1 files changed, 107 insertions, 114 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
index b2706221df99..befe013b11a7 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
@@ -42,6 +42,7 @@
 #include "sdma_common.h"
 #include "sdma_v7_0.h"
 #include "v12_structs.h"
+#include "mes_userqueue.h"
 
 MODULE_FIRMWARE("amdgpu/sdma_7_0_0.bin");
 MODULE_FIRMWARE("amdgpu/sdma_7_0_1.bin");
@@ -204,66 +205,39 @@ static uint64_t sdma_v7_0_ring_get_wptr(struct amdgpu_ring *ring)
 static void sdma_v7_0_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t *wptr_saved;
-	uint32_t *is_queue_unmap;
-	uint64_t aggregated_db_index;
-	uint32_t mqd_size = adev->mqds[AMDGPU_HW_IP_DMA].mqd_size;
 
 	DRM_DEBUG("Setting write pointer\n");
 
-	if (ring->is_mes_queue) {
-		wptr_saved = (uint32_t *)(ring->mqd_ptr + mqd_size);
-		is_queue_unmap = (uint32_t *)(ring->mqd_ptr + mqd_size +
-					      sizeof(uint32_t));
-		aggregated_db_index =
-			amdgpu_mes_get_aggregated_doorbell_index(adev,
-							 ring->hw_prio);
-
+	if (ring->use_doorbell) {
+		DRM_DEBUG("Using doorbell -- "
+			  "wptr_offs == 0x%08x "
+			  "lower_32_bits(ring->wptr) << 2 == 0x%08x "
+			  "upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+			  ring->wptr_offs,
+			  lower_32_bits(ring->wptr << 2),
+			  upper_32_bits(ring->wptr << 2));
+		/* XXX check if swapping is necessary on BE */
 		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
 			     ring->wptr << 2);
-		*wptr_saved = ring->wptr << 2;
-		if (*is_queue_unmap) {
-			WDOORBELL64(aggregated_db_index, ring->wptr << 2);
-			DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
-					ring->doorbell_index, ring->wptr << 2);
-			WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
-		} else {
-			DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
-					ring->doorbell_index, ring->wptr << 2);
-			WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
-		}
+		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
+			  ring->doorbell_index, ring->wptr << 2);
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
 	} else {
-		if (ring->use_doorbell) {
-			DRM_DEBUG("Using doorbell -- "
-				  "wptr_offs == 0x%08x "
-				  "lower_32_bits(ring->wptr) << 2 == 0x%08x "
-				  "upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
-				  ring->wptr_offs,
-				  lower_32_bits(ring->wptr << 2),
-				  upper_32_bits(ring->wptr << 2));
-			/* XXX check if swapping is necessary on BE */
-			atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
-				     ring->wptr << 2);
-			DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
-				  ring->doorbell_index, ring->wptr << 2);
-			WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
-		} else {
-			DRM_DEBUG("Not using doorbell -- "
-				  "regSDMA%i_GFX_RB_WPTR == 0x%08x "
-				  "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
-				  ring->me,
-				  lower_32_bits(ring->wptr << 2),
-				  ring->me,
-				  upper_32_bits(ring->wptr << 2));
-			WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
-								     ring->me,
-								     regSDMA0_QUEUE0_RB_WPTR),
-					lower_32_bits(ring->wptr << 2));
-			WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
-								     ring->me,
-								     regSDMA0_QUEUE0_RB_WPTR_HI),
-					upper_32_bits(ring->wptr << 2));
-		}
+		DRM_DEBUG("Not using doorbell -- "
+			  "regSDMA%i_GFX_RB_WPTR == 0x%08x "
+			  "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
+			  ring->me,
+			  lower_32_bits(ring->wptr << 2),
+			  ring->me,
+			  upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
+							     ring->me,
+							     regSDMA0_QUEUE0_RB_WPTR),
+				lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
+							     ring->me,
+							     regSDMA0_QUEUE0_RB_WPTR_HI),
+				upper_32_bits(ring->wptr << 2));
 	}
 }
 
@@ -407,11 +381,9 @@ static void sdma_v7_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
 	}
 
 	if (flags & AMDGPU_FENCE_FLAG_INT) {
-		uint32_t ctx = ring->is_mes_queue ?
-			(ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0;
 		/* generate an interrupt */
 		amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_TRAP));
-		amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx));
+		amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0));
 	}
 }
 
@@ -935,6 +907,9 @@ static int sdma_v7_0_mqd_init(struct amdgpu_device *adev, void *mqd,
 	m->sdmax_rlcx_rb_aql_cntl = 0x4000;	//regSDMA0_QUEUE0_RB_AQL_CNTL_DEFAULT;
 	m->sdmax_rlcx_dummy_reg = 0xf;	//regSDMA0_QUEUE0_DUMMY_REG_DEFAULT;
 
+	m->sdmax_rlcx_csa_addr_lo = lower_32_bits(prop->csa_addr);
+	m->sdmax_rlcx_csa_addr_hi = upper_32_bits(prop->csa_addr);
+
 	return 0;
 }
 
@@ -961,33 +936,22 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring)
 	int r;
 	u32 tmp;
 	u64 gpu_addr;
-	volatile uint32_t *cpu_ptr = NULL;
 
 	tmp = 0xCAFEDEAD;
 
-	if (ring->is_mes_queue) {
-		uint32_t offset = 0;
-		offset = amdgpu_mes_ctx_get_offs(ring,
-					 AMDGPU_MES_CTX_PADDING_OFFS);
-		gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset);
-		cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset);
-		*cpu_ptr = tmp;
-	} else {
-		r = amdgpu_device_wb_get(adev, &index);
-		if (r) {
-			dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
-			return r;
-		}
-
-		gpu_addr = adev->wb.gpu_addr + (index * 4);
-		adev->wb.wb[index] = cpu_to_le32(tmp);
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
+		return r;
 	}
 
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
 	r = amdgpu_ring_alloc(ring, 5);
 	if (r) {
 		DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r);
-		if (!ring->is_mes_queue)
-			amdgpu_device_wb_free(adev, index);
+		amdgpu_device_wb_free(adev, index);
 		return r;
 	}
 
@@ -1000,10 +964,7 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring)
 	amdgpu_ring_commit(ring);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (ring->is_mes_queue)
-			tmp = le32_to_cpu(*cpu_ptr);
-		else
-			tmp = le32_to_cpu(adev->wb.wb[index]);
+		tmp = le32_to_cpu(adev->wb.wb[index]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		if (amdgpu_emu_mode == 1)
@@ -1015,8 +976,7 @@ static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring)
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
 
-	if (!ring->is_mes_queue)
-		amdgpu_device_wb_free(adev, index);
+	amdgpu_device_wb_free(adev, index);
 
 	return r;
 }
@@ -1039,37 +999,23 @@ static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	long r;
 	u32 tmp = 0;
 	u64 gpu_addr;
-	volatile uint32_t *cpu_ptr = NULL;
 
 	tmp = 0xCAFEDEAD;
 	memset(&ib, 0, sizeof(ib));
 
-	if (ring->is_mes_queue) {
-		uint32_t offset = 0;
-		offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS);
-		ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset);
-		ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset);
-
-		offset = amdgpu_mes_ctx_get_offs(ring,
-					 AMDGPU_MES_CTX_PADDING_OFFS);
-		gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset);
-		cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset);
-		*cpu_ptr = tmp;
-	} else {
-		r = amdgpu_device_wb_get(adev, &index);
-		if (r) {
-			dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
-			return r;
-		}
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
+		return r;
+	}
 
-		gpu_addr = adev->wb.gpu_addr + (index * 4);
-		adev->wb.wb[index] = cpu_to_le32(tmp);
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
 
-		r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
-		if (r) {
-			DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
-			goto err0;
-		}
+	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
+		goto err0;
 	}
 
 	ib.ptr[0] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
@@ -1097,10 +1043,7 @@ static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		goto err1;
 	}
 
-	if (ring->is_mes_queue)
-		tmp = le32_to_cpu(*cpu_ptr);
-	else
-		tmp = le32_to_cpu(adev->wb.wb[index]);
+	tmp = le32_to_cpu(adev->wb.wb[index]);
 
 	if (tmp == 0xDEADBEEF)
 		r = 0;
@@ -1111,8 +1054,7 @@ err1:
 	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
-	if (!ring->is_mes_queue)
-		amdgpu_device_wb_free(adev, index);
+	amdgpu_device_wb_free(adev, index);
 	return r;
 }
 
@@ -1312,6 +1254,23 @@ static int sdma_v7_0_early_init(struct amdgpu_ip_block *ip_block)
 	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
+	switch (amdgpu_user_queue) {
+	case -1:
+	case 0:
+	default:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = true;
+		break;
+	case 1:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = false;
+		break;
+	case 2:
+		adev->sdma.no_user_submission = true;
+		adev->sdma.disable_uq = false;
+		break;
+	}
+
 	r = amdgpu_sdma_init_microcode(adev, 0, true);
 	if (r) {
 		DRM_ERROR("Failed to init sdma firmware!\n");
@@ -1347,6 +1306,7 @@ static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
 		ring->ring_obj = NULL;
 		ring->use_doorbell = true;
 		ring->me = i;
+		ring->no_user_submission = adev->sdma.no_user_submission;
 
 		DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i,
 				ring->use_doorbell?"true":"false");
@@ -1378,6 +1338,10 @@ static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
 	else
 		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
 
+	/* add firmware version checks here */
+	if (0 && !adev->sdma.disable_uq)
+		adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+
 	return r;
 }
 
@@ -1400,11 +1364,39 @@ static int sdma_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
 	return 0;
 }
 
+static int sdma_v7_0_set_userq_trap_interrupts(struct amdgpu_device *adev,
+					       bool enable)
+{
+	unsigned int irq_type;
+	int i, r;
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_DMA]) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			irq_type = AMDGPU_SDMA_IRQ_INSTANCE0 + i;
+			if (enable)
+				r = amdgpu_irq_get(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			else
+				r = amdgpu_irq_put(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
 static int sdma_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
-	return sdma_v7_0_start(adev);
+	r = sdma_v7_0_start(adev);
+	if (r)
+		return r;
+
+	return sdma_v7_0_set_userq_trap_interrupts(adev, true);
 }
 
 static int sdma_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
@@ -1416,6 +1408,7 @@ static int sdma_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
 
 	sdma_v7_0_ctx_switch_enable(adev, false);
 	sdma_v7_0_enable(adev, false);
+	sdma_v7_0_set_userq_trap_interrupts(adev, false);
 
 	return 0;
 }