x86util: Port all macros to cpuflags

Also do some small cosmetic changes: Drop pointless _MMX suffix from ABSD2
macro name, drop pointless check for MMX support, we always assume MMX is
available in our SIMD code, fix spelling.

libavcodec/x86/audiodsp.asm

@@ -66,7 +66,7 @@ SCALARPRODUCT
 ; %1 = number of xmm registers used
 ; %2 = number of inline load/process/store loops per asm loop
 ; %3 = process 4*mmsize (%3=0) or 8*mmsize (%3=1) bytes per loop
-; %4 = CLIPD function takes min/max as float instead of int (CLIPD_SSE2)
+; %4 = CLIPD function takes min/max as float instead of int (SSE2 version)
 ; %5 = suffix
 %macro VECTOR_CLIP_INT32 4-5
 cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
@@ -122,14 +122,11 @@ cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
 %endmacro
 
 INIT_MMX mmx
-%define CLIPD CLIPD_MMX
 VECTOR_CLIP_INT32 0, 1, 0, 0
 INIT_XMM sse2
 VECTOR_CLIP_INT32 6, 1, 0, 0, _int
-%define CLIPD CLIPD_SSE2
 VECTOR_CLIP_INT32 6, 2, 0, 1
 INIT_XMM sse4
-%define CLIPD CLIPD_SSE41
 %ifdef m8
 VECTOR_CLIP_INT32 11, 1, 1, 0
 %else

libavutil/x86/x86util.asm

@@ -174,13 +174,13 @@
 %endif
 %endmacro
 
-%macro PSIGNW_MMX 2
+%macro PSIGNW 2
+%if cpuflag(ssse3)
+    psignw     %1, %2
+%else
     pxor       %1, %2
     psubw      %1, %2
-%endmacro
-
-%macro PSIGNW_SSSE3 2
-    psignw     %1, %2
+%endif
 %endmacro
 
 %macro ABS1 2
@@ -221,7 +221,7 @@
 %endif
 %endmacro
 
-%macro ABSB 2 ; source mmreg, temp mmreg (unused for ssse3)
+%macro ABSB 2 ; source mmreg, temp mmreg (unused for SSSE3)
 %if cpuflag(ssse3)
     pabsb   %1, %1
 %else
@@ -245,7 +245,7 @@
 %endif
 %endmacro
 
-%macro ABSD2_MMX 4
+%macro ABSD2 4
     pxor    %3, %3
     pxor    %4, %4
     pcmpgtd %3, %1
@@ -290,7 +290,7 @@
 %else
     palignr %1, %2, %3
 %endif
-%elif cpuflag(mmx) ; [dst,] src1, src2, imm, tmp
+%else ; [dst,] src1, src2, imm, tmp
     %define %%dst %1
 %if %0==5
 %ifnidn %1, %2
@@ -606,37 +606,47 @@
     pminsw %1, %3
 %endmacro
 
-%macro PMINSD_MMX 3 ; dst, src, tmp
+%macro PMINSD 3 ; dst, src, tmp/unused
+%if cpuflag(sse4)
+    pminsd    %1, %2
+%elif cpuflag(sse2)
+    cvtdq2ps  %1, %1
+    minps     %1, %2
+    cvtps2dq  %1, %1
+%else
     mova      %3, %2
     pcmpgtd   %3, %1
     pxor      %1, %2
     pand      %1, %3
     pxor      %1, %2
+%endif
 %endmacro
 
-%macro PMAXSD_MMX 3 ; dst, src, tmp
+%macro PMAXSD 3 ; dst, src, tmp/unused
+%if cpuflag(sse4)
+    pmaxsd    %1, %2
+%else
     mova      %3, %1
     pcmpgtd   %3, %2
     pand      %1, %3
     pandn     %3, %2
     por       %1, %3
+%endif
 %endmacro
 
-%macro CLIPD_MMX 3-4 ; src/dst, min, max, tmp
-    PMINSD_MMX %1, %3, %4
-    PMAXSD_MMX %1, %2, %4
-%endmacro
-
-%macro CLIPD_SSE2 3-4 ; src/dst, min (float), max (float), unused
+%macro CLIPD 3-4
+%if cpuflag(sse4);  src/dst, min, max, unused
+    pminsd  %1, %3
+    pmaxsd  %1, %2
+%elif cpuflag(sse2) ; src/dst, min (float), max (float), unused
     cvtdq2ps  %1, %1
     minps     %1, %3
     maxps     %1, %2
     cvtps2dq  %1, %1
-%endmacro
-
-%macro CLIPD_SSE41 3-4 ;  src/dst, min, max, unused
-    pminsd  %1, %3
-    pmaxsd  %1, %2
+%else               ; src/dst, min, max, tmp
+    PMINSD    %1, %3, %4
+    PMAXSD    %1, %2, %4
+%endif
 %endmacro
 
 %macro VBROADCASTSS 2 ; dst xmm/ymm, src m32

libswscale/x86/scale.asm

@@ -364,15 +364,7 @@ cglobal hscale%1to%2_%4, %5, 10, %6, pos0, dst, w, srcmem, filter, fltpos, fltsi
     movd [dstq+wq*2], m0
 %endif ; %3 ==/!= X
 %else ; %2 == 19
-%if mmsize == 8
-    PMINSD_MMX    m0, m2, m4
-%elif cpuflag(sse4)
-    pminsd        m0, m2
-%else ; sse2/ssse3
-    cvtdq2ps      m0, m0
-    minps         m0, m2
-    cvtps2dq      m0, m0
-%endif ; mmx/sse2/ssse3/sse4
+    PMINSD        m0, m2, m4
 %ifnidn %3, X
     mova [dstq+wq*(4>>wshr)], m0
 %else ; %3 == X