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 0690A3858D33; Sun, 19 Nov 2023 15:14:41 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 0690A3858D33
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1700406881;
	bh=eEfLlYtRcVsUD8zoFgyAduc8A14XGHkWalzkehggfAk=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=s/VVr33/vLBXQgB48eOd4DqLz+cdrHRfVj5RLn2WmLqvI4zQNtAwpaJFQAVQUgdLc
	 UCPWFOOKY23d8i9B6xuRiEXI2fLsv5i7XF8ZAxFr36X0zT67xD+0VA9yWqE46tUSD6
	 tZ9MEcBqzeHBaeOSGOh5Gbm7tZkI3Linshx3lQcA=
From: "hubicka at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/110287] _M_check_len is expensive
Date: Sun, 19 Nov 2023 15:14:40 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: libstdc++
X-Bugzilla-Version: 13.1.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: hubicka at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
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-110287-4-LFED5WAbth@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-110287-4@http.gcc.gnu.org/bugzilla/>
References: <bug-110287-4@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=3D110287

--- Comment #8 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
With return value range propagation
https://gcc.gnu.org/pipermail/gcc-patches/2023-November/637265.html
reduces --param max-inline-insns-auto needed for _M_realloc_insert to be
inlined on my testcase from 39 to 35.

This is done by eliminating two unnecesary trow calls by propagating fact t=
hat
check_len does not return incredibly large values.

Default inline limit at -O3 is 30, so we are not that far and I think we re=
ally
ought to solve this for next release since push_back is such a common case.

Is it known that check_len can not return 0 in this situation? Adding=20
if (ret <=3D 0)
  __builtin_unreachable
saves another 2 instructions because _M_realloc_insert otherwise contain a =
code
path for case that vector gets increased to 0 elements.=