freebsd-skq/sys/net/if_edsc.c
bz 7a1c0b1ad1 Get closer to a VIMAGE network stack teardown from top to bottom rather
than removing the network interfaces first. This change is rather larger
and convoluted as the ordering requirements cannot be separated.

Move the pfil(9) framework to SI_SUB_PROTO_PFIL, move Firewalls and
related modules to their own SI_SUB_PROTO_FIREWALL.
Move initialization of "physical" interfaces to SI_SUB_DRIVERS,
move virtual (cloned) interfaces to SI_SUB_PSEUDO.
Move Multicast to SI_SUB_PROTO_MC.

Re-work parts of multicast initialisation and teardown, not taking the
huge amount of memory into account if used as a module yet.

For interface teardown we try to do as many of them as we can on
SI_SUB_INIT_IF, but for some this makes no sense, e.g., when tunnelling
over a higher layer protocol such as IP. In that case the interface
has to go along (or before) the higher layer protocol is shutdown.

Kernel hhooks need to go last on teardown as they may be used at various
higher layers and we cannot remove them before we cleaned up the higher
layers.

For interface teardown there are multiple paths:
(a) a cloned interface is destroyed (inside a VIMAGE or in the base system),
(b) any interface is moved from a virtual network stack to a different
network stack ("vmove"), or (c) a virtual network stack is being shut down.
All code paths go through if_detach_internal() where we, depending on the
vmove flag or the vnet state, make a decision on how much to shut down;
in case we are destroying a VNET the individual protocol layers will
cleanup their own parts thus we cannot do so again for each interface as
we end up with, e.g., double-frees, destroying locks twice or acquiring
already destroyed locks.
When calling into protocol cleanups we equally have to tell them
whether they need to detach upper layer protocols ("ulp") or not
(e.g., in6_ifdetach()).

Provide or enahnce helper functions to do proper cleanup at a protocol
rather than at an interface level.

Approved by:		re (hrs)
Obtained from:		projects/vnet
Reviewed by:		gnn, jhb
Sponsored by:		The FreeBSD Foundation
MFC after:		2 weeks
Differential Revision:	https://reviews.freebsd.org/D6747
2016-06-21 13:48:49 +00:00

370 lines
9.9 KiB
C

/*-
* Copyright (c) 1982, 1986, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following edsclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following edsclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE EDSCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* From: @(#)if_loop.c 8.1 (Berkeley) 6/10/93
* $FreeBSD$
*/
/*
* Discard interface driver for protocol testing and timing.
* Mimics an Ethernet device so that VLANs can be attached to it etc.
*/
#include <sys/param.h> /* types, important constants */
#include <sys/kernel.h> /* SYSINIT for load-time initializations */
#include <sys/malloc.h> /* malloc(9) */
#include <sys/module.h> /* module(9) */
#include <sys/mbuf.h> /* mbuf(9) */
#include <sys/socket.h> /* struct ifreq */
#include <sys/sockio.h> /* socket ioctl's */
/* #include <sys/systm.h> if you need printf(9) or other all-purpose globals */
#include <net/bpf.h> /* bpf(9) */
#include <net/ethernet.h> /* Ethernet related constants and types */
#include <net/if.h>
#include <net/if_var.h> /* basic part of ifnet(9) */
#include <net/if_clone.h> /* network interface cloning */
#include <net/if_types.h> /* IFT_ETHER and friends */
#include <net/if_var.h> /* kernel-only part of ifnet(9) */
#include <net/vnet.h>
static const char edscname[] = "edsc";
/*
* Software configuration of an interface specific to this device type.
*/
struct edsc_softc {
struct ifnet *sc_ifp; /* ptr to generic interface configuration */
/*
* A non-null driver can keep various things here, for instance,
* the hardware revision, cached values of write-only registers, etc.
*/
};
/*
* Attach to the interface cloning framework.
*/
static VNET_DEFINE(struct if_clone *, edsc_cloner);
#define V_edsc_cloner VNET(edsc_cloner)
static int edsc_clone_create(struct if_clone *, int, caddr_t);
static void edsc_clone_destroy(struct ifnet *);
/*
* Interface driver methods.
*/
static void edsc_init(void *dummy);
/* static void edsc_input(struct ifnet *ifp, struct mbuf *m); would be here */
static int edsc_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data);
static void edsc_start(struct ifnet *ifp);
/*
* We'll allocate softc instances from this.
*/
static MALLOC_DEFINE(M_EDSC, edscname, "Ethernet discard interface");
/*
* Create an interface instance.
*/
static int
edsc_clone_create(struct if_clone *ifc, int unit, caddr_t params)
{
struct edsc_softc *sc;
struct ifnet *ifp;
static u_char eaddr[ETHER_ADDR_LEN]; /* 0:0:0:0:0:0 */
/*
* Allocate soft and ifnet structures. Link each to the other.
*/
sc = malloc(sizeof(struct edsc_softc), M_EDSC, M_WAITOK | M_ZERO);
ifp = sc->sc_ifp = if_alloc(IFT_ETHER);
if (ifp == NULL) {
free(sc, M_EDSC);
return (ENOSPC);
}
ifp->if_softc = sc;
/*
* Get a name for this particular interface in its ifnet structure.
*/
if_initname(ifp, edscname, unit);
/*
* Typical Ethernet interface flags: we can do broadcast and
* multicast but can't hear our own broadcasts or multicasts.
*/
ifp->if_flags = IFF_BROADCAST | IFF_MULTICAST | IFF_SIMPLEX;
/*
* We can pretent we have the whole set of hardware features
* because we just discard all packets we get from the upper layer.
* However, the features are disabled initially. They can be
* enabled via edsc_ioctl() when needed.
*/
ifp->if_capabilities =
IFCAP_VLAN_MTU | IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_HWCSUM |
IFCAP_HWCSUM | IFCAP_TSO |
IFCAP_JUMBO_MTU;
ifp->if_capenable = 0;
/*
* Set the interface driver methods.
*/
ifp->if_init = edsc_init;
/* ifp->if_input = edsc_input; */
ifp->if_ioctl = edsc_ioctl;
ifp->if_start = edsc_start;
/*
* Set the maximum output queue length from the global parameter.
*/
ifp->if_snd.ifq_maxlen = ifqmaxlen;
/*
* Do ifnet initializations common to all Ethernet drivers
* and attach to the network interface framework.
* TODO: Pick a non-zero link level address.
*/
ether_ifattach(ifp, eaddr);
/*
* Now we can mark the interface as running, i.e., ready
* for operation.
*/
ifp->if_drv_flags |= IFF_DRV_RUNNING;
return (0);
}
/*
* Destroy an interface instance.
*/
static void
edsc_clone_destroy(struct ifnet *ifp)
{
struct edsc_softc *sc = ifp->if_softc;
/*
* Detach from the network interface framework.
*/
ether_ifdetach(ifp);
/*
* Free memory occupied by ifnet and softc.
*/
if_free(ifp);
free(sc, M_EDSC);
}
/*
* This method is invoked from ether_ioctl() when it's time
* to bring up the hardware.
*/
static void
edsc_init(void *dummy)
{
#if 0 /* what a hardware driver would do here... */
struct edsc_soft *sc = (struct edsc_softc *)dummy;
struct ifnet *ifp = sc->sc_ifp;
/* blah-blah-blah */
#endif
}
/*
* Network interfaces are controlled via the ioctl(2) syscall.
*/
static int
edsc_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
struct ifreq *ifr = (struct ifreq *)data;
switch (cmd) {
case SIOCSIFCAP:
#if 1
/*
* Just turn on any capabilities requested.
* The generic ifioctl() function has already made sure
* that they are supported, i.e., set in if_capabilities.
*/
ifp->if_capenable = ifr->ifr_reqcap;
#else
/*
* A h/w driver would need to analyze the requested
* bits and program the hardware, e.g.:
*/
mask = ifp->if_capenable ^ ifr->ifr_reqcap;
if (mask & IFCAP_VLAN_HWTAGGING) {
ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
if (ifp->if_capenable & IFCAP_VLAN_HWTAGGING)
/* blah-blah-blah */
else
/* etc-etc-etc */
}
#endif
break;
default:
/*
* Offload the rest onto the common Ethernet handler.
*/
return (ether_ioctl(ifp, cmd, data));
}
return (0);
}
/*
* Process the output queue.
*/
static void
edsc_start(struct ifnet *ifp)
{
struct mbuf *m;
/*
* A hardware interface driver can set IFF_DRV_OACTIVE
* in ifp->if_drv_flags:
*
* ifp->if_drv_flags |= IFF_DRV_OACTIVE;
*
* to prevent if_start from being invoked again while the
* transmission is under way. The flag is to protect the
* device's transmitter, not the method itself. The output
* queue is locked and several threads can process it in
* parallel safely, so the driver can use other means to
* serialize access to the transmitter.
*
* If using IFF_DRV_OACTIVE, the driver should clear the flag
* not earlier than the current transmission is complete, e.g.,
* upon an interrupt from the device, not just before returning
* from if_start. This method merely starts the transmission,
* which may proceed asynchronously.
*/
/*
* We loop getting packets from the queue until it's empty.
* A h/w driver would loop until the device can accept more
* data into its buffer, or while there are free transmit
* descriptors, or whatever.
*/
for (;;) {
/*
* Try to dequeue one packet. Stop if the queue is empty.
* Use IF_DEQUEUE() here if ALTQ(9) support is unneeded.
*/
IFQ_DEQUEUE(&ifp->if_snd, m);
if (m == NULL)
break;
/*
* Let bpf(9) at the packet.
*/
BPF_MTAP(ifp, m);
/*
* Update the interface counters.
*/
if_inc_counter(ifp, IFCOUNTER_OBYTES, m->m_pkthdr.len);
if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1);
/*
* Finally, just drop the packet.
* TODO: Reply to ARP requests unless IFF_NOARP is set.
*/
m_freem(m);
}
/*
* ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
* would be here only if the transmission were synchronous.
*/
}
static void
vnet_edsc_init(const void *unused __unused)
{
/*
* Connect to the network interface cloning framework.
* The last argument is the number of units to be created
* from the outset. It's also the minimum number of units
* allowed. We don't want any units created as soon as the
* driver is loaded.
*/
V_edsc_cloner = if_clone_simple(edscname, edsc_clone_create,
edsc_clone_destroy, 0);
}
VNET_SYSINIT(vnet_edsc_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
vnet_edsc_init, NULL);
static void
vnet_edsc_uninit(const void *unused __unused)
{
/*
* Disconnect from the cloning framework.
* Existing interfaces will be disposed of properly.
*/
if_clone_detach(V_edsc_cloner);
}
VNET_SYSUNINIT(vnet_edsc_uninit, SI_SUB_INIT_IF, SI_ORDER_ANY,
vnet_edsc_uninit, NULL);
/*
* This function provides handlers for module events, namely load and unload.
*/
static int
edsc_modevent(module_t mod, int type, void *data)
{
switch (type) {
case MOD_LOAD:
case MOD_UNLOAD:
break;
default:
/*
* There are other event types, but we don't handle them.
* See module(9).
*/
return (EOPNOTSUPP);
}
return (0);
}
static moduledata_t edsc_mod = {
"if_edsc", /* name */
edsc_modevent, /* event handler */
NULL /* additional data */
};
DECLARE_MODULE(if_edsc, edsc_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);