aarch64: add ',' between assembler macro arguments where missing

llvm's integrated assembler does not accept spaces as macro argument
delimiter when targeting darwin. Using a explicit delimiter is a good
idea in principle since it makes case like 'macro 4 -2' vs 'macro 4 - 2'
clear.

libavcodec/aarch64/fft_neon.S

@@ -336,7 +336,7 @@ function fft_pass_neon
 endfunc
 
 .macro  def_fft n, n2, n4
-function fft\n\()_neon  align=6
+function fft\n\()_neon, align=6
         sub             sp,  sp,  #16
         stp             x28, x30, [sp]
         add             x28, x0,  #\n4*2*8

libavcodec/aarch64/mpegaudiodsp_neon.S

@@ -24,7 +24,7 @@
 #define WFRAC_BITS  16   // fractional bits for window
 #define OUT_SHIFT (WFRAC_BITS + FRAC_BITS - 15)
 
-const   tbl_rev128.s align=4
+const   tbl_rev128.s, align=4
         .byte           12, 13, 14, 15
         .byte            8,  9, 10, 11
         .byte            4,  5,  6,  7

libavcodec/aarch64/opus_imdct_neon.S

@@ -23,7 +23,7 @@
 #include "asm-offsets.h"
 
 .macro shuffle a, b, c, d
-const shuffle_\a\b\c\d align=4
+const shuffle_\a\b\c\d, align=4
         .byte (\a * 4), (\a * 4 + 1), (\a * 4 + 2), (\a * 4 + 3)
         .byte (\b * 4), (\b * 4 + 1), (\b * 4 + 2), (\b * 4 + 3)
         .byte (\c * 4), (\c * 4 + 1), (\c * 4 + 2), (\c * 4 + 3)
@@ -344,7 +344,7 @@ function fft15_pass
         ret
 endfunc
 
-function fft30_neon  align=6
+function fft30_neon, align=6
         sub             sp,  sp,  #0x20
         stp             x20, x21, [sp]
         stp             x22, x30, [sp, #0x10]
@@ -372,7 +372,7 @@ function fft30_neon  align=6
 endfunc
 
 .macro  def_fft n, n2
-function fft\n\()_neon  align=6
+function fft\n\()_neon, align=6
         sub             sp,  sp,  #0x30
         stp             x20, x21, [sp]
         stp             x22, x30, [sp, #0x10]
@@ -641,7 +641,7 @@ function ff_celt_imdct_half_neon, export=1
 endfunc
 
 // [0] = exp(2 * i * pi / 5), [1] = exp(2 * i * pi * 2 / 5)
-const   fact5           align=4
+const   fact5,          align=4
         .float           0.30901699437494745, 0.95105651629515353
         .float          -0.80901699437494734, 0.58778525229247325
 endconst