NAME

accf_http — buffer incoming connections until a certain complete HTTP requests arrive

SYNOPSIS

options INET
pseudo-device accf_http

DESCRIPTION

This is a filter to be placed on a socket that will be using accept() to receive incoming HTTP connections.

Once installed on a listening socket, this filter is activated when a connection becomes ready to receive data (at which point accept(2) would usually return the connected descriptor to the application). The filter prevents the descriptor from being returned immediately to the application via accept(2). The descriptor is made available to the application via accept(2) only when one of the following conditions is met:

1. A complete, syntactically valid HTTP/1.0 or HTTP/1.1 HEAD or GET request has been buffered by the kernel.
2. The data buffered by the kernel cannot be part of a complete, syntactically valid HTTP 1.0 or HTTP/1.1 HEAD or GET request.

The utility of accf_http is that a server will not have to context switch several times before performing the initial parsing of the request. This effectively reduces the amount of required CPU utilization to handle incoming requests by keeping active processes in preforking servers such as Apache low and reducing the size of the file descriptor set that needs to be managed by interfaces such as select(), poll() or kevent() based servers.

EXAMPLES

If the accf_data accept filter is present in the kernel configuration, this will enable the http accept filter on the socket sok.

	struct accept_filter_arg afa;

	bzero(&afa, sizeof(afa));
	strcpy(afa.af_name, "httpready");
	setsockopt(sok, SOL_SOCKET, SO_ACCEPTFILTER, &afa, sizeof(afa));

SEE ALSO

setsockopt(2), accept_filter(9)

HISTORY

The accept filter mechanism and the accf_http filter were introduced in FreeBSD 4.0. They were ported to NetBSD by Coyote Point Systems and appeared in NetBSD 5.0.

AUTHORS

This manual page and the filter were written by Alfred Perlstein.