ARM: NEON optimized pixman_blt

NEON unit has fast access to L1/L2 caches and even simple
copy of memory buffers using NEON provides more than 1.5x
performance improvement on ARM Cortex-A8.

1 files changed, 70 insertions, 0 deletions

diff --git a/pixman/pixman-arm-neon.c b/pixman/pixman-arm-neon.c
index 2ed8b4b..9194924 100644
--- a/pixman/pixman-arm-neon.c
+++ b/pixman/pixman-arm-neon.c
@@ -292,6 +292,46 @@ pixman_fill_neon (uint32_t *bits,
     }
 }
 
+static pixman_bool_t
+pixman_blt_neon (uint32_t *src_bits,
+                 uint32_t *dst_bits,
+                 int       src_stride,
+                 int       dst_stride,
+                 int       src_bpp,
+                 int       dst_bpp,
+                 int       src_x,
+                 int       src_y,
+                 int       dst_x,
+                 int       dst_y,
+                 int       width,
+                 int       height)
+{
+    if (src_bpp != dst_bpp)
+	return FALSE;
+
+    switch (src_bpp)
+    {
+    case 16:
+	pixman_composite_src_0565_0565_asm_neon (
+		width, height,
+		(uint16_t *)(((char *) dst_bits) +
+		dst_y * dst_stride * 4 + dst_x * 2), dst_stride * 2,
+		(uint16_t *)(((char *) src_bits) +
+		src_y * src_stride * 4 + src_x * 2), src_stride * 2);
+	return TRUE;
+    case 32:
+	pixman_composite_src_8888_8888_asm_neon (
+		width, height,
+		(uint32_t *)(((char *) dst_bits) +
+		dst_y * dst_stride * 4 + dst_x * 4), dst_stride,
+		(uint32_t *)(((char *) src_bits) +
+		src_y * src_stride * 4 + src_x * 4), src_stride);
+	return TRUE;
+    default:
+	return FALSE;
+    }
+}
+
 static const pixman_fast_path_t arm_neon_fast_path_array[] =
 {
     { PIXMAN_OP_SRC,  PIXMAN_r5g6b5,   PIXMAN_null,     PIXMAN_r5g6b5,   neon_composite_src_0565_0565    },
@@ -361,6 +401,35 @@ arm_neon_composite (pixman_implementation_t *imp,
 }
 
 static pixman_bool_t
+arm_neon_blt (pixman_implementation_t *imp,
+              uint32_t *               src_bits,
+              uint32_t *               dst_bits,
+              int                      src_stride,
+              int                      dst_stride,
+              int                      src_bpp,
+              int                      dst_bpp,
+              int                      src_x,
+              int                      src_y,
+              int                      dst_x,
+              int                      dst_y,
+              int                      width,
+              int                      height)
+{
+    if (!pixman_blt_neon (
+            src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp,
+            src_x, src_y, dst_x, dst_y, width, height))
+
+    {
+	return _pixman_implementation_blt (
+	    imp->delegate,
+	    src_bits, dst_bits, src_stride, dst_stride, src_bpp, dst_bpp,
+	    src_x, src_y, dst_x, dst_y, width, height);
+    }
+
+    return TRUE;
+}
+
+static pixman_bool_t
 arm_neon_fill (pixman_implementation_t *imp,
                uint32_t *               bits,
                int                      stride,
@@ -385,6 +454,7 @@ _pixman_implementation_create_arm_neon (void)
     pixman_implementation_t *imp = _pixman_implementation_create (general);
 
     imp->composite = arm_neon_composite;
+    imp->blt = arm_neon_blt;
     imp->fill = arm_neon_fill;
 
     return imp;