This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH users/roland/osabi] Recognize GNU .note.ABI-tag values 5 (syllable) and 6 (nacl)
- From: Pedro Alves <palves at redhat dot com>
- To: Roland McGrath <mcgrathr at google dot com>
- Cc: GDB <gdb-patches at sourceware dot org>
- Date: Fri, 29 May 2015 18:07:36 +0100
- Subject: Re: [PATCH users/roland/osabi] Recognize GNU .note.ABI-tag values 5 (syllable) and 6 (nacl)
- Authentication-results: sourceware.org; auth=none
- References: <CAB=4xhpRRhMYF7-FtsYuTdCSn3fNOB1gOT_xjYaS11jJ6myGHw at mail dot gmail dot com> <55683F9E dot 7020902 at redhat dot com> <CAB=4xhr8YL3tkjWbRnUb9nRk-uhB9ohCEW7d6=hD7HtrNk=omA at mail dot gmail dot com>
On 05/29/2015 05:47 PM, Roland McGrath wrote:
> On Fri, May 29, 2015 at 3:29 AM, Pedro Alves <palves@redhat.com> wrote:
>> I think adding the constants to include/elf/common.h,
>
> I've already committed common.h based on the binutils approval.
>
>> and making
>> generic_elf_osabi_sniff_abi_tag_sections recognize syllable and nacl,
>> avoiding the internal_error (or warning after your other patch) is OK
>> and could go to the 7.9 branch. Please split that out to a separate patch.
>
> Do you mean just silently ignoring these values?
I do.
>
>> The rest of the gdb bits can be considered for trunk. I don't think we want
>> GDB_OSABI_SYLLABLE exposed to the rest of the gdb (I don't know anything
>> about Syllable either), and to users (both "set osabi" and xml target
>> descriptions). We may need to expose GDB_OSABI_NACL, but that's probably
>> best added along with the nacl port, assuming there's one.
>
> OK. I'll ignore Syllable entirely, since I was only doing it for
> completeness.
>
> For NaCl, I'll investigate further before proposing something. We have a
> patch that is mostly about x86 and needs some cleanup (or probably complete
> rewriting). NaCl's x86-64 is particularly odd and I don't want to worry
> about the weird cruft we have for that right now. At the moment, I'm
> focused on on arm-nacl. NaCl is an all-remote target (the only kind of
> "native" debugging interface available is a gdb stub). So there really is
> almost nothing to do. But when I get some time I'll look into building
> --target=arm-nacl on trunk and see what it needs.
OK. Looking forward to hear about this.
Thanks,
Pedro Alves