Re: new ibook?
On Thu, Jun 28, 2001 at 06:03:22PM +0200, Michel Dänzer wrote:
> Sven LUTHER wrote:
> >
> > On Thu, Jun 28, 2001 at 12:04:24PM +0200, Michel Dänzer wrote:
> > > Sven LUTHER wrote:
> > > >
> > > > On Wed, Jun 27, 2001 at 06:09:50PM -0400, John Hughes wrote:
> > > > > On Wednesday 27 June 2001 02:29, Peter Meilstrup wrote:
> > > > > > Video -- it's a Rage Mobility 128 chip; I assume this is supported
> > > > > > in the latest Xfree86? Is there the ability to switch from the
> > > > > > internal display to the external (RGB/composite) connectors?
> > > > >
> > > > > Worked with 4.0....but I compiled 4.1 without a hitch. With 4.1 the
> > > > > ati128 is supported AND accelerated plus support for DRI.
> > > >
> > > > So you tried it connecting an external monitor ?
> > > >
> > > > As MacOs only support mirroring, but the chip seems to be the same rage
> > > > 128 mobility as the one in the TiBook, it should be able to output a
> > > > second head also (altough it has only 8MB, right ?).
> > > >
> > > > So did you try it in Xinerama mode ? Or simply with two screens in the X
> > > > layout section ?
> > >
> > > The r128 driver can't drive both heads independently yet. But as of 4.1.0,
> > > you can choose between Panel or CRT only (which should allow to drive an
> > > external monitor at high resolutions, unless Apple somehow crippled the
> > > hardware) or mirroring.
> >
> > But it is only a software thingy, not a hardware imposibility, that is what
> > i really wanted to know.
>
> The chip can certainly do it, but maybe Apple artificially crippled it in the
> iBooks? Let's not hope so.
let's hope so, ...
> > But then, if it don't work yet in X, then there must be another way to find
> > out. BTW, i guess the problem is in the sharing of the accel engine, where a
> > trick must be used like the 2nd head support in the G400, or the trick alan
> > implemented to support the gamma on both permedia3 of the jeronimo 2000
> > board.
>
> Maybe I'm missing something, but I imagine sharing the accelerator between two
> heads shouldn't be too hard. It's a single instance of the driver, and the
> server is single threaded after all.
ah, but the nromal way of doing things, is that each instance of the driver
claims the chips it is going to use, and as thus, when the second instance get
initialized, and tries to claim the same chip for the second head, it is no
more there.
Also there are other possible conflicts, involving
synchronization/initialization fo the accel engine between access to both
head. But then the framework was already done for the G400, just copy it.
> > What about work at the fbdve level ?
>
> It would be hard if not impossible to use acceleration in X then, at least on
> one head.
erm, at first, both heads of the G400 were supported under fbdev, while X only
supportedone, and claimed that the second head's ramdac was not powerfull
enough to be worth it. At that time, you configured fbdev to support both
heads, ran accelerated X on the first head, and fbdev X on the second head,
without accel. I guess later on matrox paid PI to do the job right or
something such.
> > Or a simple test program driving the second head independently from the
> > first, but not doing usefull stuff, just to see it can work.
>
> MockOS can drive both heads independently on my Pismo, I guess that's proof of
> concept.
but not on the new ibooks, at least that is what apple claims.
> I don't really need that feature so I'm not going to work on it in the
> foreseeable future. Someone else would have to.
:)))
I would if i had one of the new ibooks. I sawone first time yesterday, but it
was dead, and i couldn't play with it :(((
Friendly,
Sven Luther
Reply to: