From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26427 invoked by alias); 18 Aug 2003 16:15:46 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 26419 invoked from network); 18 Aug 2003 16:15:45 -0000 Received: from unknown (HELO dberlin.org) (69.3.5.6) by sources.redhat.com with SMTP; 18 Aug 2003 16:15:45 -0000 Received: from [192.168.1.7] (account dberlin HELO dberlin.org) by dberlin.org (CommuniGate Pro SMTP 4.1.1) with ESMTP-TLS id 4795965; Mon, 18 Aug 2003 12:15:40 -0400 In-Reply-To: <3F40EB16.5000500@redhat.com> References: <3F1BF6D4.6090801@redhat.com> <3F2F3E9A.6050403@redhat.com> <3F2FF8F5.6070007@redhat.com> <3F3D65C5.2070005@redhat.com> <20030816151124.GA2746@nevyn.them.org> <3F3E5DA4.7080504@redhat.com> <20030816164110.GA3818@nevyn.them.org> <3F40EB16.5000500@redhat.com> Mime-Version: 1.0 (Apple Message framework v582) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <3476CD7A-D197-11D7-B29B-000A95A34564@dberlin.org> Content-Transfer-Encoding: 7bit Cc: Daniel Jacobowitz , gdb@sources.redhat.com From: Daniel Berlin Subject: Re: Terminally slow (75 seconds) on some steps Date: Mon, 18 Aug 2003 16:15:00 -0000 To: Andrew Cagney X-SW-Source: 2003-08/txt/msg00196.txt.bz2 On Monday, August 18, 2003, at 11:04 AM, Andrew Cagney wrote: >> First of all, we should be using mmap to access symbol files, at least >>> when there are no relocations. >> Been there, done that, was told to do it in BFD. > > If there are I/O performances gains to be had, GDB should share them > with binutils. Except, you can't. Even if we were to use the bfd window stuff, binutils wouldn't be using it. So i'm not sure what you mean. Whether we directly call mmap, or attempt to use the window stuff (which is harder to use for us, and not built to do what we want), we're still going to be not sharing it with binutils. IMHO, it makes more sense to add the 3 lines of code necessary to use mmap (we need not change anything else, since bfd reads on the mmap'd memory will be turned into memory reads as approriate), than to change more to use the window stuff. > Did you look at bfdwin.c (although, in light of recent patches to > modify bfd so that it can create things from target memory, the > interface may be heading for review). > The mmap usage in it not enabled unless you specifically configure it on, last i looked. In fact, last i looked, the window stuff wasn't even turned on unless you used --with-mmap. This may have changed though. > Andrew > >