Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Message-ID: <000701c00154$7bfcce60$0100000a@TIMYX18EWDT6RQ> From: "Tim Prince" To: Subject: linker mis-alignments for cygwin 1.1.3 Date: Tue, 8 Aug 2000 09:19:43 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.5600 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.5600 I see that linking my codes under cygwin 1.1.3 includes modules named _cygwin_crt0_common.o environ _fmode _impure_ptr apparently all from libcygwin.a, which take a total of 12 bytes, beginning at a properly aligned address, leaving all subsequent modules mis-aligned. Is it a total coincidence that this problem didn't occur for me prior to 1.1.3? This appears to be the likely cause of the severe performance regression which I annoyed you all by reporting. Is there a mechanism for controlling linker alignments, or is that considered impossible in the cygwin environment? Up to now, data regions have usually linked with satisfactory alignment. Of course, the .p2align scheme used by gcc has been defeated all along by the random alignments of code under cygwin. If it were not for these problems, it appears that cygwin would out-perform linux on the operations which I have been benchmarking. I will attempt to get the 1.1.4 distribution later on. As others have pointed out, simply following instructions causes nothing to be up-dated. Tim Prince -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com