unify NEON code, use preload

[sdl_omap.git] / src / video / SDL_blit_neon.S

diff --git a/src/video/SDL_blit_neon.S b/src/video/SDL_blit_neon.S
index 438d9fc..af9af36 100644 (file)
--- a/src/video/SDL_blit_neon.S
+++ b/src/video/SDL_blit_neon.S
@@ -1,5 +1,5 @@
 /*
- * (C) Gražvydas "notaz" Ignotas, 2011
+ * (C) Gražvydas "notaz" Ignotas, 2011,2012
  *
  * This work is licensed under the terms of any of these licenses
  * (at your option):
@@ -28,17 +28,24 @@
     b          0b
 .endm
 
-@ void *dst, const void *src, int count
-.macro do_argb_alpha bgr2rgb
-    mov        r3, #0xff
-    vdup.i16   q12, r3
+@ void *dst, const void *src, int count, uint global_alpha
+.macro do_argb_alpha bgr2rgb global_alpha
+    mov        r12, #0xff
+.if \global_alpha
+    vdup.16    q11, r3
+.endif
+    vdup.i16   q12, r12
 0:
+    pld        [r1, #64*2]
+    pld        [r0, #64*2]
     vld4.8     {d4-d7}, [r1]!
     vld4.8     {d0-d3}, [r0]
 .if \bgr2rgb
     vswp       d4, d6          @ BGR->RGB
 .endif
+.if !\global_alpha
     vmovl.u8   q11, d7
+.endif
     @ d = (((s-d)*a+255)>>8)+d
     vsubl.u8   q8, d4, d0
     vsubl.u8   q9, d5, d1
@@ -60,6 +67,7 @@
     b          0b
 .endm
 
+
 do_argb_finish:
     add        r2, r2, #8
     vzip.8     d4, d5          @ RRR..|GGG.. -> RGRG..
@@ -98,10 +106,18 @@ neon_ABGRtoXRGB:
 
 .global neon_ARGBtoXRGBalpha
 neon_ARGBtoXRGBalpha:
-    do_argb_alpha 0
+    do_argb_alpha 0, 0
 
 .global neon_ABGRtoXRGBalpha
 neon_ABGRtoXRGBalpha:
-    do_argb_alpha 1
+    do_argb_alpha 1, 0
+
+.global neon_ARGBtoXRGBalphaS
+neon_ARGBtoXRGBalphaS:
+    do_argb_alpha 0, 1
+
+.global neon_ABGRtoXRGBalphaS
+neon_ABGRtoXRGBalphaS:
+    do_argb_alpha 1, 1
 
 @ vim:filetype=armasm