[linuxdc-dev] Re: [linuxsh-dev] custom Dreamcast parallel port

[Paul M. <le...@li...>]

(CC list pruned.. this entire thread is barely on-topic as-is..)

On Fri, May 28, 2004 at 10:40:26PM +0100, Adrian McMenamin wrote:
> For a little while in 2002 I was perhaps the most active Dreamcast linux=
=20
> kernel hacker - writing a device driver or two and even a filesystem for =
the=20
> VMU device (though I was standing on the shoulders of giants). Change of =
job=20
> etc meant all that had to stop and there were a few others still going, b=
ut=20
> my assessment is that the DC hacking scence - both Linux and homebrew stu=
ff=20
> generally - has more or less died in the last six months.
>=20
The DC dev community has been dwindling for years, so this is nothing new.
However, it still remains a fairly accessible peice of hardware, relatively
well supported in LinuxSH, etc. so there will always be people idly hacking
on it (myself included). Further undocumented hardware and registers will
certainly keep it alive from a passive interest point of view. I'll certain=
ly
take a DC over a 7750 SolutionEngine any day.

In the 2.6 tree now the DC is still one of the best supported boards (mostly
on account of the fact that I did most of my development and testing on it),
and it should continue to be well supported in the future as well. If the
NetBSD effort is alive now, it would certainly be a good idea for someone to
keep an eye on that and see if they come up with anything interesting that
we don't support.

> A pity, but that's the way it goes - certainly no commercial support was =
ever=20
> likely and the prohibitive cost of the nic for the DC meant serious=20
> developers were always going to be thin on the ground.

Even given the cost of the BBA, it's still a very affordable and accessible
platform (especially when your other alternatives are costly reference boar=
ds).

Though in comparison to commercially viable solutions, most of the homewbrew
scene seems to have flocked to the PS2, etc.

Given the topic of drivers and such, this will all get done in time. With O=
SS
going away probably in 2.7, we don't really have much of a choice for ALSA
migration, so the AICA stuff will all be picked up and merged whenever that
happens. Maple needs to be rewritten entirely to support sub-devices, confo=
rm
with the new driver model, interoperate with the DMA API, etc. at which poi=
nt
we should finally be able to get the thing merged in mainline and never hav=
e to
touch it again, until someone decides they don't like the driver model and
decides to rewrite it again, etc.

None of this stuff is very high on my TODO list, so this obviously isn't go=
ing
to get done quickly, ideally someone else with an interest in this stuff sh=
ould
do this and just send patches (or better yet, use GNU arch so I can merge
things easier). LinuxDC still has some useful things in CVS, but it's quite
outdated, so resurrecting the linux-sh-dc stuff would probably be
a worthwhile first step (at least for 2.6, 2.4 is done at this point, so no
one should even be thinking about it).