/*      $NetBSD: bpf_stub.c,v 1.8 2018/06/25 03:22:14 msaitoh Exp $     */

/*
* Copyright (c) 2010 The NetBSD Foundation, Inc.
* 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 disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. 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 DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION 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.
*/

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: bpf_stub.c,v 1.8 2018/06/25 03:22:14 msaitoh Exp $");

#include <sys/param.h>
#include <sys/kmem.h>
#include <sys/mbuf.h>

#include <net/bpf.h>

struct laglist {
       struct ifnet *lag_ifp;
       u_int lag_dlt;
       u_int lag_hlen;
       struct bpf_if **lag_drvp;

       TAILQ_ENTRY(laglist) lag_entries;
};

static TAILQ_HEAD(, laglist) lagdrvs = TAILQ_HEAD_INITIALIZER(lagdrvs);

static void bpf_stub_attach(struct ifnet *, u_int, u_int, struct bpf_if **);
static void bpf_stub_detach(struct ifnet *);

static void bpf_stub_null(void);
static void bpf_stub_warn(void);

static kmutex_t handovermtx;
static kcondvar_t handovercv;
static bool handover;

struct bpf_ops bpf_ops_stub = {
       .bpf_attach =           bpf_stub_attach,
       .bpf_detach =           bpf_stub_detach,
       .bpf_change_type =      (void *)bpf_stub_null,

       .bpf_mtap =             (void *)bpf_stub_warn,
       .bpf_mtap2 =            (void *)bpf_stub_warn,
       .bpf_mtap_af =          (void *)bpf_stub_warn,
       .bpf_mtap_sl_in =       (void *)bpf_stub_warn,
       .bpf_mtap_sl_out =      (void *)bpf_stub_warn,

       .bpf_mtap_softint_init =        (void *)bpf_stub_null,
       .bpf_mtap_softint =             (void *)bpf_stub_warn,
};
struct bpf_ops *bpf_ops;

static void
bpf_stub_attach(struct ifnet *ifp, u_int dlt, u_int hlen, struct bpf_if **drvp)
{
       struct laglist *lag;
       bool storeattach = true;

       lag = kmem_alloc(sizeof(*lag), KM_SLEEP);
       lag->lag_ifp = ifp;
       lag->lag_dlt = dlt;
       lag->lag_hlen = hlen;
       lag->lag_drvp = drvp;

       mutex_enter(&handovermtx);
       /*
        * If handover is in progress, wait for it to finish and complete
        * attach after that.  Otherwise record ourselves.
        */
       while (handover) {
               storeattach = false;
               cv_wait(&handovercv, &handovermtx);
       }

       if (storeattach == false) {
               mutex_exit(&handovermtx);
               kmem_free(lag, sizeof(*lag));
               KASSERT(bpf_ops != &bpf_ops_stub); /* revisit when unloadable */
               bpf_ops->bpf_attach(ifp, dlt, hlen, drvp);
       } else {
               *drvp = NULL;
               TAILQ_INSERT_TAIL(&lagdrvs, lag, lag_entries);
               mutex_exit(&handovermtx);
       }
}

static void
bpf_stub_detach(struct ifnet *ifp)
{
       TAILQ_HEAD(, laglist) rmlist;
       struct laglist *lag, *lag_next;
       bool didhand;

       TAILQ_INIT(&rmlist);

       didhand = false;
       mutex_enter(&handovermtx);
       while (handover) {
               didhand = true;
               cv_wait(&handovercv, &handovermtx);
       }

       if (didhand == false) {
               /* atomically remove all */
               for (lag = TAILQ_FIRST(&lagdrvs); lag; lag = lag_next) {
                       lag_next = TAILQ_NEXT(lag, lag_entries);
                       if (lag->lag_ifp == ifp) {
                               TAILQ_REMOVE(&lagdrvs, lag, lag_entries);
                               TAILQ_INSERT_HEAD(&rmlist, lag, lag_entries);
                       }
               }
               mutex_exit(&handovermtx);
               while ((lag = TAILQ_FIRST(&rmlist)) != NULL) {
                       TAILQ_REMOVE(&rmlist, lag, lag_entries);
                       kmem_free(lag, sizeof(*lag));
               }
       } else {
               mutex_exit(&handovermtx);
               KASSERT(bpf_ops != &bpf_ops_stub); /* revisit when unloadable */
               bpf_ops->bpf_detach(ifp);
       }
}

static void
bpf_stub_null(void)
{

}

static void
bpf_stub_warn(void)
{

#ifdef DEBUG
       panic("bpf method called without attached bpf_if");
#endif
#ifdef DIAGNOSTIC
       printf("bpf method called without attached bpf_if\n");
#endif
}

void
bpf_setops(void)
{

       mutex_init(&handovermtx, MUTEX_DEFAULT, IPL_NONE);
       cv_init(&handovercv, "bpfops");
       bpf_ops = &bpf_ops_stub;
}

/*
* Party's over, prepare for handover.
* It needs to happen *before* bpf_ops is set to make it atomic
* to callers (see also stub implementations, which wait if
* called during handover).  The likelyhood of seeing a full
* attach-detach *during* handover comes close to astronomical,
* but handle it anyway since it's relatively easy.
*/
void
bpf_ops_handover_enter(struct bpf_ops *newops)
{
       struct laglist *lag;

       mutex_enter(&handovermtx);
       handover = true;

       while ((lag = TAILQ_FIRST(&lagdrvs)) != NULL) {
               TAILQ_REMOVE(&lagdrvs, lag, lag_entries);
               mutex_exit(&handovermtx);
               newops->bpf_attach(lag->lag_ifp, lag->lag_dlt,
                   lag->lag_hlen, lag->lag_drvp);
               kmem_free(lag, sizeof(*lag));
               mutex_enter(&handovermtx);
       }
       mutex_exit(&handovermtx);
}

/* hangover done */
void
bpf_ops_handover_exit(void)
{

       mutex_enter(&handovermtx);
       handover = false;
       cv_broadcast(&handovercv);
       mutex_exit(&handovermtx);
}