From owner-freebsd-hackers Wed Nov 27 5:44:21 2002 Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 18AB337B401 for ; Wed, 27 Nov 2002 05:44:19 -0800 (PST) Received: from van-laarhoven.org (ap-z-5ab8.adsl.wanadoo.nl [212.129.218.184]) by mx1.FreeBSD.org (Postfix) with SMTP id 40DF043E4A for ; Wed, 27 Nov 2002 05:44:12 -0800 (PST) (envelope-from n_hibma@van-laarhoven.org) Received: (qmail 276 invoked from network); 27 Nov 2002 13:44:10 -0000 Received: from uitsmijter.van-laarhoven.org (n?hibma@10.66.0.1) by uitsmijter.van-laarhoven.org with SMTP; 27 Nov 2002 13:44:10 -0000 Date: Wed, 27 Nov 2002 14:44:10 +0100 (CET) From: Nick Hibma To: Tomas Pluskal Cc: "freebsd-hackers@freebsd.org" Subject: Re: umass driver speed In-Reply-To: <20021127141637.A343-100000@localhost.localdomain> Message-ID: <20021127144232.N99600-100000@uitsmijter.van-laarhoven.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG You would have to switch on UMASS_DEBUG in your kernel config to see whether the driver is actually asked for large chunks. If you ask CAM for 64k blocks it will produce 64k requests, which are passed down to the USB layer, which doesn't cut them up. So in that case there is something else wrong. Nick > > Hello, > > It seems to me that the slowness of umass driver is not caused by msdosfs. > I know about the msdosfs problems, for example when I run dd from my ZIP > drive (with FAT) with request size of 2048, it's extremely slow. When I > change request size to 64K, it gains normal speed (about 1MB/s). > > But with my camera (umass), I get: > > # dd if=/dev/da0s1 of=/tmp/nikon bs=2048 count=10 > 10+0 records in > 10+0 records out > 20480 bytes transferred in 0.379587 secs (53953 bytes/sec) > > # dd if=/dev/da0s1 of=/tmp/nikon bs=4096 count=10 > 10+0 records in > 10+0 records out > 40960 bytes transferred in 0.719477 secs (56930 bytes/sec) > > # dd if=/dev/da0s1 of=/tmp/nikon bs=8192 count=10 > 10+0 records in > 10+0 records out > 81920 bytes transferred in 1.389069 secs (58975 bytes/sec) > > # dd if=/dev/da0s1 of=/tmp/nikon bs=65536 count=10 > 10+0 records in > 10+0 records out > 655360 bytes transferred in 10.832928 secs (60497 bytes/sec) > > # dd if=/dev/da0s1 of=/tmp/nikon bs=524288 count=1 > 1+0 records in > 1+0 records out > 524288 bytes transferred in 8.668605 secs (60481 bytes/sec) > > The speed is nearly the same, no matter about the request size. > > What do you think about it ? > > Tomas > > On Sat, 24 Aug 2002, Nick Hibma wrote: > > > > > Yes, unfortunately it is a known issue in the msdos file system. The > > requests it makes are too small for the driver to take advantage of the > > speed of USB. > > > > Nick > > -- n_hibma@van-laarhoven.org http://www.van-laarhoven.org/ n_hibma@FreeBSD.org http://www.etla.net/~n_hibma/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message