drm/radeon/kms: add support for the CONST IB to the CS ioctl

This adds a new chunk id to the CS ioctl to support the
INDIRECT_BUFFER_CONST packet.

On SI, the CP adds a new engine called the CE (Constant Engine)
which runs simulatenously with the DE (Drawing Engine, formerly
called the ME).  This allows the CP to process two related IBs
simultaneously.  The CE is tasked with loading the constant data
(constant buffers, resource descriptors, samplers, etc.) while
the DE loads context register state and issues drawing commands.
It's up to the userspace application to sychronize the CE and the
DE using special synchronization packets.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drivers/gpu/drm/radeon/radeon.h

@@ -632,6 +632,7 @@ struct radeon_ib {
 	uint32_t		*ptr;
 	struct radeon_fence	*fence;
 	unsigned		vm_id;
+	bool			is_const_ib;
 };
 
 /*
@@ -836,7 +837,9 @@ struct radeon_cs_parser {
 	int			chunk_ib_idx;
 	int			chunk_relocs_idx;
 	int			chunk_flags_idx;
+	int			chunk_const_ib_idx;
 	struct radeon_ib	*ib;
+	struct radeon_ib	*const_ib;
 	void			*track;
 	unsigned		family;
 	int			parser_error;

drivers/gpu/drm/radeon/radeon_cs.c

@@ -170,6 +170,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	p->chunk_ib_idx = -1;
 	p->chunk_relocs_idx = -1;
 	p->chunk_flags_idx = -1;
+	p->chunk_const_ib_idx = -1;
 	p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (p->chunks_array == NULL) {
 		return -ENOMEM;
@@ -208,6 +209,12 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
+		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) {
+			p->chunk_const_ib_idx = i;
+			/* zero length CONST IB isn't useful */
+			if (p->chunks[i].length_dw == 0)
+				return -EINVAL;
+		}
 		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
 			p->chunk_flags_idx = i;
 			/* zero length flags aren't useful */
@@ -389,6 +396,32 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
 		return 0;
 
+	if ((rdev->family >= CHIP_TAHITI) &&
+	    (parser->chunk_const_ib_idx != -1)) {
+		ib_chunk = &parser->chunks[parser->chunk_const_ib_idx];
+		if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
+			DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw);
+			return -EINVAL;
+		}
+		r =  radeon_ib_get(rdev, parser->ring, &parser->const_ib,
+				   ib_chunk->length_dw * 4);
+		if (r) {
+			DRM_ERROR("Failed to get const ib !\n");
+			return r;
+		}
+		parser->const_ib->is_const_ib = true;
+		parser->const_ib->length_dw = ib_chunk->length_dw;
+		/* Copy the packet into the IB */
+		if (DRM_COPY_FROM_USER(parser->const_ib->ptr, ib_chunk->user_ptr,
+				       ib_chunk->length_dw * 4)) {
+			return -EFAULT;
+		}
+		r = radeon_ring_ib_parse(rdev, parser->ring, parser->const_ib);
+		if (r) {
+			return r;
+		}
+	}
+
 	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
 	if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 		DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
@@ -424,11 +457,25 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("Failed to synchronize rings !\n");
 	}
+
+	if ((rdev->family >= CHIP_TAHITI) &&
+	    (parser->chunk_const_ib_idx != -1)) {
+		parser->const_ib->vm_id = vm->id;
+		/* ib pool is bind at 0 in virtual address space to gpu_addr is the
+		 * offset inside the pool bo
+		 */
+		parser->const_ib->gpu_addr = parser->const_ib->sa_bo.offset;
+		r = radeon_ib_schedule(rdev, parser->const_ib);
+		if (r)
+			goto out;
+	}
+
 	parser->ib->vm_id = vm->id;
 	/* ib pool is bind at 0 in virtual address space to gpu_addr is the
 	 * offset inside the pool bo
 	 */
 	parser->ib->gpu_addr = parser->ib->sa_bo.offset;
+	parser->ib->is_const_ib = false;
 	r = radeon_ib_schedule(rdev, parser->ib);
 out:
 	if (!r) {

drivers/gpu/drm/radeon/radeon_ring.c

@@ -133,6 +133,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
 				(*ib)->gpu_addr += (*ib)->sa_bo.offset;
 				(*ib)->fence = fence;
 				(*ib)->vm_id = 0;
+				(*ib)->is_const_ib = false;
 				/* ib are most likely to be allocated in a ring fashion
 				 * thus rdev->ib_pool.head_id should be the id of the
 				 * oldest ib

include/drm/radeon_drm.h

@@ -908,6 +908,7 @@ struct drm_radeon_gem_va {
 #define RADEON_CHUNK_ID_RELOCS	0x01
 #define RADEON_CHUNK_ID_IB	0x02
 #define RADEON_CHUNK_ID_FLAGS	0x03
+#define RADEON_CHUNK_ID_CONST_IB	0x04
 
 /* The first dword of RADEON_CHUNK_ID_FLAGS is a uint32 of these flags: */
 #define RADEON_CS_KEEP_TILING_FLAGS 0x01