aboutsummaryrefslogtreecommitdiff
path: root/bfd/peXXigen.c
diff options
context:
space:
mode:
authorNick Clifton <nickc@redhat.com>2014-06-03 17:18:27 +0100
committerNick Clifton <nickc@redhat.com>2014-06-03 17:18:27 +0100
commitc32abae8456a2cb959862626b5ff9ebdd1543514 (patch)
treecfa08f37f86ee7006d0b25d3f94f4c80728ce408 /bfd/peXXigen.c
parent1e2ccb612d2b61014bb7e9fef3eb58e4947b9d2b (diff)
downloadgdb-c32abae8456a2cb959862626b5ff9ebdd1543514.zip
gdb-c32abae8456a2cb959862626b5ff9ebdd1543514.tar.gz
gdb-c32abae8456a2cb959862626b5ff9ebdd1543514.tar.bz2
This is to fix a further problem with merging resource sections. It turns
out that the section is supposed to be page-aligned, but the newly merged section was not being padded out to a page boundary. This meant that when the executable was stripped a badly sized .rsrc section was written out. PR ld/16807 * peXXigen.c (rsrc_process_section): Page align the new contents befgore writing out.
Diffstat (limited to 'bfd/peXXigen.c')
-rw-r--r--bfd/peXXigen.c32
1 files changed, 30 insertions, 2 deletions
diff --git a/bfd/peXXigen.c b/bfd/peXXigen.c
index dba0edd..3a0dd13 100644
--- a/bfd/peXXigen.c
+++ b/bfd/peXXigen.c
@@ -2398,7 +2398,16 @@ rsrc_print_section (bfd * abfd, void * vfile)
if (data == (regions.section_end - 4))
data = regions.section_end;
else if (data < regions.section_end)
- fprintf (file, _("\nWARNING: Extra data in .rsrc section - it will be ignored by Windows:\n"));
+ {
+ /* If the extra data is all zeros then do not complain.
+ This is just padding so that the section meets the
+ page size requirements. */
+ while (data ++ < regions.section_end)
+ if (*data != 0)
+ break;
+ if (data < regions.section_end)
+ fprintf (file, _("\nWARNING: Extra data in .rsrc section - it will be ignored by Windows:\n"));
+ }
}
}
@@ -3911,6 +3920,7 @@ rsrc_process_section (bfd * abfd,
data = bfd_malloc (size);
if (data == NULL)
return;
+
datastart = data;
if (! bfd_get_section_contents (abfd, sec, data, 0, size))
@@ -4052,7 +4062,7 @@ rsrc_process_section (bfd * abfd,
starts on an 8-byte boundary. FIXME: Is this correct ? */
sizeof_strings = (sizeof_strings + 7) & ~ 7;
- new_data = bfd_malloc (size);
+ new_data = bfd_zalloc (abfd, size);
if (new_data == NULL)
goto end;
@@ -4069,6 +4079,24 @@ rsrc_process_section (bfd * abfd,
/* Step five: Replace the old contents with the new.
We recompute the size as we may have lost entries due to mergeing. */
size = ((write_data.next_data - new_data) + 3) & ~ 3;
+
+ {
+ int page_size;
+
+ if (coff_data (abfd)->link_info)
+ {
+ page_size = pe_data (abfd)->pe_opthdr.FileAlignment;
+
+ /* If no file alignment has been set, default to one.
+ This repairs 'ld -r' for arm-wince-pe target. */
+ if (page_size == 0)
+ page_size = 1;
+ }
+ else
+ page_size = PE_DEF_FILE_ALIGNMENT;
+ size = (size + page_size - 1) & - page_size;
+ }
+
bfd_set_section_contents (pfinfo->output_bfd, sec, new_data, 0, size);
sec->size = sec->rawsize = size;