Fix tic54x regression

One of the tic54x testcases looks for a section alignment of 1.  After
9136aa49 the alignment became 0.  While it happens that an alignment
of 0 is treated as an alignment of 1, there is no reason to not apply
the explicit alignment.

	* write.c (record_alignment): Revert 2016-02-18 change.

diff --git a/gas/ChangeLog b/gas/ChangeLog
index 9c2b9ec..24cf393 100644 (file)
--- a/gas/ChangeLog
+++ b/gas/ChangeLog
@@ -1,5 +1,9 @@
 2016-03-22  Alan Modra  <amodra@gmail.com>
 
+       * write.c (record_alignment): Revert 2016-02-18 change.
+
+2016-03-22  Alan Modra  <amodra@gmail.com>
+
        * config/tc-alpha.c (load_expression): Replace alloca with xmalloc.
        (emit_jsrjmp, tc_gen_reloc): Likewise.
        * config/tc-i370.c (i370_macro): Likewise.

diff --git a/gas/write.c b/gas/write.c
index 24cd6ca..15330cf 100644 (file)
--- a/gas/write.c
+++ b/gas/write.c
@@ -360,9 +360,6 @@ record_alignment (/* Segment to which alignment pertains.  */
   if (seg == absolute_section)
     return;
 
-  if (align <= OCTETS_PER_BYTE_POWER)
-    return;
-
   if (align > bfd_get_section_alignment (stdoutput, seg))
     bfd_set_section_alignment (stdoutput, seg, align);
 }