Re: [PATCH] Change handling of '!' operator in stap probes

[Sergio Durigan Junior via Gdb-patches <gdb-patches@sourceware.org>]

On Sunday, January 03 2020, Sergio Durigan Junior via Gdb-patches wrote:

> On Thursday, December 31 2020, Tom Tromey wrote:
>
>> While working on the expression rewrite, I noticed that the '!'
>> operator is handled as if it were a binary operator.  However, this
>> seems incorrect.  Instead, I think it should be handled alongside the
>> other unary operators in stap_parse_single_operand.
>>
>> This patch makes this change.

Ping.

>
> Thanks for the patch, Tom.
>
> This code was written based on the information contained here:
>
>   https://sourceware.org/binutils/docs/as/Infix-Ops.html
>
> You can see that it says:
>
>   !
>     Bitwise Or Not
>
> The problem is that, when we examine gas' source code, we see that it
> doesn't treat '!' as "bitwise or not" (fortunately!), but as the
> "logical not":
>
>   https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gas/symbols.c;h=f6e4b718ef719eef8a7e6e29dbadf3605058403d;hb=HEAD#l1128
>
> OK, this means that our stap-probe.c is not completely wrong: the enum
> should indeed be named UNOP_LOGICAL_NOT.
>
> Now, when we check to see if the precedence assigned to UNOP_LOGICAL_NOT
> is the right one, we see that it is:
>
>   https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gas/expr.c;h=1b420d928cde50456af0137529b8215294ba521d;hb=HEAD#l1481
>
>
>   /* Rank	Examples
>      0	operand, (expression)
>      1	||
>      2	&&
>      3	== <> < <= >= >
>      4	+ -
>      5	used for * / % in MRI mode
>      6	& ^ ! |
>   ...
>
> Huh, interesting.  Anyway, this whole investigation (plus the fact that,
> as you mentioned, we don't have tests to exercise this part of the code)
> made me uncover a few latent bugs on the stap parser code, so I'm
> sending a patch that incorporates part of your changes and also fixes
> the issues I found.  As a bonus, I'm proposing a new testcase
> (amd64-only) to make sure that we expand the coverage of this code.
>
> WDYT?
>
> -- 
> Sergio
> GPG key ID: 237A 54B1 0287 28BF 00EF  31F4 D0EB 7628 65FC 5E36
> Please send encrypted e-mail if possible
> https://sergiodj.net/
>
> From da02db7cacef62ce085d5715b1339b24d8f2260c Mon Sep 17 00:00:00 2001
> From: Sergio Durigan Junior <sergiodj@sergiodj.net>
> Date: Sun, 3 Jan 2021 02:42:52 -0500
> Subject: [PATCH] Fix a few stap parser issues and add a new test for probe
>  expressions
>
> The creation of this patch was motivated by Tom's "Change handling of
> '!' operator in stap probes" patch.
>
> While reviewing his patch, I stumbled upon a few issues with the stap
> expression parser.  They are:
>
> - As it turns out, even with Tom's patch applied the parser doesn't
>   properly handle the '!' operator.  The underlying issue was the fact
>   that stap_parse_argument_conditionally also needed to be patched in
>   order to recognize '!' as an operator that is part of a single
>   operand, and parse it accordingly.
>
> - While writing the testcase I'm proposing on this patch, I found that
>   parenthesized sub-expressions were not being parsed correctly when
>   there was another term after them.  For example:
>
>     1 - (2 + 3) + 4
>
>   In this case, the parser was considering "1" to be the left-side of
>   the expression, and "(2 + 3) + 4" to be the right-side.  The patch
>   fixes the parser by making it identify whether a parenthesized
>   sub-expression has just been parsed, and act accordingly.
>
> I've tested this on my Debian testing amd64, and everything seems OK.
>
> gdb/ChangeLog:
> 2021-01-03  Sergio Durigan Junior  <sergiodj@sergiodj.net>
> 	    Tom Tromey <tom@tromey.com>
>
> 	* stap-probe.c (stap_parse_single_operand): Handle '!'
> 	operator.
> 	(stap_parse_argument_conditionally): Likewise.
> 	Skip spaces after processing open-parenthesis sub-expression.
> 	(stap_parse_argument_1): Skip spaces after call to
> 	stap_parse_argument_conditionally.
> 	Handle case when right-side expression is a parenthesized
> 	sub-expression.
> 	Skip spaces after call to stap_parse_argument_1.
>
> gdb/testsuite/ChangeLog:
> 2021-01-03  Sergio Durigan Junior  <sergiodj@sergiodj.net>
>
> 	* gdb.arch/amd64-stap-expressions.S: New file.
> 	* gdb.arch/amd64-stap-expressions.exp: New file.
> ---
>  gdb/stap-probe.c                              | 31 +++++++--
>  .../gdb.arch/amd64-stap-expressions.S         | 43 ++++++++++++
>  .../gdb.arch/amd64-stap-expressions.exp       | 68 +++++++++++++++++++
>  3 files changed, 137 insertions(+), 5 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-stap-expressions.S
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-stap-expressions.exp
>
> diff --git a/gdb/stap-probe.c b/gdb/stap-probe.c
> index c2ddd047f6..224dd5714f 100644
> --- a/gdb/stap-probe.c
> +++ b/gdb/stap-probe.c
> @@ -870,7 +870,7 @@ stap_parse_single_operand (struct stap_parse_info *p)
>        return;
>      }
>  
> -  if (*p->arg == '-' || *p->arg == '~' || *p->arg == '+')
> +  if (*p->arg == '-' || *p->arg == '~' || *p->arg == '+' || *p->arg == '!')
>      {
>        char c = *p->arg;
>        /* We use this variable to do a lookahead.  */
> @@ -924,6 +924,8 @@ stap_parse_single_operand (struct stap_parse_info *p)
>  	    write_exp_elt_opcode (&p->pstate, UNOP_NEG);
>  	  else if (c == '~')
>  	    write_exp_elt_opcode (&p->pstate, UNOP_COMPLEMENT);
> +	  else if (c == '!')
> +	    write_exp_elt_opcode (&p->pstate, UNOP_LOGICAL_NOT);
>  	}
>      }
>    else if (isdigit (*p->arg))
> @@ -1012,7 +1014,7 @@ stap_parse_argument_conditionally (struct stap_parse_info *p)
>  {
>    gdb_assert (gdbarch_stap_is_single_operand_p (p->gdbarch));
>  
> -  if (*p->arg == '-' || *p->arg == '~' || *p->arg == '+' /* Unary.  */
> +  if (*p->arg == '-' || *p->arg == '~' || *p->arg == '+' || *p->arg == '!'
>        || isdigit (*p->arg)
>        || gdbarch_stap_is_single_operand (p->gdbarch, p->arg))
>      stap_parse_single_operand (p);
> @@ -1027,11 +1029,12 @@ stap_parse_argument_conditionally (struct stap_parse_info *p)
>  
>        stap_parse_argument_1 (p, 0, STAP_OPERAND_PREC_NONE);
>  
> -      --p->inside_paren_p;
> +      p->arg = skip_spaces (p->arg);
>        if (*p->arg != ')')
> -	error (_("Missign close-paren on expression `%s'."),
> +	error (_("Missign close-parenthesis on expression `%s'."),
>  	       p->saved_arg);
>  
> +      --p->inside_paren_p;
>        ++p->arg;
>        if (p->inside_paren_p)
>  	p->arg = skip_spaces (p->arg);
> @@ -1067,6 +1070,9 @@ stap_parse_argument_1 (struct stap_parse_info *p, bool has_lhs,
>        stap_parse_argument_conditionally (p);
>      }
>  
> +  if (p->inside_paren_p)
> +    p->arg = skip_spaces (p->arg);
> +
>    /* Start to parse the right-side, and to "join" left and right sides
>       depending on the operation specified.
>  
> @@ -1104,8 +1110,21 @@ stap_parse_argument_1 (struct stap_parse_info *p, bool has_lhs,
>        if (p->inside_paren_p)
>  	p->arg = skip_spaces (p->arg);
>  
> -      /* Parse the right-side of the expression.  */
> +      /* Parse the right-side of the expression.
> +
> +	 We save whether the right-side is a parenthesized
> +	 subexpression because, if it is, we will have to finish
> +	 processing this part of the expression before continuing.  */
> +      bool paren_subexp = *p->arg == '(';
> +
>        stap_parse_argument_conditionally (p);
> +      if (p->inside_paren_p)
> +	p->arg = skip_spaces (p->arg);
> +      if (paren_subexp)
> +	{
> +	  write_exp_elt_opcode (&p->pstate, opcode);
> +	  continue;
> +	}
>  
>        /* While we still have operators, try to parse another
>  	 right-side, but using the current right-side as a left-side.  */
> @@ -1130,6 +1149,8 @@ stap_parse_argument_1 (struct stap_parse_info *p, bool has_lhs,
>  	  /* Parse the right-side of the expression, but since we already
>  	     have a left-side at this point, set `has_lhs' to 1.  */
>  	  stap_parse_argument_1 (p, 1, lookahead_prec);
> +	  if (p->inside_paren_p)
> +	    p->arg = skip_spaces (p->arg);
>  	}
>  
>        write_exp_elt_opcode (&p->pstate, opcode);
> diff --git a/gdb/testsuite/gdb.arch/amd64-stap-expressions.S b/gdb/testsuite/gdb.arch/amd64-stap-expressions.S
> new file mode 100644
> index 0000000000..76a47aa9b5
> --- /dev/null
> +++ b/gdb/testsuite/gdb.arch/amd64-stap-expressions.S
> @@ -0,0 +1,43 @@
> +/* Copyright (C) 2021 Free Software Foundation, Inc.
> +
> +   This file is part of GDB.
> +
> +   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 <sys/sdt.h>
> +
> +	.file	"amd64-stap-expressions.S"
> +	.text
> +	.globl	main
> +main:
> +	/* We use a nop here because we don't want the first probe to
> +	   be placed at the same location as the main label.  */
> +	nop
> +
> +	/* Single operands.  */
> +	STAP_PROBE1(probe, log_neg, 8@!($0+$1))
> +	STAP_PROBE1(probe, minus, -8@-($3+$4))
> +	STAP_PROBE1(probe, bit_neg, -8@~$22)
> +
> +	/* Arithmetic expressions.  */
> +	STAP_PROBE1(probe, plus1, 8@$3+($10-$8)-$1)
> +	STAP_PROBE1(probe, plus2, 8@$100-(  ($8+$10) -$50)+$3)
> +	STAP_PROBE1(probe, plus3, 8@$100-(($8+$10)-$50)+((($8 - $9) + $40) - $4)+$4)
> +
> +	/* Bitwise expressions.  */
> +	STAP_PROBE1(probe, and, 8@$128&$128)
> +	STAP_PROBE1(probe, or, 8@$8|$4)
> +
> +	xor	%rax,%rax
> +	ret
> diff --git a/gdb/testsuite/gdb.arch/amd64-stap-expressions.exp b/gdb/testsuite/gdb.arch/amd64-stap-expressions.exp
> new file mode 100644
> index 0000000000..5e3cb60a9e
> --- /dev/null
> +++ b/gdb/testsuite/gdb.arch/amd64-stap-expressions.exp
> @@ -0,0 +1,68 @@
> +# 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/>.
> +
> +standard_testfile ".S"
> +
> +if { ![istarget "x86_64-*-*"] || ![is_lp64_target] } {
> +    verbose "Skipping $testfile.exp"
> +    return
> +}
> +
> +if { [prepare_for_testing "failed to prepare" $testfile $srcfile] } {
> +    return -1
> +}
> +
> +# Helper procedure to go to probe NAME
> +
> +proc goto_probe { name } {
> +    global decimal hex
> +
> +    gdb_test "break -pstap $name" "Breakpoint $decimal at $hex"
> +    gdb_test "continue" "Breakpoint $decimal, main \\(\\) at .*\r\n.*STAP_PROBE1.*${name},.*\\)"
> +}
> +
> +# Helper procedure to test the probe's argument
> +
> +proc test_probe_value { value } {
> +    gdb_test "print \$_probe_argc" "= 1"
> +    gdb_test "print \$_probe_arg0" "= $value"
> +}
> +
> +if { ![runto_main] } {
> +    return -1
> +}
> +
> +# Name and expected value for each probe.
> +set probe_names_and_values {
> +    { "log_neg" "0" }
> +    { "minus" "-7" }
> +    { "bit_neg" "-23" }
> +
> +    { "plus1" "4" }
> +    { "plus2" "135" }
> +    { "plus3" "171" }
> +
> +    { "and" "128" }
> +    { "or" "12" }
> +}
> +
> +foreach probe_info $probe_names_and_values {
> +    set name [lindex $probe_info 0]
> +    set value [lindex $probe_info 1]
> +    with_test_prefix $name {
> +	goto_probe $name
> +	test_probe_value $value
> +    }
> +}
> -- 
>
> 2.29.2
>
>

-- 
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF  31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
https://sergiodj.net/