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 2E0E33858D37; Tue, 19 Mar 2024 16:49:59 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 2E0E33858D37
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1710866999;
	bh=sEydaK41W4IcBAqfPrrKUG2fqk5/fNfQGrV8KY1fsQU=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=Pa0y4sM1pN2jGV7qx5Xf/sQwYFKLQVokiPRDXqUYI6kEHiJvWJhwu5dpmmT4eNJx7
	 g576afMtWuRrOKLsTqEsg1bNWqNbPEu5RKPapqxbqhqF2WbbXkKj3pwBW7Hw8wKtHT
	 Dw3rkFYAK7VTP15KV4vdnrI5zsvZMtOY+koYBz3Q=
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/112499] [11/12/13/14 Regression] ICE in
 output_operand: invalid expression as operand with addition of 2 differences
 of address of labels
Date: Tue, 19 Mar 2024 16:49:58 +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: 14.0
X-Bugzilla-Keywords: ice-on-invalid-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub 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: 11.5
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-112499-4-bCxJ3oC1cw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-112499-4@http.gcc.gnu.org/bugzilla/>
References: <bug-112499-4@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=3D112499

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Or consider
typedef unsigned __INTPTR_TYPE__ uintptr_t;
volatile int v;

int
foo (int x)
{
  static uintptr_t a =3D ((char *)&&l2 - (char *)&&l1) + ((char *)&&l4 - (c=
har
*)&&l3) + ((char *)&&l6 - (char *)&&l5);
  void *b[] =3D { &&l1, &&l2, &&l3, &&l4, &&l5, &&l6 };
l1:
  ++v;
l2:
  ++v;
l3:
  ++v;
l4:
  ++v;
l5:
  ++v;
l6:
  ++v;
  if (x >=3D 0 && x <=3D 5)
    {
      x +=3D 42;
      goto *b[x - 42];
    }
  return a;
}
That is even older.
r0-79351-g7771bb621aad43c7d0e21ccec4922dd3beac9cb6
still used to reject that:
error: initializer element is not computable at load time
r0-79390-g7ffb5e78794b6d8f8094d7c928e6f0e94b62cad6
already accepts but assembler doesn't handle that:
pr112499-2.s:58: Error: invalid operands (*ABS* and .text sections) for `-'
pr112499-2.s:58: Error: invalid operands (*ABS* and .text sections) for `-'
pr112499-2.s:58: Error: invalid operands (.text and .text sections) for `+'
        .quad   .L3-.L6-.L4-.L2+.L7+.L5
Guess we can only accept a subtraction of 2 labels, not addition of 2 label=
s or
something more complex.=