From owner-freebsd-hackers Thu Feb 12 15:17:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA26632 for hackers-outgoing; Thu, 12 Feb 1998 15:17:39 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from dingo.cdrom.com (dingo.cdrom.com [204.216.28.145]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA26612 for ; Thu, 12 Feb 1998 15:17:28 -0800 (PST) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost [127.0.0.1]) by dingo.cdrom.com (8.8.8/8.8.5) with ESMTP id PAA04707; Thu, 12 Feb 1998 15:16:01 -0800 (PST) Message-Id: <199802122316.PAA04707@dingo.cdrom.com> X-Mailer: exmh version 2.0zeta 7/24/97 To: Brian McGovern cc: Mike Smith , dg@root.com, hackers@FreeBSD.ORG Subject: Re: Mapping phyical memory in to the PCI address range... In-reply-to: Your message of "Wed, 11 Feb 1998 19:52:01 EST." <199802120052.TAA06972@bmcgover-pc.cisco.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 12 Feb 1998 15:16:00 -0800 From: Mike Smith Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > The UARTs they use on the board have a 64 byte buffer. By Cyclades > estimates, about 32 bytes would be available about the time the > firmware would be ready to move the data into whatever buffers > it uses. Does the firmware buffer in the onboard card memory? How does it use DMA to the host's memory? Particularly, does it require a separate DMA setup for every transfer, or can you just give it an address range to work in and not have to continually reprogram it? (I don't have the 9060 docco to hand here, so I can't check that.) > Again, I ask from the perspective that I'm trying to minimize HOST CPU > usage for moving the data, and figured a RAM (buffer) to RAM (clist) copy on > the motherboard would be cheaper/faster for the host than a RAM (buffer) to > RAM (clist) copy over the PCI bus. I'd therefore also expect that the inverse > would be true. But, from most of what I've heard, there should be little > difference at a cost of extra PLX9060 programming (which looks easy on > paper). I think the bottom line is going to be how much administrative overhead there is in reprogramming the 9060. How many 9060 register accesses are required per DMA? How many bytes of serial data are transferred per DMA? If the first is significantly less than a quarter of the second, then you have the opportunity to win. Although I think that was probably already obvious. > -Brian > -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe hackers" in the body of the message