public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: John Baldwin <jhb@FreeBSD.org>
To: Simon Marchi <simark@simark.ca>, gdb-patches@sourceware.org
Cc: Aleksandar Paunovic <aleksandar.paunovic@intel.com>
Subject: Re: [PATCH v4 12/13] gdbserver: Read offsets of the XSAVE extended region via CPUID
Date: Wed, 12 Apr 2023 14:07:59 -0700	[thread overview]
Message-ID: <80ddaec3-f871-d318-57e6-22669a4f3887@FreeBSD.org> (raw)
In-Reply-To: <f69a8d3a-193f-7d22-43f3-364763eceb76@simark.ca>

On 4/12/23 12:11 PM, Simon Marchi wrote:
> On 3/17/23 21:09, John Baldwin wrote:
>> From: Aleksandar Paunovic <aleksandar.paunovic@intel.com>
> Finally, do you think this class could eventually be shared with GDB,
> and could help simplify the implementation there?  The i387_xsave could
> take a reference to an existing x86_xsave_layout object to take the
> offsets from, rather than referring to the global instance as it does
> currently.

What I would really like to do (and planned to try as a followup after
this series) is try to use regcache_map_entry arrays to describe the
various XSAVE regions.  Those arrays could be shared between GDB and
gdbserver (might need to make some of the regcache_map_entry stuff
be shared instead of GDB-only).  I want to do this for eventual FreeBSD
support in gdbserver as well where I want to share the regcache maps
used in the foo-fbsd-tdep.c files with the equivalent low.cc file over
in gdbserver to describe the register sets available via ptrace.

But using regcache_map_entry would mean you would do something like this
for the AVX region:

/* The AVX region includes the YMMH registers.  */
const struct regcache_map_entry xsave_avx_regmap[] =
{
   { 16, 0, 16 },  /* ymmh0 ... ymmh15 */
};

const struct regset xsave_avx_regset =
{
   xsave_avx_regmap,
   regcache_collect_regset, regcache_supply_regset
};

Similar to how aarch64_fbsd_tls_regset is handled in aarch64-fbsd-tdep.c,
but using I387_YMMH0_REGNUM (tdep) instead of tdep->tls_base_regnum as
the base register number.

Just as in the case of aarch64-fbsd I share the same regset for TLS between
both the tdep.c and nat.c, the regsets for XSAVE regions could be shared
between GDB and gdbserver.

The gdbserver patches here I have only modified lightly from Aleksandar's
original work.  I can work on them further, but I think if I want to unify
them (probably using regcache maps) I'd like to do that as a followup series.

-- 
John Baldwin


  reply	other threads:[~2023-04-12 21:08 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-18  1:08 [PATCH v4 00/13] Handle variable XSAVE layouts John Baldwin
2023-03-18  1:08 ` [PATCH v4 01/13] x86: Add an x86_xsave_layout structure to handle " John Baldwin
2023-03-28 11:35   ` George, Jini Susan
2023-04-10 19:28     ` John Baldwin
2023-04-06 19:09   ` Simon Marchi
2023-04-10 20:03     ` John Baldwin
2023-04-11  1:19       ` Simon Marchi
2023-04-11 14:23     ` Pedro Alves
2023-04-11 16:02       ` John Baldwin
2023-03-18  1:08 ` [PATCH v4 02/13] core: Support fetching TARGET_OBJECT_X86_XSAVE_LAYOUT from architectures John Baldwin
2023-04-06 19:28   ` Simon Marchi
2023-04-10 20:42     ` John Baldwin
2023-04-11  1:49       ` Simon Marchi
2023-04-11 16:06         ` John Baldwin
2023-04-11 16:21           ` Simon Marchi
2023-04-11 23:59             ` John Baldwin
2023-03-18  1:08 ` [PATCH v4 03/13] nat/x86-cpuid.h: Add x86_cpuid_count wrapper around __get_cpuid_count John Baldwin
2023-04-06 19:33   ` Simon Marchi
2023-04-10 20:49     ` John Baldwin
2023-04-11  1:49       ` Simon Marchi
2023-03-18  1:08 ` [PATCH v4 04/13] x86 nat: Add helper functions to save the XSAVE layout for the host John Baldwin
2023-04-06 19:40   ` Simon Marchi
2023-04-10 21:00     ` John Baldwin
2023-04-11  1:51       ` Simon Marchi
2023-03-18  1:08 ` [PATCH v4 05/13] gdb: Update x86 FreeBSD architectures to support XSAVE layouts John Baldwin
2023-04-06 19:54   ` Simon Marchi
2023-04-10 21:02     ` John Baldwin
2023-04-11  1:55       ` Simon Marchi
2023-03-18  1:08 ` [PATCH v4 06/13] gdb: Support XSAVE layouts for the current host in the FreeBSD x86 targets John Baldwin
2023-04-06 20:18   ` Simon Marchi
2023-04-10 21:27     ` John Baldwin
2023-04-11  2:23       ` Simon Marchi
2023-04-11 16:19         ` John Baldwin
2023-04-11 16:46           ` Simon Marchi
2023-04-11 21:37             ` John Baldwin
2023-04-11 22:35               ` John Baldwin
2023-04-12 14:35                 ` Simon Marchi
2023-03-18  1:08 ` [PATCH v4 07/13] gdb: Update x86 Linux architectures to support XSAVE layouts John Baldwin
2023-04-07  1:43   ` Simon Marchi
2023-04-10 21:29     ` John Baldwin
2023-03-18  1:09 ` [PATCH v4 08/13] gdb: Support XSAVE layouts for the current host in the Linux x86 targets John Baldwin
2023-04-07  1:54   ` Simon Marchi
2023-03-18  1:09 ` [PATCH v4 09/13] gdb: Use x86_xstate_layout to parse the XSAVE extended state area John Baldwin
2023-04-07  2:13   ` Simon Marchi
2023-04-10 21:40     ` John Baldwin
2023-03-18  1:09 ` [PATCH v4 10/13] gdbserver: Add a function to set the XSAVE mask and size John Baldwin
2023-04-12 15:08   ` Simon Marchi
2023-04-27 17:24     ` John Baldwin
2023-03-18  1:09 ` [PATCH v4 11/13] gdbserver: Refactor the legacy region within the xsave struct John Baldwin
2023-04-12 18:34   ` Simon Marchi
2023-04-27 19:51     ` John Baldwin
2023-03-18  1:09 ` [PATCH v4 12/13] gdbserver: Read offsets of the XSAVE extended region via CPUID John Baldwin
2023-04-11 14:46   ` Pedro Alves
2023-04-11 16:25     ` John Baldwin
2023-04-12 19:11   ` Simon Marchi
2023-04-12 21:07     ` John Baldwin [this message]
2023-04-13 15:07       ` Simon Marchi
2023-03-18  1:09 ` [PATCH v4 13/13] x86: Remove X86_XSTATE_SIZE and related constants John Baldwin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=80ddaec3-f871-d318-57e6-22669a4f3887@FreeBSD.org \
    --to=jhb@freebsd.org \
    --cc=aleksandar.paunovic@intel.com \
    --cc=gdb-patches@sourceware.org \
    --cc=simark@simark.ca \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).