netatm: plan for removal unless an active maintainer is found

[Hans Petter Selasky]

Hi,

Thanks for CC'ing me. I cannot subscribe to all the mailing lists around :-)

On Wednesday 15 March 2006 19:03, M. Warner Losh wrote:
> In message: <20060315185803.1956114c at Magellan.Leidinger.net>
>
>             Alexander Leidinger <Alexander at Leidinger.net> writes:
> : Am Wed, 15 Mar 2006 10:29:03 -0700 (MST)
> :
> : schrieb "M. Warner Losh" <imp at bsdimp.com>:
> : > : > now available to work on the capi integration, and hopefully will
> : > : > do the SMP safety work as part of that.  If not, it's also on the
> : > : > chopping block.  It's a significant piece of otherwise unmaintained
> : > : > code, and something that's not trivially testable (at least, not by
> : > : > me or anyone I've talked to lately :-).  I don't want to see it
> : > : > leave the tree, but it needs to be updated so that it can run
> : > : > MPSAFE before 7.0.
> : > :
> : > : I may add, that Hans-Petter Selasky has a MPSAFE replacement (written
> : > : from scratch it seems) for I4B (AFAIK including capi) and the USB
> : > : stack. I have tested or reviewed neither of them, but as far as I can
> : > : read in the mailinglists, he adresses not only the issues you mention
> : > : here, but he also provides bugfixes and additional features compared
> : > : to our current code base.
> : >
> : > The problem is that this code isn't busdma safe at the moment.  It was
> : > posted for review on the NetBSD lists and this was the biggest set of
> : > comments on tech-kern at netbsd.org.  Since it isn't busdma safe, we'd
> : > lose usb on sparc64 (and maybe arm) when this code is brought into the
> : > tree.  There have also been signficant concerns about the locking
> : > that's done in the code as well, but I've not reviewed it recently.

Again, all USB drivers are driven like if 32-bit addressing is used. Actually 
the issue was about a function having "u_int32_t" as return type or 
"bus_size_t". I will change that.

> : >
> : > There's been a lot of work done here, and that work is generally good,
> : > but last time I looked at the code it wasn't ready to be integrated to
> : > the tree.

Yes, right. I am still working on the USB system, but I am only one person, 
and I really could need some help.

> :
> : The questions are:
> :  - What's less work to do?
> :  - Is there someone who is willing to do the work?
>
> Agreed.  Hans-Petter's work has great potential, but I think someone
> with a lot of time and knowledge of FreeBSD specific issues is going
> to need to work with him to properly integrate it into the tree.  It
> isn't a drop in right now, but could be with some work.
>
> If usb abd usbHPS can co-exist in the tree, we might be able to do
> some of this in-tree.  But usb is a very important subsystem and
> transitioning to a new code base is a high-risk thing.

I would suggest that we freeze "/sys/dev/usb" or make a copy of it. Then we [a 
few people] start moving all USB device drivers over to the new USB API, 
which will end up under "/sys/dev/usb2". Sure this can be in the tree. Then 
if people want to activate it, they have to run, maybe something like:

cd /sys/dev/usb2
make package S=src.temp # maybe this is not needed if all the 
      # files are in place
make install

Or we can just use my SVN account at turbocat.net.

--HPS