Re: Broken ep0

Mark Murray wrote:
> Maxime Henrion writes:
> > > Can you have a look at this please?
> > >
> > > The symptoms are an ep0 that works for a few packets and then
> > > completely stops working. It doesn't respond to pings, nothing.
> >
> > I think you and/or Warner got mistaken.  My last round of changes
> > to if_ep.c _fixed_ some race conditions that previously weren't a
> > problem, but became a problem once Warner flag'd the pccbb interrupts
> > INTR_MPSAFE.  It's very possible that there are other problems left
> > in the ep(4) driver, but I highly doubt my changes to if_ep.c are the
> > culprit.  I'm sorry but I can't do anything for you since I don't
> > even have ep(4) hardware here ; I was able to fix the race conditions
> > mentioned above because people reported panics with stacktrace that
> > made the problem obvious.
> 
> OK - problem.
> 
> I have a broken driver, with a dodgy patch to "fix" it, and you guys are
> both saying "not my problem".

You should mention that the dodgy patch to "fix" this is to backout
_Warner_'s commit, not mine.  I think it's a sufficient proof that I
have nothing to do with this, and I'm quite amazed to hear that Warner
thinks my commits are the culprit.  He didn't even talk to me about
this, are you sure you didn't get mistaken?

> I'm quite happy to make it my problem, but I'll need a bit of help,
> please.  Who is it going to be?

As I said, I'd be happy to help you, but it's hard when I don't even
have the hardware to test with...

Cheers,
Maxime