Fw: Re: Implemented OPLOCK for FreeBsd

[Greg Lehey]

On Friday,  7 September 2001 at 17:08:14 -0700, Jeremy Allison wrote:
> On Fri, Sep 07, 2001 at 05:31:04PM -0700, Julian Elischer wrote:
>
>> so Linus is apparently considering making an event handling abstraction
>> like Kqueue. Want to lay bets on how likely it will be  that it is
>> neadlessly different form kqueue?
>
> Juliam, I don't care. Linus listens to application writers and
> engages them in dialogue. This is the first time any of the
> FreeBSD people have even considered the needs of Samba, and we
> already have a working interface for this.

To be fair, Tridge and I discussed the oplock issue a couple of years
ago.  I didn't know about kqueue at the time, and we came to the
conclusion that, although the Linux implementation had some merits, it
would probably be difficult to get the FreeBSD community to agree with
it.  At least that has proven to be correct :-)

BTW, about time lines: Tridge and I first discussed the issue in
mid-September 1999, significantly before the FreeBSDCon.  I can't see
any difference between what he told me about at the time and what is
now implemented.  Have I missed something?

> How many more times do we have to re-learn the pointless API
> differentiation wars of the 1980's ?

I think we should be careful not to go too far in the opposite
direction.  If nobody ever deviates from the standards, we get almost
no progress (only when the committee can decide).  kqueue promises
significantly higher performance than real time signals.  Just because
Samba doesn't need this level of performance (yet) doesn't mean that
we should choose an inferior implementation.

What I'm seeing here, though, is a confusion between interfaces and
implementations.  Obviously we should support a real-time signal
interface.  That doesn't mean that the implementation should be the
same.

Greg
--
See complete headers for address and phone numbers