From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id B28D13858416; Thu, 29 Jun 2023 20:46:36 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B28D13858416
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1688071596;
	bh=GLVseLMY0bpedRpipQ8WvUlO91QZT3j1LHXfUuJe8pE=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=T2gxk6om4FvUKGsM3UPxiQ0hmci4qHzv8o61DqbqCCuiAB7Ka9j/RF9AL7Z1ZBAgY
	 we9UFeDkBMNtcMv8kiSTri0m30xwjSTohimwvynHKI+jTouLOFbxJ5jGmuU/W3VC+6
	 AAi/uW37M/rcdCHNlQtjB6bspG+sd+A8ZcHbGzdw=
From: "cvs-commit at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/109849] suboptimal code for vector walking loop
Date: Thu, 29 Jun 2023 20:46:34 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: middle-end
X-Bugzilla-Version: 13.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: cvs-commit at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-109849-4-8AGftFGd38@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109849-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109849-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D109849

--- Comment #19 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jan Hubicka <hubicka@gcc.gnu.org>:

https://gcc.gnu.org/g:9dc18fca431626404b0692c689a2e103666e7adb

commit r14-2202-g9dc18fca431626404b0692c689a2e103666e7adb
Author: Jan Hubicka <jh@suse.cz>
Date:   Thu Jun 29 22:45:37 2023 +0200

    Compute ipa-predicates for conditionals involving __builtin_expect_p

    std::vector allocator looks as follows:

    __attribute__((nodiscard))
    struct pair * std::__new_allocator<std::pair<unsigned int, unsigned int>
>::allocate (struct __new_allocator * const this, size_type __n, const void=
 *
D.27753)
    {
      bool _1;
      long int _2;
      long int _3;
      long unsigned int _5;
      struct pair * _9;

      <bb 2> [local count: 1073741824]:
      _1 =3D __n_7(D) > 1152921504606846975;
      _2 =3D (long int) _1;
      _3 =3D __builtin_expect (_2, 0);
      if (_3 !=3D 0)
        goto <bb 3>; [10.00%]
      else
        goto <bb 6>; [90.00%]

      <bb 3> [local count: 107374184]:
      if (__n_7(D) > 2305843009213693951)
        goto <bb 4>; [50.00%]
      else
        goto <bb 5>; [50.00%]

      <bb 4> [local count: 53687092]:
      std::__throw_bad_array_new_length ();

      <bb 5> [local count: 53687092]:
      std::__throw_bad_alloc ();

      <bb 6> [local count: 966367641]:
      _5 =3D __n_7(D) * 8;
      _9 =3D operator new (_5);
      return _9;
    }

    So there is check for allocated block size being greater than max_size
which is
    wrapper in __builtin_expect.  This makes ipa-fnsummary to give up analy=
zing
    predicates and it will miss the fact that the two different calls to
__throw
    will be optimized out if __n is larady smaller than 1152921504606846975
which
    it is after _M_check_len.

    This patch extends ipa-fnsummary to understand functions that return th=
eir
    parameter.

    gcc/ChangeLog:

            PR tree-optimization/109849
            * ipa-fnsummary.cc (decompose_param_expr): Skip
            functions returning its parameter.
            (set_cond_stmt_execution_predicate): Return early
            if predicate was constructed.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/109849
            * gcc.dg/ipa/pr109849.c: New test.=