> I apologise, publically, to Theo Deraadt for airing this in public.
> That was not, ever, my intention; and I apologise profusely for doing
> so.

you should apologize for it from the start:

1) there was a message on port-pmax about write buffers and flushing
   them.

2) i replied to Jonathan privately saying that i was surprised at the
   current pmax/locore.S code; and that a book i have says (that on IDT
   R3xx) it should be done by reading from an uncached address.

   i was really surprised at the current pmax code; for the an idt3081
   port i am working on i had to change that (following Per's example in
   his x3081 port).  the IDT method may not give as much performance,
   but device drivers are simpler to write.

   thanks to Per for explaining (in detail) how the cpu's used in the
   DEC machines are different.  i wasn't aware that such a substantial
   difference existed in the r3xxx family.

3) you send back a series (6) of increasingly louder mail calling me
   an asshole, bullheaded, and confused. you tried to explain things,
   but i don't normally read mail past the first insult.

4) i explained carefully that the book said so; as well i am pretty
   sure IDT made the full series of Rxxx series processors (as a
   second source) and well, i don't really care -- my first mail was
   just a comment of surprise.

5) after the 6 pieces of mail where you goaded me, and wouldn't stop
   sending me mail even when i told you to, i told you to stop shoving
   your cock down my throat.

6) you sent mail about this last message, completely ignoring the context
   or that you had called me an asshole, etc. -- to my boss.  DON'T
   EVER DO THAT. NetBSD and my job have nothing to do with each other.

you haven't won any friends here. with this post, i am not either, but
perhaps i can cut my losses.

back to technical discussions.  has anyone disassembled the ultrix
locore.o to see what it does?

(btw, the pmax codebase was made available about the same time as the
sparc code was. get a move-on guys!)