From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <92f0483b4252b9cd48e240d6d4122c65@plan9.bell-labs.com> From: jmk@plan9.bell-labs.com To: 9fans@cse.psu.edu Subject: Re: [9fans] ati radeon support In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Date: Fri, 3 Oct 2003 14:28:09 -0400 Topicbox-Message-UUID: 5cc21718-eacc-11e9-9e20-41e7f4b1d025 On Fri Oct 3 13:57:55 EDT 2003, forsyth@vitanuova.com wrote: > yes, the problem with graphics cards is really the strange card-specific stuff > needed to put it into a reasonable large, deep linear colour mode. > if we had only to write code for the hw cursor for each one, > that would be much less of a problem. even some of the hardware > drawing code doesn't look hard to exploit ... once the card is in a sensible mode. > > the cursor code always seemed fairly similar, so it might not be too hard > to have a table of ids just for that if necessary, and certainly card-specific > code for that is fairly easy to write. one of the changes i > was making to the software cursor in inferno is to bring it > into a common interface with the hardware one > (allowing for the need to interact a bit with devdraw re. screen access). The reason the software cursor disappeared form Plan 9 was, by the time devdraw went in, the cursor and mouse code had accumulated so much cruft after years of adapting to new graphics models and hardware that it was a constant source of lockups and crashes. There was an attempt when changing to devdraw to pare all that back and simplify it. The software cursor was the cause of many of the problems. The software cursor is OK for getting started, but I'd hate to use it in earnest. The flicker when scrolling text was awful and I'm trying to imagine what it would be like with vnc.