From owner-freebsd-multimedia Wed May 19 3:48:26 1999 Delivered-To: freebsd-multimedia@freebsd.org Received: from lion.plab.ku.dk (lion.plab.ku.dk [130.225.105.49]) by hub.freebsd.org (Postfix) with ESMTP id 1CA4615414 for ; Wed, 19 May 1999 03:48:21 -0700 (PDT) (envelope-from tobez@lion.plab.ku.dk) Received: (from tobez@localhost) by lion.plab.ku.dk (8.9.3/8.9.1) id MAA18148; Wed, 19 May 1999 12:47:01 +0200 (CEST) To: multimedia@FreeBSD.org Cc: tobez@plab.ku.dk, voland@plab.ku.dk Subject: Matrox Meteor PPB/RGB Rev.C troubles From: Anton Berezin Date: 19 May 1999 12:47:01 +0200 Message-ID: <86d7zx2v6i.fsf@lion.plab.ku.dk> Lines: 53 X-Mailer: Gnus v5.3/Emacs 19.34 - on FreeBSD 4.0-current Sender: owner-freebsd-multimedia@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Amancio Hasty recommended me to contact this list on the problem I have. So I am doing. We've been using one PPB/RGB on a FreeBSD workstation for quite some time, and it works flawlessly (the driver *and* the card, I mean) under heavy usage. Now more about a problem. :-( Some time ago we decided to install the grabber on several new workstations (we are doing recording of microscopic images of living cells at the laboratory). So we've bought three more boards from the Matrox distributor here in Denmark. To my surprise and disappointment, they did not work! More precisely, they *do* work fine on frame sizes up to 256x256 (we use PAL/RGB24/DEV1, continuous asynchronous capture with occasional single captures to write an image file). Beyond that resolution, and we used to use 768x576, there is a constant flow of even/odd FIFO overflows. The old board, being installed to the very same machine, works, so it is not a computer hardware problem. I contacted Matrox, and they told me that I've got three newer ``Rev.C'' boards, and that the only thing that changed in 'em is an updated PCI bridge chip (though the driver still reports SAA 7116). They even have sent me a piece of Windows code they use in their driver to initialize PCI bus. There was not much difference with FreeBSD driver; anyway, I put everything they do in that piece of code (not much, really, about three writes into 7116 registers FreeBSD driver does not do) into FreeBSD driver, with the same results. There is not much to say beyond this point. I have downloaded specifications on SAA 7196 from Philips website and tried (almost randomly; I am by no means a driver man) to play with its registers. No luck. I wanted to do the same with 7116 chip, but it turned out that there is currently no mention of this chip on Philips website. Very weird. Oh, and I have also tried SuSE Linux, with the same disappointing results; no wonder - the code of Linux driver is IMHO almost identical to the FreeBSD one, subject to different driver models. My best guess is that there is some clocking problem, but I am afraid I cannot solve it myself; that's why I am writing to this list and am asking for a little help. I would be very grateful if someone can provide me with any clues/advice. Probably someone already had and solved this problem before. Thank you in advance, -- Anton Berezin The Protein Laboratory, University of Copenhagen To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-multimedia" in the body of the message