Mesa (master): radeonsi: implement SDMA-based buffer clearing for SI

[Marek Olšák]

Module: Mesa
Branch: master
Commit: cba9d59362130e1b44cd9cfc4f38ad3773111442
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=cba9d59362130e1b44cd9cfc4f38ad3773111442

Author: Marek Olšák <marek.olsak at amd.com>
Date:   Tue Dec 27 17:51:43 2016 +0100

radeonsi: implement SDMA-based buffer clearing for SI

Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>

---

 src/gallium/drivers/radeonsi/si_cp_dma.c |  2 +-
 src/gallium/drivers/radeonsi/si_dma.c    | 40 ++++++++++++++++++++++++++++++++
 2 files changed, 41 insertions(+), 1 deletion(-)

diff --git a/src/gallium/drivers/radeonsi/si_cp_dma.c b/src/gallium/drivers/radeonsi/si_cp_dma.c
index 722ac4d..596e32c 100644
--- a/src/gallium/drivers/radeonsi/si_cp_dma.c
+++ b/src/gallium/drivers/radeonsi/si_cp_dma.c
@@ -196,7 +196,7 @@ static void si_clear_buffer(struct pipe_context *ctx, struct pipe_resource *dst,
 
 	/* dma_clear_buffer can use clear_buffer on failure. Make sure that
 	 * doesn't happen. We don't want an infinite recursion: */
-	if (sctx->b.chip_class >= CIK && sctx->b.dma.cs &&
+	if (sctx->b.dma.cs &&
 	    /* CP DMA is very slow. Always use SDMA for big clears. This
 	     * alone improves DeusEx:MD performance by 70%. */
 	    (size > 128 * 1024 ||
diff --git a/src/gallium/drivers/radeonsi/si_dma.c b/src/gallium/drivers/radeonsi/si_dma.c
index b6aab00..9dbee3a 100644
--- a/src/gallium/drivers/radeonsi/si_dma.c
+++ b/src/gallium/drivers/radeonsi/si_dma.c
@@ -78,6 +78,45 @@ static void si_dma_copy_buffer(struct si_context *ctx,
 	}
 }
 
+static void si_dma_clear_buffer(struct pipe_context *ctx,
+				struct pipe_resource *dst,
+				uint64_t offset,
+				uint64_t size,
+				unsigned clear_value)
+{
+	struct si_context *sctx = (struct si_context *)ctx;
+	struct radeon_winsys_cs *cs = sctx->b.dma.cs;
+	unsigned i, ncopy, csize;
+	struct r600_resource *rdst = r600_resource(dst);
+
+	if (!cs || offset % 4 != 0 || size % 4 != 0) {
+		ctx->clear_buffer(ctx, dst, offset, size, &clear_value, 4);
+		return;
+	}
+
+	/* Mark the buffer range of destination as valid (initialized),
+	 * so that transfer_map knows it should wait for the GPU when mapping
+	 * that range. */
+	util_range_add(&rdst->valid_buffer_range, offset, offset + size);
+
+	offset += rdst->gpu_address;
+
+	/* the same maximum size as for copying */
+	ncopy = DIV_ROUND_UP(size, SI_DMA_COPY_MAX_DWORD_ALIGNED_SIZE);
+	r600_need_dma_space(&sctx->b, ncopy * 4, rdst, NULL);
+
+	for (i = 0; i < ncopy; i++) {
+		csize = MIN2(size, SI_DMA_COPY_MAX_DWORD_ALIGNED_SIZE);
+		radeon_emit(cs, SI_DMA_PACKET(SI_DMA_PACKET_CONSTANT_FILL, 0,
+					      csize / 4));
+		radeon_emit(cs, offset);
+		radeon_emit(cs, clear_value);
+		radeon_emit(cs, (offset >> 32) << 16);
+		offset += csize;
+		size -= csize;
+	}
+}
+
 static void si_dma_copy_tile(struct si_context *ctx,
 			     struct pipe_resource *dst,
 			     unsigned dst_level,
@@ -285,4 +324,5 @@ fallback:
 void si_init_dma_functions(struct si_context *sctx)
 {
 	sctx->b.dma_copy = si_dma_copy;
+	sctx->b.dma_clear_buffer = si_dma_clear_buffer;
 }