armv6: Add over_n_8888 fast path (disabled)

This new fast path is initially disabled by putting the entries in the
lookup table after the sentinel. The compiler cannot tell the new code
is not used, so it cannot eliminate the code. Also the lookup table size
will include the new fast path. When the follow-up patch then enables
the new fast path, the binary layout (alignments, size, etc.) will stay
the same compared to the disabled case.

Keeping the binary layout identical is important for benchmarking on
Raspberry Pi 1. The addresses at which functions are loaded will have a
significant impact on benchmark results, causing unexpected performance
changes. Keeping all function addresses the same across the patch
enabling a new fast path improves the reliability of benchmarks.

Benchmark results are included in the patch enabling this fast path.

[Pekka: disabled the fast path, commit message]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

2 files changed, 48 insertions, 0 deletions

diff --git a/pixman/pixman-arm-simd-asm.S b/pixman/pixman-arm-simd-asm.S
index 7b0727b4..a74a0a8f 100644
--- a/pixman/pixman-arm-simd-asm.S
+++ b/pixman/pixman-arm-simd-asm.S
@@ -1136,3 +1136,44 @@ generate_composite_function \
     in_reverse_8888_8888_process_tail
 
 /******************************************************************************/
+
+.macro over_n_8888_init
+        ldr     SRC, [sp, #ARGS_STACK_OFFSET]
+        /* Hold loop invariant in MASK */
+        ldr     MASK, =0x00800080
+        /* Hold multiplier for destination in STRIDE_M */
+        mov     STRIDE_M, #255
+        sub     STRIDE_M, STRIDE_M, SRC, lsr #24
+        /* Set GE[3:0] to 0101 so SEL instructions do what we want */
+        uadd8   SCRATCH, MASK, MASK
+.endm
+
+.macro over_n_8888_process_head  cond, numbytes, firstreg, unaligned_src, unaligned_mask, preload
+        pixld   , numbytes, firstreg, DST, 0
+.endm
+
+.macro over_n_8888_1pixel dst
+        mul_8888_8  WK&dst, STRIDE_M, SCRATCH, MASK
+        uqadd8  WK&dst, WK&dst, SRC
+.endm
+
+.macro over_n_8888_process_tail  cond, numbytes, firstreg
+ .set PROCESS_REG, firstreg
+ .rept numbytes / 4
+        over_n_8888_1pixel %(PROCESS_REG)
+  .set PROCESS_REG, PROCESS_REG+1
+ .endr
+        pixst   , numbytes, firstreg, DST
+.endm
+
+generate_composite_function \
+    pixman_composite_over_n_8888_asm_armv6, 0, 0, 32 \
+    FLAG_DST_READWRITE | FLAG_BRANCH_OVER | FLAG_PROCESS_DOES_STORE \
+    2, /* prefetch distance */ \
+    over_n_8888_init, \
+    nop_macro, /* newline */ \
+    nop_macro, /* cleanup */ \
+    over_n_8888_process_head, \
+    over_n_8888_process_tail
+
+/******************************************************************************/
diff --git a/pixman/pixman-arm-simd.c b/pixman/pixman-arm-simd.c
index f40ff36f..62c0f41d 100644
--- a/pixman/pixman-arm-simd.c
+++ b/pixman/pixman-arm-simd.c
@@ -51,6 +51,8 @@ PIXMAN_ARM_BIND_FAST_PATH_SRC_DST (armv6, over_8888_8888,
 PIXMAN_ARM_BIND_FAST_PATH_SRC_DST (armv6, in_reverse_8888_8888,
                                    uint32_t, 1, uint32_t, 1)
 
+PIXMAN_ARM_BIND_FAST_PATH_N_DST (SKIP_ZERO_SRC, armv6, over_n_8888,
+                                 uint32_t, 1)
 PIXMAN_ARM_BIND_FAST_PATH_N_DST (0, armv6, over_reverse_n_8888,
                                  uint32_t, 1)
 
@@ -271,6 +273,11 @@ static const pixman_fast_path_t arm_simd_fast_paths[] =
     SIMPLE_NEAREST_FAST_PATH (SRC, x8b8g8r8, x8b8g8r8, armv6_8888_8888),
 
     { PIXMAN_OP_NONE },
+
+    PIXMAN_STD_FAST_PATH (OVER, solid, null, a8r8g8b8, armv6_composite_over_n_8888),
+    PIXMAN_STD_FAST_PATH (OVER, solid, null, x8r8g8b8, armv6_composite_over_n_8888),
+    PIXMAN_STD_FAST_PATH (OVER, solid, null, a8b8g8r8, armv6_composite_over_n_8888),
+    PIXMAN_STD_FAST_PATH (OVER, solid, null, x8b8g8r8, armv6_composite_over_n_8888),
 };
 
 pixman_implementation_t *