET_HW_RESOURCES
-{
-    struct
-    {
-        union MES_API_HEADER                header;
-        uint32                              vmid_mask_mmhub;
-        uint32                              vmid_mask_gfxhub;
-        uint32                              gds_size;
-        uint32                              paging_vmid;
-        uint32                              compute_hqd_mask[MAX_COMPUTE_PIPES];
-        uint32                              gfx_hqd_mask[MAX_GFX_PIPES];
-        uint32                              sdma_hqd_mask[MAX_SDMA_PIPES];
-        uint32                              agreegated_doorbells[AMD_PRIORITY_NUM_LEVELS];
-        uint64                              g_sch_ctx_gpu_mc_ptr;
-        struct MES_API_STATUS               api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI_SET_HW_RESOURCES {
+	struct {
+		union MES_API_HEADER	header;
+		uint32_t		vmid_mask_mmhub;
+		uint32_t		vmid_mask_gfxhub;
+		uint32_t		gds_size;
+		uint32_t		paging_vmid;
+		uint32_t		compute_hqd_mask[MAX_COMPUTE_PIPES];
+		uint32_t		gfx_hqd_mask[MAX_GFX_PIPES];
+		uint32_t		sdma_hqd_mask[MAX_SDMA_PIPES];
+		uint32_t		agreegated_doorbells[AMD_PRIORITY_NUM_LEVELS];
+		uint64_t		g_sch_ctx_gpu_mc_ptr;
+		uint64_t		query_status_fence_gpu_mc_ptr;
+		struct MES_API_STATUS	api_status;
+		union {
+			struct {
+				uint32_t disable_reset	: 1;
+				uint32_t reserved	: 31;
+			};
+			uint32_t	uint32_t_all;
+		};
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__ADD_QUEUE
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        uint32                      process_id;
-        uint64                      page_table_base_addr;
-        uint64                      process_va_start;
-        uint64                      process_va_end;
-        uint64                      process_quantum;
-        uint64                      process_context_addr;
-        uint64                      gang_quantum;
-        uint64                      gang_context_addr;
-        uint32                      inprocess_gang_priority;
-        enum MES_AMD_PRIORITY_LEVEL gang_global_priority_level;
-        uint32                      doorbell_offset;
-        uint64                      mqd_addr;
-        uint64                      wptr_addr;
-        enum MES_QUEUE_TYPE         queue_type;
-        uint32                      gds_base;
-        uint32                      gds_size;
-        uint32                      gws_base;
-        uint32                      gws_size;
-        uint32                      oa_mask;
-
-        struct
-        {
-            uint32 paging      : 1;
-            uint32 program_gds : 1;
-            uint32 reserved    : 30;
-        };
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__ADD_QUEUE {
+	struct {
+		union MES_API_HEADER		header;
+		uint32_t			process_id;
+		uint64_t			page_table_base_addr;
+		uint64_t			process_va_start;
+		uint64_t			process_va_end;
+		uint64_t			process_quantum;
+		uint64_t			process_context_addr;
+		uint64_t			gang_quantum;
+		uint64_t			gang_context_addr;
+		uint32_t			inprocess_gang_priority;
+		enum MES_AMD_PRIORITY_LEVEL	gang_global_priority_level;
+		uint32_t			doorbell_offset;
+		uint64_t			mqd_addr;
+		uint64_t			wptr_addr;
+		enum MES_QUEUE_TYPE		queue_type;
+		uint32_t			gds_base;
+		uint32_t			gds_size;
+		uint32_t			gws_base;
+		uint32_t			gws_size;
+		uint32_t			oa_mask;
+
+		struct {
+			uint32_t paging			: 1;
+			uint32_t debug_vmid		: 4;
+			uint32_t program_gds		: 1;
+			uint32_t is_gang_suspended	: 1;
+			uint32_t is_tmz_queue		: 1;
+			uint32_t reserved		: 24;
+		};
+		struct MES_API_STATUS		api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__REMOVE_QUEUE
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        uint32                      doorbell_offset;
-        uint64                      gang_context_addr;
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
+union MESAPI__REMOVE_QUEUE {
+	struct {
+		union MES_API_HEADER	header;
+		uint32_t		doorbell_offset;
+		uint64_t		gang_context_addr;
 
-union MESAPI__SET_SCHEDULING_CONFIG
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        // Grace period when preempting another priority band for this priority band.
-        // The value for idle priority band is ignored, as it never preempts other bands.
-        uint64                      grace_period_other_levels[AMD_PRIORITY_NUM_LEVELS];
+		struct {
+			uint32_t unmap_legacy_gfx_queue	: 1;
+			uint32_t reserved		: 31;
+		};
+		struct MES_API_STATUS	api_status;
+	};
 
-        // Default quantum for scheduling across processes within a priority band.
-        uint64                      process_quantum_for_level[AMD_PRIORITY_NUM_LEVELS];
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+};
 
-        // Default grace period for processes that preempt each other within a priority band.
-        uint64                      process_grace_period_same_level[AMD_PRIORITY_NUM_LEVELS];
+union MESAPI__SET_SCHEDULING_CONFIG {
+	struct {
+		union MES_API_HEADER	header;
+		/* Grace period when preempting another priority band for this
+		 * priority band. The value for idle priority band is ignored,
+		 * as it never preempts other bands.
+		 */
+		uint64_t		grace_period_other_levels[AMD_PRIORITY_NUM_LEVELS];
+		/* Default quantum for scheduling across processes within
+		 * a priority band.
+		 */
+		uint64_t		process_quantum_for_level[AMD_PRIORITY_NUM_LEVELS];
+		/* Default grace period for processes that preempt each other
+		 * within a priority band.
+		 */
+		uint64_t		process_grace_period_same_level[AMD_PRIORITY_NUM_LEVELS];
+		/* For normal level this field specifies the target GPU
+		 * percentage in situations when it's starved by the high level.
+		 * Valid values are between 0 and 50, with the default being 10.
+		 */
+		uint32_t		normal_yield_percent;
+		struct MES_API_STATUS	api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+};
 
-        // For normal level this field specifies the target GPU percentage in situations when it's starved by the high level.
-        // Valid values are between 0 and 50, with the default being 10.
-        uint32                      normal_yield_percent;
+union MESAPI__PERFORM_YIELD {
+	struct {
+		union MES_API_HEADER	header;
+		uint32_t		dummy;
+		struct MES_API_STATUS	api_status;
+	};
 
-        struct MES_API_STATUS       api_status;
-    };
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+};
 
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__CHANGE_GANG_PRIORITY_LEVEL {
+	struct {
+		union MES_API_HEADER		header;
+		uint32_t			inprocess_gang_priority;
+		enum MES_AMD_PRIORITY_LEVEL	gang_global_priority_level;
+		uint64_t			gang_quantum;
+		uint64_t			gang_context_addr;
+		struct MES_API_STATUS		api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__PERFORM_YIELD
-{
-    struct
-    {
-        union MES_API_HEADER    header;
-        uint32                  dummy;
-        struct MES_API_STATUS   api_status;
-    };
+union MESAPI__SUSPEND {
+	struct {
+		union MES_API_HEADER	header;
+		/* false - suspend all gangs; true - specific gang */
+		struct {
+			uint32_t suspend_all_gangs	: 1;
+			uint32_t reserved		: 31;
+		};
+		/* gang_context_addr is valid only if suspend_all = false */
+		uint64_t		gang_context_addr;
 
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
+		uint64_t		suspend_fence_addr;
+		uint32_t		suspend_fence_value;
+
+		struct MES_API_STATUS	api_status;
+	};
 
-union MESAPI__CHANGE_GANG_PRIORITY_LEVEL
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        uint32                      inprocess_gang_priority;
-        enum MES_AMD_PRIORITY_LEVEL gang_global_priority_level;
-        uint64                      gang_quantum;
-        uint64                      gang_context_addr;
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__SUSPEND
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        //false - suspend all gangs; true - specific gang
-        struct
-        {
-            uint32                  suspend_all_gangs : 1;
-            uint32                  reserved : 31;
-        };
-        //gang_context_addr is valid only if suspend_all = false
-        uint64                      gang_context_addr;
-
-        uint64                      suspend_fence_addr;
-        uint32                      suspend_fence_value;
-
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__RESUME {
+	struct {
+		union MES_API_HEADER	header;
+		/* false - resume all gangs; true - specified gang */
+		struct {
+			uint32_t resume_all_gangs	: 1;
+			uint32_t reserved		: 31;
+		};
+		/* valid only if resume_all_gangs = false */
+		uint64_t		gang_context_addr;
+
+		struct MES_API_STATUS	api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__RESUME
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        //false - resume all gangs; true - specified gang
-        struct
-        {
-            uint32                  resume_all_gangs : 1;
-            uint32                  reserved : 31;
-        };
-        //valid only if resume_all_gangs = false
-        uint64                      gang_context_addr;
-
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__RESET {
+	struct {
+		union MES_API_HEADER	header;
+
+		struct {
+			uint32_t reset_queue	: 1;
+			uint32_t reserved	: 31;
+		};
+
+		uint64_t		gang_context_addr;
+		uint32_t		doorbell_offset; /* valid only if reset_queue = true */
+		struct MES_API_STATUS	api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__RESET
-{
-    struct
-    {
-        union MES_API_HEADER        header;
+union MESAPI__SET_LOGGING_BUFFER {
+	struct {
+		union MES_API_HEADER	header;
+		/* There are separate log buffers for each queue type */
+		enum MES_QUEUE_TYPE	log_type;
+		/* Log buffer GPU Address */
+		uint64_t		logging_buffer_addr;
+		/* number of entries in the log buffer */
+		uint32_t		number_of_entries;
+		/* Entry index at which CPU interrupt needs to be signalled */
+		uint32_t		interrupt_entry;
+
+		struct MES_API_STATUS	api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+};
 
-        struct
-        {
-            uint32                  reset_queue : 1;
-            uint32                  reserved : 31;
-        };
+union MESAPI__QUERY_MES_STATUS {
+	struct {
+		union MES_API_HEADER	header;
+		bool			mes_healthy; /* 0 - not healthy, 1 - healthy */
+		struct MES_API_STATUS	api_status;
+	};
 
-        uint64                      gang_context_addr;
-        uint32                      doorbell_offset; //valid only if reset_queue = true
-        struct MES_API_STATUS       api_status;
-    };
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+};
 
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__PROGRAM_GDS {
+	struct {
+		union MES_API_HEADER	header;
+		uint64_t		process_context_addr;
+		uint32_t		gds_base;
+		uint32_t		gds_size;
+		uint32_t		gws_base;
+		uint32_t		gws_size;
+		uint32_t		oa_mask;
+		struct MES_API_STATUS	api_status;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__SET_LOGGING_BUFFER
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        //There are separate log buffers for each queue type
-        enum MES_QUEUE_TYPE         log_type;
-        //Log buffer GPU Address
-        uint64                      logging_buffer_addr;
-        //number of entries in the log buffer
-        uint32                      number_of_entries;
-        //Entry index at which CPU interrupt needs to be signalled
-        uint32                      interrupt_entry;
-
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+union MESAPI__SET_DEBUG_VMID {
+	struct {
+		union MES_API_HEADER	header;
+		struct MES_API_STATUS	api_status;
+		union {
+			struct {
+				uint32_t use_gds	: 1;
+				uint32_t reserved	: 31;
+			} flags;
+			uint32_t	u32All;
+		};
+		uint32_t		reserved;
+		uint32_t		debug_vmid;
+		uint64_t		process_context_addr;
+		uint64_t		page_table_base_addr;
+		uint64_t		process_va_start;
+		uint64_t		process_va_end;
+		uint32_t		gds_base;
+		uint32_t		gds_size;
+		uint32_t		gws_base;
+		uint32_t		gws_size;
+		uint32_t		oa_mask;
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
-union MESAPI__QUERY_MES_STATUS
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        bool                        mes_healthy; //0 - not healthy, 1 - healthy
-        struct MES_API_STATUS       api_status;
-    };
+enum MESAPI_MISC_OPCODE {
+	MESAPI_MISC__MODIFY_REG,
+	MESAPI_MISC__MAX,
+};
 
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+enum MODIFY_REG_SUBCODE {
+	MODIFY_REG__OVERWRITE,
+	MODIFY_REG__RMW_OR,
+	MODIFY_REG__RMW_AND,
+	MODIFY_REG__MAX,
 };
 
-union MESAPI__PROGRAM_GDS
-{
-    struct
-    {
-        union MES_API_HEADER        header;
-        uint64                      process_context_addr;
-        uint32                      gds_base;
-        uint32                      gds_size;
-        uint32                      gws_base;
-        uint32                      gws_size;
-        uint32                      oa_mask;
-        struct MES_API_STATUS       api_status;
-    };
-
-    uint32 max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
+enum { MISC_DATA_MAX_SIZE_IN_DWORDS = 20 };
+
+union MESAPI__MISC {
+	struct {
+		union MES_API_HEADER	header;
+		enum MESAPI_MISC_OPCODE	opcode;
+		struct MES_API_STATUS	api_status;
+
+		union {
+			struct {
+				enum MODIFY_REG_SUBCODE	subcode;
+				uint32_t		reg_offset;
+				uint32_t		reg_value;
+			} modify_reg;
+			uint32_t	data[MISC_DATA_MAX_SIZE_IN_DWORDS];
+		};
+	};
+
+	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
 };
 
 #pragma pack(pop)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 130/207] drm/amdgpu/mes: add status fence memory definitions

[Alex Deucher]

From: Le Ma <le.ma@amd.com>

Update for new member query_status_fence_gpu_mc_ptr in MESAPI_SET_HW_RESOURCES.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h | 3 +++
 drivers/gpu/drm/amd/amdgpu/mes_v10_1.c  | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
index 4da9d8f50578..7334982ea702 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
@@ -82,6 +82,9 @@ struct amdgpu_mes {
 	uint32_t                        sch_ctx_offs;
 	uint64_t			sch_ctx_gpu_addr;
 	uint64_t			*sch_ctx_ptr;
+	uint32_t			query_status_fence_offs;
+	uint64_t			query_status_fence_gpu_addr;
+	uint64_t			*query_status_fence_ptr;
 
 	/* ip specific functions */
 	const struct amdgpu_mes_funcs   *funcs;
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
index 447bee159089..6ba0c04f7fb5 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
@@ -237,6 +237,8 @@ static int mes_v10_1_set_hw_resources(struct amdgpu_mes *mes)
 	mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
 	mes_set_hw_res_pkt.paging_vmid = 0;
 	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr;
+	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
+		mes->query_status_fence_gpu_addr;
 
 	for (i = 0; i < MAX_COMPUTE_PIPES; i++)
 		mes_set_hw_res_pkt.compute_hqd_mask[i] =
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 131/207] drm/amdgpu/mes: allocate memory slots for hw resource setting

[Alex Deucher]

From: Le Ma <le.ma@amd.com>

Pass a piece of memory to MES ucode to fill contents.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/mes_v10_1.c | 36 ++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
index 6ba0c04f7fb5..5c28868f7adc 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
@@ -562,6 +562,35 @@ static int mes_v10_1_allocate_eop_buf(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int mes_v10_1_allocate_mem_slots(struct amdgpu_device *adev)
+{
+	int r;
+
+	r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs);
+	if (r) {
+		dev_err(adev->dev,
+			"(%d) mes sch_ctx_offs wb alloc failed\n", r);
+		return r;
+	}
+	adev->mes.sch_ctx_gpu_addr =
+		adev->wb.gpu_addr + (adev->mes.sch_ctx_offs * 4);
+	adev->mes.sch_ctx_ptr =
+		(uint64_t *)&adev->wb.wb[adev->mes.sch_ctx_offs];
+
+	r = amdgpu_device_wb_get(adev, &adev->mes.query_status_fence_offs);
+	if (r) {
+		dev_err(adev->dev,
+			"(%d) query_status_fence_offs wb alloc failed\n", r);
+		return r;
+	}
+	adev->mes.query_status_fence_gpu_addr =
+		adev->wb.gpu_addr + (adev->mes.query_status_fence_offs * 4);
+	adev->mes.query_status_fence_ptr =
+		(uint64_t *)&adev->wb.wb[adev->mes.query_status_fence_offs];
+
+	return 0;
+}
+
 static int mes_v10_1_mqd_init(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
@@ -876,6 +905,10 @@ static int mes_v10_1_sw_init(void *handle)
 	if (r)
 		return r;
 
+	r = mes_v10_1_allocate_mem_slots(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -883,6 +916,9 @@ static int mes_v10_1_sw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
+	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
+
 	kfree(adev->mes.mqd_backup);
 
 	amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj,
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 132/207] drm/amdgpu: skip VM inv eng assignment for mes ring

[Alex Deucher]

From: Le Ma <le.ma@amd.com>

Statically allocated VM inv eng of gfxhub on sienna_cichlid is used up.
Also VM inv eng is no need for mes ring.

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index acabb57aa8af..34cbd6f6a56b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -357,6 +357,9 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 		ring = adev->rings[i];
 		vmhub = ring->funcs->vmhub;
 
+		if (ring == &adev->mes.ring)
+			continue;
+
 		inv_eng = ffs(vm_inv_engs[vmhub]);
 		if (!inv_eng) {
 			dev_err(adev->dev, "no VM inv eng for ring %s\n",
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 133/207] drm/amd/powerplay: enable athub pg

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

enable athub pg and the status can be checked in
ATHUB_MISC_CNTL.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c                    | 3 ++-
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index f4d4d91ebee2..6ca4c6f043c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -725,7 +725,8 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_IH_CG |
 			AMD_CG_SUPPORT_MC_LS;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
-			AMD_PG_SUPPORT_JPEG;
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_ATHUB;
 		adev->external_rev_id = adev->rev_id + 0x28;
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 2843e1be4289..f7e67c446ae0 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -332,6 +332,9 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	if (adev->pm.pp_feature & PP_SCLK_DEEP_SLEEP_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DS_GFXCLK_BIT);
 
+	if (smu->adev->pg_flags & AMD_PG_SUPPORT_ATHUB)
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_ATHUB_PG_BIT);
+
 	if (smu->adev->pg_flags & AMD_PG_SUPPORT_VCN ||
 	    smu->adev->pg_flags & AMD_PG_SUPPORT_JPEG)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_MM_DPM_PG_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 134/207] drm/amd/powerplay: enable mmhub pg

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

mmhub pg can be obvserved from PCTL_CTRL

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c                    | 3 ++-
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 6ca4c6f043c6..936950fbafda 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -704,7 +704,8 @@ static int nv_common_early_init(void *handle)
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_VCN_DPG |
 			AMD_PG_SUPPORT_JPEG |
-			AMD_PG_SUPPORT_ATHUB;
+			AMD_PG_SUPPORT_ATHUB |
+			AMD_PG_SUPPORT_MMHUB;
 		/* guest vm gets 0xffffffff when reading RCC_DEV0_EPF0_STRAP0,
 		 * as a consequence, the rev_id and external_rev_id are wrong.
 		 * workaround it by hardcoding rev_id to 0 (default value).
diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index f7e67c446ae0..8fb08cae5c24 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -335,6 +335,9 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	if (smu->adev->pg_flags & AMD_PG_SUPPORT_ATHUB)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_ATHUB_PG_BIT);
 
+	if (smu->adev->pg_flags & AMD_PG_SUPPORT_MMHUB)
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_MMHUB_PG_BIT);
+
 	if (smu->adev->pg_flags & AMD_PG_SUPPORT_VCN ||
 	    smu->adev->pg_flags & AMD_PG_SUPPORT_JPEG)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_MM_DPM_PG_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 135/207] drm/amd/powerplay: enable GPO

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

GPO is graphics power optimizer.
SMU calculates the 16 gfxclk V/F points according to the CU numbers
and memory activity.RLC picks one of them according to the memory
speed requirements for the data transmission.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/inc/smu_types.h      | 1 +
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
index 3efadf2cff9c..dff2295705be 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h
@@ -260,6 +260,7 @@ enum smu_clk_type {
        __SMU_DUMMY_MAP(MMHUB_PG),                      	\
        __SMU_DUMMY_MAP(ATHUB_PG),                      	\
        __SMU_DUMMY_MAP(APCC_DFLL),                     	\
+       __SMU_DUMMY_MAP(DPM_GFX_GPO),                    \
        __SMU_DUMMY_MAP(WAFL_CG),
 
 #undef __SMU_DUMMY_MAP
diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 8fb08cae5c24..9dbf29e96312 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -126,6 +126,7 @@ static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_clk_map[SMU_CLK_COUNT] =
 static struct smu_11_0_cmn2aisc_mapping sienna_cichlid_feature_mask_map[SMU_FEATURE_COUNT] = {
 	FEA_MAP(DPM_PREFETCHER),
 	FEA_MAP(DPM_GFXCLK),
+	FEA_MAP(DPM_GFX_GPO),
 	FEA_MAP(DPM_UCLK),
 	FEA_MAP(DPM_SOCCLK),
 	FEA_MAP(DPM_MP0CLK),
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 136/207] drm/amd/powerplay: bundle GPO with gfx DPM

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

Bundle GPO with gfx DPM and enable it since gfxclk dpm
should work first then GPO works.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 9dbf29e96312..45707772e755 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -312,8 +312,10 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
 				| FEATURE_MASK(FEATURE_THERMAL_BIT);
 
-	if (adev->pm.pp_feature & PP_SCLK_DPM_MASK)
+	if (adev->pm.pp_feature & PP_SCLK_DPM_MASK) {
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_GFXCLK_BIT);
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_GFX_GPO_BIT);
+	}
 
 	if (adev->pm.pp_feature & PP_MCLK_DPM_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_UCLK_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 137/207] drm/amdgpu: update golden setting for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Update golden setting for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 6c52363d5662..8935b9b81d68 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3104,7 +3104,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER8_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER9_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00800000)
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00a00000)
 };
 
 static const struct soc15_reg_golden golden_settings_gc_10_3_sienna_cichlid[] =
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 138/207] drm/amd/powerplay: enable RSMU SMN PG for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable RSMU SMN PG for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 45707772e755..9ac562b7c5cf 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -303,6 +303,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_DS_FCLK_BIT)
 				| FEATURE_MASK(FEATURE_FW_DSTATE_BIT)
 				| FEATURE_MASK(FEATURE_DF_CSTATE_BIT)
+				| FEATURE_MASK(FEATURE_RSMU_SMN_CG_BIT)
 				| FEATURE_MASK(FEATURE_GFX_SS_BIT)
 				| FEATURE_MASK(FEATURE_VR0HOT_BIT)
 				| FEATURE_MASK(FEATURE_PPT_BIT)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 139/207] drm/am/powerplay: enable OUT OF BAND MONITER for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable OUT OF BAND MONITER for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 9ac562b7c5cf..545b4efc2bfa 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -311,7 +311,8 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_BACO_BIT)
 				| FEATURE_MASK(FEATURE_APCC_DFLL_BIT)
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
-				| FEATURE_MASK(FEATURE_THERMAL_BIT);
+				| FEATURE_MASK(FEATURE_THERMAL_BIT)
+				| FEATURE_MASK(FEATURE_OUT_OF_BAND_MONITOR_BIT);
 
 	if (adev->pm.pp_feature & PP_SCLK_DPM_MASK) {
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_GFXCLK_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 140/207] drm/amdgpu: add clock gating DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add vcn_v3_0_clock_gating_dpg_mode to enabling clock gating in DPG mode for VCN3.0

V2: Separate from previous patch-0002, and update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 48 +++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 340cb344a35d..b0fc589ac88f 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -613,6 +613,54 @@ static void vcn_v3_0_disable_clock_gating(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data);
 }
 
+static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
+		uint8_t sram_sel, int inst_idx, uint8_t indirect)
+{
+	uint32_t reg_data = 0;
+
+	/* enable sw clock gating control */
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		reg_data = 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	reg_data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_CM_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_IT_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_DB_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MP_MODE_MASK |
+		 UVD_CGC_CTRL__SYS_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MODE_MASK |
+		 UVD_CGC_CTRL__MPEG2_MODE_MASK |
+		 UVD_CGC_CTRL__REGS_MODE_MASK |
+		 UVD_CGC_CTRL__RBC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_MC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
+		 UVD_CGC_CTRL__IDCT_MODE_MASK |
+		 UVD_CGC_CTRL__MPRD_MODE_MASK |
+		 UVD_CGC_CTRL__MPC_MODE_MASK |
+		 UVD_CGC_CTRL__LBSI_MODE_MASK |
+		 UVD_CGC_CTRL__LRBBM_MODE_MASK |
+		 UVD_CGC_CTRL__WCB_MODE_MASK |
+		 UVD_CGC_CTRL__VCPU_MODE_MASK |
+		 UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_CGC_CTRL), reg_data, sram_sel, indirect);
+
+	/* turn off clock gating */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_CGC_GATE), 0, sram_sel, indirect);
+
+	/* turn on SUVD clock gating */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
+
+	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
+}
+
 /**
  * vcn_v3_0_enable_clock_gating - enable VCN clock gating
  *
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 141/207] drm/amdgpu: add mc resume DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add vcn_v3_0_mc_resume_dpg_mode to resume memory controller in DPG mode for VCN3.0

V2: Separate from previous patch-0002, and update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 89 +++++++++++++++++++++++++++
 1 file changed, 89 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index b0fc589ac88f..86db36537371 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -378,6 +378,95 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
 }
 
+static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	uint32_t offset;
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo), 0, indirect);
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi), 0, indirect);
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+				VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			VCN, inst_idx, mmUVD_VCPU_NONCACHE_SIZE0), 0, 0, indirect);
+}
+
 static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int inst)
 {
 	uint32_t data = 0;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 142/207] drm/amdgpu: add start DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add vcn_v3_0_start_dpg_mode to setup and start VCN block in DPG mode for VCN3.0

V2: Separate from previous patch-0002, and update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 141 ++++++++++++++++++++++++++
 1 file changed, 141 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 86db36537371..5578af5cb399 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -818,6 +818,142 @@ static void vcn_v3_0_enable_clock_gating(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data);
 }
 
+static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+	struct amdgpu_ring *ring;
+	uint32_t rb_bufsz, tmp;
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 1,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, inst_idx, mmUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, mmUVD_POWER_STATUS, tmp);
+
+	if (indirect)
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t*)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+
+	/* enable clock gating */
+	vcn_v3_0_clock_gating_dpg_mode(adev, 0, inst_idx, indirect);
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	tmp |= UVD_VCPU_CNTL__BLK_RST_MASK;
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interupt */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup mmUVD_LMI_CTRL */
+	tmp = (0x8 | UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_LMI_CTRL), tmp, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MPC_CNTL),
+		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MPC_SET_MUXA0),
+		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MPC_SET_MUXB0),
+		 ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MPC_SET_MUX),
+		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
+
+	vcn_v3_0_mc_resume_dpg_mode(adev, inst_idx, indirect);
+
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_RBC_XX_IB_REG_CHECK), 0x3, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_LMI_CTRL2), 0, 0, indirect);
+
+	/* unblock VCPU register access */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_RB_ARB_CTRL), 0, 0, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+	/* add nop to workaround PSP size check */
+	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	if (indirect)
+		psp_update_vcn_sram(adev, inst_idx, adev->vcn.inst[inst_idx].dpg_sram_gpu_addr,
+			(uint32_t)((uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_curr_addr -
+				(uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr));
+
+	ring = &adev->vcn.inst[inst_idx].ring_dec;
+	/* force RBC into idle state */
+	rb_bufsz = order_base_2(ring->ring_size);
+	tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
+	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_CNTL, tmp);
+
+	/* set the write pointer delay */
+	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR_CNTL, 0);
+
+	/* set the wb address */
+	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_RPTR_ADDR,
+		(upper_32_bits(ring->gpu_addr) >> 2));
+
+	/* programm the RB_BASE for ring buffer */
+	WREG32_SOC15(VCN, inst_idx, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, inst_idx, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_RPTR, 0);
+
+	WREG32_SOC15(VCN, inst_idx, mmUVD_SCRATCH2, 0);
+
+	ring->wptr = RREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_RPTR);
+	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR,
+		lower_32_bits(ring->wptr));
+
+	return 0;
+}
+
 static int vcn_v3_0_start(struct amdgpu_device *adev)
 {
 	struct amdgpu_ring *ring;
@@ -831,6 +967,11 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG){
+			r = vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram);
+			continue;
+		}
+
 		/* disable VCN power gating */
 		vcn_v3_0_disable_static_power_gating(adev, i);
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 143/207] drm/amdgpu: add stop DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add vcn_v3_0_stop_dpg_mode to power off in DPG mode for VCN3.0

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 34 +++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 5578af5cb399..c1aaa9423f5d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -1128,6 +1128,35 @@ static int vcn_v3_0_start(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+{
+	int ret_code = 0;
+	uint32_t tmp;
+
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_RB_RPTR, tmp, 0xFFFFFFFF, ret_code);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR2);
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_RB_RPTR2, tmp, 0xFFFFFFFF, ret_code);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR) & 0x7FFFFFFF;
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_RBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code);
+
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	return 0;
+}
+
 static int vcn_v3_0_stop(struct amdgpu_device *adev)
 {
 	uint32_t tmp;
@@ -1137,6 +1166,11 @@ static int vcn_v3_0_stop(struct amdgpu_device *adev)
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+			r = vcn_v3_0_stop_dpg_mode(adev, i);
+			continue;
+		}
+
 		/* wait for vcn idle */
 		SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, r);
 		if (r)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 144/207] drm/amdgpu: add pause DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add vcn_v3_0_pause_dpg_mode to pause/unpause DPG mode for VCN3.0

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 66 +++++++++++++++++++++++++++
 1 file changed, 66 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index c1aaa9423f5d..f47136d9bfe8 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -53,6 +53,8 @@ static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev);
 static int vcn_v3_0_set_powergating_state(void *handle,
 			enum amd_powergating_state state);
+static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
+			int inst_idx, struct dpg_pause_state *new_state);
 
 static int amdgpu_ih_clientid_vcns[] = {
 	SOC15_IH_CLIENTID_VCN,
@@ -189,6 +191,9 @@ static int vcn_v3_0_sw_init(void *handle)
 		}
 	}
 
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		adev->vcn.pause_dpg_mode = vcn_v3_0_pause_dpg_mode;
+
 	return 0;
 }
 
@@ -1232,6 +1237,67 @@ static int vcn_v3_0_stop(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
+		   int inst_idx, struct dpg_pause_state *new_state)
+{
+	struct amdgpu_ring *ring;
+	uint32_t reg_data = 0;
+	int ret_code;
+
+	/* pause/unpause if state is changed */
+	if (adev->vcn.inst[inst_idx].pause_state.fw_based != new_state->fw_based) {
+		DRM_DEBUG("dpg pause state changed %d -> %d",
+			adev->vcn.inst[inst_idx].pause_state.fw_based,	new_state->fw_based);
+		reg_data = RREG32_SOC15(VCN, inst_idx, mmUVD_DPG_PAUSE) &
+			(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+		if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+			ret_code = 0;
+			SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS, 0x1,
+				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
+
+			if (!ret_code) {
+				/* pause DPG */
+				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+				WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_PAUSE, reg_data);
+
+				/* wait for ACK */
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_DPG_PAUSE,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK, ret_code);
+
+				/* Restore */
+				ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO, ring->gpu_addr);
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE, ring->ring_size / 4);
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+
+				ring = &adev->vcn.inst[inst_idx].ring_enc[1];
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE2, ring->ring_size / 4);
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+
+				WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR,
+					RREG32_SOC15(VCN, inst_idx, mmUVD_SCRATCH2) & 0x7FFFFFFF);
+
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS,
+					UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
+			}
+		} else {
+			/* unpause dpg, no need to wait */
+			reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_PAUSE, reg_data);
+		}
+		adev->vcn.inst[inst_idx].pause_state.fw_based = new_state->fw_based;
+	}
+
+	return 0;
+}
+
 /**
  * vcn_v3_0_dec_ring_get_rptr - get read pointer
  *
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 145/207] drm/amdgpu: set indirect sram mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Use indirect sram for secure DPG mode

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index fff77b629d04..15ff30c53e24 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -111,6 +111,9 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 		break;
 	case CHIP_SIENNA_CICHLID:
 		fw_name = FIRMWARE_SIENNA_CICHLID;
+		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
+		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
+			adev->vcn.indirect_sram = true;
 		break;
 	default:
 		return -EINVAL;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 146/207] drm/amdgpu: add internal reg offset translation for VCN inst 1

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Add range for vcn instance 1 for translation for internal register offset, which
is needed for VCN3.0

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 7a2d5f8d1247..25a07704cf75 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -53,7 +53,9 @@
 #define VCN_ENC_CMD_REG_WAIT		0x0000000c
 
 #define VCN_VID_SOC_ADDRESS_2_0 	0x1fa00
+#define VCN1_VID_SOC_ADDRESS_3_0 	0x48200
 #define VCN_AON_SOC_ADDRESS_2_0 	0x1f800
+#define VCN1_AON_SOC_ADDRESS_3_0 	0x48000
 #define VCN_VID_IP_ADDRESS_2_0		0x0
 #define VCN_AON_IP_ADDRESS_2_0		0x30000
 
@@ -89,20 +91,30 @@
 #define SOC15_DPG_MODE_OFFSET_2_0(ip, inst_idx, reg) 						\
 	({											\
 		uint32_t internal_reg_offset, addr;						\
-		bool video_range, aon_range;							\
+		bool video_range, video1_range, aon_range, aon1_range;				\
 												\
 		addr = (adev->reg_offset[ip##_HWIP][inst_idx][reg##_BASE_IDX] + reg);		\
 		addr <<= 2; 									\
 		video_range = ((((0xFFFFF & addr) >= (VCN_VID_SOC_ADDRESS_2_0)) && 		\
 				((0xFFFFF & addr) < ((VCN_VID_SOC_ADDRESS_2_0 + 0x2600)))));	\
+		video1_range = ((((0xFFFFF & addr) >= (VCN1_VID_SOC_ADDRESS_3_0)) && 		\
+				((0xFFFFF & addr) < ((VCN1_VID_SOC_ADDRESS_3_0 + 0x2600)))));	\
 		aon_range   = ((((0xFFFFF & addr) >= (VCN_AON_SOC_ADDRESS_2_0)) && 		\
 				((0xFFFFF & addr) < ((VCN_AON_SOC_ADDRESS_2_0 + 0x600)))));	\
+		aon1_range   = ((((0xFFFFF & addr) >= (VCN1_AON_SOC_ADDRESS_3_0)) && 		\
+				((0xFFFFF & addr) < ((VCN1_AON_SOC_ADDRESS_3_0 + 0x600)))));	\
 		if (video_range) 								\
 			internal_reg_offset = ((0xFFFFF & addr) - (VCN_VID_SOC_ADDRESS_2_0) + 	\
 				(VCN_VID_IP_ADDRESS_2_0));					\
 		else if (aon_range)								\
 			internal_reg_offset = ((0xFFFFF & addr) - (VCN_AON_SOC_ADDRESS_2_0) + 	\
 				(VCN_AON_IP_ADDRESS_2_0));					\
+		else if (video1_range) 								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN1_VID_SOC_ADDRESS_3_0) + 	\
+				(VCN_VID_IP_ADDRESS_2_0));					\
+		else if (aon1_range)								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN1_AON_SOC_ADDRESS_3_0) + 	\
+				(VCN_AON_IP_ADDRESS_2_0));					\
 		else										\
 			internal_reg_offset = (0xFFFFF & addr);					\
 												\
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 147/207] drm/amdgpu: rename macro for VCN1.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Rename RREG32_SOC15_DPG_MODE and WREG32_SOC15_DPG_MODE for VCN1.0
These two macros are used specifically for VCN1.0, therefore rename
it from general name to VCN1.0 specific name.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h |  4 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c   | 96 ++++++++++++-------------
 2 files changed, 50 insertions(+), 50 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 25a07704cf75..f54e5ccabb42 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -67,7 +67,7 @@
 /* 1 second timeout */
 #define VCN_IDLE_TIMEOUT	msecs_to_jiffies(1000)
 
-#define RREG32_SOC15_DPG_MODE(ip, inst_idx, reg, mask, sram_sel) 			\
+#define RREG32_SOC15_DPG_MODE_1_0(ip, inst_idx, reg, mask, sram_sel) 			\
 	({	WREG32_SOC15(ip, inst_idx, mmUVD_DPG_LMA_MASK, mask); 			\
 		WREG32_SOC15(ip, inst_idx, mmUVD_DPG_LMA_CTL, 				\
 			UVD_DPG_LMA_CTL__MASK_EN_MASK | 				\
@@ -77,7 +77,7 @@
 		RREG32_SOC15(ip, inst_idx, mmUVD_DPG_LMA_DATA); 			\
 	})
 
-#define WREG32_SOC15_DPG_MODE(ip, inst_idx, reg, value, mask, sram_sel) 		\
+#define WREG32_SOC15_DPG_MODE_1_0(ip, inst_idx, reg, value, mask, sram_sel) 		\
 	do { 										\
 		WREG32_SOC15(ip, inst_idx, mmUVD_DPG_LMA_DATA, value); 			\
 		WREG32_SOC15(ip, inst_idx, mmUVD_DPG_LMA_MASK, mask); 			\
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index 1ad79155ed00..b5088b5dd4eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -360,68 +360,68 @@ static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev)
 
 	/* cache window 0: fw */
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
 			     (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo),
 			     0xFFFFFFFF, 0);
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
 			     (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi),
 			     0xFFFFFFFF, 0);
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0,
 			     0xFFFFFFFF, 0);
 		offset = 0;
 	} else {
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
 			lower_32_bits(adev->vcn.inst->gpu_addr), 0xFFFFFFFF, 0);
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
 			upper_32_bits(adev->vcn.inst->gpu_addr), 0xFFFFFFFF, 0);
 		offset = size;
-		WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
+		WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
 			     AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0xFFFFFFFF, 0);
 	}
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size, 0xFFFFFFFF, 0);
 
 	/* cache window 1: stack */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
 		     lower_32_bits(adev->vcn.inst->gpu_addr + offset), 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
 		     upper_32_bits(adev->vcn.inst->gpu_addr + offset), 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0,
 			     0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE,
 			     0xFFFFFFFF, 0);
 
 	/* cache window 2: context */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
 		     lower_32_bits(adev->vcn.inst->gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
 			     0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
 		     upper_32_bits(adev->vcn.inst->gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
 			     0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE,
 			     0xFFFFFFFF, 0);
 
 	/* VCN global tiling registers */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_UDEC_ADDR_CONFIG,
 			adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG,
 			adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG,
 			adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG,
 		adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
 }
 
@@ -636,9 +636,9 @@ static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t s
 		reg_data = 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
 	reg_data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
 	reg_data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmJPEG_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmJPEG_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
 
 	/* enable sw clock gating control */
 	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
@@ -667,16 +667,16 @@ static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t s
 		 UVD_CGC_CTRL__WCB_MODE_MASK |
 		 UVD_CGC_CTRL__VCPU_MODE_MASK |
 		 UVD_CGC_CTRL__SCPU_MODE_MASK);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
 
 	/* turn off clock gating */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
 
 	/* turn on SUVD clock gating */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_GATE, 1, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_SUVD_CGC_GATE, 1, 0xFFFFFFFF, sram_sel);
 
 	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_CTRL, 0, 0xFFFFFFFF, sram_sel);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_SUVD_CGC_CTRL, 0, 0xFFFFFFFF, sram_sel);
 }
 
 static void vcn_1_0_disable_static_power_gating(struct amdgpu_device *adev)
@@ -972,14 +972,14 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
 	tmp |= UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK;
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CNTL, tmp, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_VCPU_CNTL, tmp, 0xFFFFFFFF, 0);
 
 	/* disable interupt */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MASTINT_EN,
 			0, UVD_MASTINT_EN__VCPU_EN_MASK, 0);
 
 	/* initialize VCN memory controller */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_CTRL,
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
 		UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
 		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
@@ -993,48 +993,48 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 	/* swap (8 in 32) RB and IB */
 	lmi_swap_cntl = 0xa;
 #endif
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl, 0xFFFFFFFF, 0);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_CNTL,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MPC_CNTL,
 		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0xFFFFFFFF, 0);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUXA0,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MPC_SET_MUXA0,
 		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0xFFFFFFFF, 0);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUXB0,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MPC_SET_MUXB0,
 		((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0xFFFFFFFF, 0);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUX,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MPC_SET_MUX,
 		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0xFFFFFFFF, 0);
 
 	vcn_v1_0_mc_resume_dpg_mode(adev);
 
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_REG_XX_MASK, 0x10, 0xFFFFFFFF, 0);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, 0x3, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_REG_XX_MASK, 0x10, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, 0x3, 0xFFFFFFFF, 0);
 
 	/* boot up the VCPU */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SOFT_RESET, 0, 0xFFFFFFFF, 0);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_SOFT_RESET, 0, 0xFFFFFFFF, 0);
 
 	/* enable UMC */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL2,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_CTRL2,
 		0x1F << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT,
 		0xFFFFFFFF, 0);
 
 	/* enable master interrupt */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MASTINT_EN,
 			UVD_MASTINT_EN__VCPU_EN_MASK, UVD_MASTINT_EN__VCPU_EN_MASK, 0);
 
 	vcn_v1_0_clock_gating_dpg_mode(adev, 1);
 	/* setup mmUVD_LMI_CTRL */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_CTRL,
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
 		UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
 		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
@@ -1046,11 +1046,11 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 
 	tmp = adev->gfx.config.gb_addr_config;
 	/* setup VCN global tiling registers */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_JPEG_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
 
 	/* enable System Interrupt for JRBC */
-	WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SYS_INT_EN,
+	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_SYS_INT_EN,
 									UVD_SYS_INT_EN__UVD_JRBC_EN_MASK, 0xFFFFFFFF, 1);
 
 	/* force RBC into idle state */
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 148/207] drm/amdgpu: rename macro for VCN2.0 2.5 and 3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Rename SOC15_DPG_MODE_OFFSET_2_0, RREG32_SOC15_DPG_MODE_2_0 and
WREG32_SOC15_DPG_MODE_2_0 for VCN2.0, VCN2.5 and VCN3.0.
These three macros are used VCN2.0, VCN2.5 and VCN3.0, therefore rename
it to be a general name.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h |  6 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c   | 86 +++++++++++-----------
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c   | 94 ++++++++++++-------------
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c   | 88 +++++++++++------------
 4 files changed, 137 insertions(+), 137 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index f54e5ccabb42..e125e8bfac54 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -88,7 +88,7 @@
 			(sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); 		\
 	} while (0)
 
-#define SOC15_DPG_MODE_OFFSET_2_0(ip, inst_idx, reg) 						\
+#define SOC15_DPG_MODE_OFFSET(ip, inst_idx, reg) 						\
 	({											\
 		uint32_t internal_reg_offset, addr;						\
 		bool video_range, video1_range, aon_range, aon1_range;				\
@@ -121,7 +121,7 @@
 		internal_reg_offset >>= 2;							\
 	})
 
-#define RREG32_SOC15_DPG_MODE_2_0(inst_idx, offset, mask_en) 					\
+#define RREG32_SOC15_DPG_MODE(inst_idx, offset, mask_en) 					\
 	({											\
 		WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_CTL, 					\
 			(0x0 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT |				\
@@ -130,7 +130,7 @@
 		RREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA);				\
 	})
 
-#define WREG32_SOC15_DPG_MODE_2_0(inst_idx, offset, value, mask_en, indirect)			\
+#define WREG32_SOC15_DPG_MODE(inst_idx, offset, value, mask_en, indirect)			\
 	do {											\
 		if (!indirect) {								\
 			WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA, value);			\
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
index 90ed773695ea..060d133db55d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
@@ -382,91 +382,91 @@ static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_device *adev, bool indirec
 	/* cache window 0: fw */
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		if (!indirect) {
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		} else {
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 				UVD, 0, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		}
 		offset = 0;
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst->gpu_addr), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst->gpu_addr), 0, indirect);
 		offset = size;
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_VCPU_CACHE_OFFSET0),
 			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
 	}
 
 	if (!indirect)
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
 	else
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
 
 	/* cache window 1: stack */
 	if (!indirect) {
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst->gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst->gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 			UVD, 0, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	}
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
 
 	/* cache window 2: context */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
 		lower_32_bits(adev->vcn.inst->gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
 		upper_32_bits(adev->vcn.inst->gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
 
 	/* non-cache window */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
 		lower_32_bits(adev->vcn.inst->fw_shared_gpu_addr), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
 		upper_32_bits(adev->vcn.inst->fw_shared_gpu_addr), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_NONCACHE_SIZE0),
 		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)), 0, indirect);
 
 	/* VCN global tiling registers */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
 }
 
@@ -615,19 +615,19 @@ static void vcn_v2_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
 		 UVD_CGC_CTRL__WCB_MODE_MASK |
 		 UVD_CGC_CTRL__VCPU_MODE_MASK |
 		 UVD_CGC_CTRL__SCPU_MODE_MASK);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_CGC_CTRL), reg_data, sram_sel, indirect);
 
 	/* turn off clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_CGC_GATE), 0, sram_sel, indirect);
 
 	/* turn on SUVD clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
 
 	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
 }
 
@@ -810,11 +810,11 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
 	tmp |= UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK;
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	/* disable master interupt */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MASTINT_EN), 0, 0, indirect);
 
 	/* setup mmUVD_LMI_CTRL */
@@ -826,28 +826,28 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
 		0x00100000L);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_CTRL), tmp, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MPC_CNTL),
 		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MPC_SET_MUXA0),
 		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MPC_SET_MUXB0),
 		((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MPC_SET_MUX),
 		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
@@ -855,22 +855,22 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 
 	vcn_v2_0_mc_resume_dpg_mode(adev, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK), 0x3, 0, indirect);
 
 	/* release VCPU reset to boot */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_SOFT_RESET), 0, 0, indirect);
 
 	/* enable LMI MC and UMC channels */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_CTRL2),
 		0x1F << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT, 0, indirect);
 
 	/* enable master interrupt */
-	WREG32_SOC15_DPG_MODE_2_0(0, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_MASTINT_EN),
 		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index 3c6eafb62ee6..261afbb504bd 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -443,7 +443,7 @@ static void vcn_v2_5_mc_resume(struct amdgpu_device *adev)
 	}
 }
 
-static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
 {
 	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
 	uint32_t offset;
@@ -451,91 +451,91 @@ static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 	/* cache window 0: fw */
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		if (!indirect) {
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		} else {
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, 0, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		}
 		offset = 0;
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
 		offset = size;
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_VCPU_CACHE_OFFSET0),
 			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
 	}
 
 	if (!indirect)
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
 	else
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
 
 	/* cache window 1: stack */
 	if (!indirect) {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, 0, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	}
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
 
 	/* cache window 2: context */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
 		lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
 		upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
 
 	/* non-cache window */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
 		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
 		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_NONCACHE_SIZE0),
 		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)), 0, indirect);
 
 	/* VCN global tiling registers */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_GFX8_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
 }
 
@@ -657,7 +657,7 @@ static void vcn_v2_5_disable_clock_gating(struct amdgpu_device *adev)
 	}
 }
 
-static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev,
+void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev,
 		uint8_t sram_sel, int inst_idx, uint8_t indirect)
 {
 	uint32_t reg_data = 0;
@@ -689,19 +689,19 @@ static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev,
 		 UVD_CGC_CTRL__WCB_MODE_MASK |
 		 UVD_CGC_CTRL__VCPU_MODE_MASK |
 		 UVD_CGC_CTRL__MMSCH_MODE_MASK);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_CGC_CTRL), reg_data, sram_sel, indirect);
 
 	/* turn off clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_CGC_GATE), 0, sram_sel, indirect);
 
 	/* turn on SUVD clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
 
 	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
 }
 
@@ -792,11 +792,11 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
 	tmp |= UVD_VCPU_CNTL__BLK_RST_MASK;
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	/* disable master interupt */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MASTINT_EN), 0, 0, indirect);
 
 	/* setup mmUVD_LMI_CTRL */
@@ -808,28 +808,28 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
 		0x00100000L);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_CTRL), tmp, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MPC_CNTL),
 		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MPC_SET_MUXA0),
 		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MPC_SET_MUXB0),
 		((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MPC_SET_MUX),
 		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
@@ -837,26 +837,26 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 
 	vcn_v2_5_mc_resume_dpg_mode(adev, inst_idx, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_RBC_XX_IB_REG_CHECK), 0x3, 0, indirect);
 
 	/* enable LMI MC and UMC channels */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_CTRL2), 0, 0, indirect);
 
 	/* unblock VCPU register access */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_RB_ARB_CTRL), 0, 0, indirect);
 
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	/* enable master interrupt */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_MASTINT_EN),
 		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
 
@@ -1300,7 +1300,7 @@ static int vcn_v2_5_sriov_start(struct amdgpu_device *adev)
 	return vcn_v2_5_mmsch_start(adev, &adev->virt.mm_table);
 }
 
-static int vcn_v2_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+int vcn_v2_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
 {
 	int ret_code = 0;
 	uint32_t tmp;
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index f47136d9bfe8..c9b5d5a3e239 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -391,84 +391,84 @@ static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 	/* cache window 0: fw */
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		if (!indirect) {
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi), 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		} else {
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
-			WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 				VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
 		}
 		offset = 0;
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
 		offset = size;
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET0),
 			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
 	}
 
 	if (!indirect)
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
 	else
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
 
 	/* cache window 1: stack */
 	if (!indirect) {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	} else {
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
-		WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
 	}
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
 
 	/* cache window 2: context */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
 			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
 			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
 
 	/* non-cache window */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_NONCACHE_SIZE0), 0, 0, indirect);
 }
 
@@ -739,19 +739,19 @@ static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
 		 UVD_CGC_CTRL__WCB_MODE_MASK |
 		 UVD_CGC_CTRL__VCPU_MODE_MASK |
 		 UVD_CGC_CTRL__MMSCH_MODE_MASK);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_CGC_CTRL), reg_data, sram_sel, indirect);
 
 	/* turn off clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_CGC_GATE), 0, sram_sel, indirect);
 
 	/* turn on SUVD clock gating */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
 
 	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
 }
 
@@ -847,11 +847,11 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
 	tmp |= UVD_VCPU_CNTL__BLK_RST_MASK;
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	/* disable master interupt */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MASTINT_EN), 0, 0, indirect);
 
 	/* setup mmUVD_LMI_CTRL */
@@ -863,28 +863,28 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
 		0x00100000L);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_LMI_CTRL), tmp, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MPC_CNTL),
 		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MPC_SET_MUXA0),
 		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MPC_SET_MUXB0),
 		 ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
 		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
 		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MPC_SET_MUX),
 		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
@@ -892,31 +892,31 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 
 	vcn_v3_0_mc_resume_dpg_mode(adev, inst_idx, indirect);
 
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_RBC_XX_IB_REG_CHECK), 0x3, 0, indirect);
 
 	/* enable LMI MC and UMC channels */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_LMI_CTRL2), 0, 0, indirect);
 
 	/* unblock VCPU register access */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_RB_ARB_CTRL), 0, 0, indirect);
 
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
 	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	/* enable master interrupt */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_MASTINT_EN),
 		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
 
 	/* add nop to workaround PSP size check */
-	WREG32_SOC15_DPG_MODE_2_0(inst_idx, SOC15_DPG_MODE_OFFSET_2_0(
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
 	if (indirect)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 149/207] drm/amdgpu: add workaround for issue in DPG for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

To workaround an issue in DPG

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index c9b5d5a3e239..98ba6ddde823 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -1340,6 +1340,10 @@ static void vcn_v3_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		WREG32_SOC15(VCN, ring->me, mmUVD_SCRATCH2,
+			lower_32_bits(ring->wptr) | 0x80000000);
+
 	if (ring->use_doorbell) {
 		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 150/207] drm/amdgpu: enable DPG mode for VCN3.0

[Alex Deucher]

From: Boyuan Zhang <boyuan.zhang@amd.com>

Enable DPG mode for VCN3.0 by updating related flag.

V2: update description.

Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: James Zhu <james.zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nv.c       | 1 +
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 7 +++++--
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 936950fbafda..ef3f07dc89cd 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -726,6 +726,7 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_IH_CG |
 			AMD_CG_SUPPORT_MC_LS;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
 			AMD_PG_SUPPORT_JPEG |
 			AMD_PG_SUPPORT_ATHUB;
 		adev->external_rev_id = adev->rev_id + 0x28;
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 98ba6ddde823..875bdc877c3d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -254,7 +254,8 @@ static int vcn_v3_0_hw_init(void *handle)
 
 done:
 	if (!r)
-		DRM_INFO("VCN decode and encode initialized successfully.\n");
+		DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
+			(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
 
 	return r;
 }
@@ -278,7 +279,9 @@ static int vcn_v3_0_hw_fini(void *handle)
 
 		ring = &adev->vcn.inst[i].ring_dec;
 
-		if (RREG32_SOC15(VCN, i, mmUVD_STATUS))
+		if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
+			(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
+			RREG32_SOC15(VCN, i, mmUVD_STATUS)))
 			vcn_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
 
 		ring->sched.ready = false;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 151/207] drm/amdgpu/mes10.1: add no scheduler flag for mes

[Alex Deucher]

We don't want a gpu scheduler for mes.

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/mes_v10_1.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
index 5c28868f7adc..4b746584a797 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
@@ -846,6 +846,7 @@ static int mes_v10_1_ring_init(struct amdgpu_device *adev)
 	ring->use_doorbell = true;
 	ring->doorbell_index = adev->doorbell_index.mes_ring << 1;
 	ring->eop_gpu_addr = adev->mes.eop_gpu_addr;
+	ring->no_scheduler = true;
 	sprintf(ring->name, "mes_%d.%d.%d", ring->me, ring->pipe, ring->queue);
 
 	r = amdgpu_ring_init(adev, ring, 1024, NULL, 0, AMDGPU_RING_PRIO_DEFAULT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 152/207] drm/amdgpu/vcn3.0: schedule instance 0 for decode and 1 for encode

[Alex Deucher]

VCN3 has 2 unsymmetrical instances, i.e there're less codecs
on instance 1, we use 0 for decode and 1 for encode for now

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 875bdc877c3d..2497f29bbb35 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -167,6 +167,8 @@ static int vcn_v3_0_sw_init(void *handle)
 		ring = &adev->vcn.inst[i].ring_dec;
 		ring->use_doorbell = true;
 		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i;
+		if (i != 0)
+			ring->no_scheduler = true;
 		sprintf(ring->name, "vcn_dec_%d", i);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
 				     AMDGPU_RING_PRIO_DEFAULT);
@@ -183,6 +185,8 @@ static int vcn_v3_0_sw_init(void *handle)
 			ring = &adev->vcn.inst[i].ring_enc[j];
 			ring->use_doorbell = true;
 			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + j + 8 * i;
+			if (i != 1)
+				ring->no_scheduler = true;
 			sprintf(ring->name, "vcn_enc_%d.%d", i, j);
 			r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
 					     AMDGPU_RING_PRIO_DEFAULT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 153/207] drm/amd/powerplay: enable ULCK DS for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable uclk deep sleep for sienna_cichlid.
Df cstate kicks in first, then df triggers uclk ds with the sideband.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 545b4efc2bfa..8e3fe192e50a 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -301,6 +301,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_DS_SOCCLK_BIT)
 				| FEATURE_MASK(FEATURE_DS_DCEFCLK_BIT)
 				| FEATURE_MASK(FEATURE_DS_FCLK_BIT)
+				| FEATURE_MASK(FEATURE_DS_UCLK_BIT)
 				| FEATURE_MASK(FEATURE_FW_DSTATE_BIT)
 				| FEATURE_MASK(FEATURE_DF_CSTATE_BIT)
 				| FEATURE_MASK(FEATURE_RSMU_SMN_CG_BIT)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 154/207] drm/amdgpu/sriov : Use kiq to do tlb invalidation for gfx10 on sriov

[Alex Deucher]

From: shaoyunl <shaoyun.liu@amd.com>

On SRIOV run time, driver shouldn't directly access invalidation registers through MMIO.
Use kiq to submit wait_reg_mem package for the invalidation

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c |  3 +++
 drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 18 ++++++++++++++++++
 2 files changed, 21 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 8935b9b81d68..f0955b325798 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3480,6 +3480,9 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev)
 		    (adev->gfx.mec_feature_version >= 27))
 			adev->gfx.cp_fw_write_wait = true;
 		break;
+	case CHIP_SIENNA_CICHLID:
+		adev->gfx.cp_fw_write_wait = true;
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index ce2f72430403..061900e8afd3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -350,6 +350,24 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* flush hdp cache */
 	adev->nbio.funcs->hdp_flush(adev, NULL);
 
+	/* For SRIOV run time, driver shouldn't access the register through MMIO
+	 * Directly use kiq to do the vm invalidation instead
+	 */
+	if (adev->gfx.kiq.ring.sched.ready &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
+	    !adev->in_gpu_reset) {
+
+		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
+		const unsigned eng = 17;
+		u32 inv_req = gmc_v10_0_get_invalidate_req(vmid, flush_type);
+		u32 req = hub->vm_inv_eng0_req + eng;
+		u32 ack = hub->vm_inv_eng0_ack + eng;
+
+		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
+				1 << vmid);
+		return;
+	}
+
 	mutex_lock(&adev->mman.gtt_window_lock);
 
 	if (vmhub == AMDGPU_MMHUB_0) {
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 155/207] drm/amd/powerplay: and smc dpm info struct for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

And atom_smc_dpm_info_v4_9 struct for sienna_cichlid use.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/include/atomfirmware.h | 123 +++++++++++++++++++++
 1 file changed, 123 insertions(+)

diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h
index f7e98855cac7..b852248b2da8 100644
--- a/drivers/gpu/drm/amd/include/atomfirmware.h
+++ b/drivers/gpu/drm/amd/include/atomfirmware.h
@@ -2016,6 +2016,129 @@ struct atom_smc_dpm_info_v4_7
   uint32_t     BoardReserved[5];
 };
 
+struct smudpm_i2c_controller_config_v3
+{
+  uint8_t   Enabled;
+  uint8_t   Speed;
+  uint8_t   SlaveAddress;
+  uint8_t   ControllerPort;
+  uint8_t   ControllerName;
+  uint8_t   ThermalThrotter;
+  uint8_t   I2cProtocol;
+  uint8_t   PaddingConfig;
+};
+
+struct atom_smc_dpm_info_v4_9
+{
+  struct   atom_common_table_header  table_header;
+
+  //SECTION: Gaming Clocks
+  //uint32_t     GamingClk[6];
+
+  // SECTION: I2C Control
+  struct smudpm_i2c_controller_config_v3  I2cControllers[16];     
+
+  uint8_t      GpioScl;  // GPIO Number for SCL Line, used only for CKSVII2C1
+  uint8_t      GpioSda;  // GPIO Number for SDA Line, used only for CKSVII2C1
+  uint8_t      FchUsbPdSlaveAddr; //For requesting USB PD controller S-states via FCH I2C when entering PME turn off
+  uint8_t      I2cSpare;
+
+  // SECTION: SVI2 Board Parameters
+  uint8_t      VddGfxVrMapping;   // Use VR_MAPPING* bitfields
+  uint8_t      VddSocVrMapping;   // Use VR_MAPPING* bitfields
+  uint8_t      VddMem0VrMapping;  // Use VR_MAPPING* bitfields
+  uint8_t      VddMem1VrMapping;  // Use VR_MAPPING* bitfields
+
+  uint8_t      GfxUlvPhaseSheddingMask; // set this to 1 to set PSI0/1 to 1 in ULV mode
+  uint8_t      SocUlvPhaseSheddingMask; // set this to 1 to set PSI0/1 to 1 in ULV mode
+  uint8_t      VddciUlvPhaseSheddingMask; // set this to 1 to set PSI0/1 to 1 in ULV mode
+  uint8_t      MvddUlvPhaseSheddingMask; // set this to 1 to set PSI0/1 to 1 in ULV mode
+
+  // SECTION: Telemetry Settings
+  uint16_t     GfxMaxCurrent;   // in Amps
+  uint8_t      GfxOffset;       // in Amps
+  uint8_t      Padding_TelemetryGfx;
+
+  uint16_t     SocMaxCurrent;   // in Amps
+  uint8_t      SocOffset;       // in Amps
+  uint8_t      Padding_TelemetrySoc;
+
+  uint16_t     Mem0MaxCurrent;   // in Amps
+  uint8_t      Mem0Offset;       // in Amps
+  uint8_t      Padding_TelemetryMem0;
+  
+  uint16_t     Mem1MaxCurrent;   // in Amps
+  uint8_t      Mem1Offset;       // in Amps
+  uint8_t      Padding_TelemetryMem1;
+
+  uint32_t     MvddRatio; // This is used for MVDD  Svi2 Div Ratio workaround. It has 16 fractional bits (Q16.16)
+  
+  // SECTION: GPIO Settings
+  uint8_t      AcDcGpio;        // GPIO pin configured for AC/DC switching
+  uint8_t      AcDcPolarity;    // GPIO polarity for AC/DC switching
+  uint8_t      VR0HotGpio;      // GPIO pin configured for VR0 HOT event
+  uint8_t      VR0HotPolarity;  // GPIO polarity for VR0 HOT event
+
+  uint8_t      VR1HotGpio;      // GPIO pin configured for VR1 HOT event 
+  uint8_t      VR1HotPolarity;  // GPIO polarity for VR1 HOT event 
+  uint8_t      GthrGpio;        // GPIO pin configured for GTHR Event
+  uint8_t      GthrPolarity;    // replace GPIO polarity for GTHR
+
+  // LED Display Settings
+  uint8_t      LedPin0;         // GPIO number for LedPin[0]
+  uint8_t      LedPin1;         // GPIO number for LedPin[1]
+  uint8_t      LedPin2;         // GPIO number for LedPin[2]
+  uint8_t      LedEnableMask;
+
+  uint8_t      LedPcie;        // GPIO number for PCIE results
+  uint8_t      LedError;       // GPIO number for Error Cases
+  uint8_t      LedSpare1[2];
+
+  // SECTION: Clock Spread Spectrum
+  
+  // GFXCLK PLL Spread Spectrum
+  uint8_t      PllGfxclkSpreadEnabled;   // on or off
+  uint8_t      PllGfxclkSpreadPercent;   // Q4.4
+  uint16_t     PllGfxclkSpreadFreq;      // kHz
+
+  // GFXCLK DFLL Spread Spectrum
+  uint8_t      DfllGfxclkSpreadEnabled;   // on or off
+  uint8_t      DfllGfxclkSpreadPercent;   // Q4.4
+  uint16_t     DfllGfxclkSpreadFreq;      // kHz
+  
+  // UCLK Spread Spectrum
+  uint8_t      UclkSpreadEnabled;   // on or off
+  uint8_t      UclkSpreadPercent;   // Q4.4
+  uint16_t     UclkSpreadFreq;      // kHz
+
+  // FCLK Spread Spectrum
+  uint8_t      FclkSpreadEnabled;   // on or off
+  uint8_t      FclkSpreadPercent;   // Q4.4
+  uint16_t     FclkSpreadFreq;      // kHz
+  
+  // Section: Memory Config
+  uint32_t     MemoryChannelEnabled; // For DRAM use only, Max 32 channels enabled bit mask. 
+  
+  uint8_t      DramBitWidth; // For DRAM use only.  See Dram Bit width type defines
+  uint8_t      PaddingMem1[3];
+
+  // Section: Total Board Power
+  uint16_t     TotalBoardPower;     //Only needed for TCP Estimated case, where TCP = TGP+Total Board Power
+  uint16_t     BoardPowerPadding; 
+  
+  // SECTION: XGMI Training
+  uint8_t      XgmiLinkSpeed   [4];
+  uint8_t      XgmiLinkWidth   [4];
+
+  uint16_t     XgmiFclkFreq    [4];
+  uint16_t     XgmiSocVoltage  [4];
+
+  // SECTION: Board Reserved
+
+  uint32_t     BoardReserved[16];
+
+};
+
 /* 
   ***************************************************************************
     Data Table asic_profiling_info  structure
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 156/207] drm/amd/powerplay: append pptable for sienna_cichlid (v2)

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add function to append powerplay table from vbios for sienna_cichlid.

v2: squash in warning fix

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../drm/amd/powerplay/sienna_cichlid_ppt.c    | 101 ++++++++++++++++++
 1 file changed, 101 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 8e3fe192e50a..5589838a2238 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -358,6 +358,107 @@ static int sienna_cichlid_check_powerplay_table(struct smu_context *smu)
 
 static int sienna_cichlid_append_powerplay_table(struct smu_context *smu)
 {
+	struct smu_table_context *table_context = &smu->smu_table;
+	PPTable_t *smc_pptable = table_context->driver_pptable;
+	struct atom_smc_dpm_info_v4_9 *smc_dpm_table;
+	int index, ret;
+	int i;
+
+	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+					    smc_dpm_info);
+
+	ret = smu_get_atom_data_table(smu, index, NULL, NULL, NULL,
+				      (uint8_t **)&smc_dpm_table);
+	if (ret)
+		return ret;
+
+	memcpy(smc_pptable->I2cControllers, smc_dpm_table->I2cControllers,
+	       sizeof(I2cControllerConfig_t) * NUM_I2C_CONTROLLERS);
+
+	/* SVI2 Board Parameters */
+	smc_pptable->VddGfxVrMapping = smc_dpm_table->VddGfxVrMapping;
+	smc_pptable->VddSocVrMapping = smc_dpm_table->VddSocVrMapping;
+	smc_pptable->VddMem0VrMapping = smc_dpm_table->VddMem0VrMapping;
+	smc_pptable->VddMem1VrMapping = smc_dpm_table->VddMem1VrMapping;
+	smc_pptable->GfxUlvPhaseSheddingMask = smc_dpm_table->GfxUlvPhaseSheddingMask;
+	smc_pptable->SocUlvPhaseSheddingMask = smc_dpm_table->SocUlvPhaseSheddingMask;
+	smc_pptable->VddciUlvPhaseSheddingMask = smc_dpm_table->VddciUlvPhaseSheddingMask;
+	smc_pptable->MvddUlvPhaseSheddingMask = smc_dpm_table->MvddUlvPhaseSheddingMask;
+
+	/* Telemetry Settings */
+	smc_pptable->GfxMaxCurrent = smc_dpm_table->GfxMaxCurrent;
+	smc_pptable->GfxOffset = smc_dpm_table->GfxOffset;
+	smc_pptable->Padding_TelemetryGfx = smc_dpm_table->Padding_TelemetryGfx;
+	smc_pptable->SocMaxCurrent = smc_dpm_table->SocMaxCurrent;
+	smc_pptable->SocOffset = smc_dpm_table->SocOffset;
+	smc_pptable->Padding_TelemetrySoc = smc_dpm_table->Padding_TelemetrySoc;
+	smc_pptable->Mem0MaxCurrent = smc_dpm_table->Mem0MaxCurrent;
+	smc_pptable->Mem0Offset = smc_dpm_table->Mem0Offset;
+	smc_pptable->Padding_TelemetryMem0 = smc_dpm_table->Padding_TelemetryMem0;
+	smc_pptable->Mem1MaxCurrent = smc_dpm_table->Mem1MaxCurrent;
+	smc_pptable->Mem1Offset = smc_dpm_table->Mem1Offset;
+	smc_pptable->Padding_TelemetryMem1 = smc_dpm_table->Padding_TelemetryMem1;
+	smc_pptable->MvddRatio = smc_dpm_table->MvddRatio;
+
+	/* GPIO Settings */
+	smc_pptable->AcDcGpio = smc_dpm_table->AcDcGpio;
+	smc_pptable->AcDcPolarity = smc_dpm_table->AcDcPolarity;
+	smc_pptable->VR0HotGpio = smc_dpm_table->VR0HotGpio;
+	smc_pptable->VR0HotPolarity = smc_dpm_table->VR0HotPolarity;
+	smc_pptable->VR1HotGpio = smc_dpm_table->VR1HotGpio;
+	smc_pptable->VR1HotPolarity = smc_dpm_table->VR1HotPolarity;
+	smc_pptable->GthrGpio = smc_dpm_table->GthrGpio;
+	smc_pptable->GthrPolarity = smc_dpm_table->GthrPolarity;
+
+	/* LED Display Settings */
+	smc_pptable->LedPin0 = smc_dpm_table->LedPin0;
+	smc_pptable->LedPin1 = smc_dpm_table->LedPin1;
+	smc_pptable->LedPin2 = smc_dpm_table->LedPin2;
+	smc_pptable->LedEnableMask = smc_dpm_table->LedEnableMask;
+	smc_pptable->LedPcie = smc_dpm_table->LedPcie;
+	smc_pptable->LedError = smc_dpm_table->LedError;
+	smc_pptable->LedSpare1[0] = smc_dpm_table->LedSpare1[0];
+	smc_pptable->LedSpare1[1] = smc_dpm_table->LedSpare1[1];
+
+	/* GFXCLK PLL Spread Spectrum */
+	smc_pptable->PllGfxclkSpreadEnabled = smc_dpm_table->PllGfxclkSpreadEnabled;
+	smc_pptable->PllGfxclkSpreadPercent = smc_dpm_table->PllGfxclkSpreadPercent;
+	smc_pptable->PllGfxclkSpreadFreq = smc_dpm_table->PllGfxclkSpreadFreq;
+
+	/* GFXCLK DFLL Spread Spectrum */
+	smc_pptable->DfllGfxclkSpreadEnabled = smc_dpm_table->DfllGfxclkSpreadEnabled;
+	smc_pptable->DfllGfxclkSpreadPercent = smc_dpm_table->DfllGfxclkSpreadPercent;
+	smc_pptable->DfllGfxclkSpreadFreq = smc_dpm_table->DfllGfxclkSpreadFreq;
+
+	/* UCLK Spread Spectrum */
+	smc_pptable->UclkSpreadEnabled = smc_dpm_table->UclkSpreadEnabled;
+	smc_pptable->UclkSpreadPercent = smc_dpm_table->UclkSpreadPercent;
+	smc_pptable->UclkSpreadFreq = smc_dpm_table->UclkSpreadFreq;
+
+	/* FCLK Spred Spectrum */
+	smc_pptable->FclkSpreadEnabled = smc_dpm_table->FclkSpreadEnabled;
+	smc_pptable->FclkSpreadPercent = smc_dpm_table->FclkSpreadPercent;
+	smc_pptable->FclkSpreadFreq = smc_dpm_table->FclkSpreadFreq;
+
+	/* Memory Config */
+	smc_pptable->MemoryChannelEnabled = smc_dpm_table->MemoryChannelEnabled;
+	smc_pptable->DramBitWidth = smc_dpm_table->DramBitWidth;
+	smc_pptable->PaddingMem1[0] = smc_dpm_table->PaddingMem1[0];
+	smc_pptable->PaddingMem1[1] = smc_dpm_table->PaddingMem1[1];
+	smc_pptable->PaddingMem1[2] = smc_dpm_table->PaddingMem1[2];
+
+	/* Total board power */
+	smc_pptable->TotalBoardPower = smc_dpm_table->TotalBoardPower;
+	smc_pptable->BoardPowerPadding = smc_dpm_table->BoardPowerPadding;
+
+	/* XGMI Training */
+	for (i = 0; i < NUM_XGMI_PSTATE_LEVELS; i++) {
+		smc_pptable->XgmiLinkSpeed[i] = smc_dpm_table->XgmiLinkSpeed[i];
+		smc_pptable->XgmiLinkWidth[i] = smc_dpm_table->XgmiLinkWidth[i];
+		smc_pptable->XgmiFclkFreq[i] = smc_dpm_table->XgmiFclkFreq[i];
+		smc_pptable->XgmiSocVoltage[i] = smc_dpm_table->XgmiSocVoltage[i];
+	}
+
 	return 0;
 }
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 157/207] drm/amd/powerplay: enable VDDCI and MVDD for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable VDDCI and MVDD if PP_MCLK_DPM_MASK was enable for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 5589838a2238..6223ecf3c11b 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -321,7 +321,9 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	}
 
 	if (adev->pm.pp_feature & PP_MCLK_DPM_MASK)
-		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_UCLK_BIT);
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_UCLK_BIT)
+					| FEATURE_MASK(FEATURE_MEM_VDDCI_SCALING_BIT)
+					| FEATURE_MASK(FEATURE_MEM_MVDD_SCALING_BIT);
 
 	if (adev->pm.pp_feature & PP_PCIE_DPM_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DPM_LINK_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 158/207] drm/amdgpu: skip GPU scheduler setup for KIQ and MES ring

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Fix the coding error to skip GPU scheduler setup for KIQ and MES ring.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 8712a2e1b869..0ce82686c983 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -470,7 +470,7 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
 		return -ENOMEM;
 
 	/* No need to setup the GPU scheduler for KIQ and MES ring */
-	if (ring->funcs->type != AMDGPU_RING_TYPE_KIQ ||
+	if (ring->funcs->type != AMDGPU_RING_TYPE_KIQ &&
 	    ring->funcs->type != AMDGPU_RING_TYPE_MES) {
 		switch (ring->funcs->type) {
 		case AMDGPU_RING_TYPE_GFX:
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 159/207] drm/amd/powerplay: enable fw ctf

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

fw ctf can be triggered if the temperature can't be throttled below the limit.
then the gpu will be powered off and the whole system will hang.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 6223ecf3c11b..512fa07ab2b2 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -311,6 +311,7 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 				| FEATURE_MASK(FEATURE_TDC_BIT)
 				| FEATURE_MASK(FEATURE_BACO_BIT)
 				| FEATURE_MASK(FEATURE_APCC_DFLL_BIT)
+				| FEATURE_MASK(FEATURE_FW_CTF_BIT)
 				| FEATURE_MASK(FEATURE_FAN_CONTROL_BIT)
 				| FEATURE_MASK(FEATURE_THERMAL_BIT)
 				| FEATURE_MASK(FEATURE_OUT_OF_BAND_MONITOR_BIT);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 160/207] drm/amdgpu: disable runtime pm for sienna_cichlid temporarily

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Disable runtime pm for sienna_cichlid temporarily as BACO regression issue.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 3d85fe509deb..179b75742251 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -174,6 +174,7 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags)
 		 (adev->asic_type >= CHIP_TOPAZ) &&
 		 (adev->asic_type != CHIP_VEGA10) &&
 		 (adev->asic_type != CHIP_VEGA20) &&
+		 (adev->asic_type != CHIP_SIENNA_CICHLID) &&
 		 (adev->asic_type != CHIP_ARCTURUS)) /* enable runpm on VI+ */
 		adev->runpm = true;
 	else if (amdgpu_device_supports_baco(dev) &&
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 161/207] drm/amd/powerplay: drop jpeg instance1 dpm setup

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

VCN removed JPEG for instance 1, so drop jpeg instance1 dpm setup.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 512fa07ab2b2..667c912e47fd 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -645,9 +645,6 @@ static int sienna_cichlid_dpm_set_jpeg_enable(struct smu_context *smu, bool enab
 			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0, NULL);
 			if (ret)
 				return ret;
-			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerUpJpeg, 0x10000, NULL);
-			if (ret)
-				return ret;
 		}
 		power_gate->jpeg_gated = false;
 	} else {
@@ -655,9 +652,6 @@ static int sienna_cichlid_dpm_set_jpeg_enable(struct smu_context *smu, bool enab
 			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0, NULL);
 			if (ret)
 				return ret;
-			ret = smu_send_smc_msg_with_param(smu, SMU_MSG_PowerDownJpeg, 0x10000, NULL);
-			if (ret)
-				return ret;
 		}
 		power_gate->jpeg_gated = true;
 	}
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 162/207] drm/amdgpu: only use one gfx pipe for Sienna_Cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Only enable one gfx pipe for sienna_cichlid currently.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Acked-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index f0955b325798..cc1835cd0c91 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -55,7 +55,7 @@
  * 2. Async ring
  */
 #define GFX10_NUM_GFX_RINGS_NV1X	1
-#define GFX10_NUM_GFX_RINGS_Sienna_Cichlid	2
+#define GFX10_NUM_GFX_RINGS_Sienna_Cichlid	1
 #define GFX10_MEC_HPD_SIZE	2048
 
 #define F32_CE_PROGRAM_RAM_SIZE		65536
@@ -4232,7 +4232,7 @@ static int gfx_v10_0_sw_init(void *handle)
 		break;
 	case CHIP_SIENNA_CICHLID:
 		adev->gfx.me.num_me = 1;
-		adev->gfx.me.num_pipe_per_me = 2;
+		adev->gfx.me.num_pipe_per_me = 1;
 		adev->gfx.me.num_queue_per_pipe = 1;
 		adev->gfx.mec.num_mec = 2;
 		adev->gfx.mec.num_pipe_per_mec = 4;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 163/207] drm/amdgpu/sriov : Add sriov detection for sienna_cichlid

[Alex Deucher]

From: shaoyunl <shaoyun.liu@amd.com>

This is a regression due to the rebase , add sienna_cichlid sriov detection back

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index f3b38c9e04ca..bab9286021a7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -321,6 +321,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 	case CHIP_VEGA20:
 	case CHIP_NAVI10:
 	case CHIP_NAVI12:
+	case CHIP_SIENNA_CICHLID:
 	case CHIP_ARCTURUS:
 		reg = RREG32(mmRCC_IOV_FUNC_IDENTIFIER);
 		break;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 164/207] drm/amd/powerplay: support mclk socclk limit value set for sienna_cichlid.

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add support to force and unforce MCLK or SOCCLK to dpm limit value.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 667c912e47fd..ef8532ff8e30 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -945,6 +945,8 @@ static int sienna_cichlid_force_dpm_limit_value(struct smu_context *smu, bool hi
 
 	enum smu_clk_type clks[] = {
 		SMU_GFXCLK,
+		SMU_MCLK,
+		SMU_SOCCLK,
 	};
 
 	for (i = 0; i < ARRAY_SIZE(clks); i++) {
@@ -970,6 +972,8 @@ static int sienna_cichlid_unforce_dpm_levels(struct smu_context *smu)
 
 	enum smu_clk_type clks[] = {
 		SMU_GFXCLK,
+		SMU_MCLK,
+		SMU_SOCCLK,
 	};
 
 	for (i = 0; i < ARRAY_SIZE(clks); i++) {
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 165/207] drm/amd/amdgpu: disable gfxoff to retrieve gfxclk

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

For Sienna_Cichlid, GFXOFF state puts gfx dpm into standby mode, then the
gfxclk can't be retireved.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../drm/amd/powerplay/sienna_cichlid_ppt.c    | 71 ++++++++++++++++---
 1 file changed, 60 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index ef8532ff8e30..2ac30d2ddf5e 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -706,6 +706,9 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 	uint32_t mark_index = 0;
 	uint32_t gen_speed, lane_width;
 
+	if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
+		amdgpu_gfx_off_ctrl(adev, false);
+
 	switch (clk_type) {
 	case SMU_GFXCLK:
 	case SMU_SCLK:
@@ -716,20 +719,20 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 	case SMU_DCEFCLK:
 		ret = smu_get_current_clk_freq(smu, clk_type, &cur_value);
 		if (ret)
-			return size;
+			goto print_clk_out;
 
 		/* 10KHz -> MHz */
 		cur_value = cur_value / 100;
 
 		ret = smu_get_dpm_level_count(smu, clk_type, &count);
 		if (ret)
-			return size;
+			goto print_clk_out;
 
 		if (!sienna_cichlid_is_support_fine_grained_dpm(smu, clk_type)) {
 			for (i = 0; i < count; i++) {
 				ret = smu_get_dpm_freq_by_index(smu, clk_type, i, &value);
 				if (ret)
-					return size;
+					goto print_clk_out;
 
 				size += sprintf(buf + size, "%d: %uMhz %s\n", i, value,
 						cur_value == value ? "*" : "");
@@ -737,10 +740,10 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 		} else {
 			ret = smu_get_dpm_freq_by_index(smu, clk_type, 0, &freq_values[0]);
 			if (ret)
-				return size;
+				goto print_clk_out;
 			ret = smu_get_dpm_freq_by_index(smu, clk_type, count - 1, &freq_values[2]);
 			if (ret)
-				return size;
+				goto print_clk_out;
 
 			freq_values[1] = cur_value;
 			mark_index = cur_value == freq_values[0] ? 0 :
@@ -783,19 +786,26 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 		break;
 	}
 
+print_clk_out:
+	if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
+		amdgpu_gfx_off_ctrl(adev, true);
+
 	return size;
 }
 
 static int sienna_cichlid_force_clk_levels(struct smu_context *smu,
 				   enum smu_clk_type clk_type, uint32_t mask)
 {
-
+	struct amdgpu_device *adev = smu->adev;
 	int ret = 0, size = 0;
 	uint32_t soft_min_level = 0, soft_max_level = 0, min_freq = 0, max_freq = 0;
 
 	soft_min_level = mask ? (ffs(mask) - 1) : 0;
 	soft_max_level = mask ? (fls(mask) - 1) : 0;
 
+	if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
+		amdgpu_gfx_off_ctrl(adev, false);
+
 	switch (clk_type) {
 	case SMU_GFXCLK:
 	case SMU_SCLK:
@@ -812,20 +822,24 @@ static int sienna_cichlid_force_clk_levels(struct smu_context *smu,
 
 		ret = smu_get_dpm_freq_by_index(smu, clk_type, soft_min_level, &min_freq);
 		if (ret)
-			return size;
+			goto forec_level_out;
 
 		ret = smu_get_dpm_freq_by_index(smu, clk_type, soft_max_level, &max_freq);
 		if (ret)
-			return size;
+			goto forec_level_out;
 
 		ret = smu_set_soft_freq_range(smu, clk_type, min_freq, max_freq, false);
 		if (ret)
-			return size;
+			goto forec_level_out;
 		break;
 	default:
 		break;
 	}
 
+forec_level_out:
+	if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
+		amdgpu_gfx_off_ctrl(adev, true);
+
 	return size;
 }
 
@@ -1265,6 +1279,7 @@ static int sienna_cichlid_get_profiling_clk_mask(struct smu_context *smu,
 					 uint32_t *mclk_mask,
 					 uint32_t *soc_mask)
 {
+	struct amdgpu_device *adev = smu->adev;
 	int ret = 0;
 	uint32_t level_count = 0;
 
@@ -1276,7 +1291,9 @@ static int sienna_cichlid_get_profiling_clk_mask(struct smu_context *smu,
 			*mclk_mask = 0;
 	} else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) {
 		if(sclk_mask) {
+			amdgpu_gfx_off_ctrl(adev, false);
 			ret = smu_get_dpm_level_count(smu, SMU_SCLK, &level_count);
+			amdgpu_gfx_off_ctrl(adev, true);
 			if (ret)
 				return ret;
 			*sclk_mask = level_count - 1;
@@ -1698,6 +1715,38 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 	return 0;
 }
 
+int sienna_cichlid_get_dpm_ultimate_freq(struct smu_context *smu,
+				enum smu_clk_type clk_type,
+				uint32_t *min, uint32_t *max)
+{
+	struct amdgpu_device *adev = smu->adev;
+	int ret;
+
+	if (clk_type == SMU_GFXCLK)
+		amdgpu_gfx_off_ctrl(adev, false);
+	ret = smu_v11_0_get_dpm_ultimate_freq(smu, clk_type, min, max);
+	if (clk_type == SMU_GFXCLK)
+		amdgpu_gfx_off_ctrl(adev, true);
+
+	return ret;
+}
+
+int sienna_cichlid_set_soft_freq_limited_range(struct smu_context *smu,
+				      enum smu_clk_type clk_type,
+				      uint32_t min, uint32_t max)
+{
+	struct amdgpu_device *adev = smu->adev;
+	int ret;
+
+	if (clk_type == SMU_GFXCLK)
+		amdgpu_gfx_off_ctrl(adev, false);
+	ret = smu_v11_0_set_soft_freq_limited_range(smu, clk_type, min, max);
+	if (clk_type == SMU_GFXCLK)
+		amdgpu_gfx_off_ctrl(adev, true);
+
+	return ret;
+}
+
 static void sienna_cichlid_dump_pptable(struct smu_context *smu)
 {
 	struct smu_table_context *table_context = &smu->smu_table;
@@ -2452,8 +2501,8 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.baco_set_state = smu_v11_0_baco_set_state,
 	.baco_enter = smu_v11_0_baco_enter,
 	.baco_exit = smu_v11_0_baco_exit,
-	.get_dpm_ultimate_freq = smu_v11_0_get_dpm_ultimate_freq,
-	.set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range,
+	.get_dpm_ultimate_freq = sienna_cichlid_get_dpm_ultimate_freq,
+	.set_soft_freq_limited_range = sienna_cichlid_set_soft_freq_limited_range,
 	.override_pcie_parameters = smu_v11_0_override_pcie_parameters,
 	.get_pptable_power_limit = sienna_cichlid_get_pptable_power_limit,
 };
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 166/207] drm/amdgpu: enable gfxoff for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Enable GFXOFF for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c             | 4 +---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 3 +++
 drivers/gpu/drm/amd/powerplay/smu_v11_0.c          | 1 +
 3 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index cc1835cd0c91..5ddce490247d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3537,9 +3537,6 @@ static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev)
 		if (!gfx_v10_0_navi10_gfxoff_should_enable(adev))
 			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 		break;
-	case CHIP_SIENNA_CICHLID:
-		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
-		break;
 	default:
 		break;
 	}
@@ -7441,6 +7438,7 @@ static int gfx_v10_0_set_powergating_state(void *handle,
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
 	case CHIP_NAVI12:
+	case CHIP_SIENNA_CICHLID:
 		amdgpu_gfx_off_ctrl(adev, enable);
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 2ac30d2ddf5e..10114132fdd8 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -341,6 +341,9 @@ sienna_cichlid_get_allowed_feature_mask(struct smu_context *smu,
 	if (adev->pm.pp_feature & PP_SCLK_DEEP_SLEEP_MASK)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_DS_GFXCLK_BIT);
 
+	if (adev->pm.pp_feature & PP_GFXOFF_MASK)
+		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_GFXOFF_BIT);
+
 	if (smu->adev->pg_flags & AMD_PG_SUPPORT_ATHUB)
 		*(uint64_t *)feature_mask |= FEATURE_MASK(FEATURE_ATHUB_PG_BIT);
 
diff --git a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
index b9575f1c22a3..560f78a805b9 100644
--- a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
@@ -1353,6 +1353,7 @@ int smu_v11_0_gfx_off_control(struct smu_context *smu, bool enable)
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
 	case CHIP_NAVI12:
+	case CHIP_SIENNA_CICHLID:
 		if (!(adev->pm.pp_feature & PP_GFXOFF_MASK))
 			return 0;
 		if (enable)
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 167/207] drm/amd/powerplay: show gfxclk=0 in gfxoff state

[Alex Deucher]

From: Kenneth Feng <kenneth.feng@amd.com>

The instant retrieved gfxclk value should be 0 in gfxoff state.
This can be fetched with gfxoff enabled.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index 10114132fdd8..c924ecf2899f 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -709,9 +709,6 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 	uint32_t mark_index = 0;
 	uint32_t gen_speed, lane_width;
 
-	if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
-		amdgpu_gfx_off_ctrl(adev, false);
-
 	switch (clk_type) {
 	case SMU_GFXCLK:
 	case SMU_SCLK:
@@ -727,6 +724,10 @@ static int sienna_cichlid_print_clk_levels(struct smu_context *smu,
 		/* 10KHz -> MHz */
 		cur_value = cur_value / 100;
 
+		/* no need to disable gfxoff when retrieving the current gfxclk */
+		if ((clk_type == SMU_GFXCLK) || (clk_type == SMU_SCLK))
+			amdgpu_gfx_off_ctrl(adev, false);
+
 		ret = smu_get_dpm_level_count(smu, clk_type, &count);
 		if (ret)
 			goto print_clk_out;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 168/207] drm/amdgpu/psp: add structure to support PSP SPL

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add support for PSP SPL (Security patch level) table to support
anti-rollback of FW loaded by Trusted OS.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h   | 2 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h | 9 +++++++++
 2 files changed, 11 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index e19b98d48c98..e66958a01458 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -222,10 +222,12 @@ struct psp_context
 	uint32_t			sos_bin_size;
 	uint32_t			toc_bin_size;
 	uint32_t			kdb_bin_size;
+	uint32_t			spl_bin_size;
 	uint8_t				*sys_start_addr;
 	uint8_t				*sos_start_addr;
 	uint8_t				*toc_start_addr;
 	uint8_t				*kdb_start_addr;
+	uint8_t				*spl_start_addr;
 
 	/* tmr buffer */
 	struct amdgpu_bo		*tmr_bo;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
index b0e656409c03..df402c7b3233 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
@@ -99,6 +99,14 @@ struct psp_firmware_header_v1_2 {
 	uint32_t kdb_size_bytes;
 };
 
+/* version_major=1, version_minor=3 */
+struct psp_firmware_header_v1_3 {
+	struct psp_firmware_header_v1_1 v1_1;
+	uint32_t spl_header_version;
+	uint32_t spl_offset_bytes;
+	uint32_t spl_size_bytes;
+};
+
 /* version_major=1, version_minor=0 */
 struct ta_firmware_header_v1_0 {
 	struct common_firmware_header header;
@@ -266,6 +274,7 @@ union amdgpu_firmware_header {
 	struct smc_firmware_header_v2_0 smc_v2_0;
 	struct psp_firmware_header_v1_0 psp;
 	struct psp_firmware_header_v1_1 psp_v1_1;
+	struct psp_firmware_header_v1_3 psp_v1_3;
 	struct ta_firmware_header_v1_0 ta;
 	struct gfx_firmware_header_v1_0 gfx;
 	struct rlc_firmware_header_v1_0 rlc;
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 169/207] drm/amdgpu/psp: initialization PSP SPL fw

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Support for psp firmware header version v1_3 initialization and
information print.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c   | 13 ++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c | 24 +++++++++++++++++++++++
 2 files changed, 37 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 2fd40d0d7a0e..cd52b6e05793 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -2108,6 +2108,7 @@ int psp_init_sos_microcode(struct psp_context *psp,
 	const struct psp_firmware_header_v1_0 *sos_hdr;
 	const struct psp_firmware_header_v1_1 *sos_hdr_v1_1;
 	const struct psp_firmware_header_v1_2 *sos_hdr_v1_2;
+	const struct psp_firmware_header_v1_3 *sos_hdr_v1_3;
 	int err = 0;
 
 	if (!chip_name) {
@@ -2152,6 +2153,18 @@ int psp_init_sos_microcode(struct psp_context *psp,
 			adev->psp.kdb_start_addr = (uint8_t *)adev->psp.sys_start_addr +
 						    le32_to_cpu(sos_hdr_v1_2->kdb_offset_bytes);
 		}
+		if (sos_hdr->header.header_version_minor == 3) {
+			sos_hdr_v1_3 = (const struct psp_firmware_header_v1_3 *)adev->psp.sos_fw->data;
+			adev->psp.toc_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.toc_size_bytes);
+			adev->psp.toc_start_addr = (uint8_t *)adev->psp.sys_start_addr +
+				le32_to_cpu(sos_hdr_v1_3->v1_1.toc_offset_bytes);
+			adev->psp.kdb_bin_size = le32_to_cpu(sos_hdr_v1_3->v1_1.kdb_size_bytes);
+			adev->psp.kdb_start_addr = (uint8_t *)adev->psp.sys_start_addr +
+				le32_to_cpu(sos_hdr_v1_3->v1_1.kdb_offset_bytes);
+			adev->psp.spl_bin_size = le32_to_cpu(sos_hdr_v1_3->spl_size_bytes);
+			adev->psp.spl_start_addr = (uint8_t *)adev->psp.sys_start_addr +
+				le32_to_cpu(sos_hdr_v1_3->spl_offset_bytes);
+		}
 		break;
 	default:
 		dev_err(adev->dev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index cf3438696fce..744404a05fee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -279,6 +279,30 @@ void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr)
 			DRM_DEBUG("kdb_size_bytes: %u\n",
 				  le32_to_cpu(psp_hdr_v1_2->kdb_size_bytes));
 		}
+		if (version_minor == 3) {
+			const struct psp_firmware_header_v1_1 *psp_hdr_v1_1 =
+				container_of(psp_hdr, struct psp_firmware_header_v1_1, v1_0);
+			const struct psp_firmware_header_v1_3 *psp_hdr_v1_3 =
+				container_of(psp_hdr_v1_1, struct psp_firmware_header_v1_3, v1_1);
+			DRM_DEBUG("toc_header_version: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.toc_header_version));
+			DRM_DEBUG("toc_offset_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.toc_offset_bytes));
+			DRM_DEBUG("toc_size_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.toc_size_bytes));
+			DRM_DEBUG("kdb_header_version: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_header_version));
+			DRM_DEBUG("kdb_offset_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_offset_bytes));
+			DRM_DEBUG("kdb_size_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->v1_1.kdb_size_bytes));
+			DRM_DEBUG("spl_header_version: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->spl_header_version));
+			DRM_DEBUG("spl_offset_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->spl_offset_bytes));
+			DRM_DEBUG("spl_size_bytes: %u\n",
+				  le32_to_cpu(psp_hdr_v1_3->spl_size_bytes));
+		}
 	} else {
 		DRM_ERROR("Unknown PSP ucode version: %u.%u\n",
 			  version_major, version_minor);
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 170/207] drm/amdgpu/psp: support for loading PSP SPL fw

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add support for loading SPL firmware.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c |  8 ++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h |  4 +++
 drivers/gpu/drm/amd/amdgpu/psp_v11_0.c  | 34 +++++++++++++++++++++++++
 3 files changed, 46 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index cd52b6e05793..cdd65b577986 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -1355,6 +1355,14 @@ static int psp_hw_start(struct psp_context *psp)
 			}
 		}
 
+		if (psp->spl_bin_size) {
+			ret = psp_bootloader_load_spl(psp);
+			if (ret) {
+				DRM_ERROR("PSP load spl failed!\n");
+				return ret;
+			}
+		}
+
 		ret = psp_bootloader_load_sysdrv(psp);
 		if (ret) {
 			DRM_ERROR("PSP load sysdrv failed!\n");
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index e66958a01458..1513887e7a68 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -50,6 +50,7 @@ enum psp_bootloader_cmd {
 	PSP_BL__LOAD_KEY_DATABASE	= 0x80000,
 	PSP_BL__DRAM_LONG_TRAIN		= 0x100000,
 	PSP_BL__DRAM_SHORT_TRAIN	= 0x200000,
+	PSP_BL__LOAD_TOS_SPL_TABLE	= 0x10000000,
 };
 
 enum psp_ring_type
@@ -84,6 +85,7 @@ struct psp_funcs
 {
 	int (*init_microcode)(struct psp_context *psp);
 	int (*bootloader_load_kdb)(struct psp_context *psp);
+	int (*bootloader_load_spl)(struct psp_context *psp);
 	int (*bootloader_load_sysdrv)(struct psp_context *psp);
 	int (*bootloader_load_sos)(struct psp_context *psp);
 	int (*ring_init)(struct psp_context *psp, enum psp_ring_type ring_type);
@@ -298,6 +300,8 @@ struct amdgpu_psp_funcs {
 		((psp)->funcs->init_microcode ? (psp)->funcs->init_microcode((psp)) : 0)
 #define psp_bootloader_load_kdb(psp) \
 		((psp)->funcs->bootloader_load_kdb ? (psp)->funcs->bootloader_load_kdb((psp)) : 0)
+#define psp_bootloader_load_spl(psp) \
+		((psp)->funcs->bootloader_load_spl ? (psp)->funcs->bootloader_load_spl((psp)) : 0)
 #define psp_bootloader_load_sysdrv(psp) \
 		((psp)->funcs->bootloader_load_sysdrv ? (psp)->funcs->bootloader_load_sysdrv((psp)) : 0)
 #define psp_bootloader_load_sos(psp) \
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index 234ba127ec2e..423386272920 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -252,6 +252,39 @@ static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp)
 	return ret;
 }
 
+static int psp_v11_0_bootloader_load_spl(struct psp_context *psp)
+{
+	int ret;
+	uint32_t psp_gfxdrv_command_reg = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	/* Check tOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	if (psp_v11_0_is_sos_alive(psp))
+		return 0;
+
+	ret = psp_v11_0_wait_for_bootloader(psp);
+	if (ret)
+		return ret;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+
+	/* Copy PSP SPL binary to memory */
+	memcpy(psp->fw_pri_buf, psp->spl_start_addr, psp->spl_bin_size);
+
+	/* Provide the PSP SPL to bootloader */
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36,
+	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
+	psp_gfxdrv_command_reg = PSP_BL__LOAD_TOS_SPL_TABLE;
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
+	       psp_gfxdrv_command_reg);
+
+	ret = psp_v11_0_wait_for_bootloader(psp);
+
+	return ret;
+}
+
 static int psp_v11_0_bootloader_load_sysdrv(struct psp_context *psp)
 {
 	int ret;
@@ -785,6 +818,7 @@ static int psp_v11_0_read_usbc_pd_fw(struct psp_context *psp, uint32_t *fw_ver)
 static const struct psp_funcs psp_v11_0_funcs = {
 	.init_microcode = psp_v11_0_init_microcode,
 	.bootloader_load_kdb = psp_v11_0_bootloader_load_kdb,
+	.bootloader_load_spl = psp_v11_0_bootloader_load_spl,
 	.bootloader_load_sysdrv = psp_v11_0_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v11_0_bootloader_load_sos,
 	.ring_init = psp_v11_0_ring_init,
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 171/207] drm/amdgpu: update golden setting for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Update golden setting for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 5ddce490247d..e9045dde5b24 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -3079,8 +3079,10 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_EXCEPTION_CONTROL, 0x7fff0f1f, 0x00b80000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Sienna_Cichlid, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGE_PC_CNTL, 0x003fffff, 0x00280400),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xffffffcf),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CM_CTRL1, 0xff8fff0f, 0x580f1008),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x10f80988),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf17fffff, 0x01200007),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_TIMEOUT_COUNTER, 0xffffffff, 0x00000800),
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 172/207] drm/amd/powerplay: update smu function for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add function to check whether baco is support for sienna cichlid.
Remove fucntion of get clock by type with latency as it will not be
called.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 .../drm/amd/powerplay/sienna_cichlid_ppt.c    | 101 ++++++++----------
 1 file changed, 45 insertions(+), 56 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
index c924ecf2899f..ce16cabb0780 100644
--- a/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/sienna_cichlid_ppt.c
@@ -21,7 +21,6 @@
  *
  */
 
-#include "pp_debug.h"
 #include <linux/firmware.h>
 #include <linux/pci.h>
 #include "amdgpu.h"
@@ -36,8 +35,9 @@
 #include "sienna_cichlid_ppt.h"
 #include "smu_v11_0_pptable.h"
 #include "smu_v11_0_7_ppsmc.h"
-
+#include "nbio/nbio_2_3_offset.h"
 #include "nbio/nbio_2_3_sh_mask.h"
+
 #include "asic_reg/mp/mp_11_0_sh_mask.h"
 
 #define FEATURE_MASK(feature) (1ULL << feature)
@@ -516,6 +516,10 @@ static int sienna_cichlid_tables_init(struct smu_context *smu, struct smu_table
 		return -ENOMEM;
 	smu_table->metrics_time = 0;
 
+	smu_table->watermarks_table = kzalloc(sizeof(Watermarks_t), GFP_KERNEL);
+	if (!smu_table->watermarks_table)
+		return -ENOMEM;
+
 	return 0;
 }
 
@@ -867,40 +871,6 @@ static int sienna_cichlid_populate_umd_state_clk(struct smu_context *smu)
 	return ret;
 }
 
-static int sienna_cichlid_get_clock_by_type_with_latency(struct smu_context *smu,
-						 enum smu_clk_type clk_type,
-						 struct pp_clock_levels_with_latency *clocks)
-{
-	int ret = 0, i = 0;
-	uint32_t level_count = 0, freq = 0;
-
-	switch (clk_type) {
-	case SMU_GFXCLK:
-	case SMU_DCEFCLK:
-	case SMU_SOCCLK:
-		ret = smu_get_dpm_level_count(smu, clk_type, &level_count);
-		if (ret)
-			return ret;
-
-		level_count = min(level_count, (uint32_t)MAX_NUM_CLOCKS);
-		clocks->num_levels = level_count;
-
-		for (i = 0; i < level_count; i++) {
-			ret = smu_get_dpm_freq_by_index(smu, clk_type, i, &freq);
-			if (ret)
-				return ret;
-
-			clocks->data[i].clocks_in_khz = freq * 1000;
-			clocks->data[i].latency_in_us = 0;
-		}
-		break;
-	default:
-		break;
-	}
-
-	return ret;
-}
-
 static int sienna_cichlid_pre_display_config_changed(struct smu_context *smu)
 {
 	int ret = 0;
@@ -930,23 +900,13 @@ static int sienna_cichlid_display_config_changed(struct smu_context *smu)
 {
 	int ret = 0;
 
-	if ((smu->watermarks_bitmap & WATERMARKS_EXIST) &&
-	    !(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
-		ret = smu_write_watermarks_table(smu);
-		if (ret)
-			return ret;
-
-		smu->watermarks_bitmap |= WATERMARKS_LOADED;
-	}
-
 	if ((smu->watermarks_bitmap & WATERMARKS_EXIST) &&
 	    smu_feature_is_supported(smu, SMU_FEATURE_DPM_DCEFCLK_BIT) &&
 	    smu_feature_is_supported(smu, SMU_FEATURE_DPM_SOCCLK_BIT)) {
-		/* Sienna_Cichlid do not support to change display num currently */
-		ret = 0;
 #if 0
 		ret = smu_send_smc_msg_with_param(smu, SMU_MSG_NumOfDisplays,
-						  smu->display_config->num_display, NULL);
+						  smu->display_config->num_display,
+						  NULL);
 #endif
 		if (ret)
 			return ret;
@@ -1211,8 +1171,6 @@ static int sienna_cichlid_set_power_profile_mode(struct smu_context *smu, long *
 	}
 
 	if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
-		if (size < 0)
-			return -EINVAL;
 
 		ret = smu_update_table(smu,
 				       SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
@@ -1338,8 +1296,14 @@ static int sienna_cichlid_notify_smc_display_config(struct smu_context *smu)
 		ret = smu_v11_0_display_clock_voltage_request(smu, &clock_req);
 		if (!ret) {
 			if (smu_feature_is_supported(smu, SMU_FEATURE_DS_DCEFCLK_BIT)) {
-				pr_err("Attempt to set divider for DCEFCLK Failed as it not support currently!");
-				return ret;
+				ret = smu_send_smc_msg_with_param(smu,
+								  SMU_MSG_SetMinDeepSleepDcefclk,
+								  min_clocks.dcef_clock_in_sr/100,
+								  NULL);
+				if (ret) {
+					pr_err("Attempt to set divider for DCEFCLK Failed!");
+					return ret;
+				}
 			}
 		} else {
 			pr_info("Attempt to set Hard Min for DCEFCLK Failed!");
@@ -1363,6 +1327,7 @@ static int sienna_cichlid_set_watermarks_table(struct smu_context *smu,
 				       *clock_ranges)
 {
 	int i;
+	int ret = 0;
 	Watermarks_t *table = watermarks;
 
 	if (!table || !clock_ranges)
@@ -1414,6 +1379,17 @@ static int sienna_cichlid_set_watermarks_table(struct smu_context *smu,
 				clock_ranges->wm_mcif_clocks_ranges[i].wm_set_id;
         }
 
+	smu->watermarks_bitmap |= WATERMARKS_EXIST;
+
+	if (!(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
+		ret = smu_write_watermarks_table(smu);
+		if (ret) {
+			pr_err("Failed to update WMTABLE!");
+			return ret;
+		}
+		smu->watermarks_bitmap |= WATERMARKS_LOADED;
+	}
+
 	return 0;
 }
 
@@ -1705,8 +1681,10 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 					pcie_width_cap);
 
 		ret = smu_send_smc_msg_with_param(smu,
-						  SMU_MSG_OverridePcieParameters,
-						  smu_pcie_arg, NULL);
+					  SMU_MSG_OverridePcieParameters,
+					  smu_pcie_arg,
+					  NULL);
+
 		if (ret)
 			return ret;
 
@@ -1751,6 +1729,18 @@ int sienna_cichlid_set_soft_freq_limited_range(struct smu_context *smu,
 	return ret;
 }
 
+static bool sienna_cichlid_is_baco_supported(struct smu_context *smu)
+{
+	struct amdgpu_device *adev = smu->adev;
+	uint32_t val;
+
+	if (!smu_v11_0_baco_is_support(smu))
+		return false;
+
+	val = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP0);
+	return (val & RCC_BIF_STRAP0__STRAP_PX_CAPABLE_MASK) ? true : false;
+}
+
 static void sienna_cichlid_dump_pptable(struct smu_context *smu)
 {
 	struct smu_table_context *table_context = &smu->smu_table;
@@ -2438,7 +2428,6 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.print_clk_levels = sienna_cichlid_print_clk_levels,
 	.force_clk_levels = sienna_cichlid_force_clk_levels,
 	.populate_umd_state_clk = sienna_cichlid_populate_umd_state_clk,
-	.get_clock_by_type_with_latency = sienna_cichlid_get_clock_by_type_with_latency,
 	.pre_display_config_changed = sienna_cichlid_pre_display_config_changed,
 	.display_config_changed = sienna_cichlid_display_config_changed,
 	.notify_smc_display_config = sienna_cichlid_notify_smc_display_config,
@@ -2500,7 +2489,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.register_irq_handler = smu_v11_0_register_irq_handler,
 	.set_azalia_d3_pme = smu_v11_0_set_azalia_d3_pme,
 	.get_max_sustainable_clocks_by_dc = smu_v11_0_get_max_sustainable_clocks_by_dc,
-	.baco_is_support= smu_v11_0_baco_is_support,
+	.baco_is_support= sienna_cichlid_is_baco_supported,
 	.baco_get_state = smu_v11_0_baco_get_state,
 	.baco_set_state = smu_v11_0_baco_set_state,
 	.baco_enter = smu_v11_0_baco_enter,
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 173/207] drm/amd/powerplay: drop sienna_cichlid hardcode of using pptable

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Drop the hardcode of sienna_cichlid which will force to use softpptable,
so that it can use pptable on vbios once the value of pp_table_id get
from vbios is 0.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/smu_v11_0.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
index 560f78a805b9..06b376a1396d 100644
--- a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
@@ -366,8 +366,7 @@ int smu_v11_0_setup_pptable(struct smu_context *smu)
 	hdr = (const struct smc_firmware_header_v1_0 *) adev->pm.fw->data;
 	version_major = le16_to_cpu(hdr->header.header_version_major);
 	version_minor = le16_to_cpu(hdr->header.header_version_minor);
-	if ((version_major == 2 && smu->smu_table.boot_values.pp_table_id > 0) ||
-	    adev->asic_type == CHIP_SIENNA_CICHLID) {
+	if (version_major == 2 && smu->smu_table.boot_values.pp_table_id > 0) {
 		pr_info("use driver provided pptable %d\n", smu->smu_table.boot_values.pp_table_id);
 		switch (version_minor) {
 		case 0:
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 174/207] drm/amd/powerplay: let PMFW to handle the features disablement on BACO in V2

[Alex Deucher]

From: Evan Quan <evan.quan@amd.com>

For Sienna_Cichlid, PMFW will handle the features disablement on BACO in. No
need to have driver stepped in.

V2: limit this for baco really

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/powerplay/amdgpu_smu.c | 24 ++++++++++++++++------
 1 file changed, 18 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
index 870fe32437b6..343139a5c7a6 100644
--- a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
@@ -1532,6 +1532,11 @@ static int smu_suspend(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	struct smu_context *smu = &adev->smu;
 	int ret;
+	bool use_baco = !smu->is_apu &&
+		((adev->in_gpu_reset &&
+		  (amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO)) ||
+		 (adev->in_runpm && amdgpu_asic_supports_baco(adev)));
+
 
 	if (amdgpu_sriov_vf(adev)&& !amdgpu_sriov_is_pp_one_vf(adev))
 		return 0;
@@ -1549,15 +1554,22 @@ static int smu_suspend(void *handle)
 		return ret;
 	}
 
-	ret = smu_disable_dpm(smu);
-	if (ret)
-		return ret;
+	/*
+	 * For Sienna_Cichlid, PMFW will handle the features disablement properly
+	 * on BACO in. Driver involvement is unnecessary.
+	 */
+	if ((adev->asic_type != CHIP_SIENNA_CICHLID) || !use_baco) {
+		ret = smu_disable_dpm(smu);
+		if (ret)
+			return ret;
+
+		if (adev->asic_type >= CHIP_NAVI10 &&
+		    adev->gfx.rlc.funcs->stop)
+			adev->gfx.rlc.funcs->stop(adev);
+	}
 
 	smu->watermarks_bitmap &= ~(WATERMARKS_LOADED);
 
-	if (adev->asic_type >= CHIP_NAVI10 &&
-	    adev->gfx.rlc.funcs->stop)
-		adev->gfx.rlc.funcs->stop(adev);
 	if (smu->is_apu)
 		smu_set_gfx_cgpg(&adev->smu, false);
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 175/207] drm/amdgpu: reserve fb according to return value from vbios

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Query reserved tmr size through atom firmwareinfo for Sienna_Cichlid and onwards
for all the use cases (IP discovery/G6 memory
training/profiling/diagnostic data.etc), otherwise, fallback to legacy
approach to check and reserve tmr block for ip discovery data and G6
memory training data respectively

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 127 +++++++++++++-----------
 1 file changed, 71 insertions(+), 56 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index b5e050b82172..12675cb7dde4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1831,46 +1831,85 @@ static u64 amdgpu_ttm_training_get_c2p_offset(u64 vram_size)
        return ALIGN(vram_size, SZ_1M);
 }
 
-/**
- * amdgpu_ttm_training_reserve_vram_init - create bo vram reservation from memory training
- *
- * @adev: amdgpu_device pointer
- *
- * create bo vram reservation from memory training.
+static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
+{
+	struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx;
+
+	memset(ctx, 0, sizeof(*ctx));
+
+	ctx->c2p_train_data_offset =
+		amdgpu_ttm_training_get_c2p_offset(adev->gmc.mc_vram_size);
+	ctx->p2c_train_data_offset =
+		(adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET);
+	ctx->train_data_size =
+		GDDR6_MEM_TRAINING_DATA_SIZE_IN_BYTES;
+	
+	DRM_DEBUG("train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
+			ctx->train_data_size,
+			ctx->p2c_train_data_offset,
+			ctx->c2p_train_data_offset);
+}
+
+/*
+ * reserve TMR memory at the top of VRAM which holds
+ * IP Discovery data and is protected by PSP.
  */
-static int amdgpu_ttm_training_reserve_vram_init(struct amdgpu_device *adev)
+static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 {
 	int ret;
 	struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx;
+	bool mem_train_support = false;
 
-	memset(ctx, 0, sizeof(*ctx));
-	if (!adev->fw_vram_usage.mem_train_support) {
-		DRM_DEBUG("memory training does not support!\n");
-		return 0;
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->fw_vram_usage.mem_train_support) {
+			mem_train_support = true;
+			amdgpu_ttm_training_data_block_init(adev);
+		} else
+			DRM_DEBUG("memory training does not support!\n");
 	}
 
-	ctx->c2p_train_data_offset = amdgpu_ttm_training_get_c2p_offset(adev->gmc.mc_vram_size);
-	ctx->p2c_train_data_offset = (adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET);
-	ctx->train_data_size = GDDR6_MEM_TRAINING_DATA_SIZE_IN_BYTES;
-
-	DRM_DEBUG("train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
-		  ctx->train_data_size,
-		  ctx->p2c_train_data_offset,
-		  ctx->c2p_train_data_offset);
+	/*
+	 * Query reserved tmr size through atom firmwareinfo for Sienna_Cichlid and onwards for all
+	 * the use cases (IP discovery/G6 memory training/profiling/diagnostic data.etc)
+	 *
+	 * Otherwise, fallback to legacy approach to check and reserve tmr block for ip
+	 * discovery data and G6 memory training data respectively
+	 */
+	adev->discovery_tmr_size =
+		amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
+	if (!adev->discovery_tmr_size) {
+		adev->discovery_tmr_size = DISCOVERY_TMR_SIZE;
+		if (mem_train_support) {
+			/* reserve vram for mem train indepently */
+			ret = amdgpu_bo_create_kernel_at(adev,
+						 ctx->c2p_train_data_offset,
+						 ctx->train_data_size,
+						 AMDGPU_GEM_DOMAIN_VRAM,
+						 &ctx->c2p_bo,
+						 NULL);
+			if (ret) {
+				DRM_ERROR("alloc c2p_bo failed(%d)!\n", ret);
+				amdgpu_ttm_training_reserve_vram_fini(adev);
+				return ret;
+			}
+		}
+	}
 
 	ret = amdgpu_bo_create_kernel_at(adev,
-					 ctx->c2p_train_data_offset,
-					 ctx->train_data_size,
-					 AMDGPU_GEM_DOMAIN_VRAM,
-					 &ctx->c2p_bo,
-					 NULL);
+				adev->gmc.real_vram_size - adev->discovery_tmr_size,
+				adev->discovery_tmr_size,
+				AMDGPU_GEM_DOMAIN_VRAM,
+				&adev->discovery_memory,
+				NULL);
 	if (ret) {
-		DRM_ERROR("alloc c2p_bo failed(%d)!\n", ret);
-		amdgpu_ttm_training_reserve_vram_fini(adev);
+		DRM_ERROR("alloc tmr failed(%d)!\n", ret);
+		amdgpu_bo_free_kernel(&adev->discovery_memory, NULL, NULL);
 		return ret;
 	}
 
-	ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS;
+	if (mem_train_support)
+		ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS;
+
 	return 0;
 }
 
@@ -1938,11 +1977,12 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	}
 
 	/*
-	 *The reserved vram for memory training must be pinned to the specified
-	 *place on the VRAM, so reserve it early.
+	 * only NAVI10 and onwards ASIC support for IP discovery.
+	 * If IP discovery enabled, a block of memory should be
+	 * reserved for IP discovey.
 	 */
-	if (!amdgpu_sriov_vf(adev)) {
-		r = amdgpu_ttm_training_reserve_vram_init(adev);
+	if (adev->asic_type >= CHIP_NAVI10 && amdgpu_discovery) {
+		r = amdgpu_ttm_reserve_tmr(adev);
 		if (r)
 			return r;
 	}
@@ -1958,31 +1998,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	/*
-	 * reserve TMR memory at the top of VRAM which holds
-	 * IP Discovery data and is protected by PSP.
-	 */
-	adev->discovery_tmr_size =
-		amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
-	if (!adev->discovery_tmr_size &&
-	    adev->asic_type >= CHIP_NAVI10 &&
-	    amdgpu_discovery) {
-		/* if fw_reserved_fb_size is 0 from vbios,
-		 * then fallback to the default tmr_size */
-		adev->discovery_tmr_size = DISCOVERY_TMR_SIZE;
-	}
-
-	if (adev->discovery_tmr_size > 0) {
-		r = amdgpu_bo_create_kernel_at(adev,
-					adev->gmc.real_vram_size - adev->discovery_tmr_size,
-					adev->discovery_tmr_size,
-					AMDGPU_GEM_DOMAIN_VRAM,
-					&adev->discovery_memory,
-					NULL);
-		if (r)
-			return r;
-	}
-
 	DRM_INFO("amdgpu: %uM of VRAM memory ready\n",
 		 (unsigned) (adev->gmc.real_vram_size / (1024 * 1024)));
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 176/207] drm/amdgpu: support memory training for sienna_cichlid

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

Add memory training support for sienna_cichlid.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index a9adccfda4c0..2b55ccd5cb44 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -532,6 +532,7 @@ static int gddr6_mem_train_support(struct amdgpu_device *adev)
 		switch (hw_v) {
 		case HW_REV(11, 0, 0):
 		case HW_REV(11, 0, 5):
+		case HW_REV(11, 0, 7):
 			ret = 1;
 			break;
 		default:
@@ -561,7 +562,8 @@ int amdgpu_atomfirmware_get_mem_train_info(struct amdgpu_device *adev)
 	adev->fw_vram_usage.mem_train_support = false;
 
 	if (adev->asic_type != CHIP_NAVI10 &&
-	    adev->asic_type != CHIP_NAVI14)
+	    adev->asic_type != CHIP_NAVI14 &&
+	    adev->asic_type != CHIP_SIENNA_CICHLID)
 		return 0;
 
 	if (amdgpu_sriov_vf(adev))
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 177/207] drm/amdgpu: remove unnecessary check for mem train

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

a.Check whether mem train support when try to reserve related memory.
b.Remove ASIC check and atom firmware table version check as the check
of firmware capability is enough to achieve that purpose.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h           |  4 --
 drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c  |  5 --
 .../gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c  | 47 +------------------
 .../gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h  |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c       |  7 ++-
 5 files changed, 7 insertions(+), 58 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 6e74ff9990e3..a4a0296b168d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -655,10 +655,6 @@ struct amdgpu_fw_vram_usage {
 	u64 size;
 	struct amdgpu_bo *reserved_bo;
 	void *va;
-
-	/* GDDR6 training support flag.
-	*/
-	bool mem_train_support;
 };
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index fdd52d86a4d7..c687432da426 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -2022,11 +2022,6 @@ int amdgpu_atombios_init(struct amdgpu_device *adev)
 	if (adev->is_atom_fw) {
 		amdgpu_atomfirmware_scratch_regs_init(adev);
 		amdgpu_atomfirmware_allocate_fb_scratch(adev);
-		ret = amdgpu_atomfirmware_get_mem_train_info(adev);
-		if (ret) {
-			DRM_ERROR("Failed to get mem train fb location.\n");
-			return ret;
-		}
 	} else {
 		amdgpu_atombios_scratch_regs_init(adev);
 		amdgpu_atombios_allocate_fb_scratch(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 2b55ccd5cb44..e249b22fef54 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -516,7 +516,7 @@ static bool gddr6_mem_train_vbios_support(struct amdgpu_device *adev)
 	return false;
 }
 
-static int gddr6_mem_train_support(struct amdgpu_device *adev)
+int amdgpu_mem_train_support(struct amdgpu_device *adev)
 {
 	int ret;
 	uint32_t major, minor, revision, hw_v;
@@ -551,51 +551,6 @@ static int gddr6_mem_train_support(struct amdgpu_device *adev)
 	return ret;
 }
 
-int amdgpu_atomfirmware_get_mem_train_info(struct amdgpu_device *adev)
-{
-	struct atom_context *ctx = adev->mode_info.atom_context;
-	int index;
-	uint8_t frev, crev;
-	uint16_t data_offset, size;
-	int ret;
-
-	adev->fw_vram_usage.mem_train_support = false;
-
-	if (adev->asic_type != CHIP_NAVI10 &&
-	    adev->asic_type != CHIP_NAVI14 &&
-	    adev->asic_type != CHIP_SIENNA_CICHLID)
-		return 0;
-
-	if (amdgpu_sriov_vf(adev))
-		return 0;
-
-	ret = gddr6_mem_train_support(adev);
-	if (ret == -1)
-		return -EINVAL;
-	else if (ret == 0)
-		return 0;
-
-	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
-					    vram_usagebyfirmware);
-	ret = amdgpu_atom_parse_data_header(ctx, index, &size, &frev, &crev,
-					    &data_offset);
-	if (ret == 0) {
-		DRM_ERROR("parse data header failed.\n");
-		return -EINVAL;
-	}
-
-	DRM_DEBUG("atom firmware common table header size:0x%04x, frev:0x%02x,"
-		  " crev:0x%02x, data_offset:0x%04x.\n", size, frev, crev, data_offset);
-	/* only support 2.1+ */
-	if (((uint16_t)frev << 8 | crev) < 0x0201) {
-		DRM_ERROR("frev:0x%02x, crev:0x%02x < 2.1 !\n", frev, crev);
-		return -EINVAL;
-	}
-
-	adev->fw_vram_usage.mem_train_support = true;
-	return 0;
-}
-
 int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
 {
 	struct atom_context *ctx = adev->mode_info.atom_context;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
index 3a5ed339903e..9f0d4356e8df 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
@@ -31,11 +31,11 @@ void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 	int *vram_width, int *vram_type, int *vram_vendor);
-int amdgpu_atomfirmware_get_mem_train_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev);
+int amdgpu_mem_train_support(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 12675cb7dde4..d968e38d475b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1861,10 +1861,13 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 	bool mem_train_support = false;
 
 	if (!amdgpu_sriov_vf(adev)) {
-		if (adev->fw_vram_usage.mem_train_support) {
+		ret = amdgpu_mem_train_support(adev);
+		if (ret == 1) {
 			mem_train_support = true;
 			amdgpu_ttm_training_data_block_init(adev);
-		} else
+		} else if (ret == -1)
+			return -EINVAL;
+		else
 			DRM_DEBUG("memory training does not support!\n");
 	}
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 178/207] drm/amdgpu: bypass tmr when reserve c2p memory

[Alex Deucher]

From: Likun Gao <Likun.Gao@amd.com>

C2P memory reserved should not in tmr memory range.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 48 ++++++++++---------------
 1 file changed, 19 insertions(+), 29 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index d968e38d475b..5b42c50366c9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1823,14 +1823,6 @@ static int amdgpu_ttm_training_reserve_vram_fini(struct amdgpu_device *adev)
 	return 0;
 }
 
-static u64 amdgpu_ttm_training_get_c2p_offset(u64 vram_size)
-{
-       if ((vram_size & (SZ_1M - 1)) < (SZ_4K + 1) )
-               vram_size -= SZ_1M;
-
-       return ALIGN(vram_size, SZ_1M);
-}
-
 static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
 {
 	struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx;
@@ -1838,7 +1830,7 @@ static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
 	memset(ctx, 0, sizeof(*ctx));
 
 	ctx->c2p_train_data_offset =
-		amdgpu_ttm_training_get_c2p_offset(adev->gmc.mc_vram_size);
+		ALIGN((adev->gmc.mc_vram_size - adev->discovery_tmr_size - SZ_1M), SZ_1M);
 	ctx->p2c_train_data_offset =
 		(adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET);
 	ctx->train_data_size =
@@ -1862,10 +1854,9 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 
 	if (!amdgpu_sriov_vf(adev)) {
 		ret = amdgpu_mem_train_support(adev);
-		if (ret == 1) {
+		if (ret == 1)
 			mem_train_support = true;
-			amdgpu_ttm_training_data_block_init(adev);
-		} else if (ret == -1)
+		else if (ret == -1)
 			return -EINVAL;
 		else
 			DRM_DEBUG("memory training does not support!\n");
@@ -1880,22 +1871,24 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 	 */
 	adev->discovery_tmr_size =
 		amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
-	if (!adev->discovery_tmr_size) {
+	if (!adev->discovery_tmr_size)
 		adev->discovery_tmr_size = DISCOVERY_TMR_SIZE;
-		if (mem_train_support) {
-			/* reserve vram for mem train indepently */
-			ret = amdgpu_bo_create_kernel_at(adev,
-						 ctx->c2p_train_data_offset,
-						 ctx->train_data_size,
-						 AMDGPU_GEM_DOMAIN_VRAM,
-						 &ctx->c2p_bo,
-						 NULL);
-			if (ret) {
-				DRM_ERROR("alloc c2p_bo failed(%d)!\n", ret);
-				amdgpu_ttm_training_reserve_vram_fini(adev);
-				return ret;
-			}
+
+	if (mem_train_support) {
+		/* reserve vram for mem train according to TMR location */
+		amdgpu_ttm_training_data_block_init(adev);
+		ret = amdgpu_bo_create_kernel_at(adev,
+					 ctx->c2p_train_data_offset,
+					 ctx->train_data_size,
+					 AMDGPU_GEM_DOMAIN_VRAM,
+					 &ctx->c2p_bo,
+					 NULL);
+		if (ret) {
+			DRM_ERROR("alloc c2p_bo failed(%d)!\n", ret);
+			amdgpu_ttm_training_reserve_vram_fini(adev);
+			return ret;
 		}
+		ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS;
 	}
 
 	ret = amdgpu_bo_create_kernel_at(adev,
@@ -1910,9 +1903,6 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 		return ret;
 	}
 
-	if (mem_train_support)
-		ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS;
-
 	return 0;
 }
 
-- 
2.25.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 112/207] drm/amdgpu: fix SDMA hdp flush engine conflict

[Christian König]

Am 01.06.20 um 20:19 schrieb Alex Deucher:
> From: Likun Gao <Likun.Gao@amd.com>
>
> Each of HDP flush engine should be used by one ring, correct allocate of
> hdp flush engine to SDMA ring.
> Correct me value of each SDMA ring, as it was cleared when init microcode.
>
> Signed-off-by: Likun Gao <Likun.Gao@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Reviewed-by: Christian König <christian.koenig@amd.com>

> ---
>   drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c | 6 ++----
>   1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
> index 52206050adb9..f072cef28b60 100644
> --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
> +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
> @@ -391,10 +391,7 @@ static void sdma_v5_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
>   	u32 ref_and_mask = 0;
>   	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
>   
> -	if (ring->me == 0)
> -		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0;
> -	else
> -		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma1;
> +	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
>   
>   	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
>   			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
> @@ -1224,6 +1221,7 @@ static int sdma_v5_2_sw_init(void *handle)
>   		ring = &adev->sdma.instance[i].ring;
>   		ring->ring_obj = NULL;
>   		ring->use_doorbell = true;
> +		ring->me = i;
>   
>   		DRM_INFO("use_doorbell being set to: [%s]\n",
>   				ring->use_doorbell?"true":"false");

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx