ARM: added 'neon_composite_src_0888_0565_rev' fast path

This is ARM NEON optimized conversion of native RGB format used by
GTK/GDK into r5g6b5 format.

diff --git a/pixman/pixman-arm-neon-asm.S b/pixman/pixman-arm-neon-asm.S
index bb68be6..8010e80 100644 (file)
--- a/pixman/pixman-arm-neon-asm.S
+++ b/pixman/pixman-arm-neon-asm.S
@@ -1370,3 +1370,41 @@ generate_composite_function \
     0, /* dst_r_basereg */ \
     0, /* src_basereg   */ \
     0  /* mask_basereg  */
+
+/******************************************************************************/
+
+.macro pixman_composite_src_0888_0565_rev_process_pixblock_head
+    vshll.u8    q8, d1, #8
+    vshll.u8    q9, d2, #8
+.endm
+
+.macro pixman_composite_src_0888_0565_rev_process_pixblock_tail
+    vshll.u8    q14, d0, #8
+    vsri.u16    q14, q8, #5
+    vsri.u16    q14, q9, #11
+.endm
+
+.macro pixman_composite_src_0888_0565_rev_process_pixblock_tail_head
+        vshll.u8    q14, d0, #8
+    vld3.8 {d0, d1, d2}, [SRC]!
+        vsri.u16    q14, q8, #5
+        vsri.u16    q14, q9, #11
+    vshll.u8    q8, d1, #8
+        vst1.16 {d28, d29}, [DST_W, :128]!
+    vshll.u8    q9, d2, #8
+.endm
+
+generate_composite_function \
+    pixman_composite_src_0888_0565_rev_asm_neon, 24, 0, 16, \
+    FLAG_DST_WRITEONLY, \
+    8, /* number of pixels, processed in a single block */ \
+    10, /* prefetch distance */ \
+    default_init, \
+    default_cleanup, \
+    pixman_composite_src_0888_0565_rev_process_pixblock_head, \
+    pixman_composite_src_0888_0565_rev_process_pixblock_tail, \
+    pixman_composite_src_0888_0565_rev_process_pixblock_tail_head, \
+    28, /* dst_w_basereg */ \
+    0, /* dst_r_basereg */ \
+    0, /* src_basereg   */ \
+    0  /* mask_basereg  */

diff --git a/pixman/pixman-arm-neon.c b/pixman/pixman-arm-neon.c
index ac0c558..21ed436 100644 (file)
--- a/pixman/pixman-arm-neon.c
+++ b/pixman/pixman-arm-neon.c
@@ -256,6 +256,7 @@ BIND_SRC_NULL_DST(src_0888_0888, uint8_t, 3, uint8_t, 3)
 BIND_SRC_NULL_DST(src_8888_0565, uint32_t, 1, uint16_t, 1)
 BIND_SRC_NULL_DST(src_0565_8888, uint16_t, 1, uint32_t, 1)
 BIND_SRC_NULL_DST(src_0888_8888_rev, uint8_t, 3, uint32_t, 1)
+BIND_SRC_NULL_DST(src_0888_0565_rev, uint8_t, 3, uint16_t, 1)
 BIND_SRC_NULL_DST(add_8000_8000, uint8_t, 1, uint8_t, 1)
 BIND_SRC_NULL_DST(add_8888_8888, uint32_t, 1, uint32_t, 1)
 
@@ -399,6 +400,7 @@ static const pixman_fast_path_t arm_neon_fast_path_array[] =
     { PIXMAN_OP_SRC,  PIXMAN_x8b8g8r8, PIXMAN_null,     PIXMAN_x8b8g8r8, neon_composite_src_8888_8888    },
     { PIXMAN_OP_SRC,  PIXMAN_r8g8b8,   PIXMAN_null,     PIXMAN_r8g8b8,   neon_composite_src_0888_0888    },
     { PIXMAN_OP_SRC,  PIXMAN_b8g8r8,   PIXMAN_null,     PIXMAN_x8r8g8b8, neon_composite_src_0888_8888_rev },
+    { PIXMAN_OP_SRC,  PIXMAN_b8g8r8,   PIXMAN_null,     PIXMAN_r5g6b5,   neon_composite_src_0888_0565_rev },
     { PIXMAN_OP_OVER, PIXMAN_solid,    PIXMAN_a8,       PIXMAN_r5g6b5,   neon_composite_over_n_8_0565    },
     { PIXMAN_OP_OVER, PIXMAN_solid,    PIXMAN_a8,       PIXMAN_b5g6r5,   neon_composite_over_n_8_0565    },
     { PIXMAN_OP_OVER, PIXMAN_solid,    PIXMAN_a8,       PIXMAN_a8r8g8b8, neon_composite_over_n_8_8888    },