From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
To: Simon Marchi <simark@simark.ca>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
Subject: RE: [PATCH v8 10/10] btrace: Extend ptwrite event decoding.
Date: Fri, 31 Mar 2023 10:58:14 +0000 [thread overview]
Message-ID: <MN2PR11MB45664092DB454BDA0772FD838E8F9@MN2PR11MB4566.namprd11.prod.outlook.com> (raw)
In-Reply-To: <8763b1f0-a9ae-861e-6216-47749d9558f3@simark.ca>
> -----Original Message-----
> From: Simon Marchi <simark@simark.ca>
> Sent: Freitag, 24. März 2023 16:41
> To: Willgerodt, Felix <felix.willgerodt@intel.com>; gdb-
> patches@sourceware.org
> Subject: Re: [PATCH v8 10/10] btrace: Extend ptwrite event decoding.
>
> On 3/21/23 11:46, Felix Willgerodt via Gdb-patches wrote:
> > Call the ptwrite filter function whenever a ptwrite event is decoded.
> > The returned string is written to the aux_data string table and a
> > corresponding auxiliary instruction is appended to the function segment.
> > ---
> > gdb/NEWS | 7 +
> > gdb/btrace.c | 54 +++
> > gdb/config.in | 3 +
> > gdb/configure | 11 +
> > gdb/doc/python.texi | 150 ++++++
> > gdb/testsuite/gdb.btrace/i386-ptwrite.S | 550
> ++++++++++++++++++++++
> > gdb/testsuite/gdb.btrace/ptwrite.c | 39 ++
> > gdb/testsuite/gdb.btrace/ptwrite.exp | 200 ++++++++
> > gdb/testsuite/gdb.btrace/x86_64-ptwrite.S | 544
> +++++++++++++++++++++
> > gdb/testsuite/lib/gdb.exp | 72 +++
> > gdbsupport/common.m4 | 2 +
> > gdbsupport/config.in | 3 +
> > gdbsupport/configure | 11 +
> > 13 files changed, 1646 insertions(+)
> > create mode 100644 gdb/testsuite/gdb.btrace/i386-ptwrite.S
> > create mode 100644 gdb/testsuite/gdb.btrace/ptwrite.c
> > create mode 100644 gdb/testsuite/gdb.btrace/ptwrite.exp
> > create mode 100644 gdb/testsuite/gdb.btrace/x86_64-ptwrite.S
> >
> > diff --git a/gdb/NEWS b/gdb/NEWS
> > index cc262f1f8a6..5dd05867f2a 100644
> > --- a/gdb/NEWS
> > +++ b/gdb/NEWS
> > @@ -106,6 +106,13 @@ show always-read-ctf
> >
> > *** Changes in GDB 13
> >
> > +* GDB now supports printing of ptwrite payloads from the Intel Processor
> > + Trace during 'record instruction-history', 'record function-call-history'
> > + and all stepping commands. The payload is also accessible in Python as a
> > + RecordAuxiliary object. Printing is customizable via a ptwrite filter
> > + function in Python. By default, the raw ptwrite payload is printed for
> > + each ptwrite that is encountered.
> > +
> > * MI version 1 is deprecated, and will be removed in GDB 14.
> >
> > * GDB now supports dumping memory tag data for AArch64 MTE. It also
> supports
> > diff --git a/gdb/btrace.c b/gdb/btrace.c
> > index 37dd0b666d8..db0d0e291d9 100644
> > --- a/gdb/btrace.c
> > +++ b/gdb/btrace.c
> > @@ -1253,6 +1253,54 @@ handle_pt_insn_events (struct
> btrace_thread_info *btinfo,
> > bfun->insn_offset - 1, offset);
> >
> > break;
> > +#if defined (HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE)
> > + case ptev_ptwrite:
> > + {
> > + uint64_t ip = 0;
> > + std::string ptw_string;
> > + btrace_insn_flags flags = 0;
> > +
> > + /* Lookup the ip if available. */
> > + if (event.ip_suppressed == 0)
> > + ip = event.variant.ptwrite.ip;
> > +
> > + if (btinfo->ptw_callback_fun != nullptr)
> > + ptw_string
> > + = btinfo->ptw_callback_fun (event.variant.ptwrite.payload,
> > + ip, btinfo->ptw_context);
>
> If ptw_callback_fun is nullptr, the string will always be empty, and the
> data not shown? This means that in a build without Python, the data
> will never be shown?
>
> I think a better approach would be to handle the default case (print as
> hex) here, instead of in a Python function. I think that
> ptw_callback_fun should return a gdb::optional<std::string>. If the
> optional is instantiated (including containing an empty string), use
> that value. If ptw_callback_fun is nullptr (perhaps because GDB was
> built with no Python support), or if the returned optional is not
> instantiated (perhaps because the user did not register any filter),
> generate the default (print as hex) here.
>
We wanted the user to be able to register None as a listener. But I think
that might still be from a time where we didn't implement
the /a modifiers for the instruction/function-history CLI commands to
enable not printing the strings.
I think your suggestion makes sense. I guess in that case we should try
to get rid of the current default filter in Python. To not have two places
in the code that do the same thing.
> > +
> > + if (ptw_string.empty ())
> > + break;
> > +
> > + btinfo->aux_data.emplace_back (ptw_string);
>
> Since you don't need ptw_string below, you can std::move it in the
> vector.
Right, I will do that.
> > +@end defun
> > +
> > +@findex gdb.ptwrite.get_filter
> > +@defun get_filter ()
> > +Return the currently active @code{PTWRITE} filter function.
> > +@end defun
> > +
> > +@findex gdb.ptwrite.default_filter
> > +@defun default_filter (@var{payload}, @var{ip})
> > +The filter function active by default. It prints the payload in hexadecimal
> > +format.
> > +@end defun
> > +
> > +@value{GDBN} creates a new copy of the filter function for each thread
> to
> > +allow for independent internal states. There is no support for registering
>
> Ok, I guess that answers my question about deepcopy. I think it would
> be good to say that the copy is done using the deepcopy module/function.
>
I will add that.
Thanks,
Felix
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de <http://www.intel.de>
Managing Directors: Christin Eisenschmid, Sharon Heck, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
next prev parent reply other threads:[~2023-03-31 10:58 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-21 15:46 [PATCH v8 00/10] Extensions for PTWRITE Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 02/10] btrace: Enable auxiliary instructions in record instruction-history Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 03/10] btrace: Enable auxiliary instructions in record function-call-history Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 04/10] btrace: Handle stepping and goto for auxiliary instructions Felix Willgerodt
2023-03-24 14:09 ` Simon Marchi
2023-03-31 10:58 ` Willgerodt, Felix
2023-03-21 15:46 ` [PATCH v8 05/10] python: Introduce gdb.RecordAuxiliary class Felix Willgerodt
2023-03-24 14:27 ` Simon Marchi
2023-03-31 10:58 ` Willgerodt, Felix
2023-04-03 19:06 ` Simon Marchi
2023-04-04 6:57 ` Metzger, Markus T
2023-04-04 14:17 ` Simon Marchi
2023-04-04 14:26 ` Willgerodt, Felix
2023-03-21 15:46 ` [PATCH v8 06/10] python: Add clear() to gdb.Record Felix Willgerodt
2023-03-24 14:36 ` Simon Marchi
2023-03-31 10:58 ` Willgerodt, Felix
2023-03-21 15:46 ` [PATCH v8 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 08/10] btrace, linux: Enable ptwrite packets Felix Willgerodt
2023-03-21 15:46 ` [PATCH v8 09/10] btrace, python: Enable ptwrite filter registration Felix Willgerodt
2023-03-24 15:23 ` Simon Marchi
2023-03-31 10:58 ` Willgerodt, Felix
2023-04-03 20:44 ` Simon Marchi
2023-04-04 14:42 ` Willgerodt, Felix
2023-04-04 15:06 ` Simon Marchi
2023-04-05 10:20 ` Willgerodt, Felix
2023-04-05 20:27 ` Simon Marchi
2023-04-06 9:44 ` Willgerodt, Felix
2023-03-21 15:46 ` [PATCH v8 10/10] btrace: Extend ptwrite event decoding Felix Willgerodt
2023-03-24 15:40 ` Simon Marchi
2023-03-31 10:58 ` Willgerodt, Felix [this message]
2023-04-04 14:23 ` Simon Marchi
2023-03-24 13:56 ` [PATCH v8 00/10] Extensions for PTWRITE Simon Marchi
2023-03-24 18:23 ` Tom Tromey
2023-03-24 18:28 ` Simon Marchi
2023-03-24 22:29 ` Tom Tromey
2023-03-31 10:57 ` Willgerodt, Felix
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=MN2PR11MB45664092DB454BDA0772FD838E8F9@MN2PR11MB4566.namprd11.prod.outlook.com \
--to=felix.willgerodt@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).