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 A22FD38582A3; Wed,  8 Mar 2023 03:37:29 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org A22FD38582A3
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1678246649;
	bh=mZG3NK2lNOmajAaQIdc+WWWX0P/nXHWlLoD3uzCZcAo=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=cY5GHdc8/EBU3Q77ON3bR4HeUjdFVCidgn17M9K3p4yUI14vHFoisthVJd/UrZ6H5
	 +nBZLXfG+0BYg7/s/TDWQCwv/uYQFAmJndZqP7Fa8B9Ysz0vDIcbCPkso9UgUmqXjL
	 qqQVnjPlCjQT7oOmfNhokDyiw+LZ0GvtuKj8bF5M=
From: "xry111 at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/109035] meaningless memory store on RISC-V and
 LoongArch
Date: Wed, 08 Mar 2023 03:37:29 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: rtl-optimization
X-Bugzilla-Version: 13.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: xry111 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-109035-4-OK23DFZVJI@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109035-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109035-4@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=3D109035

--- Comment #3 from Xi Ruoyao <xry111 at gcc dot gnu.org> ---
(In reply to chenglulu from comment #2)
> I think this is most likely caused by the implementation of the public co=
de.

Agree, so I filled the component as rtl-optimization.

I tracked a (non root) cause to the line 1944 in ira-costs.cc:

          if (i >=3D first_moveable_pseudo && i < last_moveable_pseudo)
            i_mem_cost =3D 0;
          else if (equiv_savings < 0)
            i_mem_cost =3D -equiv_savings;
          else if (equiv_savings > 0)
            {
              i_mem_cost =3D 0;    // <=3D=3D=3D=3D=3D=3D HERE
              for (k =3D cost_classes_ptr->num - 1; k >=3D 0; k--)
                i_costs[k] +=3D equiv_savings;
            }

I don't really understand why we should prefer the memory if there is a
REG_EQUIV note, nor why this does not happen with -fPIE.=