Toolchain/binutils-gdb
1
0
Fork 0
forked from Imagelibrary/binutils-gdb
Code Pull Requests Activity

PR25569, PDP11 ld -s clobbers last data byte

Browse Source 
This patch fixes an ancient wart in aout support, in that text and
data section sizes are rounded up for alignment rather that just the
corresponding header sizes.  Changing section sizes could conceivably
result in buffer overflows if section contents were held in memory.
Also, keeping the original section sizes allows this PR to be fixed
nicely.

bfd/
	PR 25569
	* aoutx.h (adjust_o_magic, adjust_z_magic, adjust_n_magic): Use
	"text", "data" and "bss" section pointer vars.  Don't update
	section size, just exec header sizes.
	(adjust_sizes_and_vmas): Don't update text section size.  Set
	initial exec header a_text.  Print exec headers sizes.
	* pdp11.c (adjust_o_magic, adjust_z_magic, adjust_n_magic),
	(adjust_sizes_and_vmas): Similarly.  Formatting.
	(final_link): Correct final file extension.
gas/
	PR 25569
	* config/obj-aout.c (obj_aout_frob_file_before_fix): Don't loop
	on section size adjustment, instead perform another write if
	exec header size is larger than section size.
This commit is contained in:
Alan Modra
2020-02-20 21:53:44 +10:30
parent 6565bf67ad
commit dda2980f54
5 changed files with 263 additions and 241 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
gas/config/obj-aout.c
View File

@@ -113,31 +113,39 @@ obj_aout_frob_symbol (symbolS *sym, int *punt ATTRIBUTE_UNUSED)
S_GET_NAME (sym));
}
/* Relocation processing may require knowing the VMAs of the sections.
Writing to a section will cause the BFD back end to compute the
VMAs. This function also ensures that file size is large enough
to cover a_text and a_data should text or data be the last section
in the file. */
void
obj_aout_frob_file_before_fix (void)
{
/* Relocation processing may require knowing the VMAs of the sections.
Since writing to a section will cause the BFD back end to compute the
VMAs, fake it out here....
Writing to the end of the section ensures the file contents
extend to cover the entire aligned size. We possibly won't know
the aligned size until after VMAs and sizes are set on the first
bfd_set_section_contents call, so it might be necessary to repeat. */
asection *sec = NULL;
if (data_section->size != 0)
sec = data_section;
else if (text_section->size != 0)
sec = text_section;
if (sec)
asection *sec;
bfd_vma *sizep = NULL;
if ((sec = data_section)->size != 0)
sizep = &exec_hdr (stdoutput)->a_data;
else if ((sec = text_section)->size != 0)
sizep = &exec_hdr (stdoutput)->a_text;
if (sizep)
{
bfd_size_type size;
do
bfd_size_type size = sec->size;
bfd_byte b = 0;
gas_assert (bfd_set_section_contents (stdoutput, sec, &b, size - 1, 1));
/* We don't know the aligned size until after VMAs and sizes are
set on the bfd_set_section_contents call. If that size is
larger than the section then write again to ensure the file
contents extend to cover the aligned size. */
if (*sizep > size)
{
bfd_byte b = 0;
size = sec->size;
gas_assert (bfd_set_section_contents (stdoutput, sec, &b,
size - 1, (bfd_size_type) 1));
} while (size != sec->size);
file_ptr pos = sec->filepos + *sizep;
gas_assert (bfd_seek (stdoutput, pos - 1, SEEK_SET) == 0
&& bfd_bwrite (&b, 1, stdoutput) == 1);
}
}
}