drm/radeon: add VCE 1.0 support v4

Initial support for VCE 1.0 using newest firmware.

v2: rebased
v3: fix for TN
v4: fix FW size calculation

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

8 files changed, 274 insertions, 3 deletions

diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index a6d940fc44e8..9269e188cd92 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -2041,6 +2041,25 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
 
+	if (rdev->family == CHIP_ARUBA) {
+		r = radeon_vce_resume(rdev);
+		if (!r)
+			r = vce_v1_0_resume(rdev);
+
+		if (!r)
+			r = radeon_fence_driver_start_ring(rdev,
+							   TN_RING_TYPE_VCE1_INDEX);
+		if (!r)
+			r = radeon_fence_driver_start_ring(rdev,
+							   TN_RING_TYPE_VCE2_INDEX);
+
+		if (r) {
+			dev_err(rdev->dev, "VCE init error (%d).\n", r);
+			rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size = 0;
+			rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_size = 0;
+		}
+	}
+
 	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
 	if (r) {
 		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
@@ -2118,6 +2137,19 @@ static int cayman_startup(struct radeon_device *rdev)
 			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
 	}
 
+	ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX];
+	if (ring->ring_size)
+		r = radeon_ring_init(rdev, ring, ring->ring_size, 0, 0x0);
+
+	ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX];
+	if (ring->ring_size)
+		r = radeon_ring_init(rdev, ring, ring->ring_size, 0, 0x0);
+
+	if (!r)
+		r = vce_v1_0_init(rdev);
+	else if (r != -ENOENT)
+		DRM_ERROR("radeon: failed initializing VCE (%d).\n", r);
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -2273,6 +2305,19 @@ int cayman_init(struct radeon_device *rdev)
 		r600_ring_init(rdev, ring, 4096);
 	}
 
+	if (rdev->family == CHIP_ARUBA) {
+		r = radeon_vce_init(rdev);
+		if (!r) {
+			ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX];
+			ring->ring_obj = NULL;
+			r600_ring_init(rdev, ring, 4096);
+
+			ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX];
+			ring->ring_obj = NULL;
+			r600_ring_init(rdev, ring, 4096);
+		}
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -2326,6 +2371,7 @@ void cayman_fini(struct radeon_device *rdev)
 	radeon_irq_kms_fini(rdev);
 	uvd_v1_0_fini(rdev);
 	radeon_uvd_fini(rdev);
+	radeon_vce_fini(rdev);
 	cayman_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 3d74fdcac4bf..ccebf6279c05 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1719,6 +1719,7 @@ struct radeon_vce {
 	struct drm_file		*filp[RADEON_MAX_VCE_HANDLES];
 	unsigned		img_size[RADEON_MAX_VCE_HANDLES];
 	struct delayed_work	idle_work;
+	uint32_t		keyselect;
 };
 
 int radeon_vce_init(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index b37b22bdf579..eaf909e7feb3 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -1761,6 +1761,19 @@ static struct radeon_asic cayman_asic = {
 	},
 };
 
+static struct radeon_asic_ring trinity_vce_ring = {
+	.ib_execute = &radeon_vce_ib_execute,
+	.emit_fence = &radeon_vce_fence_emit,
+	.emit_semaphore = &radeon_vce_semaphore_emit,
+	.cs_parse = &radeon_vce_cs_parse,
+	.ring_test = &radeon_vce_ring_test,
+	.ib_test = &radeon_vce_ib_test,
+	.is_lockup = &radeon_ring_test_lockup,
+	.get_rptr = &vce_v1_0_get_rptr,
+	.get_wptr = &vce_v1_0_get_wptr,
+	.set_wptr = &vce_v1_0_set_wptr,
+};
+
 static struct radeon_asic trinity_asic = {
 	.init = &cayman_init,
 	.fini = &cayman_fini,
@@ -1794,6 +1807,8 @@ static struct radeon_asic trinity_asic = {
 		[R600_RING_TYPE_DMA_INDEX] = &cayman_dma_ring,
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = &cayman_dma_ring,
 		[R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring,
+		[TN_RING_TYPE_VCE1_INDEX] = &trinity_vce_ring,
+		[TN_RING_TYPE_VCE2_INDEX] = &trinity_vce_ring,
 	},
 	.irq = {
 		.set = &evergreen_irq_set,
@@ -1930,6 +1945,8 @@ static struct radeon_asic si_asic = {
 		[R600_RING_TYPE_DMA_INDEX] = &si_dma_ring,
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = &si_dma_ring,
 		[R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring,
+		[TN_RING_TYPE_VCE1_INDEX] = &trinity_vce_ring,
+		[TN_RING_TYPE_VCE2_INDEX] = &trinity_vce_ring,
 	},
 	.irq = {
 		.set = &si_irq_set,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 629f2910a33e..e0aa33262eac 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -972,6 +972,9 @@ uint32_t vce_v1_0_get_wptr(struct radeon_device *rdev,
 			   struct radeon_ring *ring);
 void vce_v1_0_set_wptr(struct radeon_device *rdev,
 		       struct radeon_ring *ring);
+int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data);
+unsigned vce_v1_0_bo_size(struct radeon_device *rdev);
+int vce_v1_0_resume(struct radeon_device *rdev);
 int vce_v1_0_init(struct radeon_device *rdev);
 int vce_v1_0_start(struct radeon_device *rdev);
 
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index aa5d9bac70fa..574f62bbd215 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -38,8 +38,10 @@
 #define VCE_IDLE_TIMEOUT_MS	1000
 
 /* Firmware Names */
+#define FIRMWARE_TAHITI	"radeon/TAHITI_vce.bin"
 #define FIRMWARE_BONAIRE	"radeon/BONAIRE_vce.bin"
 
+MODULE_FIRMWARE(FIRMWARE_TAHITI);
 MODULE_FIRMWARE(FIRMWARE_BONAIRE);
 
 static void radeon_vce_idle_work_handler(struct work_struct *work);
@@ -63,6 +65,14 @@ int radeon_vce_init(struct radeon_device *rdev)
 	INIT_DELAYED_WORK(&rdev->vce.idle_work, radeon_vce_idle_work_handler);
 
 	switch (rdev->family) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+	case CHIP_ARUBA:
+		fw_name = FIRMWARE_TAHITI;
+		break;
+
 	case CHIP_BONAIRE:
 	case CHIP_KAVERI:
 	case CHIP_KABINI:
@@ -125,7 +135,10 @@ int radeon_vce_init(struct radeon_device *rdev)
 
 	/* allocate firmware, stack and heap BO */
 
-	size = vce_v2_0_bo_size(rdev);
+	if (rdev->family < CHIP_BONAIRE)
+		size = vce_v1_0_bo_size(rdev);
+	else
+		size = vce_v2_0_bo_size(rdev);
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
 			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, NULL,
 			     &rdev->vce.vcpu_bo);
@@ -226,13 +239,17 @@ int radeon_vce_resume(struct radeon_device *rdev)
 		return r;
 	}
 
-	memcpy(cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size);
+	memset(cpu_addr, 0, radeon_bo_size(rdev->vce.vcpu_bo));
+	if (rdev->family < CHIP_BONAIRE)
+		r = vce_v1_0_load_fw(rdev, cpu_addr);
+	else
+		memcpy(cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size);
 
 	radeon_bo_kunmap(rdev->vce.vcpu_bo);
 
 	radeon_bo_unreserve(rdev->vce.vcpu_bo);
 
-	return 0;
+	return r;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 0c04f6d1faf8..34c3739c87cf 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -6907,6 +6907,22 @@ static int si_startup(struct radeon_device *rdev)
 			rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
 	}
 
+	r = radeon_vce_resume(rdev);
+	if (!r) {
+		r = vce_v1_0_resume(rdev);
+		if (!r)
+			r = radeon_fence_driver_start_ring(rdev,
+							   TN_RING_TYPE_VCE1_INDEX);
+		if (!r)
+			r = radeon_fence_driver_start_ring(rdev,
+							   TN_RING_TYPE_VCE2_INDEX);
+	}
+	if (r) {
+		dev_err(rdev->dev, "VCE init error (%d).\n", r);
+		rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size = 0;
+		rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_size = 0;
+	}
+
 	/* Enable IRQ */
 	if (!rdev->irq.installed) {
 		r = radeon_irq_kms_init(rdev);
@@ -6975,6 +6991,23 @@ static int si_startup(struct radeon_device *rdev)
 		}
 	}
 
+	r = -ENOENT;
+
+	ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX];
+	if (ring->ring_size)
+		r = radeon_ring_init(rdev, ring, ring->ring_size, 0,
+				     VCE_CMD_NO_OP);
+
+	ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX];
+	if (ring->ring_size)
+		r = radeon_ring_init(rdev, ring, ring->ring_size, 0,
+				     VCE_CMD_NO_OP);
+
+	if (!r)
+		r = vce_v1_0_init(rdev);
+	else if (r != -ENOENT)
+		DRM_ERROR("radeon: failed initializing VCE (%d).\n", r);
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -7033,6 +7066,7 @@ int si_suspend(struct radeon_device *rdev)
 	if (rdev->has_uvd) {
 		uvd_v1_0_fini(rdev);
 		radeon_uvd_suspend(rdev);
+		radeon_vce_suspend(rdev);
 	}
 	si_fini_pg(rdev);
 	si_fini_cg(rdev);
@@ -7140,6 +7174,17 @@ int si_init(struct radeon_device *rdev)
 		}
 	}
 
+	r = radeon_vce_init(rdev);
+	if (!r) {
+		ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+
+		ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -7191,6 +7236,7 @@ void si_fini(struct radeon_device *rdev)
 	if (rdev->has_uvd) {
 		uvd_v1_0_fini(rdev);
 		radeon_uvd_fini(rdev);
+		radeon_vce_fini(rdev);
 	}
 	si_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 1630440c5136..4823a075f0c3 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -1879,6 +1879,7 @@
 #define VCE_VCPU_CACHE_SIZE1				0x20030
 #define VCE_VCPU_CACHE_OFFSET2				0x20034
 #define VCE_VCPU_CACHE_SIZE2				0x20038
+#define VCE_VCPU_SCRATCH7				0x200dc
 #define VCE_SOFT_RESET					0x20120
 #define 	VCE_ECPU_SOFT_RESET			(1 << 0)
 #define 	VCE_FME_SOFT_RESET			(1 << 2)
diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c
index b44d9c842f7b..81dd39b8de1c 100644
--- a/drivers/gpu/drm/radeon/vce_v1_0.c
+++ b/drivers/gpu/drm/radeon/vce_v1_0.c
@@ -31,6 +31,23 @@
 #include "radeon_asic.h"
 #include "sid.h"
 
+#define VCE_V1_0_FW_SIZE	(256 * 1024)
+#define VCE_V1_0_STACK_SIZE	(64 * 1024)
+#define VCE_V1_0_DATA_SIZE	(7808 * (RADEON_MAX_VCE_HANDLES + 1))
+
+struct vce_v1_0_fw_signature
+{
+	int32_t off;
+	uint32_t len;
+	int32_t num;
+	struct {
+		uint32_t chip_id;
+		uint32_t keyselect;
+		uint32_t nonce[4];
+		uint32_t sigval[4];
+	} val[8];
+};
+
 /**
  * vce_v1_0_get_rptr - get read pointer
  *
@@ -82,6 +99,129 @@ void vce_v1_0_set_wptr(struct radeon_device *rdev,
 		WREG32(VCE_RB_WPTR2, ring->wptr);
 }
 
+int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data)
+{
+	struct vce_v1_0_fw_signature *sign = (void*)rdev->vce_fw->data;
+	uint32_t chip_id;
+	int i;
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		chip_id = 0x01000014;
+		break;
+	case CHIP_VERDE:
+		chip_id = 0x01000015;
+		break;
+	case CHIP_PITCAIRN:
+	case CHIP_OLAND:
+		chip_id = 0x01000016;
+		break;
+	case CHIP_ARUBA:
+		chip_id = 0x01000017;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	for (i = 0; i < sign->num; ++i) {
+		if (sign->val[i].chip_id == chip_id)
+			break;
+	}
+
+	if (i == sign->num)
+		return -EINVAL;
+
+	data += (256 - 64) / 4;
+	data[0] = sign->val[i].nonce[0];
+	data[1] = sign->val[i].nonce[1];
+	data[2] = sign->val[i].nonce[2];
+	data[3] = sign->val[i].nonce[3];
+	data[4] = sign->len + 64;
+
+	memset(&data[5], 0, 44);
+	memcpy(&data[16], &sign[1], rdev->vce_fw->size - sizeof(*sign));
+
+	data += data[4] / 4;
+	data[0] = sign->val[i].sigval[0];
+	data[1] = sign->val[i].sigval[1];
+	data[2] = sign->val[i].sigval[2];
+	data[3] = sign->val[i].sigval[3];
+
+	rdev->vce.keyselect = sign->val[i].keyselect;
+
+	return 0;
+}
+
+unsigned vce_v1_0_bo_size(struct radeon_device *rdev)
+{
+	WARN_ON(VCE_V1_0_FW_SIZE < rdev->vce_fw->size);
+	return VCE_V1_0_FW_SIZE + VCE_V1_0_STACK_SIZE + VCE_V1_0_DATA_SIZE;
+}
+
+int vce_v1_0_resume(struct radeon_device *rdev)
+{
+	uint64_t addr = rdev->vce.gpu_addr;
+	uint32_t size;
+	int i;
+
+	WREG32_P(VCE_CLOCK_GATING_A, 0, ~(1 << 16));
+	WREG32_P(VCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
+	WREG32_P(VCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
+	WREG32(VCE_CLOCK_GATING_B, 0);
+
+	WREG32_P(VCE_LMI_FW_PERIODIC_CTRL, 0x4, ~0x4);
+
+	WREG32(VCE_LMI_CTRL, 0x00398000);
+	WREG32_P(VCE_LMI_CACHE_CTRL, 0x0, ~0x1);
+	WREG32(VCE_LMI_SWAP_CNTL, 0);
+	WREG32(VCE_LMI_SWAP_CNTL1, 0);
+	WREG32(VCE_LMI_VM_CTRL, 0);
+
+	WREG32(VCE_VCPU_SCRATCH7, RADEON_MAX_VCE_HANDLES);
+
+	addr += 256;
+	size = VCE_V1_0_FW_SIZE;
+	WREG32(VCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff);
+	WREG32(VCE_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = VCE_V1_0_STACK_SIZE;
+	WREG32(VCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff);
+	WREG32(VCE_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = VCE_V1_0_DATA_SIZE;
+	WREG32(VCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff);
+	WREG32(VCE_VCPU_CACHE_SIZE2, size);
+
+	WREG32_P(VCE_LMI_CTRL2, 0x0, ~0x100);
+
+	WREG32(VCE_LMI_FW_START_KEYSEL, rdev->vce.keyselect);
+
+	for (i = 0; i < 10; ++i) {
+		mdelay(10);
+		if (RREG32(VCE_FW_REG_STATUS) & VCE_FW_REG_STATUS_DONE)
+			break;
+	}
+
+	if (i == 10)
+		return -ETIMEDOUT;
+
+	if (!(RREG32(VCE_FW_REG_STATUS) & VCE_FW_REG_STATUS_PASS))
+		return -EINVAL;
+
+	for (i = 0; i < 10; ++i) {
+		mdelay(10);
+		if (!(RREG32(VCE_FW_REG_STATUS) & VCE_FW_REG_STATUS_BUSY))
+			break;
+	}
+
+	if (i == 10)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
 /**
  * vce_v1_0_start - start VCE block
  *