This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
RE: Ping: [PATCH v2] x86-64: fix ZMM register state tracking
- From: "Metzger, Markus T" <markus dot t dot metzger at intel dot com>
- To: Jan Beulich <JBeulich at suse dot com>, Simon Marchi <simon dot marchi at polymtl dot ca>
- Cc: GDB <gdb-patches at sourceware dot org>, "Pedro Alves (palves at redhat dot com)" <palves at redhat dot com>
- Date: Wed, 7 Nov 2018 13:18:48 +0000
- Subject: RE: Ping: [PATCH v2] x86-64: fix ZMM register state tracking
- References: <5B8FD8B302000078001E5940@prv1-mh.provo.novell.com> <5BD6E19202000078001F5BC4@prv1-mh.provo.novell.com> <8bafd220c79c3936de8ce3d6421a9c01@polymtl.ca> <A78C989F6D9628469189715575E55B236B380E92@IRSMSX104.ger.corp.intel.com> <5BE2AC5802000078001F8F7D@prv1-mh.provo.novell.com>
> >>> On 07.11.18 at 10:07, <markus.t.metzger@intel.com> wrote:
> >> On 2018-10-29 06:31, Jan Beulich wrote:
> >> >>>> On 10.10.18 at 17:12, wrote:
> >> >> The three AVX512 state components are entirely independent - one
> >> >> being in its "init state" has no implication whatsoever on either
> >> >> of the other two. Fully separate X86_XSTATE_ZMM_H and
> >> >> X86_XSTATE_ZMM handling, to prevent upper halves of the upper 16
> >> >> ZMM registers to display as if they were zero (when they aren't) after e.g.
> >> >> VZEROALL/VZEROUPPER.
> >> >>
> >> >> gdb/
> >> >> 2018-10-10 Jan Beulich <jbeulich@suse.com>
> >> >>
> >> >> * i387-tdep.c (i387_supply_xsave): Split handling of
> >> >> X86_XSTATE_ZMM_H and X86_XSTATE_ZMM.
> >> >> (i387_collect_xsave): Likewise.
> >> >>
> >> >> gdb/testsuite/
> >> >> 2018-10-10 Simon Marchi <simon.marchi@polymtl.ca>
> >> >>
> >> >> * testsuite/gdb.arch/i386-avx512.c,
> >> >> testsuite/gdb.arch/i386-avx512.exp: Add 7th test.
> >> >>
> >> >> ---
> >> >> v2: Attach comments to zmm_endlo_regnum declarations. Add testcase
> >> >> provided by Simon.
> >>
> >> The testcase obviously LGTM. I will let Markus approve the other changes.
> >
> > The code already looked good to me in v1. Thanks for adding comments.
>
> So can I translate this into an ack for me to commit the change?
> Or else, who would be the one to give the go-ahead?
Simon can approve your patch.
IIRC Pedro had a question regarding gdbserver. From a first look, it seems to get the
feature bits right but does not distinguish 32-bit and 64-bit mode regarding the number
of available registers.
Did you run the new test you added also in remote configuration?
Markus.
Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928