use fistpll mnemonic instead of fistpq (more widely supported) on x86_64 too

this was fixed previously on i386 but the corresponding code on x86_64
was missed.

src/math/x86_64/llrintl.s

@@ -2,6 +2,6 @@
 .type llrintl,@function
 llrintl:
 	fldt 8(%rsp)
-	fistpq 8(%rsp)
+	fistpll 8(%rsp)
 	mov 8(%rsp),%rax
 	ret