I found the function bfd_find_nearest_line() has a memory leak bug in
bfd/dwarf2.c of binutils-2.19. The variable "stash->dwarf_str_buffer"
in the funciton read_indirect_string() has memory allocated by
read_section() and it is never freed, as follows.
% grep -n dwarf_str_buffer binutils-2.19/bfd/dwarf2.c
129: bfd_byte *dwarf_str_buffer;
555: &stash->dwarf_str_buffer, &stash->dwarf_str_size))
558: str = (char *) stash->dwarf_str_buffer + offset;
% cat binutils-2.19/bfd/dwarf2.c
...
84 struct dwarf2_debug
85 {
...
128 /* Pointer to the .debug_str section loaded into memory. */
129 bfd_byte *dwarf_str_buffer;
...
172 };
...
406 static bfd_boolean
407 read_section (bfd *abfd,
408 const char* section_name, const char*
compressed_section_name,
409 asymbol** syms, bfd_uint64_t offset,
410 bfd_byte
**section_buffer, bfd_size_type *section_size)
411 {
...
415 /* read_section is a noop if the section has already been read. */
416 if (*section_buffer)
417 return TRUE;
...
432 if (syms)
433 {
...
435 *section_buffer
436 = bfd_simple_get_relocated_section_contents (abfd, msec,
NULL, syms);
...
439 }
440 else
441 {
...
443 *section_buffer = bfd_malloc (*section_size);
...
449 }
...
472 }
...
538 static char *
539 read_indirect_string (struct comp_unit* unit,
540 bfd_byte *buf,
541 unsigned
int *bytes_read_ptr)
542 {
...
553 if (! read_section (unit->abfd, ".debug_str", ".zdebug_str",
554 0, offset,
555 &stash->dwarf_str_buffer, &stash-
>dwarf_str_size))
556 return 0;
557
558 str = (char *) stash-
>dwarf_str_buffer + offset;
...
562 }
...
Binutils-2.18 does not have the same problem, because the memory for
"stash->dwarf_str_buffer" is allocated by bfd_alloc().
--
Summary: The function bfd_find_nearest_line() leaks memory.
Product: binutils
Version: 2.19
Status: NEW
Severity: normal
Priority: P2
Component: binutils
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: Tomohisa dot Tanaka at gmail dot com
CC: bug-binutils at gnu dot org
http://sourceware.org/bugzilla/show_bug.cgi?id=7012
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
bug-binutils mailing list
bug-binutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-binutils
