From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18101 invoked by alias); 25 Jun 2010 09:15:01 -0000 Received: (qmail 17928 invoked by uid 48); 25 Jun 2010 09:14:34 -0000 Date: Fri, 25 Jun 2010 09:15:00 -0000 Message-ID: <20100625091434.17927.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug middle-end/43866] [4.3/4.4/4.5/4.6 Regression] wrong code with -fbounds-check -funswitch-loops In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "jakub at gcc dot gnu dot org" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2010-06/txt/msg02434.txt.bz2 ------- Comment #18 from jakub at gcc dot gnu dot org 2010-06-25 09:14 ------- That part comes from the questionable testcase, which does a_sp => matrix%local_data_sp before the loop unconditionally, eventhough matrix%local_data_sp is uninitialized unless use_sp is .true. Without the unswitching the values from the uninitialized struct matrix%local_data_sp are read, but not tested when use_sp is .false. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43866