2 Socket Usage

2.1  Introduction

The socket interface (module) is basically a "thin" layer on top of the OS socket interface. It is assumed that, unless you have special needs, gen_[tcp|udp|sctp] should be sufficent (when they become available).

Note that just because we have a documented and described option, it does not mean that the OS supports it. So its recommended that the user reads the platform specific documentation for the option used.

Asynchronous calls

Some functions allow for an asynchronous call (accept/2, connect/3, recv/3,4, recvfrom/3,4, recvmsg/2,3,5, send/3,4, sendmsg/3,4 and sendto/4,5). This is achieved by setting the Timeout argument to nowait. For instance, if calling the recv/3 function with Timeout set to nowait (i.e. recv(Sock, 0, nowait)) when there is actually nothing to read, it will return with {select, SelectInfo} (SelectInfo contains the SelectRef). When data eventually arrives a 'select message' will be sent to the caller:

{'$socket', socket(), select, SelectRef}

The caller can then make another call to the recv function and now expect data.

Note that all other users are locked out until the 'current user' has called the function (recv in this case). So either immediately call the function or cancel.

The user must also be prepared to receive an abort message:

{'$socket', socket(), abort, Info}

If the operation is aborted for whatever reason (e.g. if the socket is closed "by someone else"). The Info part contains the abort reason (in this case that the socket has been closed Info = {SelectRef, closed}).

The general form of the 'socket' message is:

{'$socket', Sock :: socket(), Tag :: atom(), Info :: term()}

Where the format of Info is a function of Tag:

Tag Info value type
select select_ref()
abort {select_ref(), Reason :: term()}

Table 2.1:   socket message info value type

The select_ref() is the same as was received in the SelectInfo.

2.2  Socket Registry

The socket registry is how we keep track of sockets. There are two functions that can be used for interaction: socket:number_of/0 and socket:which_sockets/1.

In systems which create and delete many sockets dynamically, it (the socket registry) could become a bottleneck. For such systems, there are a couple of ways to control the use of the socket registry.

Firstly, its possible to effect the global default value when building OTP from source with the two configure options:

--enable-esock-socket-registry (default) | --disable-esock-socket-registry

Second, its possible to effect the global default value by setting the environment variable ESOCK_USE_SOCKET_REGISTRY (boolean) before starting the erlang.

Third, its possible to alter the global default value in runtime by calling the function use_registry/1.

And finally, its possible to override the global default when creating a socket (with open/2 and open/4) by providing the attribute use_registry (boolean) in the their Opts argument (which effects that specific socket).

2.3  Socket Options

Options for level otp:

Option Name Value Type Set Get Other Requirements and comments
assoc_id integer() no yes type = seqpacket, protocol = sctp, is an association
debug boolean() yes yes none
iow boolean() yes yes none
controlling_process pid() yes yes none
rcvbuf default | pos_integer() | {pos_integer(), pos_ineteger()} yes yes 'default' only valid for set. The tuple form is only valid for type 'stream' and protocol 'tcp'.
rcvctrlbuf default | pos_integer() yes yes default only valid for set
sndctrlbuf default | pos_integer() yes yes default only valid for set
fd integer() no yes none
use_registry boolean() no yes the value is set when the socket is created, by a call to open/2 or open/4.

Table 2.2:   option levels

Options for level socket:

Option Name Value Type Set Get Other Requirements and comments
acceptconn boolean() no yes none
bindtodevice string() yes yes Before Linux 3.8, this socket option could be set, but not get. Only works for some socket types (e.g. inet). If empty value is set, the binding is removed.
broadcast boolean() yes yes type = dgram
debug integer() yes yes may require admin capability
domain domain() no yes Not on FreeBSD (for instance)
dontroute boolean() yes yes none
keepalive boolean() yes yes none
linger abort | linger() yes yes none
oobinline boolean() yes yes none
peek_off integer() yes yes domain = local (unix). Currently disabled due to a possible infinite loop when calling recv([peek]) the second time.
priority integer() yes yes none
protocol protocol() no yes Not on (some) Darwin (for instance)
rcvbuf non_neg_integer() yes yes none
rcvlowat non_neg_integer() yes yes none
rcvtimeo timeval() yes yes This option is not normally supported (see why below). OTP has to be explicitly built with the --enable-esock-rcvsndtime configure option for this to be available. Since our implementation is nonblocking, its unknown if and how this option works, or even if it may cause malfunctions. Therefor, we do not recommend setting this option. Instead, use the Timeout argument to, for instance, the recv/3 function.
reuseaddr boolean() yes yes none
reuseport boolean() yes yes domain = inet | inet6
sndbuf non_neg_integer() yes yes none
sndlowat non_neg_integer() yes yes not changeable on Linux
sndtimeo timeval() yes yes This option is not normally supported (see why below). OTP has to be explicitly built with the --enable-esock-rcvsndtime configure option for this to be available. Since our implementation is nonblocking, its unknown if and how this option works, or even if it may cause malfunctions. Therefor, we do not recommend setting this option. Instead, use the Timeout argument to, for instance, the send/3 function.
timestamp boolean() yes yes none
type type() no yes none

Table 2.3:   socket options

Options for level ip:

Option Name Value Type Set Get Other Requirements and comments
add_membership ip_mreq() yes no none
add_source_membership ip_mreq_source() yes no none
block_source ip_mreq_source() yes no none
drop_membership ip_mreq() yes no none
drop_source_membership ip_mreq_source() yes no none
freebind boolean() yes yes none
hdrincl boolean() yes yes type = raw
minttl integer() yes yes type = raw
msfilter null | ip_msfilter() yes no none
mtu integer() no yes type = raw
mtu_discover ip_pmtudisc() yes yes none
multicast_all boolean() yes yes none
multicast_if any | ip4_address() yes yes none
multicast_loop boolean() yes yes none
multicast_ttl uint8() yes yes none
nodefrag boolean() yes yes type = raw
pktinfo boolean() yes yes type = dgram
recvdstaddr boolean() yes yes type = dgram
recverr boolean() yes yes none
recvif boolean() yes yes type = dgram | raw
recvopts boolean() yes yes type =/= stream
recvorigdstaddr boolean() yes yes none
recvttl boolean() yes yes type =/= stream
retopts boolean() yes yes type =/= stream
router_alert integer() yes yes type = raw
sendsrcaddr boolean() yes yes none
tos ip_tos() yes yes some high-priority levels may require superuser capability
transparent boolean() yes yes requires admin capability
ttl integer() yes yes none
unblock_source ip_mreq_source() yes no none

Table 2.4:   ip options

Options for level ipv6:

Option Name Value Type Set Get Other Requirements and comments
addrform inet yes no allowed only for IPv6 sockets that are connected and bound to a v4-mapped-on-v6 address
add_membership ipv6_mreq() yes no none
authhdr boolean() yes yes type = dgram | raw, obsolete?
drop_membership ipv6_mreq() yes no none
dstopts boolean() yes yes type = dgram | raw, requires superuser privileges to update
flowinfo boolean() yes yes type = dgram | raw, requires superuser privileges to update
hoplimit boolean() yes yes type = dgram | raw. On some platforms (e.g. FreeBSD) is used to set in order to get hoplimit as a control message heeader. On others (e.g. Linux), recvhoplimit is set in order to get hoplimit.
hopopts boolean() yes yes type = dgram | raw, requires superuser privileges to update
mtu boolean() yes yes Get: Only after the socket has been connected
mtu_discover ipv6_pmtudisc() yes yes none
multicast_hops default | uint8() yes yes none
multicast_if integer() yes yes type = dgram | raw
multicast_loop boolean() yes yes none
recverr boolean() yes yes none
recvhoplimit boolean() yes yes type = dgram | raw. On some platforms (e.g. Linux), recvhoplimit is set in order to get hoplimit
recvpktinfo | pktinfo boolean() yes yes type = dgram | raw. On some platforms (e.g. FreeBSD) is used to set in order to get hoplimit as a control message heeader. On others (e.g. Linux), recvhoplimit is set in order to get hoplimit.
recvtclass boolean() yes yes type = dgram | raw. On some platforms is used to set (=true) in order to get the tclass control message heeader. On others, tclass is set in order to get tclass control message heeader.
router_alert integer() yes yes type = raw
rthdr boolean() yes yes type = dgram | raw, requires superuser privileges to update
tclass boolean() yes yes type = dgram | raw. On some platforms is used to set (=true) in order to get the tclass control message heeader. On others, recvtclass is set in order to get tclass control message heeader.
unicast_hops default | uint8() yes yes none
v6only boolean() yes no none

Table 2.5:   ipv6 options

Options for level tcp:

Option Name Value Type Set Get Other Requirements and comments
congestion string() yes yes none
maxseg integer() yes yes Set not allowed on all platforms.
nodelay boolean() yes yes none

Table 2.6:   tcp options

Options for level udp:

Option Name Value Type Set Get Other Requirements and comments
cork boolean() yes yes none

Table 2.7:   udp options

Options for level sctp:

Option Name Value Type Set Get Other Requirements and comments
associnfo sctp_assocparams() yes yes none
autoclose non_neg_integer() yes yes none
disable_fragments boolean() yes yes none
events sctp_event_subscribe() yes no none
initmsg sctp_initmsg() yes yes none
maxseg non_neg_integer() yes yes none
nodelay boolean() yes yes none
rtoinfo sctp_rtoinfo() yes yes none

Table 2.8:   sctp options