DragonFly kernel List (threaded) for 2006-04
Re: Development stalling?
On 4/23/06, Justin C. Sherrill <justin@xxxxxxxxxxxxxxxxxx> wrote:
> There was discussion around AMD64 and also a PowerPC port, which was
> enough to create a FAQ item. The end result was that any port work that
> happens now would then have to get revamped anyway as we work towards a
> new system, so it's better to wait.
Is that so? I was under the impression the truly low-level stuff was
done. However there is still an IPI task on the Projects page.
> This would be very useful. corecode's regular builds help catch build
> problems, but (for obvious reasons) not usage.
Then that's what I'll do. I'll move the dual-core out of the critical
path of my network and use it for development and testing (development
of my own projects, that is, unless I find something I can do well
enough for DragonFly).
> I'd say the opposite has happened - we've reached completion on a lot of
> projects recently. Pkgsrc has gone from being an experiment to becoming
> _far_ more functional than I thought would happen in this timeframe (much
> credit to Joerg on that). A good amount of what happened in 1.0 and 1.2
> was fixing things that people always wanted to fix in FreeBSD 4. The Big
> Things are just starting to happen now, like Matt working on cache
> coherency, or the move to ZFS.
I say stalling because in the past, catching up with other code bases
like pf, NetBSD's USB, etc. used to happen quickly, but now it doesn't
seem to happen at all. With pf it's fine already, but USB is
definitely verging on broken. I'll see what more I can diagnose about
DragonFly ignoring my mass storage device. In my case, it could just
be the motherboard acting up - now after rebooting NetBSD USB doesn't
work at *all*, which is either a -current regression or something to
complain to Asus about. I chose this motherboard for good BSD support
but that's a moot point when it doesn't work properly on the hardware
> And, people who aren't me are looking at the docs, which is fantastic as
> far as I'm concerned.
Always good. The effort for good documentation is fantastic, a big
part of the reason people move to BSD in the first place.
-- Dmitri Nikulin