ARM: Tiny improvement in over_n_8888_8888_ca_process_pixblock_head

Instead of two

	mvn d24, d24
	mvn d25, d25

use just one

	mvn q12, q12

Also move another vmvn instruction into the created pipeline bubble,
as pointed out by Siarhei.

diff --git a/pixman/pixman-arm-neon-asm.S b/pixman/pixman-arm-neon-asm.S
index 71b30ac..1d3e64e 100644 (file)
--- a/pixman/pixman-arm-neon-asm.S
+++ b/pixman/pixman-arm-neon-asm.S
@@ -1358,11 +1358,10 @@ generate_composite_function \
      *
      * output: updated dest in {d28, d29, d30, d31}
      */
-    vmvn.8      d24, d24
-    vmvn.8      d25, d25
+    vmvn.8      q12, q12
+    vmvn.8      d26, d26
     vmull.u8    q8,  d24, d4
     vmull.u8    q9,  d25, d5
-    vmvn.8      d26, d26
     vmvn.8      d27, d3
     vmull.u8    q10, d26, d6
     vmull.u8    q11, d27, d7