patch for bfd/dwarf2.c, support for 16-bit addresses in DWARF-2

Nick Clifton nickc@cygnus.com
Wed Feb 23 14:25:00 GMT 2000


Hi Stephane,

: > There is code in _bfd_dwarf2_find_nearest_line () that also checks to
: > see if addr_size is 4 or 8.  Why have you not patched this function
: > as well ?
: > 
: Because it is correct.
: 
: There is a confusion in bfd and gdb about addr_size.
: 
: In _bfd_dwarf2_find_nearest_line(), the 'addr_size' corresponds to the
: ELF format size: 4 for ELF32 and 8 for ELF64. It is used to represent
: size/offset/index within the ELF sections.
: 
: This 'addr_size' has nothing to do with the address size of the target 
: specified by the DW_AT_address_class. It should better be named
: 'elf_addr_size'
: or something like this.

Thanks for the explanation.


: 2000-02-23  Stephane Carrez  <stcarrez@worldnet.fr>
: 
: 	* dwarf2read.c (dwarf2_build_psymtabs_hard): Use
: 	bfd_arch_bits_per_address to get the size of addresses.
: 	(read_address): Read 16-bits addresses.

Umm - your patch does not include any modification to
dwarf2_build_psymtabs_hard. 

But it does include a modification to parse_comp_unit, which is not
mentioned.

However that is just nitpiking, I can pick the correct ChangeLog entry
up from your previous patch.

So I think the patch is OK now, and I will check it in for you.

Cheers
	Nick


More information about the Binutils mailing list