This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
RE: [RFC-v5] Fix .text section offset for windows DLL (was Calling __stdcall functions in the inferior)
- From: "Pierre Muller" <pierre dot muller at ics-cnrs dot unistra dot fr>
- To: "'Pedro Alves'" <palves at redhat dot com>, "'GDB Patches'" <gdb-patches at sourceware dot org>
- Date: Thu, 13 Dec 2012 15:56:04 +0100
- Subject: RE: [RFC-v5] Fix .text section offset for windows DLL (was Calling __stdcall functions in the inferior)
- References: <20121024194517.GK3555@adacore.com> <011901cdb2ab$48076b90$d81642b0$@muller@ics-cnrs.unistra.fr> <20121105171121.GA2972@adacore.com> <50991f5f.8382440a.1100.ffff82abSMTPIN_ADDED@mx.google.com> <509ABA17.30507@redhat.com> <000301cdbd96$f5cd9f10$e168dd30$@muller@ics-cnrs.unistra.fr> <20121122173019.GF9964@adacore.com> <15690.5992342674$1353883881@news.gmane.org> <87624si9ur.fsf@fleche.redhat.com> <001501cdccaf$ad85e9b0$0891bd10$@muller@ics-cnrs.unistra.fr> <20121207071035.GG31477@adacore.com> <50C20A66.70002@gmail.com> <29545.4593528577$1354894901@news.gmane.org> <50C21696.7040006@gmail.com> <50c21914.a750420a.2ec3.ffffe4ffSMTPIN_ADDED_BROKEN@mx.google.com> <50C222D5.4000802@redhat.com> <50C22C20.8090906@redhat.com> <008101cdd920$907e7580$b17b6080$@muller@ics-cnrs.unistra.fr> <50C9B6E0.1050709@redhat.com> <00a201cdd931$b0ee13f0$12ca3bd0$@muller@ics-cnrs.unistra.fr> <50C9E742.1030207@redhat.com>
> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pedro Alves
> Envoyà : jeudi 13 dÃcembre 2012 15:34
> Ã : Pierre Muller; GDB Patches
> Objet : Re: [RFC-v5] Fix .text section offset for windows DLL (was Calling
> __stdcall functions in the inferior)
>
> On 12/13/2012 12:59 PM, Pierre Muller wrote:
> >> > There was also the point that section names in PE headers are stored in
> 8
> >> > character arrays, and are not necessarily zero-terminated --
> >> > get_pe_section_index
> >> > does strcmp without accounting for this (grep for SCNNMLEN in bfd).
> > the section_name are copied using xstrdup from sec_name
> > char array of length 9, (indexes 0 to 8)
> > which is reset using memset for each section.
> > only the indexes 0to 7 are possibly modified by the bfd_bread call,
> > so that sec_name[8] always should remain a valid termination of the
> string, no?
>
> I see now, thanks. I think this would make the code a tiny bit clearer.
> WDYT?
Yes, you are right...
Maybe removing the memset before and replacing it by a
sec_name[SCNNMLEN] = '\0';
after the call to bfd_bread
would make it even more clear that you
always get a nicely zero terminated string...
Pierre
> 2012-12-13 Pedro Alves <palves@redhat.com>
>
> * coff-pe-read.c (read_pe_exported_syms): Use SCNNMLEN instead of
> hardcoded 8.
>
> ---
>
> gdb/coff-pe-read.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/gdb/coff-pe-read.c b/gdb/coff-pe-read.c
> index 10bba78..aa4903a 100644
> --- a/gdb/coff-pe-read.c
> +++ b/gdb/coff-pe-read.c
> @@ -464,12 +464,12 @@ read_pe_exported_syms (struct objfile *objfile)
> unsigned long vsize = pe_get32 (dll, secptr1 + 8);
> unsigned long vaddr = pe_get32 (dll, secptr1 + 12);
> unsigned long characteristics = pe_get32 (dll, secptr1 + 36);
> - char sec_name[9];
> + char sec_name[SCNNMLEN + 1];
> int sectix;
>
> memset (sec_name, 0, sizeof (sec_name));
> bfd_seek (dll, (file_ptr) secptr1 + 0, SEEK_SET);
> - bfd_bread (sec_name, (bfd_size_type) 8, dll);
> + bfd_bread (sec_name, (bfd_size_type) SCNNMLEN, dll);
>
> sectix = read_pe_section_index (sec_name);
>