Re: bfd_get_full_section_contents memory leak, plus

["H.J. Lu" <hjl dot tools at gmail dot com>]

On Sat, Oct 20, 2012 at 7:45 PM, Alan Modra <amodra@gmail.com> wrote:
> On Sat, Oct 20, 2012 at 05:30:12PM -0700, H.J. Lu wrote:
>> On Sat, Oct 20, 2012 at 5:26 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
>> > On Sat, Oct 20, 2012 at 5:21 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
>> >> On Sat, Oct 20, 2012 at 3:17 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
>> >>> It breaks objdump.  I checked in this tectase to show the
>> >>> error.
>
>> diff --git a/bfd/compress.c b/bfd/compress.c
>> index 294bfd3..0502f80 100644
>> --- a/bfd/compress.c
>> +++ b/bfd/compress.c
>> @@ -166,6 +166,11 @@ bfd_get_full_section_contents (bfd *abfd, sec_ptr
>> sec, bfd_byte **ptr)
>>    bfd_byte *compressed_buffer;
>>  #endif
>>
>> +  if ((sec->flags & SEC_IN_MEMORY) != 0
>> +      && p != NULL
>> +      && p == sec->contents)
>> +    return TRUE;
>> +
>>    if (abfd->direction != write_direction && sec->rawsize != 0)
>>      sz = sec->rawsize;
>>    else
>
> I think I prefer to put the extra hack where objdump is twiddling
> things behind BFD's back.  I also can't see why sec->size needs to be
> set.  What do you think?
>
>         * objdump.c (load_specific_debug_section): Update compress_status
>         along with setting sec->contents.  No need to set sec->size.
>
> Index: binutils/objdump.c
> ===================================================================
> RCS file: /cvs/src/src/binutils/objdump.c,v
> retrieving revision 1.197
> diff -u -p -r1.197 objdump.c
> --- binutils/objdump.c  18 Oct 2012 17:42:24 -0000      1.197
> +++ binutils/objdump.c  21 Oct 2012 02:41:05 -0000
> @@ -2276,9 +2276,10 @@ load_specific_debug_section (enum dwarf_
>           decompressed), so we store a pointer to the data in
>           the bfd_section, and tell it that the contents are
>           already in memory.  */
> +      if (sec->compress_status == DECOMPRESS_SECTION_SIZED)
> +       sec->compress_status = COMPRESS_SECTION_DONE;
>        sec->contents = section->start;
>        sec->flags |= SEC_IN_MEMORY;
> -      sec->size = section->size;
>
>        ret = bfd_simple_get_relocated_section_contents (abfd,
>                                                        sec,

I agree that we don't need to set size in objdump.  On the other hand,
can we move checking and setting compress_status to compress.c to
make compressing code more robust?

Thanks.

-- 
H.J.