From 3d7c63849072747eaba6b5d35671bd9cd2d002c1 Mon Sep 17 00:00:00 2001
From: Chunming Zhou <David1.Zhou@amd.com>
Date: Fri, 15 Jul 2016 11:28:30 +0800
Subject: drm/amdgpu: implement gfx8 check_soft_reset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 133 ++++++++++++++++++++--------------
 1 file changed, 78 insertions(+), 55 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index bff8668e9e6d..87232397dbd0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5047,11 +5047,11 @@ static int gfx_v8_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int gfx_v8_0_soft_reset(void *handle)
+static int gfx_v8_0_check_soft_reset(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	/* GRBM_STATUS */
 	tmp = RREG32(mmGRBM_STATUS);
@@ -5060,16 +5060,12 @@ static int gfx_v8_0_soft_reset(void *handle)
 		   GRBM_STATUS__TA_BUSY_MASK | GRBM_STATUS__VGT_BUSY_MASK |
 		   GRBM_STATUS__DB_BUSY_MASK | GRBM_STATUS__CB_BUSY_MASK |
 		   GRBM_STATUS__GDS_BUSY_MASK | GRBM_STATUS__SPI_BUSY_MASK |
-		   GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK)) {
+		   GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK |
+		   GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) {
 		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
 						GRBM_SOFT_RESET, SOFT_RESET_CP, 1);
 		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
 						GRBM_SOFT_RESET, SOFT_RESET_GFX, 1);
-	}
-
-	if (tmp & (GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) {
-		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
-						GRBM_SOFT_RESET, SOFT_RESET_CP, 1);
 		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
 						SRBM_SOFT_RESET, SOFT_RESET_GRBM, 1);
 	}
@@ -5080,73 +5076,99 @@ static int gfx_v8_0_soft_reset(void *handle)
 		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
 						GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
 
+	if (REG_GET_FIELD(tmp, GRBM_STATUS2, CPF_BUSY) ||
+	    REG_GET_FIELD(tmp, GRBM_STATUS2, CPC_BUSY) ||
+	    REG_GET_FIELD(tmp, GRBM_STATUS2, CPG_BUSY)) {
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+						SOFT_RESET_CPF, 1);
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+						SOFT_RESET_CPC, 1);
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+						SOFT_RESET_CPG, 1);
+		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET,
+						SOFT_RESET_GRBM, 1);
+	}
+
 	/* SRBM_STATUS */
 	tmp = RREG32(mmSRBM_STATUS);
 	if (REG_GET_FIELD(tmp, SRBM_STATUS, GRBM_RQ_PENDING))
 		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
 						SRBM_SOFT_RESET, SOFT_RESET_GRBM, 1);
+	if (REG_GET_FIELD(tmp, SRBM_STATUS, SEM_BUSY))
+		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
+						SRBM_SOFT_RESET, SOFT_RESET_SEM, 1);
 
 	if (grbm_soft_reset || srbm_soft_reset) {
-		/* stop the rlc */
-		gfx_v8_0_rlc_stop(adev);
+		adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang = true;
+		adev->gfx.grbm_soft_reset = grbm_soft_reset;
+		adev->gfx.srbm_soft_reset = srbm_soft_reset;
+	} else {
+		adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang = false;
+		adev->gfx.grbm_soft_reset = 0;
+		adev->gfx.srbm_soft_reset = 0;
+	}
 
-		/* Disable GFX parsing/prefetching */
-		gfx_v8_0_cp_gfx_enable(adev, false);
+	return 0;
+}
 
-		/* Disable MEC parsing/prefetching */
-		gfx_v8_0_cp_compute_enable(adev, false);
+static int gfx_v8_0_soft_reset(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
 
-		if (grbm_soft_reset || srbm_soft_reset) {
-			tmp = RREG32(mmGMCON_DEBUG);
-			tmp = REG_SET_FIELD(tmp,
-					    GMCON_DEBUG, GFX_STALL, 1);
-			tmp = REG_SET_FIELD(tmp,
-					    GMCON_DEBUG, GFX_CLEAR, 1);
-			WREG32(mmGMCON_DEBUG, tmp);
+	if (!adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang)
+		return 0;
 
-			udelay(50);
-		}
+	grbm_soft_reset = adev->gfx.grbm_soft_reset;
+	srbm_soft_reset = adev->gfx.srbm_soft_reset;
 
-		if (grbm_soft_reset) {
-			tmp = RREG32(mmGRBM_SOFT_RESET);
-			tmp |= grbm_soft_reset;
-			dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
-			WREG32(mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32(mmGRBM_SOFT_RESET);
+	if (grbm_soft_reset || srbm_soft_reset) {
+		tmp = RREG32(mmGMCON_DEBUG);
+		tmp = REG_SET_FIELD(tmp, GMCON_DEBUG, GFX_STALL, 1);
+		tmp = REG_SET_FIELD(tmp, GMCON_DEBUG, GFX_CLEAR, 1);
+		WREG32(mmGMCON_DEBUG, tmp);
+		udelay(50);
+	}
 
-			udelay(50);
+	if (grbm_soft_reset) {
+		tmp = RREG32(mmGRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmGRBM_SOFT_RESET);
 
-			tmp &= ~grbm_soft_reset;
-			WREG32(mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32(mmGRBM_SOFT_RESET);
-		}
+		udelay(50);
 
-		if (srbm_soft_reset) {
-			tmp = RREG32(mmSRBM_SOFT_RESET);
-			tmp |= srbm_soft_reset;
-			dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
-			WREG32(mmSRBM_SOFT_RESET, tmp);
-			tmp = RREG32(mmSRBM_SOFT_RESET);
+		tmp &= ~grbm_soft_reset;
+		WREG32(mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmGRBM_SOFT_RESET);
+	}
 
-			udelay(50);
+	if (srbm_soft_reset) {
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
 
-			tmp &= ~srbm_soft_reset;
-			WREG32(mmSRBM_SOFT_RESET, tmp);
-			tmp = RREG32(mmSRBM_SOFT_RESET);
-		}
+		udelay(50);
 
-		if (grbm_soft_reset || srbm_soft_reset) {
-			tmp = RREG32(mmGMCON_DEBUG);
-			tmp = REG_SET_FIELD(tmp,
-					    GMCON_DEBUG, GFX_STALL, 0);
-			tmp = REG_SET_FIELD(tmp,
-					    GMCON_DEBUG, GFX_CLEAR, 0);
-			WREG32(mmGMCON_DEBUG, tmp);
-		}
+		tmp &= ~srbm_soft_reset;
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+	}
 
-		/* Wait a little for things to settle down */
-		udelay(50);
+	if (grbm_soft_reset || srbm_soft_reset) {
+		tmp = RREG32(mmGMCON_DEBUG);
+		tmp = REG_SET_FIELD(tmp, GMCON_DEBUG, GFX_STALL, 0);
+		tmp = REG_SET_FIELD(tmp, GMCON_DEBUG, GFX_CLEAR, 0);
+		WREG32(mmGMCON_DEBUG, tmp);
 	}
+
+	/* Wait a little for things to settle down */
+	udelay(50);
+
 	return 0;
 }
 
@@ -6334,6 +6356,7 @@ const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
 	.resume = gfx_v8_0_resume,
 	.is_idle = gfx_v8_0_is_idle,
 	.wait_for_idle = gfx_v8_0_wait_for_idle,
+	.check_soft_reset = gfx_v8_0_check_soft_reset,
 	.soft_reset = gfx_v8_0_soft_reset,
 	.set_clockgating_state = gfx_v8_0_set_clockgating_state,
 	.set_powergating_state = gfx_v8_0_set_powergating_state,
-- 
cgit 


From 1057f20c2252c5eba17e4a9bbf74ea990411f46c Mon Sep 17 00:00:00 2001
From: Chunming Zhou <David1.Zhou@amd.com>
Date: Fri, 15 Jul 2016 15:18:25 +0800
Subject: drm/amdgpu: add gfx8 pre soft reset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 58 +++++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 87232397dbd0..62ba7e550aee 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5111,6 +5111,63 @@ static int gfx_v8_0_check_soft_reset(void *handle)
 	return 0;
 }
 
+static void gfx_v8_0_inactive_hqd(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring)
+{
+	int i;
+
+	vi_srbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+	if (RREG32(mmCP_HQD_ACTIVE) & CP_HQD_ACTIVE__ACTIVE_MASK) {
+		u32 tmp;
+		tmp = RREG32(mmCP_HQD_DEQUEUE_REQUEST);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_DEQUEUE_REQUEST,
+				    DEQUEUE_REQ, 2);
+		WREG32(mmCP_HQD_DEQUEUE_REQUEST, tmp);
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32(mmCP_HQD_ACTIVE) & CP_HQD_ACTIVE__ACTIVE_MASK))
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static int gfx_v8_0_pre_soft_reset(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+
+	if (!adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang)
+		return 0;
+
+	grbm_soft_reset = adev->gfx.grbm_soft_reset;
+	srbm_soft_reset = adev->gfx.srbm_soft_reset;
+
+	/* stop the rlc */
+	gfx_v8_0_rlc_stop(adev);
+
+	if (REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX))
+		/* Disable GFX parsing/prefetching */
+		gfx_v8_0_cp_gfx_enable(adev, false);
+
+	if (REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPF) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPC) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPG)) {
+		int i;
+
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
+
+			gfx_v8_0_inactive_hqd(adev, ring);
+		}
+		/* Disable MEC parsing/prefetching */
+		gfx_v8_0_cp_compute_enable(adev, false);
+	}
+
+       return 0;
+}
+
 static int gfx_v8_0_soft_reset(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -6357,6 +6414,7 @@ const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
 	.is_idle = gfx_v8_0_is_idle,
 	.wait_for_idle = gfx_v8_0_wait_for_idle,
 	.check_soft_reset = gfx_v8_0_check_soft_reset,
+	.pre_soft_reset = gfx_v8_0_pre_soft_reset,
 	.soft_reset = gfx_v8_0_soft_reset,
 	.set_clockgating_state = gfx_v8_0_set_clockgating_state,
 	.set_powergating_state = gfx_v8_0_set_powergating_state,
-- 
cgit 


From e4ae0fc3363191f31fb9627fff9f88d43523aac7 Mon Sep 17 00:00:00 2001
From: Chunming Zhou <David1.Zhou@amd.com>
Date: Fri, 15 Jul 2016 16:24:25 +0800
Subject: drm/amdgpu: implement gfx8 post_soft_reset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c |  1 -
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c      | 44 ++++++++++++++++++++++++++++++
 2 files changed, 44 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 2bd2b19d4666..1e553663e47d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1978,7 +1978,6 @@ static bool amdgpu_need_full_reset(struct amdgpu_device *adev)
 	if (adev->ip_block_status[AMD_IP_BLOCK_TYPE_GMC].hang ||
 	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_IH].hang ||
 	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_SMC].hang ||
-	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang ||
 	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_SDMA].hang ||
 	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_UVD].hang ||
 	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_VCE].hang ||
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 62ba7e550aee..af0efa250d80 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5229,6 +5229,49 @@ static int gfx_v8_0_soft_reset(void *handle)
 	return 0;
 }
 
+static void gfx_v8_0_init_hqd(struct amdgpu_device *adev,
+			      struct amdgpu_ring *ring)
+{
+	vi_srbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+	WREG32(mmCP_HQD_DEQUEUE_REQUEST, 0);
+	WREG32(mmCP_HQD_PQ_RPTR, 0);
+	WREG32(mmCP_HQD_PQ_WPTR, 0);
+	vi_srbm_select(adev, 0, 0, 0, 0);
+}
+
+static int gfx_v8_0_post_soft_reset(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+
+	if (!adev->ip_block_status[AMD_IP_BLOCK_TYPE_GFX].hang)
+		return 0;
+
+	grbm_soft_reset = adev->gfx.grbm_soft_reset;
+	srbm_soft_reset = adev->gfx.srbm_soft_reset;
+
+	if (REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX))
+		gfx_v8_0_cp_gfx_resume(adev);
+
+	if (REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPF) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPC) ||
+	    REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPG)) {
+		int i;
+
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
+
+			gfx_v8_0_init_hqd(adev, ring);
+		}
+		gfx_v8_0_cp_compute_resume(adev);
+	}
+	gfx_v8_0_rlc_start(adev);
+
+	return 0;
+}
+
 /**
  * gfx_v8_0_get_gpu_clock_counter - return GPU clock counter snapshot
  *
@@ -6416,6 +6459,7 @@ const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
 	.check_soft_reset = gfx_v8_0_check_soft_reset,
 	.pre_soft_reset = gfx_v8_0_pre_soft_reset,
 	.soft_reset = gfx_v8_0_soft_reset,
+	.post_soft_reset = gfx_v8_0_post_soft_reset,
 	.set_clockgating_state = gfx_v8_0_set_clockgating_state,
 	.set_powergating_state = gfx_v8_0_set_powergating_state,
 };
-- 
cgit 


From 78bbbd9c0fcde51b56c5bec99dd286cbad698712 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 14 Dec 2015 13:23:57 +0100
Subject: drm/amdgpu: pin shared GWS, GDS and OA resources
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

They can't move anyway, but just to be clean here.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 21 +++++++++------------
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 21 +++++++++------------
 2 files changed, 18 insertions(+), 24 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index d869d058ef24..f4fbec3e224e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -4465,24 +4465,21 @@ static int gfx_v7_0_sw_init(void *handle)
 	}
 
 	/* reserve GDS, GWS and OA resource for gfx */
-	r = amdgpu_bo_create(adev, adev->gds.mem.gfx_partition_size,
-			PAGE_SIZE, true,
-			AMDGPU_GEM_DOMAIN_GDS, 0,
-			NULL, NULL, &adev->gds.gds_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.mem.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GDS,
+				    &adev->gds.gds_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
-	r = amdgpu_bo_create(adev, adev->gds.gws.gfx_partition_size,
-		PAGE_SIZE, true,
-		AMDGPU_GEM_DOMAIN_GWS, 0,
-		NULL, NULL, &adev->gds.gws_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.gws.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GWS,
+				    &adev->gds.gws_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
-	r = amdgpu_bo_create(adev, adev->gds.oa.gfx_partition_size,
-			PAGE_SIZE, true,
-			AMDGPU_GEM_DOMAIN_OA, 0,
-			NULL, NULL, &adev->gds.oa_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.oa.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_OA,
+				    &adev->gds.oa_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index af0efa250d80..ce0929bbf2e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -2078,24 +2078,21 @@ static int gfx_v8_0_sw_init(void *handle)
 	}
 
 	/* reserve GDS, GWS and OA resource for gfx */
-	r = amdgpu_bo_create(adev, adev->gds.mem.gfx_partition_size,
-			PAGE_SIZE, true,
-			AMDGPU_GEM_DOMAIN_GDS, 0, NULL,
-			NULL, &adev->gds.gds_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.mem.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GDS,
+				    &adev->gds.gds_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
-	r = amdgpu_bo_create(adev, adev->gds.gws.gfx_partition_size,
-		PAGE_SIZE, true,
-		AMDGPU_GEM_DOMAIN_GWS, 0, NULL,
-		NULL, &adev->gds.gws_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.gws.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GWS,
+				    &adev->gds.gws_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
-	r = amdgpu_bo_create(adev, adev->gds.oa.gfx_partition_size,
-			PAGE_SIZE, true,
-			AMDGPU_GEM_DOMAIN_OA, 0, NULL,
-			NULL, &adev->gds.oa_gfx_bo);
+	r = amdgpu_bo_create_kernel(adev, adev->gds.oa.gfx_partition_size,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_OA,
+				    &adev->gds.oa_gfx_bo, NULL, NULL);
 	if (r)
 		return r;
 
-- 
cgit 


From 5765a36d4c88d7ea2ff53e35cccca751c74716d8 Mon Sep 17 00:00:00 2001
From: Rex Zhu <Rex.Zhu@amd.com>
Date: Fri, 29 Jul 2016 19:19:26 +0800
Subject: drm/amdgpu: AMD SW workaround for certain partner boards

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index ce0929bbf2e7..7e8cddb6d512 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -699,7 +699,10 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev)
 						 polaris10_golden_common_all,
 						 (const u32)ARRAY_SIZE(polaris10_golden_common_all));
 		WREG32_SMC(ixCG_ACLK_CNTL, 0x0000001C);
-		if (adev->pdev->revision == 0xc7) {
+		if (adev->pdev->revision == 0xc7 &&
+		    ((adev->pdev->subsystem_device == 0xb37 && adev->pdev->subsystem_vendor == 0x1002) ||
+		     (adev->pdev->subsystem_device == 0x4a8 && adev->pdev->subsystem_vendor == 0x1043) ||
+		     (adev->pdev->subsystem_device == 0x9480 && adev->pdev->subsystem_vendor == 0x1682))) {
 			amdgpu_atombios_i2c_channel_trans(adev, 0x10, 0x96, 0x1E, 0xDD);
 			amdgpu_atombios_i2c_channel_trans(adev, 0x10, 0x96, 0x1F, 0xD0);
 		}
-- 
cgit 


From 62d2ce4b09f62c4be62ba7b2cab81d164403bf85 Mon Sep 17 00:00:00 2001
From: Tom St Denis <tom.stdenis@amd.com>
Date: Tue, 9 Aug 2016 10:07:36 -0400
Subject: drm/amd/amdgpu: Correct whitespace in GFX v8
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fix various whitespace issues in gfx v8 driver.

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 41 ++++++++++++++---------------------
 1 file changed, 16 insertions(+), 25 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 7e8cddb6d512..ac29fe6f785c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -1232,10 +1232,9 @@ static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev)
 	if (adev->gfx.rlc.clear_state_obj) {
 		r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
 		if (unlikely(r != 0))
-			dev_warn(adev->dev, "(%d) reserve RLC c bo failed\n", r);
+			dev_warn(adev->dev, "(%d) reserve RLC cbs bo failed\n", r);
 		amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
 		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-
 		amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj);
 		adev->gfx.rlc.clear_state_obj = NULL;
 	}
@@ -1247,7 +1246,6 @@ static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev)
 			dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r);
 		amdgpu_bo_unpin(adev->gfx.rlc.cp_table_obj);
 		amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
-
 		amdgpu_bo_unref(&adev->gfx.rlc.cp_table_obj);
 		adev->gfx.rlc.cp_table_obj = NULL;
 	}
@@ -1289,14 +1287,14 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 				  &adev->gfx.rlc.clear_state_gpu_addr);
 		if (r) {
 			amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-			dev_warn(adev->dev, "(%d) pin RLC c bo failed\n", r);
+			dev_warn(adev->dev, "(%d) pin RLC cbs bo failed\n", r);
 			gfx_v8_0_rlc_fini(adev);
 			return r;
 		}
 
 		r = amdgpu_bo_kmap(adev->gfx.rlc.clear_state_obj, (void **)&adev->gfx.rlc.cs_ptr);
 		if (r) {
-			dev_warn(adev->dev, "(%d) map RLC c bo failed\n", r);
+			dev_warn(adev->dev, "(%d) map RLC cbs bo failed\n", r);
 			gfx_v8_0_rlc_fini(adev);
 			return r;
 		}
@@ -1331,7 +1329,7 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 				  &adev->gfx.rlc.cp_table_gpu_addr);
 		if (r) {
 			amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
-			dev_warn(adev->dev, "(%d) pin RLC cp_table bo failed\n", r);
+			dev_warn(adev->dev, "(%d) pin RLC cp table bo failed\n", r);
 			return r;
 		}
 		r = amdgpu_bo_kmap(adev->gfx.rlc.cp_table_obj, (void **)&adev->gfx.rlc.cp_table_ptr);
@@ -1344,7 +1342,6 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 
 		amdgpu_bo_kunmap(adev->gfx.rlc.cp_table_obj);
 		amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
-
 	}
 
 	return 0;
@@ -1360,7 +1357,6 @@ static void gfx_v8_0_mec_fini(struct amdgpu_device *adev)
 			dev_warn(adev->dev, "(%d) reserve HPD EOP bo failed\n", r);
 		amdgpu_bo_unpin(adev->gfx.mec.hpd_eop_obj);
 		amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj);
-
 		amdgpu_bo_unref(&adev->gfx.mec.hpd_eop_obj);
 		adev->gfx.mec.hpd_eop_obj = NULL;
 	}
@@ -2123,9 +2119,7 @@ static int gfx_v8_0_sw_fini(void *handle)
 		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
 
 	gfx_v8_0_mec_fini(adev);
-
 	gfx_v8_0_rlc_fini(adev);
-
 	gfx_v8_0_free_microcode(adev);
 
 	return 0;
@@ -3581,7 +3575,6 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
 	WREG32(mmDMIF_ADDR_CALC, adev->gfx.config.gb_addr_config);
 
 	gfx_v8_0_tiling_mode_table_init(adev);
-
 	gfx_v8_0_setup_rb(adev);
 	gfx_v8_0_get_cu_info(adev);
 
@@ -3994,14 +3987,13 @@ static int gfx_v8_0_rlc_resume(struct amdgpu_device *adev)
 	/* disable CG */
 	WREG32(mmRLC_CGCG_CGLS_CTRL, 0);
 	if (adev->asic_type == CHIP_POLARIS11 ||
-		adev->asic_type == CHIP_POLARIS10)
+	    adev->asic_type == CHIP_POLARIS10)
 		WREG32(mmRLC_CGCG_CGLS_CTRL_3D, 0);
 
 	/* disable PG */
 	WREG32(mmRLC_PG_CNTL, 0);
 
 	gfx_v8_0_rlc_reset(adev);
-
 	gfx_v8_0_init_pg(adev);
 
 	if (!adev->pp_enabled) {
@@ -4976,7 +4968,6 @@ static int gfx_v8_0_hw_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	gfx_v8_0_init_golden_registers(adev);
-
 	gfx_v8_0_gpu_init(adev);
 
 	r = gfx_v8_0_rlc_resume(adev);
@@ -5548,15 +5539,15 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev,
 
 	data = RREG32(mmRLC_SERDES_WR_CTRL);
 	if (adev->asic_type == CHIP_STONEY)
-			data &= ~(RLC_SERDES_WR_CTRL__WRITE_COMMAND_MASK |
-			RLC_SERDES_WR_CTRL__READ_COMMAND_MASK |
-			RLC_SERDES_WR_CTRL__P1_SELECT_MASK |
-			RLC_SERDES_WR_CTRL__P2_SELECT_MASK |
-			RLC_SERDES_WR_CTRL__RDDATA_RESET_MASK |
-			RLC_SERDES_WR_CTRL__POWER_DOWN_MASK |
-			RLC_SERDES_WR_CTRL__POWER_UP_MASK |
-			RLC_SERDES_WR_CTRL__SHORT_FORMAT_MASK |
-			RLC_SERDES_WR_CTRL__SRBM_OVERRIDE_MASK);
+		data &= ~(RLC_SERDES_WR_CTRL__WRITE_COMMAND_MASK |
+			  RLC_SERDES_WR_CTRL__READ_COMMAND_MASK |
+			  RLC_SERDES_WR_CTRL__P1_SELECT_MASK |
+			  RLC_SERDES_WR_CTRL__P2_SELECT_MASK |
+			  RLC_SERDES_WR_CTRL__RDDATA_RESET_MASK |
+			  RLC_SERDES_WR_CTRL__POWER_DOWN_MASK |
+			  RLC_SERDES_WR_CTRL__POWER_UP_MASK |
+			  RLC_SERDES_WR_CTRL__SHORT_FORMAT_MASK |
+			  RLC_SERDES_WR_CTRL__SRBM_OVERRIDE_MASK);
 	else
 		data &= ~(RLC_SERDES_WR_CTRL__WRITE_COMMAND_MASK |
 			  RLC_SERDES_WR_CTRL__READ_COMMAND_MASK |
@@ -6089,9 +6080,9 @@ static void gfx_v8_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	amdgpu_ring_write(ring,
 #ifdef __BIG_ENDIAN
-					  (2 << 0) |
+				(2 << 0) |
 #endif
-					  (ib->gpu_addr & 0xFFFFFFFC));
+				(ib->gpu_addr & 0xFFFFFFFC));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
 	amdgpu_ring_write(ring, control);
 }
-- 
cgit 


From 5003f2785a20d9caf4040d9d9039644f6ee41892 Mon Sep 17 00:00:00 2001
From: Tom St Denis <tom.stdenis@amd.com>
Date: Tue, 9 Aug 2016 10:10:30 -0400
Subject: drm/amd/amdgpu: Simplify various gfx v8 functions
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 61 ++++++++++++-----------------------
 1 file changed, 20 insertions(+), 41 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index ac29fe6f785c..f0a813c57164 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -3455,19 +3455,16 @@ static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev,
 	else
 		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX, instance);
 
-	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff)) {
-		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_BROADCAST_WRITES, 1);
-		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES, 1);
-	} else if (se_num == 0xffffffff) {
-		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_INDEX, sh_num);
+	if (se_num == 0xffffffff)
 		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES, 1);
-	} else if (sh_num == 0xffffffff) {
-		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_BROADCAST_WRITES, 1);
+	else
 		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX, se_num);
-	} else {
+
+	if (sh_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_BROADCAST_WRITES, 1);
+	else
 		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_INDEX, sh_num);
-		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX, se_num);
-	}
+
 	WREG32(mmGRBM_GFX_INDEX, data);
 }
 
@@ -3480,11 +3477,10 @@ static u32 gfx_v8_0_get_rb_active_bitmap(struct amdgpu_device *adev)
 {
 	u32 data, mask;
 
-	data = RREG32(mmCC_RB_BACKEND_DISABLE);
-	data |= RREG32(mmGC_USER_RB_BACKEND_DISABLE);
+	data =  RREG32(mmCC_RB_BACKEND_DISABLE) |
+		RREG32(mmGC_USER_RB_BACKEND_DISABLE);
 
-	data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK;
-	data >>= GC_USER_RB_BACKEND_DISABLE__BACKEND_DISABLE__SHIFT;
+	data = REG_GET_FIELD(data, GC_USER_RB_BACKEND_DISABLE, BACKEND_DISABLE);
 
 	mask = gfx_v8_0_create_bitmask(adev->gfx.config.max_backends_per_se /
 				       adev->gfx.config.max_sh_per_se);
@@ -4288,12 +4284,10 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
 	gfx_v8_0_cp_gfx_start(adev);
 	ring->ready = true;
 	r = amdgpu_ring_test_ring(ring);
-	if (r) {
+	if (r)
 		ring->ready = false;
-		return r;
-	}
 
-	return 0;
+	return r;
 }
 
 static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
@@ -4975,8 +4969,6 @@ static int gfx_v8_0_hw_init(void *handle)
 		return r;
 
 	r = gfx_v8_0_cp_resume(adev);
-	if (r)
-		return r;
 
 	return r;
 }
@@ -5024,15 +5016,12 @@ static bool gfx_v8_0_is_idle(void *handle)
 static int gfx_v8_0_wait_for_idle(void *handle)
 {
 	unsigned i;
-	u32 tmp;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		/* read MC_STATUS */
-		tmp = RREG32(mmGRBM_STATUS) & GRBM_STATUS__GUI_ACTIVE_MASK;
-
-		if (!REG_GET_FIELD(tmp, GRBM_STATUS, GUI_ACTIVE))
+		if (gfx_v8_0_is_idle(handle))
 			return 0;
+
 		udelay(1);
 	}
 	return -ETIMEDOUT;
@@ -5963,25 +5952,18 @@ static int gfx_v8_0_set_clockgating_state(void *handle,
 
 static u32 gfx_v8_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
 {
-	u32 rptr;
-
-	rptr = ring->adev->wb.wb[ring->rptr_offs];
-
-	return rptr;
+	return ring->adev->wb.wb[ring->rptr_offs];
 }
 
 static u32 gfx_v8_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	u32 wptr;
 
 	if (ring->use_doorbell)
 		/* XXX check if swapping is necessary on BE */
-		wptr = ring->adev->wb.wb[ring->wptr_offs];
+		return ring->adev->wb.wb[ring->wptr_offs];
 	else
-		wptr = RREG32(mmCP_RB0_WPTR);
-
-	return wptr;
+		return RREG32(mmCP_RB0_WPTR);
 }
 
 static void gfx_v8_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
@@ -6591,15 +6573,12 @@ static u32 gfx_v8_0_get_cu_active_bitmap(struct amdgpu_device *adev)
 {
 	u32 data, mask;
 
-	data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG);
-	data |= RREG32(mmGC_USER_SHADER_ARRAY_CONFIG);
-
-	data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
-	data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data =  RREG32(mmCC_GC_SHADER_ARRAY_CONFIG) |
+		RREG32(mmGC_USER_SHADER_ARRAY_CONFIG);
 
 	mask = gfx_v8_0_create_bitmask(adev->gfx.config.max_cu_per_sh);
 
-	return (~data) & mask;
+	return ~REG_GET_FIELD(data, CC_GC_SHADER_ARRAY_CONFIG, INACTIVE_CUS) & mask;
 }
 
 static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev)
-- 
cgit 


From 61cb8cef526d238de2ff278d9e562a6239d8f98a Mon Sep 17 00:00:00 2001
From: Tom St Denis <tom.stdenis@amd.com>
Date: Tue, 9 Aug 2016 10:13:21 -0400
Subject: drm/amd/amdgpu: Simplify bitfield operations in gfx v8
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This patch introduces a new macro WREG32_FIELD which is used
to write to a register with a new value in a field.  It's designed
to replace the pattern:

tmp = RREG32(mmFoo);
tmp &= ~REG__FIELD_MASK;
tmp |= new_value << REG__FIELD__SHIFT;
WREG32(mmFoo, tmp)

with:

WREG32_FIELD(Foo, FIELD, new_value);

Unlike WREG32_P() it understands offsets/masks and doesn't
require the caller to shift the value (or mask properly).

It's applied where suitable in the gfx_v8_0.c driver to start
with.

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h   |   3 +
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 275 ++++++----------------------------
 2 files changed, 48 insertions(+), 230 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 46f250c80a31..5a4af87bb209 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -2210,6 +2210,9 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
 #define REG_GET_FIELD(value, reg, field)				\
 	(((value) & REG_FIELD_MASK(reg, field)) >> REG_FIELD_SHIFT(reg, field))
 
+#define WREG32_FIELD(reg, field, val)	\
+	WREG32(mm##reg, (RREG32(mm##reg) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
+
 /*
  * BIOS helpers.
  */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index f0a813c57164..6022208ead31 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -3562,10 +3562,7 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
 	u32 tmp;
 	int i;
 
-	tmp = RREG32(mmGRBM_CNTL);
-	tmp = REG_SET_FIELD(tmp, GRBM_CNTL, READ_TIMEOUT, 0xff);
-	WREG32(mmGRBM_CNTL, tmp);
-
+	WREG32_FIELD(GRBM_CNTL, READ_TIMEOUT, 0xFF);
 	WREG32(mmGB_ADDR_CONFIG, adev->gfx.config.gb_addr_config);
 	WREG32(mmHDP_ADDR_CONFIG, adev->gfx.config.gb_addr_config);
 	WREG32(mmDMIF_ADDR_CALC, adev->gfx.config.gb_addr_config);
@@ -3754,9 +3751,7 @@ static int gfx_v8_0_init_save_restore_list(struct amdgpu_device *adev)
 				sizeof(indirect_start_offsets)/sizeof(int));
 
 	/* save and restore list */
-	temp = RREG32(mmRLC_SRM_CNTL);
-	temp |= RLC_SRM_CNTL__AUTO_INCR_ADDR_MASK;
-	WREG32(mmRLC_SRM_CNTL, temp);
+	WREG32_FIELD(RLC_SRM_CNTL, AUTO_INCR_ADDR, 1);
 
 	WREG32(mmRLC_SRM_ARAM_ADDR, 0);
 	for (i = 0; i < adev->gfx.rlc.reg_list_size_bytes >> 2; i++)
@@ -3793,11 +3788,7 @@ static int gfx_v8_0_init_save_restore_list(struct amdgpu_device *adev)
 
 static void gfx_v8_0_enable_save_restore_machine(struct amdgpu_device *adev)
 {
-	uint32_t data;
-
-	data = RREG32(mmRLC_SRM_CNTL);
-	data |= RLC_SRM_CNTL__SRM_ENABLE_MASK;
-	WREG32(mmRLC_SRM_CNTL, data);
+	WREG32_FIELD(RLC_SRM_CNTL, SRM_ENABLE, 1);
 }
 
 static void gfx_v8_0_init_power_gating(struct amdgpu_device *adev)
@@ -3807,75 +3798,34 @@ static void gfx_v8_0_init_power_gating(struct amdgpu_device *adev)
 	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
 			      AMD_PG_SUPPORT_GFX_SMG |
 			      AMD_PG_SUPPORT_GFX_DMG)) {
-		data = RREG32(mmCP_RB_WPTR_POLL_CNTL);
-		data &= ~CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK;
-		data |= (0x60 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT);
-		WREG32(mmCP_RB_WPTR_POLL_CNTL, data);
-
-		data = 0;
-		data |= (0x10 << RLC_PG_DELAY__POWER_UP_DELAY__SHIFT);
-		data |= (0x10 << RLC_PG_DELAY__POWER_DOWN_DELAY__SHIFT);
-		data |= (0x10 << RLC_PG_DELAY__CMD_PROPAGATE_DELAY__SHIFT);
-		data |= (0x10 << RLC_PG_DELAY__MEM_SLEEP_DELAY__SHIFT);
-		WREG32(mmRLC_PG_DELAY, data);
+		WREG32_FIELD(CP_RB_WPTR_POLL_CNTL, IDLE_POLL_COUNT, 0x60);
 
-		data = RREG32(mmRLC_PG_DELAY_2);
-		data &= ~RLC_PG_DELAY_2__SERDES_CMD_DELAY_MASK;
-		data |= (0x3 << RLC_PG_DELAY_2__SERDES_CMD_DELAY__SHIFT);
-		WREG32(mmRLC_PG_DELAY_2, data);
+		data = REG_SET_FIELD(0, RLC_PG_DELAY, POWER_UP_DELAY, 0x10);
+		data = REG_SET_FIELD(data, RLC_PG_DELAY, POWER_DOWN_DELAY, 0x10);
+		data = REG_SET_FIELD(data, RLC_PG_DELAY, CMD_PROPAGATE_DELAY, 0x10);
+		data = REG_SET_FIELD(data, RLC_PG_DELAY, MEM_SLEEP_DELAY, 0x10);
+		WREG32(mmRLC_PG_DELAY, data);
 
-		data = RREG32(mmRLC_AUTO_PG_CTRL);
-		data &= ~RLC_AUTO_PG_CTRL__GRBM_REG_SAVE_GFX_IDLE_THRESHOLD_MASK;
-		data |= (0x55f0 << RLC_AUTO_PG_CTRL__GRBM_REG_SAVE_GFX_IDLE_THRESHOLD__SHIFT);
-		WREG32(mmRLC_AUTO_PG_CTRL, data);
+		WREG32_FIELD(RLC_PG_DELAY_2, SERDES_CMD_DELAY, 0x3);
+		WREG32_FIELD(RLC_AUTO_PG_CTRL, GRBM_REG_SAVE_GFX_IDLE_THRESHOLD, 0x55f0);
 	}
 }
 
 static void cz_enable_sck_slow_down_on_power_up(struct amdgpu_device *adev,
 						bool enable)
 {
-	u32 data, orig;
-
-	orig = data = RREG32(mmRLC_PG_CNTL);
-
-	if (enable)
-		data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK;
-
-	if (orig != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, SMU_CLK_SLOWDOWN_ON_PU_ENABLE, enable ? 1 : 0);
 }
 
 static void cz_enable_sck_slow_down_on_power_down(struct amdgpu_device *adev,
 						  bool enable)
 {
-	u32 data, orig;
-
-	orig = data = RREG32(mmRLC_PG_CNTL);
-
-	if (enable)
-		data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK;
-
-	if (orig != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, SMU_CLK_SLOWDOWN_ON_PD_ENABLE, enable ? 1 : 0);
 }
 
 static void cz_enable_cp_power_gating(struct amdgpu_device *adev, bool enable)
 {
-	u32 data, orig;
-
-	orig = data = RREG32(mmRLC_PG_CNTL);
-
-	if (enable)
-		data &= ~RLC_PG_CNTL__CP_PG_DISABLE_MASK;
-	else
-		data |= RLC_PG_CNTL__CP_PG_DISABLE_MASK;
-
-	if (orig != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, CP_PG_DISABLE, enable ? 1 : 0);
 }
 
 static void gfx_v8_0_init_pg(struct amdgpu_device *adev)
@@ -3914,34 +3864,24 @@ static void gfx_v8_0_init_pg(struct amdgpu_device *adev)
 
 void gfx_v8_0_rlc_stop(struct amdgpu_device *adev)
 {
-	u32 tmp = RREG32(mmRLC_CNTL);
-
-	tmp = REG_SET_FIELD(tmp, RLC_CNTL, RLC_ENABLE_F32, 0);
-	WREG32(mmRLC_CNTL, tmp);
+	WREG32_FIELD(RLC_CNTL, RLC_ENABLE_F32, 0);
 
 	gfx_v8_0_enable_gui_idle_interrupt(adev, false);
-
 	gfx_v8_0_wait_for_rlc_serdes(adev);
 }
 
 static void gfx_v8_0_rlc_reset(struct amdgpu_device *adev)
 {
-	u32 tmp = RREG32(mmGRBM_SOFT_RESET);
-
-	tmp = REG_SET_FIELD(tmp, GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
-	WREG32(mmGRBM_SOFT_RESET, tmp);
+	WREG32_FIELD(GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
 	udelay(50);
-	tmp = REG_SET_FIELD(tmp, GRBM_SOFT_RESET, SOFT_RESET_RLC, 0);
-	WREG32(mmGRBM_SOFT_RESET, tmp);
+
+	WREG32_FIELD(GRBM_SOFT_RESET, SOFT_RESET_RLC, 0);
 	udelay(50);
 }
 
 static void gfx_v8_0_rlc_start(struct amdgpu_device *adev)
 {
-	u32 tmp = RREG32(mmRLC_CNTL);
-
-	tmp = REG_SET_FIELD(tmp, RLC_CNTL, RLC_ENABLE_F32, 1);
-	WREG32(mmRLC_CNTL, tmp);
+	WREG32_FIELD(RLC_CNTL, RLC_ENABLE_F32, 1);
 
 	/* carrizo do enable cp interrupt after cp inited */
 	if (!(adev->flags & AMD_IS_APU))
@@ -5367,8 +5307,6 @@ static int gfx_v8_0_late_init(void *handle)
 static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev,
 						       bool enable)
 {
-	uint32_t data, temp;
-
 	if (adev->asic_type == CHIP_POLARIS11)
 		/* Send msg to SMU via Powerplay */
 		amdgpu_set_powergating_state(adev,
@@ -5376,83 +5314,35 @@ static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *ade
 					     enable ?
 					     AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE);
 
-	temp = data = RREG32(mmRLC_PG_CNTL);
-	/* Enable static MGPG */
-	if (enable)
-		data |= RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK;
-
-	if (temp != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, STATIC_PER_CU_PG_ENABLE, enable ? 1 : 0);
 }
 
 static void gfx_v8_0_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *adev,
 							bool enable)
 {
-	uint32_t data, temp;
-
-	temp = data = RREG32(mmRLC_PG_CNTL);
-	/* Enable dynamic MGPG */
-	if (enable)
-		data |= RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK;
-
-	if (temp != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, DYN_PER_CU_PG_ENABLE, enable ? 1 : 0);
 }
 
 static void polaris11_enable_gfx_quick_mg_power_gating(struct amdgpu_device *adev,
 		bool enable)
 {
-	uint32_t data, temp;
-
-	temp = data = RREG32(mmRLC_PG_CNTL);
-	/* Enable quick PG */
-	if (enable)
-		data |= RLC_PG_CNTL__QUICK_PG_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__QUICK_PG_ENABLE_MASK;
-
-	if (temp != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, QUICK_PG_ENABLE, enable ? 1 : 0);
 }
 
 static void cz_enable_gfx_cg_power_gating(struct amdgpu_device *adev,
 					  bool enable)
 {
-	u32 data, orig;
-
-	orig = data = RREG32(mmRLC_PG_CNTL);
-
-	if (enable)
-		data |= RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
-
-	if (orig != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, GFX_POWER_GATING_ENABLE, enable ? 1 : 0);
 }
 
 static void cz_enable_gfx_pipeline_power_gating(struct amdgpu_device *adev,
 						bool enable)
 {
-	u32 data, orig;
-
-	orig = data = RREG32(mmRLC_PG_CNTL);
-
-	if (enable)
-		data |= RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK;
-	else
-		data &= ~RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK;
-
-	if (orig != data)
-		WREG32(mmRLC_PG_CNTL, data);
+	WREG32_FIELD(RLC_PG_CNTL, GFX_PIPELINE_PG_ENABLE, enable ? 1 : 0);
 
 	/* Read any GFX register to wake up GFX. */
 	if (!enable)
-		data = RREG32(mmDB_RENDER_CONTROL);
+		RREG32(mmDB_RENDER_CONTROL);
 }
 
 static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev,
@@ -5559,10 +5449,10 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev,
 
 #define MSG_ENTER_RLC_SAFE_MODE     1
 #define MSG_EXIT_RLC_SAFE_MODE      0
-
-#define RLC_GPR_REG2__REQ_MASK           0x00000001
-#define RLC_GPR_REG2__MESSAGE__SHIFT     0x00000001
-#define RLC_GPR_REG2__MESSAGE_MASK       0x0000001e
+#define RLC_GPR_REG2__REQ_MASK 0x00000001
+#define RLC_GPR_REG2__REQ__SHIFT 0
+#define RLC_GPR_REG2__MESSAGE__SHIFT 0x00000001
+#define RLC_GPR_REG2__MESSAGE_MASK 0x0000001e
 
 static void cz_enter_rlc_safe_mode(struct amdgpu_device *adev)
 {
@@ -5592,7 +5482,7 @@ static void cz_enter_rlc_safe_mode(struct amdgpu_device *adev)
 		}
 
 		for (i = 0; i < adev->usec_timeout; i++) {
-			if ((RREG32(mmRLC_GPR_REG2) & RLC_GPR_REG2__REQ_MASK) == 0)
+			if (!REG_GET_FIELD(RREG32(mmRLC_GPR_REG2), RLC_GPR_REG2, REQ))
 				break;
 			udelay(1);
 		}
@@ -5620,7 +5510,7 @@ static void cz_exit_rlc_safe_mode(struct amdgpu_device *adev)
 	}
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if ((RREG32(mmRLC_GPR_REG2) & RLC_GPR_REG2__REQ_MASK) == 0)
+		if (!REG_GET_FIELD(RREG32(mmRLC_GPR_REG2), RLC_GPR_REG2, REQ))
 			break;
 		udelay(1);
 	}
@@ -5652,7 +5542,7 @@ static void iceland_enter_rlc_safe_mode(struct amdgpu_device *adev)
 		}
 
 		for (i = 0; i < adev->usec_timeout; i++) {
-			if ((RREG32(mmRLC_SAFE_MODE) & RLC_SAFE_MODE__CMD_MASK) == 0)
+			if (!REG_GET_FIELD(RREG32(mmRLC_SAFE_MODE), RLC_SAFE_MODE, CMD))
 				break;
 			udelay(1);
 		}
@@ -5679,7 +5569,7 @@ static void iceland_exit_rlc_safe_mode(struct amdgpu_device *adev)
 	}
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if ((RREG32(mmRLC_SAFE_MODE) & RLC_SAFE_MODE__CMD_MASK) == 0)
+		if (!REG_GET_FIELD(RREG32(mmRLC_SAFE_MODE), RLC_SAFE_MODE, CMD))
 			break;
 		udelay(1);
 	}
@@ -5720,21 +5610,12 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 	/* It is disabled by HW by default */
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
 		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) {
-			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_RLC_LS) {
+			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_RLC_LS)
 				/* 1 - RLC memory Light sleep */
-				temp = data = RREG32(mmRLC_MEM_SLP_CNTL);
-				data |= RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK;
-				if (temp != data)
-					WREG32(mmRLC_MEM_SLP_CNTL, data);
-			}
+				WREG32_FIELD(RLC_MEM_SLP_CNTL, RLC_MEM_LS_EN, 1);
 
-			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS) {
-				/* 2 - CP memory Light sleep */
-				temp = data = RREG32(mmCP_MEM_SLP_CNTL);
-				data |= CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK;
-				if (temp != data)
-					WREG32(mmCP_MEM_SLP_CNTL, data);
-			}
+			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS)
+				WREG32_FIELD(CP_MEM_SLP_CNTL, CP_MEM_LS_EN, 1);
 		}
 
 		/* 3 - RLC_CGTT_MGCG_OVERRIDE */
@@ -6209,33 +6090,14 @@ static void gfx_v8_0_ring_emit_fence_compute(struct amdgpu_ring *ring,
 static void gfx_v8_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
-	u32 cp_int_cntl;
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-					    TIME_STAMP_INT_ENABLE, 0);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl =
-			REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-				      TIME_STAMP_INT_ENABLE, 1);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	default:
-		break;
-	}
+	WREG32_FIELD(CP_INT_CNTL_RING0, TIME_STAMP_INT_ENABLE,
+		     state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
 }
 
 static void gfx_v8_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
 						     int me, int pipe,
 						     enum amdgpu_interrupt_state state)
 {
-	u32 mec_int_cntl, mec_int_cntl_reg;
-
 	/*
 	 * amdgpu controls only pipe 0 of MEC1. That's why this function only
 	 * handles the setting of interrupts for this specific pipe. All other
@@ -6245,7 +6107,6 @@ static void gfx_v8_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
 	if (me == 1) {
 		switch (pipe) {
 		case 0:
-			mec_int_cntl_reg = mmCP_ME1_PIPE0_INT_CNTL;
 			break;
 		default:
 			DRM_DEBUG("invalid pipe %d\n", pipe);
@@ -6256,22 +6117,8 @@ static void gfx_v8_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
 		return;
 	}
 
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		mec_int_cntl = RREG32(mec_int_cntl_reg);
-		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
-					     TIME_STAMP_INT_ENABLE, 0);
-		WREG32(mec_int_cntl_reg, mec_int_cntl);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		mec_int_cntl = RREG32(mec_int_cntl_reg);
-		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
-					     TIME_STAMP_INT_ENABLE, 1);
-		WREG32(mec_int_cntl_reg, mec_int_cntl);
-		break;
-	default:
-		break;
-	}
+	WREG32_FIELD(CP_ME1_PIPE0_INT_CNTL, TIME_STAMP_INT_ENABLE,
+		     state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
 }
 
 static int gfx_v8_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
@@ -6279,24 +6126,8 @@ static int gfx_v8_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
 					     unsigned type,
 					     enum amdgpu_interrupt_state state)
 {
-	u32 cp_int_cntl;
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-					    PRIV_REG_INT_ENABLE, 0);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-					    PRIV_REG_INT_ENABLE, 1);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	default:
-		break;
-	}
+	WREG32_FIELD(CP_INT_CNTL_RING0, PRIV_REG_INT_ENABLE,
+		     state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
 
 	return 0;
 }
@@ -6306,24 +6137,8 @@ static int gfx_v8_0_set_priv_inst_fault_state(struct amdgpu_device *adev,
 					      unsigned type,
 					      enum amdgpu_interrupt_state state)
 {
-	u32 cp_int_cntl;
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-					    PRIV_INSTR_INT_ENABLE, 0);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		cp_int_cntl = RREG32(mmCP_INT_CNTL_RING0);
-		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
-					    PRIV_INSTR_INT_ENABLE, 1);
-		WREG32(mmCP_INT_CNTL_RING0, cp_int_cntl);
-		break;
-	default:
-		break;
-	}
+	WREG32_FIELD(CP_INT_CNTL_RING0, PRIV_INSTR_INT_ENABLE,
+		     state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
 
 	return 0;
 }
-- 
cgit 


From c2167a659cc91075417018b88437da86633bb455 Mon Sep 17 00:00:00 2001
From: Monk Liu <Monk.Liu@amd.com>
Date: Fri, 26 Aug 2016 14:12:37 +0800
Subject: drm/amdgpu:add switch buffer to end of CS (v2)

sync switch buffer scheme with windows kmd for gfx v8,
step1: append a switch_buffer to the end of CS

v2:rebase on latest staging

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h    | 2 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 2 ++
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c  | 7 +++++++
 3 files changed, 11 insertions(+)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 938ef1cb68cc..e31ffe819ec7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -319,6 +319,7 @@ struct amdgpu_ring_funcs {
 	/* note usage for clock and power gating */
 	void (*begin_use)(struct amdgpu_ring *ring);
 	void (*end_use)(struct amdgpu_ring *ring);
+	void (*emit_switch_buffer) (struct amdgpu_ring *ring);
 };
 
 /*
@@ -2254,6 +2255,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_emit_gds_switch(r, v, db, ds, wb, ws, ab, as) (r)->funcs->emit_gds_switch((r), (v), (db), (ds), (wb), (ws), (ab), (as))
 #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r))
 #define amdgpu_ring_emit_hdp_invalidate(r) (r)->funcs->emit_hdp_invalidate((r))
+#define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r))
 #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))
 #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index f5810f700668..11f2fba4653d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -210,6 +210,8 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		amdgpu_ring_patch_cond_exec(ring, patch_offset);
 
 	ring->current_ctx = ctx;
+	if (ring->funcs->emit_switch_buffer)
+		amdgpu_ring_emit_switch_buffer(ring);
 	amdgpu_ring_commit(ring);
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 6022208ead31..b8bc413a8e32 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -6087,6 +6087,12 @@ static void gfx_v8_0_ring_emit_fence_compute(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, upper_32_bits(seq));
 }
 
+static void gfx_v8_ring_emit_sb(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+	amdgpu_ring_write(ring, 0);
+}
+
 static void gfx_v8_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
@@ -6268,6 +6274,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 	.test_ib = gfx_v8_0_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_switch_buffer = gfx_v8_ring_emit_sb,
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
-- 
cgit 


From e7706b422c087da851448c420c90821622ee1721 Mon Sep 17 00:00:00 2001
From: Tom St Denis <tom.stdenis@amd.com>
Date: Fri, 2 Sep 2016 08:15:42 -0400
Subject: drm/amd/amdgpu: Merge get_rptr functions in gfx8
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index b8bc413a8e32..24480f1f8b5c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5831,7 +5831,7 @@ static int gfx_v8_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static u32 gfx_v8_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
+static u32 gfx_v8_0_ring_get_rptr(struct amdgpu_ring *ring)
 {
 	return ring->adev->wb.wb[ring->rptr_offs];
 }
@@ -6047,11 +6047,6 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	}
 }
 
-static u32 gfx_v8_0_ring_get_rptr_compute(struct amdgpu_ring *ring)
-{
-	return ring->adev->wb.wb[ring->rptr_offs];
-}
-
 static u32 gfx_v8_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
 {
 	return ring->adev->wb.wb[ring->wptr_offs];
@@ -6259,7 +6254,7 @@ const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
-	.get_rptr = gfx_v8_0_ring_get_rptr_gfx,
+	.get_rptr = gfx_v8_0_ring_get_rptr,
 	.get_wptr = gfx_v8_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v8_0_ring_set_wptr_gfx,
 	.parse_cs = NULL,
@@ -6278,7 +6273,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
-	.get_rptr = gfx_v8_0_ring_get_rptr_compute,
+	.get_rptr = gfx_v8_0_ring_get_rptr,
 	.get_wptr = gfx_v8_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v8_0_ring_set_wptr_compute,
 	.parse_cs = NULL,
-- 
cgit 


From d4946ccfd0373610ffa6181cffae236bf07d3aff Mon Sep 17 00:00:00 2001
From: Monk Liu <Monk.Liu@amd.com>
Date: Fri, 26 Aug 2016 14:15:03 +0800
Subject: drm/amdgpu: new method to sync ce&de
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

sync switch buffer scheme with windows kmd for gfx v8,
step2:
Insert 128NOP after&before VM flush to prevent CE vm fault.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 24 ++++++------------------
 1 file changed, 6 insertions(+), 18 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 24480f1f8b5c..c49d06b12e8c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5911,12 +5911,6 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 {
 	u32 header, control = 0;
 
-	/* insert SWITCH_BUFFER packet before first IB in the ring frame */
-	if (ctx_switch) {
-		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
-		amdgpu_ring_write(ring, 0);
-	}
-
 	if (ib->flags & AMDGPU_IB_FLAG_CE)
 		header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2);
 	else
@@ -5986,14 +5980,6 @@ static void gfx_v8_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, seq);
 	amdgpu_ring_write(ring, 0xffffffff);
 	amdgpu_ring_write(ring, 4); /* poll interval */
-
-	if (usepfp) {
-		/* synce CE with ME to prevent CE fetch CEIB before context switch done */
-		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
-		amdgpu_ring_write(ring, 0);
-		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
-		amdgpu_ring_write(ring, 0);
-	}
 }
 
 static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
@@ -6001,6 +5987,10 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 {
 	int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX);
 
+	/* GFX8 emits 128 dw nop to prevent DE do vm_flush before CE finish CEIB */
+	if (usepfp)
+		amdgpu_ring_insert_nop(ring, 128);
+
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)) |
@@ -6040,10 +6030,8 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 		/* sync PFP to ME, otherwise we might get invalid PFP reads */
 		amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
 		amdgpu_ring_write(ring, 0x0);
-		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
-		amdgpu_ring_write(ring, 0);
-		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
-		amdgpu_ring_write(ring, 0);
+		/* GFX8 emits 128 dw nop to prevent CE access VM before vm_flush finish */
+		amdgpu_ring_insert_nop(ring, 128);
 	}
 }
 
-- 
cgit 


From 8640faed5a8a8f55d6decd2d9c9f35ae687b1e4c Mon Sep 17 00:00:00 2001
From: Junwei Zhang <Jerry.Zhang@amd.com>
Date: Wed, 7 Sep 2016 17:14:46 +0800
Subject: drm/amdgpu: free the BO in kernel by helper amdgpu_bo_free_kernel()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c   | 16 +++-------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c | 19 ++++---------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c  | 15 +++------------
 drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c    |  6 +++---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c    |  6 +++---
 5 files changed, 16 insertions(+), 46 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
index 5ebb3f43feb6..3ab4c65ecc8b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
@@ -119,8 +119,6 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
  */
 void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
 {
-	int r;
-
 	if (adev->irq.ih.use_bus_addr) {
 		if (adev->irq.ih.ring) {
 			/* add 8 bytes for the rptr/wptr shadows and
@@ -132,17 +130,9 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
 			adev->irq.ih.ring = NULL;
 		}
 	} else {
-		if (adev->irq.ih.ring_obj) {
-			r = amdgpu_bo_reserve(adev->irq.ih.ring_obj, false);
-			if (likely(r == 0)) {
-				amdgpu_bo_kunmap(adev->irq.ih.ring_obj);
-				amdgpu_bo_unpin(adev->irq.ih.ring_obj);
-				amdgpu_bo_unreserve(adev->irq.ih.ring_obj);
-			}
-			amdgpu_bo_unref(&adev->irq.ih.ring_obj);
-			adev->irq.ih.ring = NULL;
-			adev->irq.ih.ring_obj = NULL;
-		}
+		amdgpu_bo_free_kernel(&adev->irq.ih.ring_obj,
+				      &adev->irq.ih.gpu_addr,
+				      (void **)&adev->irq.ih.ring);
 		amdgpu_wb_free(adev, adev->irq.ih.wptr_offs);
 		amdgpu_wb_free(adev, adev->irq.ih.rptr_offs);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index 242ba04bfde6..777f11b63b4c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -252,28 +252,17 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
  */
 void amdgpu_ring_fini(struct amdgpu_ring *ring)
 {
-	int r;
-	struct amdgpu_bo *ring_obj;
-
-	ring_obj = ring->ring_obj;
 	ring->ready = false;
-	ring->ring = NULL;
-	ring->ring_obj = NULL;
 
 	amdgpu_wb_free(ring->adev, ring->cond_exe_offs);
 	amdgpu_wb_free(ring->adev, ring->fence_offs);
 	amdgpu_wb_free(ring->adev, ring->rptr_offs);
 	amdgpu_wb_free(ring->adev, ring->wptr_offs);
 
-	if (ring_obj) {
-		r = amdgpu_bo_reserve(ring_obj, false);
-		if (likely(r == 0)) {
-			amdgpu_bo_kunmap(ring_obj);
-			amdgpu_bo_unpin(ring_obj);
-			amdgpu_bo_unreserve(ring_obj);
-		}
-		amdgpu_bo_unref(&ring_obj);
-	}
+	amdgpu_bo_free_kernel(&ring->ring_obj,
+			      &ring->gpu_addr,
+			      (void **)&ring->ring);
+
 	amdgpu_debugfs_ring_fini(ring);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index cc766cc53a87..a71efeb5083b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -249,22 +249,13 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 
 int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 {
-	int r;
-
 	kfree(adev->uvd.saved_bo);
 
 	amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity);
 
-	if (adev->uvd.vcpu_bo) {
-		r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false);
-		if (!r) {
-			amdgpu_bo_kunmap(adev->uvd.vcpu_bo);
-			amdgpu_bo_unpin(adev->uvd.vcpu_bo);
-			amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
-		}
-
-		amdgpu_bo_unref(&adev->uvd.vcpu_bo);
-	}
+	amdgpu_bo_free_kernel(&adev->uvd.vcpu_bo,
+			      &adev->uvd.gpu_addr,
+			      (void **)&adev->uvd.cpu_addr);
 
 	amdgpu_ring_fini(&adev->uvd.ring);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index 1c3bb2da406e..a93a803b659e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -4490,9 +4490,9 @@ static int gfx_v7_0_sw_fini(void *handle)
 	int i;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	amdgpu_bo_unref(&adev->gds.oa_gfx_bo);
-	amdgpu_bo_unref(&adev->gds.gws_gfx_bo);
-	amdgpu_bo_unref(&adev->gds.gds_gfx_bo);
+	amdgpu_bo_free_kernel(&adev->gds.oa_gfx_bo, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gds.gws_gfx_bo, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gds.gds_gfx_bo, NULL, NULL);
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index c49d06b12e8c..44915056297b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -2109,9 +2109,9 @@ static int gfx_v8_0_sw_fini(void *handle)
 	int i;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	amdgpu_bo_unref(&adev->gds.oa_gfx_bo);
-	amdgpu_bo_unref(&adev->gds.gws_gfx_bo);
-	amdgpu_bo_unref(&adev->gds.gds_gfx_bo);
+	amdgpu_bo_free_kernel(&adev->gds.oa_gfx_bo, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gds.gws_gfx_bo, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gds.gds_gfx_bo, NULL, NULL);
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
-- 
cgit 


From 753ad49c9fdfc732972b0d03f2889f473ed35e59 Mon Sep 17 00:00:00 2001
From: Monk Liu <Monk.Liu@amd.com>
Date: Fri, 26 Aug 2016 13:28:28 +0800
Subject: drm/amdgpu:implement CONTEXT_CONTROL (v5)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

v1:
for gfx8, use CONTEXT_CONTROL package to dynamically
skip preamble CEIB and other load_xxx command in sequence.

v2:
support GFX7 as well.
remove cntxcntl in compute ring funcs because CPC doesn't
support this packet.

v3: fix reduntant judgement in cntxcntl.
v4: some cleanups, don't change cs_submit()
v5: keep old MESA supported & bump up KMS version.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Ack-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  8 ++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  8 ++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c |  3 ++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 12 +++++++++++-
 drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   | 20 ++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   | 30 ++++++++++++++++++++++++++++++
 6 files changed, 79 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 10ec29c50077..717c3b4e1d54 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -320,6 +320,7 @@ struct amdgpu_ring_funcs {
 	void (*begin_use)(struct amdgpu_ring *ring);
 	void (*end_use)(struct amdgpu_ring *ring);
 	void (*emit_switch_buffer) (struct amdgpu_ring *ring);
+	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
 };
 
 /*
@@ -966,6 +967,7 @@ struct amdgpu_ctx {
 	spinlock_t		ring_lock;
 	struct fence            **fences;
 	struct amdgpu_ctx_ring	rings[AMDGPU_MAX_RINGS];
+	bool preamble_presented;
 };
 
 struct amdgpu_ctx_mgr {
@@ -1231,6 +1233,10 @@ struct amdgpu_cs_parser {
 	struct amdgpu_bo_list_entry	uf_entry;
 };
 
+#define AMDGPU_PREAMBLE_IB_PRESENT          (1 << 0) /* bit set means command submit involves a preamble IB */
+#define AMDGPU_PREAMBLE_IB_PRESENT_FIRST    (1 << 1) /* bit set means preamble IB is first presented in belonging context */
+#define AMDGPU_HAVE_CTX_SWITCH              (1 << 2) /* bit set means context switch occured */
+
 struct amdgpu_job {
 	struct amd_sched_job    base;
 	struct amdgpu_device	*adev;
@@ -1239,6 +1245,7 @@ struct amdgpu_job {
 	struct amdgpu_sync	sync;
 	struct amdgpu_ib	*ibs;
 	struct fence		*fence; /* the hw fence */
+	uint32_t		preamble_status;
 	uint32_t		num_ibs;
 	void			*owner;
 	uint64_t		fence_ctx; /* the fence_context this job uses */
@@ -2276,6 +2283,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r))
 #define amdgpu_ring_emit_hdp_invalidate(r) (r)->funcs->emit_hdp_invalidate((r))
 #define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r))
+#define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
 #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))
 #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 56bde6436a1f..61b7e25535bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -850,6 +850,14 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 		if (r)
 			return r;
 
+		if (ib->flags & AMDGPU_IB_FLAG_PREAMBLE) {
+			parser->job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT;
+			if (!parser->ctx->preamble_presented) {
+				parser->job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT_FIRST;
+				parser->ctx->preamble_presented = true;
+			}
+		}
+
 		if (parser->job->ring && parser->job->ring != ring)
 			return -EINVAL;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 3bbc0faf48c8..ca3d87aac7fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -55,9 +55,10 @@
  * - 3.3.0 - Add VM support for UVD on supported hardware.
  * - 3.4.0 - Add AMDGPU_INFO_NUM_EVICTIONS.
  * - 3.5.0 - Add support for new UVD_NO_OP register.
+ * - 3.6.0 - kmd involves use CONTEXT_CONTROL in ring buffer.
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	5
+#define KMS_DRIVER_MINOR	6
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 04263f0fd1af..2aa741c2a64c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -125,6 +125,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
 	uint64_t fence_ctx;
+	uint32_t status = 0;
 
 	unsigned i;
 	int r = 0;
@@ -176,11 +177,20 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 
 	skip_preamble = ring->current_ctx == fence_ctx;
 	need_ctx_switch = ring->current_ctx != fence_ctx;
+	if (job && ring->funcs->emit_cntxcntl) {
+		if (need_ctx_switch)
+			status |= AMDGPU_HAVE_CTX_SWITCH;
+		status |= job->preamble_status;
+		amdgpu_ring_emit_cntxcntl(ring, status);
+	}
+
 	for (i = 0; i < num_ibs; ++i) {
 		ib = &ibs[i];
 
 		/* drop preamble IBs if we don't have a context switch */
-		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) && skip_preamble)
+		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) &&
+			skip_preamble &&
+			!(status & AMDGPU_PREAMBLE_IB_PRESENT_FIRST))
 			continue;
 
 		amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index a93a803b659e..8c780f6c1276 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -2096,6 +2096,25 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, control);
 }
 
+static void gfx_v7_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
+{
+	uint32_t dw2 = 0;
+
+	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
+	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		/* set load_global_config & load_global_uconfig */
+		dw2 |= 0x8001;
+		/* set load_cs_sh_regs */
+		dw2 |= 0x01000000;
+		/* set load_per_context_state & load_gfx_sh_regs */
+		dw2 |= 0x10002;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, dw2);
+	amdgpu_ring_write(ring, 0);
+}
+
 /**
  * gfx_v7_0_ring_test_ib - basic ring IB test
  *
@@ -4938,6 +4957,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = {
 	.test_ib = gfx_v7_0_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_cntxcntl = gfx_v7_ring_emit_cntxcntl,
 };
 
 static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 44915056297b..dca8b368728c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -6076,6 +6076,35 @@ static void gfx_v8_ring_emit_sb(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0);
 }
 
+static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
+{
+	uint32_t dw2 = 0;
+
+	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
+	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		/* set load_global_config & load_global_uconfig */
+		dw2 |= 0x8001;
+		/* set load_cs_sh_regs */
+		dw2 |= 0x01000000;
+		/* set load_per_context_state & load_gfx_sh_regs for GFX */
+		dw2 |= 0x10002;
+
+		/* set load_ce_ram if preamble presented */
+		if (AMDGPU_PREAMBLE_IB_PRESENT & flags)
+			dw2 |= 0x10000000;
+	} else {
+		/* still load_ce_ram if this is the first time preamble presented
+		 * although there is no context switch happens.
+		 */
+		if (AMDGPU_PREAMBLE_IB_PRESENT_FIRST & flags)
+			dw2 |= 0x10000000;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, dw2);
+	amdgpu_ring_write(ring, 0);
+}
+
 static void gfx_v8_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
@@ -6258,6 +6287,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_switch_buffer = gfx_v8_ring_emit_sb,
+	.emit_cntxcntl = gfx_v8_ring_emit_cntxcntl,
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
-- 
cgit 


From 761c2e82054fda665bcdec95b9daf3d468c5fd5b Mon Sep 17 00:00:00 2001
From: Baoyou Xie <baoyou.xie@linaro.org>
Date: Sat, 3 Sep 2016 13:57:14 +0800
Subject: drm/amdgpu: mark symbols static where possible
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We get a few warnings when building kernel with W=1:
drivers/gpu/drm/amd/amdgpu/cz_smc.c:51:5: warning: no previous prototype for 'cz_send_msg_to_smc_async' [-Wmissing-prototypes]
drivers/gpu/drm/amd/amdgpu/cz_smc.c:143:5: warning: no previous prototype for 'cz_write_smc_sram_dword' [-Wmissing-prototypes]
drivers/gpu/drm/amd/amdgpu/iceland_smc.c:124:6: warning: no previous prototype for 'iceland_start_smc' [-Wmissing-prototypes]
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c:3926:6: warning: no previous prototype for 'gfx_v8_0_rlc_stop' [-Wmissing-prototypes]
drivers/gpu/drm/amd/amdgpu/amdgpu_job.c:94:6: warning: no previous prototype for 'amdgpu_job_free_cb' [-Wmissing-prototypes]
....

In fact, these functions are only used in the file in which they are
declared and don't need a declaration, but can be made static.
So this patch marks these functions with 'static'.

Reviewed-by: Christian König <christian.koenig@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Baoyou Xie <baoyou.xie@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c  | 4 ++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c   | 2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_job.c  | 2 +-
 drivers/gpu/drm/amd/amdgpu/cz_smc.c      | 4 ++--
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c    | 2 +-
 drivers/gpu/drm/amd/amdgpu/iceland_smc.c | 8 ++++----
 6 files changed, 11 insertions(+), 11 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
index 040993c998d4..f1c53a2b09c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
@@ -616,7 +616,7 @@ static int amdgpu_cgs_irq_put(struct cgs_device *cgs_device, unsigned src_id, un
 	return amdgpu_irq_put(adev, adev->irq.sources[src_id], type);
 }
 
-int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
+static int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
 				  enum amd_ip_block_type block_type,
 				  enum amd_clockgating_state state)
 {
@@ -637,7 +637,7 @@ int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
 	return r;
 }
 
-int amdgpu_cgs_set_powergating_state(struct cgs_device *cgs_device,
+static int amdgpu_cgs_set_powergating_state(struct cgs_device *cgs_device,
 				  enum amd_ip_block_type block_type,
 				  enum amd_powergating_state state)
 {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index f7dccb10d965..9480be45b45f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -431,7 +431,7 @@ static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p,
 	return false;
 }
 
-int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
+static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
 			    struct list_head *validated)
 {
 	struct amdgpu_bo_list_entry *lobj;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index ac8d40168a47..8c5807994073 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -91,7 +91,7 @@ void amdgpu_job_free_resources(struct amdgpu_job *job)
 		amdgpu_ib_free(job->adev, &job->ibs[i], f);
 }
 
-void amdgpu_job_free_cb(struct amd_sched_job *s_job)
+static void amdgpu_job_free_cb(struct amd_sched_job *s_job)
 {
 	struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/cz_smc.c b/drivers/gpu/drm/amd/amdgpu/cz_smc.c
index 69ac373c4d0f..95887e484c51 100644
--- a/drivers/gpu/drm/amd/amdgpu/cz_smc.c
+++ b/drivers/gpu/drm/amd/amdgpu/cz_smc.c
@@ -50,7 +50,7 @@ static struct cz_smu_private_data *cz_smu_get_priv(struct amdgpu_device *adev)
 	return priv;
 }
 
-int cz_send_msg_to_smc_async(struct amdgpu_device *adev, u16 msg)
+static int cz_send_msg_to_smc_async(struct amdgpu_device *adev, u16 msg)
 {
 	int i;
 	u32 content = 0, tmp;
@@ -142,7 +142,7 @@ int cz_read_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
 	return 0;
 }
 
-int cz_write_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
+static int cz_write_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
 						u32 value, u32 limit)
 {
 	int ret;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index dca8b368728c..e2ed71d5fd53 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -3862,7 +3862,7 @@ static void gfx_v8_0_init_pg(struct amdgpu_device *adev)
 	}
 }
 
-void gfx_v8_0_rlc_stop(struct amdgpu_device *adev)
+static void gfx_v8_0_rlc_stop(struct amdgpu_device *adev)
 {
 	WREG32_FIELD(RLC_CNTL, RLC_ENABLE_F32, 0);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c
index 211839913728..ef7c27d7356a 100644
--- a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c
+++ b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c
@@ -121,7 +121,7 @@ out:
 	return result;
 }
 
-void iceland_start_smc(struct amdgpu_device *adev)
+static void iceland_start_smc(struct amdgpu_device *adev)
 {
 	uint32_t val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
 
@@ -129,7 +129,7 @@ void iceland_start_smc(struct amdgpu_device *adev)
 	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
 }
 
-void iceland_reset_smc(struct amdgpu_device *adev)
+static void iceland_reset_smc(struct amdgpu_device *adev)
 {
 	uint32_t val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
 
@@ -145,7 +145,7 @@ static int iceland_program_jump_on_start(struct amdgpu_device *adev)
 	return 0;
 }
 
-void iceland_stop_smc_clock(struct amdgpu_device *adev)
+static void iceland_stop_smc_clock(struct amdgpu_device *adev)
 {
 	uint32_t val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
 
@@ -153,7 +153,7 @@ void iceland_stop_smc_clock(struct amdgpu_device *adev)
 	WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, val);
 }
 
-void iceland_start_smc_clock(struct amdgpu_device *adev)
+static void iceland_start_smc_clock(struct amdgpu_device *adev)
 {
 	uint32_t val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
 
-- 
cgit 


From 7512ad16b532eee288a85a55a41ebf7d0eb8080d Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 16 Sep 2016 10:57:39 -0400
Subject: drm/amdgpu/gfx8: add ring callbacks for ib and dma frame size
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 40 +++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index e2ed71d5fd53..1e7c615c054e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -6105,6 +6105,42 @@ static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 	amdgpu_ring_write(ring, 0);
 }
 
+static unsigned gfx_v8_0_ring_get_emit_ib_size_gfx(struct amdgpu_ring *ring)
+{
+	return
+		4; /* gfx_v8_0_ring_emit_ib_gfx */
+}
+
+static unsigned gfx_v8_0_ring_get_dma_frame_size_gfx(struct amdgpu_ring *ring)
+{
+	return
+		20 + /* gfx_v8_0_ring_emit_gds_switch */
+		7 + /* gfx_v8_0_ring_emit_hdp_flush */
+		5 + /* gfx_v8_0_ring_emit_hdp_invalidate */
+		6 + 6 + 6 +/* gfx_v8_0_ring_emit_fence_gfx x3 for user fence, vm fence */
+		7 + /* gfx_v8_0_ring_emit_pipeline_sync */
+		256 + 19 + /* gfx_v8_0_ring_emit_vm_flush */
+		2 + /* gfx_v8_ring_emit_sb */
+		3; /* gfx_v8_ring_emit_cntxcntl */
+}
+
+static unsigned gfx_v8_0_ring_get_emit_ib_size_compute(struct amdgpu_ring *ring)
+{
+	return
+		4; /* gfx_v8_0_ring_emit_ib_compute */
+}
+
+static unsigned gfx_v8_0_ring_get_dma_frame_size_compute(struct amdgpu_ring *ring)
+{
+	return
+		20 + /* gfx_v8_0_ring_emit_gds_switch */
+		7 + /* gfx_v8_0_ring_emit_hdp_flush */
+		5 + /* gfx_v8_0_ring_emit_hdp_invalidate */
+		7 + /* gfx_v8_0_ring_emit_pipeline_sync */
+		17 + /* gfx_v8_0_ring_emit_vm_flush */
+		7 + 7 + 7; /* gfx_v8_0_ring_emit_fence_compute x3 for user fence, vm fence */
+}
+
 static void gfx_v8_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
@@ -6288,6 +6324,8 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_switch_buffer = gfx_v8_ring_emit_sb,
 	.emit_cntxcntl = gfx_v8_ring_emit_cntxcntl,
+	.get_emit_ib_size = gfx_v8_0_ring_get_emit_ib_size_gfx,
+	.get_dma_frame_size = gfx_v8_0_ring_get_dma_frame_size_gfx,
 };
 
 static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
@@ -6306,6 +6344,8 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
 	.test_ib = gfx_v8_0_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.get_emit_ib_size = gfx_v8_0_ring_get_emit_ib_size_compute,
+	.get_dma_frame_size = gfx_v8_0_ring_get_dma_frame_size_compute,
 };
 
 static void gfx_v8_0_set_ring_funcs(struct amdgpu_device *adev)
-- 
cgit 


From 167ac5733c70683c4886a8b3ef347cc6c93a76a6 Mon Sep 17 00:00:00 2001
From: Huang Rui <ray.huang@amd.com>
Date: Wed, 14 Sep 2016 17:14:16 +0800
Subject: drm/amdgpu: implement raster configuration for gfx v8

This patch is to implement the raster configuration and harvested
configuration of gfx v8.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 168 +++++++++++++++++++++++++++++++++-
 drivers/gpu/drm/amd/amdgpu/vid.h      |  37 ++++++++
 2 files changed, 204 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 1e7c615c054e..6da7d94925eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -3488,13 +3488,163 @@ static u32 gfx_v8_0_get_rb_active_bitmap(struct amdgpu_device *adev)
 	return (~data) & mask;
 }
 
+static void
+gfx_v8_0_raster_config(struct amdgpu_device *adev, u32 *rconf, u32 *rconf1)
+{
+	switch (adev->asic_type) {
+	case CHIP_FIJI:
+		*rconf |= RB_MAP_PKR0(2) | RB_MAP_PKR1(2) |
+			  RB_XSEL2(1) | PKR_MAP(2) |
+			  PKR_XSEL(1) | PKR_YSEL(1) |
+			  SE_MAP(2) | SE_XSEL(2) | SE_YSEL(3);
+		*rconf1 |= SE_PAIR_MAP(2) | SE_PAIR_XSEL(3) |
+			   SE_PAIR_YSEL(2);
+		break;
+	case CHIP_TONGA:
+	case CHIP_POLARIS10:
+		*rconf |= RB_MAP_PKR0(2) | RB_XSEL2(1) | SE_MAP(2) |
+			  SE_XSEL(1) | SE_YSEL(1);
+		*rconf1 |= SE_PAIR_MAP(2) | SE_PAIR_XSEL(2) |
+			   SE_PAIR_YSEL(2);
+		break;
+	case CHIP_TOPAZ:
+	case CHIP_CARRIZO:
+		*rconf |= RB_MAP_PKR0(2);
+		*rconf1 |= 0x0;
+		break;
+	case CHIP_POLARIS11:
+		*rconf |= RB_MAP_PKR0(2) | RB_XSEL2(1) | SE_MAP(2) |
+			  SE_XSEL(1) | SE_YSEL(1);
+		*rconf1 |= 0x0;
+		break;
+	case CHIP_STONEY:
+		*rconf |= 0x0;
+		*rconf1 |= 0x0;
+		break;
+	default:
+		DRM_ERROR("unknown asic: 0x%x\n", adev->asic_type);
+		break;
+	}
+}
+
+static void
+gfx_v8_0_write_harvested_raster_configs(struct amdgpu_device *adev,
+					u32 raster_config, u32 raster_config_1,
+					unsigned rb_mask, unsigned num_rb)
+{
+	unsigned sh_per_se = max_t(unsigned, adev->gfx.config.max_sh_per_se, 1);
+	unsigned num_se = max_t(unsigned, adev->gfx.config.max_shader_engines, 1);
+	unsigned rb_per_pkr = min_t(unsigned, num_rb / num_se / sh_per_se, 2);
+	unsigned rb_per_se = num_rb / num_se;
+	unsigned se_mask[4];
+	unsigned se;
+
+	se_mask[0] = ((1 << rb_per_se) - 1) & rb_mask;
+	se_mask[1] = (se_mask[0] << rb_per_se) & rb_mask;
+	se_mask[2] = (se_mask[1] << rb_per_se) & rb_mask;
+	se_mask[3] = (se_mask[2] << rb_per_se) & rb_mask;
+
+	WARN_ON(!(num_se == 1 || num_se == 2 || num_se == 4));
+	WARN_ON(!(sh_per_se == 1 || sh_per_se == 2));
+	WARN_ON(!(rb_per_pkr == 1 || rb_per_pkr == 2));
+
+	if ((num_se > 2) && ((!se_mask[0] && !se_mask[1]) ||
+			     (!se_mask[2] && !se_mask[3]))) {
+		raster_config_1 &= ~SE_PAIR_MAP_MASK;
+
+		if (!se_mask[0] && !se_mask[1]) {
+			raster_config_1 |=
+				SE_PAIR_MAP(RASTER_CONFIG_SE_PAIR_MAP_3);
+		} else {
+			raster_config_1 |=
+				SE_PAIR_MAP(RASTER_CONFIG_SE_PAIR_MAP_0);
+		}
+	}
+
+	for (se = 0; se < num_se; se++) {
+		unsigned raster_config_se = raster_config;
+		unsigned pkr0_mask = ((1 << rb_per_pkr) - 1) << (se * rb_per_se);
+		unsigned pkr1_mask = pkr0_mask << rb_per_pkr;
+		int idx = (se / 2) * 2;
+
+		if ((num_se > 1) && (!se_mask[idx] || !se_mask[idx + 1])) {
+			raster_config_se &= ~SE_MAP_MASK;
+
+			if (!se_mask[idx]) {
+				raster_config_se |= SE_MAP(RASTER_CONFIG_SE_MAP_3);
+			} else {
+				raster_config_se |= SE_MAP(RASTER_CONFIG_SE_MAP_0);
+			}
+		}
+
+		pkr0_mask &= rb_mask;
+		pkr1_mask &= rb_mask;
+		if (rb_per_se > 2 && (!pkr0_mask || !pkr1_mask)) {
+			raster_config_se &= ~PKR_MAP_MASK;
+
+			if (!pkr0_mask) {
+				raster_config_se |= PKR_MAP(RASTER_CONFIG_PKR_MAP_3);
+			} else {
+				raster_config_se |= PKR_MAP(RASTER_CONFIG_PKR_MAP_0);
+			}
+		}
+
+		if (rb_per_se >= 2) {
+			unsigned rb0_mask = 1 << (se * rb_per_se);
+			unsigned rb1_mask = rb0_mask << 1;
+
+			rb0_mask &= rb_mask;
+			rb1_mask &= rb_mask;
+			if (!rb0_mask || !rb1_mask) {
+				raster_config_se &= ~RB_MAP_PKR0_MASK;
+
+				if (!rb0_mask) {
+					raster_config_se |=
+						RB_MAP_PKR0(RASTER_CONFIG_RB_MAP_3);
+				} else {
+					raster_config_se |=
+						RB_MAP_PKR0(RASTER_CONFIG_RB_MAP_0);
+				}
+			}
+
+			if (rb_per_se > 2) {
+				rb0_mask = 1 << (se * rb_per_se + rb_per_pkr);
+				rb1_mask = rb0_mask << 1;
+				rb0_mask &= rb_mask;
+				rb1_mask &= rb_mask;
+				if (!rb0_mask || !rb1_mask) {
+					raster_config_se &= ~RB_MAP_PKR1_MASK;
+
+					if (!rb0_mask) {
+						raster_config_se |=
+							RB_MAP_PKR1(RASTER_CONFIG_RB_MAP_3);
+					} else {
+						raster_config_se |=
+							RB_MAP_PKR1(RASTER_CONFIG_RB_MAP_0);
+					}
+				}
+			}
+		}
+
+		/* GRBM_GFX_INDEX has a different offset on VI */
+		gfx_v8_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff);
+		WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se);
+		WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1);
+	}
+
+	/* GRBM_GFX_INDEX has a different offset on VI */
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+}
+
 static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 {
 	int i, j;
 	u32 data;
+	u32 raster_config = 0, raster_config_1 = 0;
 	u32 active_rbs = 0;
 	u32 rb_bitmap_width_per_sh = adev->gfx.config.max_backends_per_se /
 					adev->gfx.config.max_sh_per_se;
+	unsigned num_rb_pipes;
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
@@ -3506,10 +3656,26 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 		}
 	}
 	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
-	mutex_unlock(&adev->grbm_idx_mutex);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
 	adev->gfx.config.num_rbs = hweight32(active_rbs);
+
+	num_rb_pipes = min_t(unsigned, adev->gfx.config.max_backends_per_se *
+			     adev->gfx.config.max_shader_engines, 16);
+
+	gfx_v8_0_raster_config(adev, &raster_config, &raster_config_1);
+
+	if (!adev->gfx.config.backend_enable_mask ||
+			adev->gfx.config.num_rbs >= num_rb_pipes) {
+		WREG32(mmPA_SC_RASTER_CONFIG, raster_config);
+		WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1);
+	} else {
+		gfx_v8_0_write_harvested_raster_configs(adev, raster_config, raster_config_1,
+							adev->gfx.config.backend_enable_mask,
+							num_rb_pipes);
+	}
+
+	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
 /**
diff --git a/drivers/gpu/drm/amd/amdgpu/vid.h b/drivers/gpu/drm/amd/amdgpu/vid.h
index f62b261660d4..11746f22d0c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/vid.h
+++ b/drivers/gpu/drm/amd/amdgpu/vid.h
@@ -373,4 +373,41 @@
 #define VCE_CMD_WAIT_GE         0x00000106
 #define VCE_CMD_UPDATE_PTB      0x00000107
 #define VCE_CMD_FLUSH_TLB       0x00000108
+
+/* mmPA_SC_RASTER_CONFIG mask */
+#define RB_MAP_PKR0(x)				((x) << 0)
+#define RB_MAP_PKR0_MASK			(0x3 << 0)
+#define RB_MAP_PKR1(x)				((x) << 2)
+#define RB_MAP_PKR1_MASK			(0x3 << 2)
+#define RB_XSEL2(x)				((x) << 4)
+#define RB_XSEL2_MASK				(0x3 << 4)
+#define RB_XSEL					(1 << 6)
+#define RB_YSEL					(1 << 7)
+#define PKR_MAP(x)				((x) << 8)
+#define PKR_MAP_MASK				(0x3 << 8)
+#define PKR_XSEL(x)				((x) << 10)
+#define PKR_XSEL_MASK				(0x3 << 10)
+#define PKR_YSEL(x)				((x) << 12)
+#define PKR_YSEL_MASK				(0x3 << 12)
+#define SC_MAP(x)				((x) << 16)
+#define SC_MAP_MASK				(0x3 << 16)
+#define SC_XSEL(x)				((x) << 18)
+#define SC_XSEL_MASK				(0x3 << 18)
+#define SC_YSEL(x)				((x) << 20)
+#define SC_YSEL_MASK				(0x3 << 20)
+#define SE_MAP(x)				((x) << 24)
+#define SE_MAP_MASK				(0x3 << 24)
+#define SE_XSEL(x)				((x) << 26)
+#define SE_XSEL_MASK				(0x3 << 26)
+#define SE_YSEL(x)				((x) << 28)
+#define SE_YSEL_MASK				(0x3 << 28)
+
+/* mmPA_SC_RASTER_CONFIG_1 mask */
+#define SE_PAIR_MAP(x)				((x) << 0)
+#define SE_PAIR_MAP_MASK			(0x3 << 0)
+#define SE_PAIR_XSEL(x)				((x) << 2)
+#define SE_PAIR_XSEL_MASK			(0x3 << 2)
+#define SE_PAIR_YSEL(x)				((x) << 4)
+#define SE_PAIR_YSEL_MASK			(0x3 << 4)
+
 #endif
-- 
cgit 


From a8ca34136453b5a570b514e466f8b0b9efd71df2 Mon Sep 17 00:00:00 2001
From: Rex Zhu <Rex.Zhu@amd.com>
Date: Sun, 18 Sep 2016 16:55:00 +0800
Subject: drm/amdgpu: set gfx clock gating for tonga/polaris.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 77 +++++++++++++++++++++++++++++++++++
 1 file changed, 77 insertions(+)

(limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 6da7d94925eb..f490691d4b6c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5979,6 +5979,76 @@ static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int gfx_v8_0_tonga_update_gfx_clock_gating(struct amdgpu_device *adev,
+					  enum amd_clockgating_state state)
+{
+	uint32_t msg_id, pp_state;
+	void *pp_handle = adev->powerplay.pp_handle;
+
+	if (state == AMD_CG_STATE_UNGATE)
+		pp_state = 0;
+	else
+		pp_state = PP_STATE_CG | PP_STATE_LS;
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_CG,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_MG,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	return 0;
+}
+
+static int gfx_v8_0_polaris_update_gfx_clock_gating(struct amdgpu_device *adev,
+					  enum amd_clockgating_state state)
+{
+	uint32_t msg_id, pp_state;
+	void *pp_handle = adev->powerplay.pp_handle;
+
+	if (state == AMD_CG_STATE_UNGATE)
+		pp_state = 0;
+	else
+		pp_state = PP_STATE_CG | PP_STATE_LS;
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_CG,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_3D,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_MG,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_RLC,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	msg_id = PP_CG_MSG_ID(PP_GROUP_GFX,
+			PP_BLOCK_GFX_CP,
+			PP_STATE_SUPPORT_CG | PP_STATE_SUPPORT_LS,
+			pp_state);
+	amd_set_clockgating_by_smu(pp_handle, msg_id);
+
+	return 0;
+}
+
 static int gfx_v8_0_set_clockgating_state(void *handle,
 					  enum amd_clockgating_state state)
 {
@@ -5991,6 +6061,13 @@ static int gfx_v8_0_set_clockgating_state(void *handle,
 		gfx_v8_0_update_gfx_clock_gating(adev,
 						 state == AMD_CG_STATE_GATE ? true : false);
 		break;
+	case CHIP_TONGA:
+		gfx_v8_0_tonga_update_gfx_clock_gating(adev, state);
+		break;
+	case CHIP_POLARIS10:
+	case CHIP_POLARIS11:
+		gfx_v8_0_polaris_update_gfx_clock_gating(adev, state);
+		break;
 	default:
 		break;
 	}
-- 
cgit