Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 19 Aug 1998 10:32:16 -0700 (PDT)
From:      Marc Slemko <marcs@znep.com>
To:        Andre Oppermann <oppermann@pipeline.ch>
Cc:        hackers@FreeBSD.ORG
Subject:   Re: combined accept&read
Message-ID:  <Pine.GSO.4.02A.9808191026500.13979-100000@redfish>
In-Reply-To: <35DB0422.B6447353@pipeline.ch>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 19 Aug 1998, Andre Oppermann wrote:

> I was chatting a little bit with the Zeus developers to ask them how
> to make FreeBSD the best/fastest platform for them.
> 
>  "accept&read combined call is something else we're toying with. Very
>  often we get a poll/select broken because there's data coming in on a
>  socket, and most of the time there's already data in the socket
>  buffers.  Being able to do the accept() and read() in a single call
>  should be beneifital."

Note that this is similar to NT's AcceptEx() call.  I'm not fully
convinced of how significant the impact of this is on a good
architecture.

> 
>  They also wish to have a few thousand open fd's without slowing
>  down stuff like get_first_free_fd().

Yes, whatever you do you need efficient handling of a few thousand fds
just for clients.  If you can efficiently handle a few thousand more, you
can then afford to use a descriptor cache, which looks to be the nicest
way to implement a cache if the path from file descriptor to network is
short enough.  If it isn't, then actually keeping a copy of the data in
your userland memory can be an annoying win.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.02A.9808191026500.13979-100000>