public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
* [PATCH v6 00/10] Extensions for PTWRITE
@ 2022-09-16 11:36 Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
                   ` (10 more replies)
  0 siblings, 11 replies; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

The older revisions can be found here:
V1: https://sourceware.org/pipermail/gdb-patches/2019-May/157933.html
V2: https://sourceware.org/pipermail/gdb-patches/2021-June/179908.html
V3: https://sourceware.org/pipermail/gdb-patches/2021-June/180035.html
v4: https://sourceware.org/pipermail/gdb-patches/2022-May/188772.html
v5: https://sourceware.org/pipermail/gdb-patches/2022-June/190236.html

The documentation parts basically didn't change since Eli approved them.
I only made a very minor documentation change in patch 7 to reflect the
change to a boolean.

Here is the status of each patch:
1: Approved.
2: Fixed minor formatting comment.
3: Renamed a define based on comments from Markus.
4: Approved. Only fixed a nit.
5: Fixed minor formatting comment.
6: Approved by Markus, with a request that other maintainers might want
   to look at the Python parts.
7: Changed Qbtrace-conf:pt:ptwrite to a boolean.
8: Approved. Only fixed a nit.
9: Reworked GIL, error messages and DECREF logic in py-record-btrace.c.
   Reworked ptwrite.py according to Markus comments.
10: Approved. Only fixed a nit.

Regards,
Felix


Felix Willgerodt (10):
  btrace: Introduce auxiliary instructions.
  btrace: Enable auxiliary instructions in record instruction-history.
  btrace: Enable auxiliary instructions in record function-call-history.
  btrace: Handle stepping and goto for auxiliary instructions.
  python: Introduce gdb.RecordAuxiliary class.
  python: Add clear() to gdb.Record.
  btrace, gdbserver: Add ptwrite to btrace_config_pt.
  btrace, linux: Enable ptwrite packets.
  btrace, python: Enable ptwrite filter registration.
  btrace: Extend ptwrite event decoding.

 gdb/NEWS                                      |   6 +
 gdb/btrace.c                                  |  68 ++-
 gdb/btrace.h                                  |  38 +-
 gdb/config.in                                 |   3 +
 gdb/configure                                 |  11 +
 gdb/data-directory/Makefile.in                |   1 +
 gdb/disasm-flags.h                            |   1 +
 gdb/doc/gdb.texinfo                           |  33 +-
 gdb/doc/python.texi                           | 168 ++++++
 gdb/extension-priv.h                          |   5 +
 gdb/extension.c                               |  13 +
 gdb/extension.h                               |   3 +
 gdb/features/btrace-conf.dtd                  |   1 +
 gdb/guile/guile.c                             |   1 +
 gdb/nat/linux-btrace.c                        |  29 +
 gdb/python/lib/gdb/ptwrite.py                 |  80 +++
 gdb/python/py-record-btrace.c                 | 136 ++++-
 gdb/python/py-record-btrace.h                 |  11 +
 gdb/python/py-record.c                        |  89 ++-
 gdb/python/py-record.h                        |   3 +
 gdb/python/python-internal.h                  |   3 +
 gdb/python/python.c                           |   2 +
 gdb/record-btrace.c                           | 107 +++-
 gdb/record.c                                  |  10 +
 gdb/record.h                                  |   5 +-
 gdb/remote.c                                  |  30 +
 gdb/testsuite/gdb.btrace/i386-ptwrite.S       | 550 ++++++++++++++++++
 gdb/testsuite/gdb.btrace/ptwrite.c            |  39 ++
 gdb/testsuite/gdb.btrace/ptwrite.exp          | 209 +++++++
 gdb/testsuite/gdb.btrace/x86_64-ptwrite.S     | 544 +++++++++++++++++
 gdb/testsuite/gdb.python/py-record-btrace.exp |   6 +-
 gdb/testsuite/lib/gdb.exp                     |  74 +++
 gdbserver/linux-low.cc                        |   2 +
 gdbserver/server.cc                           |  18 +
 gdbsupport/btrace-common.h                    |   6 +
 gdbsupport/common.m4                          |   2 +
 gdbsupport/config.in                          |   3 +
 gdbsupport/configure                          |  11 +
 38 files changed, 2286 insertions(+), 35 deletions(-)
 create mode 100644 gdb/python/lib/gdb/ptwrite.py
 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

-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 01/10] btrace: Introduce auxiliary instructions.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:46   ` Eli Zaretskii
  2022-09-16 11:36 ` [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history Felix Willgerodt
                   ` (9 subsequent siblings)
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Auxiliary instructions are pseudo instructions pointing to auxiliary data.
This auxiliary data can be printed in all commands displaying (record
function-call-history, record instruction-history) or stepping through
(stepi etc.) the execution history, which will be introduced in the next
commits.

This patch is in preparation for the new ptwrite feature, which is based on
auxiliary instructions.
---
 gdb/btrace.c        |  2 ++
 gdb/btrace.h        | 24 +++++++++++++++++++++---
 gdb/doc/gdb.texinfo |  4 ++++
 3 files changed, 27 insertions(+), 3 deletions(-)

diff --git a/gdb/btrace.c b/gdb/btrace.c
index c4f0d4911ab..3305ebfb58f 100644
--- a/gdb/btrace.c
+++ b/gdb/btrace.c
@@ -1823,6 +1823,8 @@ btrace_clear_history (struct btrace_thread_info *btinfo)
   btinfo->insn_history = NULL;
   btinfo->call_history = NULL;
   btinfo->replay = NULL;
+
+  btinfo->aux_data.clear ();
 }
 
 /* Clear the branch trace maintenance histories in BTINFO.  */
diff --git a/gdb/btrace.h b/gdb/btrace.h
index 0fcf669597a..be4ce424ed5 100644
--- a/gdb/btrace.h
+++ b/gdb/btrace.h
@@ -52,7 +52,10 @@ enum btrace_insn_class
   BTRACE_INSN_RETURN,
 
   /* The instruction is an unconditional jump.  */
-  BTRACE_INSN_JUMP
+  BTRACE_INSN_JUMP,
+
+  /* The instruction is a pseudo instruction containing auxiliary data.  */
+  BTRACE_INSN_AUX
 };
 
 /* Instruction flags.  */
@@ -68,8 +71,19 @@ DEF_ENUM_FLAGS_TYPE (enum btrace_insn_flag, btrace_insn_flags);
    This represents a single instruction in a branch trace.  */
 struct btrace_insn
 {
-  /* The address of this instruction.  */
-  CORE_ADDR pc;
+  union
+  {
+    /* The address of this instruction.  Applies to btrace_insn with
+       iclass == BTRACE_INSN_OTHER or
+       iclass == BTRACE_INSN_CALL or
+       iclass == BTRACE_INSN_RETURN or
+       iclass == BTRACE_INSN_JUMP.  */
+    CORE_ADDR pc;
+
+    /* Index into btrace_info::aux_data.  Applies to btrace_insn with
+       iclass == BTRACE_INSN_AUX.  */
+    uint64_t aux_data_index;
+  };
 
   /* The size of this instruction in bytes.  */
   gdb_byte size;
@@ -330,6 +344,10 @@ struct btrace_thread_info
      function segment i will be at index (i - 1).  */
   std::vector<btrace_function> functions;
 
+  /* Optional auxiliary information that is printed in all commands
+     displaying or stepping through the execution history.  */
+  std::vector<std::string> aux_data;
+
   /* The function level offset.  When added to each function's LEVEL,
      this normalizes the function levels such that the smallest level
      becomes zero.  */
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 238a49b027d..0f769f5e24c 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86 (i386/amd64) running
 GNU/Linux.  Process record and replay can be used both when native
 debugging, and when remote debugging via @code{gdbserver}.
 
+When recording an inferior, @value{GDBN} may print additional auxiliary
+information during stepping commands and commands displaying the execution
+history.
+
 For architecture environments that support process record and replay,
 @value{GDBN} provides the following commands:
 
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:50   ` Eli Zaretskii
  2022-09-16 11:36 ` [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history Felix Willgerodt
                   ` (8 subsequent siblings)
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Print the auxiliary data when a btrace_insn of type BTRACE_INSN_AUX
is encountered in the instruction-history.  Printing is active by default,
it can be silenced with the /a modifier.

This patch is in preparation for the new ptwrite feature, which is based on
auxiliary instructions.
---
 gdb/disasm-flags.h  |  1 +
 gdb/doc/gdb.texinfo |  3 +++
 gdb/record-btrace.c | 14 ++++++++++++++
 gdb/record.c        |  5 +++++
 4 files changed, 23 insertions(+)

diff --git a/gdb/disasm-flags.h b/gdb/disasm-flags.h
index 025b6893941..4c920971d99 100644
--- a/gdb/disasm-flags.h
+++ b/gdb/disasm-flags.h
@@ -33,6 +33,7 @@ enum gdb_disassembly_flag
     DISASSEMBLY_OMIT_PC = (0x1 << 4),
     DISASSEMBLY_SOURCE = (0x1 << 5),
     DISASSEMBLY_SPECULATIVE = (0x1 << 6),
+    DISASSEMBLY_OMIT_AUX_INSN = (0x1 << 7),
   };
 DEF_ENUM_FLAGS_TYPE (enum gdb_disassembly_flag, gdb_disassembly_flags);
 
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 0f769f5e24c..42cb574d25e 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -7961,6 +7961,9 @@ To better align the printed instructions when the trace contains
 instructions from more than one function, the function name may be
 omitted by specifying the @code{/f} modifier.
 
+Printing auxiliary information is enabled by default and can be
+omitted with the @code{/a} modifier.
+
 Speculatively executed instructions are prefixed with @samp{?}.  This
 feature is not available for all recording formats.
 
diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
index 3a56925f63e..43e9dfa1008 100644
--- a/gdb/record-btrace.c
+++ b/gdb/record-btrace.c
@@ -821,6 +821,20 @@ btrace_insn_history (struct ui_out *uiout,
 	  btrace_ui_out_decode_error (uiout, btrace_insn_get_error (&it),
 				      conf->format);
 	}
+      else if (insn->iclass == BTRACE_INSN_AUX)
+	{
+	  if ((flags & DISASSEMBLY_OMIT_AUX_INSN) != 0)
+	    continue;
+
+	  uiout->field_fmt ("insn-number", "%u", btrace_insn_number (&it));
+	  uiout->text ("\t");
+	  uiout->spaces (3);
+	  uiout->text ("[");
+	  uiout->field_fmt ("aux-data", "%s",
+			    it.btinfo->aux_data.at
+			     (insn->aux_data_index).c_str ());
+	  uiout->text ("]\n");
+	}
       else
 	{
 	  struct disasm_insn dinsn;
diff --git a/gdb/record.c b/gdb/record.c
index 17a5df262bd..94b8c7b7ed3 100644
--- a/gdb/record.c
+++ b/gdb/record.c
@@ -486,6 +486,9 @@ get_insn_history_modifiers (const char **arg)
 
 	  switch (*args)
 	    {
+	    case 'a':
+	      modifiers |= DISASSEMBLY_OMIT_AUX_INSN;
+	      break;
 	    case 'm':
 	    case 's':
 	      modifiers |= DISASSEMBLY_SOURCE;
@@ -853,6 +856,8 @@ With a /m or /s modifier, source lines are included (if available).\n\
 With a /r modifier, raw instructions in hex are included.\n\
 With a /f modifier, function names are omitted.\n\
 With a /p modifier, current position markers are omitted.\n\
+With a /a modifier, omits output of auxiliary data, which is enabled \
+by default.\n\
 With no argument, disassembles ten more instructions after the previous \
 disassembly.\n\
 \"record instruction-history -\" disassembles ten instructions before a \
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:43   ` Eli Zaretskii
  2022-09-16 11:36 ` [PATCH v6 04/10] btrace: Handle stepping and goto for auxiliary instructions Felix Willgerodt
                   ` (7 subsequent siblings)
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Print the auxiliary data when a btrace_insn of type BTRACE_INSN_AUX
is encountered in the function-call-history.  Printing is
active by default, it can be silenced with the /a modifier.

This patch is in preparation for the new ptwrite feature, which is based on
auxiliary instructions.
---
 gdb/btrace.h        |  6 +++++-
 gdb/doc/gdb.texinfo |  5 +++--
 gdb/record-btrace.c | 21 +++++++++++++++++++++
 gdb/record.c        |  5 +++++
 gdb/record.h        |  5 ++++-
 5 files changed, 38 insertions(+), 4 deletions(-)

diff --git a/gdb/btrace.h b/gdb/btrace.h
index be4ce424ed5..36870d9b2ec 100644
--- a/gdb/btrace.h
+++ b/gdb/btrace.h
@@ -105,7 +105,11 @@ enum btrace_function_flag
 
   /* The 'up' link points to a tail call.  This obviously only makes sense
      if bfun_up_links_to_ret is clear.  */
-  BFUN_UP_LINKS_TO_TAILCALL = (1 << 1)
+  BFUN_UP_LINKS_TO_TAILCALL = (1 << 1),
+
+  /* Indicates that at least one auxiliary instruction is in the current
+     function segment.  */
+  BFUN_CONTAINS_AUX = (1 << 2)
 };
 DEF_ENUM_FLAGS_TYPE (enum btrace_function_flag, btrace_function_flags);
 
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 42cb574d25e..d5c4718a8f9 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -8017,8 +8017,9 @@ that function, the source lines for this instruction sequence (if the
 @code{/l} modifier is specified), and the instructions numbers that form
 the sequence (if the @code{/i} modifier is specified).  The function names
 are indented to reflect the call stack depth if the @code{/c} modifier is
-specified.  The @code{/l}, @code{/i}, and @code{/c} modifiers can be given
-together.
+specified.  Printing auxiliary information is enabled by default and can be
+omitted with the @code{/a} modifier.  The @code{/l}, @code{/i}, @code{/a},
+and @code{/c} modifiers can be given together.
 
 @smallexample
 (@value{GDBP}) @b{list 1, 10}
diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
index 43e9dfa1008..4d738d2ac67 100644
--- a/gdb/record-btrace.c
+++ b/gdb/record-btrace.c
@@ -1153,6 +1153,23 @@ btrace_get_bfun_name (const struct btrace_function *bfun)
     return "??";
 }
 
+static void
+btrace_print_aux_insn (struct ui_out *uiout,
+		       const struct btrace_function *bfun,
+		       const struct btrace_thread_info *btinfo)
+{
+  for (const btrace_insn &insn : bfun->insn)
+    {
+      if (insn.iclass == BTRACE_INSN_AUX)
+	{
+	  uiout->text ("\t\t[");
+	  uiout->field_fmt ("aux-data", "%s",
+			    btinfo->aux_data.at (insn.aux_data_index).c_str ());
+	  uiout->text ("]\n");
+	}
+    }
+}
+
 /* Disassemble a section of the recorded function trace.  */
 
 static void
@@ -1228,6 +1245,10 @@ btrace_call_history (struct ui_out *uiout,
 	}
 
       uiout->text ("\n");
+
+      if (((flags & RECORD_DONT_PRINT_AUX) == 0)
+	  && ((bfun->flags & BFUN_CONTAINS_AUX) != 0))
+	btrace_print_aux_insn(uiout, bfun, btinfo);
     }
 }
 
diff --git a/gdb/record.c b/gdb/record.c
index 94b8c7b7ed3..956475b57c7 100644
--- a/gdb/record.c
+++ b/gdb/record.c
@@ -636,6 +636,9 @@ get_call_history_modifiers (const char **arg)
 	    case 'c':
 	      modifiers |= RECORD_PRINT_INDENT_CALLS;
 	      break;
+	    case 'a':
+	      modifiers |= RECORD_DONT_PRINT_AUX;
+	      break;
 	    default:
 	      error (_("Invalid modifier: %c."), *args);
 	    }
@@ -880,6 +883,8 @@ Without modifiers, it prints the function name.\n\
 With a /l modifier, the source file and line number range is included.\n\
 With a /i modifier, the instruction number range is included.\n\
 With a /c modifier, the output is indented based on the call stack depth.\n\
+With a /a modifier, omits output of auxiliary data, which is enabled \
+by default.\n\
 With no argument, prints ten more lines after the previous ten-line print.\n\
 \"record function-call-history -\" prints ten lines before a previous ten-line \
 print.\n\
diff --git a/gdb/record.h b/gdb/record.h
index 0fbca9d44ce..d882f313c1d 100644
--- a/gdb/record.h
+++ b/gdb/record.h
@@ -62,7 +62,10 @@ enum record_print_flag
   RECORD_PRINT_INSN_RANGE = (1 << 1),
 
   /* Indent based on call stack depth (if applicable).  */
-  RECORD_PRINT_INDENT_CALLS = (1 << 2)
+  RECORD_PRINT_INDENT_CALLS = (1 << 2),
+
+  /* Deactivate printing auxiliary data (if applicable).  */
+  RECORD_DONT_PRINT_AUX = (1 << 3)
 };
 DEF_ENUM_FLAGS_TYPE (enum record_print_flag, record_print_flags);
 
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 04/10] btrace: Handle stepping and goto for auxiliary instructions.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (2 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class Felix Willgerodt
                   ` (6 subsequent siblings)
  10 siblings, 0 replies; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Print the auxiliary data when stepping. Don't allow to goto an auxiliary
instruction.

This patch is in preparation for the new ptwrite feature, which is based on
auxiliary instructions.
---
 gdb/record-btrace.c | 67 ++++++++++++++++++++++++++++++++++++---------
 1 file changed, 54 insertions(+), 13 deletions(-)

diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
index 4d738d2ac67..4e1ab60691f 100644
--- a/gdb/record-btrace.c
+++ b/gdb/record-btrace.c
@@ -2375,9 +2375,13 @@ record_btrace_single_step_forward (struct thread_info *tp)
     return btrace_step_stopped ();
 
   /* Skip gaps during replay.  If we end up at a gap (at the end of the trace),
-     jump back to the instruction at which we started.  */
+     jump back to the instruction at which we started.  If we're stepping a
+     BTRACE_INSN_AUX instruction, print the auxiliary data and skip the
+     instruction.  */
+
   start = *replay;
-  do
+
+  for (;;)
     {
       unsigned int steps;
 
@@ -2385,12 +2389,27 @@ record_btrace_single_step_forward (struct thread_info *tp)
 	 of the execution history.  */
       steps = btrace_insn_next (replay, 1);
       if (steps == 0)
+        {
+          *replay = start;
+          return btrace_step_no_history ();
+        }
+
+      const struct btrace_insn *insn = btrace_insn_get (replay);
+      if (insn == nullptr)
+	continue;
+
+      /* If we're stepping a BTRACE_INSN_AUX instruction, print the auxiliary
+	 data and skip the instruction.  */
+      if (insn->iclass == BTRACE_INSN_AUX)
 	{
-	  *replay = start;
-	  return btrace_step_no_history ();
+	  gdb_printf ("[%s]\n",
+		      btinfo->aux_data.at (insn->aux_data_index).c_str ());
+	  continue;
 	}
+
+      /* We have an instruction, we are done.  */
+      break;
     }
-  while (btrace_insn_get (replay) == NULL);
 
   /* Determine the end of the instruction trace.  */
   btrace_insn_end (&end, btinfo);
@@ -2421,9 +2440,12 @@ record_btrace_single_step_backward (struct thread_info *tp)
 
   /* If we can't step any further, we reached the end of the history.
      Skip gaps during replay.  If we end up at a gap (at the beginning of
-     the trace), jump back to the instruction at which we started.  */
+     the trace), jump back to the instruction at which we started.
+     If we're stepping a BTRACE_INSN_AUX instruction, print the auxiliary
+     data and skip the instruction.  */
   start = *replay;
-  do
+
+  for (;;)
     {
       unsigned int steps;
 
@@ -2433,8 +2455,22 @@ record_btrace_single_step_backward (struct thread_info *tp)
 	  *replay = start;
 	  return btrace_step_no_history ();
 	}
+
+      const struct btrace_insn *insn = btrace_insn_get (replay);
+      if (insn == nullptr)
+	continue;
+
+      /* Check if we're stepping a BTRACE_INSN_AUX instruction and skip it.  */
+      if (insn->iclass == BTRACE_INSN_AUX)
+	{
+	  gdb_printf ("[%s]\n",
+		      btinfo->aux_data.at (insn->aux_data_index).c_str ());
+	  continue;
+	}
+
+      /* We have an instruction, we are done.  */
+      break;
     }
-  while (btrace_insn_get (replay) == NULL);
 
   /* Check if we're stepping a breakpoint.
 
@@ -2856,25 +2892,30 @@ record_btrace_target::goto_record_end ()
 /* The goto_record method of target record-btrace.  */
 
 void
-record_btrace_target::goto_record (ULONGEST insn)
+record_btrace_target::goto_record (ULONGEST insn_number)
 {
   struct thread_info *tp;
   struct btrace_insn_iterator it;
   unsigned int number;
   int found;
 
-  number = insn;
+  number = insn_number;
 
   /* Check for wrap-arounds.  */
-  if (number != insn)
+  if (number != insn_number)
     error (_("Instruction number out of range."));
 
   tp = require_btrace_thread ();
 
   found = btrace_find_insn_by_number (&it, &tp->btrace, number);
 
-  /* Check if the instruction could not be found or is a gap.  */
-  if (found == 0 || btrace_insn_get (&it) == NULL)
+  /* Check if the instruction could not be found or is a gap or an
+     auxilliary instruction.  */
+  if (found == 0)
+    error (_("No such instruction."));
+
+  const struct btrace_insn *insn = btrace_insn_get (&it);
+  if ((insn == NULL) || (insn->iclass == BTRACE_INSN_AUX))
     error (_("No such instruction."));
 
   record_btrace_set_replay (tp, &it);
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (3 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 04/10] btrace: Handle stepping and goto for auxiliary instructions Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 12:04   ` Eli Zaretskii
  2022-09-16 11:36 ` [PATCH v6 06/10] python: Add clear() to gdb.Record Felix Willgerodt
                   ` (5 subsequent siblings)
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Auxiliary instructions are no real instructions and get their own object
class, similar to gaps. gdb.Record.instruction_history is now possibly a
list of gdb.RecordInstruction, gdb.RecordGap or gdb.RecordAuxiliary
objects.

This patch is in preparation for the new ptwrite feature, which is based on
auxiliary instructions.
---
 gdb/doc/python.texi           | 13 +++++++
 gdb/python/py-record-btrace.c | 35 +++++++++++------
 gdb/python/py-record.c        | 73 ++++++++++++++++++++++++++++++++++-
 gdb/python/py-record.h        |  3 ++
 4 files changed, 111 insertions(+), 13 deletions(-)

diff --git a/gdb/doc/python.texi b/gdb/doc/python.texi
index 7aa9e853d85..a8b890d5a54 100644
--- a/gdb/doc/python.texi
+++ b/gdb/doc/python.texi
@@ -3892,6 +3892,19 @@ the current recording method.
 A human readable string with the reason for the gap.
 @end defvar
 
+Some @value{GDBN} features write auxiliary information into the execution
+history.  This information is represented by a @code{gdb.RecordAuxiliary} object
+in the instruction list.  It has the following attributes:
+
+@defvar RecordAuxiliary.number
+An integer identifying this auxiliary.  @code{number} corresponds to the numbers
+seen in @code{record instruction-history} (@pxref{Process Record and Replay}).
+@end defvar
+
+@defvar RecordAuxiliary.data
+A string representation of the auxiliary data.
+@end defvar
+
 A @code{gdb.RecordFunctionSegment} object has the following attributes:
 
 @defvar RecordFunctionSegment.number
diff --git a/gdb/python/py-record-btrace.c b/gdb/python/py-record-btrace.c
index 85401010f0a..1dab097335b 100644
--- a/gdb/python/py-record-btrace.c
+++ b/gdb/python/py-record-btrace.c
@@ -45,7 +45,8 @@ struct btpy_list_object {
   /* Stride size.  */
   Py_ssize_t step;
 
-  /* Either &BTPY_CALL_TYPE or &RECPY_INSN_TYPE.  */
+  /* Either &recpy_func_type, &recpy_insn_type, &recpy_aux_type or
+     &recpy_gap_type.  */
   PyTypeObject* element_type;
 };
 
@@ -141,10 +142,11 @@ btrace_func_from_recpy_func (const PyObject * const pyobject)
 }
 
 /* Looks at the recorded item with the number NUMBER and create a
-   gdb.RecordInstruction or gdb.RecordGap object for it accordingly.  */
+   gdb.RecordInstruction, gdb.RecordGap or gdb.RecordAuxiliary object
+   for it accordingly.  */
 
 static PyObject *
-btpy_insn_or_gap_new (thread_info *tinfo, Py_ssize_t number)
+btpy_item_new (thread_info *tinfo, Py_ssize_t number)
 {
   btrace_insn_iterator iter;
   int err_code;
@@ -163,6 +165,14 @@ btpy_insn_or_gap_new (thread_info *tinfo, Py_ssize_t number)
       return recpy_gap_new (err_code, err_string, number);
     }
 
+  const struct btrace_insn *insn = btrace_insn_get (&iter);
+  gdb_assert (insn != nullptr);
+
+  if (insn->iclass == BTRACE_INSN_AUX)
+    return recpy_aux_new
+      (iter.btinfo->aux_data.at (insn->aux_data_index).c_str (), number);
+
+
   return recpy_insn_new (tinfo, RECORD_METHOD_BTRACE, number);
 }
 
@@ -441,8 +451,10 @@ btpy_list_length (PyObject *self)
 }
 
 /* Implementation of
-   BtraceList.__getitem__ (self, key) -> BtraceInstruction and
-   BtraceList.__getitem__ (self, key) -> BtraceFunctionCall.  */
+   BtraceList.__getitem__ (self, key) -> BtraceInstruction,
+   BtraceList.__getitem__ (self, key) -> BtraceFunctionCall,
+   BtraceList.__getitem__ (self, key) -> BtraceAuxilliary and
+   BtraceList.__getitem__ (self, key) -> BtraceGap.  */
 
 static PyObject *
 btpy_list_item (PyObject *self, Py_ssize_t index)
@@ -456,10 +468,10 @@ btpy_list_item (PyObject *self, Py_ssize_t index)
 
   number = obj->first + (obj->step * index);
 
-  if (obj->element_type == &recpy_insn_type)
-    return recpy_insn_new (obj->thread, RECORD_METHOD_BTRACE, number);
-  else
+  if (obj->element_type == &recpy_func_type)
     return recpy_func_new (obj->thread, RECORD_METHOD_BTRACE, number);
+  else
+    return btpy_item_new (obj->thread, number);
 }
 
 /* Implementation of BtraceList.__getitem__ (self, slice) -> BtraceList.  */
@@ -646,8 +658,7 @@ recpy_bt_replay_position (PyObject *self, void *closure)
   if (tinfo->btrace.replay == NULL)
     Py_RETURN_NONE;
 
-  return btpy_insn_or_gap_new (tinfo,
-			       btrace_insn_number (tinfo->btrace.replay));
+  return btpy_item_new (tinfo, btrace_insn_number (tinfo->btrace.replay));
 }
 
 /* Implementation of
@@ -669,7 +680,7 @@ recpy_bt_begin (PyObject *self, void *closure)
     Py_RETURN_NONE;
 
   btrace_insn_begin (&iterator, &tinfo->btrace);
-  return btpy_insn_or_gap_new (tinfo, btrace_insn_number (&iterator));
+  return btpy_item_new (tinfo, btrace_insn_number (&iterator));
 }
 
 /* Implementation of
@@ -691,7 +702,7 @@ recpy_bt_end (PyObject *self, void *closure)
     Py_RETURN_NONE;
 
   btrace_insn_end (&iterator, &tinfo->btrace);
-  return btpy_insn_or_gap_new (tinfo, btrace_insn_number (&iterator));
+  return btpy_item_new (tinfo, btrace_insn_number (&iterator));
 }
 
 /* Implementation of
diff --git a/gdb/python/py-record.c b/gdb/python/py-record.c
index 51084dfac72..47bbd290144 100644
--- a/gdb/python/py-record.c
+++ b/gdb/python/py-record.c
@@ -49,6 +49,12 @@ static PyTypeObject recpy_gap_type = {
   PyVarObject_HEAD_INIT (NULL, 0)
 };
 
+/* Python RecordAuxiliary type.  */
+
+PyTypeObject recpy_aux_type = {
+  PyVarObject_HEAD_INIT (nullptr, 0)
+};
+
 /* Python RecordGap object.  */
 struct recpy_gap_object
 {
@@ -64,6 +70,18 @@ struct recpy_gap_object
   Py_ssize_t number;
 };
 
+/* Python RecordAuxiliary object.  */
+typedef struct
+{
+  PyObject_HEAD
+
+  /* Auxiliary data.  */
+  const char *data;
+
+  /* Element number.  */
+  Py_ssize_t number;
+} recpy_aux_object;
+
 /* Implementation of record.method.  */
 
 static PyObject *
@@ -477,6 +495,43 @@ recpy_gap_reason_string (PyObject *self, void *closure)
   return PyUnicode_FromString (obj->reason_string);
 }
 
+/* Create a new gdb.Auxiliary object.  */
+
+PyObject *
+recpy_aux_new (const char *data, Py_ssize_t number)
+{
+  recpy_aux_object * const obj = PyObject_New (recpy_aux_object,
+					       &recpy_aux_type);
+
+  if (obj == nullptr)
+   return nullptr;
+
+  obj->data = data;
+  obj->number = number;
+
+  return (PyObject *) obj;
+}
+
+/* Implementation of Auxiliary.number [int].  */
+
+static PyObject *
+recpy_aux_number (PyObject *self, void *closure)
+{
+  const recpy_aux_object * const obj = (const recpy_aux_object *) self;
+
+  return gdb_py_object_from_longest (obj->number).release ();
+}
+
+/* Implementation of Auxiliary.data [str].  */
+
+static PyObject *
+recpy_aux_data (PyObject *self, void *closure)
+{
+  const recpy_aux_object * const obj = (const recpy_aux_object *) self;
+
+  return PyUnicode_FromString (obj->data);
+}
+
 /* Record method list.  */
 
 static PyMethodDef recpy_record_methods[] = {
@@ -542,6 +597,14 @@ static gdb_PyGetSetDef recpy_gap_getset[] = {
   { NULL }
 };
 
+/* RecordAuxiliary member list.  */
+
+static gdb_PyGetSetDef recpy_aux_getset[] = {
+  { "number", recpy_aux_number, nullptr, "element number", nullptr},
+  { "data", recpy_aux_data, nullptr, "data", nullptr},
+  { nullptr }
+};
+
 /* Sets up the record API in the gdb module.  */
 
 int
@@ -581,10 +644,18 @@ gdbpy_initialize_record (void)
   recpy_gap_type.tp_doc = "GDB recorded gap object";
   recpy_gap_type.tp_getset = recpy_gap_getset;
 
+  recpy_aux_type.tp_new = PyType_GenericNew;
+  recpy_aux_type.tp_flags = Py_TPFLAGS_DEFAULT;
+  recpy_aux_type.tp_basicsize = sizeof (recpy_aux_object);
+  recpy_aux_type.tp_name = "gdb.RecordAuxiliary";
+  recpy_aux_type.tp_doc = "GDB recorded auxiliary object";
+  recpy_aux_type.tp_getset = recpy_aux_getset;
+
   if (PyType_Ready (&recpy_record_type) < 0
       || PyType_Ready (&recpy_insn_type) < 0
       || PyType_Ready (&recpy_func_type) < 0
-      || PyType_Ready (&recpy_gap_type) < 0)
+      || PyType_Ready (&recpy_gap_type) < 0
+      || PyType_Ready (&recpy_aux_type) < 0)
     return -1;
   else
     return 0;
diff --git a/gdb/python/py-record.h b/gdb/python/py-record.h
index 2c8cfd6e4a5..d1d784e758e 100644
--- a/gdb/python/py-record.h
+++ b/gdb/python/py-record.h
@@ -71,4 +71,7 @@ extern PyObject *recpy_func_new (thread_info *thread, enum record_method method,
 extern PyObject *recpy_gap_new (int reason_code, const char *reason_string,
 				Py_ssize_t number);
 
+/* Create a new gdb.RecordGap object.  */
+extern PyObject *recpy_aux_new (const char *data, Py_ssize_t number);
+
 #endif /* PYTHON_PY_RECORD_H */
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 06/10] python: Add clear() to gdb.Record.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (4 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:49   ` Eli Zaretskii
  2022-09-16 11:36 ` [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
                   ` (4 subsequent siblings)
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

This function allows to clear the trace data from python, forcing to
re-decode the trace for successive commands.
---
 gdb/doc/python.texi                           |  5 +++++
 gdb/python/py-record-btrace.c                 | 13 +++++++++++++
 gdb/python/py-record-btrace.h                 |  3 +++
 gdb/python/py-record.c                        | 16 ++++++++++++++++
 gdb/testsuite/gdb.python/py-record-btrace.exp |  6 +++++-
 5 files changed, 42 insertions(+), 1 deletion(-)

diff --git a/gdb/doc/python.texi b/gdb/doc/python.texi
index a8b890d5a54..b963202b7b9 100644
--- a/gdb/doc/python.texi
+++ b/gdb/doc/python.texi
@@ -3836,6 +3836,11 @@ A @code{gdb.Record} object has the following methods:
 Move the replay position to the given @var{instruction}.
 @end defun
 
+@defun Record.clear ()
+Clear the trace data of the current recording.  This forces re-decoding of the
+trace for successive commands.
+@end defun
+
 The common @code{gdb.Instruction} class that recording method specific
 instruction objects inherit from, has the following attributes:
 
diff --git a/gdb/python/py-record-btrace.c b/gdb/python/py-record-btrace.c
index 1dab097335b..c09ba6271b5 100644
--- a/gdb/python/py-record-btrace.c
+++ b/gdb/python/py-record-btrace.c
@@ -802,6 +802,19 @@ recpy_bt_goto (PyObject *self, PyObject *args)
   Py_RETURN_NONE;
 }
 
+/* Implementation of BtraceRecord.clear (self) -> None.  */
+
+PyObject *
+recpy_bt_clear (PyObject *self, PyObject *args)
+{
+  const recpy_record_object * const record = (recpy_record_object *) self;
+  thread_info *const tinfo = record->thread;
+
+  btrace_clear (tinfo);
+
+  Py_RETURN_NONE;
+}
+
 /* BtraceList methods.  */
 
 static PyMethodDef btpy_list_methods[] =
diff --git a/gdb/python/py-record-btrace.h b/gdb/python/py-record-btrace.h
index 8ef26c860db..b25b121d255 100644
--- a/gdb/python/py-record-btrace.h
+++ b/gdb/python/py-record-btrace.h
@@ -31,6 +31,9 @@ extern PyObject *recpy_bt_format (PyObject *self, void *closure);
 /* Implementation of record.goto (instruction) -> None.  */
 extern PyObject *recpy_bt_goto (PyObject *self, PyObject *value);
 
+/* Implementation of BtraceRecord.clear (self) -> None.  */
+extern PyObject *recpy_bt_clear (PyObject *self, PyObject *args);
+
 /* Implementation of record.instruction_history [list].  */
 extern PyObject *recpy_bt_instruction_history (PyObject *self, void *closure);
 
diff --git a/gdb/python/py-record.c b/gdb/python/py-record.c
index 47bbd290144..0aef0e0ae1b 100644
--- a/gdb/python/py-record.c
+++ b/gdb/python/py-record.c
@@ -127,6 +127,19 @@ recpy_goto (PyObject *self, PyObject *value)
   return PyErr_Format (PyExc_NotImplementedError, _("Not implemented."));
 }
 
+/* Implementation of record.clear () -> None.  */
+
+static PyObject *
+recpy_clear (PyObject *self, PyObject *value)
+{
+  const recpy_record_object * const obj = (recpy_record_object *) self;
+
+  if (obj->method == RECORD_METHOD_BTRACE)
+    return recpy_bt_clear (self, value);
+
+  return PyErr_Format (PyExc_NotImplementedError, _("Not implemented."));
+}
+
 /* Implementation of record.replay_position [instruction]  */
 
 static PyObject *
@@ -538,6 +551,9 @@ static PyMethodDef recpy_record_methods[] = {
   { "goto", recpy_goto, METH_VARARGS,
     "goto (instruction|function_call) -> None.\n\
 Rewind to given location."},
+  { "clear", recpy_clear, METH_VARARGS,
+    "clear () -> None.\n\
+Clears the trace."},
   { NULL }
 };
 
diff --git a/gdb/testsuite/gdb.python/py-record-btrace.exp b/gdb/testsuite/gdb.python/py-record-btrace.exp
index e9fdd06154e..82d887dbd18 100644
--- a/gdb/testsuite/gdb.python/py-record-btrace.exp
+++ b/gdb/testsuite/gdb.python/py-record-btrace.exp
@@ -88,7 +88,11 @@ with_test_prefix "instruction " {
     gdb_test "python print(repr(i.data))" "<memory at $hex>"
     gdb_test "python print(i.decoded)" ".*"
     gdb_test "python print(i.size)" "$decimal"
-    gdb_test "python print(i.is_speculative)" "False"
+    gdb_test "python print(i.is_speculative)" "False"    
+    gdb_test_no_output "python r.clear()"
+    gdb_test "python insn = r.instruction_history"
+    gdb_test_no_output "python i = insn\[0\]"
+    gdb_test "python print(i.size)" "$decimal"
 }
 
 with_test_prefix "function call" {
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (5 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 06/10] python: Add clear() to gdb.Record Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:49   ` Eli Zaretskii
  2022-10-19 14:41   ` Metzger, Markus T
  2022-09-16 11:36 ` [PATCH v6 08/10] btrace, linux: Enable ptwrite packets Felix Willgerodt
                   ` (3 subsequent siblings)
  10 siblings, 2 replies; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

This enables gdb and gdbserver to communicate about ptwrite support.  If
ptwrite support would be enabled unconditionally, GDBs with older libipt
versions would break.
---
 gdb/btrace.c                 |  8 +++++++-
 gdb/doc/gdb.texinfo          | 21 +++++++++++++++++++++
 gdb/features/btrace-conf.dtd |  1 +
 gdb/remote.c                 | 30 ++++++++++++++++++++++++++++++
 gdbserver/linux-low.cc       |  2 ++
 gdbserver/server.cc          | 18 ++++++++++++++++++
 gdbsupport/btrace-common.h   |  6 ++++++
 7 files changed, 85 insertions(+), 1 deletion(-)

diff --git a/gdb/btrace.c b/gdb/btrace.c
index 3305ebfb58f..b5ccb2c8264 100644
--- a/gdb/btrace.c
+++ b/gdb/btrace.c
@@ -2273,7 +2273,7 @@ parse_xml_btrace_conf_pt (struct gdb_xml_parser *parser,
 			  std::vector<gdb_xml_value> &attributes)
 {
   struct btrace_config *conf;
-  struct gdb_xml_value *size;
+  struct gdb_xml_value *size, *ptwrite;
 
   conf = (struct btrace_config *) user_data;
   conf->format = BTRACE_FORMAT_PT;
@@ -2282,10 +2282,16 @@ parse_xml_btrace_conf_pt (struct gdb_xml_parser *parser,
   size = xml_find_attribute (attributes, "size");
   if (size != NULL)
     conf->pt.size = (unsigned int) *(ULONGEST *) size->value.get ();
+
+  ptwrite = xml_find_attribute (attributes, "ptwrite");
+  if (ptwrite != nullptr)
+    conf->pt.ptwrite = (bool) *(ULONGEST *) ptwrite->value.get ();
 }
 
 static const struct gdb_xml_attribute btrace_conf_pt_attributes[] = {
   { "size", GDB_XML_AF_OPTIONAL, gdb_xml_parse_attr_ulongest, NULL },
+  { "ptwrite", GDB_XML_AF_OPTIONAL, gdb_xml_parse_attr_enum,
+    gdb_xml_enums_boolean },
   { NULL, GDB_XML_AF_NONE, NULL, NULL }
 };
 
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index d5c4718a8f9..cbc2d301c92 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -43012,6 +43012,11 @@ These are the currently defined stub features and their properties:
 @tab @samp{-}
 @tab Yes
 
+@item @samp{Qbtrace-conf:pt:ptwrite}
+@tab Yes
+@tab @samp{-}
+@tab Yes
+
 @item @samp{QNonStop}
 @tab No
 @tab @samp{-}
@@ -43323,6 +43328,9 @@ The remote stub understands the @samp{Qbtrace-conf:bts:size} packet.
 @item Qbtrace-conf:pt:size
 The remote stub understands the @samp{Qbtrace-conf:pt:size} packet.
 
+@item Qbtrace-conf:pt:ptwrite
+The remote stub understands the @samp{Qbtrace-conf:pt:ptwrite} packet.
+
 @item swbreak
 The remote stub reports the @samp{swbreak} stop reason for memory
 breakpoints.
@@ -43824,6 +43832,18 @@ The ring buffer size has been set.
 A badly formed request or an error was encountered.
 @end table
 
+@item Qbtrace-conf:pt:ptwrite=@var{(yes|no)}
+Indicate support for ptwrite packets.  This allows for backwards-
+compatibility.
+
+Reply:
+@table @samp
+@item OK
+The ptwrite config parameter has been set.
+@item E.errtext
+A badly formed request or an error was encountered.
+@end table
+
 @end table
 
 @node Architecture-Specific Protocol Details
@@ -46466,6 +46486,7 @@ The formal DTD for the branch trace configuration format is given below:
 
 <!ELEMENT pt	EMPTY>
 <!ATTLIST pt	size	CDATA	#IMPLIED>
+<!ATTLIST pt	ptwrite	(yes | no)	#IMPLIED>
 @end smallexample
 
 @include agentexpr.texi
diff --git a/gdb/features/btrace-conf.dtd b/gdb/features/btrace-conf.dtd
index c71b11f2087..6571515c264 100644
--- a/gdb/features/btrace-conf.dtd
+++ b/gdb/features/btrace-conf.dtd
@@ -12,3 +12,4 @@
 
 <!ELEMENT pt	EMPTY>
 <!ATTLIST pt	size	CDATA	#IMPLIED>
+<!ATTLIST pt	ptwrite	(yes | no) #IMPLIED>
diff --git a/gdb/remote.c b/gdb/remote.c
index 70f918a7362..f8cca2330aa 100644
--- a/gdb/remote.c
+++ b/gdb/remote.c
@@ -2204,6 +2204,9 @@ enum {
   /* Support for the Qbtrace-conf:pt:size packet.  */
   PACKET_Qbtrace_conf_pt_size,
 
+  /* Support for the Qbtrace-conf:pt:ptwrite packet.  */
+  PACKET_Qbtrace_conf_pt_ptwrite,
+
   /* Support for exec events.  */
   PACKET_exec_event_feature,
 
@@ -5422,6 +5425,8 @@ static const struct protocol_feature remote_protocol_features[] = {
     PACKET_exec_event_feature },
   { "Qbtrace-conf:pt:size", PACKET_DISABLE, remote_supported_packet,
     PACKET_Qbtrace_conf_pt_size },
+  { "Qbtrace-conf:pt:ptwrite", PACKET_DISABLE, remote_supported_packet,
+    PACKET_Qbtrace_conf_pt_ptwrite },
   { "vContSupported", PACKET_DISABLE, remote_supported_packet, PACKET_vContSupported },
   { "QThreadEvents", PACKET_DISABLE, remote_supported_packet, PACKET_QThreadEvents },
   { "no-resumed", PACKET_DISABLE, remote_supported_packet, PACKET_no_resumed },
@@ -14106,6 +14111,28 @@ remote_target::btrace_sync_conf (const btrace_config *conf)
 
       rs->btrace_config.pt.size = conf->pt.size;
     }
+
+  packet = &remote_protocol_packets[PACKET_Qbtrace_conf_pt_ptwrite];
+  if (packet_config_support (packet) == PACKET_ENABLE
+      && conf->pt.ptwrite != rs->btrace_config.pt.ptwrite)
+    {
+      pos = buf;
+      const char *ptw = conf->pt.ptwrite ? "yes" : "no";
+      pos += xsnprintf (pos, endbuf - pos, "%s=\"%s\"", packet->name, ptw);
+
+      putpkt (buf);
+      getpkt (&rs->buf, 0);
+
+      if (packet_ok (buf, packet) == PACKET_ERROR)
+	{
+	  if (buf[0] == 'E' && buf[1] == '.')
+	    error (_("Failed to sync ptwrite config: %s"), buf + 2);
+	  else
+	    error (_("Failed to sync ptwrite config."));
+	}
+
+      rs->btrace_config.pt.ptwrite = conf->pt.ptwrite;
+    }
 }
 
 /* Read TP's btrace configuration from the target and store it into CONF.  */
@@ -15324,6 +15351,9 @@ Show the maximum size of the address (in bits) in a memory packet."), NULL,
   add_packet_config_cmd (&remote_protocol_packets[PACKET_Qbtrace_conf_pt_size],
        "Qbtrace-conf:pt:size", "btrace-conf-pt-size", 0);
 
+  add_packet_config_cmd (&remote_protocol_packets[PACKET_Qbtrace_conf_pt_ptwrite],
+       "Qbtrace-conf:pt:ptwrite", "btrace-conf-pt-ptwrite", 0);
+
   add_packet_config_cmd (&remote_protocol_packets[PACKET_vContSupported],
 			 "vContSupported", "verbose-resume-supported", 0);
 
diff --git a/gdbserver/linux-low.cc b/gdbserver/linux-low.cc
index 2f71360d3bd..1333ec13dbd 100644
--- a/gdbserver/linux-low.cc
+++ b/gdbserver/linux-low.cc
@@ -6806,6 +6806,8 @@ linux_process_target::read_btrace_conf (const btrace_target_info *tinfo,
 	case BTRACE_FORMAT_PT:
 	  buffer_xml_printf (buffer, "<pt");
 	  buffer_xml_printf (buffer, " size=\"0x%x\"", conf->pt.size);
+	  const char *ptw = conf->pt.ptwrite ? "yes" : "no";
+	  buffer_xml_printf (buffer, " ptwrite=\"%s\"", ptw);
 	  buffer_xml_printf (buffer, "/>\n");
 	  break;
 	}
diff --git a/gdbserver/server.cc b/gdbserver/server.cc
index c619206d5d2..0ea7b1a9d57 100644
--- a/gdbserver/server.cc
+++ b/gdbserver/server.cc
@@ -542,6 +542,23 @@ handle_btrace_conf_general_set (char *own_buf)
 
       current_btrace_conf.pt.size = (unsigned int) size;
     }
+  else if (strncmp (op, "pt:ptwrite=", strlen ("pt:ptwrite=")) == 0)
+    {
+      bool ptwrite;
+
+      op += strlen ("pt:ptwrite=");
+      if (strncmp (op, "\"yes\"", strlen ("\"yes\"")) == 0)
+	ptwrite = true;
+      else if (strncmp (op, "\"no\"", strlen ("\"no\"")) == 0)
+	ptwrite = false;
+      else
+	{
+	  strcpy (own_buf, "E.Bad ptwrite value.");
+	  return -1;
+	}
+
+      current_btrace_conf.pt.ptwrite = ptwrite;
+    }
   else
     {
       strcpy (own_buf, "E.Bad Qbtrace configuration option.");
@@ -2162,6 +2179,7 @@ supported_btrace_packets (char *buf)
   strcat (buf, ";Qbtrace-conf:bts:size+");
   strcat (buf, ";Qbtrace:pt+");
   strcat (buf, ";Qbtrace-conf:pt:size+");
+  strcat (buf, ";Qbtrace-conf:pt:ptwrite+");
   strcat (buf, ";Qbtrace:off+");
   strcat (buf, ";qXfer:btrace:read+");
   strcat (buf, ";qXfer:btrace-conf:read+");
diff --git a/gdbsupport/btrace-common.h b/gdbsupport/btrace-common.h
index 3c55ae92185..74708891923 100644
--- a/gdbsupport/btrace-common.h
+++ b/gdbsupport/btrace-common.h
@@ -117,6 +117,12 @@ struct btrace_config_pt
      This is unsigned int and not size_t since it is registered as
      control variable for "set record btrace pt buffer-size".  */
   unsigned int size;
+
+  /* Configuration bit for ptwrite packets.
+
+     If both gdb and gdbserver support this, gdb will try to enable ptwrite
+     packets when tracing is started.  */
+  bool ptwrite;
 };
 
 /* A branch tracing configuration.
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 08/10] btrace, linux: Enable ptwrite packets.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (6 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 11:36 ` [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration Felix Willgerodt
                   ` (2 subsequent siblings)
  10 siblings, 0 replies; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

Enable ptwrite in the PT config, if it is supported by the kernel.
---
 gdb/nat/linux-btrace.c | 29 +++++++++++++++++++++++++++++
 gdb/record-btrace.c    |  5 +++++
 2 files changed, 34 insertions(+)

diff --git a/gdb/nat/linux-btrace.c b/gdb/nat/linux-btrace.c
index c31fb5ffe43..166994e85e2 100644
--- a/gdb/nat/linux-btrace.c
+++ b/gdb/nat/linux-btrace.c
@@ -415,6 +415,29 @@ cpu_supports_bts (void)
     }
 }
 
+/* Check whether the linux target supports Intel Processor Trace PTWRITE.  */
+
+static bool
+linux_supports_ptwrite ()
+{
+  static const char filename[]
+      = "/sys/bus/event_source/devices/intel_pt/caps/ptwrite";
+  gdb_file_up file = gdb_fopen_cloexec (filename, "r");
+
+  if (file.get () == nullptr)
+    return false;
+
+  int status, found = fscanf (file.get (), "%d", &status);
+
+  if (found != 1)
+    {
+      warning (_("Failed to determine ptwrite support from %s."), filename);
+      return false;
+    }
+
+  return status == 1;
+}
+
 /* The perf_event_open syscall failed.  Try to print a helpful error
    message.  */
 
@@ -624,6 +647,12 @@ linux_enable_pt (ptid_t ptid, const struct btrace_config_pt *conf)
   pt->attr.exclude_hv = 1;
   pt->attr.exclude_idle = 1;
 
+  if (conf->ptwrite && linux_supports_ptwrite ())
+    {
+      pt->attr.config |= 0x1000;
+      tinfo->conf.pt.ptwrite = conf->ptwrite;
+    }
+
   errno = 0;
   scoped_fd fd (syscall (SYS_perf_event_open, &pt->attr, pid, -1, -1, 0));
   if (fd.get () < 0)
diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
index 4e1ab60691f..a6ddf8237e1 100644
--- a/gdb/record-btrace.c
+++ b/gdb/record-btrace.c
@@ -3288,4 +3288,9 @@ to see the actual buffer size."), NULL, show_record_pt_buffer_size_value,
 
   record_btrace_conf.bts.size = 64 * 1024;
   record_btrace_conf.pt.size = 16 * 1024;
+#if (LIBIPT_VERSION >= 0x200)
+  record_btrace_conf.pt.ptwrite = true;
+#else
+  record_btrace_conf.pt.ptwrite = false;
+#endif
 }
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (7 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 08/10] btrace, linux: Enable ptwrite packets Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-10-19 14:41   ` Metzger, Markus T
  2022-09-16 11:36 ` [PATCH v6 10/10] btrace: Extend ptwrite event decoding Felix Willgerodt
  2022-10-19 14:45 ` [PATCH v6 00/10] Extensions for PTWRITE Metzger, Markus T
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

With this patch a default ptwrite filter is registered upon start of GDB.
It prints the plain ptwrite payload as hex.  The default filter can be
overwritten by registering a custom filter in python or by registering
"None", for no output at all.  Registering a filter function creates per
thread copies to allow unique internal states per thread.
---
 gdb/btrace.c                   |  4 ++
 gdb/btrace.h                   |  8 ++++
 gdb/data-directory/Makefile.in |  1 +
 gdb/extension-priv.h           |  5 ++
 gdb/extension.c                | 13 +++++
 gdb/extension.h                |  3 ++
 gdb/guile/guile.c              |  1 +
 gdb/python/lib/gdb/ptwrite.py  | 80 +++++++++++++++++++++++++++++++
 gdb/python/py-record-btrace.c  | 88 ++++++++++++++++++++++++++++++++++
 gdb/python/py-record-btrace.h  |  8 ++++
 gdb/python/python-internal.h   |  3 ++
 gdb/python/python.c            |  2 +
 12 files changed, 216 insertions(+)
 create mode 100644 gdb/python/lib/gdb/ptwrite.py

diff --git a/gdb/btrace.c b/gdb/btrace.c
index b5ccb2c8264..14739d1d922 100644
--- a/gdb/btrace.c
+++ b/gdb/btrace.c
@@ -34,6 +34,7 @@
 #include "gdbsupport/rsp-low.h"
 #include "gdbcmd.h"
 #include "cli/cli-utils.h"
+#include "extension.h"
 #include "gdbarch.h"
 
 /* For maintenance commands.  */
@@ -1317,6 +1318,9 @@ ftrace_add_pt (struct btrace_thread_info *btinfo,
   uint64_t offset;
   int status;
 
+  /* Register the ptwrite filter.  */
+  apply_ext_lang_ptwrite_filter (btinfo);
+
   for (;;)
     {
       struct pt_insn insn;
diff --git a/gdb/btrace.h b/gdb/btrace.h
index 36870d9b2ec..f256988cf2c 100644
--- a/gdb/btrace.h
+++ b/gdb/btrace.h
@@ -352,6 +352,14 @@ struct btrace_thread_info
      displaying or stepping through the execution history.  */
   std::vector<std::string> aux_data;
 
+  /* Function pointer to the ptwrite callback.  Returns the string returned
+     by the ptwrite filter function.  */
+  std::string (*ptw_callback_fun) (const uint64_t payload, const uint64_t ip,
+				   const void *ptw_context);
+
+  /* Context for the ptw_callback_fun.  */
+  void *ptw_context = nullptr;
+
   /* The function level offset.  When added to each function's LEVEL,
      this normalizes the function levels such that the smallest level
      becomes zero.  */
diff --git a/gdb/data-directory/Makefile.in b/gdb/data-directory/Makefile.in
index cf5226f3961..506c16af4b5 100644
--- a/gdb/data-directory/Makefile.in
+++ b/gdb/data-directory/Makefile.in
@@ -75,6 +75,7 @@ PYTHON_FILE_LIST = \
 	gdb/frames.py \
 	gdb/printing.py \
 	gdb/prompt.py \
+	gdb/ptwrite.py \
 	gdb/styling.py \
 	gdb/types.py \
 	gdb/unwinder.py \
diff --git a/gdb/extension-priv.h b/gdb/extension-priv.h
index 7c74e721c57..b3bed3c88ab 100644
--- a/gdb/extension-priv.h
+++ b/gdb/extension-priv.h
@@ -183,6 +183,11 @@ struct extension_language_ops
      enum ext_lang_frame_args args_type,
      struct ui_out *out, int frame_low, int frame_high);
 
+  /* Used for registering the ptwrite filter to the current thread.  */
+  void (*apply_ptwrite_filter)
+    (const struct extension_language_defn *extlang,
+     struct btrace_thread_info *btinfo);
+
   /* Update values held by the extension language when OBJFILE is discarded.
      New global types must be created for every such value, which must then be
      updated to use the new types.
diff --git a/gdb/extension.c b/gdb/extension.c
index 5a805bea00e..0f6e3e02f0e 100644
--- a/gdb/extension.c
+++ b/gdb/extension.c
@@ -550,6 +550,19 @@ apply_ext_lang_frame_filter (struct frame_info *frame,
   return EXT_LANG_BT_NO_FILTERS;
 }
 
+/* Used for registering the ptwrite filter to the current thread.  */
+
+void
+apply_ext_lang_ptwrite_filter (btrace_thread_info *btinfo)
+{
+  for (const struct extension_language_defn *extlang : extension_languages)
+    {
+      if (extlang->ops != nullptr
+	  && extlang->ops->apply_ptwrite_filter != nullptr)
+	extlang->ops->apply_ptwrite_filter (extlang, btinfo);
+    }
+}
+
 /* Update values held by the extension language when OBJFILE is discarded.
    New global types must be created for every such value, which must then be
    updated to use the new types.
diff --git a/gdb/extension.h b/gdb/extension.h
index 47839ea50be..bd05d61cd50 100644
--- a/gdb/extension.h
+++ b/gdb/extension.h
@@ -295,6 +295,9 @@ extern enum ext_lang_bt_status apply_ext_lang_frame_filter
    enum ext_lang_frame_args args_type,
    struct ui_out *out, int frame_low, int frame_high);
 
+extern void apply_ext_lang_ptwrite_filter
+  (struct btrace_thread_info *btinfo);
+
 extern void preserve_ext_lang_values (struct objfile *, htab_t copied_types);
 
 extern const struct extension_language_defn *get_breakpoint_cond_ext_lang
diff --git a/gdb/guile/guile.c b/gdb/guile/guile.c
index e5565b627d9..0ddb1567d2d 100644
--- a/gdb/guile/guile.c
+++ b/gdb/guile/guile.c
@@ -124,6 +124,7 @@ static const struct extension_language_ops guile_extension_ops =
   gdbscm_apply_val_pretty_printer,
 
   NULL, /* gdbscm_apply_frame_filter, */
+  NULL, /* gdbscm_load_ptwrite_filter, */
 
   gdbscm_preserve_values,
 
diff --git a/gdb/python/lib/gdb/ptwrite.py b/gdb/python/lib/gdb/ptwrite.py
new file mode 100644
index 00000000000..0f5b0473023
--- /dev/null
+++ b/gdb/python/lib/gdb/ptwrite.py
@@ -0,0 +1,80 @@
+# Ptwrite utilities.
+# Copyright (C) 2022 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+"""Utilities for working with ptwrite filters."""
+
+from copy import deepcopy
+import gdb
+
+
+def default_filter(payload, ip):
+    """Default filter that is active upon starting GDB."""
+    return "{:x}".format(payload)
+
+
+# This dict contains the per thread copies of the filter function and the
+# global template filter, from which the copies are created.
+_ptwrite_filter = {"global": default_filter}
+
+
+def ptwrite_exit_handler(event):
+    """Exit handler to prune _ptwrite_filter on inferior exit."""
+    for key in list(_ptwrite_filter.keys()):
+        if key.startswith(f"{event.inferior.pid}."):
+            del _ptwrite_filter[key]
+
+
+gdb.events.exited.connect(ptwrite_exit_handler)
+
+
+def _clear_traces(thread_list):
+    """Helper function to clear the trace of all threads in THREAD_LIST."""
+    current_thread = gdb.selected_thread()
+    recording = gdb.current_recording()
+
+    if recording is not None:
+        for thread in thread_list:
+            thread.switch()
+            recording.clear()
+
+        current_thread.switch()
+
+
+def register_filter(filter_):
+    """Register the ptwrite filter function."""
+    if filter_ is not None and not callable(filter_):
+        raise TypeError("The filter must be callable or 'None'.")
+
+    # Clear the traces of all threads to force re-decoding with
+    # the new filter.
+    thread_list = gdb.selected_inferior().threads()
+    _clear_traces(thread_list)
+
+    _ptwrite_filter.clear()
+    _ptwrite_filter["global"] = filter_
+
+
+def get_filter():
+    """Returns the filters of the current thread."""
+    # The key is of this format to enable an per-inferior cleanup when an
+    # inferior exits.
+    key = f"{gdb.selected_inferior().pid}.{gdb.selected_thread().ptid[1]}"
+
+    # Create a new filter for new threads.
+    if key not in _ptwrite_filter.keys():
+        _ptwrite_filter[key] = deepcopy(_ptwrite_filter["global"])
+
+    return _ptwrite_filter[key]
diff --git a/gdb/python/py-record-btrace.c b/gdb/python/py-record-btrace.c
index c09ba6271b5..6234124e3b1 100644
--- a/gdb/python/py-record-btrace.c
+++ b/gdb/python/py-record-btrace.c
@@ -763,6 +763,94 @@ recpy_bt_function_call_history (PyObject *self, void *closure)
   return btpy_list_new (tinfo, first, last, 1, &recpy_func_type);
 }
 
+/* Helper function that calls PTW_FILTER with PAYLOAD and IP as arguments.
+   Returns the string that will be printed.  */
+std::string
+recpy_call_filter (const uint64_t payload, const uint64_t ip,
+		   const void *ptw_filter)
+{
+  std::string result;
+
+  if ((PyObject *) ptw_filter == nullptr)
+    error (_("No valid ptwrite filter."));
+  if ((PyObject *) ptw_filter == Py_None)
+    return result;
+
+  gdbpy_enter enter_py;
+
+  gdbpy_ref<> py_payload (PyLong_FromUnsignedLongLong (payload));
+  gdbpy_ref<> py_ip (PyLong_FromUnsignedLongLong (ip));
+
+  if (ip == 0)
+    py_ip = gdbpy_ref<>::new_reference (Py_None);
+
+  gdbpy_ref<> py_result (PyObject_CallFunctionObjArgs ((PyObject *) ptw_filter,
+							py_payload.get (),
+							py_ip.get (),
+							nullptr));
+
+  if (PyErr_Occurred ())
+    {
+      gdbpy_print_stack ();
+      gdbpy_error (_("Couldn't call the ptwrite filter."));
+    }
+
+  /* Py_None is valid and results in no output.  */
+  if (py_result == Py_None)
+    return result;
+
+  result = gdbpy_obj_to_string (py_result.get ()).get ();
+
+  if (PyErr_Occurred ())
+    {
+      gdbpy_print_stack ();
+      gdbpy_error (_("The ptwrite filter didn't return a string."));
+    }
+
+  return result;
+}
+
+/* Helper function returning the current ptwrite filter.  */
+
+PyObject *
+get_ptwrite_filter ()
+{
+  gdbpy_ref<> module (PyImport_ImportModule ("gdb.ptwrite"));
+
+  if (PyErr_Occurred ())
+  {
+    gdbpy_print_stack ();
+    return nullptr;
+  }
+
+  /* We need to keep the reference count.  */
+  gdbpy_ref<> ptw_filter (PyObject_CallMethod (module.get (), "get_filter",
+					       nullptr));
+
+  if (PyErr_Occurred ())
+    {
+      gdbpy_print_stack ();
+      gdbpy_error (_("Couldn't get the ptwrite filter."));
+    }
+
+  return ptw_filter.get();
+}
+
+/* Used for registering the default ptwrite filter to the current thread.  A
+   pointer to this function is stored in the python extension interface.  */
+
+void
+gdbpy_load_ptwrite_filter (const struct extension_language_defn *extlang,
+			   struct btrace_thread_info *btinfo)
+{
+  gdb_assert (btinfo != nullptr);
+
+  gdbpy_enter enter_py;
+
+  btinfo->ptw_callback_fun = &recpy_call_filter;
+  btinfo->ptw_context= get_ptwrite_filter ();
+}
+
 /* Implementation of BtraceRecord.goto (self, BtraceInstruction) -> None.  */
 
 PyObject *
diff --git a/gdb/python/py-record-btrace.h b/gdb/python/py-record-btrace.h
index b25b121d255..96abb07eaef 100644
--- a/gdb/python/py-record-btrace.h
+++ b/gdb/python/py-record-btrace.h
@@ -91,4 +91,12 @@ extern PyObject *recpy_bt_func_prev (PyObject *self, void *closure);
 /* Implementation of RecordFunctionSegment.next [RecordFunctionSegment].  */
 extern PyObject *recpy_bt_func_next (PyObject *self, void *closure);
 
+/* Helper function returning the current ptwrite filter.  */
+extern PyObject *get_ptwrite_filter ();
+
+/* Helper function to call the ptwrite filter.  */
+extern std::string recpy_call_filter (const uint64_t payload,
+				      const uint64_t ip,
+				      const void *ptw_filter);
+
 #endif /* PYTHON_PY_RECORD_BTRACE_H */
diff --git a/gdb/python/python-internal.h b/gdb/python/python-internal.h
index d624b23fdc5..7b0088847dd 100644
--- a/gdb/python/python-internal.h
+++ b/gdb/python/python-internal.h
@@ -364,6 +364,9 @@ extern enum ext_lang_rc gdbpy_apply_val_pretty_printer
    struct ui_file *stream, int recurse,
    const struct value_print_options *options,
    const struct language_defn *language);
+extern void gdbpy_load_ptwrite_filter
+  (const struct extension_language_defn *extlang,
+   struct btrace_thread_info *btinfo);
 extern enum ext_lang_bt_status gdbpy_apply_frame_filter
   (const struct extension_language_defn *,
    struct frame_info *frame, frame_filter_flags flags,
diff --git a/gdb/python/python.c b/gdb/python/python.c
index 516e2c983cf..f8211258756 100644
--- a/gdb/python/python.c
+++ b/gdb/python/python.c
@@ -151,6 +151,8 @@ static const struct extension_language_ops python_extension_ops =
 
   gdbpy_apply_frame_filter,
 
+  gdbpy_load_ptwrite_filter,
+
   gdbpy_preserve_values,
 
   gdbpy_breakpoint_has_cond,
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (8 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration Felix Willgerodt
@ 2022-09-16 11:36 ` Felix Willgerodt
  2022-09-16 12:01   ` Eli Zaretskii
  2022-10-19 14:45 ` [PATCH v6 00/10] Extensions for PTWRITE Metzger, Markus T
  10 siblings, 1 reply; 33+ messages in thread
From: Felix Willgerodt @ 2022-09-16 11:36 UTC (permalink / raw)
  To: gdb-patches, markus.t.metzger

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                                  |   6 +
 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      | 209 ++++++++
 gdb/testsuite/gdb.btrace/x86_64-ptwrite.S | 544 +++++++++++++++++++++
 gdb/testsuite/lib/gdb.exp                 |  74 +++
 gdbsupport/common.m4                      |   2 +
 gdbsupport/config.in                      |   3 +
 gdbsupport/configure                      |  11 +
 13 files changed, 1656 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 555ef2ddf77..d75027ad64d 100644
--- a/gdb/NEWS
+++ b/gdb/NEWS
@@ -3,6 +3,12 @@
 
 *** Changes since GDB 12
 
+* GDB now supports printing of ptwrite payloads from the Intel Processor
+  Trace during 'record instruction-history', 'record function-call-history',
+  all stepping commands and in  Python.  Printing is customizable via a
+  ptwrite filter function in  Python.  By default, the raw ptwrite
+  payload is printed for each ptwrite that is encountered.
+
 * GDB now supports dumping memory tag data for AArch64 MTE.  It also supports
   reading memory tag data for AArch64 MTE from core files generated by
   the gcore command or the Linux kernel.
diff --git a/gdb/btrace.c b/gdb/btrace.c
index 14739d1d922..0dd816035e7 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_string.empty ())
+	      break;
+
+	    btinfo->aux_data.emplace_back (ptw_string);
+
+	    if (!btinfo->functions.empty ()
+		&& !btinfo->functions.back ().insn.empty ())
+	      flags = btinfo->functions.back ().insn.back ().flags;
+
+	    /* Update insn list with ptw payload insn.  */
+	    struct btrace_insn ptw_insn;
+	    ptw_insn.aux_data_index = btinfo->aux_data.size () - 1;
+	    ptw_insn.size = 0;
+	    ptw_insn.iclass = BTRACE_INSN_AUX;
+	    ptw_insn.flags = flags;
+
+	    if (ip != 0)
+	      bfun = ftrace_update_function (btinfo, ip);
+	    else
+	      {
+		if (btinfo->functions.empty ())
+		  bfun = ftrace_new_function (btinfo, NULL, NULL);
+		else
+		  bfun = &btinfo->functions.back ();
+	      }
+
+	    bfun->flags |= BFUN_CONTAINS_AUX;
+	    ftrace_update_insns (bfun, ptw_insn);
+
+	    break;
+	  }
+#endif /* defined (HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE) */
 	}
     }
 #endif /* defined (HAVE_PT_INSN_EVENT) */
@@ -2981,6 +3029,12 @@ pt_print_packet (const struct pt_packet *packet)
     case ppt_mnt:
       gdb_printf (("mnt %" PRIx64 ""), packet->payload.mnt.payload);
       break;
+
+#if defined (HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE)
+    case ppt_ptw:
+      gdb_printf (("ptw %" PRIx64 ""), packet->payload.ptw.payload);
+      break;
+#endif /* defined (HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE) */
     }
 }
 
diff --git a/gdb/config.in b/gdb/config.in
index a4975c68aad..d572c51ec79 100644
--- a/gdb/config.in
+++ b/gdb/config.in
@@ -460,6 +460,9 @@
 /* Define to 1 if `pl_tdname' is a member of `struct ptrace_lwpinfo'. */
 #undef HAVE_STRUCT_PTRACE_LWPINFO_PL_TDNAME
 
+/* Define to 1 if `variant.ptwrite' is a member of `struct pt_event'. */
+#undef HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE
+
 /* Define to 1 if `enabled' is a member of `struct pt_insn'. */
 #undef HAVE_STRUCT_PT_INSN_ENABLED
 
diff --git a/gdb/configure b/gdb/configure
index 4dbd0c3b13c..b1ad2a001ec 100755
--- a/gdb/configure
+++ b/gdb/configure
@@ -15349,6 +15349,17 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+fi
+
+      ac_fn_c_check_member "$LINENO" "struct pt_event" "variant.ptwrite" "ac_cv_member_struct_pt_event_variant_ptwrite" "#include <intel-pt.h>
+"
+if test "x$ac_cv_member_struct_pt_event_variant_ptwrite" = xyes; then :
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE 1
+_ACEOF
+
+
 fi
 
       LIBS=$save_LIBS
diff --git a/gdb/doc/python.texi b/gdb/doc/python.texi
index b963202b7b9..7d600677077 100644
--- a/gdb/doc/python.texi
+++ b/gdb/doc/python.texi
@@ -7115,6 +7115,7 @@ registering objfile-specific pretty-printers and frame-filters.
 * gdb.printing::       Building and registering pretty-printers.
 * gdb.types::          Utilities for working with types.
 * gdb.prompt::         Utilities for prompt value substitution.
+* gdb.ptwrite::        Utilities for ptwrite filter registration.
 @end menu
 
 @node gdb.printing
@@ -7305,3 +7306,152 @@ substitute_prompt ("frame: \f, args: \p@{print frame-arguments@}")
 "frame: main, args: scalars"
 @end smallexample
 @end table
+
+@node gdb.ptwrite
+@subsubsection gdb.ptwrite
+@cindex gdb.ptwrite
+
+This module provides additional functionality for recording programs that
+make use of the @code{PTWRITE} instruction.  @code{PTWRITE} is a x86
+instruction that allows to write values into the Intel Processor Trace
+(@pxref{Process Record and Replay}).
+The @value{NGCC} built-in functions for it are:
+@smallexample
+void __builtin_ia32_ptwrite32 (unsigned);
+void __builtin_ia32_ptwrite64 (unsigned long long);
+@end smallexample
+
+If an inferior uses the instruction, @value{GDBN} by default inserts the
+raw payload value as auxiliary information into the execution history.
+Auxiliary information is by default printed during
+@code{record instruction-history}, @code{record function-call-history},
+and all stepping commands and is accessible in Python as a
+@code{RecordAuxiliary} object.
+
+@exdent Sample program:
+@smallexample
+@group
+void
+ptwrite64 (unsigned long long value)
+@{
+  __builtin_ia32_ptwrite64 (value);
+@}
+@end group
+
+@group
+int
+main (void)
+@{
+  ptwrite64 (0x42);
+  return 0; /* break here.  */
+@}
+@end group
+@end smallexample
+
+
+@exdent @value{GDBN} output after recording the sample program in pt format:
+@smallexample
+@group
+(gdb) record instruction-history 12,14
+12         0x0040074c <ptwrite64+16>:   ptwrite %rbx
+13         [42]
+14         0x00400751 <ptwrite64+21>:   mov -0x8(%rbp),%rbx
+(gdb) record function-call-history
+1       main
+2       ptwrite64
+                [42]
+3       main
+@end group
+@end smallexample
+
+The @code{gdb.ptwrite} module allows customizing the default output of ptwrite
+auxiliary information.  A custom Python function can be registered via
+@code{gdb.ptwrite.register_filter} as the ptwrite filter function.
+This function will be called with the ptwrite payload and PC as arguments
+during trace decoding.
+
+@findex gdb.ptwrite.register_filter
+@defun register_filter (@var{filter})
+Used to register the ptwrite filter.  The filter can be any callable
+object that accepts two arguments, the payload and PC.  It can return
+a string, which will be printed by @value{GDBN} during the aforementioned
+commands, or @code{None}, resulting in no output.  @code{None} can also be
+registered to deactivate printing.
+@end defun
+
+@findex gdb.ptwrite.get_filter
+@defun get_filter ()
+Returns the currently active 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
+different filters for different threads.  The filter can however
+distinguish between multiple threads with the help of
+@code{gdb.selected_thread().global_num} (@pxref{Threads In Python}) or
+similar.  For example:
+
+@smallexample
+@group
+(gdb) python-interactive
+>>> class my_filter():
+...    def __init__(self):
+...        self.var = 0
+...    def __call__(self, payload, ip):
+...        if gdb.selected_thread().global_num == 1:
+...            self.var += 1
+...            return f"counter: @{self.var@}, ip: @{ip:#x@}"
+...        else:
+...            return None
+...
+>>> import gdb.ptwrite
+>>> gdb.ptwrite.register_filter(my_filter())
+>>>
+@end group
+
+@group
+(gdb) record function-call-history 59,64
+59      pthread_create@@GLIBC_2.2.5
+60      job()
+61      task(void*)
+62      ptwrite64(unsigned long)
+                [counter: 1, ip: 0x401156]
+63      task(void*)
+64      ptwrite32(unsigned int)
+                [counter: 2, ip: 0x40116c]
+@end group
+
+@group
+(gdb) info threads 
+* 1    Thread 0x7ffff7fd8740 (LWP 25796) "ptw_threads" task ()
+    at bin/ptwrite/ptw_threads.c:45
+  2    Thread 0x7ffff6eb8700 (LWP 25797) "ptw_threads" task ()
+    at bin/ptwrite/ptw_threads.c:45
+@end group
+
+@group
+(gdb) thread 2
+[Switching to thread 2 (Thread 0x7ffff6eb8700 (LWP 25797))]
+#0  task (arg=0x0) at ptwrite_threads.c:45
+45        return NULL;
+@end group
+
+@group
+(gdb) record function-call-history 10,14
+10    start_thread
+11    task(void*)
+12    ptwrite64(unsigned long)
+13    task(void*)
+14    ptwrite32(unsigned int)
+@end group
+@end smallexample
+
+This @value{GDBN} feature is dependent on hardware and operating system
+support and requires the Intel Processor Trace decoder library in version
+2.0.0 or newer.
diff --git a/gdb/testsuite/gdb.btrace/i386-ptwrite.S b/gdb/testsuite/gdb.btrace/i386-ptwrite.S
new file mode 100644
index 00000000000..efafb4b847b
--- /dev/null
+++ b/gdb/testsuite/gdb.btrace/i386-ptwrite.S
@@ -0,0 +1,550 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2022 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+
+   This file has been generated using gcc version 10.3.1 20210422
+   (Red Hat 10.3.1-1):
+   gcc -S -dA -g -m32 -mptwrite ptwrite.c -o i386-ptwrite.S.  */
+
+
+	.file	"ptwrite.c"
+	.text
+.Ltext0:
+	.globl	ptwrite1
+	.type	ptwrite1, @function
+ptwrite1:
+.LFB4021:
+	.file 1 "ptwrite.c"
+	# ptwrite.c:22:1
+	.loc 1 22 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushl	%ebp
+	.cfi_def_cfa_offset 8
+	.cfi_offset 5, -8
+	movl	%esp, %ebp
+	.cfi_def_cfa_register 5
+	subl	$16, %esp
+	# ptwrite.c:23:3
+	.loc 1 23 3
+	movl	8(%ebp), %eax
+	movl	%eax, -4(%ebp)
+.LBB6:
+.LBB7:
+	.file 2 "/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h"
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:307:3
+	.loc 2 307 3
+	movl	-4(%ebp), %eax
+	ptwrite	%eax
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:308:1
+	.loc 2 308 1
+	nop
+.LBE7:
+.LBE6:
+	# ptwrite.c:24:1
+	.loc 1 24 1
+	nop
+	leave
+	.cfi_restore 5
+	.cfi_def_cfa 4, 4
+# SUCC: EXIT [always]
+	ret
+	.cfi_endproc
+.LFE4021:
+	.size	ptwrite1, .-ptwrite1
+	.globl	ptwrite2
+	.type	ptwrite2, @function
+ptwrite2:
+.LFB4022:
+	# ptwrite.c:28:1
+	.loc 1 28 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushl	%ebp
+	.cfi_def_cfa_offset 8
+	.cfi_offset 5, -8
+	movl	%esp, %ebp
+	.cfi_def_cfa_register 5
+	subl	$16, %esp
+	# ptwrite.c:29:3
+	.loc 1 29 3
+	movl	8(%ebp), %eax
+	movl	%eax, -4(%ebp)
+.LBB8:
+.LBB9:
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:307:3
+	.loc 2 307 3
+	movl	-4(%ebp), %eax
+	ptwrite	%eax
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:308:1
+	.loc 2 308 1
+	nop
+.LBE9:
+.LBE8:
+	# ptwrite.c:30:1
+	.loc 1 30 1
+	nop
+	leave
+	.cfi_restore 5
+	.cfi_def_cfa 4, 4
+# SUCC: EXIT [always]
+	ret
+	.cfi_endproc
+.LFE4022:
+	.size	ptwrite2, .-ptwrite2
+	.globl	main
+	.type	main, @function
+main:
+.LFB4023:
+	# ptwrite.c:34:1
+	.loc 1 34 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushl	%ebp
+	.cfi_def_cfa_offset 8
+	.cfi_offset 5, -8
+	movl	%esp, %ebp
+	.cfi_def_cfa_register 5
+	# ptwrite.c:35:3
+	.loc 1 35 3
+	pushl	$66
+	call	ptwrite1
+	addl	$4, %esp
+	# ptwrite.c:36:3
+	.loc 1 36 3
+	pushl	$67
+	call	ptwrite2
+	addl	$4, %esp
+	# ptwrite.c:38:10
+	.loc 1 38 10
+	movl	$0, %eax
+	# ptwrite.c:39:1
+	.loc 1 39 1
+	leave
+	.cfi_restore 5
+	.cfi_def_cfa 4, 4
+# SUCC: EXIT [always]
+	ret
+	.cfi_endproc
+.LFE4023:
+	.size	main, .-main
+.Letext0:
+	.section	.debug_info,"",@progbits
+.Ldebug_info0:
+	.long	0x129	# Length of Compilation Unit Info
+	.value	0x4	# DWARF version number
+	.long	.Ldebug_abbrev0	# Offset Into Abbrev. Section
+	.byte	0x4	# Pointer Size (in bytes)
+	.uleb128 0x1	# (DIE (0xb) DW_TAG_compile_unit)
+	.long	.LASF15	# DW_AT_producer: "GNU C17 10.3.1 20210422 (Red Hat 10.3.1-1) -m32 -mptwrite -mtune=generic -march=i686 -g"
+	.byte	0xc	# DW_AT_language
+	.long	.LASF16	# DW_AT_name: "ptwrite.c"
+	.long	.LASF17	# DW_AT_comp_dir: "gdb/gdb/testsuite/gdb.btrace"
+	.long	.Ltext0	# DW_AT_low_pc
+	.long	.Letext0-.Ltext0	# DW_AT_high_pc
+	.long	.Ldebug_line0	# DW_AT_stmt_list
+	.uleb128 0x2	# (DIE (0x25) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.ascii "int\0"	# DW_AT_name
+	.uleb128 0x3	# (DIE (0x2c) DW_TAG_base_type)
+	.byte	0x2	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF0	# DW_AT_name: "short int"
+	.uleb128 0x3	# (DIE (0x33) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x6	# DW_AT_encoding
+	.long	.LASF1	# DW_AT_name: "char"
+	.uleb128 0x3	# (DIE (0x3a) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF2	# DW_AT_name: "long long int"
+	.uleb128 0x3	# (DIE (0x41) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x4	# DW_AT_encoding
+	.long	.LASF3	# DW_AT_name: "float"
+	.uleb128 0x3	# (DIE (0x48) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF4	# DW_AT_name: "unsigned int"
+	.uleb128 0x3	# (DIE (0x4f) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF5	# DW_AT_name: "long int"
+	.uleb128 0x3	# (DIE (0x56) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x8	# DW_AT_encoding
+	.long	.LASF6	# DW_AT_name: "unsigned char"
+	.uleb128 0x3	# (DIE (0x5d) DW_TAG_base_type)
+	.byte	0x2	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF7	# DW_AT_name: "short unsigned int"
+	.uleb128 0x3	# (DIE (0x64) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF8	# DW_AT_name: "long unsigned int"
+	.uleb128 0x3	# (DIE (0x6b) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x6	# DW_AT_encoding
+	.long	.LASF9	# DW_AT_name: "signed char"
+	.uleb128 0x3	# (DIE (0x72) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF10	# DW_AT_name: "long long unsigned int"
+	.uleb128 0x3	# (DIE (0x79) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x4	# DW_AT_encoding
+	.long	.LASF11	# DW_AT_name: "double"
+	.uleb128 0x4	# (DIE (0x80) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF18	# DW_AT_name: "main"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x21	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.long	0x25	# DW_AT_type
+	.long	.LFB4023	# DW_AT_low_pc
+	.long	.LFE4023-.LFB4023	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_tail_call_sites
+	.uleb128 0x5	# (DIE (0x96) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF12	# DW_AT_name: "ptwrite2"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x1b	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.long	.LFB4022	# DW_AT_low_pc
+	.long	.LFE4022-.LFB4022	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_call_sites
+	.long	0xd5	# DW_AT_sibling
+	.uleb128 0x6	# (DIE (0xac) DW_TAG_formal_parameter)
+	.long	.LASF14	# DW_AT_name: "value"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x1b	# DW_AT_decl_line
+	.byte	0xf	# DW_AT_decl_column
+	.long	0x25	# DW_AT_type
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 0
+	.uleb128 0x7	# (DIE (0xbb) DW_TAG_inlined_subroutine)
+	.long	0x114	# DW_AT_abstract_origin
+	.long	.LBB8	# DW_AT_low_pc
+	.long	.LBE8-.LBB8	# DW_AT_high_pc
+	.byte	0x1	# DW_AT_call_file (ptwrite.c)
+	.byte	0x1d	# DW_AT_call_line
+	.byte	0x3	# DW_AT_call_column
+	.uleb128 0x8	# (DIE (0xcb) DW_TAG_formal_parameter)
+	.long	0x11e	# DW_AT_abstract_origin
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -12
+	.byte	0	# end of children of DIE 0xbb
+	.byte	0	# end of children of DIE 0x96
+	.uleb128 0x5	# (DIE (0xd5) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF13	# DW_AT_name: "ptwrite1"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x15	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.long	.LFB4021	# DW_AT_low_pc
+	.long	.LFE4021-.LFB4021	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_call_sites
+	.long	0x114	# DW_AT_sibling
+	.uleb128 0x6	# (DIE (0xeb) DW_TAG_formal_parameter)
+	.long	.LASF14	# DW_AT_name: "value"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x15	# DW_AT_decl_line
+	.byte	0xf	# DW_AT_decl_column
+	.long	0x25	# DW_AT_type
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 0
+	.uleb128 0x7	# (DIE (0xfa) DW_TAG_inlined_subroutine)
+	.long	0x114	# DW_AT_abstract_origin
+	.long	.LBB6	# DW_AT_low_pc
+	.long	.LBE6-.LBB6	# DW_AT_high_pc
+	.byte	0x1	# DW_AT_call_file (ptwrite.c)
+	.byte	0x17	# DW_AT_call_line
+	.byte	0x3	# DW_AT_call_column
+	.uleb128 0x8	# (DIE (0x10a) DW_TAG_formal_parameter)
+	.long	0x11e	# DW_AT_abstract_origin
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -12
+	.byte	0	# end of children of DIE 0xfa
+	.byte	0	# end of children of DIE 0xd5
+	.uleb128 0x9	# (DIE (0x114) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF19	# DW_AT_name: "_ptwrite32"
+	.byte	0x2	# DW_AT_decl_file (/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h)
+	.value	0x131	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.byte	0x3	# DW_AT_inline
+			# DW_AT_artificial
+	.uleb128 0xa	# (DIE (0x11e) DW_TAG_formal_parameter)
+	.ascii "__B\0"	# DW_AT_name
+	.byte	0x2	# DW_AT_decl_file (/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h)
+	.value	0x131	# DW_AT_decl_line
+	.byte	0x16	# DW_AT_decl_column
+	.long	0x48	# DW_AT_type
+	.byte	0	# end of children of DIE 0x114
+	.byte	0	# end of children of DIE 0xb
+	.section	.debug_abbrev,"",@progbits
+.Ldebug_abbrev0:
+	.uleb128 0x1	# (abbrev code)
+	.uleb128 0x11	# (TAG: DW_TAG_compile_unit)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x25	# (DW_AT_producer)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x13	# (DW_AT_language)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x1b	# (DW_AT_comp_dir)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x6	# (DW_FORM_data4)
+	.uleb128 0x10	# (DW_AT_stmt_list)
+	.uleb128 0x17	# (DW_FORM_sec_offset)
+	.byte	0
+	.byte	0
+	.uleb128 0x2	# (abbrev code)
+	.uleb128 0x24	# (TAG: DW_TAG_base_type)
+	.byte	0	# DW_children_no
+	.uleb128 0xb	# (DW_AT_byte_size)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3e	# (DW_AT_encoding)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0x8	# (DW_FORM_string)
+	.byte	0
+	.byte	0
+	.uleb128 0x3	# (abbrev code)
+	.uleb128 0x24	# (TAG: DW_TAG_base_type)
+	.byte	0	# DW_children_no
+	.uleb128 0xb	# (DW_AT_byte_size)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3e	# (DW_AT_encoding)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.byte	0
+	.byte	0
+	.uleb128 0x4	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0	# DW_children_no
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x6	# (DW_FORM_data4)
+	.uleb128 0x40	# (DW_AT_frame_base)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.uleb128 0x2116	# (DW_AT_GNU_all_tail_call_sites)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.byte	0
+	.byte	0
+	.uleb128 0x5	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x6	# (DW_FORM_data4)
+	.uleb128 0x40	# (DW_AT_frame_base)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.uleb128 0x2117	# (DW_AT_GNU_all_call_sites)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x1	# (DW_AT_sibling)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.byte	0
+	.byte	0
+	.uleb128 0x6	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x2	# (DW_AT_location)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.byte	0
+	.byte	0
+	.uleb128 0x7	# (abbrev code)
+	.uleb128 0x1d	# (TAG: DW_TAG_inlined_subroutine)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x31	# (DW_AT_abstract_origin)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x6	# (DW_FORM_data4)
+	.uleb128 0x58	# (DW_AT_call_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x59	# (DW_AT_call_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x57	# (DW_AT_call_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.byte	0
+	.byte	0
+	.uleb128 0x8	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x31	# (DW_AT_abstract_origin)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x2	# (DW_AT_location)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.byte	0
+	.byte	0
+	.uleb128 0x9	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0x5	# (DW_FORM_data2)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x20	# (DW_AT_inline)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x34	# (DW_AT_artificial)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.byte	0
+	.byte	0
+	.uleb128 0xa	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0x8	# (DW_FORM_string)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0x5	# (DW_FORM_data2)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.byte	0
+	.byte	0
+	.byte	0
+	.section	.debug_aranges,"",@progbits
+	.long	0x1c	# Length of Address Ranges Info
+	.value	0x2	# DWARF aranges version
+	.long	.Ldebug_info0	# Offset of Compilation Unit Info
+	.byte	0x4	# Size of Address
+	.byte	0	# Size of Segment Descriptor
+	.value	0	# Pad to 8 byte boundary
+	.value	0
+	.long	.Ltext0	# Address
+	.long	.Letext0-.Ltext0	# Length
+	.long	0
+	.long	0
+	.section	.debug_line,"",@progbits
+.Ldebug_line0:
+	.section	.debug_str,"MS",@progbits,1
+.LASF2:
+	.string	"long long int"
+.LASF14:
+	.string	"value"
+.LASF19:
+	.string	"_ptwrite32"
+.LASF4:
+	.string	"unsigned int"
+.LASF17:
+	.string	"gdb/gdb/testsuite/gdb.btrace"
+.LASF18:
+	.string	"main"
+.LASF16:
+	.string	"ptwrite.c"
+.LASF8:
+	.string	"long unsigned int"
+.LASF10:
+	.string	"long long unsigned int"
+.LASF13:/
+	.string	"ptwrite1"
+.LASF12:
+	.string	"ptwrite2"
+.LASF6:
+	.string	"unsigned char"
+.LASF1:
+	.string	"char"
+.LASF5:
+	.string	"long int"
+.LASF15:
+	.string	"GNU C17 10.3.1 20210422 (Red Hat 10.3.1-1) -m32 -mptwrite -mtune=generic -march=i686 -g"
+.LASF11:
+	.string	"double"
+.LASF7:
+	.string	"short unsigned int"
+.LASF9:
+	.string	"signed char"
+.LASF3:
+	.string	"float"
+.LASF0:
+	.string	"short int"
+	.ident	"GCC: (GNU) 10.3.1 20210422 (Red Hat 10.3.1-1)"
+	.section	.note.GNU-stack,"",@progbits
diff --git a/gdb/testsuite/gdb.btrace/ptwrite.c b/gdb/testsuite/gdb.btrace/ptwrite.c
new file mode 100644
index 00000000000..c85b656919b
--- /dev/null
+++ b/gdb/testsuite/gdb.btrace/ptwrite.c
@@ -0,0 +1,39 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2021 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#include <immintrin.h>
+
+void
+ptwrite1 (int value)
+{
+  _ptwrite32 (value);
+}
+
+void
+ptwrite2 (int value)
+{
+  _ptwrite32 (value);
+}
+
+int
+main (void)
+{
+  ptwrite1 (0x42);
+  ptwrite2 (0x43);
+
+  return 0;
+}
diff --git a/gdb/testsuite/gdb.btrace/ptwrite.exp b/gdb/testsuite/gdb.btrace/ptwrite.exp
new file mode 100644
index 00000000000..569760087a4
--- /dev/null
+++ b/gdb/testsuite/gdb.btrace/ptwrite.exp
@@ -0,0 +1,209 @@
+# This testcase is part of GDB, the GNU debugger.
+#
+# Copyright 2021 Free Software Foundation, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+load_lib gdb-python.exp
+
+if { [skip_btrace_ptw_tests] } {
+    unsupported "No support for ptwrite decoding."
+    return -1
+}
+
+set opts {}
+
+if [info exists COMPILE] {
+    # make check RUNTESTFLAGS="gdb.btrace/ptwrite.exp COMPILE=1"
+    standard_testfile ptwrite.c
+    lappend opts debug additional_flags=-mptwrite
+} elseif {[istarget "i?86-*-*"] || [istarget "x86_64-*-*"]} {
+    if {[is_amd64_regs_target]} {
+	standard_testfile x86_64-ptwrite.S
+    } else {
+	standard_testfile i386-ptwrite.S
+    }
+} else {
+    unsupported "target architecture not supported"
+    return -1
+}
+
+if [prepare_for_testing "failed to prepare" $testfile $srcfile $opts] {
+    return -1
+}
+
+if { ![runto_main] } {
+    untested "failed to run to main"
+    return -1
+}
+
+# This needs to be after runto_main
+if { [skip_python_tests] } {
+    unsupported "Configuration doesn't support python scripting."
+    return -1
+}
+
+### 1. Default testrun
+
+# Setup recording
+gdb_test_no_output "set record instruction-history-size unlimited"
+gdb_test_no_output "record btrace pt"
+gdb_test "next" ".*" "next"
+gdb_test "next" ".*" "next 2"
+
+with_test_prefix "Default" {
+    # Test record instruction-history
+    gdb_test "record instruction-history 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[42\\\]" \
+	".*\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[43\\\].*" \
+	]
+
+    gdb_test "record instruction-history /a 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	".*\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+.*" \
+	]
+
+    # Test function call history
+    gdb_test "record function-call-history 1,4" [multi_line \
+	"1\tmain" \
+	"2\tptwrite1" \
+	"\t\t\\\[42\\\]" \
+	"3\tmain" \
+	"4\tptwrite2" \
+	"\t\t\\\[43\\\]" \
+	]
+
+    gdb_test "record function-call-history /a 1,4" [multi_line \
+	"1\tmain" \
+	"2\tptwrite1" \
+	"3\tmain" \
+	"4\tptwrite2" \
+	]
+}
+
+# Test payload printing during stepping
+with_test_prefix "Stepping" {
+    gdb_test "record goto 10" "No such instruction\."
+    gdb_test "record goto 9" ".*ptwrite.* at .*"
+    gdb_test "stepi" ".*\\\[42\\\].*"
+    gdb_test "reverse-stepi" ".*\\\[42\\\].*"
+    gdb_test "continue" [multi_line \
+	    ".*\\\[42\\\]" \
+	    "\\\[43\\\].*" \
+	    ]
+    gdb_test "reverse-continue" [multi_line \
+	    ".*\\\[43\\\]" \
+	    "\\\[42\\\].*" \
+	    ]
+}
+
+# Test auxiliary type in python
+gdb_test_multiline "auxiliary type in python" \
+    "python" "" \
+    "h = gdb.current_recording().instruction_history" "" \
+    "for insn in h:" "" \
+    "    if hasattr(insn, 'decoded'):" "" \
+    "        print(insn.decoded.decode())" "" \
+    "    elif hasattr(insn, 'data'):" "" \
+    "        print(insn.data)" "" \
+    "end" \
+    [multi_line \
+	".*mov    -0x4\\\(%(e|r)bp\\\),%(e|r)ax" \
+	"ptwrite %eax" \
+	"42" \
+	"nop.*" \
+	"mov    -0x4\\\(%(e|r)bp\\\),%(e|r)ax" \
+	"ptwrite %eax" \
+	"43" \
+	"nop.*"
+    ]
+
+
+### 2. Test filter registration
+### 2.1 Custom filter
+with_test_prefix "Custom" {
+    gdb_test_multiline "register filter in python" \
+	"python" "" \
+	"def my_filter(payload, ip):" "" \
+	"    if  payload == 66:" "" \
+	"        return \"payload: {0}, ip: {1:#x}\".format(payload, ip)" "" \
+	"    else:" "" \
+	"        return None" "" \
+	"import gdb.ptwrite" "" \
+	"gdb.ptwrite.register_filter(my_filter)" "" \
+	"end" ""
+
+    gdb_test "record instruction-history 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[payload: 66, ip: $hex\\\]" \
+	".*\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:.*" \
+	]
+}
+
+### 2.2 None as filter
+with_test_prefix "None" {
+    gdb_test_multiline "register filter in python" \
+	"python" "" \
+	"import gdb.ptwrite" "" \
+	"gdb.ptwrite.register_filter(None)" "" \
+	"end" ""
+
+    gdb_test "record instruction-history 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:.*" \
+	"\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:.*" \
+	]
+}
+
+### 2.3 Lambdas as filter
+with_test_prefix "Lambdas" {
+    gdb_test_multiline "register filter in python" \
+	"python" "" \
+	"import gdb.ptwrite" "" \
+	"gdb.ptwrite.register_filter(lambda payload, ip: \"{}\".format(payload + 2))" "" \
+	"end" ""
+
+    gdb_test "record instruction-history 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[68\\\]" \
+	".*\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[69\\\].*" \
+	] "Lambdas: record instruction-history 1"
+}
+
+### 2.4 Functors as filter
+with_test_prefix "Functors" {
+    gdb_test_multiline "register filter in python" \
+	"python" "" \
+	"import gdb.ptwrite" "" \
+	"class foobar(object):" "" \
+	"    def __init__(self):" "" \
+	"        self.variable = 0" "" \
+	"    def __call__(self, payload, ip):" "" \
+	"        self.variable += 1" "" \
+	"        return \"{}, {}\".format(self.variable, payload)" "" \
+	"gdb.ptwrite.register_filter(foobar())" "" \
+	"end" ""
+
+    gdb_test "record instruction-history 1" [multi_line \
+	".*\[0-9\]+\t   $hex <ptwrite1\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[1, 66\\\]" \
+	".*\[0-9\]+\t   $hex <ptwrite2\\+\[0-9\]+>:\tptwrite %\[a-z\]+" \
+	"\[0-9\]+\t   \\\[2, 67\\\].*" \
+	] "Functors: record instruction-history 1"
+}
diff --git a/gdb/testsuite/gdb.btrace/x86_64-ptwrite.S b/gdb/testsuite/gdb.btrace/x86_64-ptwrite.S
new file mode 100644
index 00000000000..6fe64013399
--- /dev/null
+++ b/gdb/testsuite/gdb.btrace/x86_64-ptwrite.S
@@ -0,0 +1,544 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2022 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+
+   This file has been generated using gcc version 10.3.1 20210422
+   (Red Hat 10.3.1-1):
+   gcc -S -dA -g -mptwrite ptwrite.c -o x86_64-ptwrite.S.  */
+
+	.file	"ptwrite.c"
+	.text
+.Ltext0:
+	.globl	ptwrite1
+	.type	ptwrite1, @function
+ptwrite1:
+.LFB4096:
+	.file 1 "ptwrite.c"
+	# ptwrite.c:22:1
+	.loc 1 22 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushq	%rbp
+	.cfi_def_cfa_offset 16
+	.cfi_offset 6, -16
+	movq	%rsp, %rbp
+	.cfi_def_cfa_register 6
+	movl	%edi, -20(%rbp)
+	# ptwrite.c:23:3
+	.loc 1 23 3
+	movl	-20(%rbp), %eax
+	movl	%eax, -4(%rbp)
+.LBB6:
+.LBB7:
+	.file 2 "/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h"
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:307:3
+	.loc 2 307 3
+	movl	-4(%rbp), %eax
+	ptwrite	%eax
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:308:1
+	.loc 2 308 1
+	nop
+.LBE7:
+.LBE6:
+	# ptwrite.c:24:1
+	.loc 1 24 1
+	nop
+	popq	%rbp
+	.cfi_def_cfa 7, 8
+# SUCC: EXIT [always] 
+	ret
+	.cfi_endproc
+.LFE4096:
+	.size	ptwrite1, .-ptwrite1
+	.globl	ptwrite2
+	.type	ptwrite2, @function
+ptwrite2:
+.LFB4097:
+	# ptwrite.c:28:1
+	.loc 1 28 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushq	%rbp
+	.cfi_def_cfa_offset 16
+	.cfi_offset 6, -16
+	movq	%rsp, %rbp
+	.cfi_def_cfa_register 6
+	movl	%edi, -20(%rbp)
+	# ptwrite.c:29:3
+	.loc 1 29 3
+	movl	-20(%rbp), %eax
+	movl	%eax, -4(%rbp)
+.LBB8:
+.LBB9:
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:307:3
+	.loc 2 307 3
+	movl	-4(%rbp), %eax
+	ptwrite	%eax
+	# /usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h:308:1
+	.loc 2 308 1
+	nop
+.LBE9:
+.LBE8:
+	# ptwrite.c:30:1
+	.loc 1 30 1
+	nop
+	popq	%rbp
+	.cfi_def_cfa 7, 8
+# SUCC: EXIT [always] 
+	ret
+	.cfi_endproc
+.LFE4097:
+	.size	ptwrite2, .-ptwrite2
+	.globl	main
+	.type	main, @function
+main:
+.LFB4098:
+	# ptwrite.c:34:1
+	.loc 1 34 1
+	.cfi_startproc
+# BLOCK 2 seq:0
+# PRED: ENTRY (FALLTHRU)
+	pushq	%rbp
+	.cfi_def_cfa_offset 16
+	.cfi_offset 6, -16
+	movq	%rsp, %rbp
+	.cfi_def_cfa_register 6
+	# ptwrite.c:35:3
+	.loc 1 35 3
+	movl	$66, %edi
+	call	ptwrite1
+	# ptwrite.c:36:3
+	.loc 1 36 3
+	movl	$67, %edi
+	call	ptwrite2
+	# ptwrite.c:38:10
+	.loc 1 38 10
+	movl	$0, %eax
+	# ptwrite.c:39:1
+	.loc 1 39 1
+	popq	%rbp
+	.cfi_def_cfa 7, 8
+# SUCC: EXIT [always] 
+	ret
+	.cfi_endproc
+.LFE4098:
+	.size	main, .-main
+.Letext0:
+	.section	.debug_info,"",@progbits
+.Ldebug_info0:
+	.long	0x159	# Length of Compilation Unit Info
+	.value	0x4	# DWARF version number
+	.long	.Ldebug_abbrev0	# Offset Into Abbrev. Section
+	.byte	0x8	# Pointer Size (in bytes)
+	.uleb128 0x1	# (DIE (0xb) DW_TAG_compile_unit)
+	.long	.LASF15	# DW_AT_producer: "GNU C17 10.3.1 20210422 (Red Hat 10.3.1-1) -mptwrite -mtune=generic -march=x86-64 -g"
+	.byte	0xc	# DW_AT_language
+	.long	.LASF16	# DW_AT_name: "ptwrite.c"
+	.long	.LASF17	# DW_AT_comp_dir: "gdb/gdb/testsuite/gdb.btrace"
+	.quad	.Ltext0	# DW_AT_low_pc
+	.quad	.Letext0-.Ltext0	# DW_AT_high_pc
+	.long	.Ldebug_line0	# DW_AT_stmt_list
+	.uleb128 0x2	# (DIE (0x2d) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.ascii "int\0"	# DW_AT_name
+	.uleb128 0x3	# (DIE (0x34) DW_TAG_base_type)
+	.byte	0x2	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF0	# DW_AT_name: "short int"
+	.uleb128 0x3	# (DIE (0x3b) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x6	# DW_AT_encoding
+	.long	.LASF1	# DW_AT_name: "char"
+	.uleb128 0x3	# (DIE (0x42) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF2	# DW_AT_name: "long long int"
+	.uleb128 0x3	# (DIE (0x49) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x4	# DW_AT_encoding
+	.long	.LASF3	# DW_AT_name: "float"
+	.uleb128 0x3	# (DIE (0x50) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF4	# DW_AT_name: "long unsigned int"
+	.uleb128 0x3	# (DIE (0x57) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x5	# DW_AT_encoding
+	.long	.LASF5	# DW_AT_name: "long int"
+	.uleb128 0x3	# (DIE (0x5e) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x8	# DW_AT_encoding
+	.long	.LASF6	# DW_AT_name: "unsigned char"
+	.uleb128 0x3	# (DIE (0x65) DW_TAG_base_type)
+	.byte	0x2	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF7	# DW_AT_name: "short unsigned int"
+	.uleb128 0x3	# (DIE (0x6c) DW_TAG_base_type)
+	.byte	0x4	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF8	# DW_AT_name: "unsigned int"
+	.uleb128 0x3	# (DIE (0x73) DW_TAG_base_type)
+	.byte	0x1	# DW_AT_byte_size
+	.byte	0x6	# DW_AT_encoding
+	.long	.LASF9	# DW_AT_name: "signed char"
+	.uleb128 0x3	# (DIE (0x7a) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x7	# DW_AT_encoding
+	.long	.LASF10	# DW_AT_name: "long long unsigned int"
+	.uleb128 0x3	# (DIE (0x81) DW_TAG_base_type)
+	.byte	0x8	# DW_AT_byte_size
+	.byte	0x4	# DW_AT_encoding
+	.long	.LASF11	# DW_AT_name: "double"
+	.uleb128 0x4	# (DIE (0x88) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF18	# DW_AT_name: "main"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x21	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.long	0x2d	# DW_AT_type
+	.quad	.LFB4098	# DW_AT_low_pc
+	.quad	.LFE4098-.LFB4098	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_tail_call_sites
+	.uleb128 0x5	# (DIE (0xa6) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF12	# DW_AT_name: "ptwrite2"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x1b	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.quad	.LFB4097	# DW_AT_low_pc
+	.quad	.LFE4097-.LFB4097	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_call_sites
+	.long	0xf5	# DW_AT_sibling
+	.uleb128 0x6	# (DIE (0xc4) DW_TAG_formal_parameter)
+	.long	.LASF14	# DW_AT_name: "value"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x1b	# DW_AT_decl_line
+	.byte	0xf	# DW_AT_decl_column
+	.long	0x2d	# DW_AT_type
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -36
+	.uleb128 0x7	# (DIE (0xd3) DW_TAG_inlined_subroutine)
+	.long	0x144	# DW_AT_abstract_origin
+	.quad	.LBB8	# DW_AT_low_pc
+	.quad	.LBE8-.LBB8	# DW_AT_high_pc
+	.byte	0x1	# DW_AT_call_file (ptwrite.c)
+	.byte	0x1d	# DW_AT_call_line
+	.byte	0x3	# DW_AT_call_column
+	.uleb128 0x8	# (DIE (0xeb) DW_TAG_formal_parameter)
+	.long	0x14e	# DW_AT_abstract_origin
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -20
+	.byte	0	# end of children of DIE 0xd3
+	.byte	0	# end of children of DIE 0xa6
+	.uleb128 0x5	# (DIE (0xf5) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF13	# DW_AT_name: "ptwrite1"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x15	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.quad	.LFB4096	# DW_AT_low_pc
+	.quad	.LFE4096-.LFB4096	# DW_AT_high_pc
+	.uleb128 0x1	# DW_AT_frame_base
+	.byte	0x9c	# DW_OP_call_frame_cfa
+			# DW_AT_GNU_all_call_sites
+	.long	0x144	# DW_AT_sibling
+	.uleb128 0x6	# (DIE (0x113) DW_TAG_formal_parameter)
+	.long	.LASF14	# DW_AT_name: "value"
+	.byte	0x1	# DW_AT_decl_file (ptwrite.c)
+	.byte	0x15	# DW_AT_decl_line
+	.byte	0xf	# DW_AT_decl_column
+	.long	0x2d	# DW_AT_type
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -36
+	.uleb128 0x7	# (DIE (0x122) DW_TAG_inlined_subroutine)
+	.long	0x144	# DW_AT_abstract_origin
+	.quad	.LBB6	# DW_AT_low_pc
+	.quad	.LBE6-.LBB6	# DW_AT_high_pc
+	.byte	0x1	# DW_AT_call_file (ptwrite.c)
+	.byte	0x17	# DW_AT_call_line
+	.byte	0x3	# DW_AT_call_column
+	.uleb128 0x8	# (DIE (0x13a) DW_TAG_formal_parameter)
+	.long	0x14e	# DW_AT_abstract_origin
+	.uleb128 0x2	# DW_AT_location
+	.byte	0x91	# DW_OP_fbreg
+	.sleb128 -20
+	.byte	0	# end of children of DIE 0x122
+	.byte	0	# end of children of DIE 0xf5
+	.uleb128 0x9	# (DIE (0x144) DW_TAG_subprogram)
+			# DW_AT_external
+	.long	.LASF19	# DW_AT_name: "_ptwrite32"
+	.byte	0x2	# DW_AT_decl_file (/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h)
+	.value	0x131	# DW_AT_decl_line
+	.byte	0x1	# DW_AT_decl_column
+			# DW_AT_prototyped
+	.byte	0x3	# DW_AT_inline
+			# DW_AT_artificial
+	.uleb128 0xa	# (DIE (0x14e) DW_TAG_formal_parameter)
+	.ascii "__B\0"	# DW_AT_name
+	.byte	0x2	# DW_AT_decl_file (/usr/lib/gcc/x86_64-redhat-linux/10/include/immintrin.h)
+	.value	0x131	# DW_AT_decl_line
+	.byte	0x16	# DW_AT_decl_column
+	.long	0x6c	# DW_AT_type
+	.byte	0	# end of children of DIE 0x144
+	.byte	0	# end of children of DIE 0xb
+	.section	.debug_abbrev,"",@progbits
+.Ldebug_abbrev0:
+	.uleb128 0x1	# (abbrev code)
+	.uleb128 0x11	# (TAG: DW_TAG_compile_unit)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x25	# (DW_AT_producer)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x13	# (DW_AT_language)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x1b	# (DW_AT_comp_dir)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x7	# (DW_FORM_data8)
+	.uleb128 0x10	# (DW_AT_stmt_list)
+	.uleb128 0x17	# (DW_FORM_sec_offset)
+	.byte	0
+	.byte	0
+	.uleb128 0x2	# (abbrev code)
+	.uleb128 0x24	# (TAG: DW_TAG_base_type)
+	.byte	0	# DW_children_no
+	.uleb128 0xb	# (DW_AT_byte_size)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3e	# (DW_AT_encoding)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0x8	# (DW_FORM_string)
+	.byte	0
+	.byte	0
+	.uleb128 0x3	# (abbrev code)
+	.uleb128 0x24	# (TAG: DW_TAG_base_type)
+	.byte	0	# DW_children_no
+	.uleb128 0xb	# (DW_AT_byte_size)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3e	# (DW_AT_encoding)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.byte	0
+	.byte	0
+	.uleb128 0x4	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0	# DW_children_no
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x7	# (DW_FORM_data8)
+	.uleb128 0x40	# (DW_AT_frame_base)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.uleb128 0x2116	# (DW_AT_GNU_all_tail_call_sites)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.byte	0
+	.byte	0
+	.uleb128 0x5	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x7	# (DW_FORM_data8)
+	.uleb128 0x40	# (DW_AT_frame_base)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.uleb128 0x2117	# (DW_AT_GNU_all_call_sites)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x1	# (DW_AT_sibling)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.byte	0
+	.byte	0
+	.uleb128 0x6	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x2	# (DW_AT_location)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.byte	0
+	.byte	0
+	.uleb128 0x7	# (abbrev code)
+	.uleb128 0x1d	# (TAG: DW_TAG_inlined_subroutine)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x31	# (DW_AT_abstract_origin)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x11	# (DW_AT_low_pc)
+	.uleb128 0x1	# (DW_FORM_addr)
+	.uleb128 0x12	# (DW_AT_high_pc)
+	.uleb128 0x7	# (DW_FORM_data8)
+	.uleb128 0x58	# (DW_AT_call_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x59	# (DW_AT_call_line)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x57	# (DW_AT_call_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.byte	0
+	.byte	0
+	.uleb128 0x8	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x31	# (DW_AT_abstract_origin)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.uleb128 0x2	# (DW_AT_location)
+	.uleb128 0x18	# (DW_FORM_exprloc)
+	.byte	0
+	.byte	0
+	.uleb128 0x9	# (abbrev code)
+	.uleb128 0x2e	# (TAG: DW_TAG_subprogram)
+	.byte	0x1	# DW_children_yes
+	.uleb128 0x3f	# (DW_AT_external)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0xe	# (DW_FORM_strp)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0x5	# (DW_FORM_data2)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x27	# (DW_AT_prototyped)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.uleb128 0x20	# (DW_AT_inline)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x34	# (DW_AT_artificial)
+	.uleb128 0x19	# (DW_FORM_flag_present)
+	.byte	0
+	.byte	0
+	.uleb128 0xa	# (abbrev code)
+	.uleb128 0x5	# (TAG: DW_TAG_formal_parameter)
+	.byte	0	# DW_children_no
+	.uleb128 0x3	# (DW_AT_name)
+	.uleb128 0x8	# (DW_FORM_string)
+	.uleb128 0x3a	# (DW_AT_decl_file)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x3b	# (DW_AT_decl_line)
+	.uleb128 0x5	# (DW_FORM_data2)
+	.uleb128 0x39	# (DW_AT_decl_column)
+	.uleb128 0xb	# (DW_FORM_data1)
+	.uleb128 0x49	# (DW_AT_type)
+	.uleb128 0x13	# (DW_FORM_ref4)
+	.byte	0
+	.byte	0
+	.byte	0
+	.section	.debug_aranges,"",@progbits
+	.long	0x2c	# Length of Address Ranges Info
+	.value	0x2	# DWARF aranges version
+	.long	.Ldebug_info0	# Offset of Compilation Unit Info
+	.byte	0x8	# Size of Address
+	.byte	0	# Size of Segment Descriptor
+	.value	0	# Pad to 16 byte boundary
+	.value	0
+	.quad	.Ltext0	# Address
+	.quad	.Letext0-.Ltext0	# Length
+	.quad	0
+	.quad	0
+	.section	.debug_line,"",@progbits
+.Ldebug_line0:
+	.section	.debug_str,"MS",@progbits,1
+.LASF2:
+	.string	"long long int"
+.LASF14:
+	.string	"value"
+.LASF19:
+	.string	"_ptwrite32"
+.LASF8:
+	.string	"unsigned int"
+.LASF17:
+	.string	"gdb/gdb/testsuite/gdb.btrace"
+.LASF18:
+	.string	"main"
+.LASF16:
+	.string	"ptwrite.c"
+.LASF4:
+	.string	"long unsigned int"
+.LASF10:
+	.string	"long long unsigned int"
+.LASF15:
+	.string	"GNU C17 10.3.1 20210422 (Red Hat 10.3.1-1) -mptwrite -mtune=generic -march=x86-64 -g"
+.LASF13:
+	.string	"ptwrite1"
+.LASF12:
+	.string	"ptwrite2"
+.LASF6:
+	.string	"unsigned char"
+.LASF1:
+	.string	"char"
+.LASF5:
+	.string	"long int"
+.LASF11:
+	.string	"double"
+.LASF7:
+	.string	"short unsigned int"
+.LASF9:
+	.string	"signed char"
+.LASF3:
+	.string	"float"
+.LASF0:
+	.string	"short int"
+	.ident	"GCC: (GNU) 10.3.1 20210422 (Red Hat 10.3.1-1)"
+	.section	.note.GNU-stack,"",@progbits
diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
index 0f6bb20b49c..99039e9e6c4 100644
--- a/gdb/testsuite/lib/gdb.exp
+++ b/gdb/testsuite/lib/gdb.exp
@@ -3687,6 +3687,80 @@ gdb_caching_proc skip_btrace_pt_tests {
     return $skip_btrace_tests
 }
 
+# Run a test on the target to see if it supports ptwrite instructions and
+# if GDB can decode ptwrite events.  Return 0 if so, 1 if it does not.
+
+gdb_caching_proc skip_btrace_ptw_tests {
+    global srcdir subdir gdb_prompt inferior_exited_re decimal
+
+    set me "skip_ptw_tests"
+    if { [skip_btrace_pt_tests] } {
+	verbose "$me:  target does not support btrace pt, returning 1" 2
+	return 1
+    }
+
+    set src {
+	int
+	main ()
+	{
+	  asm volatile ("PTWRITE %0;" : : "b"(0x42));
+	  return 0;
+	}
+    }
+
+    if {![gdb_simple_compile $me $src executable]} {
+        return 1
+    }
+
+    gdb_exit
+    gdb_start
+    gdb_reinitialize_dir $srcdir/$subdir
+    gdb_load "$obj"
+    if ![runto_main] {
+        return 1
+    }
+
+    gdb_test_no_output "record btrace pt" "$me: record btrace pt"
+
+    set skip_ptw_tests 2
+    gdb_test_multiple "next" "$me: next" {
+        -re -wrap ".*Illegal instruction.*" {
+            verbose -log "$me:  ptwrite instruction support not detected."
+            set skip_ptw_tests 2
+        }
+        -re -wrap ".*$inferior_exited_re normally.*" {
+            verbose -log "$me:  ptwrite support not detected."
+            set skip_ptw_tests 2
+        }
+        -re -wrap "$decimal.*(at|in).*" {
+            set skip_ptw_tests 1
+        }
+    }
+
+    if { $skip_ptw_tests == 1 } {
+	# Show the func-call-history to get the packet trace.
+	gdb_test "record function-call-history" ".*"
+
+	gdb_test_multiple "maintenance btrace packet-history 0,1000" \
+	    "$me: check decoding support" {
+	    -re  "ptw" {
+		verbose -log "$me:  ptwrite decoding support detected."
+		set skip_ptw_tests 0
+	    }
+	    -re ".*${gdb_prompt} $" {
+		verbose -log "$me:  ptwrite decoding support not detected."
+	    }
+	}
+    }
+
+    gdb_exit
+    remote_file build delete $obj
+
+    verbose "$me:  returning $skip_ptw_tests" 2
+    return $skip_ptw_tests
+}
+
+
 # Run a test on the target to see if it supports Aarch64 SVE hardware.
 # Return 0 if so, 1 if it does not.  Note this causes a restart of GDB.
 
diff --git a/gdbsupport/common.m4 b/gdbsupport/common.m4
index 0fed186ae1d..526f7d47c05 100644
--- a/gdbsupport/common.m4
+++ b/gdbsupport/common.m4
@@ -179,6 +179,8 @@ AC_DEFUN([GDB_AC_COMMON], [
       AC_CHECK_FUNCS(pt_insn_event)
       AC_CHECK_MEMBERS([struct pt_insn.enabled, struct pt_insn.resynced], [], [],
 		       [#include <intel-pt.h>])
+      AC_CHECK_MEMBERS([struct pt_event.variant.ptwrite], [], [],
+		       [#include <intel-pt.h>])
       LIBS=$save_LIBS
     fi
   fi
diff --git a/gdbsupport/config.in b/gdbsupport/config.in
index 577866c97b3..56a314195f3 100644
--- a/gdbsupport/config.in
+++ b/gdbsupport/config.in
@@ -235,6 +235,9 @@
 /* Define to 1 if you have the <string.h> header file. */
 #undef HAVE_STRING_H
 
+/* Define to 1 if `variant.ptwrite' is a member of `struct pt_event'. */
+#undef HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE
+
 /* Define to 1 if `enabled' is a member of `struct pt_insn'. */
 #undef HAVE_STRUCT_PT_INSN_ENABLED
 
diff --git a/gdbsupport/configure b/gdbsupport/configure
index 1b0752609e5..5b1d9bde715 100755
--- a/gdbsupport/configure
+++ b/gdbsupport/configure
@@ -9605,6 +9605,17 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+fi
+
+      ac_fn_c_check_member "$LINENO" "struct pt_event" "variant.ptwrite" "ac_cv_member_struct_pt_event_variant_ptwrite" "#include <intel-pt.h>
+"
+if test "x$ac_cv_member_struct_pt_event_variant_ptwrite" = xyes; then :
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_STRUCT_PT_EVENT_VARIANT_PTWRITE 1
+_ACEOF
+
+
 fi
 
       LIBS=$save_LIBS
-- 
2.34.3

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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history.
  2022-09-16 11:36 ` [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history Felix Willgerodt
@ 2022-09-16 11:43   ` Eli Zaretskii
  0 siblings, 0 replies; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 11:43 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:39 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> Print the auxiliary data when a btrace_insn of type BTRACE_INSN_AUX
> is encountered in the function-call-history.  Printing is
> active by default, it can be silenced with the /a modifier.
> 
> This patch is in preparation for the new ptwrite feature, which is based on
> auxiliary instructions.
> ---
>  gdb/btrace.h        |  6 +++++-
>  gdb/doc/gdb.texinfo |  5 +++--
>  gdb/record-btrace.c | 21 +++++++++++++++++++++
>  gdb/record.c        |  5 +++++
>  gdb/record.h        |  5 ++++-
>  5 files changed, 38 insertions(+), 4 deletions(-)

OK for the documentation part.

Thanks.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 01/10] btrace: Introduce auxiliary instructions.
  2022-09-16 11:36 ` [PATCH v6 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
@ 2022-09-16 11:46   ` Eli Zaretskii
  2022-09-16 14:01     ` Willgerodt, Felix
  0 siblings, 1 reply; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 11:46 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:37 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
> index 238a49b027d..0f769f5e24c 100644
> --- a/gdb/doc/gdb.texinfo
> +++ b/gdb/doc/gdb.texinfo
> @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86 (i386/amd64) running
>  GNU/Linux.  Process record and replay can be used both when native
>  debugging, and when remote debugging via @code{gdbserver}.
>  
> +When recording an inferior, @value{GDBN} may print additional auxiliary
> +information during stepping commands and commands displaying the execution
> +history.
> +
>  For architecture environments that support process record and replay,
>  @value{GDBN} provides the following commands:

"Auxiliary" and "additional" both say almost the same thing, so I'd
lose the "additional" part.

And I wonder whether this additional sentence really adds something
important to the text, since it doesn't say what and when will be
displayed in addition.  So maybe we don't need this addition at all?

Thanks.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-09-16 11:36 ` [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
@ 2022-09-16 11:49   ` Eli Zaretskii
  2022-09-16 14:02     ` Willgerodt, Felix
  2022-10-19 14:41   ` Metzger, Markus T
  1 sibling, 1 reply; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 11:49 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:43 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> +@item Qbtrace-conf:pt:ptwrite=@var{(yes|no)}
> +Indicate support for ptwrite packets.  This allows for backwards-
> +compatibility.

I think "ptwrite" should be in @code, since it's a program symbol.

Also, we leave hyphenation to TeX, so please don't break lines at the
hyphen, as you did with "backwards-compatibility".

> +Reply:
> +@table @samp
> +@item OK
> +The ptwrite config parameter has been set.
       ^^^^^^^
@code{ptwrite}

Thanks.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 06/10] python: Add clear() to gdb.Record.
  2022-09-16 11:36 ` [PATCH v6 06/10] python: Add clear() to gdb.Record Felix Willgerodt
@ 2022-09-16 11:49   ` Eli Zaretskii
  0 siblings, 0 replies; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 11:49 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:42 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> This function allows to clear the trace data from python, forcing to
> re-decode the trace for successive commands.
> ---
>  gdb/doc/python.texi                           |  5 +++++
>  gdb/python/py-record-btrace.c                 | 13 +++++++++++++
>  gdb/python/py-record-btrace.h                 |  3 +++
>  gdb/python/py-record.c                        | 16 ++++++++++++++++
>  gdb/testsuite/gdb.python/py-record-btrace.exp |  6 +++++-
>  5 files changed, 42 insertions(+), 1 deletion(-)

Thanks, the python.texi part is OK.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history.
  2022-09-16 11:36 ` [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history Felix Willgerodt
@ 2022-09-16 11:50   ` Eli Zaretskii
  0 siblings, 0 replies; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 11:50 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:38 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> Print the auxiliary data when a btrace_insn of type BTRACE_INSN_AUX
> is encountered in the instruction-history.  Printing is active by default,
> it can be silenced with the /a modifier.
> 
> This patch is in preparation for the new ptwrite feature, which is based on
> auxiliary instructions.
> ---
>  gdb/disasm-flags.h  |  1 +
>  gdb/doc/gdb.texinfo |  3 +++
>  gdb/record-btrace.c | 14 ++++++++++++++
>  gdb/record.c        |  5 +++++
>  4 files changed, 23 insertions(+)

OK for the gdb.texinfo part.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 11:36 ` [PATCH v6 10/10] btrace: Extend ptwrite event decoding Felix Willgerodt
@ 2022-09-16 12:01   ` Eli Zaretskii
  2022-09-16 14:02     ` Willgerodt, Felix
  0 siblings, 1 reply; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 12:01 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:46 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> --- a/gdb/NEWS
> +++ b/gdb/NEWS
> @@ -3,6 +3,12 @@
>  
>  *** Changes since GDB 12
>  
> +* GDB now supports printing of ptwrite payloads from the Intel Processor
> +  Trace during 'record instruction-history', 'record function-call-history',
> +  all stepping commands and in  Python.  Printing is customizable via a

Can we rephrase the "and in Python" part to be less confusing and
ambiguous?  (Also, there's an extraneous blank character there.)

> +If an inferior uses the instruction, @value{GDBN} by default inserts the
> +raw payload value as auxiliary information into the execution history.
> +Auxiliary information is by default printed during
> +@code{record instruction-history}, @code{record function-call-history},
> +and all stepping commands and is accessible in Python as a
                            ^
Comma missing there.

Btw, it sounds like this functionality is only available if GDB was
built with Python?  If so, I think other places where you mention this
in the manual should indicate that the feature is only available
through Python.

> +This function will be called with the ptwrite payload and PC as arguments
> +during trace decoding.                ^^^^^^^

@code{ptwrite}.

Also, please decide whether you want to use "ptwrite" in lower-case or
PTWRITE in upper-case, and please use the same convention everywhere
in the manual.

I also think talking about a "payload" obfuscates the description for
no good reason.  It's just a value of some data type supported by the
'ptwrite' instruction, right?  If so, I think "value" is much clearer
and less mysterious.

> +@findex gdb.ptwrite.register_filter
> +@defun register_filter (@var{filter})
> +Used to register the ptwrite filter.  The filter can be any callable
                        ^^^^^^^
@code{ptwrite}

> +@findex gdb.ptwrite.get_filter
> +@defun get_filter ()
> +Returns the currently active ptwrite filter function.

I think our style is to say "Return", not "Returns".

Thanks.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class.
  2022-09-16 11:36 ` [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class Felix Willgerodt
@ 2022-09-16 12:04   ` Eli Zaretskii
  0 siblings, 0 replies; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 12:04 UTC (permalink / raw)
  To: Felix Willgerodt; +Cc: gdb-patches, markus.t.metzger

> Date: Fri, 16 Sep 2022 13:36:41 +0200
> From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> 
> Auxiliary instructions are no real instructions and get their own object
> class, similar to gaps. gdb.Record.instruction_history is now possibly a
> list of gdb.RecordInstruction, gdb.RecordGap or gdb.RecordAuxiliary
> objects.
> 
> This patch is in preparation for the new ptwrite feature, which is based on
> auxiliary instructions.
> ---
>  gdb/doc/python.texi           | 13 +++++++
>  gdb/python/py-record-btrace.c | 35 +++++++++++------
>  gdb/python/py-record.c        | 73 ++++++++++++++++++++++++++++++++++-
>  gdb/python/py-record.h        |  3 ++
>  4 files changed, 111 insertions(+), 13 deletions(-)

Thanks, the python.texi part is OK.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 01/10] btrace: Introduce auxiliary instructions.
  2022-09-16 11:46   ` Eli Zaretskii
@ 2022-09-16 14:01     ` Willgerodt, Felix
  2022-09-16 14:18       ` Eli Zaretskii
  0 siblings, 1 reply; 33+ messages in thread
From: Willgerodt, Felix @ 2022-09-16 14:01 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb-patches, Metzger, Markus T

Thanks for reviewing this again.
In the cover letter I mentioned that basically all doc parts have already been
approved by you. As I didn't want to waste your time, having you review the
same text again and again. Is there a better way to let you know that you
approved this already? Or did you read the cover letter and intentionally reviewed it?
(Which is totally fine.)

I just don't want to waste your time, I do appreciate the feedback!

> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Freitag, 16. September 2022 13:46
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org; Metzger, Markus T
> <markus.t.metzger@intel.com>
> Subject: Re: [PATCH v6 01/10] btrace: Introduce auxiliary instructions.
> 
> > Date: Fri, 16 Sep 2022 13:36:37 +0200
> > From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> >
> > diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
> > index 238a49b027d..0f769f5e24c 100644
> > --- a/gdb/doc/gdb.texinfo
> > +++ b/gdb/doc/gdb.texinfo
> > @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86
> (i386/amd64) running
> >  GNU/Linux.  Process record and replay can be used both when native
> >  debugging, and when remote debugging via @code{gdbserver}.
> >
> > +When recording an inferior, @value{GDBN} may print additional auxiliary
> > +information during stepping commands and commands displaying the
> execution
> > +history.
> > +
> >  For architecture environments that support process record and replay,
> >  @value{GDBN} provides the following commands:
> 
> "Auxiliary" and "additional" both say almost the same thing, so I'd
> lose the "additional" part.

Thanks, I removed it locally.

> And I wonder whether this additional sentence really adds something
> important to the text, since it doesn't say what and when will be
> displayed in addition.  So maybe we don't need this addition at all?

The auxiliary instructions are designed to be generic. To be
able to use them for other features in the future. I think a "definition"
of the concept is needed, as it is user exposed via the /a modifiers later
in patch 2 and 3. Which are also documented and basically reference
this "definition" here.

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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-09-16 11:49   ` Eli Zaretskii
@ 2022-09-16 14:02     ` Willgerodt, Felix
  0 siblings, 0 replies; 33+ messages in thread
From: Willgerodt, Felix @ 2022-09-16 14:02 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb-patches, Metzger, Markus T


> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Freitag, 16. September 2022 13:49
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org; Metzger, Markus T
> <markus.t.metzger@intel.com>
> Subject: Re: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to
> btrace_config_pt.
> 
> > Date: Fri, 16 Sep 2022 13:36:43 +0200
> > From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> >
> > +@item Qbtrace-conf:pt:ptwrite=@var{(yes|no)}
> > +Indicate support for ptwrite packets.  This allows for backwards-
> > +compatibility.
> 
> I think "ptwrite" should be in @code, since it's a program symbol.
>

My intention was to use @code{PTWRITE} for the instruction. And lower case for
every ptwrite related "thing". E.g. the ptwrite packet, the ptwrite payload,
the ptwrite filter, the ptwrite auxiliary information etc.
But I will change it if that is preferred. I don't have a strong preference.

> Also, we leave hyphenation to TeX, so please don't break lines at the
> hyphen, as you did with "backwards-compatibility".

Thanks. I fixed it locally.

> > +Reply:
> > +@table @samp
> > +@item OK
> > +The ptwrite config parameter has been set.
>        ^^^^^^^
> @code{ptwrite}
> 

Here I am not so sure if it makes sense.
Here it is the config parameter in the RSP packet. Not the instruction.
I could wrap it in @samp{}? I see that other packets are wrapped like 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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 12:01   ` Eli Zaretskii
@ 2022-09-16 14:02     ` Willgerodt, Felix
  2022-09-16 14:21       ` Eli Zaretskii
  0 siblings, 1 reply; 33+ messages in thread
From: Willgerodt, Felix @ 2022-09-16 14:02 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb-patches, Metzger, Markus T

> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Freitag, 16. September 2022 14:02
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org; Metzger, Markus T
> <markus.t.metzger@intel.com>
> Subject: Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
> 
> > Date: Fri, 16 Sep 2022 13:36:46 +0200
> > From: Felix Willgerodt via Gdb-patches <gdb-patches@sourceware.org>
> >
> > --- a/gdb/NEWS
> > +++ b/gdb/NEWS
> > @@ -3,6 +3,12 @@
> >
> >  *** Changes since GDB 12
> >
> > +* GDB now supports printing of ptwrite payloads from the Intel Processor
> > +  Trace during 'record instruction-history', 'record function-call-history',
> > +  all stepping commands and in  Python.  Printing is customizable via a
> 
> Can we rephrase the "and in Python" part to be less confusing and
> ambiguous?  (Also, there's an extraneous blank character there.)

The NEWS file very often says "in Python".
But it is a fair point. How about "and in the Python API?"
I will fix the extra space.


> > +If an inferior uses the instruction, @value{GDBN} by default inserts the
> > +raw payload value as auxiliary information into the execution history.
> > +Auxiliary information is by default printed during
> > +@code{record instruction-history}, @code{record function-call-history},
> > +and all stepping commands and is accessible in Python as a
>                             ^
> Comma missing there.

Thanks, fixed locally.

> Btw, it sounds like this functionality is only available if GDB was
> built with Python?  If so, I think other places where you mention this
> in the manual should indicate that the feature is only available
> through Python.

That is true. Yet most of the documentation is in python.texi for that reason.
Do you think that is not enough?

The other documentation changes (outside of this patch) were for:
1) the btrace config parameter in RSP:
The new RSP parameter will be available regardless of Python.
I don't see a need or a good place to add anything here.
See patch 7.
2) about auxiliary instructions:
These are also python independent. That is by design, to be able
to use the infrastructure for other things in the future.
See patches 1-3.
3) New auxiliary instruction in Python. This is also in python.texi and clear
enough imo. See patch 5.
4) Document record.clear(). This is also only in python.texi and clear enough imo.
See patch 6.

If you think I should mention it in any of these, please let me know.
I don't see where it would make sense.

> > +This function will be called with the ptwrite payload and PC as arguments
> > +during trace decoding.                ^^^^^^^
> 
> @code{ptwrite}.
> 
> Also, please decide whether you want to use "ptwrite" in lower-case or
> PTWRITE in upper-case, and please use the same convention everywhere
> in the manual.

My intention was to use PTWRITE for the instruction. And lower case for
every ptwrite related "thing". E.g. the ptwrite packet, the ptwrite payload,
the ptwrite filter etc.
I will change it to @code{PTWRITE} everywhere.

> I also think talking about a "payload" obfuscates the description for
> no good reason.  It's just a value of some data type supported by the
> 'ptwrite' instruction, right?  If so, I think "value" is much clearer
> and less mysterious.

The Intel Software Developer Manual calls this value the "payload".
Mainly the chapter "Software Trace Instrumentation with PTWRITE".
I would rather have us stick to that name. Talking about just a "value"
or a "PTWRITE value" is more confusing in my eyes, as it isn't specific
and not the official name.

> > +@findex gdb.ptwrite.register_filter
> > +@defun register_filter (@var{filter})
> > +Used to register the ptwrite filter.  The filter can be any callable
>                         ^^^^^^^
> @code{ptwrite}

See my answer above.

> > +@findex gdb.ptwrite.get_filter
> > +@defun get_filter ()
> > +Returns the currently active ptwrite filter function.
> 
> I think our style is to say "Return", not "Returns".

Thanks, will change it.

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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 01/10] btrace: Introduce auxiliary instructions.
  2022-09-16 14:01     ` Willgerodt, Felix
@ 2022-09-16 14:18       ` Eli Zaretskii
  0 siblings, 0 replies; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 14:18 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches, markus.t.metzger

> From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>, "Metzger,
>  Markus T" <markus.t.metzger@intel.com>
> Date: Fri, 16 Sep 2022 14:01:57 +0000
> 
> Thanks for reviewing this again.
> In the cover letter I mentioned that basically all doc parts have already been
> approved by you. As I didn't want to waste your time, having you review the
> same text again and again. Is there a better way to let you know that you
> approved this already? Or did you read the cover letter and intentionally reviewed it?
> (Which is totally fine.)

I guess I missed that part, sorry.  And I didn't remember reviewing
this.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 14:02     ` Willgerodt, Felix
@ 2022-09-16 14:21       ` Eli Zaretskii
  2022-09-16 14:34         ` Willgerodt, Felix
  0 siblings, 1 reply; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 14:21 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches, markus.t.metzger

> From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>, "Metzger,
>  Markus T" <markus.t.metzger@intel.com>
> Date: Fri, 16 Sep 2022 14:02:07 +0000
> 
> > >  *** Changes since GDB 12
> > >
> > > +* GDB now supports printing of ptwrite payloads from the Intel Processor
> > > +  Trace during 'record instruction-history', 'record function-call-history',
> > > +  all stepping commands and in  Python.  Printing is customizable via a
> > 
> > Can we rephrase the "and in Python" part to be less confusing and
> > ambiguous?  (Also, there's an extraneous blank character there.)
> 
> The NEWS file very often says "in Python".
> But it is a fair point. How about "and in the Python API?"

I think it is better to use the same wording as you did in the manual.

> > Btw, it sounds like this functionality is only available if GDB was
> > built with Python?  If so, I think other places where you mention this
> > in the manual should indicate that the feature is only available
> > through Python.
> 
> That is true. Yet most of the documentation is in python.texi for that reason.
> Do you think that is not enough?

I was talking about stuff outside of python.texi.

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 14:21       ` Eli Zaretskii
@ 2022-09-16 14:34         ` Willgerodt, Felix
  2022-09-16 15:29           ` Eli Zaretskii
  0 siblings, 1 reply; 33+ messages in thread
From: Willgerodt, Felix @ 2022-09-16 14:34 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb-patches, Metzger, Markus T

> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Freitag, 16. September 2022 16:21
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org; Metzger, Markus T
> <markus.t.metzger@intel.com>
> Subject: Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
> 
> > From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
> > CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
> "Metzger,
> >  Markus T" <markus.t.metzger@intel.com>
> > Date: Fri, 16 Sep 2022 14:02:07 +0000
> >
> > > >  *** Changes since GDB 12
> > > >
> > > > +* GDB now supports printing of ptwrite payloads from the Intel
> Processor
> > > > +  Trace during 'record instruction-history', 'record function-call-history',
> > > > +  all stepping commands and in  Python.  Printing is customizable via a
> > >
> > > Can we rephrase the "and in Python" part to be less confusing and
> > > ambiguous?  (Also, there's an extraneous blank character there.)
> >
> > The NEWS file very often says "in Python".
> > But it is a fair point. How about "and in the Python API?"
> 
> I think it is better to use the same wording as you did in the manual.

Which is? I can't find anything similar.

> > > Btw, it sounds like this functionality is only available if GDB was
> > > built with Python?  If so, I think other places where you mention this
> > > in the manual should indicate that the feature is only available
> > > through Python.
> >
> > That is true. Yet most of the documentation is in python.texi for that
> reason.
> > Do you think that is not enough?
> 
> I was talking about stuff outside of python.texi.

I listed all the stuff outside of python.texi in my previous answer and mentioned why
I don't see any place where your suggestion would make sense.

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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 14:34         ` Willgerodt, Felix
@ 2022-09-16 15:29           ` Eli Zaretskii
  2022-09-19 11:23             ` Willgerodt, Felix
  0 siblings, 1 reply; 33+ messages in thread
From: Eli Zaretskii @ 2022-09-16 15:29 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches, markus.t.metzger

> From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>, "Metzger,
>  Markus T" <markus.t.metzger@intel.com>
> Date: Fri, 16 Sep 2022 14:34:39 +0000
> 
> > > > Can we rephrase the "and in Python" part to be less confusing and
> > > > ambiguous?  (Also, there's an extraneous blank character there.)
> > >
> > > The NEWS file very often says "in Python".
> > > But it is a fair point. How about "and in the Python API?"
> > 
> > I think it is better to use the same wording as you did in the manual.
> 
> Which is? I can't find anything similar.

 Auxiliary information is by default printed during
 @code{record instruction-history}, @code{record function-call-history},
 and all stepping commands and is accessible in Python as a
 @code{RecordAuxiliary} object.

> > > > Btw, it sounds like this functionality is only available if GDB was
> > > > built with Python?  If so, I think other places where you mention this
> > > > in the manual should indicate that the feature is only available
> > > > through Python.
> > >
> > > That is true. Yet most of the documentation is in python.texi for that
> > reason.
> > > Do you think that is not enough?
> > 
> > I was talking about stuff outside of python.texi.
> 
> I listed all the stuff outside of python.texi in my previous answer and mentioned why
> I don't see any place where your suggestion would make sense.

Here:

  --- a/gdb/doc/gdb.texinfo
  +++ b/gdb/doc/gdb.texinfo
  @@ -8017,8 +8017,9 @@ that function, the source lines for this instruction sequence (if the
   @code{/l} modifier is specified), and the instructions numbers that form
   the sequence (if the @code{/i} modifier is specified).  The function names
   are indented to reflect the call stack depth if the @code{/c} modifier is
  -specified.  The @code{/l}, @code{/i}, and @code{/c} modifiers can be given
  -together.
  +specified.  Printing auxiliary information is enabled by default and can be
  +omitted with the @code{/a} modifier.  The @code{/l}, @code{/i}, @code{/a},
  +and @code{/c} modifiers can be given together.

Here:

  --- a/gdb/doc/gdb.texinfo
  +++ b/gdb/doc/gdb.texinfo
  @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86 (i386/amd64) running
   GNU/Linux.  Process record and replay can be used both when native
   debugging, and when remote debugging via @code{gdbserver}.

  +When recording an inferior, @value{GDBN} may print additional auxiliary
  +information during stepping commands and commands displaying the execution
  +history.

And here:

  --- a/gdb/doc/gdb.texinfo
  +++ b/gdb/doc/gdb.texinfo
  @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86 (i386/amd64) running
   GNU/Linux.  Process record and replay can be used both when native
   debugging, and when remote debugging via @code{gdbserver}.

  +When recording an inferior, @value{GDBN} may print additional auxiliary
  +information during stepping commands and commands displaying the execution
  +history.


^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
  2022-09-16 15:29           ` Eli Zaretskii
@ 2022-09-19 11:23             ` Willgerodt, Felix
  0 siblings, 0 replies; 33+ messages in thread
From: Willgerodt, Felix @ 2022-09-19 11:23 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb-patches, Metzger, Markus T

> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Freitag, 16. September 2022 17:30
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org; Metzger, Markus T
> <markus.t.metzger@intel.com>
> Subject: Re: [PATCH v6 10/10] btrace: Extend ptwrite event decoding.
> 
> > From: "Willgerodt, Felix" <felix.willgerodt@intel.com>
> > CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
> "Metzger,
> >  Markus T" <markus.t.metzger@intel.com>
> > Date: Fri, 16 Sep 2022 14:34:39 +0000
> >
> > > > > Can we rephrase the "and in Python" part to be less confusing and
> > > > > ambiguous?  (Also, there's an extraneous blank character there.)
> > > >
> > > > The NEWS file very often says "in Python".
> > > > But it is a fair point. How about "and in the Python API?"
> > >
> > > I think it is better to use the same wording as you did in the manual.
> >
> > Which is? I can't find anything similar.
> 
>  Auxiliary information is by default printed during
>  @code{record instruction-history}, @code{record function-call-history},
>  and all stepping commands and is accessible in Python as a
>  @code{RecordAuxiliary} object.

Thanks! I will change it in the next version to:

" 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."

> > > > > Btw, it sounds like this functionality is only available if GDB was
> > > > > built with Python?  If so, I think other places where you mention this
> > > > > in the manual should indicate that the feature is only available
> > > > > through Python.
> > > >
> > > > That is true. Yet most of the documentation is in python.texi for that
> > > reason.
> > > > Do you think that is not enough?
> > >
> > > I was talking about stuff outside of python.texi.
> >
> > I listed all the stuff outside of python.texi in my previous answer and
> mentioned why
> > I don't see any place where your suggestion would make sense.
> 
> Here:
> 
>   --- a/gdb/doc/gdb.texinfo
>   +++ b/gdb/doc/gdb.texinfo
>   @@ -8017,8 +8017,9 @@ that function, the source lines for this instruction
> sequence (if the
>    @code{/l} modifier is specified), and the instructions numbers that form
>    the sequence (if the @code{/i} modifier is specified).  The function names
>    are indented to reflect the call stack depth if the @code{/c} modifier is
>   -specified.  The @code{/l}, @code{/i}, and @code{/c} modifiers can be
> given
>   -together.
>   +specified.  Printing auxiliary information is enabled by default and can be
>   +omitted with the @code{/a} modifier.  The @code{/l}, @code{/i},
> @code{/a},
>   +and @code{/c} modifiers can be given together.
> 
> Here:
> 
>   --- a/gdb/doc/gdb.texinfo
>   +++ b/gdb/doc/gdb.texinfo
>   @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86
> (i386/amd64) running
>    GNU/Linux.  Process record and replay can be used both when native
>    debugging, and when remote debugging via @code{gdbserver}.
> 
>   +When recording an inferior, @value{GDBN} may print additional auxiliary
>   +information during stepping commands and commands displaying the
> execution
>   +history.
> 
> And here:
> 
>   --- a/gdb/doc/gdb.texinfo
>   +++ b/gdb/doc/gdb.texinfo
>   @@ -7545,6 +7545,10 @@ Moxie, PowerPC, PowerPC64, S/390, and x86
> (i386/amd64) running
>    GNU/Linux.  Process record and replay can be used both when native
>    debugging, and when remote debugging via @code{gdbserver}.
> 
>   +When recording an inferior, @value{GDBN} may print additional auxiliary
>   +information during stepping commands and commands displaying the
> execution
>   +history.

All of the three are about auxiliary information. As stated previously, that is a
generic infrastructure that will be available independent of Python.
Ptwrite might be the only user right now, but afaik there has
already been discussion about using that infrastructure for some other
feature on the mailing list. Therefore, I still wouldn't add anything to any of
these three.

Regards,
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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-09-16 11:36 ` [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
  2022-09-16 11:49   ` Eli Zaretskii
@ 2022-10-19 14:41   ` Metzger, Markus T
  2022-10-20 11:49     ` Willgerodt, Felix
  1 sibling, 1 reply; 33+ messages in thread
From: Metzger, Markus T @ 2022-10-19 14:41 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches

Hello Felix,

>This enables gdb and gdbserver to communicate about ptwrite support.  If
>ptwrite support would be enabled unconditionally, GDBs with older libipt
>versions would break.
>---
> gdb/btrace.c                 |  8 +++++++-
> gdb/doc/gdb.texinfo          | 21 +++++++++++++++++++++
> gdb/features/btrace-conf.dtd |  1 +
> gdb/remote.c                 | 30 ++++++++++++++++++++++++++++++
> gdbserver/linux-low.cc       |  2 ++
> gdbserver/server.cc          | 18 ++++++++++++++++++
> gdbsupport/btrace-common.h   |  6 ++++++
> 7 files changed, 85 insertions(+), 1 deletion(-)

Looks good to me with one nit:

>--- a/gdb/features/btrace-conf.dtd
>+++ b/gdb/features/btrace-conf.dtd
>@@ -12,3 +12,4 @@
>
> <!ELEMENT pt	EMPTY>
> <!ATTLIST pt	size	CDATA	#IMPLIED>
>+<!ATTLIST pt	ptwrite	(yes | no) #IMPLIED>

Should there be a space between 'ptwrite' and '(yes|no)'?

>diff --git a/gdbserver/linux-low.cc b/gdbserver/linux-low.cc
>index 2f71360d3bd..1333ec13dbd 100644
>--- a/gdbserver/linux-low.cc
>+++ b/gdbserver/linux-low.cc
>@@ -6806,6 +6806,8 @@ linux_process_target::read_btrace_conf (const
>btrace_target_info *tinfo,
> 	case BTRACE_FORMAT_PT:
> 	  buffer_xml_printf (buffer, "<pt");
> 	  buffer_xml_printf (buffer, " size=\"0x%x\"", conf->pt.size);
>+	  const char *ptw = conf->pt.ptwrite ? "yes" : "no";
>+	  buffer_xml_printf (buffer, " ptwrite=\"%s\"", ptw);

We don't really need the PTW variable.

thanks,
markus.
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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration.
  2022-09-16 11:36 ` [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration Felix Willgerodt
@ 2022-10-19 14:41   ` Metzger, Markus T
  2022-10-20 11:49     ` Willgerodt, Felix
  0 siblings, 1 reply; 33+ messages in thread
From: Metzger, Markus T @ 2022-10-19 14:41 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches

Hello Felix,

>With this patch a default ptwrite filter is registered upon start of GDB.
>It prints the plain ptwrite payload as hex.  The default filter can be
>overwritten by registering a custom filter in python or by registering
>"None", for no output at all.  Registering a filter function creates per
>thread copies to allow unique internal states per thread.
>---
> gdb/btrace.c                   |  4 ++
> gdb/btrace.h                   |  8 ++++
> gdb/data-directory/Makefile.in |  1 +
> gdb/extension-priv.h           |  5 ++
> gdb/extension.c                | 13 +++++
> gdb/extension.h                |  3 ++
> gdb/guile/guile.c              |  1 +
> gdb/python/lib/gdb/ptwrite.py  | 80 +++++++++++++++++++++++++++++++
> gdb/python/py-record-btrace.c  | 88 ++++++++++++++++++++++++++++++++++
> gdb/python/py-record-btrace.h  |  8 ++++
> gdb/python/python-internal.h   |  3 ++
> gdb/python/python.c            |  2 +
> 12 files changed, 216 insertions(+)
> create mode 100644 gdb/python/lib/gdb/ptwrite.py

Looks good to me with one small nit and a question.

>+  /* Function pointer to the ptwrite callback.  Returns the string returned
>+     by the ptwrite filter function.  */
>+  std::string (*ptw_callback_fun) (const uint64_t payload, const uint64_t ip,
>+				   const void *ptw_context);

This should probably be initialized to nullptr.

>+def get_filter():
>+    """Returns the filters of the current thread."""
>+    # The key is of this format to enable an per-inferior cleanup when an
>+    # inferior exits.
>+    key = f"{gdb.selected_inferior().pid}.{gdb.selected_thread().ptid[1]}"
>+
>+    # Create a new filter for new threads.
>+    if key not in _ptwrite_filter.keys():
>+        _ptwrite_filter[key] = deepcopy(_ptwrite_filter["global"])

Do we even need this "global" filter in the dict?  Could we just use the
default_filter, here?

regards,
markus.
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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 00/10] Extensions for PTWRITE
  2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
                   ` (9 preceding siblings ...)
  2022-09-16 11:36 ` [PATCH v6 10/10] btrace: Extend ptwrite event decoding Felix Willgerodt
@ 2022-10-19 14:45 ` Metzger, Markus T
  10 siblings, 0 replies; 33+ messages in thread
From: Metzger, Markus T @ 2022-10-19 14:45 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches

Hello Felix,

>The documentation parts basically didn't change since Eli approved them.
>I only made a very minor documentation change in patch 7 to reflect the
>change to a boolean.
>
>Here is the status of each patch:
>1: Approved.
>2: Fixed minor formatting comment.
>3: Renamed a define based on comments from Markus.
>4: Approved. Only fixed a nit.
>5: Fixed minor formatting comment.
>6: Approved by Markus, with a request that other maintainers might want
>   to look at the Python parts.
>7: Changed Qbtrace-conf:pt:ptwrite to a boolean.
>8: Approved. Only fixed a nit.
>9: Reworked GIL, error messages and DECREF logic in py-record-btrace.c.
>   Reworked ptwrite.py according to Markus comments.
>10: Approved. Only fixed a nit.

I re-reviewed patches 3, 7, and 9.  Btrace bits are OK with the remaining
comments addressed.

For the python bits, you need a global maintainer's approval.

regards,
markus.
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


^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-10-19 14:41   ` Metzger, Markus T
@ 2022-10-20 11:49     ` Willgerodt, Felix
  2022-10-20 12:31       ` Metzger, Markus T
  0 siblings, 1 reply; 33+ messages in thread
From: Willgerodt, Felix @ 2022-10-20 11:49 UTC (permalink / raw)
  To: Metzger, Markus T; +Cc: gdb-patches

> -----Original Message-----
> From: Metzger, Markus T <markus.t.metzger@intel.com>
> Sent: Mittwoch, 19. Oktober 2022 16:42
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org
> Subject: RE: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to
> btrace_config_pt.
> 
> Hello Felix,
> 
> >This enables gdb and gdbserver to communicate about ptwrite support.  If
> >ptwrite support would be enabled unconditionally, GDBs with older libipt
> >versions would break.
> >---
> > gdb/btrace.c                 |  8 +++++++-
> > gdb/doc/gdb.texinfo          | 21 +++++++++++++++++++++
> > gdb/features/btrace-conf.dtd |  1 +
> > gdb/remote.c                 | 30 ++++++++++++++++++++++++++++++
> > gdbserver/linux-low.cc       |  2 ++
> > gdbserver/server.cc          | 18 ++++++++++++++++++
> > gdbsupport/btrace-common.h   |  6 ++++++
> > 7 files changed, 85 insertions(+), 1 deletion(-)
> 
> Looks good to me with one nit:
> 
> >--- a/gdb/features/btrace-conf.dtd
> >+++ b/gdb/features/btrace-conf.dtd
> >@@ -12,3 +12,4 @@
> >
> > <!ELEMENT pt	EMPTY>
> > <!ATTLIST pt	size	CDATA	#IMPLIED>
> >+<!ATTLIST pt	ptwrite	(yes | no) #IMPLIED>
> 
> Should there be a space between 'ptwrite' and '(yes|no)'?

There is a tab there. That matches with the alignment of the file.
The other lines also have tabs.
I am not quite sure what you want changed.

> >diff --git a/gdbserver/linux-low.cc b/gdbserver/linux-low.cc
> >index 2f71360d3bd..1333ec13dbd 100644
> >--- a/gdbserver/linux-low.cc
> >+++ b/gdbserver/linux-low.cc
> >@@ -6806,6 +6806,8 @@ linux_process_target::read_btrace_conf (const
> >btrace_target_info *tinfo,
> > 	case BTRACE_FORMAT_PT:
> > 	  buffer_xml_printf (buffer, "<pt");
> > 	  buffer_xml_printf (buffer, " size=\"0x%x\"", conf->pt.size);
> >+	  const char *ptw = conf->pt.ptwrite ? "yes" : "no";
> >+	  buffer_xml_printf (buffer, " ptwrite=\"%s\"", ptw);
> 
> We don't really need the PTW variable.

Right. I will inline it.

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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration.
  2022-10-19 14:41   ` Metzger, Markus T
@ 2022-10-20 11:49     ` Willgerodt, Felix
  2022-10-20 12:33       ` Metzger, Markus T
  0 siblings, 1 reply; 33+ messages in thread
From: Willgerodt, Felix @ 2022-10-20 11:49 UTC (permalink / raw)
  To: Metzger, Markus T; +Cc: gdb-patches

> -----Original Message-----
> From: Metzger, Markus T <markus.t.metzger@intel.com>
> Sent: Mittwoch, 19. Oktober 2022 16:42
> To: Willgerodt, Felix <felix.willgerodt@intel.com>
> Cc: gdb-patches@sourceware.org
> Subject: RE: [PATCH v6 09/10] btrace, python: Enable ptwrite filter
> registration.
> 
> Hello Felix,
> 
> >With this patch a default ptwrite filter is registered upon start of GDB.
> >It prints the plain ptwrite payload as hex.  The default filter can be
> >overwritten by registering a custom filter in python or by registering
> >"None", for no output at all.  Registering a filter function creates per
> >thread copies to allow unique internal states per thread.
> >---
> > gdb/btrace.c                   |  4 ++
> > gdb/btrace.h                   |  8 ++++
> > gdb/data-directory/Makefile.in |  1 +
> > gdb/extension-priv.h           |  5 ++
> > gdb/extension.c                | 13 +++++
> > gdb/extension.h                |  3 ++
> > gdb/guile/guile.c              |  1 +
> > gdb/python/lib/gdb/ptwrite.py  | 80
> +++++++++++++++++++++++++++++++
> > gdb/python/py-record-btrace.c  | 88
> ++++++++++++++++++++++++++++++++++
> > gdb/python/py-record-btrace.h  |  8 ++++
> > gdb/python/python-internal.h   |  3 ++
> > gdb/python/python.c            |  2 +
> > 12 files changed, 216 insertions(+)
> > create mode 100644 gdb/python/lib/gdb/ptwrite.py
> 
> Looks good to me with one small nit and a question.
> 
> >+  /* Function pointer to the ptwrite callback.  Returns the string returned
> >+     by the ptwrite filter function.  */
> >+  std::string (*ptw_callback_fun) (const uint64_t payload, const uint64_t ip,
> >+				   const void *ptw_context);
> 
> This should probably be initialized to nullptr.

Right, I will do that.

> >+def get_filter():
> >+    """Returns the filters of the current thread."""
> >+    # The key is of this format to enable an per-inferior cleanup when an
> >+    # inferior exits.
> >+    key = f"{gdb.selected_inferior().pid}.{gdb.selected_thread().ptid[1]}"
> >+
> >+    # Create a new filter for new threads.
> >+    if key not in _ptwrite_filter.keys():
> >+        _ptwrite_filter[key] = deepcopy(_ptwrite_filter["global"])
> 
> Do we even need this "global" filter in the dict?  Could we just use the
> default_filter, here?

I don't see how. In case a custom filter is registered and
a new thread spawns, we need "global". We can't copy default_filter and
we can't copy from other threads, as they might have internal state already.

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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt.
  2022-10-20 11:49     ` Willgerodt, Felix
@ 2022-10-20 12:31       ` Metzger, Markus T
  0 siblings, 0 replies; 33+ messages in thread
From: Metzger, Markus T @ 2022-10-20 12:31 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches

Hello Felix,

>> > <!ELEMENT pt	EMPTY>
>> > <!ATTLIST pt	size	CDATA	#IMPLIED>
>> >+<!ATTLIST pt	ptwrite	(yes | no) #IMPLIED>
>>
>> Should there be a space between 'ptwrite' and '(yes|no)'?
>
>There is a tab there. That matches with the alignment of the file.
>The other lines also have tabs.
>I am not quite sure what you want changed.

It looks in the email like it was one word 'ptwrite(yes' without any separation.

regards,
markus.
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

^ permalink raw reply	[flat|nested] 33+ messages in thread

* RE: [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration.
  2022-10-20 11:49     ` Willgerodt, Felix
@ 2022-10-20 12:33       ` Metzger, Markus T
  0 siblings, 0 replies; 33+ messages in thread
From: Metzger, Markus T @ 2022-10-20 12:33 UTC (permalink / raw)
  To: Willgerodt, Felix; +Cc: gdb-patches

Hello Felix,

>> >+def get_filter():
>> >+    """Returns the filters of the current thread."""
>> >+    # The key is of this format to enable an per-inferior cleanup when an
>> >+    # inferior exits.
>> >+    key = f"{gdb.selected_inferior().pid}.{gdb.selected_thread().ptid[1]}"
>> >+
>> >+    # Create a new filter for new threads.
>> >+    if key not in _ptwrite_filter.keys():
>> >+        _ptwrite_filter[key] = deepcopy(_ptwrite_filter["global"])
>>
>> Do we even need this "global" filter in the dict?  Could we just use the
>> default_filter, here?
>
>I don't see how. In case a custom filter is registered and
>a new thread spawns, we need "global". We can't copy default_filter and
>we can't copy from other threads, as they might have internal state already.

That was the question whether we could copy default_filter.

regards,
markus.
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

^ permalink raw reply	[flat|nested] 33+ messages in thread

end of thread, other threads:[~2022-10-20 12:33 UTC | newest]

Thread overview: 33+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-09-16 11:36 [PATCH v6 00/10] Extensions for PTWRITE Felix Willgerodt
2022-09-16 11:36 ` [PATCH v6 01/10] btrace: Introduce auxiliary instructions Felix Willgerodt
2022-09-16 11:46   ` Eli Zaretskii
2022-09-16 14:01     ` Willgerodt, Felix
2022-09-16 14:18       ` Eli Zaretskii
2022-09-16 11:36 ` [PATCH v6 02/10] btrace: Enable auxiliary instructions in record instruction-history Felix Willgerodt
2022-09-16 11:50   ` Eli Zaretskii
2022-09-16 11:36 ` [PATCH v6 03/10] btrace: Enable auxiliary instructions in record function-call-history Felix Willgerodt
2022-09-16 11:43   ` Eli Zaretskii
2022-09-16 11:36 ` [PATCH v6 04/10] btrace: Handle stepping and goto for auxiliary instructions Felix Willgerodt
2022-09-16 11:36 ` [PATCH v6 05/10] python: Introduce gdb.RecordAuxiliary class Felix Willgerodt
2022-09-16 12:04   ` Eli Zaretskii
2022-09-16 11:36 ` [PATCH v6 06/10] python: Add clear() to gdb.Record Felix Willgerodt
2022-09-16 11:49   ` Eli Zaretskii
2022-09-16 11:36 ` [PATCH v6 07/10] btrace, gdbserver: Add ptwrite to btrace_config_pt Felix Willgerodt
2022-09-16 11:49   ` Eli Zaretskii
2022-09-16 14:02     ` Willgerodt, Felix
2022-10-19 14:41   ` Metzger, Markus T
2022-10-20 11:49     ` Willgerodt, Felix
2022-10-20 12:31       ` Metzger, Markus T
2022-09-16 11:36 ` [PATCH v6 08/10] btrace, linux: Enable ptwrite packets Felix Willgerodt
2022-09-16 11:36 ` [PATCH v6 09/10] btrace, python: Enable ptwrite filter registration Felix Willgerodt
2022-10-19 14:41   ` Metzger, Markus T
2022-10-20 11:49     ` Willgerodt, Felix
2022-10-20 12:33       ` Metzger, Markus T
2022-09-16 11:36 ` [PATCH v6 10/10] btrace: Extend ptwrite event decoding Felix Willgerodt
2022-09-16 12:01   ` Eli Zaretskii
2022-09-16 14:02     ` Willgerodt, Felix
2022-09-16 14:21       ` Eli Zaretskii
2022-09-16 14:34         ` Willgerodt, Felix
2022-09-16 15:29           ` Eli Zaretskii
2022-09-19 11:23             ` Willgerodt, Felix
2022-10-19 14:45 ` [PATCH v6 00/10] Extensions for PTWRITE Metzger, Markus T

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).