x86: use 32-bit source registers with movd instruction

yasm tolerates mismatch between movd/movq and source register size,
adjusting the instruction according to the register.  nasm is more
strict.
Signed-off-by: Mans Rullgard <mans@mansr.com>

libavcodec/x86/h264_deblock_10bit.asm

@@ -165,7 +165,7 @@ cglobal deblock_v_luma_10, 5,5,8*(mmsize/16)
     SUB        rsp, pad
     shl        r2d, 2
     shl        r3d, 2
-    LOAD_AB     m4, m5, r2, r3
+    LOAD_AB     m4, m5, r2d, r3d
     mov         r3, 32/mmsize
     mov         r2, r0
     sub         r0, r1
@@ -222,7 +222,7 @@ cglobal deblock_h_luma_10, 5,6,8*(mmsize/16)
     SUB        rsp, pad
     shl        r2d, 2
     shl        r3d, 2
-    LOAD_AB     m4, m5, r2, r3
+    LOAD_AB     m4, m5, r2d, r3d
     mov         r3, r1
     mova        am, m4
     add         r3, r1
@@ -351,7 +351,7 @@ cglobal deblock_v_luma_10, 5,5,15
     %define mask2 m11
     shl        r2d, 2
     shl        r3d, 2
-    LOAD_AB    m12, m13, r2, r3
+    LOAD_AB    m12, m13, r2d, r3d
     mov         r2, r0
     sub         r0, r1
     sub         r0, r1
@@ -379,7 +379,7 @@ cglobal deblock_v_luma_10, 5,5,15
 cglobal deblock_h_luma_10, 5,7,15
     shl        r2d, 2
     shl        r3d, 2
-    LOAD_AB    m12, m13, r2, r3
+    LOAD_AB    m12, m13, r2d, r3d
     mov         r2, r1
     add         r2, r1
     add         r2, r1
@@ -857,7 +857,7 @@ cglobal deblock_v_chroma_10, 5,7-(mmsize/16),8*(mmsize/16)
 .loop:
 %endif
     CHROMA_V_LOAD r5
-    LOAD_AB     m4, m5, r2, r3
+    LOAD_AB     m4, m5, r2d, r3d
     LOAD_MASK   m0, m1, m2, m3, m4, m5, m7, m6, m4
     pxor        m4, m4
     CHROMA_V_LOAD_TC m6, r4
@@ -891,7 +891,7 @@ cglobal deblock_v_chroma_intra_10, 4,6-(mmsize/16),8*(mmsize/16)
 .loop:
 %endif
     CHROMA_V_LOAD r4
-    LOAD_AB     m4, m5, r2, r3
+    LOAD_AB     m4, m5, r2d, r3d
     LOAD_MASK   m0, m1, m2, m3, m4, m5, m7, m6, m4
     CHROMA_DEBLOCK_P0_Q0_INTRA m1, m2, m0, m3, m7, m5, m6
     CHROMA_V_STORE

libavcodec/x86/rv34dsp.asm

@@ -49,7 +49,7 @@ SECTION .text
 cglobal rv34_idct_%1, 1, 2, 0
     movsx   r1, word [r0]
     IDCT_DC r1
-    movd    m0, r1
+    movd    m0, r1d
     pshufw  m0, m0, 0
     movq    [r0+ 0], m0
     movq    [r0+ 8], m0
@@ -70,7 +70,7 @@ cglobal rv34_idct_dc_add, 3, 3
     ; calculate DC
     IDCT_DC_ROUND r2
     pxor       m1, m1
-    movd       m0, r2
+    movd       m0, r2d
     psubw      m1, m0
     packuswb   m0, m0
     packuswb   m1, m1
@@ -175,7 +175,7 @@ cglobal rv34_idct_dc_add, 3, 3, 6
     pxor       m1, m1
 
     ; calculate DC
-    movd       m0, r2
+    movd       m0, r2d
     lea        r2, [r0+r1*2]
     movd       m2, [r0]
     movd       m3, [r0+r1]

libavcodec/x86/rv40dsp.asm

@@ -466,8 +466,8 @@ cglobal rv40_weight_func_%1_%2, 6, 7, 8
     add        r2, r6
     neg        r6
 
-    movd       m2, r3
-    movd       m3, r4
+    movd       m2, r3d
+    movd       m3, r4d
 %ifidn %1,rnd
 %define  RND   0
     SPLATW     m2, m2