Add MSA optimized I422ToARGBRow_MSA and I422ToRGBARow_MSA functions

include/libyuv/macros_msa.h

 /*
  *  Copyright 2016 The LibYuv Project Authors. All rights reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS. All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
 #ifndef INCLUDE_LIBYUV_MACROS_MSA_H_
 #define INCLUDE_LIBYUV_MACROS_MSA_H_
 
 #if !defined(LIBYUV_DISABLE_MSA) && defined(__mips_msa)
 #include <stdint.h>
 #include <msa.h>
 
+#if (__mips_isa_rev >= 6)
+  #define LW(psrc) ( {                     \
+    uint8 *psrc_lw_m = (uint8 *) (psrc);   \
+    uint32 val_m;                          \
+                                           \
+    asm volatile (                         \
+      "lw  %[val_m],  %[psrc_lw_m]  \n\t"  \
+                                           \

[fbarchard1, 2016/10/24 17:36:31]

remove blank line

+      : [val_m] "=r" (val_m)               \
+      : [psrc_lw_m] "m" (*psrc_lw_m)       \
+    );                                     \
+                                           \
+    val_m;                                 \
+  } )
+
+  #if (__mips == 64)
+    #define LD(psrc) ( {                     \
+      uint8 *psrc_ld_m = (uint8 *) (psrc);   \
+      uint64 val_m = 0;                      \
+                                             \
+      asm volatile (                         \
+        "ld  %[val_m],  %[psrc_ld_m]  \n\t"  \
+                                             \
+        : [val_m] "=r" (val_m)               \
+        : [psrc_ld_m] "m" (*psrc_ld_m)       \
+      );                                     \
+                                             \
+      val_m;                                 \
+    } )
+  #else  // !(__mips == 64)
+    #define LD(psrc) ( {                                      \
+      uint8 *psrc_ld_m = (uint8 *) (psrc);                    \
+      uint32 val0_m, val1_m;                                  \
+      uint64 val_m = 0;                                       \
+                                                              \
+      val0_m = LW(psrc_ld_m);                                 \
+      val1_m = LW(psrc_ld_m + 4);                             \
+                                                              \
+      val_m = (uint64) (val1_m);                              \
+      val_m = (uint64) ((val_m << 32) & 0xFFFFFFFF00000000);  \
+      val_m = (uint64) (val_m | (uint64) val0_m);             \
+                                                              \
+      val_m;                                                  \
+    } )
+  #endif  // (__mips == 64)
+#else  // !(__mips_isa_rev >= 6)
+  #define LW(psrc) ( {                      \
+    uint8 *psrc_lw_m = (uint8 *) (psrc);    \
+    uint32 val_m;                           \
+                                            \
+    asm volatile (                          \
+      "ulw  %[val_m],  %[psrc_lw_m]  \n\t"  \
+                                            \
+      : [val_m] "=r" (val_m)                \
+      : [psrc_lw_m] "m" (*psrc_lw_m)        \
+    );                                      \
+                                            \
+    val_m;                                  \
+  } )
+
+  #if (__mips == 64)
+    #define LD(psrc) ( {                      \
+      uint8 *psrc_ld_m = (uint8 *) (psrc);    \
+      uint64 val_m = 0;                       \
+                                              \
+      asm volatile (                          \
+        "uld  %[val_m],  %[psrc_ld_m]  \n\t"  \
+                                              \
+        : [val_m] "=r" (val_m)                \
+        : [psrc_ld_m] "m" (*psrc_ld_m)        \
+      );                                      \
+                                              \
+      val_m;                                  \
+    } )
+  #else  // !(__mips == 64)
+    #define LD(psrc) ( {                                      \
+      uint8 *psrc_ld_m = (uint8 *) (psrc);                    \
+      uint32 val0_m, val1_m;                                  \
+      uint64 val_m = 0;                                       \
+                                                              \
+      val0_m = LW(psrc_ld_m);                                 \
+      val1_m = LW(psrc_ld_m + 4);                             \
+                                                              \
+      val_m = (uint64) (val1_m);                              \
+      val_m = (uint64) ((val_m << 32) & 0xFFFFFFFF00000000);  \
+      val_m = (uint64) (val_m | (uint64) val0_m);             \
+                                                              \
+      val_m;                                                  \
+    } )
+  #endif  // (__mips == 64)
+#endif  // (__mips_isa_rev >= 6)
+
 #define LD_B(RTYPE, psrc) *((RTYPE*)(psrc))   /* NOLINT */
 #define LD_UB(...) LD_B(v16u8, __VA_ARGS__)
 
 #define ST_B(RTYPE, in, pdst) *((RTYPE*)(pdst)) = (in)  /* NOLINT */
 #define ST_UB(...) ST_B(v16u8, __VA_ARGS__)
 
 /* Description : Load two vectors with 16 'byte' sized elements
    Arguments   : Inputs  - psrc, stride
                  Outputs - out0, out1
                  Return Type - as per RTYPE
@@ skipping 52 matching lines @@
 */
 #define ILVRL_B2(RTYPE, in0, in1, out0, out1) {           \
   out0 = (RTYPE) __msa_ilvr_b((v16i8) in0, (v16i8) in1);  \
   out1 = (RTYPE) __msa_ilvl_b((v16i8) in0, (v16i8) in1);  \
 }
 #define ILVRL_B2_UB(...) ILVRL_B2(v16u8, __VA_ARGS__)
 
 #endif  /* !defined(LIBYUV_DISABLE_MSA) && defined(__mips_msa) */
 
 #endif  // INCLUDE_LIBYUV_MACROS_MSA_H_