Date: Tue, 29 Nov 2011 14:47:41 -0500 From: Jung-uk Kim <jkim@FreeBSD.org> To: Kostik Belousov <kostikbel@gmail.com> Cc: freebsd-x11@freebsd.org, Luigi Rizzo <rizzo@iet.unipi.it> Subject: Re: suggested xorg-compatible video HW for FreeBSD/amd64 ? Message-ID: <201111291447.43176.jkim@FreeBSD.org> In-Reply-To: <20111129183839.GU50300@deviant.kiev.zoral.com.ua> References: <20111128092008.GA58668@onelab2.iet.unipi.it> <201111291304.15998.jkim@FreeBSD.org> <20111129183839.GU50300@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday 29 November 2011 01:38 pm, Kostik Belousov wrote: > On Tue, Nov 29, 2011 at 01:04:14PM -0500, Jung-uk Kim wrote: > > On Tuesday 29 November 2011 04:00 am, Gary Jennejohn wrote: > > > On Mon, 28 Nov 2011 18:46:23 -0700 (MST) > > > > > > Warren Block <wblock@wonkity.com> wrote: > > > > On Mon, 28 Nov 2011, Kevin Oberman wrote: > > > > > Just a quick reminder as it seems to be forgotten too > > > > > often...radeonhd has not been supported or updated in a > > > > > long time and is generally a bad choice. Modern Radeon > > > > > (4xxx) is supported much better with the ati driver which > > > > > is still getting some love. > > > > > > > > > > Since Radeon is open-source, you might expect good support, > > > > > but it also depends on kernel mode setting and requires > > > > > custom KMS DRI code in the kernel which I am not sure > > > > > anyone is working on for FreeBSD. > > > > > > > > > > Maybe when the Intel KMS is completed... > > > > > > > > The Foundation has expressed interest in funding someone to > > > > work on it. http://forums.freebsd.org/showthread.php?t=27255 > > > > > > > > Spread the word. > > > > > > > > Kostik Belousov has said he won't be workin on the radeon > > > > driver. Don't know why. If anyone can be found to work on > > > > it, I'm pretty sure a bunch of people would be willing to > > > > individually contribute additional funding or hardware. > > > > > > I don't think he wants to take on TTM. Could also be that he > > > doesn't have any hardware, which would IMO be a major problem > > > for whomever takes this on considering the rate at which > > > AMD/ATI brings new chips to market. High-end cards, which are > > > usually the first to market, can be quite expensive. > > > > I believe major hurdle is porting TTM but the future of this API > > is not so bright. In fact, X.org ATI driver uses GEM API now and > > it is internally mapped to TTM calls by Linux DRM (aka "GEM-ified > > TTM manager"). Unfortunately, as always, I don't see clear plans > > from Linux/X.org developers. I can only guess few possibilities. > > > > 1. Linux/X.org folks drop GEM-ified TTM and use native GEM calls. > > 2. Linux/X.org folks drop GEM-ified TTM and use native TTM calls. > > 3. Linux/X.org folks re-invent new wheels (again). > > 4. No change. > > > > My guess is #1 is most likely scenario in the near future. Even > > if Linux/X.org folks don't do it, we may be able to implement it > > without TTM because X.org ATI driver uses GEM API already and we > > do not have AMD/ATI Catalyst driver for FreeBSD anyway. So, I > > guess we have two choices ATM: > > > > 1. Fully porting TTM, GEM-ified TTM, and KMS. > > 2. Replacing GEM-ified TTM with GEM and porting KMS. > > > > BTW, I am not volunteering. ;-) > > GEM is sort of API, which provides the interface between user and > kernel callers and the real code managing the GPU. TTM is memory > manager, which was designed for GPUs that have on-board memory and > also can access host memory. Besides memory manager, you also need > a thick layer of card-specific code. This is only for the execution > part. If my understanding is correct, "GEM-ified TTM manager" is card-specific code, isn't it? So, we may be able to replace GEM-ified TTM with Radeon-specific code *without* TTM backend. Am I wrong? > I already stated several times that I will port TTM if somebody > takes care of the card specific execution code and KMS. I found > that KMS is not very interesting for me. Or someone write card-specific execution code without TTM backend? Don't get me wrong, I am not saying that is easier. It was just a thought... ;-) > Another problem with already ported Intel driver and possibly > ported Radeon driver is the need for shared pool of people > interested in taking the updates from Linux and merging it back to > FreeBSD. This is definitely not hard to do, e.g the merge of > changes from the 3.2 merge window took ~5 hours of my time. but I > do not want to do this forever. In fact, I did just that for couple of years and I understand what you're saying. It wasn't trivial even though Linux and FreeBSD shared the same DRM sources at the time and we had anholt@ in our camp. :-( Jung-uk Kim
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201111291447.43176.jkim>