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 359633858D28; Tue,  2 May 2023 16:47:00 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 359633858D28
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1683046020;
	bh=O7a80esU871syp5nyF9a1D31NZ2VQjRmnrX3Wda21vw=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=OnFrmqWWXQ9W/UqLyQVTBYkfNguDymjByeADWe2MIA2hvmZOAYVYSXas1MU/8BE1G
	 tO0bcEvgm4jThoy8wpr3KlB1qim31n4/k46PIDfv5ObMDMxH4GqyD4q2qFG5FhfzbD
	 J54wCQ06NBkAD9Vbnp8RwgABnajAfQ9eMtTloEWs=
From: "gjl at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/109650] avr-gcc incorrect code with -Os
Date: Tue, 02 May 2023 16:46:59 +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: 12.2.0
X-Bugzilla-Keywords: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: gjl 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: cc
Message-ID: <bug-109650-4-qjHv6tvSdw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109650-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109650-4@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=3D109650

Georg-Johann Lay <gjl at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |saaadhu at gcc dot gnu.org

--- Comment #6 from Georg-Johann Lay <gjl at gcc dot gnu.org> ---
As it appears, disabling avr.cc's machine dependent reorg by means of
-fdisable-rtl-mach can stop this PR.  Maybe avr.cc::avr_reorg() is no more
correct after PR92729 because contrary to cc0, now a branch instruction may
have more than one CCmode setter.  A part of this optimization could be run
prior to .split2 when cbranch insns are still intact.  In particular
transformations like from "if (x <=3D 2)" to "if (x < 3)" must be performed=
 on
cbranch, or otherwise avr_reorg() must check that there is only one CCmode
setter, or otherwise adjust *all* CCmode setters.

This optimization should not run too early so that
avr_reorg_remove_redundant_compare can still do its job on switch-case if-e=
lse
trees.

I am allowing me to add Senthil to CC, maybe he has some ideas.=