x86: replace explicit REP_RETs with RETs

From x86inc:
> On AMD cpus <=K10, an ordinary ret is slow if it immediately follows either
> a branch or a branch target. So switch to a 2-byte form of ret in that case.
> We can automatically detect "follows a branch", but not a branch target.
> (SSSE3 is a sufficient condition to know that your cpu doesn't have this problem.)

x86inc can automatically determine whether to use REP_RET rather than
REP in most of these cases, so impact is minimal. Additionally, a few
REP_RETs were used unnecessary, despite the return being nowhere near a
branch.

The only CPUs affected were AMD K10s, made between 2007 and 2011, 16
years ago and 12 years ago, respectively.

In the future, everyone involved with x86inc should consider dropping
REP_RETs altogether.

libswresample/x86/audio_convert.asm

@@ -85,7 +85,7 @@ pack_2ch_%2_to_%1_u_int %+ SUFFIX:
     add lenq, 2*mmsize/(2<<%4)
 %endif
         jl .next
-    REP_RET
+    RET
 %endmacro
 
 %macro UNPACK_2CH 5-7
@@ -157,7 +157,7 @@ unpack_2ch_%2_to_%1_u_int %+ SUFFIX:
     add lenq, mmsize/(1<<%4)
 %endif
         jl .next
-    REP_RET
+    RET
 %endmacro
 
 %macro CONV 5-7
@@ -198,7 +198,7 @@ cglobal %2_to_%1_%3, 3, 3, 6, dst, src, len
     emms
     RET
 %else
-    REP_RET
+    RET
 %endif
 %endmacro
 
@@ -301,7 +301,7 @@ pack_6ch_%2_to_%1_u_int %+ SUFFIX:
     emms
     RET
 %else
-    REP_RET
+    RET
 %endif
 %endmacro
 
@@ -375,7 +375,7 @@ unpack_6ch_%2_to_%1_u_int %+ SUFFIX:
     add      dstq, mmsize
     sub      lend, mmsize/4
     jg .loop
-    REP_RET
+    RET
 %endmacro
 
 %define PACK_8CH_GPRS (10 * ARCH_X86_64) + ((6 + HAVE_ALIGNED_STACK) * ARCH_X86_32)
@@ -525,7 +525,7 @@ pack_8ch_%2_to_%1_u_int %+ SUFFIX:
 %endif
     sub      lend, mmsize/4
     jg .loop
-    REP_RET
+    RET
 %endmacro
 
 %macro INT16_TO_INT32_N 6

libswresample/x86/rematrix.asm

@@ -68,7 +68,7 @@ mix_2_1_float_u_int %+ SUFFIX:
     mov%1  [outq + lenq + mmsize], m2
     add        lenq, mmsize*2
         jl .next
-    REP_RET
+    RET
 %endmacro
 
 %macro MIX1_FLT 1
@@ -100,7 +100,7 @@ mix_1_1_float_u_int %+ SUFFIX:
     mov%1  [outq + lenq + mmsize], m1
     add        lenq, mmsize*2
         jl .next
-    REP_RET
+    RET
 %endmacro
 
 %macro MIX1_INT16 1
@@ -152,7 +152,7 @@ mix_1_1_int16_u_int %+ SUFFIX:
     emms
     RET
 %else
-    REP_RET
+    RET
 %endif
 %endmacro
 
@@ -218,7 +218,7 @@ mix_2_1_int16_u_int %+ SUFFIX:
     emms
     RET
 %else
-    REP_RET
+    RET
 %endif
 %endmacro