idle_timeout processing in the parent smbd?

[Volker Lendecke]

On Mon, Jun 04, 2007 at 10:23:51AM -0700, Jeremy Allison wrote:
> Depends on the API :-). Sounds useful, but we need to
> move everything to the event-based system first don't
> we ?

No, I don't think so. The particular thing in question here
is the parent smbd that needs to exit (!) after being idle
for a while. So we need to tell the difference between
events that would block the parent to exit from events that
are so unimportant that the parent might just exit.

With a general events system that has timed events this is
quite difficult to achieve I think. Thus my suggestion that
the parent might just exit when the number of clients drops
to zero.

BTW, I find the concept that some daemon creates the
listening socket for us quite strange. inetd, that's
something I see, but doing the listen and not do the fork? I
don't know...

I'm not sure I want to go through hoops to adjust to this,
I'd rather accept some inefficiency on that platform by
doing a few more exec(2) calls than strictly necessary.

Volker
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.samba.org/archive/samba-technical/attachments/20070604/e644a426/attachment.bin