[PATCH 1/7] drm/amdgpu: add new emit_reg_wait callback

[PATCH 1/7] drm/amdgpu: add new emit_reg_wait callback

[Christian König]

Allows us to wait for a register value/mask on a ring.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h      | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 ++
 2 files changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index d7930f3ead33..787f79c80b6b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1796,6 +1796,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
 #define amdgpu_ring_emit_rreg(r, d) (r)->funcs->emit_rreg((r), (d))
 #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r), (d), (v))
+#define amdgpu_ring_emit_reg_wait(r, d, v, m) (r)->funcs->emit_reg_wait((r), (d), (v), (m))
 #define amdgpu_ring_emit_tmz(r, b) (r)->funcs->emit_tmz((r), (b))
 #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
 #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index 70d05ec7bc07..867f53332305 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -150,6 +150,8 @@ struct amdgpu_ring_funcs {
 	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
 	void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg);
 	void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
+	void (*emit_reg_wait)(struct amdgpu_ring *ring, uint32_t reg,
+			      uint32_t val, uint32_t mask);
 	void (*emit_tmz)(struct amdgpu_ring *ring, bool start);
 	/* priority functions */
 	void (*set_priority) (struct amdgpu_ring *ring,
-- 
2.14.1

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

[PATCH 2/7] drm/amdgpu: implement gfx_v9_0_ring_emit_reg_wait

[Christian König]

Implement emit_reg_wait for gfx v9.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index e5d5341c459a..801d4a1dd7db 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -4003,6 +4003,12 @@ static void gfx_v9_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
 	amdgpu_ring_write(ring, val);
 }
 
+static void gfx_v9_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask)
+{
+	gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20);
+}
+
 static void gfx_v9_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
@@ -4321,6 +4327,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 	.patch_cond_exec = gfx_v9_0_ring_emit_patch_cond_exec,
 	.emit_tmz = gfx_v9_0_ring_emit_tmz,
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
@@ -4352,6 +4359,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.set_priority = gfx_v9_0_ring_set_priority_compute,
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
@@ -4379,6 +4387,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_rreg = gfx_v9_0_ring_emit_rreg,
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 };
 
 static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev)
-- 
2.14.1

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

[PATCH 3/7] drm/amdgpu: implement sdma_v4_0_ring_emit_reg_wait

[Christian König]

Add emit_reg_wait implementation for SDMA v4.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index 8505458d7041..e1ae39f86adf 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -1152,6 +1152,20 @@ static void sdma_v4_0_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
+static void sdma_v4_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					 uint32_t val, uint32_t mask)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val); /* reference */
+	amdgpu_ring_write(ring, mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
+}
+
 static int sdma_v4_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -1588,6 +1602,7 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
 	.insert_nop = sdma_v4_0_ring_insert_nop,
 	.pad_ib = sdma_v4_0_ring_pad_ib,
 	.emit_wreg = sdma_v4_0_ring_emit_wreg,
+	.emit_reg_wait = sdma_v4_0_ring_emit_reg_wait,
 };
 
 static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev)
-- 
2.14.1

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

[PATCH 4/7] drm/amdgpu: implement uvd_v7_0_(enc_|)ring_emit_reg_wait v2

[Christian König]

Add emit_reg_wait implementation for UVD v7.

v2: call new function directly from the existing code

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c | 40 +++++++++++++++++++++--------------
 1 file changed, 24 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index d317c764cc91..fcec0bea101d 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -1241,17 +1241,17 @@ static void uvd_v7_0_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, 8);
 }
 
-static void uvd_v7_0_vm_reg_wait(struct amdgpu_ring *ring,
-				uint32_t data0, uint32_t data1, uint32_t mask)
+static void uvd_v7_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask)
 {
 	struct amdgpu_device *adev = ring->adev;
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
-	amdgpu_ring_write(ring, data0);
+	amdgpu_ring_write(ring, reg << 2);
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
-	amdgpu_ring_write(ring, data1);
+	amdgpu_ring_write(ring, val);
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GP_SCRATCH8), 0));
 	amdgpu_ring_write(ring, mask);
@@ -1271,16 +1271,16 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
-	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
+	data0 = hub->ctx0_ptb_addr_lo32 + vmid * 2;
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
-	uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
+	uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
 
 	/* wait for flush */
-	data0 = (hub->vm_inv_eng0_ack + eng) << 2;
+	data0 = hub->vm_inv_eng0_ack + eng;
 	data1 = 1 << vmid;
 	mask =  1 << vmid;
-	uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
+	uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
 static void uvd_v7_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
@@ -1298,6 +1298,16 @@ static void uvd_v7_0_enc_ring_insert_end(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, HEVC_ENC_CMD_END);
 }
 
+static void uvd_v7_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
+					    uint32_t reg, uint32_t val,
+					    uint32_t mask)
+{
+	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
+	amdgpu_ring_write(ring,	reg << 2);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, val);
+}
+
 static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    unsigned int vmid, unsigned pasid,
 					    uint64_t pd_addr)
@@ -1308,16 +1318,12 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
-	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
-	amdgpu_ring_write(ring, 0xffffffff);
-	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
+	uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
+					lower_32_bits(pd_addr), 0xffffffff);
 
 	/* wait for flush */
-	amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
-	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vmid);
-	amdgpu_ring_write(ring, 1 << vmid);
+	uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
+					1 << vmid, 1 << vmid);
 }
 
 static void uvd_v7_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1676,6 +1682,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
 	.begin_use = amdgpu_uvd_ring_begin_use,
 	.end_use = amdgpu_uvd_ring_end_use,
 	.emit_wreg = uvd_v7_0_ring_emit_wreg,
+	.emit_reg_wait = uvd_v7_0_ring_emit_reg_wait,
 };
 
 static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
@@ -1704,6 +1711,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
 	.begin_use = amdgpu_uvd_ring_begin_use,
 	.end_use = amdgpu_uvd_ring_end_use,
 	.emit_wreg = uvd_v7_0_enc_ring_emit_wreg,
+	.emit_reg_wait = uvd_v7_0_enc_ring_emit_reg_wait,
 };
 
 static void uvd_v7_0_set_ring_funcs(struct amdgpu_device *adev)
-- 
2.14.1

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

[PATCH 5/7] drm/amdgpu: implement vce_v4_0_emit_reg_wait v2

[Christian König]

Add emit_reg_wait implementation for VCE v4.

v2: call new function directly from existing code

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vce_v4_0.c | 22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index e62a24b90aaf..2a4f73ddea97 100755
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -965,6 +965,15 @@ static void vce_v4_0_ring_insert_end(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, VCE_CMD_END);
 }
 
+static void vce_v4_0_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val, uint32_t mask)
+{
+	amdgpu_ring_write(ring, VCE_CMD_REG_WAIT);
+	amdgpu_ring_write(ring,	reg << 2);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, val);
+}
+
 static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 				   unsigned int vmid, unsigned pasid,
 				   uint64_t pd_addr)
@@ -975,16 +984,12 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
-	amdgpu_ring_write(ring, VCE_CMD_REG_WAIT);
-	amdgpu_ring_write(ring,	(hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
-	amdgpu_ring_write(ring, 0xffffffff);
-	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
+	vce_v4_0_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
+			       lower_32_bits(pd_addr), 0xffffffff);
 
 	/* wait for flush */
-	amdgpu_ring_write(ring, VCE_CMD_REG_WAIT);
-	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vmid);
-	amdgpu_ring_write(ring, 1 << vmid);
+	vce_v4_0_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
+			       1 << vmid, 1 << vmid);
 }
 
 static void vce_v4_0_emit_wreg(struct amdgpu_ring *ring,
@@ -1079,6 +1084,7 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
 	.begin_use = amdgpu_vce_ring_begin_use,
 	.end_use = amdgpu_vce_ring_end_use,
 	.emit_wreg = vce_v4_0_emit_wreg,
+	.emit_reg_wait = vce_v4_0_emit_reg_wait,
 };
 
 static void vce_v4_0_set_ring_funcs(struct amdgpu_device *adev)
-- 
2.14.1

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

[PATCH 6/7] drm/amdgpu: implement vcn_v1_0_(dec|enc)_ring_emit_reg_wait v2

[Christian König]

Add emit_reg_wait implementation for VCN v1.

v2: cleanup the existing code as well

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 42 +++++++++++++++++++++--------------
 1 file changed, 25 insertions(+), 17 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index 24ebc3e296a6..294a1bfb59df 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -838,17 +838,18 @@ static void vcn_v1_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
-static void vcn_v1_0_dec_vm_reg_wait(struct amdgpu_ring *ring,
-				uint32_t data0, uint32_t data1, uint32_t mask)
+static void vcn_v1_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring,
+					    uint32_t reg, uint32_t val,
+					    uint32_t mask)
 {
 	struct amdgpu_device *adev = ring->adev;
 
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
-	amdgpu_ring_write(ring, data0);
+	amdgpu_ring_write(ring, reg << 2);
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
-	amdgpu_ring_write(ring, data1);
+	amdgpu_ring_write(ring, val);
 	amdgpu_ring_write(ring,
 		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GP_SCRATCH8), 0));
 	amdgpu_ring_write(ring, mask);
@@ -868,16 +869,16 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for register write */
-	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
+	data0 = hub->ctx0_ptb_addr_lo32 + vmid * 2;
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
-	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);
+	vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
 
 	/* wait for flush */
-	data0 = (hub->vm_inv_eng0_ack + eng) << 2;
+	data0 = hub->vm_inv_eng0_ack + eng;
 	data1 = 1 << vmid;
 	mask =  1 << vmid;
-	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);
+	vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
 static void vcn_v1_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -992,6 +993,16 @@ static void vcn_v1_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
+static void vcn_v1_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
+					    uint32_t reg, uint32_t val,
+					    uint32_t mask)
+{
+	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, val);
+}
+
 static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    unsigned int vmid, unsigned pasid,
 					    uint64_t pd_addr)
@@ -1002,17 +1013,12 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
-	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
-	amdgpu_ring_write(ring,
-			  (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
-	amdgpu_ring_write(ring, 0xffffffff);
-	amdgpu_ring_write(ring, lower_32_bits(pd_addr));
+	vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
+					lower_32_bits(pd_addr), 0xffffffff);
 
 	/* wait for flush */
-	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
-	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 1 << vmid);
-	amdgpu_ring_write(ring, 1 << vmid);
+	vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
+					1 << vmid, 1 << vmid);
 }
 
 static void vcn_v1_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1114,6 +1120,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
 	.begin_use = amdgpu_vcn_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
 	.emit_wreg = vcn_v1_0_dec_ring_emit_wreg,
+	.emit_reg_wait = vcn_v1_0_dec_ring_emit_reg_wait,
 };
 
 static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
@@ -1141,6 +1148,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 	.begin_use = amdgpu_vcn_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
 	.emit_wreg = vcn_v1_0_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v1_0_enc_ring_emit_reg_wait,
 };
 
 static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev)
-- 
2.14.1

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

[PATCH 7/7] drm/amdgpu: move waiting for VM flush into gmc_v9_0_emit_flush_gpu_tlb

[Christian König]

Keep that at a common place instead of spread over all engines.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c  | 19 +++++++++----------
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c  |  4 ++++
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 18 +++---------------
 drivers/gpu/drm/amd/amdgpu/soc15.h     |  3 ++-
 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c  | 20 ++++++--------------
 drivers/gpu/drm/amd/amdgpu/vce_v4_0.c  |  9 +++------
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c  | 20 ++++++--------------
 7 files changed, 33 insertions(+), 60 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 801d4a1dd7db..f7363f821cff 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -3679,15 +3679,8 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					unsigned vmid, unsigned pasid,
 					uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
-
 	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
-	/* wait for the invalidate to complete */
-	gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, hub->vm_inv_eng0_ack + eng,
-			      0, 1 << vmid, 1 << vmid, 0x20);
-
 	/* compute doesn't have PFP */
 	if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) {
 		/* sync PFP to ME, otherwise we might get invalid PFP reads */
@@ -4295,7 +4288,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 	.emit_frame_size = /* totally 242 maximum if 16 IBs */
 		5 +  /* COND_EXEC */
 		7 +  /* PIPELINE_SYNC */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* VM_FLUSH */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* VM_FLUSH */
 		8 +  /* FENCE for VM_FLUSH */
 		20 + /* GDS switch */
 		4 + /* double SWITCH_BUFFER,
@@ -4344,7 +4339,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 		7 + /* gfx_v9_0_ring_emit_hdp_flush */
 		5 + /* hdp invalidate */
 		7 + /* gfx_v9_0_ring_emit_pipeline_sync */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v9_0_ring_emit_vm_flush */
 		8 + 8 + 8, /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */
 	.emit_ib_size =	4, /* gfx_v9_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v9_0_ring_emit_ib_compute,
@@ -4376,7 +4373,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 		7 + /* gfx_v9_0_ring_emit_hdp_flush */
 		5 + /* hdp invalidate */
 		7 + /* gfx_v9_0_ring_emit_pipeline_sync */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v9_0_ring_emit_vm_flush */
 		8 + 8 + 8, /* gfx_v9_0_ring_emit_fence_kiq x3 for user fence, vm fence */
 	.emit_ib_size =	4, /* gfx_v9_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v9_0_ring_emit_ib_compute,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 2b251df94684..2c60981d2eec 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -395,6 +395,10 @@ static uint64_t gmc_v9_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
 
 	amdgpu_ring_emit_wreg(ring, hub->vm_inv_eng0_req + eng, req);
 
+	/* wait for the invalidate to complete */
+	amdgpu_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
+				  1 << vmid, 1 << vmid);
+
 	return pd_addr;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index e1ae39f86adf..ce599fd24412 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -1126,21 +1126,7 @@ static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					 unsigned vmid, unsigned pasid,
 					 uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
-
 	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
-
-	/* wait for flush */
-	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
-			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
-			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
-	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
-	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, 1 << vmid); /* reference */
-	amdgpu_ring_write(ring, 1 << vmid); /* mask */
-	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
-			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
 }
 
 static void sdma_v4_0_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1589,7 +1575,9 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
 		6 + /* sdma_v4_0_ring_emit_hdp_flush */
 		3 + /* hdp invalidate */
 		6 + /* sdma_v4_0_ring_emit_pipeline_sync */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 6 + /* sdma_v4_0_ring_emit_vm_flush */
+		/* sdma_v4_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
 		10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
 	.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
 	.emit_ib = sdma_v4_0_ring_emit_ib,
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h
index 0f63ed674d0b..f70da8a29f86 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.h
@@ -27,7 +27,8 @@
 #include "nbio_v6_1.h"
 #include "nbio_v7_0.h"
 
-#define SOC15_FLUSH_GPU_TLB_NUM_WREG	4
+#define SOC15_FLUSH_GPU_TLB_NUM_WREG		4
+#define SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT	1
 
 extern const struct amd_ip_funcs soc15_common_ip_funcs;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index fcec0bea101d..bf16440e7258 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -1265,7 +1265,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
@@ -1275,12 +1274,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
 	uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
-
-	/* wait for flush */
-	data0 = hub->vm_inv_eng0_ack + eng;
-	data1 = 1 << vmid;
-	mask =  1 << vmid;
-	uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
 static void uvd_v7_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
@@ -1313,17 +1306,12 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
 	uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
 					lower_32_bits(pd_addr), 0xffffffff);
-
-	/* wait for flush */
-	uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
-					1 << vmid, 1 << vmid);
 }
 
 static void uvd_v7_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1669,7 +1657,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
 	.set_wptr = uvd_v7_0_ring_set_wptr,
 	.emit_frame_size =
 		6 + 6 + /* hdp flush / invalidate */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* uvd_v7_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* uvd_v7_0_ring_emit_vm_flush */
 		14 + 14, /* uvd_v7_0_ring_emit_fence x2 vm fence */
 	.emit_ib_size = 8, /* uvd_v7_0_ring_emit_ib */
 	.emit_ib = uvd_v7_0_ring_emit_ib,
@@ -1696,7 +1686,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
 	.set_wptr = uvd_v7_0_enc_ring_set_wptr,
 	.emit_frame_size =
 		3 + 3 + /* hdp flush / invalidate */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* uvd_v7_0_enc_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* uvd_v7_0_enc_ring_emit_vm_flush */
 		5 + 5 + /* uvd_v7_0_enc_ring_emit_fence x2 vm fence */
 		1, /* uvd_v7_0_enc_ring_insert_end */
 	.emit_ib_size = 5, /* uvd_v7_0_enc_ring_emit_ib */
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index 2a4f73ddea97..22c2067bd849 100755
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -979,17 +979,12 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 				   uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
 	vce_v4_0_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
 			       lower_32_bits(pd_addr), 0xffffffff);
-
-	/* wait for flush */
-	vce_v4_0_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
-			       1 << vmid, 1 << vmid);
 }
 
 static void vce_v4_0_emit_wreg(struct amdgpu_ring *ring,
@@ -1069,7 +1064,9 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
 	.set_wptr = vce_v4_0_ring_set_wptr,
 	.parse_cs = amdgpu_vce_ring_parse_cs_vm,
 	.emit_frame_size =
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vce_v4_0_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vce_v4_0_emit_vm_flush */
 		5 + 5 + /* amdgpu_vce_ring_emit_fence x2 vm fence */
 		1, /* vce_v4_0_ring_insert_end */
 	.emit_ib_size = 5, /* vce_v4_0_ring_emit_ib */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index 294a1bfb59df..d9f597c36b63 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -863,7 +863,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
@@ -873,12 +872,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	data1 = lower_32_bits(pd_addr);
 	mask = 0xffffffff;
 	vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
-
-	/* wait for flush */
-	data0 = hub->vm_inv_eng0_ack + eng;
-	data1 = 1 << vmid;
-	mask =  1 << vmid;
-	vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
 static void vcn_v1_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1008,17 +1001,12 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	unsigned eng = ring->vm_inv_eng;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
 
 	/* wait for reg writes */
 	vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
 					lower_32_bits(pd_addr), 0xffffffff);
-
-	/* wait for flush */
-	vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
-					1 << vmid, 1 << vmid);
 }
 
 static void vcn_v1_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1104,7 +1092,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
 	.set_wptr = vcn_v1_0_dec_ring_set_wptr,
 	.emit_frame_size =
 		6 + 6 + /* hdp invalidate / flush */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* vcn_v1_0_dec_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* vcn_v1_0_dec_ring_emit_vm_flush */
 		14 + 14 + /* vcn_v1_0_dec_ring_emit_fence x2 vm fence */
 		6,
 	.emit_ib_size = 8, /* vcn_v1_0_dec_ring_emit_ib */
@@ -1133,7 +1123,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 	.get_wptr = vcn_v1_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v1_0_enc_ring_set_wptr,
 	.emit_frame_size =
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vcn_v1_0_enc_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v1_0_enc_ring_emit_vm_flush */
 		5 + 5 + /* vcn_v1_0_enc_ring_emit_fence x2 vm fence */
 		1, /* vcn_v1_0_enc_ring_insert_end */
 	.emit_ib_size = 5, /* vcn_v1_0_enc_ring_emit_ib */
-- 
2.14.1

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

Re: [PATCH 7/7] drm/amdgpu: move waiting for VM flush into gmc_v9_0_emit_flush_gpu_tlb

[Alex Deucher]

On Tue, Jan 30, 2018 at 7:09 AM, Christian König
<ckoenig.leichtzumerken@gmail.com> wrote:
> Keep that at a common place instead of spread over all engines.
>
> Signed-off-by: Christian König <christian.koenig@amd.com>

Series is:
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

> ---
>  drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c  | 19 +++++++++----------
>  drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c  |  4 ++++
>  drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 18 +++---------------
>  drivers/gpu/drm/amd/amdgpu/soc15.h     |  3 ++-
>  drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c  | 20 ++++++--------------
>  drivers/gpu/drm/amd/amdgpu/vce_v4_0.c  |  9 +++------
>  drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c  | 20 ++++++--------------
>  7 files changed, 33 insertions(+), 60 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> index 801d4a1dd7db..f7363f821cff 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> @@ -3679,15 +3679,8 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                         unsigned vmid, unsigned pasid,
>                                         uint64_t pd_addr)
>  {
> -       struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
> -
>         amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
>
> -       /* wait for the invalidate to complete */
> -       gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, hub->vm_inv_eng0_ack + eng,
> -                             0, 1 << vmid, 1 << vmid, 0x20);
> -
>         /* compute doesn't have PFP */
>         if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) {
>                 /* sync PFP to ME, otherwise we might get invalid PFP reads */
> @@ -4295,7 +4288,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
>         .emit_frame_size = /* totally 242 maximum if 16 IBs */
>                 5 +  /* COND_EXEC */
>                 7 +  /* PIPELINE_SYNC */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* VM_FLUSH */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
> +               2 + /* VM_FLUSH */
>                 8 +  /* FENCE for VM_FLUSH */
>                 20 + /* GDS switch */
>                 4 + /* double SWITCH_BUFFER,
> @@ -4344,7 +4339,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
>                 7 + /* gfx_v9_0_ring_emit_hdp_flush */
>                 5 + /* hdp invalidate */
>                 7 + /* gfx_v9_0_ring_emit_pipeline_sync */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
> +               2 + /* gfx_v9_0_ring_emit_vm_flush */
>                 8 + 8 + 8, /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */
>         .emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */
>         .emit_ib = gfx_v9_0_ring_emit_ib_compute,
> @@ -4376,7 +4373,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
>                 7 + /* gfx_v9_0_ring_emit_hdp_flush */
>                 5 + /* hdp invalidate */
>                 7 + /* gfx_v9_0_ring_emit_pipeline_sync */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
> +               2 + /* gfx_v9_0_ring_emit_vm_flush */
>                 8 + 8 + 8, /* gfx_v9_0_ring_emit_fence_kiq x3 for user fence, vm fence */
>         .emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */
>         .emit_ib = gfx_v9_0_ring_emit_ib_compute,
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> index 2b251df94684..2c60981d2eec 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
> @@ -395,6 +395,10 @@ static uint64_t gmc_v9_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
>
>         amdgpu_ring_emit_wreg(ring, hub->vm_inv_eng0_req + eng, req);
>
> +       /* wait for the invalidate to complete */
> +       amdgpu_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
> +                                 1 << vmid, 1 << vmid);
> +
>         return pd_addr;
>  }
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
> index e1ae39f86adf..ce599fd24412 100644
> --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
> @@ -1126,21 +1126,7 @@ static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                          unsigned vmid, unsigned pasid,
>                                          uint64_t pd_addr)
>  {
> -       struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
> -
>         amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
> -
> -       /* wait for flush */
> -       amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
> -                         SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
> -                         SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
> -       amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
> -       amdgpu_ring_write(ring, 0);
> -       amdgpu_ring_write(ring, 1 << vmid); /* reference */
> -       amdgpu_ring_write(ring, 1 << vmid); /* mask */
> -       amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
> -                         SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
>  }
>
>  static void sdma_v4_0_ring_emit_wreg(struct amdgpu_ring *ring,
> @@ -1589,7 +1575,9 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
>                 6 + /* sdma_v4_0_ring_emit_hdp_flush */
>                 3 + /* hdp invalidate */
>                 6 + /* sdma_v4_0_ring_emit_pipeline_sync */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 6 + /* sdma_v4_0_ring_emit_vm_flush */
> +               /* sdma_v4_0_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
>                 10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
>         .emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
>         .emit_ib = sdma_v4_0_ring_emit_ib,
> diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h
> index 0f63ed674d0b..f70da8a29f86 100644
> --- a/drivers/gpu/drm/amd/amdgpu/soc15.h
> +++ b/drivers/gpu/drm/amd/amdgpu/soc15.h
> @@ -27,7 +27,8 @@
>  #include "nbio_v6_1.h"
>  #include "nbio_v7_0.h"
>
> -#define SOC15_FLUSH_GPU_TLB_NUM_WREG   4
> +#define SOC15_FLUSH_GPU_TLB_NUM_WREG           4
> +#define SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT       1
>
>  extern const struct amd_ip_funcs soc15_common_ip_funcs;
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> index fcec0bea101d..bf16440e7258 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> @@ -1265,7 +1265,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                         uint64_t pd_addr)
>  {
>         struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
>         uint32_t data0, data1, mask;
>
>         pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
> @@ -1275,12 +1274,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
>         data1 = lower_32_bits(pd_addr);
>         mask = 0xffffffff;
>         uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
> -
> -       /* wait for flush */
> -       data0 = hub->vm_inv_eng0_ack + eng;
> -       data1 = 1 << vmid;
> -       mask =  1 << vmid;
> -       uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
>  }
>
>  static void uvd_v7_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
> @@ -1313,17 +1306,12 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                             uint64_t pd_addr)
>  {
>         struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
>
>         pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
>
>         /* wait for reg writes */
>         uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
>                                         lower_32_bits(pd_addr), 0xffffffff);
> -
> -       /* wait for flush */
> -       uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
> -                                       1 << vmid, 1 << vmid);
>  }
>
>  static void uvd_v7_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
> @@ -1669,7 +1657,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
>         .set_wptr = uvd_v7_0_ring_set_wptr,
>         .emit_frame_size =
>                 6 + 6 + /* hdp flush / invalidate */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* uvd_v7_0_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
> +               8 + /* uvd_v7_0_ring_emit_vm_flush */
>                 14 + 14, /* uvd_v7_0_ring_emit_fence x2 vm fence */
>         .emit_ib_size = 8, /* uvd_v7_0_ring_emit_ib */
>         .emit_ib = uvd_v7_0_ring_emit_ib,
> @@ -1696,7 +1686,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
>         .set_wptr = uvd_v7_0_enc_ring_set_wptr,
>         .emit_frame_size =
>                 3 + 3 + /* hdp flush / invalidate */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* uvd_v7_0_enc_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
> +               4 + /* uvd_v7_0_enc_ring_emit_vm_flush */
>                 5 + 5 + /* uvd_v7_0_enc_ring_emit_fence x2 vm fence */
>                 1, /* uvd_v7_0_enc_ring_insert_end */
>         .emit_ib_size = 5, /* uvd_v7_0_enc_ring_emit_ib */
> diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
> index 2a4f73ddea97..22c2067bd849 100755
> --- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
> @@ -979,17 +979,12 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
>                                    uint64_t pd_addr)
>  {
>         struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
>
>         pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
>
>         /* wait for reg writes */
>         vce_v4_0_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
>                                lower_32_bits(pd_addr), 0xffffffff);
> -
> -       /* wait for flush */
> -       vce_v4_0_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
> -                              1 << vmid, 1 << vmid);
>  }
>
>  static void vce_v4_0_emit_wreg(struct amdgpu_ring *ring,
> @@ -1069,7 +1064,9 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
>         .set_wptr = vce_v4_0_ring_set_wptr,
>         .parse_cs = amdgpu_vce_ring_parse_cs_vm,
>         .emit_frame_size =
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vce_v4_0_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
> +               4 + /* vce_v4_0_emit_vm_flush */
>                 5 + 5 + /* amdgpu_vce_ring_emit_fence x2 vm fence */
>                 1, /* vce_v4_0_ring_insert_end */
>         .emit_ib_size = 5, /* vce_v4_0_ring_emit_ib */
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> index 294a1bfb59df..d9f597c36b63 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> @@ -863,7 +863,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                             uint64_t pd_addr)
>  {
>         struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
>         uint32_t data0, data1, mask;
>
>         pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
> @@ -873,12 +872,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
>         data1 = lower_32_bits(pd_addr);
>         mask = 0xffffffff;
>         vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
> -
> -       /* wait for flush */
> -       data0 = hub->vm_inv_eng0_ack + eng;
> -       data1 = 1 << vmid;
> -       mask =  1 << vmid;
> -       vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
>  }
>
>  static void vcn_v1_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
> @@ -1008,17 +1001,12 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
>                                             uint64_t pd_addr)
>  {
>         struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
> -       unsigned eng = ring->vm_inv_eng;
>
>         pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
>
>         /* wait for reg writes */
>         vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
>                                         lower_32_bits(pd_addr), 0xffffffff);
> -
> -       /* wait for flush */
> -       vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
> -                                       1 << vmid, 1 << vmid);
>  }
>
>  static void vcn_v1_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
> @@ -1104,7 +1092,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
>         .set_wptr = vcn_v1_0_dec_ring_set_wptr,
>         .emit_frame_size =
>                 6 + 6 + /* hdp invalidate / flush */
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* vcn_v1_0_dec_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
> +               8 + /* vcn_v1_0_dec_ring_emit_vm_flush */
>                 14 + 14 + /* vcn_v1_0_dec_ring_emit_fence x2 vm fence */
>                 6,
>         .emit_ib_size = 8, /* vcn_v1_0_dec_ring_emit_ib */
> @@ -1133,7 +1123,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
>         .get_wptr = vcn_v1_0_enc_ring_get_wptr,
>         .set_wptr = vcn_v1_0_enc_ring_set_wptr,
>         .emit_frame_size =
> -               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vcn_v1_0_enc_ring_emit_vm_flush */
> +               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
> +               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
> +               4 + /* vcn_v1_0_enc_ring_emit_vm_flush */
>                 5 + 5 + /* vcn_v1_0_enc_ring_emit_fence x2 vm fence */
>                 1, /* vcn_v1_0_enc_ring_insert_end */
>         .emit_ib_size = 5, /* vcn_v1_0_enc_ring_emit_ib */
> --
> 2.14.1
>
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 1/7] drm/amdgpu: add new emit_reg_wait callback

[Christian König]

Felix, Alex anybody brave enough to review this?

Independent of the ATC work it seems like a nice to have cleanup.

Regards,
Christian.

Am 30.01.2018 um 13:09 schrieb Christian König:
> Allows us to wait for a register value/mask on a ring.
>
> Signed-off-by: Christian König <christian.koenig@amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h      | 1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 ++
>   2 files changed, 3 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index d7930f3ead33..787f79c80b6b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1796,6 +1796,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
>   #define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
>   #define amdgpu_ring_emit_rreg(r, d) (r)->funcs->emit_rreg((r), (d))
>   #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r), (d), (v))
> +#define amdgpu_ring_emit_reg_wait(r, d, v, m) (r)->funcs->emit_reg_wait((r), (d), (v), (m))
>   #define amdgpu_ring_emit_tmz(r, b) (r)->funcs->emit_tmz((r), (b))
>   #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
>   #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> index 70d05ec7bc07..867f53332305 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> @@ -150,6 +150,8 @@ struct amdgpu_ring_funcs {
>   	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
>   	void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg);
>   	void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
> +	void (*emit_reg_wait)(struct amdgpu_ring *ring, uint32_t reg,
> +			      uint32_t val, uint32_t mask);
>   	void (*emit_tmz)(struct amdgpu_ring *ring, bool start);
>   	/* priority functions */
>   	void (*set_priority) (struct amdgpu_ring *ring,

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

Re: [PATCH 1/7] drm/amdgpu: add new emit_reg_wait callback

[Felix Kuehling]

On 2018-01-30 07:28 AM, Christian König wrote:
> Felix, Alex anybody brave enough to review this?
>
> Independent of the ATC work it seems like a nice to have cleanup.
Yes. You can add my R-B for the series.

Regards,
  Felix


>
> Regards,
> Christian.
>
> Am 30.01.2018 um 13:09 schrieb Christian König:
>> Allows us to wait for a register value/mask on a ring.
>>
>> Signed-off-by: Christian König <christian.koenig@amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu.h      | 1 +
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 ++
>>   2 files changed, 3 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> index d7930f3ead33..787f79c80b6b 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> @@ -1796,6 +1796,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
>>   #define amdgpu_ring_emit_cntxcntl(r, d)
>> (r)->funcs->emit_cntxcntl((r), (d))
>>   #define amdgpu_ring_emit_rreg(r, d) (r)->funcs->emit_rreg((r), (d))
>>   #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r),
>> (d), (v))
>> +#define amdgpu_ring_emit_reg_wait(r, d, v, m)
>> (r)->funcs->emit_reg_wait((r), (d), (v), (m))
>>   #define amdgpu_ring_emit_tmz(r, b) (r)->funcs->emit_tmz((r), (b))
>>   #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
>>   #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>> index 70d05ec7bc07..867f53332305 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>> @@ -150,6 +150,8 @@ struct amdgpu_ring_funcs {
>>       void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
>>       void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg);
>>       void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg,
>> uint32_t val);
>> +    void (*emit_reg_wait)(struct amdgpu_ring *ring, uint32_t reg,
>> +                  uint32_t val, uint32_t mask);
>>       void (*emit_tmz)(struct amdgpu_ring *ring, bool start);
>>       /* priority functions */
>>       void (*set_priority) (struct amdgpu_ring *ring,
>

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