drm/radeon: add dma engine support for vm pt updates on si (v2)

Async DMA has a special packet for contiguous pt updates
which saves overhead.

v2: rebase
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/radeon/si.c

@@ -2825,9 +2825,12 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 {
 	struct radeon_ring *ring = &rdev->ring[rdev->asic->vm.pt_ring_index];
 	uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
+	uint64_t value;
+	unsigned ndw;
 
+	if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) {
 		while (count) {
-		unsigned ndw = 2 + count * 2;
+			ndw = 2 + count * 2;
 			if (ndw > 0x3FFE)
 				ndw = 0x3FFE;
 
@@ -2837,20 +2840,73 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 			radeon_ring_write(ring, pe);
 			radeon_ring_write(ring, upper_32_bits(pe));
 			for (; ndw > 2; ndw -= 2, --count, pe += 8) {
-			uint64_t value;
 				if (flags & RADEON_VM_PAGE_SYSTEM) {
 					value = radeon_vm_map_gart(rdev, addr);
 					value &= 0xFFFFFFFFFFFFF000ULL;
-			} else if (flags & RADEON_VM_PAGE_VALID)
+				} else if (flags & RADEON_VM_PAGE_VALID) {
 					value = addr;
-			else
+				} else {
 					value = 0;
+				}
 				addr += incr;
 				value |= r600_flags;
 				radeon_ring_write(ring, value);
 				radeon_ring_write(ring, upper_32_bits(value));
 			}
 		}
+	} else {
+		/* DMA */
+		if (flags & RADEON_VM_PAGE_SYSTEM) {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
+
+				/* for non-physically contiguous pages (system) */
+				radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw));
+				radeon_ring_write(ring, pe);
+				radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
+				for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+					if (flags & RADEON_VM_PAGE_SYSTEM) {
+						value = radeon_vm_map_gart(rdev, addr);
+						value &= 0xFFFFFFFFFFFFF000ULL;
+					} else if (flags & RADEON_VM_PAGE_VALID) {
+						value = addr;
+					} else {
+						value = 0;
+					}
+					addr += incr;
+					value |= r600_flags;
+					radeon_ring_write(ring, value);
+					radeon_ring_write(ring, upper_32_bits(value));
+				}
+			}
+		} else {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
+
+				if (flags & RADEON_VM_PAGE_VALID)
+					value = addr;
+				else
+					value = 0;
+				/* for physically contiguous pages (vram) */
+				radeon_ring_write(ring, DMA_PTE_PDE_PACKET(ndw));
+				radeon_ring_write(ring, pe); /* dst addr */
+				radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
+				radeon_ring_write(ring, r600_flags); /* mask */
+				radeon_ring_write(ring, 0);
+				radeon_ring_write(ring, value); /* value */
+				radeon_ring_write(ring, upper_32_bits(value));
+				radeon_ring_write(ring, incr); /* increment size */
+				radeon_ring_write(ring, 0);
+				pe += ndw * 4;
+				addr += (ndw / 2) * incr;
+				count -= ndw / 2;
+			}
+		}
+	}
 }
 
 void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)

drivers/gpu/drm/radeon/sid.h

@@ -972,6 +972,16 @@
 					 (((t) & 0x1) << 23) |		\
 					 (((s) & 0x1) << 22) |		\
 					 (((n) & 0xFFFFF) << 0))
+
+#define DMA_IB_PACKET(cmd, vmid, n)	((((cmd) & 0xF) << 28) |	\
+					 (((vmid) & 0xF) << 20) |	\
+					 (((n) & 0xFFFFF) << 0))
+
+#define DMA_PTE_PDE_PACKET(n)		((2 << 28) |			\
+					 (1 << 26) |			\
+					 (1 << 21) |			\
+					 (((n) & 0xFFFFF) << 0))
+
 /* async DMA Packet types */
 #define	DMA_PACKET_WRITE				  0x2
 #define	DMA_PACKET_COPY					  0x3