[PATCH 01/19] drm/amdgpu: add CHIP_IP_DISCOVERY support for virtualization

[PATCH 01/19] drm/amdgpu: add CHIP_IP_DISCOVERY support for virtualization

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

For further chips we will use CHIP_IP_DISCOVERY, so add this
support for virtualization

Signed-off-by: Horace Chen <horace.chen@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 | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 9be57389301b..0707e4338e65 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -707,6 +707,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 	case CHIP_SIENNA_CICHLID:
 	case CHIP_ARCTURUS:
 	case CHIP_ALDEBARAN:
+	case CHIP_IP_DISCOVERY:
 		reg = RREG32(mmRCC_IOV_FUNC_IDENTIFIER);
 		break;
 	default: /* other chip doesn't support SRIOV */
@@ -750,6 +751,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 		case CHIP_NAVI10:
 		case CHIP_NAVI12:
 		case CHIP_SIENNA_CICHLID:
+		case CHIP_IP_DISCOVERY:
 			nv_set_virt_ops(adev);
 			/* try send GPU_INIT_DATA request to host */
 			amdgpu_virt_request_init_data(adev);
-- 
2.37.1

[PATCH 02/19] drm/amdgpu: add sriov nbio callback structure

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

[Why]
under SR-IOV, the nbio doorbell range will be defined by PF. So VF
nbio doorbell range registers will be blocked. It will cause violation
if VF access those registers directly.

[How]
create an nbio_v4_3_sriov_funcs for sriov nbio_v4_3 initialization to
skip the setting for the doorbell range registers.

Signed-off-by: Horace Chen <horace.chen@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_discovery.c |  5 ++-
 drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c        | 44 +++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h        |  1 +
 3 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 5fd56ebc3f6c..854ca8d4505b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -2242,7 +2242,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		break;
 	case IP_VERSION(4, 3, 0):
 	case IP_VERSION(4, 3, 1):
-		adev->nbio.funcs = &nbio_v4_3_funcs;
+		if (amdgpu_sriov_vf(adev))
+			adev->nbio.funcs = &nbio_v4_3_sriov_funcs;
+		else
+			adev->nbio.funcs = &nbio_v4_3_funcs;
 		adev->nbio.hdp_flush_reg = &nbio_v4_3_hdp_flush_reg;
 		break;
 	case IP_VERSION(7, 7, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c
index 982a89f841d5..15eb3658d70e 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c
@@ -488,3 +488,47 @@ const struct amdgpu_nbio_funcs nbio_v4_3_funcs = {
 	.get_rom_offset = nbio_v4_3_get_rom_offset,
 	.program_aspm = nbio_v4_3_program_aspm,
 };
+
+
+static void nbio_v4_3_sriov_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{
+}
+
+static void nbio_v4_3_sriov_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+}
+
+static void nbio_v4_3_sriov_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					 int doorbell_index, int instance)
+{
+}
+
+static void nbio_v4_3_sriov_gc_doorbell_init(struct amdgpu_device *adev)
+{
+}
+
+const struct amdgpu_nbio_funcs nbio_v4_3_sriov_funcs = {
+	.get_hdp_flush_req_offset = nbio_v4_3_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v4_3_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v4_3_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v4_3_get_pcie_data_offset,
+	.get_rev_id = nbio_v4_3_get_rev_id,
+	.mc_access_enable = nbio_v4_3_mc_access_enable,
+	.get_memsize = nbio_v4_3_get_memsize,
+	.sdma_doorbell_range = nbio_v4_3_sriov_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v4_3_sriov_vcn_doorbell_range,
+	.gc_doorbell_init = nbio_v4_3_sriov_gc_doorbell_init,
+	.enable_doorbell_aperture = nbio_v4_3_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v4_3_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v4_3_sriov_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbio_v4_3_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v4_3_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v4_3_get_clockgating_state,
+	.ih_control = nbio_v4_3_ih_control,
+	.init_registers = nbio_v4_3_init_registers,
+	.remap_hdp_registers = nbio_v4_3_remap_hdp_registers,
+	.get_rom_offset = nbio_v4_3_get_rom_offset,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h
index ade43661d7a9..711999ceedf4 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h
@@ -28,5 +28,6 @@
 
 extern const struct nbio_hdp_flush_reg nbio_v4_3_hdp_flush_reg;
 extern const struct amdgpu_nbio_funcs nbio_v4_3_funcs;
+extern const struct amdgpu_nbio_funcs nbio_v4_3_sriov_funcs;
 
 #endif
-- 
2.37.1

[PATCH 03/19] drm/amdgpu: add a compute pipe reset for RS64

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

[Why]
Under SR-IOV, we are not sure whether pipe status is
good or not when doing initialization. The compute engine
maybe fail to bringup if pipe status is bad.

[How]
Do an RS64 pipe reset for MEC before we do initialization.
Also apply to bare-metal.

Signed-off-by: Horace Chen <horace.chen@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_v11_0.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index f2d126b110e8..fa718318568e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -2725,6 +2725,21 @@ static void gfx_v11_0_config_gfx_rs64(struct amdgpu_device *adev)
 					mec_hdr->ucode_start_addr_hi >> 2);
 	}
 	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset mec pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
+
+	/* clear mec pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
 }
 
 static int gfx_v11_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev)
-- 
2.37.1

[PATCH 04/19] drm/amdgpu: enable WPTR_POLL_ENABLE for sriov on sdma_v6_0

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

[Why]
Under SR-IOV, if VF is switched out then its doorbell will be disabled,
SDMA rely on WPTR_POLL to get doorbells which was sent during VF
switched-out time.

[How]
For SR-IOV, set SDMA WPTR_POLL_ENABLE to 1.

Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
index 2bc1407e885e..fb31dc911cd7 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
@@ -594,7 +594,10 @@ static int sdma_v6_0_gfx_resume(struct amdgpu_device *adev)
 		       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
 
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
+		if (amdgpu_sriov_vf(adev))
+			rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 1);
+		else
+			rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, F32_WPTR_POLL_ENABLE, 1);
 
 		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
-- 
2.37.1

[PATCH 05/19] drm/amdgpu: refine virtualization psp fw skip check

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

SR-IOV may need to load different firmwares for different ASIC inside
VF.
So create a new function in amdgpu_virt to check whether FW load needs
to be skipped.

Signed-off-by: Horace Chen <horace.chen@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  | 17 +++-----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 29 ++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h |  2 ++
 3 files changed, 34 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index b007371df59c..d00a9fe542bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -340,11 +340,12 @@ static int psp_init_sriov_microcode(struct psp_context *psp)
 		ret = psp_init_cap_microcode(psp, "aldebaran");
 		ret &= psp_init_ta_microcode(psp, "aldebaran");
 		break;
+	case IP_VERSION(13, 0, 0):
+		break;
 	default:
 		BUG();
 		break;
 	}
-
 	return ret;
 }
 
@@ -2412,19 +2413,7 @@ static bool fw_load_skip_check(struct psp_context *psp,
 		return true;
 
 	if (amdgpu_sriov_vf(psp->adev) &&
-	   (ucode->ucode_id == AMDGPU_UCODE_ID_SDMA0
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA1
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA2
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA3
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA4
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA5
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA6
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA7
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_G
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SMC))
+	    amdgpu_virt_fw_load_skip_check(psp->adev, ucode->ucode_id))
 		/*skip ucode loading in SRIOV VF */
 		return true;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 0707e4338e65..79f1e37caaec 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -809,6 +809,35 @@ enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *ad
 	return mode;
 }
 
+bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_id)
+{
+	/* this version doesn't support sriov autoload */
+	if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 0)) {
+		if (ucode_id == AMDGPU_UCODE_ID_VCN1 ||
+		    ucode_id == AMDGPU_UCODE_ID_VCN)
+			return false;
+		else
+			return true;
+	}
+
+	if (ucode_id == AMDGPU_UCODE_ID_SDMA0
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA1
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA2
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA3
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA4
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA5
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA6
+	    || ucode_id == AMDGPU_UCODE_ID_SDMA7
+	    || ucode_id == AMDGPU_UCODE_ID_RLC_G
+	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
+	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
+	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
+	    || ucode_id == AMDGPU_UCODE_ID_SMC)
+		return true;
+
+	return false;
+}
+
 void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
 			struct amdgpu_video_codec_info *encode, uint32_t encode_array_size,
 			struct amdgpu_video_codec_info *decode, uint32_t decode_array_size)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 239f232f9c02..cd6fce05978f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -343,4 +343,6 @@ void amdgpu_sriov_wreg(struct amdgpu_device *adev,
 		       u32 acc_flags, u32 hwip);
 u32 amdgpu_sriov_rreg(struct amdgpu_device *adev,
 		      u32 offset, u32 acc_flags, u32 hwip);
+bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev,
+			uint32_t ucode_id);
 #endif
-- 
2.37.1

[PATCH 06/19] drm/amdgpu: sriov remove vcn_4_0 and jpeg_4_0

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

SRIOV needs to initialize mmsch instead of multimedia engines
directly. So currently remove them for SR-IOV until the code and
firmwares are ready.

Signed-off-by: Horace Chen <horace.chen@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_discovery.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 854ca8d4505b..06b2d1830091 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -1908,8 +1908,10 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(4, 0, 0):
 		case IP_VERSION(4, 0, 2):
 		case IP_VERSION(4, 0, 4):
-			amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block);
-			amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block);
+			if (!amdgpu_sriov_vf(adev)) {
+				amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block);
+				amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block);
+			}
 			break;
 		default:
 			dev_err(adev->dev,
-- 
2.37.1

[PATCH 07/19] drm/amdgpu: Support PSP 13.0.10 on SR-IOV

[Alex Deucher]

From: Horace Chen <horace.chen@amd.com>

Add support for PSP 13.0.10 for SR-IOV VF

Signed-off-by: Horace Chen <horace.chen@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  | 12 ++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 62 ++++++++++++++++--------
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h |  3 ++
 3 files changed, 55 insertions(+), 22 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index d00a9fe542bf..bae4154455cf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -328,19 +328,27 @@ static int psp_init_sriov_microcode(struct psp_context *psp)
 
 	switch (adev->ip_versions[MP0_HWIP][0]) {
 	case IP_VERSION(9, 0, 0):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
 		ret = psp_init_cap_microcode(psp, "vega10");
 		break;
 	case IP_VERSION(11, 0, 9):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
 		ret = psp_init_cap_microcode(psp, "navi12");
 		break;
 	case IP_VERSION(11, 0, 7):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
 		ret = psp_init_cap_microcode(psp, "sienna_cichlid");
 		break;
 	case IP_VERSION(13, 0, 2):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
 		ret = psp_init_cap_microcode(psp, "aldebaran");
 		ret &= psp_init_ta_microcode(psp, "aldebaran");
 		break;
 	case IP_VERSION(13, 0, 0):
+		adev->virt.autoload_ucode_id = 0;
+		break;
+	case IP_VERSION(13, 0, 10):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MES1_DATA;
 		break;
 	default:
 		BUG();
@@ -768,6 +776,7 @@ static bool psp_skip_tmr(struct psp_context *psp)
 	case IP_VERSION(11, 0, 9):
 	case IP_VERSION(11, 0, 7):
 	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 10):
 		return true;
 	default:
 		return false;
@@ -2414,7 +2423,6 @@ static bool fw_load_skip_check(struct psp_context *psp,
 
 	if (amdgpu_sriov_vf(psp->adev) &&
 	    amdgpu_virt_fw_load_skip_check(psp->adev, ucode->ucode_id))
-		/*skip ucode loading in SRIOV VF */
 		return true;
 
 	if (psp->autoload_supported &&
@@ -2488,7 +2496,7 @@ static int psp_load_non_psp_fw(struct psp_context *psp)
 
 		/* Start rlc autoload after psp recieved all the gfx firmware */
 		if (psp->autoload_supported && ucode->ucode_id == (amdgpu_sriov_vf(adev) ?
-		    AMDGPU_UCODE_ID_CP_MEC2 : AMDGPU_UCODE_ID_RLC_G)) {
+		    adev->virt.autoload_ucode_id : AMDGPU_UCODE_ID_RLC_G)) {
 			ret = psp_rlc_autoload_start(psp);
 			if (ret) {
 				DRM_ERROR("Failed to start rlc autoload\n");
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 79f1e37caaec..a7de2c2b0cae 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -690,7 +690,6 @@ void amdgpu_virt_exchange_data(struct amdgpu_device *adev)
 	}
 }
 
-
 void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 {
 	uint32_t reg;
@@ -811,31 +810,54 @@ enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *ad
 
 bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_id)
 {
-	/* this version doesn't support sriov autoload */
-	if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 0)) {
+	switch (adev->ip_versions[MP0_HWIP][0]) {
+	case IP_VERSION(13, 0, 0):
+		/* no vf autoload, white list */
 		if (ucode_id == AMDGPU_UCODE_ID_VCN1 ||
 		    ucode_id == AMDGPU_UCODE_ID_VCN)
 			return false;
 		else
 			return true;
+	case IP_VERSION(13, 0, 10):
+		/* white list */
+		if (ucode_id == AMDGPU_UCODE_ID_CAP
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES_DATA
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1_DATA)
+			return false;
+		else
+			return true;
+	default:
+		/* lagacy black list */
+		if (ucode_id == AMDGPU_UCODE_ID_SDMA0
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA1
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA2
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA3
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA4
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA5
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA6
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA7
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_G
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_SMC)
+			return true;
+		else
+			return false;
 	}
-
-	if (ucode_id == AMDGPU_UCODE_ID_SDMA0
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA1
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA2
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA3
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA4
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA5
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA6
-	    || ucode_id == AMDGPU_UCODE_ID_SDMA7
-	    || ucode_id == AMDGPU_UCODE_ID_RLC_G
-	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
-	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
-	    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
-	    || ucode_id == AMDGPU_UCODE_ID_SMC)
-		return true;
-
-	return false;
 }
 
 void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index cd6fce05978f..d94c31e68a14 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -253,6 +253,9 @@ struct amdgpu_virt {
 	uint32_t decode_max_frame_pixels;
 	uint32_t encode_max_dimension_pixels;
 	uint32_t encode_max_frame_pixels;
+
+	/* the ucode id to signal the autoload */
+	uint32_t autoload_ucode_id;
 };
 
 struct amdgpu_video_codec_info;
-- 
2.37.1

[PATCH 08/19] drm/amdgpu: Use PSP program IH_RB_CNTL registers under SRIOV

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
With L1 Policy applied, IH_RB_CNTL/RING cannot be accessed by VF.

[How]
Use PSP program IH_RB_CNTL in VF.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/ih_v6_0.c | 26 +++++++++++++++++++++++---
 1 file changed, 23 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
index 085e613f3646..7cd79a3844b2 100644
--- a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
@@ -105,7 +105,13 @@ force_update_wptr_for_self_int(struct amdgpu_device *adev,
 	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
 				   RB_USED_INT_THRESHOLD, threshold);
 
-	WREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1, ih_rb_cntl);
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1, ih_rb_cntl))
+			return;
+	} else {
+		WREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1, ih_rb_cntl);
+	}
+
 	WREG32_SOC15(OSSSYS, 0, regIH_CNTL2, ih_cntl);
 }
 
@@ -132,7 +138,13 @@ static int ih_v6_0_toggle_ring_interrupts(struct amdgpu_device *adev,
 	/* enable_intr field is only valid in ring0 */
 	if (ih == &adev->irq.ih)
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
-	WREG32(ih_regs->ih_rb_cntl, tmp);
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+			return -ETIMEDOUT;
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
 
 	if (enable) {
 		ih->enabled = true;
@@ -242,7 +254,15 @@ static int ih_v6_0_enable_ring(struct amdgpu_device *adev,
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
 	}
-	WREG32(ih_regs->ih_rb_cntl, tmp);
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
 
 	if (ih == &adev->irq.ih) {
 		/* set the ih ring 0 writeback address whether it's enabled or not */
-- 
2.37.1

[PATCH 09/19] drm/amdgpu: Skip the program of MMMC_VM_AGP_* in SRIOV on MMHUB v3_0_0

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
VF should not program these registers, the value were defined in the host.

[How]
Skip writing them in SRIOV environment and program them on host side.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
index bc11b2de37ae..a1d26c4d80b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
@@ -169,17 +169,17 @@ static void mmhub_v3_0_init_system_aperture_regs(struct amdgpu_device *adev)
 	uint64_t value;
 	uint32_t tmp;
 
-	/* Disable AGP. */
-	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
-	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, 0);
-	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, 0x00FFFFFF);
-
 	if (!amdgpu_sriov_vf(adev)) {
 		/*
 		 * the new L1 policy will block SRIOV guest from writing
 		 * these regs, and they will be programed at host.
 		 * so skip programing these regs.
 		 */
+		/* Disable AGP. */
+		WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+		WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, 0);
+		WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, 0x00FFFFFF);
+
 		/* Program the system aperture low logical page number. */
 		WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
 			     adev->gmc.vram_start >> 18);
-- 
2.37.1

[PATCH 10/19] drm/amdgpu: skip "Issue additional private vm invalidation to MMHUB" on SRIOV

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
vm_l2_bank_select_reserved_cid2 is a PF_only register
that cannot be programmed by VF. This feature is only
support HDP using GPUVM page tables to access FB memory
which should be disabled on SRIOV.

[How]
Disable the feature on VF.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index 497f52283a30..592729cb0f48 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -234,7 +234,8 @@ static void gmc_v11_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 
 	/* Issue additional private vm invalidation to MMHUB */
 	if ((vmhub != AMDGPU_GFXHUB_0) &&
-	    (hub->vm_l2_bank_select_reserved_cid2)) {
+	    (hub->vm_l2_bank_select_reserved_cid2) &&
+		!amdgpu_sriov_vf(adev)) {
 		inv_req = RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2);
 		/* bit 25: RSERVED_CACHE_PRIVATE_INVALIDATION */
 		inv_req |= (1 << 25);
-- 
2.37.1

[PATCH 11/19] drm/amdgpu: Skip the VRAM base offset on SRIOV

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
As VF cannot read MMMC_VM_FB_OFFSET with L1 Policy(read 0xffffffff).
It leads to driver get the incorrect vram base offset.

[How]
Since SR-IOV is dGPU only, skip reading this register and set the
fb_offest to 0.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index 592729cb0f48..846ccb6cf07d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -649,7 +649,10 @@ static void gmc_v11_0_vram_gtt_location(struct amdgpu_device *adev,
 	amdgpu_gmc_gart_location(adev, mc);
 
 	/* base offset of vram pages */
-	adev->vm_manager.vram_base_offset = adev->mmhub.funcs->get_mc_fb_offset(adev);
+	if (amdgpu_sriov_vf(adev))
+		adev->vm_manager.vram_base_offset = 0;
+	else
+		adev->vm_manager.vram_base_offset = adev->mmhub.funcs->get_mc_fb_offset(adev);
 }
 
 /**
-- 
2.37.1

[PATCH 12/19] drm/amdgpu: Skip program SDMA0_SEM_WAIT_FAIL_TIMER_CNTL under SRIOV VF

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
As SDMA0_SEM_WAIT_FAIL_TIMER_CNTL is a PF-only register,
L1 would block this register for VF access.

[How]
VF do not program it.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
index fb31dc911cd7..7ae572a08cb3 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
@@ -560,7 +560,8 @@ static int sdma_v6_0_gfx_resume(struct amdgpu_device *adev)
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
 
-		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
+		if (!amdgpu_sriov_vf(adev))
+			WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
 
 		/* Set ring buffer size in dwords */
 		rb_bufsz = order_base_2(ring->ring_size / 4);
-- 
2.37.1

[PATCH 13/19] drm/amdgpu: Use RLCG to program GRBM_GFX_CNTL during full access time

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
KIQ register init requires GRBM_GFX_CNTL to select KIQ.

[How]
As RLCG accessing registers will save the data of GRBM_GFX_CNTL and restore it.
Use RLCG indirect accessing register method to select grbm instead of mmio directly access.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/soc21.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index 54bce19a52ec..47ef671543b3 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -179,7 +179,7 @@ void soc21_grbm_select(struct amdgpu_device *adev,
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid);
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_CNTL), grbm_gfx_cntl);
+	WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, grbm_gfx_cntl);
 }
 
 static void soc21_vga_set_state(struct amdgpu_device *adev, bool state)
-- 
2.37.1

[PATCH 14/19] drm/admgpu: Skip CG/PG on SOC21 under SRIOV VF

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
There is no CG(Clock Gating)/PG(Power Gating) requirement on SRIOV VF.
For multi VF, VF should not enable any CG/PG features.
For one VF, PF will program CG/PG related registers.

[How]
Do not set any cg/pg flag bit at early init under sriov.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/soc21.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index 47ef671543b3..34a267265971 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -582,6 +582,10 @@ static int soc21_common_early_init(void *handle)
 			AMD_PG_SUPPORT_JPEG |
 			AMD_PG_SUPPORT_ATHUB |
 			AMD_PG_SUPPORT_MMHUB;
+		if (amdgpu_sriov_vf(adev)) {
+			adev->cg_flags = 0;
+			adev->pg_flags = 0;
+		}
 		adev->external_rev_id = adev->rev_id + 0x1; // TODO: need update
 		break;
 	case IP_VERSION(11, 0, 2):
-- 
2.37.1

[PATCH 15/19] drm/amd: Skip smu_v13 register irq on SRIOV VF

[Alex Deucher]

From: Yifan Zha <Yifan.Zha@amd.com>

[Why]
VF do not need to handle SMU IRQ state.
L1 Policy will block VF access THM_THERMAL_INT_CTRL and MP1_SMN_IH_SW_INT/CNTL.

[How]
Skip smu_v13 init register_irq_handler under SRIOV VF.
And add irq_src check in enable/disable thermal alert
to avoid thermal alert enable/disable fail.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c
index 4fee391f4cfa..bebcd668a766 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c
@@ -1108,6 +1108,9 @@ int smu_v13_0_enable_thermal_alert(struct smu_context *smu)
 {
 	int ret = 0;
 
+	if (!smu->irq_source.num_types)
+		return 0;
+
 	ret = amdgpu_irq_get(smu->adev, &smu->irq_source, 0);
 	if (ret)
 		return ret;
@@ -1117,6 +1120,9 @@ int smu_v13_0_enable_thermal_alert(struct smu_context *smu)
 
 int smu_v13_0_disable_thermal_alert(struct smu_context *smu)
 {
+	if (!smu->irq_source.num_types)
+		return 0;
+
 	return amdgpu_irq_put(smu->adev, &smu->irq_source, 0);
 }
 
@@ -1488,6 +1494,9 @@ int smu_v13_0_register_irq_handler(struct smu_context *smu)
 	struct amdgpu_irq_src *irq_src = &smu->irq_source;
 	int ret = 0;
 
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
 	irq_src->num_types = 1;
 	irq_src->funcs = &smu_v13_0_irq_funcs;
 
-- 
2.37.1

[PATCH 16/19] drm/amdgpu/vcn: Disable CG/PG for SRIOV

[Alex Deucher]

From: Jane Jian <Jane.Jian@amd.com>

For sriov, CG and MG are controlled from hypervisor side,
no need to manage them again in ip init

Signed-off-by: Jane Jian <Jane.Jian@amd.com>
Reviewed-by: Sonny Jiang <sonny.jiang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/soc21.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index 34a267265971..a26c5723c46e 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -637,6 +637,11 @@ static int soc21_common_early_init(void *handle)
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_VCN_DPG |
 			AMD_PG_SUPPORT_JPEG;
+		if (amdgpu_sriov_vf(adev)) {
+			/* hypervisor control CG and PG enablement */
+			adev->cg_flags = 0;
+			adev->pg_flags = 0;
+		}
 		adev->external_rev_id = adev->rev_id + 0x20;
 		break;
 	default:
-- 
2.37.1

[PATCH 17/19] drm/amdgpu/vcn: Add vcn/vcn1 in white list to load its firmware under sriov

[Alex Deucher]

From: Jane Jian <Jane.Jian@amd.com>

Previously since vcn0/vcn1 are not enabled, loading firmware
is skipped. Now add firmware loading back since vcn0/vcn1
has already been enabled on sriov

Signed-off-by: Jane Jian <Jane.Jian@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 | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index a7de2c2b0cae..e4af40b9a8aa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -835,7 +835,9 @@ bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_i
 		|| ucode_id == AMDGPU_UCODE_ID_CP_MES
 		|| ucode_id == AMDGPU_UCODE_ID_CP_MES_DATA
 		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1
-		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1_DATA)
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1_DATA
+		|| ucode_id == AMDGPU_UCODE_ID_VCN1
+		|| ucode_id == AMDGPU_UCODE_ID_VCN)
 			return false;
 		else
 			return true;
-- 
2.37.1

[PATCH 18/19] drm/amdgpu/vcn: Add sriov VCN v4_0 unified queue support

[Alex Deucher]

From: Jane Jian <Jane.Jian@amd.com>

Enable unified queue support for sriov, abandon all previous
multi-queue settings

Signed-off-by: Jane Jian <Jane.Jian@amd.com>
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c |   5 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h       |  13 +
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c         | 297 +++++++++++++++++-
 3 files changed, 300 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 06b2d1830091..9fa2a5ceb77d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -1908,10 +1908,9 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(4, 0, 0):
 		case IP_VERSION(4, 0, 2):
 		case IP_VERSION(4, 0, 4):
-			if (!amdgpu_sriov_vf(adev)) {
-				amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block);
+			amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block);
+			if (!amdgpu_sriov_vf(adev))
 				amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block);
-			}
 			break;
 		default:
 			dev_err(adev->dev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 60c608144480..80b7a6cfd026 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -161,6 +161,7 @@
 #define AMDGPU_VCN_SW_RING_FLAG		(1 << 9)
 #define AMDGPU_VCN_FW_LOGGING_FLAG	(1 << 10)
 #define AMDGPU_VCN_SMU_VERSION_INFO_FLAG (1 << 11)
+#define AMDGPU_VCN_VF_RB_SETUP_FLAG (1 << 12)
 
 #define AMDGPU_VCN_IB_FLAG_DECODE_BUFFER	0x00000001
 #define AMDGPU_VCN_CMD_FLAG_MSG_BUFFER		0x00000001
@@ -317,12 +318,24 @@ struct amdgpu_fw_shared {
 	struct amdgpu_fw_shared_smu_interface_info smu_interface_info;
 };
 
+struct amdgpu_fw_shared_rb_setup {
+	uint32_t is_rb_enabled_flags;
+	uint32_t rb_addr_lo;
+	uint32_t rb_addr_hi;
+	uint32_t  rb_size;
+	uint32_t  rb4_addr_lo;
+	uint32_t  rb4_addr_hi;
+	uint32_t  rb4_size;
+	uint32_t  reserved[6];
+};
+
 struct amdgpu_vcn4_fw_shared {
 	uint32_t present_flag_0;
 	uint8_t pad[12];
 	struct amdgpu_fw_shared_unified_queue_struct sq;
 	uint8_t pad1[8];
 	struct amdgpu_fw_shared_fw_logging fw_log;
+	struct amdgpu_fw_shared_rb_setup rb_setup;
 };
 
 struct amdgpu_vcn_fwlog {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
index fb2d74f30448..09c89faa8c27 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -30,6 +30,7 @@
 #include "soc15d.h"
 #include "soc15_hw_ip.h"
 #include "vcn_v2_0.h"
+#include "mmsch_v4_0.h"
 
 #include "vcn/vcn_4_0_0_offset.h"
 #include "vcn/vcn_4_0_0_sh_mask.h"
@@ -45,6 +46,8 @@
 #define VCN_VID_SOC_ADDRESS_2_0							0x1fb00
 #define VCN1_VID_SOC_ADDRESS_3_0						0x48300
 
+#define VCN_HARVEST_MMSCH								0
+
 #define RDECODE_MSG_CREATE							0x00000000
 #define RDECODE_MESSAGE_CREATE							0x00000001
 
@@ -53,12 +56,14 @@ static int amdgpu_ih_clientid_vcns[] = {
 	SOC15_IH_CLIENTID_VCN1
 };
 
+static int vcn_v4_0_start_sriov(struct amdgpu_device *adev);
 static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev);
 static int vcn_v4_0_set_powergating_state(void *handle,
         enum amd_powergating_state state);
 static int vcn_v4_0_pause_dpg_mode(struct amdgpu_device *adev,
         int inst_idx, struct dpg_pause_state *new_state);
+static void vcn_v4_0_unified_ring_set_wptr(struct amdgpu_ring *ring);
 
 /**
  * vcn_v4_0_early_init - set function pointers
@@ -71,6 +76,9 @@ static int vcn_v4_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (amdgpu_sriov_vf(adev))
+		adev->vcn.harvest_config = VCN_HARVEST_MMSCH;
+
 	/* re-use enc ring as unified ring */
 	adev->vcn.num_enc_rings = 1;
 
@@ -92,6 +100,7 @@ static int vcn_v4_0_sw_init(void *handle)
 	struct amdgpu_ring *ring;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int i, r;
+	int vcn_doorbell_index = 0;
 
 	r = amdgpu_vcn_sw_init(adev);
 	if (r)
@@ -103,6 +112,12 @@ static int vcn_v4_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	if (amdgpu_sriov_vf(adev)) {
+		vcn_doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 - MMSCH_DOORBELL_OFFSET;
+		/* get DWORD offset */
+		vcn_doorbell_index = vcn_doorbell_index << 1;
+	}
+
 	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
 		volatile struct amdgpu_vcn4_fw_shared *fw_shared;
 
@@ -119,7 +134,10 @@ static int vcn_v4_0_sw_init(void *handle)
 
 		ring = &adev->vcn.inst[i].ring_enc[0];
 		ring->use_doorbell = true;
-		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i;
+		if (amdgpu_sriov_vf(adev))
+			ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1) + 1;
+		else
+			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i;
 
 		sprintf(ring->name, "vcn_unified_%d", i);
 
@@ -132,10 +150,19 @@ static int vcn_v4_0_sw_init(void *handle)
 		fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
 		fw_shared->sq.is_enabled = 1;
 
+		if (amdgpu_sriov_vf(adev))
+			fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
 		if (amdgpu_vcnfw_log)
 			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
 	}
 
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_alloc_mm_table(adev);
+		if (r)
+			return r;
+	}
+
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
 		adev->vcn.pause_dpg_mode = vcn_v4_0_pause_dpg_mode;
 
@@ -169,6 +196,9 @@ static int vcn_v4_0_sw_fini(void *handle)
 		drm_dev_exit(idx);
 	}
 
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_free_mm_table(adev);
+
 	r = amdgpu_vcn_suspend(adev);
 	if (r)
 		return r;
@@ -191,18 +221,42 @@ static int vcn_v4_0_hw_init(void *handle)
 	struct amdgpu_ring *ring;
 	int i, r;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
+	if (amdgpu_sriov_vf(adev)) {
+		r = vcn_v4_0_start_sriov(adev);
+		if (r)
+			goto done;
 
-		ring = &adev->vcn.inst[i].ring_enc[0];
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
 
-		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
-				((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i), i);
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			if (amdgpu_vcn_is_disabled_vcn(adev, VCN_ENCODE_RING, i)) {
+				ring->sched.ready = false;
+				ring->no_scheduler = true;
+				dev_info(adev->dev, "ring %s is disabled by hypervisor\n", ring->name);
+			} else {
+				ring->wptr = 0;
+				ring->wptr_old = 0;
+				vcn_v4_0_unified_ring_set_wptr(ring);
+				ring->sched.ready = true;
+			}
+		}
+	} else {
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
 
-		r = amdgpu_ring_test_helper(ring);
-		if (r)
-			goto done;
+			ring = &adev->vcn.inst[i].ring_enc[0];
+
+			adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+					((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i), i);
+
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				goto done;
+
+		}
 	}
 
 done:
@@ -230,12 +284,14 @@ static int vcn_v4_0_hw_fini(void *handle)
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
-
-		if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
+		if (!amdgpu_sriov_vf(adev)) {
+			if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
                         (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
                                 RREG32_SOC15(VCN, i, regUVD_STATUS))) {
                         vcn_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+			}
 		}
+
 	}
 
 	return 0;
@@ -1107,6 +1163,214 @@ static int vcn_v4_0_start(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int vcn_v4_0_start_sriov(struct amdgpu_device *adev)
+{
+	int i;
+	struct amdgpu_ring *ring_enc;
+	uint64_t cache_addr;
+	uint64_t rb_enc_addr;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t offset, cache_size;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw;
+	uint32_t init_status;
+	uint32_t enabled_vcn;
+
+	struct mmsch_v4_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v4_0_cmd_direct_read_modify_write
+		direct_rd_mod_wt = { {0} };
+	struct mmsch_v4_0_cmd_end end = { {0} };
+	struct mmsch_v4_0_init_header header;
+
+	volatile struct amdgpu_vcn4_fw_shared *fw_shared;
+	volatile struct amdgpu_fw_shared_rb_setup *rb_setup;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	direct_rd_mod_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE;
+	end.cmd_header.command_type =
+		MMSCH_COMMAND__END;
+
+	header.version = MMSCH_VERSION;
+	header.total_size = sizeof(struct mmsch_v4_0_init_header) >> 2;
+	for (i = 0; i < AMDGPU_MAX_VCN_INSTANCES; i++) {
+		header.inst[i].init_status = 0;
+		header.inst[i].table_offset = 0;
+		header.inst[i].table_size = 0;
+	}
+
+	table_loc = (uint32_t *)table->cpu_addr;
+	table_loc += header.total_size;
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		table_size = 0;
+
+		MMSCH_V4_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_STATUS),
+			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
+
+		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo);
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi);
+			offset = 0;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_VCPU_CACHE_OFFSET0),
+				0);
+		} else {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				lower_32_bits(adev->vcn.inst[i].gpu_addr));
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				upper_32_bits(adev->vcn.inst[i].gpu_addr));
+			offset = cache_size;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_VCPU_CACHE_OFFSET0),
+				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+		}
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE0),
+			cache_size);
+
+		cache_addr = adev->vcn.inst[i].gpu_addr + offset;
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_OFFSET1),
+			0);
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE1),
+			AMDGPU_VCN_STACK_SIZE);
+
+		cache_addr = adev->vcn.inst[i].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE;
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+			lower_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+			upper_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_OFFSET2),
+			0);
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE2),
+			AMDGPU_VCN_CONTEXT_SIZE);
+
+		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+		rb_setup = &fw_shared->rb_setup;
+
+		ring_enc = &adev->vcn.inst[i].ring_enc[0];
+		ring_enc->wptr = 0;
+		rb_enc_addr = ring_enc->gpu_addr;
+
+		rb_setup->is_rb_enabled_flags |= RB_ENABLED;
+		rb_setup->rb_addr_lo = lower_32_bits(rb_enc_addr);
+		rb_setup->rb_addr_hi = upper_32_bits(rb_enc_addr);
+		rb_setup->rb_size = ring_enc->ring_size / 4;
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+
+		/* add end packet */
+		MMSCH_V4_0_INSERT_END();
+
+		/* refine header */
+		header.inst[i].init_status = 0;
+		header.inst[i].table_offset = header.total_size;
+		header.inst[i].table_size = table_size;
+		header.total_size += table_size;
+	}
+
+	/* Update init table header in memory */
+	size = sizeof(struct mmsch_v4_0_init_header);
+	table_loc = (uint32_t *)table->cpu_addr;
+	memcpy((void *)table_loc, &header, size);
+
+	/* message MMSCH (in VCN[0]) to initialize this client
+	 * 1, write to mmsch_vf_ctx_addr_lo/hi register with GPU mc addr
+	 * of memory descriptor location
+	 */
+	ctx_addr = table->gpu_addr;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+	/* 2, update vmid of descriptor */
+	tmp = RREG32_SOC15(VCN, 0, regMMSCH_VF_VMID);
+	tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+	/* use domain0 for MM scheduler */
+	tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_VMID, tmp);
+
+	/* 3, notify mmsch about the size of this descriptor */
+	size = header.total_size;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_SIZE, size);
+
+	/* 4, set resp to zero */
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP, 0);
+
+	/* 5, kick off the initialization and wait until
+	 * MMSCH_VF_MAILBOX_RESP becomes non-zero
+	 */
+	param = 0x00000001;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_HOST, param);
+	tmp = 0;
+	timeout = 1000;
+	resp = 0;
+	expected = MMSCH_VF_MAILBOX_RESP__OK;
+	while (resp != expected) {
+		resp = RREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP);
+		if (resp != 0)
+			break;
+
+		udelay(10);
+		tmp = tmp + 10;
+		if (tmp >= timeout) {
+			DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+				" waiting for regMMSCH_VF_MAILBOX_RESP "\
+				"(expected=0x%08x, readback=0x%08x)\n",
+				tmp, expected, resp);
+			return -EBUSY;
+		}
+	}
+	enabled_vcn = amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, 0) ? 1 : 0;
+	init_status = ((struct mmsch_v4_0_init_header *)(table_loc))->inst[enabled_vcn].init_status;
+	if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE
+	&& init_status != MMSCH_VF_ENGINE_STATUS__PASS)
+		DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init "\
+			"status for VCN%x: 0x%x\n", resp, enabled_vcn, init_status);
+
+	return 0;
+}
+
 /**
  * vcn_v4_0_stop_dpg_mode - VCN stop with dpg mode
  *
@@ -1596,6 +1860,15 @@ static int vcn_v4_0_set_powergating_state(void *handle, enum amd_powergating_sta
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret;
 
+	/* for SRIOV, guest should not control VCN Power-gating
+	 * MMSCH FW should control Power-gating and clock-gating
+	 * guest should avoid touching CGC and PG
+	 */
+	if (amdgpu_sriov_vf(adev)) {
+		adev->vcn.cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
 	if(state == adev->vcn.cur_state)
 		return 0;
 
-- 
2.37.1

[PATCH 19/19] drm/amdgpu/vcn: Add MMSCH v4_0 support for sriov

[Alex Deucher]

From: Jane Jian <Jane.Jian@amd.com>

These structures are basically ported from MMSCH v3_0,
besides, added RB and RB4 enablement flag to support
unified queue

Signed-off-by: Jane Jian <Jane.Jian@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h | 140 ++++++++++++++++++++++++
 1 file changed, 140 insertions(+)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h

diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
new file mode 100644
index 000000000000..f772bb499f3e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
@@ -0,0 +1,140 @@
+/*
+ * Copyright 2022 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.
+ *
+ */
+
+#ifndef __MMSCH_V4_0_H__
+#define __MMSCH_V4_0_H__
+
+#include "amdgpu_vcn.h"
+
+#define MMSCH_VERSION_MAJOR	4
+#define MMSCH_VERSION_MINOR	0
+#define MMSCH_VERSION	(MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR)
+
+#define RB_ENABLED (1 << 0)
+#define RB4_ENABLED (1 << 1)
+#define MMSCH_DOORBELL_OFFSET 0x8
+
+#define MMSCH_VF_ENGINE_STATUS__PASS 0x1
+
+#define MMSCH_VF_MAILBOX_RESP__OK 0x1
+#define MMSCH_VF_MAILBOX_RESP__INCOMPLETE 0x2
+
+enum mmsch_v4_0_command_type {
+	MMSCH_COMMAND__DIRECT_REG_WRITE = 0,
+	MMSCH_COMMAND__DIRECT_REG_POLLING = 2,
+	MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE = 3,
+	MMSCH_COMMAND__INDIRECT_REG_WRITE = 8,
+	MMSCH_COMMAND__END = 0xf
+};
+
+struct mmsch_v4_0_table_info {
+	uint32_t init_status;
+	uint32_t table_offset;
+	uint32_t table_size;
+};
+
+struct mmsch_v4_0_init_header {
+	uint32_t version;
+	uint32_t total_size;
+	struct mmsch_v4_0_table_info inst[AMDGPU_MAX_VCN_INSTANCES];
+	struct mmsch_v4_0_table_info jpegdec;
+};
+
+struct mmsch_v4_0_cmd_direct_reg_header {
+	uint32_t reg_offset   : 28;
+	uint32_t command_type : 4;
+};
+
+struct mmsch_v4_0_cmd_indirect_reg_header {
+	uint32_t reg_offset    : 20;
+	uint32_t reg_idx_space : 8;
+	uint32_t command_type  : 4;
+};
+
+struct mmsch_v4_0_cmd_direct_write {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+struct mmsch_v4_0_cmd_direct_read_modify_write {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t write_data;
+	uint32_t mask_value;
+};
+
+struct mmsch_v4_0_cmd_direct_polling {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t mask_value;
+	uint32_t wait_value;
+};
+
+struct mmsch_v4_0_cmd_end {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+};
+
+struct mmsch_v4_0_cmd_indirect_write {
+	struct mmsch_v4_0_cmd_indirect_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+#define MMSCH_V4_0_INSERT_DIRECT_RD_MOD_WT(reg, mask, data) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_read_modify_write); \
+	size_dw = size / 4; \
+	direct_rd_mod_wt.cmd_header.reg_offset = reg; \
+	direct_rd_mod_wt.mask_value = mask; \
+	direct_rd_mod_wt.write_data = data; \
+	memcpy((void *)table_loc, &direct_rd_mod_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_DIRECT_WT(reg, value) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_write); \
+	size_dw = size / 4; \
+	direct_wt.cmd_header.reg_offset = reg; \
+	direct_wt.reg_value = value; \
+	memcpy((void *)table_loc, &direct_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_DIRECT_POLL(reg, mask, wait) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_polling); \
+	size_dw = size / 4; \
+	direct_poll.cmd_header.reg_offset = reg; \
+	direct_poll.mask_value = mask; \
+	direct_poll.wait_value = wait; \
+	memcpy((void *)table_loc, &direct_poll, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_END() { \
+	size = sizeof(struct mmsch_v4_0_cmd_end); \
+	size_dw = size / 4; \
+	memcpy((void *)table_loc, &end, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#endif
-- 
2.37.1

Re: [PATCH 01/19] drm/amdgpu: add CHIP_IP_DISCOVERY support for virtualization

[Christian König]

Am 30.08.22 um 20:39 schrieb Alex Deucher:
> From: Horace Chen <horace.chen@amd.com>
>
> For further chips we will use CHIP_IP_DISCOVERY, so add this
> support for virtualization
>
> Signed-off-by: Horace Chen <horace.chen@amd.com>
> Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Acked-by: Christian König <christian.koenig@amd.com> for the entire series.

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
> index 9be57389301b..0707e4338e65 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
> @@ -707,6 +707,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
>   	case CHIP_SIENNA_CICHLID:
>   	case CHIP_ARCTURUS:
>   	case CHIP_ALDEBARAN:
> +	case CHIP_IP_DISCOVERY:
>   		reg = RREG32(mmRCC_IOV_FUNC_IDENTIFIER);
>   		break;
>   	default: /* other chip doesn't support SRIOV */
> @@ -750,6 +751,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
>   		case CHIP_NAVI10:
>   		case CHIP_NAVI12:
>   		case CHIP_SIENNA_CICHLID:
> +		case CHIP_IP_DISCOVERY:
>   			nv_set_virt_ops(adev);
>   			/* try send GPU_INIT_DATA request to host */
>   			amdgpu_virt_request_init_data(adev);