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 30F873858410; Tue, 21 Feb 2023 12:39:08 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 30F873858410
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1676983148;
	bh=S3TmeiYe+mNYM+5SRef6VZIImDf8R0GhFrZoS11zCqg=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=wlt2RaHHvk62mOmCV5eyDqnYQxS8Dt87O03rZZ0swowBYqwSSnaWKgPibJPDE5CBj
	 I2CEGPt6pbkya61nWaVyl6tShrUd2W5YVQfVoXb1KzNbXuXVFDpzXH40hWo9Q+6Rwt
	 B2nG/etfJCfCRH5qQ85U4hdXGQDdeFDtHYuOcwlI=
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug ipa/108802] [10/11/12/13 Regression] missed inlining of call
 via pointer to member function
Date: Tue, 21 Feb 2023 12:39:07 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: ipa
X-Bugzilla-Version: 13.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenth at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 10.5
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: priority cf_reconfirmed_on everconfirmed
 bug_status
Message-ID: <bug-108802-4-87uG8cjKXY@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-108802-4@http.gcc.gnu.org/bugzilla/>
References: <bug-108802-4@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=3D108802

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2
   Last reconfirmed|                            |2023-02-21
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
That means it was a conscious decision.

IPA function summary for f1()::<lambda(auto:1&&)> [with auto:1 =3D int
(A::*)(int)]/1 inlinable
  global time:     30.500000
  self size:       23
  global size:     25

the body is

  <bb 2> [local count: 1073741824]:
  _1 =3D f_18(D)->__pfn;
  _2 =3D (long int) _1;
  _3 =3D _2 & 1;
  if (_3 !=3D 0)
    goto <bb 3>; [50.00%]
  else
    goto <bb 4>; [50.00%]

  <bb 3> [local count: 536870913]:
  _4 =3D __closure_19(D)->__a;
  _5 =3D f_18(D)->__delta;
  _6 =3D (sizetype) _5;
  _7 =3D _4 + _6;
  _8 =3D *_7;
  _9 =3D _2 + -1;
  _10 =3D (sizetype) _9;
  _11 =3D _8 + _10;
  iftmp.0_20 =3D *_11;

  <bb 4> [local count: 1073741824]:
  # iftmp.0_16 =3D PHI <iftmp.0_20(3), _1(2)>
  _12 =3D __closure_19(D)->__a;
  _13 =3D f_18(D)->__delta;
  _14 =3D (sizetype) _13;
  _15 =3D _12 + _14;
  _22 =3D iftmp.0_16 (_15, 42);
  return _22;

so it is quite large.

While we can give inline incentives in IPA inlining we do not discover the
inline target there - the above has it also well obfuscated.  Maybe we
can still compute the set of functions possibly called here and perform
speculative inlining in IPA?

Maybe the frontend can also improve the initial IL which looks like

  <<cleanup_point return <retval> =3D (((long int) f->__pfn & 1) !=3D 0) ? =
(int
A::<T3f8> (struct A *, int) *) *(*((int (*) () * *) a + (sizetype) f->__del=
ta)
+ (sizetype) ((long int) f->__pfn + -1)) : f->__pfn ((struct A *) a +
(sizetype) f->__delta, 42)>>;=