X-Recipient: archive-cygwin AT delorie DOT com X-Spam-Check-By: sourceware.org Date: Mon, 23 Aug 2010 12:34:47 +0200 From: Corinna Vinschen To: cygwin AT cygwin DOT com Subject: Re: [ANNOUNCEMENT] Updated: cygwin-1.7.6-1 Message-ID: <20100823103447.GD6726@calimero.vinschen.de> Reply-To: cygwin AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com References: <4C70E660 DOT 2070600 AT alice DOT it> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <4C70E660.2070600@alice.it> User-Agent: Mutt/1.5.20 (2009-06-14) Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com On Aug 22 10:57, Angelo Graziosi wrote: > Larry Hall wrote: > >>Corinna Vinschen wrote: > >>>- Improve performance of stat and a few other functions. ls(1) should > >>> be up to 30% faster > >> > >>I have a directory (500MB, 30 files) which contains mainly 'exe' (setups for TB, FF, OO etc.). If I try 'ls -l' in this directory, the first time it take about 30 seconds to list the files. After the first time, the listing is almost without delay. The same happens also with 'ls -l /usr/bin'. > >> > >>When there is the 'hang' (30 secs.), Task Manager shows that AVG9 takes about 50% of CPU: this occurs *only* with 1.7.6 but _not_ with 1.7.5, with which 'ls -l' is almost immediate, regardless of the number and type of files. > >> > >>Obviously I have tested this, each time, with a 'fresh machine', to avoid 'cache' effects. > >> > >>The system is WinXP SP3, AMD Athlon 64X2DC 2.03GHz, 1.75GB RAM. > > > >Try a recent snapshot: > > > > > > I have tried cygwin1-20100822.dll.bz2, but same results. :( > > The first time (no cache 'effects') I do > > $ time ls -lrt /usr/bin > > the results are: > > CYGWIN_NT-5.1 1.7.5(0.225/5/3) 2010-04-12 19:07 i686 Cygwin > time ls -lrt /usr/bin > real 0m16.531s > user 0m0.108s > sys 0m0.421s > > CYGWIN_NT-5.1 1.7.6(0.230/5/3) 2010-08-16 16:06 i686 Cygwin > real 1m3.171s > user 0m0.155s > sys 0m0.702s > > CYGWIN_NT-5.1 1.7.6s(0.231/5/3) 20100822 02:25:11 i686 Cygwin > real 1m4.218s > user 0m0.280s > sys 0m0.609s I can't reproduce such a problem and I don't have AVG9 (virus scanner?). The effect is unfortunate, but the only important thing which has changed in 1.7.6 in terms of readdir and stat is the fact that the code tries to reduce the number of NtCreateFile/NtOpenFile calls by reusing a handle already opened on the file or directory before. I don't know what Cygwin can do about it, other than dropping the speedup entirely. Always assuming this is the real cause. There's no good reason that AVG9 hangs on anything Cygwin opens at all. There are only *very* few cases in which a handle is opened without allowing to share the file, and if so, it's never doing that for longer than the respective function call. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple