[Pixman] [PATCH 07/10] ARM: NEON optimization for scaled over_8888_0565 with nearest filter

[Siarhei Siamashka]

From: Siarhei Siamashka <siarhei.siamashka at nokia.com>

Benchmark from ARM Cortex-A8 @720MHz:
    == before ==
    op=3, src_fmt=20028888, dst_fmt=10020565, speed=10.29 MPix/s

    == after ==
    op=3, src_fmt=20028888, dst_fmt=10020565, speed=36.36 MPix/s

    == unscaled ==
    op=3, src_fmt=20028888, dst_fmt=10020565, speed=79.40 MPix/s
---
 pixman/pixman-arm-neon-asm.S |   14 ++++++++++++++
 pixman/pixman-arm-neon.c     |    5 +++++
 2 files changed, 19 insertions(+), 0 deletions(-)

diff --git a/pixman/pixman-arm-neon-asm.S b/pixman/pixman-arm-neon-asm.S
index b009578..ea3337a 100644
--- a/pixman/pixman-arm-neon-asm.S
+++ b/pixman/pixman-arm-neon-asm.S
@@ -1985,3 +1985,17 @@ generate_composite_function_nearest_scanline \
     pixman_composite_over_8888_8888_process_pixblock_head, \
     pixman_composite_over_8888_8888_process_pixblock_tail, \
     pixman_composite_over_8888_8888_process_pixblock_tail_head
+
+generate_composite_function_nearest_scanline \
+    pixman_scaled_nearest_scanline_8888_0565_OVER_asm_neon, 32, 0, 16, \
+    FLAG_DST_READWRITE | FLAG_DEINTERLEAVE_32BPP, \
+    8, /* number of pixels, processed in a single block */ \
+    default_init, \
+    default_cleanup, \
+    pixman_composite_over_8888_0565_process_pixblock_head, \
+    pixman_composite_over_8888_0565_process_pixblock_tail, \
+    pixman_composite_over_8888_0565_process_pixblock_tail_head, \
+    28, /* dst_w_basereg */ \
+    4,  /* dst_r_basereg */ \
+    0,  /* src_basereg   */ \
+    24  /* mask_basereg  */
diff --git a/pixman/pixman-arm-neon.c b/pixman/pixman-arm-neon.c
index 9c2838f..a200ed1 100644
--- a/pixman/pixman-arm-neon.c
+++ b/pixman/pixman-arm-neon.c
@@ -99,6 +99,8 @@ PIXMAN_ARM_BIND_FAST_PATH_SRC_MASK_DST (neon, over_0565_8_0565,
 
 PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST (neon, 8888_8888, OVER,
                                         uint32_t, uint32_t)
+PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST (neon, 8888_0565, OVER,
+                                        uint32_t, uint16_t)
 
 void
 pixman_composite_src_n_8_asm_neon (int32_t   w,
@@ -279,6 +281,9 @@ static const pixman_fast_path_t arm_neon_fast_paths[] =
     PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (OVER, a8r8g8b8, x8r8g8b8, neon_8888_8888),
     PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (OVER, a8b8g8r8, x8b8g8r8, neon_8888_8888),
 
+    PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (OVER, a8r8g8b8, r5g6b5, neon_8888_0565),
+    PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (OVER, a8b8g8r8, b5g6r5, neon_8888_0565),
+
     { PIXMAN_OP_NONE },
 };
 
-- 
1.7.2.2