From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23349 invoked by alias); 9 Jul 2004 21:21:05 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 23327 invoked by uid 48); 9 Jul 2004 21:21:04 -0000 Date: Fri, 09 Jul 2004 21:21:00 -0000 Message-ID: <20040709212104.23326.qmail@sourceware.org> From: "dnovillo at gcc dot gnu dot org" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20040709210136.16461.falk@debian.org> References: <20040709210136.16461.falk@debian.org> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug tree-optimization/16461] ICE: expected ssa_name, have integer_cst in independent_of_stmt_p, at tree-tailcall.c:210 X-Bugzilla-Reason: CC X-SW-Source: 2004-07/txt/msg00968.txt.bz2 List-Id: ------- Additional Comments From dnovillo at gcc dot gnu dot org 2004-07-09 21:21 ------- Confirmed. The bug is in tree-tailcall.c:independent_of_stmt_p. The loop that looks up PHI arguments, ought to handle gimple_min_invariants it finds in PHI operands. It should not assume that they are all SSA_NAMEs. Not quite sure what the right to do is here. We are analyzing ret_1 in statement T.2_6 = ret_1 + ret2_2; and abort as soon as we get to the first operand of # ret_1 = PHI <0(2), ret_11(1)>; Jeff, IIRC you implemented this. -- What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|unassigned at gcc dot gnu |law at gcc dot gnu dot org |dot org | Status|UNCONFIRMED |ASSIGNED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16461