FreeBSD's problems as seen by the BSDForen.de community

[Erik Osterholm]

Sorry to cold-CC you on this, yongari--please ignore if this doesn't
interest you.

On Thu, Jan 10, 2008 at 09:40:50PM +0100, Kris Kennaway wrote:
> Erik Osterholm wrote:
> >On Thu, Jan 10, 2008 at 11:56:15PM +0900, Adrian Chadd wrote:
> >>On 10/01/2008, Dominic Fandrey <kamikaze at bsdforen.de> wrote:
> >>
> >>This is the thing though. Its working for the developers, its not
> >>working for the users, so how do you think it'll get fixed?
> >>
> >>>The second big problem is the handling of regressions. PRs remain
> >>>unanswered or the reporters are told that the regressions they
> >>>report do not exist. Some of our members have even suffered the
> >>>experience that they developed a patch, but it simply was ignored
> >>>or turned down for the reason that it was a "Linux solution".
> >>>Especially frustrating for those among us who have never looked at
> >>>Linux code.
> >>Whats the PR number?
> >
> >I'm coming in in the middle of this thread, but here's one from July
> >2006:
> >kern/100839
> >
> >No one from the FreeBSD community ever responded on it.  I thought
> >that I'd even suggested removing the driver entirely, due to this
> >showstopping bug, and removing its listing as compatible, but now I
> >can't find an archived reference, so maybe it was in my head.
> >
> >I love FreeBSD, and I used it on a daily basis, but there's an
> >example, if you're genuinely interested.
> >
> >Erik
> 
> Yeah, that's a pretty good example of hardware with no real maintainer 
> in the FreeBSD community.  Actually it does look like yongari@ worked on 
> it a couple of months ago, so you might want to bring it to his attention.
> 
> Kris

I can do that, though it looks like the changes made were quite
generic to interfaces in general, and not specific to the TXP.  While
I was trying to get this to work, it looked pretty likely that the
problem was in how the kernel was talking to the device itself--the
device would get confused when it was brought down and back up.

My recollection is that the Linux driver just stops I/O to the card,
but leaves it in its online state, effectively disconnecting it from
the TCP/IP stack, in order to bring it down.  FreeBSD tries to
actually disable the interface, but doesn't re-initialize it correctly
when bringing it back up.

If someone doesn't want to take accountability for the bug, I'd really
like to see it removed from the compatibility list.  I could probably
find some hardware to donate to the cause of fixing it, if someone was
committed to fixing it, though.

Erik