From owner-freebsd-multimedia  Wed May 28 03:30:20 1997
Return-Path: <owner-multimedia>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id DAA05288
          for multimedia-outgoing; Wed, 28 May 1997 03:30:20 -0700 (PDT)
Received: from whqvax.picker.com (whqvax.picker.com [144.54.1.1])
          by hub.freebsd.org (8.8.5/8.8.5) with SMTP id DAA05283
          for <multimedia@FreeBSD.ORG>; Wed, 28 May 1997 03:30:18 -0700 (PDT)
Received: from ct.picker.com by whqvax.picker.com with SMTP;
          Wed, 28 May 1997 6:26:45 -0400 (EDT)
Received: from elmer.ct.picker.com ([144.54.57.34]) by ct.picker.com (4.1/SMI-4.1)
	id AA04330; Wed, 28 May 97 06:26:44 EDT
Received: by elmer.ct.picker.com (SMI-8.6/SMI-SVR4)
	id GAA22891; Wed, 28 May 1997 06:25:41 -0400
Message-Id: <19970528062541.49315@ct.picker.com>
Date: Wed, 28 May 1997 06:25:41 -0400
From: Randall Hopper <rhh@ct.picker.com>
To: Bernie Doehner <bad@uhf.wireless.net>
Cc: multimedia@FreeBSD.ORG
Subject: Re: DMA and fxtv.
References: <Pine.BSF.3.95.970527195528.1231A-100000@uhf.wdc.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.74
In-Reply-To: <Pine.BSF.3.95.970527195528.1231A-100000@uhf.wdc.net>; from Bernie Doehner on Tue, May 27, 1997 at 07:58:49PM -0400
Sender: owner-multimedia@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

Bernie Doehner:
 |FINALLY getting somewhere. I just ran fxtv with the -noDirectV switch and
 |it was rock solid. Interactive response time was horrible, but I didn't
 |lose any frames, picture was sharp, and NO CRASHES.
 |
 |At least I can live with this for the time being.

Good deal!  If it's eating too much CPU, bump up the FRAME_TIMER_DELAY_MS
setting (#define at the top of tvcapture.c).  This'll slow down the rate at
which the frames are grabbed from the driver buffer.

Now as far as analyzing why this doesn't crash you...hmmm.  It's
interesting.  30fps are still being kicked out by your TV card.  However,
it's not remaining only on the PCI bus, blasting onto your video card
directly; it's being forced into the memory bus and dumped in the driver's
capture buffer.  This is then being sampled periodically by fxtv, coaxed
into an XImage, and tossed to the XServer that then blits it to the frame
buffer.

What's interesting is it seems there'd be just as much PCI traffic in both
cases--even a bit more in the non-direct-video case because the occasional
upstream ximage blit has to compete with the downstream video from the TV
card.  Something just strange about PCI-to-PCI DMA on your MB chipset or
your Stealth 3D 2000 I guess.

Randall