freebsd-dev/share/man/man4/tap.4
Kyle Evans 2a345f61da tap(4): also note that we drop configured addresses
This provides a specific pointer for users of tap(4) to understand why their
interfaces are losing their addresses, and specifically how to workaround
this if they need different behavior.

This manpage received a .Dd bump earlier today in r35688, so no bump occurs
this time.

Submitted by:	sigsys@gmail.com (via IRC)
2020-01-13 18:26:27 +00:00

320 lines
7.2 KiB
Groff

.\" $FreeBSD$
.\" Based on PR#2411
.\"
.Dd January 13, 2020
.Dt TAP 4
.Os
.Sh NAME
.Nm tap ,
.Nm vmnet
.Nd Ethernet tunnel software network interface
.Sh SYNOPSIS
.Cd device tuntap
.Sh DESCRIPTION
The
.Nm
interface is a software loopback mechanism that can be loosely
described as the network interface analog of the
.Xr pty 4 ,
that is,
.Nm
does for network interfaces what the
.Xr pty 4
driver does for terminals.
.Pp
The
.Nm
driver, like the
.Xr pty 4
driver, provides two interfaces: an interface like the usual facility
it is simulating
(an Ethernet network interface in the case of
.Nm ,
or a terminal for
.Xr pty 4 ) ,
and a character-special device
.Dq control
interface.
A client program transfers Ethernet frames to or from the
.Nm
.Dq control
interface.
The
.Xr tun 4
interface provides similar functionality at the network layer:
a client will transfer IP (by default) packets to or from a
.Xr tun 4
.Dq control
interface.
.Pp
The network interfaces are named
.Dq Li tap0 ,
.Dq Li tap1 ,
etc., one for each control device that has been opened.
These Ethernet network interfaces persist until
.Pa if_tuntap.ko
module is unloaded, or until removed with "ifconfig destroy" (see below).
.Pp
.Nm
devices are created using interface cloning.
This is done using the
.Dq ifconfig tap Ns Sy N No create
command.
This is the preferred method of creating
.Nm
devices.
The same method allows removal of interfaces.
For this, use the
.Dq ifconfig tap Ns Sy N No destroy
command.
.Pp
If the
.Xr sysctl 8
variable
.Va net.link.tap.devfs_cloning
is non-zero, the
.Nm
interface
permits opens on the special control device
.Pa /dev/tap .
When this device is opened,
.Nm
will return a handle for the lowest unused
.Nm
device (use
.Xr devname 3
to determine which).
.Pp
.Bf Em
Disabling the legacy devfs cloning functionality may break existing
applications which use
.Nm ,
such as
.Tn VMware
and
.Xr ssh 1 .
It therefore defaults to being enabled until further notice.
.Ef
.Pp
Control devices (once successfully opened) persist until
.Pa if_tuntap.ko
is unloaded or the interface is destroyed.
.Pp
Each interface supports the usual Ethernet network interface
.Xr ioctl 2 Ns s
and thus can be used with
.Xr ifconfig 8
like any other Ethernet interface.
When the system chooses to transmit
an Ethernet frame on the network interface, the frame can be read from
the control device
(it appears as
.Dq input
there);
writing an Ethernet frame to the control device generates an input frame on
the network interface, as if the
(non-existent)
hardware had just received it.
.Pp
The Ethernet tunnel device, normally
.Pa /dev/tap Ns Sy N ,
is exclusive-open
(it cannot be opened if it is already open)
and is restricted to the super-user, unless the
.Xr sysctl 8
variable
.Va net.link.tap.user_open
is non-zero.
If the
.Xr sysctl 8
variable
.Va net.link.tap.up_on_open
is non-zero, the tunnel device will be marked
.Dq up
when the control device is opened.
A
.Fn read
call will return an error
.Pq Er EHOSTDOWN
if the interface is not
.Dq ready .
Once the interface is ready,
.Fn read
will return an Ethernet frame if one is available; if not, it will
either block until one is or return
.Er EWOULDBLOCK ,
depending on whether non-blocking I/O has been enabled.
If the frame
is longer than is allowed for in the buffer passed to
.Fn read ,
the extra data will be silently dropped.
.Pp
A
.Xr write 2
call passes an Ethernet frame in to be
.Dq received
on the pseudo-interface.
Each
.Fn write
call supplies exactly one frame; the frame length is taken from the
amount of data provided to
.Fn write .
Writes will not block; if the frame cannot be accepted
for a transient reason
(e.g., no buffer space available),
it is silently dropped; if the reason is not transient
(e.g., frame too large),
an error is returned.
The following
.Xr ioctl 2
calls are supported
(defined in
.In net/if_tap.h ) :
.Bl -tag -width VMIO_SIOCSETMACADDR
.It Dv TAPSIFINFO
Set network interface information (line speed and MTU).
The type must be the same as returned by
.Dv TAPGIFINFO
or set to
.Dv IFT_ETHER
else the
.Xr ioctl 2
call will fail.
The argument should be a pointer to a
.Va struct tapinfo .
.It Dv TAPGIFINFO
Retrieve network interface information (line speed, MTU and type).
The argument should be a pointer to a
.Va struct tapinfo .
.It Dv TAPSDEBUG
The argument should be a pointer to an
.Va int ;
this sets the internal debugging variable to that value.
What, if
anything, this variable controls is not documented here; see the source
code.
.It Dv TAPGDEBUG
The argument should be a pointer to an
.Va int ;
this stores the internal debugging variable's value into it.
.It Dv TAPGIFNAME
Retrieve network interface name.
The argument should be a pointer to a
.Va struct ifreq .
The interface name will be returned in the
.Va ifr_name
field.
.It Dv FIONBIO
Turn non-blocking I/O for reads off or on, according as the argument
.Va int Ns 's
value is or is not zero
(Writes are always nonblocking).
.It Dv FIOASYNC
Turn asynchronous I/O for reads
(i.e., generation of
.Dv SIGIO
when data is available to be read)
off or on, according as the argument
.Va int Ns 's
value is or is not zero.
.It Dv FIONREAD
If any frames are queued to be read, store the size of the first one into the argument
.Va int ;
otherwise, store zero.
.It Dv TIOCSPGRP
Set the process group to receive
.Dv SIGIO
signals, when asynchronous I/O is enabled, to the argument
.Va int
value.
.It Dv TIOCGPGRP
Retrieve the process group value for
.Dv SIGIO
signals into the argument
.Va int
value.
.It Dv SIOCGIFADDR
Retrieve the Media Access Control
.Pq Dv MAC
address of the
.Dq remote
side.
This command is used by the VMware port and expected to be executed on
descriptor, associated with control device
(usually
.Pa /dev/vmnet Ns Sy N
or
.Pa /dev/tap Ns Sy N ) .
The
.Va buffer ,
which is passed as the argument, is expected to have enough space to store
the
.Dv MAC
address.
At the open time both
.Dq local
and
.Dq remote
.Dv MAC
addresses are the same, so this command could be used to retrieve the
.Dq local
.Dv MAC
address.
.It Dv SIOCSIFADDR
Set the Media Access Control
.Pq Dv MAC
address of the
.Dq remote
side.
This command is used by VMware port and expected to be executed on
a descriptor, associated with control device
(usually
.Pa /dev/vmnet Ns Sy N ) .
.El
.Pp
The control device also supports
.Xr select 2
for read; selecting for write is pointless, and always succeeds, since
writes are always non-blocking.
.Pp
On the last close of the data device, the interface is
brought down
(as if with
.Dq ifconfig tap Ns Sy N No down )
and has all of its configured addresses deleted
unless the device is a
.Em VMnet
device, or has
.Dv IFF_LINK0
flag set.
All queued frames are thrown away.
If the interface is up when the data
device is not open, output frames are thrown away rather than
letting them pile up.
.Pp
The
.Nm
device can also be used with the VMware port as a replacement
for the old
.Em VMnet
device driver.
.Em VMnet
devices do not
.Xr ifconfig 8
themselves down when the
control device is closed.
Everything else is the same.
.Pp
In addition to the above mentioned
.Xr ioctl 2
calls, there is an additional one for the VMware port.
.Bl -tag -width VMIO_SIOCSETMACADDR
.It Dv VMIO_SIOCSIFFLAGS
VMware
.Dv SIOCSIFFLAGS .
.El
.Sh SEE ALSO
.Xr inet 4 ,
.Xr intro 4 ,
.Xr tun 4