recordmcount: Fix nop_mcount() function

The removal of the longjmp code in recordmcount.c mistakenly made the return
of make_nop() being negative an exit of nop_mcount(). It should not exit the
routine, but instead just not process that part of the code. By exiting with
an error code, it would cause the update of recordmcount to fail some files
which would fail the build if ftrace function tracing was enabled.

Link: http://lkml.kernel.org/r/20191009110538.5909fec6@gandalf.local.home

Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Fixes: 3f1df12 ("recordmcount: Rewrite error/success handling")
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>

scripts/recordmcount.h

@@ -389,11 +389,8 @@ static int nop_mcount(Elf_Shdr const *const relhdr,
 			mcountsym = get_mcountsym(sym0, relp, str0);
 
 		if (mcountsym == Elf_r_sym(relp) && !is_fake_mcount(relp)) {
-			if (make_nop) {
+			if (make_nop)
 				ret = make_nop((void *)ehdr, _w(shdr->sh_offset) + _w(relp->r_offset));
-				if (ret < 0)
-					return -1;
-			}
 			if (warn_on_notrace_sect && !once) {
 				printf("Section %s has mcount callers being ignored\n",
 				       txtname);