Fix typos

src/asm/mulx_mont_256-x86_64.pl

@@ -10,7 +10,7 @@
 # In general Montgomery multiplication algorithm can handle one of the
 # inputs being non-reduced and capped by 1<<radix_width, 1<<256 in this
 # case, rather than the modulus. Whether or not mul_mont_sparse_256, a
-# *taylored* implementation of the algorithm, can handle such input can
+# *tailored* implementation of the algorithm, can handle such input can
 # be circumstantial. For example, in most general case it depends on
 # similar "bit sparsity" of individual limbs of the second, fully reduced
 # multiplicand. If you can't make such assumption about the limbs, then

src/asm/x86_64-xlate.pl

@@ -1919,7 +1919,7 @@ sub process {
 #	...
 #	mov	-8(%rbp),%rbx
 #	mov	%rbp,%rsp
-# .cfi_def_cfa_regiser	%rsp
+# .cfi_def_cfa_register	%rsp
 #	pop	%rbp		# recognized by Windows
 # .cfi_pop	%rbp
 # .cfi_epilogue