[Pixman] [PATCH 09/10] ARM: NEON optimization for scaled src_0565_8888 with nearest filter

[Siarhei Siamashka]

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

Benchmark from ARM Cortex-A8 @720MHz:
    == before ==
    op=1, src_fmt=10020565, dst_fmt=20028888, speed=8.99 MPix/s

    == after ==
    op=1, src_fmt=10020565, dst_fmt=20028888, speed=76.98 MPix/s

    == unscaled ==
    op=1, src_fmt=10020565, dst_fmt=20028888, speed=137.78 MPix/s
---
 pixman/pixman-arm-neon-asm.S |   10 ++++++++++
 pixman/pixman-arm-neon.c     |   10 ++++++++++
 2 files changed, 20 insertions(+), 0 deletions(-)

diff --git a/pixman/pixman-arm-neon-asm.S b/pixman/pixman-arm-neon-asm.S
index 4d6b03b..91ec27d 100644
--- a/pixman/pixman-arm-neon-asm.S
+++ b/pixman/pixman-arm-neon-asm.S
@@ -2009,3 +2009,13 @@ generate_composite_function_nearest_scanline \
     pixman_composite_src_8888_0565_process_pixblock_head, \
     pixman_composite_src_8888_0565_process_pixblock_tail, \
     pixman_composite_src_8888_0565_process_pixblock_tail_head
+
+generate_composite_function_nearest_scanline \
+    pixman_scaled_nearest_scanline_0565_8888_SRC_asm_neon, 16, 0, 32, \
+    FLAG_DST_WRITEONLY | FLAG_DEINTERLEAVE_32BPP, \
+    8, /* number of pixels, processed in a single block */ \
+    default_init, \
+    default_cleanup, \
+    pixman_composite_src_0565_8888_process_pixblock_head, \
+    pixman_composite_src_0565_8888_process_pixblock_tail, \
+    pixman_composite_src_0565_8888_process_pixblock_tail_head
diff --git a/pixman/pixman-arm-neon.c b/pixman/pixman-arm-neon.c
index bf921e6..2f82069 100644
--- a/pixman/pixman-arm-neon.c
+++ b/pixman/pixman-arm-neon.c
@@ -103,6 +103,8 @@ PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST (neon, 8888_0565, OVER,
                                         uint32_t, uint16_t)
 PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST (neon, 8888_0565, SRC,
                                         uint32_t, uint16_t)
+PIXMAN_ARM_BIND_SCALED_NEAREST_SRC_DST (neon, 0565_8888, SRC,
+                                        uint16_t, uint32_t)
 
 void
 pixman_composite_src_n_8_asm_neon (int32_t   w,
@@ -291,6 +293,14 @@ static const pixman_fast_path_t arm_neon_fast_paths[] =
     PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (SRC, a8b8g8r8, b5g6r5, neon_8888_0565),
     PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (SRC, x8b8g8r8, b5g6r5, neon_8888_0565),
 
+    PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (SRC, b5g6r5, x8b8g8r8, neon_0565_8888),
+    PIXMAN_ARM_SIMPLE_NEAREST_FAST_PATH (SRC, r5g6b5, x8r8g8b8, neon_0565_8888),
+    /* Note: NONE repeat is not supported yet */
+    SIMPLE_NEAREST_FAST_PATH_COVER (SRC, r5g6b5, a8r8g8b8, neon_0565_8888),
+    SIMPLE_NEAREST_FAST_PATH_COVER (SRC, b5g6r5, a8b8g8r8, neon_0565_8888),
+    SIMPLE_NEAREST_FAST_PATH_PAD (SRC, r5g6b5, a8r8g8b8, neon_0565_8888),
+    SIMPLE_NEAREST_FAST_PATH_PAD (SRC, b5g6r5, a8b8g8r8, neon_0565_8888),
+
     { PIXMAN_OP_NONE },
 };
 
-- 
1.7.2.2