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 B275F386F802; Fri, 10 May 2024 17:01:38 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B275F386F802
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1715360498;
	bh=RjcorNKdCGurc4PqjnFxRhLnnZ7MPYUcOpS0qsh79UI=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=k8dILx+NlcCi8UgNjEsAar2XYyDblwrKQ4s6lPdpd7DSl0vNgK1PrN+YdgDgRqm7v
	 31UFOYaSz43XORkpvS32C6WNWTARzAWaklJKt0d7MTRKYMRp5o2yELBKEoLZv2vwBb
	 PreBOk4O2hYg0Y44d1xTLXjzMKFRkZOw/3ftkAiw=
From: "roger at nextmovesoftware dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/115021] [14/15 regression] unnecessary spill
 for vpternlog
Date: Fri, 10 May 2024 17:01:38 +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: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: roger at nextmovesoftware dot com
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: roger at nextmovesoftware dot com
X-Bugzilla-Target-Milestone: 14.2
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-115021-4-rB2rmo2pYo@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-115021-4@http.gcc.gnu.org/bugzilla/>
References: <bug-115021-4@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=3D115021

--- Comment #2 from Roger Sayle <roger at nextmovesoftware dot com> ---
Here's a reduced test case that should be unaffected by the pending changes=
 to
how V8QI shifts are expanded.  Note that the final "t -=3D t4" is required =
to
convince the register allocator to "spill".

typedef signed char v16qi __attribute__ ((__vector_size__ (16)));
// sign-extend low 3 bits to a byte.
v16qi foo (v16qi x) {
    v16qi t7 =3D (v16qi){7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7};
    v16qi t4 =3D (v16qi){4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4};
    v16qi t =3D x & t7;
    t ^=3D t4;
    t -=3D t4;
    return t;
}

which produces:

foo:    movl    $67372036, %eax
        vmovdqa %xmm0, %xmm2
        vpbroadcastd    %eax, %xmm1
        movl    $117901063, %eax
        vpbroadcastd    %eax, %xmm3
        vmovdqa %xmm1, %xmm0
        vmovdqa %xmm3, -24(%rsp)
        vmovdqa -24(%rsp), %xmm4
        vpternlogd      $120, %xmm2, %xmm4, %xmm0
        vpsubb  %xmm1, %xmm0, %xmm0
        ret=