diff options
Diffstat (limited to 'sys/net')
38 files changed, 15617 insertions, 0 deletions
diff --git a/sys/net/bpf.c b/sys/net/bpf.c new file mode 100644 index 00000000000..aa3cadb62a2 --- /dev/null +++ b/sys/net/bpf.c @@ -0,0 +1,1365 @@ +/* $NetBSD: bpf.c,v 1.23 1995/09/27 18:30:37 thorpej Exp $ */ + +/* + * Copyright (c) 1990, 1991, 1993 + * The Regents of the University of California. All rights reserved. + * + * This code is derived from the Stanford/CMU enet packet filter, + * (net/enet.c) distributed as part of 4.3BSD, and code contributed + * to Berkeley by Steven McCanne and Van Jacobson both of Lawrence + * Berkeley Laboratory. + * + * 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)bpf.c 8.2 (Berkeley) 3/28/94 + */ + +#include "bpfilter.h" + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/mbuf.h> +#include <sys/buf.h> +#include <sys/time.h> +#include <sys/proc.h> +#include <sys/user.h> +#include <sys/ioctl.h> +#include <sys/map.h> + +#include <sys/file.h> +#if defined(sparc) && BSD < 199103 +#include <sys/stream.h> +#endif +#include <sys/tty.h> +#include <sys/uio.h> + +#include <sys/protosw.h> +#include <sys/socket.h> +#include <net/if.h> + +#include <net/bpf.h> +#include <net/bpfdesc.h> + +#include <sys/errno.h> + +#include <netinet/in.h> +#include <netinet/if_arc.h> +#include <netinet/if_ether.h> +#include <sys/kernel.h> + +/* + * Older BSDs don't have kernel malloc. + */ +#if BSD < 199103 +extern bcopy(); +static caddr_t bpf_alloc(); +#include <net/bpf_compat.h> +#define BPF_BUFSIZE (MCLBYTES-8) +#define UIOMOVE(cp, len, code, uio) uiomove(cp, len, code, uio) +#else +#define BPF_BUFSIZE 4096 +#define UIOMOVE(cp, len, code, uio) uiomove(cp, len, uio) +#endif + +#define PRINET 26 /* interruptible */ + +/* + * The default read buffer size is patchable. + */ +int bpf_bufsize = BPF_BUFSIZE; + +/* + * bpf_iflist is the list of interfaces; each corresponds to an ifnet + * bpf_dtab holds the descriptors, indexed by minor device # + */ +struct bpf_if *bpf_iflist; +struct bpf_d bpf_dtab[NBPFILTER]; + +#if BSD >= 199207 || NetBSD0_9 >= 2 +/* + * bpfilterattach() is called at boot time in new systems. We do + * nothing here since old systems will not call this. + */ +/* ARGSUSED */ +void +bpfilterattach(n) + int n; +{ +} +#endif + +static int bpf_allocbufs __P((struct bpf_d *)); +static int bpf_allocbufs __P((struct bpf_d *)); +static void bpf_freed __P((struct bpf_d *)); +static void bpf_freed __P((struct bpf_d *)); +static void bpf_ifname __P((struct ifnet *, struct ifreq *)); +static void bpf_ifname __P((struct ifnet *, struct ifreq *)); +static void bpf_mcopy __P((const void *, void *, size_t)); +static int bpf_movein __P((struct uio *, int, + struct mbuf **, struct sockaddr *)); +static int bpf_setif __P((struct bpf_d *, struct ifreq *)); +static int bpf_setif __P((struct bpf_d *, struct ifreq *)); +static __inline void + bpf_wakeup __P((struct bpf_d *)); +static void catchpacket __P((struct bpf_d *, u_char *, size_t, + size_t, void (*)(const void *, void *, size_t))); +static void reset_d __P((struct bpf_d *)); + +static int +bpf_movein(uio, linktype, mp, sockp) + register struct uio *uio; + int linktype; + register struct mbuf **mp; + register struct sockaddr *sockp; +{ + struct mbuf *m; + int error; + int len; + int hlen; + + /* + * Build a sockaddr based on the data link layer type. + * We do this at this level because the ethernet header + * is copied directly into the data field of the sockaddr. + * In the case of SLIP, there is no header and the packet + * is forwarded as is. + * Also, we are careful to leave room at the front of the mbuf + * for the link level header. + */ + switch (linktype) { + + case DLT_SLIP: + sockp->sa_family = AF_INET; + hlen = 0; + break; + + case DLT_PPP: + sockp->sa_family = AF_UNSPEC; + hlen = 0; + break; + + case DLT_EN10MB: + sockp->sa_family = AF_UNSPEC; + /* XXX Would MAXLINKHDR be better? */ + hlen = sizeof(struct ether_header); + break; + + case DLT_ARCNET: + sockp->sa_family = AF_UNSPEC; + hlen = ARC_HDRLEN; + break; + + case DLT_FDDI: + sockp->sa_family = AF_UNSPEC; + /* XXX 4(FORMAC)+6(dst)+6(src)+3(LLC)+5(SNAP) */ + hlen = 24; + break; + + case DLT_NULL: + sockp->sa_family = AF_UNSPEC; + hlen = 0; + break; + + default: + return (EIO); + } + + len = uio->uio_resid; + if ((unsigned)len > MCLBYTES) + return (EIO); + + MGETHDR(m, M_WAIT, MT_DATA); + if (m == 0) + return (ENOBUFS); + m->m_pkthdr.rcvif = 0; + m->m_pkthdr.len = len - hlen; + + if (len > MHLEN) { +#if BSD >= 199103 + MCLGET(m, M_WAIT); + if ((m->m_flags & M_EXT) == 0) { +#else + MCLGET(m); + if (m->m_len != MCLBYTES) { +#endif + error = ENOBUFS; + goto bad; + } + } + m->m_len = len; + *mp = m; + /* + * Make room for link header. + */ + if (hlen != 0) { + m->m_len -= hlen; +#if BSD >= 199103 + m->m_data += hlen; /* XXX */ +#else + m->m_off += hlen; +#endif + error = UIOMOVE((caddr_t)sockp->sa_data, hlen, UIO_WRITE, uio); + if (error) + goto bad; + } + error = UIOMOVE(mtod(m, caddr_t), len - hlen, UIO_WRITE, uio); + if (!error) + return (0); + bad: + m_freem(m); + return (error); +} + +/* + * Attach file to the bpf interface, i.e. make d listen on bp. + * Must be called at splimp. + */ +static void +bpf_attachd(d, bp) + struct bpf_d *d; + struct bpf_if *bp; +{ + /* + * Point d at bp, and add d to the interface's list of listeners. + * Finally, point the driver's bpf cookie at the interface so + * it will divert packets to bpf. + */ + d->bd_bif = bp; + d->bd_next = bp->bif_dlist; + bp->bif_dlist = d; + + *bp->bif_driverp = bp; +} + +/* + * Detach a file from its interface. + */ +static void +bpf_detachd(d) + struct bpf_d *d; +{ + struct bpf_d **p; + struct bpf_if *bp; + + bp = d->bd_bif; + /* + * Check if this descriptor had requested promiscuous mode. + * If so, turn it off. + */ + if (d->bd_promisc) { + int error; + + d->bd_promisc = 0; + error = ifpromisc(bp->bif_ifp, 0); + if (error && error != EINVAL) + /* + * Something is really wrong if we were able to put + * the driver into promiscuous mode, but can't + * take it out. + */ + panic("bpf: ifpromisc failed"); + } + /* Remove d from the interface's descriptor list. */ + p = &bp->bif_dlist; + while (*p != d) { + p = &(*p)->bd_next; + if (*p == 0) + panic("bpf_detachd: descriptor not in list"); + } + *p = (*p)->bd_next; + if (bp->bif_dlist == 0) + /* + * Let the driver know that there are no more listeners. + */ + *d->bd_bif->bif_driverp = 0; + d->bd_bif = 0; +} + + +/* + * Mark a descriptor free by making it point to itself. + * This is probably cheaper than marking with a constant since + * the address should be in a register anyway. + */ +#define D_ISFREE(d) ((d) == (d)->bd_next) +#define D_MARKFREE(d) ((d)->bd_next = (d)) +#define D_MARKUSED(d) ((d)->bd_next = 0) + +/* + * Open ethernet device. Returns ENXIO for illegal minor device number, + * EBUSY if file is open by another process. + */ +/* ARGSUSED */ +int +bpfopen(dev, flag) + dev_t dev; + int flag; +{ + register struct bpf_d *d; + + if (minor(dev) >= NBPFILTER) + return (ENXIO); + /* + * Each minor can be opened by only one process. If the requested + * minor is in use, return EBUSY. + */ + d = &bpf_dtab[minor(dev)]; + if (!D_ISFREE(d)) + return (EBUSY); + + /* Mark "free" and do most initialization. */ + bzero((char *)d, sizeof(*d)); + d->bd_bufsize = bpf_bufsize; + d->bd_sig = SIGIO; + + return (0); +} + +/* + * Close the descriptor by detaching it from its interface, + * deallocating its buffers, and marking it free. + */ +/* ARGSUSED */ +int +bpfclose(dev, flag) + dev_t dev; + int flag; +{ + register struct bpf_d *d = &bpf_dtab[minor(dev)]; + register int s; + + s = splimp(); + if (d->bd_bif) + bpf_detachd(d); + splx(s); + bpf_freed(d); + + return (0); +} + +/* + * Support for SunOS, which does not have tsleep. + */ +#if BSD < 199103 +static +bpf_timeout(arg) + caddr_t arg; +{ + struct bpf_d *d = (struct bpf_d *)arg; + d->bd_timedout = 1; + wakeup(arg); +} + +#define BPF_SLEEP(chan, pri, s, t) bpf_sleep((struct bpf_d *)chan) + +int +bpf_sleep(d) + register struct bpf_d *d; +{ + register int rto = d->bd_rtout; + register int st; + + if (rto != 0) { + d->bd_timedout = 0; + timeout(bpf_timeout, (caddr_t)d, rto); + } + st = sleep((caddr_t)d, PRINET|PCATCH); + if (rto != 0) { + if (d->bd_timedout == 0) + untimeout(bpf_timeout, (caddr_t)d); + else if (st == 0) + return EWOULDBLOCK; + } + return (st != 0) ? EINTR : 0; +} +#else +#define BPF_SLEEP tsleep +#endif + +/* + * Rotate the packet buffers in descriptor d. Move the store buffer + * into the hold slot, and the free buffer into the store slot. + * Zero the length of the new store buffer. + */ +#define ROTATE_BUFFERS(d) \ + (d)->bd_hbuf = (d)->bd_sbuf; \ + (d)->bd_hlen = (d)->bd_slen; \ + (d)->bd_sbuf = (d)->bd_fbuf; \ + (d)->bd_slen = 0; \ + (d)->bd_fbuf = 0; +/* + * bpfread - read next chunk of packets from buffers + */ +int +bpfread(dev, uio) + dev_t dev; + register struct uio *uio; +{ + register struct bpf_d *d = &bpf_dtab[minor(dev)]; + int error; + int s; + + /* + * Restrict application to use a buffer the same size as + * as kernel buffers. + */ + if (uio->uio_resid != d->bd_bufsize) + return (EINVAL); + + s = splimp(); + /* + * If the hold buffer is empty, then do a timed sleep, which + * ends when the timeout expires or when enough packets + * have arrived to fill the store buffer. + */ + while (d->bd_hbuf == 0) { + if (d->bd_immediate && d->bd_slen != 0) { + /* + * A packet(s) either arrived since the previous + * read or arrived while we were asleep. + * Rotate the buffers and return what's here. + */ + ROTATE_BUFFERS(d); + break; + } + if (d->bd_rtout != -1) + error = BPF_SLEEP((caddr_t)d, PRINET|PCATCH, "bpf", + d->bd_rtout); + else + error = EWOULDBLOCK; /* User requested non-blocking I/O */ + if (error == EINTR || error == ERESTART) { + splx(s); + return (error); + } + if (error == EWOULDBLOCK) { + /* + * On a timeout, return what's in the buffer, + * which may be nothing. If there is something + * in the store buffer, we can rotate the buffers. + */ + if (d->bd_hbuf) + /* + * We filled up the buffer in between + * getting the timeout and arriving + * here, so we don't need to rotate. + */ + break; + + if (d->bd_slen == 0) { + splx(s); + return (0); + } + ROTATE_BUFFERS(d); + break; + } + } + /* + * At this point, we know we have something in the hold slot. + */ + splx(s); + + /* + * Move data from hold buffer into user space. + * We know the entire buffer is transferred since + * we checked above that the read buffer is bpf_bufsize bytes. + */ + error = UIOMOVE(d->bd_hbuf, d->bd_hlen, UIO_READ, uio); + + s = splimp(); + d->bd_fbuf = d->bd_hbuf; + d->bd_hbuf = 0; + d->bd_hlen = 0; + splx(s); + + return (error); +} + + +/* + * If there are processes sleeping on this descriptor, wake them up. + */ +static __inline void +bpf_wakeup(d) + register struct bpf_d *d; +{ + struct proc *p; + + wakeup((caddr_t)d); + if (d->bd_async && d->bd_sig) + if (d->bd_pgid > 0) + gsignal (d->bd_pgid, d->bd_sig); + else if (p = pfind (-d->bd_pgid)) + psignal (p, d->bd_sig); + +#if BSD >= 199103 + selwakeup(&d->bd_sel); + /* XXX */ + d->bd_sel.si_pid = 0; +#else + if (d->bd_selproc) { + selwakeup(d->bd_selproc, (int)d->bd_selcoll); + d->bd_selcoll = 0; + d->bd_selproc = 0; + } +#endif +} + +int +bpfwrite(dev, uio) + dev_t dev; + struct uio *uio; +{ + register struct bpf_d *d = &bpf_dtab[minor(dev)]; + struct ifnet *ifp; + struct mbuf *m; + int error, s; + static struct sockaddr dst; + + if (d->bd_bif == 0) + return (ENXIO); + + ifp = d->bd_bif->bif_ifp; + + if (uio->uio_resid == 0) + return (0); + + error = bpf_movein(uio, (int)d->bd_bif->bif_dlt, &m, &dst); + if (error) + return (error); + + if (m->m_pkthdr.len > ifp->if_mtu) + return (EMSGSIZE); + + s = splsoftnet(); +#if BSD >= 199103 + error = (*ifp->if_output)(ifp, m, &dst, (struct rtentry *)0); +#else + error = (*ifp->if_output)(ifp, m, &dst); +#endif + splx(s); + /* + * The driver frees the mbuf. + */ + return (error); +} + +/* + * Reset a descriptor by flushing its packet buffer and clearing the + * receive and drop counts. Should be called at splimp. + */ +static void +reset_d(d) + struct bpf_d *d; +{ + if (d->bd_hbuf) { + /* Free the hold buffer. */ + d->bd_fbuf = d->bd_hbuf; + d->bd_hbuf = 0; + } + d->bd_slen = 0; + d->bd_hlen = 0; + d->bd_rcount = 0; + d->bd_dcount = 0; +} + +/* + * FIONREAD Check for read packet available. + * BIOCGBLEN Get buffer len [for read()]. + * BIOCSETF Set ethernet read filter. + * BIOCFLUSH Flush read packet buffer. + * BIOCPROMISC Put interface into promiscuous mode. + * BIOCGDLT Get link layer type. + * BIOCGETIF Get interface name. + * BIOCSETIF Set interface. + * BIOCSRTIMEOUT Set read timeout. + * BIOCGRTIMEOUT Get read timeout. + * BIOCGSTATS Get packet stats. + * BIOCIMMEDIATE Set immediate mode. + * BIOCVERSION Get filter language version. + */ +/* ARGSUSED */ +int +bpfioctl(dev, cmd, addr, flag) + dev_t dev; + u_long cmd; + caddr_t addr; + int flag; +{ + register struct bpf_d *d = &bpf_dtab[minor(dev)]; + int s, error = 0; + + switch (cmd) { + + default: + error = EINVAL; + break; + + /* + * Check for read packet available. + */ + case FIONREAD: + { + int n; + + s = splimp(); + n = d->bd_slen; + if (d->bd_hbuf) + n += d->bd_hlen; + splx(s); + + *(int *)addr = n; + break; + } + + /* + * Get buffer len [for read()]. + */ + case BIOCGBLEN: + *(u_int *)addr = d->bd_bufsize; + break; + + /* + * Set buffer length. + */ + case BIOCSBLEN: +#if BSD < 199103 + error = EINVAL; +#else + if (d->bd_bif != 0) + error = EINVAL; + else { + register u_int size = *(u_int *)addr; + + if (size > BPF_MAXBUFSIZE) + *(u_int *)addr = size = BPF_MAXBUFSIZE; + else if (size < BPF_MINBUFSIZE) + *(u_int *)addr = size = BPF_MINBUFSIZE; + d->bd_bufsize = size; + } +#endif + break; + + /* + * Set link layer read filter. + */ + case BIOCSETF: + error = bpf_setf(d, (struct bpf_program *)addr); + break; + + /* + * Flush read packet buffer. + */ + case BIOCFLUSH: + s = splimp(); + reset_d(d); + splx(s); + break; + + /* + * Put interface into promiscuous mode. + */ + case BIOCPROMISC: + if (d->bd_bif == 0) { + /* + * No interface attached yet. + */ + error = EINVAL; + break; + } + s = splimp(); + if (d->bd_promisc == 0) { + error = ifpromisc(d->bd_bif->bif_ifp, 1); + if (error == 0) + d->bd_promisc = 1; + } + splx(s); + break; + + /* + * Get device parameters. + */ + case BIOCGDLT: + if (d->bd_bif == 0) + error = EINVAL; + else + *(u_int *)addr = d->bd_bif->bif_dlt; + break; + + /* + * Set interface name. + */ + case BIOCGETIF: + if (d->bd_bif == 0) + error = EINVAL; + else + bpf_ifname(d->bd_bif->bif_ifp, (struct ifreq *)addr); + break; + + /* + * Set interface. + */ + case BIOCSETIF: + error = bpf_setif(d, (struct ifreq *)addr); + break; + + /* + * Set read timeout. + */ + case BIOCSRTIMEOUT: + { + struct timeval *tv = (struct timeval *)addr; + + /* Compute number of ticks. */ + d->bd_rtout = tv->tv_sec * hz + tv->tv_usec / tick; + break; + } + + /* + * Get read timeout. + */ + case BIOCGRTIMEOUT: + { + struct timeval *tv = (struct timeval *)addr; + + tv->tv_sec = d->bd_rtout / hz; + tv->tv_usec = (d->bd_rtout % hz) * tick; + break; + } + + /* + * Get packet stats. + */ + case BIOCGSTATS: + { + struct bpf_stat *bs = (struct bpf_stat *)addr; + + bs->bs_recv = d->bd_rcount; + bs->bs_drop = d->bd_dcount; + break; + } + + /* + * Set immediate mode. + */ + case BIOCIMMEDIATE: + d->bd_immediate = *(u_int *)addr; + break; + + case BIOCVERSION: + { + struct bpf_version *bv = (struct bpf_version *)addr; + + bv->bv_major = BPF_MAJOR_VERSION; + bv->bv_minor = BPF_MINOR_VERSION; + break; + } + + + case FIONBIO: /* Non-blocking I/O */ + if (*(int *)addr) + d->bd_rtout = -1; + else + d->bd_rtout = 0; + break; + + case FIOASYNC: /* Send signal on receive packets */ + d->bd_async = *(int *)addr; + break; + + /* + * N.B. ioctl (FIOSETOWN) and fcntl (F_SETOWN) both end up doing + * the equivalent of a TIOCSPGRP and hence end up here. *However* + * TIOCSPGRP's arg is a process group if it's positive and a process + * id if it's negative. This is exactly the opposite of what the + * other two functions want! Therefore there is code in ioctl and + * fcntl to negate the arg before calling here. + */ + case TIOCSPGRP: /* Process or group to send signals to */ + d->bd_pgid = *(int *)addr; + break; + + case TIOCGPGRP: + *(int *)addr = d->bd_pgid; + break; + + case BIOCSRSIG: /* Set receive signal */ + { + u_int sig; + + sig = *(u_int *)addr; + + if (sig >= NSIG) + error = EINVAL; + else + d->bd_sig = sig; + break; + } + case BIOCGRSIG: + *(u_int *)addr = d->bd_sig; + break; + } + return (error); +} + +/* + * Set d's packet filter program to fp. If this file already has a filter, + * free it and replace it. Returns EINVAL for bogus requests. + */ +int +bpf_setf(d, fp) + struct bpf_d *d; + struct bpf_program *fp; +{ + struct bpf_insn *fcode, *old; + u_int flen, size; + int s; + + old = d->bd_filter; + if (fp->bf_insns == 0) { + if (fp->bf_len != 0) + return (EINVAL); + s = splimp(); + d->bd_filter = 0; + reset_d(d); + splx(s); + if (old != 0) + free((caddr_t)old, M_DEVBUF); + return (0); + } + flen = fp->bf_len; + if (flen > BPF_MAXINSNS) + return (EINVAL); + + size = flen * sizeof(*fp->bf_insns); + fcode = (struct bpf_insn *)malloc(size, M_DEVBUF, M_WAITOK); + if (copyin((caddr_t)fp->bf_insns, (caddr_t)fcode, size) == 0 && + bpf_validate(fcode, (int)flen)) { + s = splimp(); + d->bd_filter = fcode; + reset_d(d); + splx(s); + if (old != 0) + free((caddr_t)old, M_DEVBUF); + + return (0); + } + free((caddr_t)fcode, M_DEVBUF); + return (EINVAL); +} + +/* + * Detach a file from its current interface (if attached at all) and attach + * to the interface indicated by the name stored in ifr. + * Return an errno or 0. + */ +static int +bpf_setif(d, ifr) + struct bpf_d *d; + struct ifreq *ifr; +{ + struct bpf_if *bp; + char *cp; + int unit, s, error; + + /* + * Separate string into name part and unit number. Put a null + * byte at the end of the name part, and compute the number. + * If the a unit number is unspecified, the default is 0, + * as initialized above. XXX This should be common code. + */ + unit = 0; + cp = ifr->ifr_name; + cp[sizeof(ifr->ifr_name) - 1] = '\0'; + while (*cp++) { + if (*cp >= '0' && *cp <= '9') { + unit = *cp - '0'; + *cp++ = '\0'; + while (*cp) + unit = 10 * unit + *cp++ - '0'; + break; + } + } + /* + * Look through attached interfaces for the named one. + */ + for (bp = bpf_iflist; bp != 0; bp = bp->bif_next) { + struct ifnet *ifp = bp->bif_ifp; + + if (ifp == 0 || unit != ifp->if_unit + || strcmp(ifp->if_name, ifr->ifr_name) != 0) + continue; + /* + * We found the requested interface. + * If it's not up, return an error. + * Allocate the packet buffers if we need to. + * If we're already attached to requested interface, + * just flush the buffer. + */ + if ((ifp->if_flags & IFF_UP) == 0) + return (ENETDOWN); + + if (d->bd_sbuf == 0) { + error = bpf_allocbufs(d); + if (error != 0) + return (error); + } + s = splimp(); + if (bp != d->bd_bif) { + if (d->bd_bif) + /* + * Detach if attached to something else. + */ + bpf_detachd(d); + + bpf_attachd(d, bp); + } + reset_d(d); + splx(s); + return (0); + } + /* Not found. */ + return (ENXIO); +} + +/* + * Convert an interface name plus unit number of an ifp to a single + * name which is returned in the ifr. + */ +static void +bpf_ifname(ifp, ifr) + struct ifnet *ifp; + struct ifreq *ifr; +{ + char *s = ifp->if_name; + char *d = ifr->ifr_name; + + while (*d++ = *s++) + continue; + /* XXX Assume that unit number is less than 10. */ + *d++ = ifp->if_unit + '0'; + *d = '\0'; +} + +/* + * The new select interface passes down the proc pointer; the old select + * stubs had to grab it out of the user struct. This glue allows either case. + */ +#if BSD >= 199103 +#define bpf_select bpfselect +#else +int +bpfselect(dev, rw) + register dev_t dev; + int rw; +{ + return (bpf_select(dev, rw, u.u_procp)); +} +#endif + +/* + * Support for select() system call + * + * Return true iff the specific operation will not block indefinitely. + * Otherwise, return false but make a note that a selwakeup() must be done. + */ +int +bpf_select(dev, rw, p) + register dev_t dev; + int rw; + struct proc *p; +{ + register struct bpf_d *d; + register int s; + + if (rw != FREAD) + return (0); + /* + * An imitation of the FIONREAD ioctl code. + */ + d = &bpf_dtab[minor(dev)]; + + s = splimp(); + if (d->bd_hlen != 0 || (d->bd_immediate && d->bd_slen != 0)) { + /* + * There is data waiting. + */ + splx(s); + return (1); + } +#if BSD >= 199103 + selrecord(p, &d->bd_sel); +#else + /* + * No data ready. If there's already a select() waiting on this + * minor device then this is a collision. This shouldn't happen + * because minors really should not be shared, but if a process + * forks while one of these is open, it is possible that both + * processes could select on the same descriptor. + */ + if (d->bd_selproc && d->bd_selproc->p_wchan == (caddr_t)&selwait) + d->bd_selcoll = 1; + else + d->bd_selproc = p; +#endif + splx(s); + return (0); +} + +/* + * Incoming linkage from device drivers. Process the packet pkt, of length + * pktlen, which is stored in a contiguous buffer. The packet is parsed + * by each process' filter, and if accepted, stashed into the corresponding + * buffer. + */ +void +bpf_tap(arg, pkt, pktlen) + caddr_t arg; + register u_char *pkt; + register u_int pktlen; +{ + struct bpf_if *bp; + register struct bpf_d *d; + register size_t slen; + /* + * Note that the ipl does not have to be raised at this point. + * The only problem that could arise here is that if two different + * interfaces shared any data. This is not the case. + */ + bp = (struct bpf_if *)arg; + for (d = bp->bif_dlist; d != 0; d = d->bd_next) { + ++d->bd_rcount; + slen = bpf_filter(d->bd_filter, pkt, pktlen, pktlen); + if (slen != 0) + catchpacket(d, pkt, pktlen, slen, bcopy); + } +} + +/* + * Copy data from an mbuf chain into a buffer. This code is derived + * from m_copydata in sys/uipc_mbuf.c. + */ +static void +bpf_mcopy(src_arg, dst_arg, len) + const void *src_arg; + void *dst_arg; + register size_t len; +{ + register const struct mbuf *m; + register u_int count; + u_char *dst; + + m = src_arg; + dst = dst_arg; + while (len > 0) { + if (m == 0) + panic("bpf_mcopy"); + count = min(m->m_len, len); + bcopy(mtod(m, caddr_t), (caddr_t)dst, count); + m = m->m_next; + dst += count; + len -= count; + } +} + +/* + * Incoming linkage from device drivers, when packet is in an mbuf chain. + */ +void +bpf_mtap(arg, m) + caddr_t arg; + struct mbuf *m; +{ + struct bpf_if *bp = (struct bpf_if *)arg; + struct bpf_d *d; + size_t pktlen, slen; + struct mbuf *m0; + + pktlen = 0; + for (m0 = m; m0 != 0; m0 = m0->m_next) + pktlen += m0->m_len; + + for (d = bp->bif_dlist; d != 0; d = d->bd_next) { + ++d->bd_rcount; + slen = bpf_filter(d->bd_filter, (u_char *)m, pktlen, 0); + if (slen != 0) + catchpacket(d, (u_char *)m, pktlen, slen, bpf_mcopy); + } +} + +/* + * Move the packet data from interface memory (pkt) into the + * store buffer. Return 1 if it's time to wakeup a listener (buffer full), + * otherwise 0. "copy" is the routine called to do the actual data + * transfer. bcopy is passed in to copy contiguous chunks, while + * bpf_mcopy is passed in to copy mbuf chains. In the latter case, + * pkt is really an mbuf. + */ +static void +catchpacket(d, pkt, pktlen, snaplen, cpfn) + register struct bpf_d *d; + register u_char *pkt; + register size_t pktlen, snaplen; + register void (*cpfn) __P((const void *, void *, size_t)); +{ + register struct bpf_hdr *hp; + register int totlen, curlen; + register int hdrlen = d->bd_bif->bif_hdrlen; + /* + * Figure out how many bytes to move. If the packet is + * greater or equal to the snapshot length, transfer that + * much. Otherwise, transfer the whole packet (unless + * we hit the buffer size limit). + */ + totlen = hdrlen + min(snaplen, pktlen); + if (totlen > d->bd_bufsize) + totlen = d->bd_bufsize; + + /* + * Round up the end of the previous packet to the next longword. + */ + curlen = BPF_WORDALIGN(d->bd_slen); + if (curlen + totlen > d->bd_bufsize) { + /* + * This packet will overflow the storage buffer. + * Rotate the buffers if we can, then wakeup any + * pending reads. + */ + if (d->bd_fbuf == 0) { + /* + * We haven't completed the previous read yet, + * so drop the packet. + */ + ++d->bd_dcount; + return; + } + ROTATE_BUFFERS(d); + bpf_wakeup(d); + curlen = 0; + } + else if (d->bd_immediate) + /* + * Immediate mode is set. A packet arrived so any + * reads should be woken up. + */ + bpf_wakeup(d); + + /* + * Append the bpf header. + */ + hp = (struct bpf_hdr *)(d->bd_sbuf + curlen); +#if BSD >= 199103 + microtime(&hp->bh_tstamp); +#elif defined(sun) + uniqtime(&hp->bh_tstamp); +#else + hp->bh_tstamp = time; +#endif + hp->bh_datalen = pktlen; + hp->bh_hdrlen = hdrlen; + /* + * Copy the packet data into the store buffer and update its length. + */ + (*cpfn)(pkt, (u_char *)hp + hdrlen, (hp->bh_caplen = totlen - hdrlen)); + d->bd_slen = curlen + totlen; +} + +/* + * Initialize all nonzero fields of a descriptor. + */ +static int +bpf_allocbufs(d) + register struct bpf_d *d; +{ + d->bd_fbuf = (caddr_t)malloc(d->bd_bufsize, M_DEVBUF, M_WAITOK); + if (d->bd_fbuf == 0) + return (ENOBUFS); + + d->bd_sbuf = (caddr_t)malloc(d->bd_bufsize, M_DEVBUF, M_WAITOK); + if (d->bd_sbuf == 0) { + free(d->bd_fbuf, M_DEVBUF); + return (ENOBUFS); + } + d->bd_slen = 0; + d->bd_hlen = 0; + return (0); +} + +/* + * Free buffers currently in use by a descriptor. + * Called on close. + */ +static void +bpf_freed(d) + register struct bpf_d *d; +{ + /* + * We don't need to lock out interrupts since this descriptor has + * been detached from its interface and it yet hasn't been marked + * free. + */ + if (d->bd_sbuf != 0) { + free(d->bd_sbuf, M_DEVBUF); + if (d->bd_hbuf != 0) + free(d->bd_hbuf, M_DEVBUF); + if (d->bd_fbuf != 0) + free(d->bd_fbuf, M_DEVBUF); + } + if (d->bd_filter) + free((caddr_t)d->bd_filter, M_DEVBUF); + + D_MARKFREE(d); +} + +/* + * Attach an interface to bpf. driverp is a pointer to a (struct bpf_if *) + * in the driver's softc; dlt is the link layer type; hdrlen is the fixed + * size of the link header (variable length headers not yet supported). + */ +void +bpfattach(driverp, ifp, dlt, hdrlen) + caddr_t *driverp; + struct ifnet *ifp; + u_int dlt, hdrlen; +{ + struct bpf_if *bp; + int i; +#if BSD < 199103 + static struct bpf_if bpf_ifs[NBPFILTER]; + static int bpfifno; + + bp = (bpfifno < NBPFILTER) ? &bpf_ifs[bpfifno++] : 0; +#else + bp = (struct bpf_if *)malloc(sizeof(*bp), M_DEVBUF, M_DONTWAIT); +#endif + if (bp == 0) + panic("bpfattach"); + + bp->bif_dlist = 0; + bp->bif_driverp = (struct bpf_if **)driverp; + bp->bif_ifp = ifp; + bp->bif_dlt = dlt; + + bp->bif_next = bpf_iflist; + bpf_iflist = bp; + + *bp->bif_driverp = 0; + + /* + * Compute the length of the bpf header. This is not necessarily + * equal to SIZEOF_BPF_HDR because we want to insert spacing such + * that the network layer header begins on a longword boundary (for + * performance reasons and to alleviate alignment restrictions). + */ + bp->bif_hdrlen = BPF_WORDALIGN(hdrlen + SIZEOF_BPF_HDR) - hdrlen; + + /* + * Mark all the descriptors free if this hasn't been done. + */ + if (!D_ISFREE(&bpf_dtab[0])) + for (i = 0; i < NBPFILTER; ++i) + D_MARKFREE(&bpf_dtab[i]); + +#if 0 + printf("bpf: %s%d attached\n", ifp->if_name, ifp->if_unit); +#endif +} + +#if BSD >= 199103 +/* XXX This routine belongs in net/if.c. */ +/* + * Set/clear promiscuous mode on interface ifp based on the truth value + * of pswitch. The calls are reference counted so that only the first + * "on" request actually has an effect, as does the final "off" request. + * Results are undefined if the "off" and "on" requests are not matched. + */ +int +ifpromisc(ifp, pswitch) + struct ifnet *ifp; + int pswitch; +{ + struct ifreq ifr; + + if (pswitch) { + /* + * If the device is not configured up, we cannot put it in + * promiscuous mode. + */ + if ((ifp->if_flags & IFF_UP) == 0) + return (ENETDOWN); + if (ifp->if_pcount++ != 0) + return (0); + ifp->if_flags |= IFF_PROMISC; + } else { + if (--ifp->if_pcount > 0) + return (0); + ifp->if_flags &= ~IFF_PROMISC; + /* + * If the device is not configured up, we should not need to + * turn off promiscuous mode (device should have turned it + * off when interface went down; and will look at IFF_PROMISC + * again next time interface comes up). + */ + if ((ifp->if_flags & IFF_UP) == 0) + return (0); + } + ifr.ifr_flags = ifp->if_flags; + return ((*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr)); +} +#endif + +#if BSD < 199103 +/* + * Allocate some memory for bpf. This is temporary SunOS support, and + * is admittedly a hack. + * If resources unavaiable, return 0. + */ +static caddr_t +bpf_alloc(size, canwait) + register int size; + register int canwait; +{ + register struct mbuf *m; + + if ((unsigned)size > (MCLBYTES-8)) + return 0; + + MGET(m, canwait, MT_DATA); + if (m == 0) + return 0; + if ((unsigned)size > (MLEN-8)) { + MCLGET(m); + if (m->m_len != MCLBYTES) { + m_freem(m); + return 0; + } + } + *mtod(m, struct mbuf **) = m; + return mtod(m, caddr_t) + 8; +} +#endif diff --git a/sys/net/bpf.h b/sys/net/bpf.h new file mode 100644 index 00000000000..c2797fc3383 --- /dev/null +++ b/sys/net/bpf.h @@ -0,0 +1,259 @@ +/* $NetBSD: bpf.h,v 1.12 1995/09/27 18:30:40 thorpej Exp $ */ + +/* + * Copyright (c) 1990, 1991, 1993 + * The Regents of the University of California. All rights reserved. + * + * This code is derived from the Stanford/CMU enet packet filter, + * (net/enet.c) distributed as part of 4.3BSD, and code contributed + * to Berkeley by Steven McCanne and Van Jacobson both of Lawrence + * Berkeley Laboratory. + * + * 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)bpf.h 8.1 (Berkeley) 6/10/93 + */ + +#ifndef _NET_BPF_H_ +#define _NET_BPF_H_ + +/* + * Alignment macros. BPF_WORDALIGN rounds up to the next + * even multiple of BPF_ALIGNMENT. + */ +#define BPF_ALIGNMENT sizeof(long) +#define BPF_WORDALIGN(x) (((x)+(BPF_ALIGNMENT-1))&~(BPF_ALIGNMENT-1)) + +#define BPF_MAXINSNS 512 +#define BPF_MAXBUFSIZE 0x8000 +#define BPF_MINBUFSIZE 32 + +/* + * Structure for BIOCSETF. + */ +struct bpf_program { + u_int bf_len; + struct bpf_insn *bf_insns; +}; + +/* + * Struct returned by BIOCGSTATS. + */ +struct bpf_stat { + u_int bs_recv; /* number of packets received */ + u_int bs_drop; /* number of packets dropped */ +}; + +/* + * Struct return by BIOCVERSION. This represents the version number of + * the filter language described by the instruction encodings below. + * bpf understands a program iff kernel_major == filter_major && + * kernel_minor >= filter_minor, that is, if the value returned by the + * running kernel has the same major number and a minor number equal + * equal to or less than the filter being downloaded. Otherwise, the + * results are undefined, meaning an error may be returned or packets + * may be accepted haphazardly. + * It has nothing to do with the source code version. + */ +struct bpf_version { + u_short bv_major; + u_short bv_minor; +}; +/* Current version number of filter architecture. */ +#define BPF_MAJOR_VERSION 1 +#define BPF_MINOR_VERSION 1 + +/* + * BPF ioctls + * + * The first set is for compatibility with Sun's pcc style + * header files. If your using gcc, we assume that you + * have run fixincludes so the latter set should work. + */ +#if (defined(sun) || defined(ibm032)) && !defined(__GNUC__) +#define BIOCGBLEN _IOR(B,102, u_int) +#define BIOCSBLEN _IOWR(B,102, u_int) +#define BIOCSETF _IOW(B,103, struct bpf_program) +#define BIOCFLUSH _IO(B,104) +#define BIOCPROMISC _IO(B,105) +#define BIOCGDLT _IOR(B,106, u_int) +#define BIOCGETIF _IOR(B,107, struct ifreq) +#define BIOCSETIF _IOW(B,108, struct ifreq) +#define BIOCSRTIMEOUT _IOW(B,109, struct timeval) +#define BIOCGRTIMEOUT _IOR(B,110, struct timeval) +#define BIOCGSTATS _IOR(B,111, struct bpf_stat) +#define BIOCIMMEDIATE _IOW(B,112, u_int) +#define BIOCVERSION _IOR(B,113, struct bpf_version) +#define BIOCSRSIG _IOW(B,114, u_int) +#define BIOCGRSIG _IOR(B,115, u_int) +#else +#define BIOCGBLEN _IOR('B',102, u_int) +#define BIOCSBLEN _IOWR('B',102, u_int) +#define BIOCSETF _IOW('B',103, struct bpf_program) +#define BIOCFLUSH _IO('B',104) +#define BIOCPROMISC _IO('B',105) +#define BIOCGDLT _IOR('B',106, u_int) +#define BIOCGETIF _IOR('B',107, struct ifreq) +#define BIOCSETIF _IOW('B',108, struct ifreq) +#define BIOCSRTIMEOUT _IOW('B',109, struct timeval) +#define BIOCGRTIMEOUT _IOR('B',110, struct timeval) +#define BIOCGSTATS _IOR('B',111, struct bpf_stat) +#define BIOCIMMEDIATE _IOW('B',112, u_int) +#define BIOCVERSION _IOR('B',113, struct bpf_version) +#define BIOCSRSIG _IOW('B',114, u_int) +#define BIOCGRSIG _IOR('B',115, u_int) +#endif + +/* + * Structure prepended to each packet. + */ +struct bpf_hdr { + struct timeval bh_tstamp; /* time stamp */ + u_int32_t bh_caplen; /* length of captured portion */ + u_int32_t bh_datalen; /* original length of packet */ + u_int16_t bh_hdrlen; /* length of bpf header (this struct + plus alignment padding) */ +}; +/* + * Because the structure above is not a multiple of 4 bytes, some compilers + * will insist on inserting padding; hence, sizeof(struct bpf_hdr) won't work. + * Only the kernel needs to know about it; applications use bh_hdrlen. + */ +#ifdef _KERNEL +#define SIZEOF_BPF_HDR 18 +#endif + +/* + * Data-link level type codes. + */ +#define DLT_NULL 0 /* no link-layer encapsulation */ +#define DLT_EN10MB 1 /* Ethernet (10Mb) */ +#define DLT_EN3MB 2 /* Experimental Ethernet (3Mb) */ +#define DLT_AX25 3 /* Amateur Radio AX.25 */ +#define DLT_PRONET 4 /* Proteon ProNET Token Ring */ +#define DLT_CHAOS 5 /* Chaos */ +#define DLT_IEEE802 6 /* IEEE 802 Networks */ +#define DLT_ARCNET 7 /* ARCNET */ +#define DLT_SLIP 8 /* Serial Line IP */ +#define DLT_PPP 9 /* Point-to-point Protocol */ +#define DLT_FDDI 10 /* FDDI */ + +/* + * The instruction encondings. + */ +/* instruction classes */ +#define BPF_CLASS(code) ((code) & 0x07) +#define BPF_LD 0x00 +#define BPF_LDX 0x01 +#define BPF_ST 0x02 +#define BPF_STX 0x03 +#define BPF_ALU 0x04 +#define BPF_JMP 0x05 +#define BPF_RET 0x06 +#define BPF_MISC 0x07 + +/* ld/ldx fields */ +#define BPF_SIZE(code) ((code) & 0x18) +#define BPF_W 0x00 +#define BPF_H 0x08 +#define BPF_B 0x10 +#define BPF_MODE(code) ((code) & 0xe0) +#define BPF_IMM 0x00 +#define BPF_ABS 0x20 +#define BPF_IND 0x40 +#define BPF_MEM 0x60 +#define BPF_LEN 0x80 +#define BPF_MSH 0xa0 + +/* alu/jmp fields */ +#define BPF_OP(code) ((code) & 0xf0) +#define BPF_ADD 0x00 +#define BPF_SUB 0x10 +#define BPF_MUL 0x20 +#define BPF_DIV 0x30 +#define BPF_OR 0x40 +#define BPF_AND 0x50 +#define BPF_LSH 0x60 +#define BPF_RSH 0x70 +#define BPF_NEG 0x80 +#define BPF_JA 0x00 +#define BPF_JEQ 0x10 +#define BPF_JGT 0x20 +#define BPF_JGE 0x30 +#define BPF_JSET 0x40 +#define BPF_SRC(code) ((code) & 0x08) +#define BPF_K 0x00 +#define BPF_X 0x08 + +/* ret - BPF_K and BPF_X also apply */ +#define BPF_RVAL(code) ((code) & 0x18) +#define BPF_A 0x10 + +/* misc */ +#define BPF_MISCOP(code) ((code) & 0xf8) +#define BPF_TAX 0x00 +#define BPF_TXA 0x80 + +/* + * The instruction data structure. + */ +struct bpf_insn { + u_int16_t code; + u_char jt; + u_char jf; + int32_t k; +}; + +/* + * Macros for insn array initializers. + */ +#define BPF_STMT(code, k) { (u_int16_t)(code), 0, 0, k } +#define BPF_JUMP(code, k, jt, jf) { (u_int16_t)(code), jt, jf, k } + +#ifdef _KERNEL +int bpf_validate __P((struct bpf_insn *, int)); +int bpfopen __P((dev_t, int)); +int bpfclose __P((dev_t, int)); +int bpfread __P((dev_t, struct uio *)); +int bpfwrite __P((dev_t, struct uio *)); +int bpfioctl __P((dev_t, u_long, caddr_t, int)); +int bpf_select __P((dev_t, int, struct proc *)); +void bpf_tap __P((caddr_t, u_char *, u_int)); +void bpf_mtap __P((caddr_t, struct mbuf *)); +void bpfattach __P((caddr_t *, struct ifnet *, u_int, u_int)); +void bpfilterattach __P((int)); +u_int bpf_filter __P((struct bpf_insn *, u_char *, u_int, u_int)); +#endif + +/* + * Number of scratch memory words (for BPF_LD|BPF_MEM and BPF_ST). + */ +#define BPF_MEMWORDS 16 + +#endif /* _NET_BPF_H_ */ diff --git a/sys/net/bpf_filter.c b/sys/net/bpf_filter.c new file mode 100644 index 00000000000..d0a07dd78ec --- /dev/null +++ b/sys/net/bpf_filter.c @@ -0,0 +1,525 @@ +/* $NetBSD: bpf_filter.c,v 1.11 1995/04/22 13:26:39 cgd Exp $ */ + +/* + * Copyright (c) 1990, 1991, 1992, 1993 + * The Regents of the University of California. All rights reserved. + * + * This code is derived from the Stanford/CMU enet packet filter, + * (net/enet.c) distributed as part of 4.3BSD, and code contributed + * to Berkeley by Steven McCanne and Van Jacobson both of Lawrence + * Berkeley Laboratory. + * + * 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)bpf_filter.c 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/param.h> +#include <sys/types.h> +#include <sys/time.h> + +#ifdef sun +#include <netinet/in.h> +#endif + +#if defined(sparc) || defined(mips) || defined(ibm032) || \ + (defined(__NetBSD__) && !defined(UNALIGNED_ACCESS)) +#define BPF_ALIGN +#endif + +#ifndef BPF_ALIGN +#define EXTRACT_SHORT(p) ((u_int16_t)ntohs(*(u_int16_t *)p)) +#define EXTRACT_LONG(p) (ntohl(*(u_int32_t *)p)) +#else +#define EXTRACT_SHORT(p)\ + ((u_int16_t)\ + ((u_int16_t)*((u_char *)p+0)<<8|\ + (u_int16_t)*((u_char *)p+1)<<0)) +#define EXTRACT_LONG(p)\ + ((u_int32_t)*((u_char *)p+0)<<24|\ + (u_int32_t)*((u_char *)p+1)<<16|\ + (u_int32_t)*((u_char *)p+2)<<8|\ + (u_int32_t)*((u_char *)p+3)<<0) +#endif + +#ifdef _KERNEL +#include <sys/mbuf.h> +#define MINDEX(len, m, k) \ +{ \ + len = m->m_len; \ + while (k >= len) { \ + k -= len; \ + m = m->m_next; \ + if (m == 0) \ + return 0; \ + len = m->m_len; \ + } \ +} + +static int +m_xword(m, k, err) + register struct mbuf *m; + register int k, *err; +{ + register int len; + register u_char *cp, *np; + register struct mbuf *m0; + + MINDEX(len, m, k); + cp = mtod(m, u_char *) + k; + if (len - k >= 4) { + *err = 0; + return EXTRACT_LONG(cp); + } + m0 = m->m_next; + if (m0 == 0 || m0->m_len + len - k < 4) + goto bad; + *err = 0; + np = mtod(m0, u_char *); + switch (len - k) { + + case 1: + return (cp[0] << 24) | (np[0] << 16) | (np[1] << 8) | np[2]; + + case 2: + return (cp[0] << 24) | (cp[1] << 16) | (np[0] << 8) | np[1]; + + default: + return (cp[0] << 24) | (cp[1] << 16) | (cp[2] << 8) | np[0]; + } + bad: + *err = 1; + return 0; +} + +static int +m_xhalf(m, k, err) + register struct mbuf *m; + register int k, *err; +{ + register int len; + register u_char *cp; + register struct mbuf *m0; + + MINDEX(len, m, k); + cp = mtod(m, u_char *) + k; + if (len - k >= 2) { + *err = 0; + return EXTRACT_SHORT(cp); + } + m0 = m->m_next; + if (m0 == 0) + goto bad; + *err = 0; + return (cp[0] << 8) | mtod(m0, u_char *)[0]; + bad: + *err = 1; + return 0; +} +#endif + +#include <net/bpf.h> + +/* + * Execute the filter program starting at pc on the packet p + * wirelen is the length of the original packet + * buflen is the amount of data present + */ +u_int +bpf_filter(pc, p, wirelen, buflen) + register struct bpf_insn *pc; + register u_char *p; + u_int wirelen; + register u_int buflen; +{ + register u_int32_t A, X; + register int k; + int32_t mem[BPF_MEMWORDS]; + + if (pc == 0) + /* + * No filter means accept all. + */ + return (u_int)-1; +#ifdef lint + A = 0; + X = 0; +#endif + --pc; + while (1) { + ++pc; + switch (pc->code) { + + default: +#ifdef _KERNEL + return 0; +#else + abort(); +#endif + case BPF_RET|BPF_K: + return (u_int)pc->k; + + case BPF_RET|BPF_A: + return (u_int)A; + + case BPF_LD|BPF_W|BPF_ABS: + k = pc->k; + if (k + sizeof(int32_t) > buflen) { +#ifdef _KERNEL + int merr; + + if (buflen != 0) + return 0; + A = m_xword((struct mbuf *)p, k, &merr); + if (merr != 0) + return 0; + continue; +#else + return 0; +#endif + } + A = EXTRACT_LONG(&p[k]); + continue; + + case BPF_LD|BPF_H|BPF_ABS: + k = pc->k; + if (k + sizeof(int16_t) > buflen) { +#ifdef _KERNEL + int merr; + + if (buflen != 0) + return 0; + A = m_xhalf((struct mbuf *)p, k, &merr); + continue; +#else + return 0; +#endif + } + A = EXTRACT_SHORT(&p[k]); + continue; + + case BPF_LD|BPF_B|BPF_ABS: + k = pc->k; + if (k >= buflen) { +#ifdef _KERNEL + register struct mbuf *m; + register int len; + + if (buflen != 0) + return 0; + m = (struct mbuf *)p; + MINDEX(len, m, k); + A = mtod(m, u_char *)[k]; + continue; +#else + return 0; +#endif + } + A = p[k]; + continue; + + case BPF_LD|BPF_W|BPF_LEN: + A = wirelen; + continue; + + case BPF_LDX|BPF_W|BPF_LEN: + X = wirelen; + continue; + + case BPF_LD|BPF_W|BPF_IND: + k = X + pc->k; + if (k + sizeof(int32_t) > buflen) { +#ifdef _KERNEL + int merr; + + if (buflen != 0) + return 0; + A = m_xword((struct mbuf *)p, k, &merr); + if (merr != 0) + return 0; + continue; +#else + return 0; +#endif + } + A = EXTRACT_LONG(&p[k]); + continue; + + case BPF_LD|BPF_H|BPF_IND: + k = X + pc->k; + if (k + sizeof(int16_t) > buflen) { +#ifdef _KERNEL + int merr; + + if (buflen != 0) + return 0; + A = m_xhalf((struct mbuf *)p, k, &merr); + if (merr != 0) + return 0; + continue; +#else + return 0; +#endif + } + A = EXTRACT_SHORT(&p[k]); + continue; + + case BPF_LD|BPF_B|BPF_IND: + k = X + pc->k; + if (k >= buflen) { +#ifdef _KERNEL + register struct mbuf *m; + register int len; + + if (buflen != 0) + return 0; + m = (struct mbuf *)p; + MINDEX(len, m, k); + A = mtod(m, char *)[k]; + continue; +#else + return 0; +#endif + } + A = p[k]; + continue; + + case BPF_LDX|BPF_MSH|BPF_B: + k = pc->k; + if (k >= buflen) { +#ifdef _KERNEL + register struct mbuf *m; + register int len; + + if (buflen != 0) + return 0; + m = (struct mbuf *)p; + MINDEX(len, m, k); + X = (mtod(m, char *)[k] & 0xf) << 2; + continue; +#else + return 0; +#endif + } + X = (p[pc->k] & 0xf) << 2; + continue; + + case BPF_LD|BPF_IMM: + A = pc->k; + continue; + + case BPF_LDX|BPF_IMM: + X = pc->k; + continue; + + case BPF_LD|BPF_MEM: + A = mem[pc->k]; + continue; + + case BPF_LDX|BPF_MEM: + X = mem[pc->k]; + continue; + + case BPF_ST: + mem[pc->k] = A; + continue; + + case BPF_STX: + mem[pc->k] = X; + continue; + + case BPF_JMP|BPF_JA: + pc += pc->k; + continue; + + case BPF_JMP|BPF_JGT|BPF_K: + pc += (A > pc->k) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JGE|BPF_K: + pc += (A >= pc->k) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JEQ|BPF_K: + pc += (A == pc->k) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JSET|BPF_K: + pc += (A & pc->k) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JGT|BPF_X: + pc += (A > X) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JGE|BPF_X: + pc += (A >= X) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JEQ|BPF_X: + pc += (A == X) ? pc->jt : pc->jf; + continue; + + case BPF_JMP|BPF_JSET|BPF_X: + pc += (A & X) ? pc->jt : pc->jf; + continue; + + case BPF_ALU|BPF_ADD|BPF_X: + A += X; + continue; + + case BPF_ALU|BPF_SUB|BPF_X: + A -= X; + continue; + + case BPF_ALU|BPF_MUL|BPF_X: + A *= X; + continue; + + case BPF_ALU|BPF_DIV|BPF_X: + if (X == 0) + return 0; + A /= X; + continue; + + case BPF_ALU|BPF_AND|BPF_X: + A &= X; + continue; + + case BPF_ALU|BPF_OR|BPF_X: + A |= X; + continue; + + case BPF_ALU|BPF_LSH|BPF_X: + A <<= X; + continue; + + case BPF_ALU|BPF_RSH|BPF_X: + A >>= X; + continue; + + case BPF_ALU|BPF_ADD|BPF_K: + A += pc->k; + continue; + + case BPF_ALU|BPF_SUB|BPF_K: + A -= pc->k; + continue; + + case BPF_ALU|BPF_MUL|BPF_K: + A *= pc->k; + continue; + + case BPF_ALU|BPF_DIV|BPF_K: + A /= pc->k; + continue; + + case BPF_ALU|BPF_AND|BPF_K: + A &= pc->k; + continue; + + case BPF_ALU|BPF_OR|BPF_K: + A |= pc->k; + continue; + + case BPF_ALU|BPF_LSH|BPF_K: + A <<= pc->k; + continue; + + case BPF_ALU|BPF_RSH|BPF_K: + A >>= pc->k; + continue; + + case BPF_ALU|BPF_NEG: + A = -A; + continue; + + case BPF_MISC|BPF_TAX: + X = A; + continue; + + case BPF_MISC|BPF_TXA: + A = X; + continue; + } + } +} + +#ifdef _KERNEL +/* + * Return true if the 'fcode' is a valid filter program. + * The constraints are that each jump be forward and to a valid + * code. The code must terminate with either an accept or reject. + * 'valid' is an array for use by the routine (it must be at least + * 'len' bytes long). + * + * The kernel needs to be able to verify an application's filter code. + * Otherwise, a bogus program could easily crash the system. + */ +int +bpf_validate(f, len) + struct bpf_insn *f; + int len; +{ + register int i; + register struct bpf_insn *p; + + for (i = 0; i < len; ++i) { + /* + * Check that that jumps are forward, and within + * the code block. + */ + p = &f[i]; + if (BPF_CLASS(p->code) == BPF_JMP) { + register int from = i + 1; + + if (BPF_OP(p->code) == BPF_JA) { + if (from + p->k >= len) + return 0; + } + else if (from + p->jt >= len || from + p->jf >= len) + return 0; + } + /* + * Check that memory operations use valid addresses. + */ + if ((BPF_CLASS(p->code) == BPF_ST || + (BPF_CLASS(p->code) == BPF_LD && + (p->code & 0xe0) == BPF_MEM)) && + (p->k >= BPF_MEMWORDS || p->k < 0)) + return 0; + /* + * Check for constant division by 0. + */ + if (p->code == (BPF_ALU|BPF_DIV|BPF_K) && p->k == 0) + return 0; + } + return BPF_CLASS(f[len - 1].code) == BPF_RET; +} +#endif diff --git a/sys/net/bpfdesc.h b/sys/net/bpfdesc.h new file mode 100644 index 00000000000..a12d7ec9e06 --- /dev/null +++ b/sys/net/bpfdesc.h @@ -0,0 +1,103 @@ +/* $NetBSD: bpfdesc.h,v 1.11 1995/09/27 18:30:42 thorpej Exp $ */ + +/* + * Copyright (c) 1990, 1991, 1993 + * The Regents of the University of California. All rights reserved. + * + * This code is derived from the Stanford/CMU enet packet filter, + * (net/enet.c) distributed as part of 4.3BSD, and code contributed + * to Berkeley by Steven McCanne and Van Jacobson both of Lawrence + * Berkeley Laboratory. + * + * 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)bpfdesc.h 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/select.h> + +/* + * Descriptor associated with each open bpf file. + */ +struct bpf_d { + struct bpf_d *bd_next; /* Linked list of descriptors */ + /* + * Buffer slots: two mbuf clusters buffer the incoming packets. + * The model has three slots. Sbuf is always occupied. + * sbuf (store) - Receive interrupt puts packets here. + * hbuf (hold) - When sbuf is full, put cluster here and + * wakeup read (replace sbuf with fbuf). + * fbuf (free) - When read is done, put cluster here. + * On receiving, if sbuf is full and fbuf is 0, packet is dropped. + */ + caddr_t bd_sbuf; /* store slot */ + caddr_t bd_hbuf; /* hold slot */ + caddr_t bd_fbuf; /* free slot */ + int bd_slen; /* current length of store buffer */ + int bd_hlen; /* current length of hold buffer */ + + int bd_bufsize; /* absolute length of buffers */ + + struct bpf_if * bd_bif; /* interface descriptor */ + u_long bd_rtout; /* Read timeout in 'ticks' */ + struct bpf_insn *bd_filter; /* filter code */ + u_long bd_rcount; /* number of packets received */ + u_long bd_dcount; /* number of packets dropped */ + + u_char bd_promisc; /* true if listening promiscuously */ + u_char bd_state; /* idle, waiting, or timed out */ + u_char bd_immediate; /* true to return on packet arrival */ + int bd_async; /* non-zero if packet reception should generate signal */ + int bd_sig; /* signal to send upon packet reception */ + pid_t bd_pgid; /* process or group id for signal */ +#if BSD < 199103 + u_char bd_selcoll; /* true if selects collide */ + int bd_timedout; + struct proc * bd_selproc; /* process that last selected us */ +#else + u_char bd_pad; /* explicit alignment */ + struct selinfo bd_sel; /* bsd select info */ +#endif +}; + +/* + * Descriptor associated with each attached hardware interface. + */ +struct bpf_if { + struct bpf_if *bif_next; /* list of all interfaces */ + struct bpf_d *bif_dlist; /* descriptor list */ + struct bpf_if **bif_driverp; /* pointer into softc */ + u_int bif_dlt; /* link layer type */ + u_int bif_hdrlen; /* length of header (with padding) */ + struct ifnet *bif_ifp; /* correspoding interface */ +}; + +#ifdef _KERNEL +int bpf_setf __P((struct bpf_d *, struct bpf_program *)); +#endif diff --git a/sys/net/bsd-comp.c b/sys/net/bsd-comp.c new file mode 100644 index 00000000000..efe0f02df17 --- /dev/null +++ b/sys/net/bsd-comp.c @@ -0,0 +1,1116 @@ +/* $NetBSD: bsd-comp.c,v 1.1 1995/07/04 06:28:17 paulus Exp $ */ + +/* Because this code is derived from the 4.3BSD compress source: + * + * + * Copyright (c) 1985, 1986 The Regents of the University of California. + * All rights reserved. + * + * This code is derived from software contributed to Berkeley by + * James A. Woods, derived from original work by Spencer Thomas + * and Joseph Orost. + * + * 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + */ + +/* + * This version is for use with mbufs on BSD-derived systems. + * + * $Id: bsd-comp.c,v 1.1 1995/10/18 08:53:05 deraadt Exp $ + */ + +#include <sys/param.h> +#include <sys/types.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <net/if.h> +#include <net/if_types.h> +#include <net/ppp_defs.h> +#include <net/if_ppp.h> + +#define PACKETPTR struct mbuf * +#include <net/ppp-comp.h> + +#if DO_BSD_COMPRESS +/* + * PPP "BSD compress" compression + * The differences between this compression and the classic BSD LZW + * source are obvious from the requirement that the classic code worked + * with files while this handles arbitrarily long streams that + * are broken into packets. They are: + * + * When the code size expands, a block of junk is not emitted by + * the compressor and not expected by the decompressor. + * + * New codes are not necessarily assigned every time an old + * code is output by the compressor. This is because a packet + * end forces a code to be emitted, but does not imply that a + * new sequence has been seen. + * + * The compression ratio is checked at the first end of a packet + * after the appropriate gap. Besides simplifying and speeding + * things up, this makes it more likely that the transmitter + * and receiver will agree when the dictionary is cleared when + * compression is not going well. + */ + +/* + * A dictionary for doing BSD compress. + */ +struct bsd_db { + int totlen; /* length of this structure */ + u_int hsize; /* size of the hash table */ + u_char hshift; /* used in hash function */ + u_char n_bits; /* current bits/code */ + u_char maxbits; + u_char debug; + u_char unit; + u_int16_t seqno; /* sequence # of next packet */ + u_int hdrlen; /* header length to preallocate */ + u_int mru; + u_int maxmaxcode; /* largest valid code */ + u_int max_ent; /* largest code in use */ + u_int in_count; /* uncompressed bytes, aged */ + u_int bytes_out; /* compressed bytes, aged */ + u_int ratio; /* recent compression ratio */ + u_int checkpoint; /* when to next check the ratio */ + u_int clear_count; /* times dictionary cleared */ + u_int incomp_count; /* incompressible packets */ + u_int incomp_bytes; /* incompressible bytes */ + u_int uncomp_count; /* uncompressed packets */ + u_int uncomp_bytes; /* uncompressed bytes */ + u_int comp_count; /* compressed packets */ + u_int comp_bytes; /* compressed bytes */ + u_int16_t *lens; /* array of lengths of codes */ + struct bsd_dict { + union { /* hash value */ + u_int32_t fcode; + struct { +#if BYTE_ORDER == LITTLE_ENDIAN + u_int16_t prefix; /* preceding code */ + u_char suffix; /* last character of new code */ + u_char pad; +#else + u_char pad; + u_char suffix; /* last character of new code */ + u_int16_t prefix; /* preceding code */ +#endif + } hs; + } f; + u_int16_t codem1; /* output of hash table -1 */ + u_int16_t cptr; /* map code to hash table entry */ + } dict[1]; +}; + +#define BSD_OVHD 2 /* BSD compress overhead/packet */ +#define BSD_INIT_BITS BSD_MIN_BITS + +static void *bsd_comp_alloc __P((u_char *options, int opt_len)); +static void *bsd_decomp_alloc __P((u_char *options, int opt_len)); +static void bsd_free __P((void *state)); +static int bsd_comp_init __P((void *state, u_char *options, int opt_len, + int unit, int hdrlen, int debug)); +static int bsd_decomp_init __P((void *state, u_char *options, int opt_len, + int unit, int hdrlen, int mru, int debug)); +static int bsd_compress __P((void *state, struct mbuf **mret, + struct mbuf *mp, int slen, int maxolen)); +static void bsd_incomp __P((void *state, struct mbuf *dmsg)); +static int bsd_decompress __P((void *state, struct mbuf *cmp, + struct mbuf **dmpp)); +static void bsd_reset __P((void *state)); +static void bsd_comp_stats __P((void *state, struct compstat *stats)); + +/* + * Procedures exported to if_ppp.c. + */ +struct compressor ppp_bsd_compress = { + CI_BSD_COMPRESS, /* compress_proto */ + bsd_comp_alloc, /* comp_alloc */ + bsd_free, /* comp_free */ + bsd_comp_init, /* comp_init */ + bsd_reset, /* comp_reset */ + bsd_compress, /* compress */ + bsd_comp_stats, /* comp_stat */ + bsd_decomp_alloc, /* decomp_alloc */ + bsd_free, /* decomp_free */ + bsd_decomp_init, /* decomp_init */ + bsd_reset, /* decomp_reset */ + bsd_decompress, /* decompress */ + bsd_incomp, /* incomp */ + bsd_comp_stats, /* decomp_stat */ +}; + +/* + * the next two codes should not be changed lightly, as they must not + * lie within the contiguous general code space. + */ +#define CLEAR 256 /* table clear output code */ +#define FIRST 257 /* first free entry */ +#define LAST 255 + +#define MAXCODE(b) ((1 << (b)) - 1) +#define BADCODEM1 MAXCODE(BSD_MAX_BITS) + +#define BSD_HASH(prefix,suffix,hshift) ((((u_int32_t)(suffix)) << (hshift)) \ + ^ (u_int32_t)(prefix)) +#define BSD_KEY(prefix,suffix) ((((u_int32_t)(suffix)) << 16) \ + + (u_int32_t)(prefix)) + +#define CHECK_GAP 10000 /* Ratio check interval */ + +#define RATIO_SCALE_LOG 8 +#define RATIO_SCALE (1<<RATIO_SCALE_LOG) +#define RATIO_MAX (0x7fffffff>>RATIO_SCALE_LOG) + +/* + * clear the dictionary + */ +static void +bsd_clear(db) + struct bsd_db *db; +{ + db->clear_count++; + db->max_ent = FIRST-1; + db->n_bits = BSD_INIT_BITS; + db->ratio = 0; + db->bytes_out = 0; + db->in_count = 0; + db->incomp_count = 0; + db->checkpoint = CHECK_GAP; +} + +/* + * If the dictionary is full, then see if it is time to reset it. + * + * Compute the compression ratio using fixed-point arithmetic + * with 8 fractional bits. + * + * Since we have an infinite stream instead of a single file, + * watch only the local compression ratio. + * + * Since both peers must reset the dictionary at the same time even in + * the absence of CLEAR codes (while packets are incompressible), they + * must compute the same ratio. + */ +static int /* 1=output CLEAR */ +bsd_check(db) + struct bsd_db *db; +{ + u_int new_ratio; + + if (db->in_count >= db->checkpoint) { + /* age the ratio by limiting the size of the counts */ + if (db->in_count >= RATIO_MAX + || db->bytes_out >= RATIO_MAX) { + db->in_count -= db->in_count/4; + db->bytes_out -= db->bytes_out/4; + } + + db->checkpoint = db->in_count + CHECK_GAP; + + if (db->max_ent >= db->maxmaxcode) { + /* Reset the dictionary only if the ratio is worse, + * or if it looks as if it has been poisoned + * by incompressible data. + * + * This does not overflow, because + * db->in_count <= RATIO_MAX. + */ + new_ratio = db->in_count << RATIO_SCALE_LOG; + if (db->bytes_out != 0) + new_ratio /= db->bytes_out; + + if (new_ratio < db->ratio || new_ratio < 1 * RATIO_SCALE) { + bsd_clear(db); + return 1; + } + db->ratio = new_ratio; + } + } + return 0; +} + +/* + * Return statistics. + */ +static void +bsd_comp_stats(state, stats) + void *state; + struct compstat *stats; +{ + struct bsd_db *db = (struct bsd_db *) state; + u_int out; + + stats->unc_bytes = db->uncomp_bytes; + stats->unc_packets = db->uncomp_count; + stats->comp_bytes = db->comp_bytes; + stats->comp_packets = db->comp_count; + stats->inc_bytes = db->incomp_bytes; + stats->inc_packets = db->incomp_count; + stats->ratio = db->in_count; + out = db->bytes_out; + if (stats->ratio <= 0x7fffff) + stats->ratio <<= 8; + else + out >>= 8; + if (out != 0) + stats->ratio /= out; +} + +/* + * Reset state, as on a CCP ResetReq. + */ +static void +bsd_reset(state) + void *state; +{ + struct bsd_db *db = (struct bsd_db *) state; + + db->seqno = 0; + bsd_clear(db); + db->clear_count = 0; +} + +/* + * Allocate space for a (de) compressor. + */ +static void * +bsd_alloc(options, opt_len, decomp) + u_char *options; + int opt_len, decomp; +{ + int bits; + u_int newlen, hsize, hshift, maxmaxcode; + struct bsd_db *db; + + if (opt_len != CILEN_BSD_COMPRESS || options[0] != CI_BSD_COMPRESS + || options[1] != CILEN_BSD_COMPRESS + || BSD_VERSION(options[2]) != BSD_CURRENT_VERSION) + return NULL; + bits = BSD_NBITS(options[2]); + switch (bits) { + case 9: /* needs 82152 for both directions */ + case 10: /* needs 84144 */ + case 11: /* needs 88240 */ + case 12: /* needs 96432 */ + hsize = 5003; + hshift = 4; + break; + case 13: /* needs 176784 */ + hsize = 9001; + hshift = 5; + break; + case 14: /* needs 353744 */ + hsize = 18013; + hshift = 6; + break; + case 15: /* needs 691440 */ + hsize = 35023; + hshift = 7; + break; + case 16: /* needs 1366160--far too much, */ + /* hsize = 69001; */ /* and 69001 is too big for cptr */ + /* hshift = 8; */ /* in struct bsd_db */ + /* break; */ + default: + return NULL; + } + + maxmaxcode = MAXCODE(bits); + newlen = sizeof(*db) + (hsize-1) * (sizeof(db->dict[0])); + MALLOC(db, struct bsd_db *, newlen, M_DEVBUF, M_NOWAIT); + if (!db) + return NULL; + bzero(db, sizeof(*db) - sizeof(db->dict)); + + if (!decomp) { + db->lens = NULL; + } else { + MALLOC(db->lens, u_int16_t *, (maxmaxcode+1) * sizeof(db->lens[0]), + M_DEVBUF, M_NOWAIT); + if (!db->lens) { + FREE(db, M_DEVBUF); + return NULL; + } + } + + db->totlen = newlen; + db->hsize = hsize; + db->hshift = hshift; + db->maxmaxcode = maxmaxcode; + db->maxbits = bits; + + return (void *) db; +} + +static void +bsd_free(state) + void *state; +{ + struct bsd_db *db = (struct bsd_db *) state; + + if (db->lens) + FREE(db->lens, M_DEVBUF); + FREE(db, M_DEVBUF); +} + +static void * +bsd_comp_alloc(options, opt_len) + u_char *options; + int opt_len; +{ + return bsd_alloc(options, opt_len, 0); +} + +static void * +bsd_decomp_alloc(options, opt_len) + u_char *options; + int opt_len; +{ + return bsd_alloc(options, opt_len, 1); +} + +/* + * Initialize the database. + */ +static int +bsd_init(db, options, opt_len, unit, hdrlen, mru, debug, decomp) + struct bsd_db *db; + u_char *options; + int opt_len, unit, hdrlen, mru, debug, decomp; +{ + int i; + + if (opt_len != CILEN_BSD_COMPRESS || options[0] != CI_BSD_COMPRESS + || options[1] != CILEN_BSD_COMPRESS + || BSD_VERSION(options[2]) != BSD_CURRENT_VERSION + || BSD_NBITS(options[2]) != db->maxbits + || decomp && db->lens == NULL) + return 0; + + if (decomp) { + i = LAST+1; + while (i != 0) + db->lens[--i] = 1; + } + i = db->hsize; + while (i != 0) { + db->dict[--i].codem1 = BADCODEM1; + db->dict[i].cptr = 0; + } + + db->unit = unit; + db->hdrlen = hdrlen; + db->mru = mru; +#ifndef DEBUG + if (debug) +#endif + db->debug = 1; + + bsd_reset(db); + + return 1; +} + +static int +bsd_comp_init(state, options, opt_len, unit, hdrlen, debug) + void *state; + u_char *options; + int opt_len, unit, hdrlen, debug; +{ + return bsd_init((struct bsd_db *) state, options, opt_len, + unit, hdrlen, 0, debug, 0); +} + +static int +bsd_decomp_init(state, options, opt_len, unit, hdrlen, mru, debug) + void *state; + u_char *options; + int opt_len, unit, hdrlen, mru, debug; +{ + return bsd_init((struct bsd_db *) state, options, opt_len, + unit, hdrlen, mru, debug, 1); +} + + +/* + * compress a packet + * One change from the BSD compress command is that when the + * code size expands, we do not output a bunch of padding. + */ +int /* new slen */ +bsd_compress(state, mret, mp, slen, maxolen) + void *state; + struct mbuf **mret; /* return compressed mbuf chain here */ + struct mbuf *mp; /* from here */ + int slen; /* uncompressed length */ + int maxolen; /* max compressed length */ +{ + struct bsd_db *db = (struct bsd_db *) state; + int hshift = db->hshift; + u_int max_ent = db->max_ent; + u_int n_bits = db->n_bits; + u_int bitno = 32; + u_int32_t accm = 0, fcode; + struct bsd_dict *dictp; + u_char c; + int hval, disp, ent, ilen; + struct mbuf *np; + u_char *rptr, *wptr; + u_char *cp_end; + int olen; + struct mbuf *m, **mnp; + +#define PUTBYTE(v) { \ + ++olen; \ + if (wptr) { \ + *wptr++ = (v); \ + if (wptr >= cp_end) { \ + m->m_len = wptr - mtod(m, u_char *); \ + MGET(m->m_next, M_DONTWAIT, MT_DATA); \ + m = m->m_next; \ + if (m) { \ + m->m_len = 0; \ + if (maxolen - olen > MLEN) \ + MCLGET(m, M_DONTWAIT); \ + wptr = mtod(m, u_char *); \ + cp_end = wptr + M_TRAILINGSPACE(m); \ + } else \ + wptr = NULL; \ + } \ + } \ +} + +#define OUTPUT(ent) { \ + bitno -= n_bits; \ + accm |= ((ent) << bitno); \ + do { \ + PUTBYTE(accm >> 24); \ + accm <<= 8; \ + bitno += 8; \ + } while (bitno <= 24); \ +} + + /* + * If the protocol is not in the range we're interested in, + * just return without compressing the packet. If it is, + * the protocol becomes the first byte to compress. + */ + rptr = mtod(mp, u_char *); + ent = PPP_PROTOCOL(rptr); + if (ent < 0x21 || ent > 0xf9) { + *mret = NULL; + return slen; + } + + /* Don't generate compressed packets which are larger than + the uncompressed packet. */ + if (maxolen > slen) + maxolen = slen; + + /* Allocate one mbuf to start with. */ + MGET(m, M_DONTWAIT, MT_DATA); + *mret = m; + if (m != NULL) { + m->m_len = 0; + if (maxolen + db->hdrlen > MLEN) + MCLGET(m, M_DONTWAIT); + m->m_data += db->hdrlen; + wptr = mtod(m, u_char *); + cp_end = wptr + M_TRAILINGSPACE(m); + } else + wptr = cp_end = NULL; + + /* + * Copy the PPP header over, changing the protocol, + * and install the 2-byte packet sequence number. + */ + if (wptr) { + *wptr++ = PPP_ADDRESS(rptr); /* assumes the ppp header is */ + *wptr++ = PPP_CONTROL(rptr); /* all in one mbuf */ + *wptr++ = 0; /* change the protocol */ + *wptr++ = PPP_COMP; + *wptr++ = db->seqno >> 8; + *wptr++ = db->seqno; + } + ++db->seqno; + + olen = 0; + rptr += PPP_HDRLEN; + slen = mp->m_len - PPP_HDRLEN; + ilen = slen + 1; + for (;;) { + if (slen <= 0) { + mp = mp->m_next; + if (!mp) + break; + rptr = mtod(mp, u_char *); + slen = mp->m_len; + if (!slen) + continue; /* handle 0-length buffers */ + ilen += slen; + } + + slen--; + c = *rptr++; + fcode = BSD_KEY(ent, c); + hval = BSD_HASH(ent, c, hshift); + dictp = &db->dict[hval]; + + /* Validate and then check the entry. */ + if (dictp->codem1 >= max_ent) + goto nomatch; + if (dictp->f.fcode == fcode) { + ent = dictp->codem1+1; + continue; /* found (prefix,suffix) */ + } + + /* continue probing until a match or invalid entry */ + disp = (hval == 0) ? 1 : hval; + do { + hval += disp; + if (hval >= db->hsize) + hval -= db->hsize; + dictp = &db->dict[hval]; + if (dictp->codem1 >= max_ent) + goto nomatch; + } while (dictp->f.fcode != fcode); + ent = dictp->codem1 + 1; /* finally found (prefix,suffix) */ + continue; + + nomatch: + OUTPUT(ent); /* output the prefix */ + + /* code -> hashtable */ + if (max_ent < db->maxmaxcode) { + struct bsd_dict *dictp2; + /* expand code size if needed */ + if (max_ent >= MAXCODE(n_bits)) + db->n_bits = ++n_bits; + + /* Invalidate old hash table entry using + * this code, and then take it over. + */ + dictp2 = &db->dict[max_ent+1]; + if (db->dict[dictp2->cptr].codem1 == max_ent) + db->dict[dictp2->cptr].codem1 = BADCODEM1; + dictp2->cptr = hval; + dictp->codem1 = max_ent; + dictp->f.fcode = fcode; + + db->max_ent = ++max_ent; + } + ent = c; + } + + OUTPUT(ent); /* output the last code */ + db->bytes_out += olen; + db->in_count += ilen; + if (bitno < 32) + ++db->bytes_out; /* count complete bytes */ + + if (bsd_check(db)) + OUTPUT(CLEAR); /* do not count the CLEAR */ + + /* + * Pad dribble bits of last code with ones. + * Do not emit a completely useless byte of ones. + */ + if (bitno != 32) + PUTBYTE((accm | (0xff << (bitno-8))) >> 24); + + if (m != NULL) { + m->m_len = wptr - mtod(m, u_char *); + m->m_next = NULL; + } + + /* + * Increase code size if we would have without the packet + * boundary and as the decompressor will. + */ + if (max_ent >= MAXCODE(n_bits) && max_ent < db->maxmaxcode) + db->n_bits++; + + db->uncomp_bytes += ilen; + ++db->uncomp_count; + if (olen + PPP_HDRLEN + BSD_OVHD > maxolen) { + /* throw away the compressed stuff if it is longer than uncompressed */ + if (*mret != NULL) { + m_freem(*mret); + *mret = NULL; + } + ++db->incomp_count; + db->incomp_bytes += ilen; + } else { + ++db->comp_count; + db->comp_bytes += olen + BSD_OVHD; + } + + return olen + PPP_HDRLEN + BSD_OVHD; +#undef OUTPUT +#undef PUTBYTE +} + + +/* + * Update the "BSD Compress" dictionary on the receiver for + * incompressible data by pretending to compress the incoming data. + */ +static void +bsd_incomp(state, dmsg) + void *state; + struct mbuf *dmsg; +{ + struct bsd_db *db = (struct bsd_db *) state; + u_int hshift = db->hshift; + u_int max_ent = db->max_ent; + u_int n_bits = db->n_bits; + struct bsd_dict *dictp; + u_int32_t fcode; + u_char c; + u_int32_t hval, disp; + int slen, ilen; + u_int bitno = 7; + u_char *rptr; + u_int ent; + + /* + * If the protocol is not in the range we're interested in, + * just return without looking at the packet. If it is, + * the protocol becomes the first byte to "compress". + */ + rptr = mtod(dmsg, u_char *); + ent = PPP_PROTOCOL(rptr); + if (ent < 0x21 || ent > 0xf9) + return; + + db->incomp_count++; + db->seqno++; + ilen = 1; /* count the protocol as 1 byte */ + rptr += PPP_HDRLEN; + slen = dmsg->m_len - PPP_HDRLEN; + for (;;) { + if (slen <= 0) { + dmsg = dmsg->m_next; + if (!dmsg) + break; + rptr = mtod(dmsg, u_char *); + slen = dmsg->m_len; + continue; + } + ilen += slen; + + do { + c = *rptr++; + fcode = BSD_KEY(ent, c); + hval = BSD_HASH(ent, c, hshift); + dictp = &db->dict[hval]; + + /* validate and then check the entry */ + if (dictp->codem1 >= max_ent) + goto nomatch; + if (dictp->f.fcode == fcode) { + ent = dictp->codem1+1; + continue; /* found (prefix,suffix) */ + } + + /* continue probing until a match or invalid entry */ + disp = (hval == 0) ? 1 : hval; + do { + hval += disp; + if (hval >= db->hsize) + hval -= db->hsize; + dictp = &db->dict[hval]; + if (dictp->codem1 >= max_ent) + goto nomatch; + } while (dictp->f.fcode != fcode); + ent = dictp->codem1+1; + continue; /* finally found (prefix,suffix) */ + + nomatch: /* output (count) the prefix */ + bitno += n_bits; + + /* code -> hashtable */ + if (max_ent < db->maxmaxcode) { + struct bsd_dict *dictp2; + /* expand code size if needed */ + if (max_ent >= MAXCODE(n_bits)) + db->n_bits = ++n_bits; + + /* Invalidate previous hash table entry + * assigned this code, and then take it over. + */ + dictp2 = &db->dict[max_ent+1]; + if (db->dict[dictp2->cptr].codem1 == max_ent) + db->dict[dictp2->cptr].codem1 = BADCODEM1; + dictp2->cptr = hval; + dictp->codem1 = max_ent; + dictp->f.fcode = fcode; + + db->max_ent = ++max_ent; + db->lens[max_ent] = db->lens[ent]+1; + } + ent = c; + } while (--slen != 0); + } + bitno += n_bits; /* output (count) the last code */ + db->bytes_out += bitno/8; + db->in_count += ilen; + (void)bsd_check(db); + + ++db->incomp_count; + db->incomp_bytes += ilen; + ++db->uncomp_count; + db->uncomp_bytes += ilen; + + /* Increase code size if we would have without the packet + * boundary and as the decompressor will. + */ + if (max_ent >= MAXCODE(n_bits) && max_ent < db->maxmaxcode) + db->n_bits++; +} + + +/* + * Decompress "BSD Compress". + * + * Because of patent problems, we return DECOMP_ERROR for errors + * found by inspecting the input data and for system problems, but + * DECOMP_FATALERROR for any errors which could possibly be said to + * be being detected "after" decompression. For DECOMP_ERROR, + * we can issue a CCP reset-request; for DECOMP_FATALERROR, we may be + * infringing a patent of Motorola's if we do, so we take CCP down + * instead. + * + * Given that the frame has the correct sequence number and a good FCS, + * errors such as invalid codes in the input most likely indicate a + * bug, so we return DECOMP_FATALERROR for them in order to turn off + * compression, even though they are detected by inspecting the input. + */ +int +bsd_decompress(state, cmp, dmpp) + void *state; + struct mbuf *cmp, **dmpp; +{ + struct bsd_db *db = (struct bsd_db *) state; + u_int max_ent = db->max_ent; + u_int32_t accm = 0; + u_int bitno = 32; /* 1st valid bit in accm */ + u_int n_bits = db->n_bits; + u_int tgtbitno = 32-n_bits; /* bitno when we have a code */ + struct bsd_dict *dictp; + int explen, i, seq, len; + u_int incode, oldcode, finchar; + u_char *p, *rptr, *wptr; + struct mbuf *m, *dmp, *mret; + int adrs, ctrl, ilen; + int space, codelen, extra; + struct mbuf *last; + + /* + * Save the address/control from the PPP header + * and then get the sequence number. + */ + *dmpp = NULL; + rptr = mtod(cmp, u_char *); + adrs = PPP_ADDRESS(rptr); + ctrl = PPP_CONTROL(rptr); + rptr += PPP_HDRLEN; + len = cmp->m_len - PPP_HDRLEN; + seq = 0; + for (i = 0; i < 2; ++i) { + while (len <= 0) { + cmp = cmp->m_next; + if (cmp == NULL) + return DECOMP_ERROR; + rptr = mtod(cmp, u_char *); + len = cmp->m_len; + } + seq = (seq << 8) + *rptr++; + --len; + } + + /* + * Check the sequence number and give up if it differs from + * the value we're expecting. + */ + if (seq != db->seqno) { + if (db->debug) + printf("bsd_decomp%d: bad sequence # %d, expected %d\n", + db->unit, seq, db->seqno - 1); + return DECOMP_ERROR; + } + ++db->seqno; + + /* + * Allocate one mbuf to start with. + */ + MGETHDR(dmp, M_DONTWAIT, MT_DATA); + if (dmp == NULL) + return DECOMP_ERROR; + mret = dmp; + dmp->m_len = 0; + dmp->m_next = NULL; + MCLGET(dmp, M_DONTWAIT); + dmp->m_data += db->hdrlen; + wptr = mtod(dmp, u_char *); + space = M_TRAILINGSPACE(dmp) - PPP_HDRLEN + 1; + + /* + * Fill in the ppp header, but not the last byte of the protocol + * (that comes from the decompressed data). + */ + wptr[0] = adrs; + wptr[1] = ctrl; + wptr[2] = 0; + wptr += PPP_HDRLEN - 1; + + ilen = len; + oldcode = CLEAR; + explen = 0; + for (;;) { + if (len == 0) { + cmp = cmp->m_next; + if (!cmp) /* quit at end of message */ + break; + rptr = mtod(cmp, u_char *); + len = cmp->m_len; + ilen += len; + continue; /* handle 0-length buffers */ + } + + /* + * Accumulate bytes until we have a complete code. + * Then get the next code, relying on the 32-bit, + * unsigned accm to mask the result. + */ + bitno -= 8; + accm |= *rptr++ << bitno; + --len; + if (tgtbitno < bitno) + continue; + incode = accm >> tgtbitno; + accm <<= n_bits; + bitno += n_bits; + + if (incode == CLEAR) { + /* + * The dictionary must only be cleared at + * the end of a packet. But there could be an + * empty mbuf at the end. + */ + if (len > 0 || cmp->m_next != NULL) { + while ((cmp = cmp->m_next) != NULL) + len += cmp->m_len; + if (len > 0) { + m_freem(mret); + if (db->debug) + printf("bsd_decomp%d: bad CLEAR\n", db->unit); + return DECOMP_FATALERROR; /* probably a bug */ + } + } + bsd_clear(db); + explen = ilen = 0; + break; + } + + if (incode > max_ent + 2 || incode > db->maxmaxcode + || incode > max_ent && oldcode == CLEAR) { + m_freem(mret); + if (db->debug) { + printf("bsd_decomp%d: bad code 0x%x oldcode=0x%x ", + db->unit, incode, oldcode); + printf("max_ent=0x%x explen=%d seqno=%d\n", + max_ent, explen, db->seqno); + } + return DECOMP_FATALERROR; /* probably a bug */ + } + + /* Special case for KwKwK string. */ + if (incode > max_ent) { + finchar = oldcode; + extra = 1; + } else { + finchar = incode; + extra = 0; + } + + codelen = db->lens[finchar]; + explen += codelen + extra; + if (explen > db->mru + 1) { + m_freem(mret); + if (db->debug) { + printf("bsd_decomp%d: ran out of mru\n", db->unit); +#ifdef DEBUG + while ((cmp = cmp->m_next) != NULL) + len += cmp->m_len; + printf(" len=%d, finchar=0x%x, codelen=%d, explen=%d\n", + len, finchar, codelen, explen); +#endif + } + return DECOMP_FATALERROR; + } + + /* + * For simplicity, the decoded characters go in a single mbuf, + * so we allocate a single extra cluster mbuf if necessary. + */ + if ((space -= codelen + extra) < 0) { + dmp->m_len = wptr - mtod(dmp, u_char *); + MGET(m, M_DONTWAIT, MT_DATA); + if (m == NULL) { + m_freem(mret); + return DECOMP_ERROR; + } + m->m_len = 0; + m->m_next = NULL; + dmp->m_next = m; + MCLGET(m, M_DONTWAIT); + space = M_TRAILINGSPACE(m) - (codelen + extra); + if (space < 0) { + /* now that's what I call *compression*. */ + m_freem(mret); + return DECOMP_ERROR; + } + dmp = m; + wptr = mtod(dmp, u_char *); + } + + /* + * Decode this code and install it in the decompressed buffer. + */ + p = (wptr += codelen); + while (finchar > LAST) { + dictp = &db->dict[db->dict[finchar].cptr]; +#ifdef DEBUG + if (--codelen <= 0 || dictp->codem1 != finchar-1) + goto bad; +#endif + *--p = dictp->f.hs.suffix; + finchar = dictp->f.hs.prefix; + } + *--p = finchar; + +#ifdef DEBUG + if (--codelen != 0) + printf("bsd_decomp%d: short by %d after code 0x%x, max_ent=0x%x\n", + db->unit, codelen, incode, max_ent); +#endif + + if (extra) /* the KwKwK case again */ + *wptr++ = finchar; + + /* + * If not first code in a packet, and + * if not out of code space, then allocate a new code. + * + * Keep the hash table correct so it can be used + * with uncompressed packets. + */ + if (oldcode != CLEAR && max_ent < db->maxmaxcode) { + struct bsd_dict *dictp2; + u_int32_t fcode; + u_int32_t hval, disp; + + fcode = BSD_KEY(oldcode,finchar); + hval = BSD_HASH(oldcode,finchar,db->hshift); + dictp = &db->dict[hval]; + + /* look for a free hash table entry */ + if (dictp->codem1 < max_ent) { + disp = (hval == 0) ? 1 : hval; + do { + hval += disp; + if (hval >= db->hsize) + hval -= db->hsize; + dictp = &db->dict[hval]; + } while (dictp->codem1 < max_ent); + } + + /* + * Invalidate previous hash table entry + * assigned this code, and then take it over + */ + dictp2 = &db->dict[max_ent+1]; + if (db->dict[dictp2->cptr].codem1 == max_ent) { + db->dict[dictp2->cptr].codem1 = BADCODEM1; + } + dictp2->cptr = hval; + dictp->codem1 = max_ent; + dictp->f.fcode = fcode; + + db->max_ent = ++max_ent; + db->lens[max_ent] = db->lens[oldcode]+1; + + /* Expand code size if needed. */ + if (max_ent >= MAXCODE(n_bits) && max_ent < db->maxmaxcode) { + db->n_bits = ++n_bits; + tgtbitno = 32-n_bits; + } + } + oldcode = incode; + } + dmp->m_len = wptr - mtod(dmp, u_char *); + + /* + * Keep the checkpoint right so that incompressible packets + * clear the dictionary at the right times. + */ + db->bytes_out += ilen; + db->in_count += explen; + if (bsd_check(db) && db->debug) { + printf("bsd_decomp%d: peer should have cleared dictionary\n", + db->unit); + } + + ++db->comp_count; + db->comp_bytes += ilen + BSD_OVHD; + ++db->uncomp_count; + db->uncomp_bytes += explen; + + *dmpp = mret; + return DECOMP_OK; + +#ifdef DEBUG + bad: + if (codelen <= 0) { + printf("bsd_decomp%d: fell off end of chain ", db->unit); + printf("0x%x at 0x%x by 0x%x, max_ent=0x%x\n", + incode, finchar, db->dict[finchar].cptr, max_ent); + } else if (dictp->codem1 != finchar-1) { + printf("bsd_decomp%d: bad code chain 0x%x finchar=0x%x ", + db->unit, incode, finchar); + printf("oldcode=0x%x cptr=0x%x codem1=0x%x\n", oldcode, + db->dict[finchar].cptr, dictp->codem1); + } + m_freem(mret); + return DECOMP_FATALERROR; +#endif /* DEBUG */ +} +#endif /* DO_BSD_COMPRESS */ diff --git a/sys/net/if.c b/sys/net/if.c new file mode 100644 index 00000000000..8fd2138f43f --- /dev/null +++ b/sys/net/if.c @@ -0,0 +1,654 @@ +/* $NetBSD: if.c,v 1.23 1995/08/12 23:59:19 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if.c 8.3 (Berkeley) 1/4/94 + */ + +#include <sys/param.h> +#include <sys/mbuf.h> +#include <sys/systm.h> +#include <sys/proc.h> +#include <sys/socket.h> +#include <sys/socketvar.h> +#include <sys/protosw.h> +#include <sys/kernel.h> +#include <sys/ioctl.h> + +#include <net/if.h> +#include <net/if_dl.h> +#include <net/if_types.h> + +int ifqmaxlen = IFQ_MAXLEN; +void if_slowtimo __P((void *arg)); + +/* + * Network interface utility routines. + * + * Routines with ifa_ifwith* names take sockaddr *'s as + * parameters. + */ +void +ifinit() +{ + register struct ifnet *ifp; + + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) + if (ifp->if_snd.ifq_maxlen == 0) + ifp->if_snd.ifq_maxlen = ifqmaxlen; + if_slowtimo(NULL); +} + +int if_index = 0; +struct ifaddr **ifnet_addrs; +static char *sprint_d __P((u_int, char *, int)); + +/* + * Attach an interface to the + * list of "active" interfaces. + */ +void +if_attach(ifp) + struct ifnet *ifp; +{ + unsigned socksize, ifasize; + int namelen, unitlen, masklen; + char workbuf[12], *unitname; + register struct sockaddr_dl *sdl; + register struct ifaddr *ifa; + static int if_indexlim = 8; + extern void link_rtrequest(); + + if (if_index == 0) + TAILQ_INIT(&ifnet); + TAILQ_INIT(&ifp->if_addrlist); + TAILQ_INSERT_TAIL(&ifnet, ifp, if_list); + ifp->if_index = ++if_index; + if (ifnet_addrs == 0 || if_index >= if_indexlim) { + unsigned n = (if_indexlim <<= 1) * sizeof(ifa); + struct ifaddr **q = (struct ifaddr **) + malloc(n, M_IFADDR, M_WAITOK); + if (ifnet_addrs) { + bcopy((caddr_t)ifnet_addrs, (caddr_t)q, n/2); + free((caddr_t)ifnet_addrs, M_IFADDR); + } + ifnet_addrs = q; + } + /* + * create a Link Level name for this device + */ + unitname = sprint_d((u_int)ifp->if_unit, workbuf, sizeof(workbuf)); + namelen = strlen(ifp->if_name); + unitlen = strlen(unitname); +#define _offsetof(t, m) ((int)((caddr_t)&((t *)0)->m)) + masklen = _offsetof(struct sockaddr_dl, sdl_data[0]) + + unitlen + namelen; + socksize = masklen + ifp->if_addrlen; +#define ROUNDUP(a) (1 + (((a) - 1) | (sizeof(long) - 1))) + if (socksize < sizeof(*sdl)) + socksize = sizeof(*sdl); + socksize = ROUNDUP(socksize); + ifasize = sizeof(*ifa) + 2 * socksize; + ifa = (struct ifaddr *)malloc(ifasize, M_IFADDR, M_WAITOK); + bzero((caddr_t)ifa, ifasize); + sdl = (struct sockaddr_dl *)(ifa + 1); + sdl->sdl_len = socksize; + sdl->sdl_family = AF_LINK; + bcopy(ifp->if_name, sdl->sdl_data, namelen); + bcopy(unitname, namelen + (caddr_t)sdl->sdl_data, unitlen); + sdl->sdl_nlen = (namelen += unitlen); + sdl->sdl_index = ifp->if_index; + sdl->sdl_type = ifp->if_type; + ifnet_addrs[if_index - 1] = ifa; + ifa->ifa_ifp = ifp; + ifa->ifa_rtrequest = link_rtrequest; + TAILQ_INSERT_HEAD(&ifp->if_addrlist, ifa, ifa_list); + ifa->ifa_addr = (struct sockaddr *)sdl; + sdl = (struct sockaddr_dl *)(socksize + (caddr_t)sdl); + ifa->ifa_netmask = (struct sockaddr *)sdl; + sdl->sdl_len = masklen; + while (namelen != 0) + sdl->sdl_data[--namelen] = 0xff; +} +/* + * Locate an interface based on a complete address. + */ +/*ARGSUSED*/ +struct ifaddr * +ifa_ifwithaddr(addr) + register struct sockaddr *addr; +{ + register struct ifnet *ifp; + register struct ifaddr *ifa; + +#define equal(a1, a2) \ + (bcmp((caddr_t)(a1), (caddr_t)(a2), ((struct sockaddr *)(a1))->sa_len) == 0) + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) { + if (ifa->ifa_addr->sa_family != addr->sa_family) + continue; + if (equal(addr, ifa->ifa_addr)) + return (ifa); + if ((ifp->if_flags & IFF_BROADCAST) && ifa->ifa_broadaddr && + equal(ifa->ifa_broadaddr, addr)) + return (ifa); + } + return ((struct ifaddr *)0); +} +/* + * Locate the point to point interface with a given destination address. + */ +/*ARGSUSED*/ +struct ifaddr * +ifa_ifwithdstaddr(addr) + register struct sockaddr *addr; +{ + register struct ifnet *ifp; + register struct ifaddr *ifa; + + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) + if (ifp->if_flags & IFF_POINTOPOINT) + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) { + if (ifa->ifa_addr->sa_family != addr->sa_family) + continue; + if (equal(addr, ifa->ifa_dstaddr)) + return (ifa); + } + return ((struct ifaddr *)0); +} + +/* + * Find an interface on a specific network. If many, choice + * is most specific found. + */ +struct ifaddr * +ifa_ifwithnet(addr) + struct sockaddr *addr; +{ + register struct ifnet *ifp; + register struct ifaddr *ifa; + struct ifaddr *ifa_maybe = 0; + u_int af = addr->sa_family; + char *addr_data = addr->sa_data, *cplim; + + if (af == AF_LINK) { + register struct sockaddr_dl *sdl = (struct sockaddr_dl *)addr; + if (sdl->sdl_index && sdl->sdl_index <= if_index) + return (ifnet_addrs[sdl->sdl_index - 1]); + } + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) { + register char *cp, *cp2, *cp3; + + if (ifa->ifa_addr->sa_family != af || + ifa->ifa_netmask == 0) + next: continue; + cp = addr_data; + cp2 = ifa->ifa_addr->sa_data; + cp3 = ifa->ifa_netmask->sa_data; + cplim = (char *)ifa->ifa_netmask + + ifa->ifa_netmask->sa_len; + while (cp3 < cplim) + if ((*cp++ ^ *cp2++) & *cp3++) + goto next; + if (ifa_maybe == 0 || + rn_refines((caddr_t)ifa->ifa_netmask, + (caddr_t)ifa_maybe->ifa_netmask)) + ifa_maybe = ifa; + } + return (ifa_maybe); +} + +/* + * Find an interface using a specific address family + */ +struct ifaddr * +ifa_ifwithaf(af) + register int af; +{ + register struct ifnet *ifp; + register struct ifaddr *ifa; + + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) + if (ifa->ifa_addr->sa_family == af) + return (ifa); + return ((struct ifaddr *)0); +} + +/* + * Find an interface address specific to an interface best matching + * a given address. + */ +struct ifaddr * +ifaof_ifpforaddr(addr, ifp) + struct sockaddr *addr; + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + register char *cp, *cp2, *cp3; + register char *cplim; + struct ifaddr *ifa_maybe = 0; + u_int af = addr->sa_family; + + if (af >= AF_MAX) + return (0); + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) { + if (ifa->ifa_addr->sa_family != af) + continue; + ifa_maybe = ifa; + if (ifa->ifa_netmask == 0) { + if (equal(addr, ifa->ifa_addr) || + (ifa->ifa_dstaddr && equal(addr, ifa->ifa_dstaddr))) + return (ifa); + continue; + } + cp = addr->sa_data; + cp2 = ifa->ifa_addr->sa_data; + cp3 = ifa->ifa_netmask->sa_data; + cplim = ifa->ifa_netmask->sa_len + (char *)ifa->ifa_netmask; + for (; cp3 < cplim; cp3++) + if ((*cp++ ^ *cp2++) & *cp3) + break; + if (cp3 == cplim) + return (ifa); + } + return (ifa_maybe); +} + +#include <net/route.h> + +/* + * Default action when installing a route with a Link Level gateway. + * Lookup an appropriate real ifa to point to. + * This should be moved to /sys/net/link.c eventually. + */ +void +link_rtrequest(cmd, rt, sa) + int cmd; + register struct rtentry *rt; + struct sockaddr *sa; +{ + register struct ifaddr *ifa; + struct sockaddr *dst; + struct ifnet *ifp; + + if (cmd != RTM_ADD || ((ifa = rt->rt_ifa) == 0) || + ((ifp = ifa->ifa_ifp) == 0) || ((dst = rt_key(rt)) == 0)) + return; + if (ifa = ifaof_ifpforaddr(dst, ifp)) { + IFAFREE(rt->rt_ifa); + rt->rt_ifa = ifa; + ifa->ifa_refcnt++; + if (ifa->ifa_rtrequest && ifa->ifa_rtrequest != link_rtrequest) + ifa->ifa_rtrequest(cmd, rt, sa); + } +} + +/* + * Mark an interface down and notify protocols of + * the transition. + * NOTE: must be called at splsoftnet or equivalent. + */ +void +if_down(ifp) + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + + ifp->if_flags &= ~IFF_UP; + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) + pfctlinput(PRC_IFDOWN, ifa->ifa_addr); + if_qflush(&ifp->if_snd); + rt_ifmsg(ifp); +} + +/* + * Mark an interface up and notify protocols of + * the transition. + * NOTE: must be called at splsoftnet or equivalent. + */ +void +if_up(ifp) + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + + ifp->if_flags |= IFF_UP; +#ifdef notyet + /* this has no effect on IP, and will kill all ISO connections XXX */ + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; ifa = ifa->ifa_list.tqe_next) + pfctlinput(PRC_IFUP, ifa->ifa_addr); +#endif + rt_ifmsg(ifp); +} + +/* + * Flush an interface queue. + */ +void +if_qflush(ifq) + register struct ifqueue *ifq; +{ + register struct mbuf *m, *n; + + n = ifq->ifq_head; + while (m = n) { + n = m->m_act; + m_freem(m); + } + ifq->ifq_head = 0; + ifq->ifq_tail = 0; + ifq->ifq_len = 0; +} + +/* + * Handle interface watchdog timer routines. Called + * from softclock, we decrement timers (if set) and + * call the appropriate interface routine on expiration. + */ +void +if_slowtimo(arg) + void *arg; +{ + register struct ifnet *ifp; + int s = splimp(); + + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) { + if (ifp->if_timer == 0 || --ifp->if_timer) + continue; + if (ifp->if_watchdog) + (*ifp->if_watchdog)(ifp->if_unit); + } + splx(s); + timeout(if_slowtimo, NULL, hz / IFNET_SLOWHZ); +} + +/* + * Map interface name to + * interface structure pointer. + */ +struct ifnet * +ifunit(name) + register char *name; +{ + register char *cp; + register struct ifnet *ifp; + int unit; + unsigned len; + char *ep, c; + + for (cp = name; cp < name + IFNAMSIZ && *cp; cp++) + if (*cp >= '0' && *cp <= '9') + break; + if (*cp == '\0' || cp == name + IFNAMSIZ) + return ((struct ifnet *)0); + /* + * Save first char of unit, and pointer to it, + * so we can put a null there to avoid matching + * initial substrings of interface names. + */ + len = cp - name + 1; + c = *cp; + ep = cp; + for (unit = 0; *cp >= '0' && *cp <= '9'; ) + unit = unit * 10 + *cp++ - '0'; + *ep = 0; + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) { + if (bcmp(ifp->if_name, name, len)) + continue; + if (unit == ifp->if_unit) + break; + } + *ep = c; + return (ifp); +} + +/* + * Interface ioctls. + */ +int +ifioctl(so, cmd, data, p) + struct socket *so; + u_long cmd; + caddr_t data; + struct proc *p; +{ + register struct ifnet *ifp; + register struct ifreq *ifr; + int error; + + switch (cmd) { + + case SIOCGIFCONF: + case OSIOCGIFCONF: + return (ifconf(cmd, data)); + } + ifr = (struct ifreq *)data; + ifp = ifunit(ifr->ifr_name); + if (ifp == 0) + return (ENXIO); + switch (cmd) { + + case SIOCGIFFLAGS: + ifr->ifr_flags = ifp->if_flags; + break; + + case SIOCGIFMETRIC: + ifr->ifr_metric = ifp->if_metric; + break; + + case SIOCSIFFLAGS: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + if (ifp->if_flags & IFF_UP && (ifr->ifr_flags & IFF_UP) == 0) { + int s = splimp(); + if_down(ifp); + splx(s); + } + if (ifr->ifr_flags & IFF_UP && (ifp->if_flags & IFF_UP) == 0) { + int s = splimp(); + if_up(ifp); + splx(s); + } + ifp->if_flags = (ifp->if_flags & IFF_CANTCHANGE) | + (ifr->ifr_flags &~ IFF_CANTCHANGE); + if (ifp->if_ioctl) + (void) (*ifp->if_ioctl)(ifp, cmd, data); + break; + + case SIOCSIFMETRIC: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + ifp->if_metric = ifr->ifr_metric; + break; + + case SIOCADDMULTI: + case SIOCDELMULTI: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + if (ifp->if_ioctl == 0) + return (EOPNOTSUPP); + return ((*ifp->if_ioctl)(ifp, cmd, data)); + + default: + if (so->so_proto == 0) + return (EOPNOTSUPP); +#ifndef COMPAT_43 + return ((*so->so_proto->pr_usrreq)(so, PRU_CONTROL, + cmd, data, ifp)); +#else + { + int ocmd = cmd; + + switch (cmd) { + + case SIOCSIFDSTADDR: + case SIOCSIFADDR: + case SIOCSIFBRDADDR: + case SIOCSIFNETMASK: +#if BYTE_ORDER != BIG_ENDIAN + if (ifr->ifr_addr.sa_family == 0 && + ifr->ifr_addr.sa_len < 16) { + ifr->ifr_addr.sa_family = ifr->ifr_addr.sa_len; + ifr->ifr_addr.sa_len = 16; + } +#else + if (ifr->ifr_addr.sa_len == 0) + ifr->ifr_addr.sa_len = 16; +#endif + break; + + case OSIOCGIFADDR: + cmd = SIOCGIFADDR; + break; + + case OSIOCGIFDSTADDR: + cmd = SIOCGIFDSTADDR; + break; + + case OSIOCGIFBRDADDR: + cmd = SIOCGIFBRDADDR; + break; + + case OSIOCGIFNETMASK: + cmd = SIOCGIFNETMASK; + } + error = ((*so->so_proto->pr_usrreq)(so, PRU_CONTROL, + cmd, data, ifp)); + switch (ocmd) { + + case OSIOCGIFADDR: + case OSIOCGIFDSTADDR: + case OSIOCGIFBRDADDR: + case OSIOCGIFNETMASK: + *(u_int16_t *)&ifr->ifr_addr = ifr->ifr_addr.sa_family; + } + return (error); + + } +#endif + } + return (0); +} + +/* + * Return interface configuration + * of system. List may be used + * in later ioctl's (above) to get + * other information. + */ +/*ARGSUSED*/ +int +ifconf(cmd, data) + u_long cmd; + caddr_t data; +{ + register struct ifconf *ifc = (struct ifconf *)data; + register struct ifnet *ifp; + register struct ifaddr *ifa; + register char *cp, *ep; + struct ifreq ifr, *ifrp; + int space = ifc->ifc_len, error = 0; + + ifrp = ifc->ifc_req; + ep = ifr.ifr_name + sizeof (ifr.ifr_name) - 2; + for (ifp = ifnet.tqh_first; + space > sizeof (ifr) && ifp != 0; ifp = ifp->if_list.tqe_next) { + strncpy(ifr.ifr_name, ifp->if_name, sizeof(ifr.ifr_name) - 2); + for (cp = ifr.ifr_name; cp < ep && *cp; cp++) + continue; + *cp++ = '0' + ifp->if_unit; *cp = '\0'; + if ((ifa = ifp->if_addrlist.tqh_first) == 0) { + bzero((caddr_t)&ifr.ifr_addr, sizeof(ifr.ifr_addr)); + error = copyout((caddr_t)&ifr, (caddr_t)ifrp, + sizeof(ifr)); + if (error) + break; + space -= sizeof (ifr), ifrp++; + } else + for (; space > sizeof (ifr) && ifa != 0; ifa = ifa->ifa_list.tqe_next) { + register struct sockaddr *sa = ifa->ifa_addr; +#ifdef COMPAT_43 + if (cmd == OSIOCGIFCONF) { + struct osockaddr *osa = + (struct osockaddr *)&ifr.ifr_addr; + ifr.ifr_addr = *sa; + osa->sa_family = sa->sa_family; + error = copyout((caddr_t)&ifr, (caddr_t)ifrp, + sizeof (ifr)); + ifrp++; + } else +#endif + if (sa->sa_len <= sizeof(*sa)) { + ifr.ifr_addr = *sa; + error = copyout((caddr_t)&ifr, (caddr_t)ifrp, + sizeof (ifr)); + ifrp++; + } else { + space -= sa->sa_len - sizeof(*sa); + if (space < sizeof (ifr)) + break; + error = copyout((caddr_t)&ifr, (caddr_t)ifrp, + sizeof (ifr.ifr_name)); + if (error == 0) + error = copyout((caddr_t)sa, + (caddr_t)&ifrp->ifr_addr, sa->sa_len); + ifrp = (struct ifreq *) + (sa->sa_len + (caddr_t)&ifrp->ifr_addr); + } + if (error) + break; + space -= sizeof (ifr); + } + } + ifc->ifc_len -= space; + return (error); +} + +static char * +sprint_d(n, buf, buflen) + u_int n; + char *buf; + int buflen; +{ + register char *cp = buf + buflen - 1; + + *cp = 0; + do { + cp--; + *cp = "0123456789"[n % 10]; + n /= 10; + } while (n != 0); + return (cp); +} diff --git a/sys/net/if.h b/sys/net/if.h new file mode 100644 index 00000000000..dada588d55b --- /dev/null +++ b/sys/net/if.h @@ -0,0 +1,355 @@ +/* $NetBSD: if.h,v 1.19 1995/06/19 21:57:28 cgd Exp $ */ + +/* + * Copyright (c) 1982, 1986, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if.h 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/queue.h> + +/* + * Structures defining a network interface, providing a packet + * transport mechanism (ala level 0 of the PUP protocols). + * + * Each interface accepts output datagrams of a specified maximum + * length, and provides higher level routines with input datagrams + * received from its medium. + * + * Output occurs when the routine if_output is called, with four parameters: + * (*ifp->if_output)(ifp, m, dst, rt) + * Here m is the mbuf chain to be sent and dst is the destination address. + * The output routine encapsulates the supplied datagram if necessary, + * and then transmits it on its medium. + * + * On input, each interface unwraps the data received by it, and either + * places it on the input queue of a internetwork datagram routine + * and posts the associated software interrupt, or passes the datagram to a raw + * packet input routine. + * + * Routines exist for locating interfaces by their addresses + * or for locating a interface on a certain network, as well as more general + * routing and gateway routines maintaining information used to locate + * interfaces. These routines live in the files if.c and route.c + */ +/* XXX fast fix for SNMP, going away soon */ +#include <sys/time.h> + +struct mbuf; +struct proc; +struct rtentry; +struct socket; +struct ether_header; + +/* + * Structure defining statistics and other data kept regarding a network + * interface. + */ +struct if_data { + /* generic interface information */ + u_char ifi_type; /* ethernet, tokenring, etc. */ + u_char ifi_addrlen; /* media address length */ + u_char ifi_hdrlen; /* media header length */ + u_long ifi_mtu; /* maximum transmission unit */ + u_long ifi_metric; /* routing metric (external only) */ + u_long ifi_baudrate; /* linespeed */ + /* volatile statistics */ + u_long ifi_ipackets; /* packets received on interface */ + u_long ifi_ierrors; /* input errors on interface */ + u_long ifi_opackets; /* packets sent on interface */ + u_long ifi_oerrors; /* output errors on interface */ + u_long ifi_collisions; /* collisions on csma interfaces */ + u_long ifi_ibytes; /* total number of octets received */ + u_long ifi_obytes; /* total number of octets sent */ + u_long ifi_imcasts; /* packets received via multicast */ + u_long ifi_omcasts; /* packets sent via multicast */ + u_long ifi_iqdrops; /* dropped on input, this interface */ + u_long ifi_noproto; /* destined for unsupported protocol */ + struct timeval ifi_lastchange; /* last updated */ +}; + +/* + * Structure defining a queue for a network interface. + * + * (Would like to call this struct ``if'', but C isn't PL/1.) + */ +TAILQ_HEAD(ifnet_head, ifnet); /* the actual queue head */ + +struct ifnet { /* and the entries */ + char *if_name; /* name, e.g. ``en'' or ``lo'' */ + TAILQ_ENTRY(ifnet) if_list; /* all struct ifnets are chained */ + TAILQ_HEAD(, ifaddr) if_addrlist; /* linked list of addresses per if */ + int if_pcount; /* number of promiscuous listeners */ + caddr_t if_bpf; /* packet filter structure */ + u_short if_index; /* numeric abbreviation for this if */ + short if_unit; /* sub-unit for lower level driver */ + short if_timer; /* time 'til if_watchdog called */ + short if_flags; /* up/down, broadcast, etc. */ + struct if_data if_data; /* statistics and other data about if */ +/* procedure handles */ + int (*if_output) /* output routine (enqueue) */ + __P((struct ifnet *, struct mbuf *, struct sockaddr *, + struct rtentry *)); + void (*if_start) /* initiate output routine */ + __P((struct ifnet *)); + int (*if_ioctl) /* ioctl routine */ + __P((struct ifnet *, u_long, caddr_t)); + int (*if_reset) /* XXX bus reset routine */ + __P((int)); + void (*if_watchdog) /* timer routine */ + __P((int)); + struct ifqueue { + struct mbuf *ifq_head; + struct mbuf *ifq_tail; + int ifq_len; + int ifq_maxlen; + int ifq_drops; + } if_snd; /* output queue */ +}; +#define if_mtu if_data.ifi_mtu +#define if_type if_data.ifi_type +#define if_addrlen if_data.ifi_addrlen +#define if_hdrlen if_data.ifi_hdrlen +#define if_metric if_data.ifi_metric +#define if_baudrate if_data.ifi_baudrate +#define if_ipackets if_data.ifi_ipackets +#define if_ierrors if_data.ifi_ierrors +#define if_opackets if_data.ifi_opackets +#define if_oerrors if_data.ifi_oerrors +#define if_collisions if_data.ifi_collisions +#define if_ibytes if_data.ifi_ibytes +#define if_obytes if_data.ifi_obytes +#define if_imcasts if_data.ifi_imcasts +#define if_omcasts if_data.ifi_omcasts +#define if_iqdrops if_data.ifi_iqdrops +#define if_noproto if_data.ifi_noproto +#define if_lastchange if_data.ifi_lastchange + +#define IFF_UP 0x1 /* interface is up */ +#define IFF_BROADCAST 0x2 /* broadcast address valid */ +#define IFF_DEBUG 0x4 /* turn on debugging */ +#define IFF_LOOPBACK 0x8 /* is a loopback net */ +#define IFF_POINTOPOINT 0x10 /* interface is point-to-point link */ +#define IFF_NOTRAILERS 0x20 /* avoid use of trailers */ +#define IFF_RUNNING 0x40 /* resources allocated */ +#define IFF_NOARP 0x80 /* no address resolution protocol */ +#define IFF_PROMISC 0x100 /* receive all packets */ +#define IFF_ALLMULTI 0x200 /* receive all multicast packets */ +#define IFF_OACTIVE 0x400 /* transmission in progress */ +#define IFF_SIMPLEX 0x800 /* can't hear own transmissions */ +#define IFF_LINK0 0x1000 /* per link layer defined bit */ +#define IFF_LINK1 0x2000 /* per link layer defined bit */ +#define IFF_LINK2 0x4000 /* per link layer defined bit */ +#define IFF_MULTICAST 0x8000 /* supports multicast */ + +/* flags set internally only: */ +#define IFF_CANTCHANGE \ + (IFF_BROADCAST|IFF_POINTOPOINT|IFF_RUNNING|IFF_OACTIVE|\ + IFF_SIMPLEX|IFF_MULTICAST|IFF_ALLMULTI) + +/* + * Output queues (ifp->if_snd) and internetwork datagram level (pup level 1) + * input routines have queues of messages stored on ifqueue structures + * (defined above). Entries are added to and deleted from these structures + * by these macros, which should be called with ipl raised to splimp(). + */ +#define IF_QFULL(ifq) ((ifq)->ifq_len >= (ifq)->ifq_maxlen) +#define IF_DROP(ifq) ((ifq)->ifq_drops++) +#define IF_ENQUEUE(ifq, m) { \ + (m)->m_nextpkt = 0; \ + if ((ifq)->ifq_tail == 0) \ + (ifq)->ifq_head = m; \ + else \ + (ifq)->ifq_tail->m_nextpkt = m; \ + (ifq)->ifq_tail = m; \ + (ifq)->ifq_len++; \ +} +#define IF_PREPEND(ifq, m) { \ + (m)->m_nextpkt = (ifq)->ifq_head; \ + if ((ifq)->ifq_tail == 0) \ + (ifq)->ifq_tail = (m); \ + (ifq)->ifq_head = (m); \ + (ifq)->ifq_len++; \ +} +#define IF_DEQUEUE(ifq, m) { \ + (m) = (ifq)->ifq_head; \ + if (m) { \ + if (((ifq)->ifq_head = (m)->m_nextpkt) == 0) \ + (ifq)->ifq_tail = 0; \ + (m)->m_nextpkt = 0; \ + (ifq)->ifq_len--; \ + } \ +} + +#define IFQ_MAXLEN 50 +#define IFNET_SLOWHZ 1 /* granularity is 1 second */ + +/* + * The ifaddr structure contains information about one address + * of an interface. They are maintained by the different address families, + * are allocated and attached when an address is set, and are linked + * together so all addresses for an interface can be located. + */ +struct ifaddr { + struct sockaddr *ifa_addr; /* address of interface */ + struct sockaddr *ifa_dstaddr; /* other end of p-to-p link */ +#define ifa_broadaddr ifa_dstaddr /* broadcast address interface */ + struct sockaddr *ifa_netmask; /* used to determine subnet */ + struct ifnet *ifa_ifp; /* back-pointer to interface */ + TAILQ_ENTRY(ifaddr) ifa_list; /* list of addresses for interface */ + void (*ifa_rtrequest)(); /* check or clean routes (+ or -)'d */ + u_short ifa_flags; /* mostly rt_flags for cloning */ + short ifa_refcnt; /* count of references */ + int ifa_metric; /* cost of going out this interface */ +}; +#define IFA_ROUTE RTF_UP /* route installed */ + +/* + * Message format for use in obtaining information about interfaces + * from sysctl and the routing socket. + */ +struct if_msghdr { + u_short ifm_msglen; /* to skip over non-understood messages */ + u_char ifm_version; /* future binary compatability */ + u_char ifm_type; /* message type */ + int ifm_addrs; /* like rtm_addrs */ + int ifm_flags; /* value of if_flags */ + u_short ifm_index; /* index for associated ifp */ + struct if_data ifm_data;/* statistics and other data about if */ +}; + +/* + * Message format for use in obtaining information about interface addresses + * from sysctl and the routing socket. + */ +struct ifa_msghdr { + u_short ifam_msglen; /* to skip over non-understood messages */ + u_char ifam_version; /* future binary compatability */ + u_char ifam_type; /* message type */ + int ifam_addrs; /* like rtm_addrs */ + int ifam_flags; /* value of ifa_flags */ + u_short ifam_index; /* index for associated ifp */ + int ifam_metric; /* value of ifa_metric */ +}; + +/* + * Interface request structure used for socket + * ioctl's. All interface ioctl's must have parameter + * definitions which begin with ifr_name. The + * remainder may be interface specific. + */ +struct ifreq { +#define IFNAMSIZ 16 + char ifr_name[IFNAMSIZ]; /* if name, e.g. "en0" */ + union { + struct sockaddr ifru_addr; + struct sockaddr ifru_dstaddr; + struct sockaddr ifru_broadaddr; + short ifru_flags; + int ifru_metric; + caddr_t ifru_data; + } ifr_ifru; +#define ifr_addr ifr_ifru.ifru_addr /* address */ +#define ifr_dstaddr ifr_ifru.ifru_dstaddr /* other end of p-to-p link */ +#define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */ +#define ifr_flags ifr_ifru.ifru_flags /* flags */ +#define ifr_metric ifr_ifru.ifru_metric /* metric */ +#define ifr_data ifr_ifru.ifru_data /* for use by interface */ +}; + +struct ifaliasreq { + char ifra_name[IFNAMSIZ]; /* if name, e.g. "en0" */ + struct sockaddr ifra_addr; + struct sockaddr ifra_broadaddr; + struct sockaddr ifra_mask; +}; + +/* + * Structure used in SIOCGIFCONF request. + * Used to retrieve interface configuration + * for machine (useful for programs which + * must know all networks accessible). + */ +struct ifconf { + int ifc_len; /* size of associated buffer */ + union { + caddr_t ifcu_buf; + struct ifreq *ifcu_req; + } ifc_ifcu; +#define ifc_buf ifc_ifcu.ifcu_buf /* buffer address */ +#define ifc_req ifc_ifcu.ifcu_req /* array of structures returned */ +}; + +#include <net/if_arp.h> + +#ifdef _KERNEL +#define IFAFREE(ifa) \ + if ((ifa)->ifa_refcnt <= 0) \ + ifafree(ifa); \ + else \ + (ifa)->ifa_refcnt--; + +struct ifnet_head ifnet; + +void ether_ifattach __P((struct ifnet *)); +void ether_input __P((struct ifnet *, struct ether_header *, struct mbuf *)); +int ether_output __P((struct ifnet *, + struct mbuf *, struct sockaddr *, struct rtentry *)); +char *ether_sprintf __P((u_char *)); + +void if_attach __P((struct ifnet *)); +void if_down __P((struct ifnet *)); +void if_qflush __P((struct ifqueue *)); +void if_slowtimo __P((void *)); +void if_up __P((struct ifnet *)); +int ifconf __P((u_long, caddr_t)); +void ifinit __P((void)); +int ifioctl __P((struct socket *, u_long, caddr_t, struct proc *)); +int ifpromisc __P((struct ifnet *, int)); +struct ifnet *ifunit __P((char *)); + +struct ifaddr *ifa_ifwithaddr __P((struct sockaddr *)); +struct ifaddr *ifa_ifwithaf __P((int)); +struct ifaddr *ifa_ifwithdstaddr __P((struct sockaddr *)); +struct ifaddr *ifa_ifwithnet __P((struct sockaddr *)); +struct ifaddr *ifa_ifwithroute __P((int, struct sockaddr *, + struct sockaddr *)); +struct ifaddr *ifaof_ifpforaddr __P((struct sockaddr *, struct ifnet *)); +void ifafree __P((struct ifaddr *)); +void link_rtrequest __P((int, struct rtentry *, struct sockaddr *)); + +int loioctl __P((struct ifnet *, u_long, caddr_t)); +void loopattach __P((int)); +int looutput __P((struct ifnet *, + struct mbuf *, struct sockaddr *, struct rtentry *)); +void lortrequest __P((int, struct rtentry *, struct sockaddr *)); +#endif /* _KERNEL */ diff --git a/sys/net/if_arcsubr.c b/sys/net/if_arcsubr.c new file mode 100644 index 00000000000..d4130d69395 --- /dev/null +++ b/sys/net/if_arcsubr.c @@ -0,0 +1,611 @@ +/* $NetBSD: if_arcsubr.c,v 1.5 1995/07/12 08:27:26 cgd Exp $ */ + +/* + * Copyright (c) 1994, 1995 Ignatios Souvatzis + * Copyright (c) 1982, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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: NetBSD: if_ethersubr.c,v 1.9 1994/06/29 06:36:11 cgd Exp + * @(#)if_ethersubr.c 8.1 (Berkeley) 6/10/93 + * + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/kernel.h> +#include <sys/malloc.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/syslog.h> +#include <sys/errno.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/netisr.h> +#include <net/route.h> +#include <net/if_dl.h> +#include <net/if_types.h> + +#ifdef INET +#include <netinet/in.h> +#include <netinet/in_var.h> +#endif +#include <netinet/if_arc.h> + +#ifndef ARC_PHDSMTU +#define ARC_PHDSMTU 1500 +#endif + +/* why isnt this in /sys/sys/mbuf.h?? */ +extern struct mbuf *m_split __P((struct mbuf *, int, int)); +static struct mbuf *arc_defrag __P((struct ifnet *, struct mbuf *)); + +/* + * RC1201 requires us to have this configurable. We have it only per + * machine at the moment... there is no generic "set mtu" ioctl, AFAICS. + * Anyway, it is possible to binpatch this or set it per kernel config + * option. + */ +#if ARC_PHDSMTU > 60480 +ERROR: The arc_phdsmtu is ARC_PHDSMTU, but must not exceed 60480. +#endif +u_int16_t arc_phdsmtu = ARC_PHDSMTU; +u_int8_t arcbroadcastaddr = 0; + +#define senderr(e) { error = (e); goto bad;} +#define SIN(s) ((struct sockaddr_in *)s) + +/* + * ARCnet output routine. + * Encapsulate a packet of type family for the local net. + * Assumes that ifp is actually pointer to arccom structure. + */ +int +arc_output(ifp, m0, dst, rt0) + register struct ifnet *ifp; + struct mbuf *m0; + struct sockaddr *dst; + struct rtentry *rt0; +{ + struct mbuf *m, *m1, *mcopy; + struct rtentry *rt; + struct arccom *ac; + register struct arc_header *ah; + int off, len; + int s, error, newencoding; + u_int8_t atype, adst; + int tfrags, sflag, fsflag, rsflag; + + if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING)) + senderr(ENETDOWN); + + error = newencoding = 0; + ac = (struct arccom *)ifp; + m = m0; + len = m->m_pkthdr.len; + mcopy = m1 = NULL; + + ifp->if_lastchange = time; + if (rt = rt0) { + if ((rt->rt_flags & RTF_UP) == 0) { + if (rt0 = rt = rtalloc1(dst, 1)) + rt->rt_refcnt--; + else + senderr(EHOSTUNREACH); + } + if (rt->rt_flags & RTF_GATEWAY) { + if (rt->rt_gwroute == 0) + goto lookup; + if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) { + rtfree(rt); rt = rt0; + lookup: rt->rt_gwroute = rtalloc1(rt->rt_gateway, 1); + if ((rt = rt->rt_gwroute) == 0) + senderr(EHOSTUNREACH); + } + } + if (rt->rt_flags & RTF_REJECT) + if (rt->rt_rmx.rmx_expire == 0 || + time.tv_sec < rt->rt_rmx.rmx_expire) + senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH); + } + + switch (dst->sa_family) { +#ifdef INET + case AF_INET: + + /* + * For now, use the simple IP addr -> ARCnet addr mapping + */ + if (m->m_flags & M_BCAST) + adst = arcbroadcastaddr; /* ARCnet broadcast address */ + else + adst = ntohl(SIN(dst)->sin_addr.s_addr) & 0xFF; + + /* If broadcasting on a simplex interface, loopback a copy */ + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX)) + mcopy = m_copy(m, 0, (int)M_COPYALL); + off = m->m_pkthdr.len - m->m_len; + if (ifp->if_flags & IFF_LINK0) { + atype = ARCTYPE_IP; + newencoding = 1; + } else { + atype = ARCTYPE_IP_OLD; + newencoding = 0; + } + break; +#endif + + case AF_UNSPEC: + ah = (struct arc_header *)dst->sa_data; + adst = ah->arc_dhost; + atype = ah->arc_type; + break; + + default: + printf("%s%d: can't handle af%d\n", ifp->if_name, ifp->if_unit, + dst->sa_family); + senderr(EAFNOSUPPORT); + } + + if (mcopy) + (void) looutput(ifp, mcopy, dst, rt); + + /* + * Add local net header. If no space in first mbuf, + * allocate another. + * + * For ARCnet, this is just symbolic. The header changes + * form and position on its way into the hardware and out of + * the wire. At this point, it contains source, destination and + * packet type. + */ + if (newencoding) { + ++ac->ac_seqid; /* make the seqid unique */ + + tfrags = (len + 503) / 504; + fsflag = 2*tfrags-3; + sflag = 0; + rsflag = fsflag; + + while (sflag < fsflag) { + /* we CAN'T have short packets here */ + m1 = m_split(m, 504, M_DONTWAIT); + if (m1 == 0) + senderr(ENOBUFS); + + M_PREPEND(m, ARC_HDRNEWLEN, M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + m = m_pullup(m,ARC_HDRNEWLEN); + if (m == 0) + senderr(ENOBUFS); + + ah = mtod(m, struct arc_header *); + ah->arc_type = atype; + ah->arc_dhost = adst; + ah->arc_shost = ac->ac_anaddr; + ah->arc_flag = rsflag; + ah->arc_seqid = ac->ac_seqid; + + s = splimp(); + /* + * Queue message on interface, and start output if + * interface not yet active. + */ + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + splx(s); + senderr(ENOBUFS); + } + IF_ENQUEUE(&ifp->if_snd, m); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + splx(s); + + /* we don't count the hardwares lenght bytes here */ + ifp->if_obytes += 504 + ARC_HDRNEWLEN; + + m = m1; + len -= 504; + sflag += 2; + rsflag = sflag; + } + m1 = NULL; + + M_PREPEND(m, ARC_HDRNEWLEN, M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + m = m_pullup(m, ARC_HDRNEWLEN); + if (m == 0) + senderr(ENOBUFS); + + ah = mtod(m, struct arc_header *); + ah->arc_type = atype; + ah->arc_flag = sflag; + ah->arc_seqid= ac->ac_seqid; + + /* here we can have small, especially forbidden packets */ + + if ((len >= ARC_MIN_FORBID_LEN - ARC_HDRNEWLEN + 2) && + (len <= ARC_MAX_FORBID_LEN - ARC_HDRNEWLEN + 2)) { + + M_PREPEND(m, 4, M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + + m = m_pullup(m, ARC_HDRNEWLEN); + if (m == 0) + senderr(ENOBUFS); + + ah = mtod(m, struct arc_header *); + ah->arc_type = atype; + ah->arc_flag = 0xFF; + ah->arc_seqid= 0xFFFF; + len += 4; + } + + ah->arc_dhost= adst; + ah->arc_shost= ac->ac_anaddr; + + s = splimp(); + /* + * Queue message on interface, and start output if interface + * not yet active. + */ + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + splx(s); + senderr(ENOBUFS); + } + IF_ENQUEUE(&ifp->if_snd, m); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + splx(s); + + ifp->if_obytes += len + ARC_HDRNEWLEN; + + } else { + M_PREPEND(m, ARC_HDRLEN, M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + ah = mtod(m, struct arc_header *); + ah->arc_type = atype; + ah->arc_dhost= adst; + ah->arc_shost= ac->ac_anaddr; + s = splimp(); + /* + * Queue message on interface, and start output if interface + * not yet active. + */ + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + splx(s); + senderr(ENOBUFS); + } + IF_ENQUEUE(&ifp->if_snd, m); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + splx(s); + + ifp->if_obytes += len + ARC_HDRLEN; + } + return (error); + +bad: + if (m1) + m_freem(m1); + if (m) + m_freem(m); + return (error); +} + +/* + * Defragmenter. Returns mbuf if last packet found, else + * NULL. frees imcoming mbuf as necessary. + */ + +__inline struct mbuf * +arc_defrag(ifp, m) + struct ifnet *ifp; + struct mbuf *m; +{ + struct arc_header *ah, *ah1; + struct arccom *ac; + struct ac_frag *af; + struct mbuf *m1; + char *s; + int newflen; + u_char src,dst,typ; + + ac = (struct arccom *)ifp; + + m = m_pullup(m, ARC_HDRNEWLEN); + if (m == NULL) { + ++ifp->if_ierrors; + return NULL; + } + + ah = mtod(m, struct arc_header *); + typ = ah->arc_type; + + if (!arc_isphds(typ)) + return m; + + src = ah->arc_shost; + dst = ah->arc_dhost; + + if (ah->arc_flag == 0xff) { + m_adj(m, 4); + + m = m_pullup(m, ARC_HDRNEWLEN); + if (m == NULL) { + ++ifp->if_ierrors; + return NULL; + } + + ah = mtod(m, struct arc_header *); + ah->arc_shost = src; + ah->arc_dhost = dst; + ah->arc_type = typ; + } + + af = &ac->ac_fragtab[src]; + m1 = af->af_packet; + s = "debug code error"; + + if (ah->arc_flag & 1) { + /* + * first fragment. We always initialize, which is + * about the right thing to do, as we only want to + * accept one fragmented packet per src at a time. + */ + if (m1 != NULL) + m_freem(m1); + + af->af_packet = m; + m1 = m; + af->af_maxflag = ah->arc_flag; + af->af_lastseen = 0; + af->af_seqid = ah->arc_seqid; + + return NULL; + /* notreached */ + } else { + /* check for unfragmented packet */ + if (ah->arc_flag == 0) + return m; + + /* do we have a first packet from that src? */ + if (m1 == NULL) { + s = "no first frag"; + goto outofseq; + } + + ah1 = mtod(m1, struct arc_header *); + + if (ah->arc_seqid != ah1->arc_seqid) { + s = "seqid differs"; + goto outofseq; + } + + if (ah->arc_type != ah1->arc_type) { + s = "type differs"; + goto outofseq; + } + + if (ah->arc_dhost != ah1->arc_dhost) { + s = "dest host differs"; + goto outofseq; + } + + /* typ, seqid and dst are ok here. */ + + if (ah->arc_flag == af->af_lastseen) { + m_freem(m); + return NULL; + } + + if (ah->arc_flag == af->af_lastseen + 2) { + /* ok, this is next fragment */ + af->af_lastseen = ah->arc_flag; + m_adj(m,ARC_HDRNEWLEN); + + /* + * m_cat might free the first mbuf (with pkthdr) + * in 2nd chain; therefore: + */ + + newflen = m->m_pkthdr.len; + + m_cat(m1,m); + + m1->m_pkthdr.len += newflen; + + /* is it the last one? */ + if (af->af_lastseen > af->af_maxflag) { + af->af_packet = NULL; + return(m1); + } else + return NULL; + } + s = "other reason"; + /* if all else fails, it is out of sequence, too */ + } +outofseq: + if (m1) { + m_freem(m1); + af->af_packet = NULL; + } + + if (m) + m_freem(m); + + log(LOG_INFO,"%s%d: got out of seq. packet: %s\n", + ifp->if_name, ifp->if_unit, s); + + return NULL; +} + +/* + * return 1 if Packet Header Definition Standard, else 0. + * For now: old IP, old ARP aren't obviously. Lacking correct information, + * we guess that besides new IP and new ARP also IPX and APPLETALK are PHDS. + * (Apple and Novell corporations were involved, among others, in PHDS work). + * Easiest is to assume that everybody else uses that, too. + */ +int +arc_isphds(type) + int type; +{ + return ((type != ARCTYPE_IP_OLD && + type != ARCTYPE_ARP_OLD)); +} + +/* + * Process a received Arcnet packet; + * the packet is in the mbuf chain m with + * the ARCnet header. + */ +void +arc_input(ifp, m) + struct ifnet *ifp; + struct mbuf *m; +{ + register struct arc_header *ah; + register struct ifqueue *inq; + u_int8_t atype; + int s; + + if ((ifp->if_flags & IFF_UP) == 0) { + m_freem(m); + return; + } + + /* possibly defragment: */ + m = arc_defrag(ifp, m); + if (m == NULL) + return; + + ah = mtod(m, struct arc_header *); + + ifp->if_lastchange = time; + ifp->if_ibytes += m->m_pkthdr.len; + + if (arcbroadcastaddr == ah->arc_dhost) { + m->m_flags |= M_BCAST; + ifp->if_imcasts++; + } + + atype = ah->arc_type; + switch (atype) { +#ifdef INET + case ARCTYPE_IP: + m_adj(m,ARC_HDRNEWLEN); + schednetisr(NETISR_IP); + inq = &ipintrq; + break; + + case ARCTYPE_IP_OLD: + m_adj(m,ARC_HDRLEN); + schednetisr(NETISR_IP); + inq = &ipintrq; + break; +#endif + default: + m_freem(m); + return; + } + + s = splimp(); + if (IF_QFULL(inq)) { + IF_DROP(inq); + m_freem(m); + } else + IF_ENQUEUE(inq, m); + splx(s); +} + +/* + * Convert Arcnet address to printable (loggable) representation. + */ +static char digits[] = "0123456789abcdef"; +char * +arc_sprintf(ap) + register u_int8_t *ap; +{ + static char arcbuf[3]; + register char *cp = arcbuf; + + *cp++ = digits[*ap >> 4]; + *cp++ = digits[*ap++ & 0xf]; + *cp = 0; + return (arcbuf); +} + +/* + * Perform common duties while attaching to interface list + */ +void +arc_ifattach(ifp) + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + register struct sockaddr_dl *sdl; + register struct arccom *ac; + + ifp->if_type = IFT_ARCNET; + ifp->if_addrlen = 1; + ifp->if_hdrlen = ARC_HDRLEN; + if (ifp->if_flags & IFF_LINK0 && arc_phdsmtu > 60480) + log(LOG_ERR, + "%s%d: arc_phdsmtu is %d, but must not exceed 60480", + ifp->if_name, ifp->if_unit, arc_phdsmtu); + + ifp->if_mtu = (ifp->if_flags & IFF_LINK0 ? arc_phdsmtu : ARCMTU); + ac = (struct arccom *)ifp; + ac->ac_seqid = (time.tv_sec) & 0xFFFF; /* try to make seqid unique */ + if (ac->ac_anaddr == 0) { + /* XXX this message isn't entirely clear, to me -- cgd */ + log(LOG_ERR,"%s%d: link address 0 reserved for broadcasts. Please change it and ifconfig %s%d down up\n", + ifp->if_name,ifp->if_unit,ifp->if_name,ifp->if_unit); + } + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; + ifa = ifa->ifa_list.tqe_next) + if ((sdl = (struct sockaddr_dl *)ifa->ifa_addr) && + sdl->sdl_family == AF_LINK) { + sdl->sdl_type = IFT_ARCNET; + sdl->sdl_alen = ifp->if_addrlen; + bcopy((caddr_t)&((struct arccom *)ifp)->ac_anaddr, + LLADDR(sdl), ifp->if_addrlen); + break; + } +} diff --git a/sys/net/if_arp.h b/sys/net/if_arp.h new file mode 100644 index 00000000000..a15af81ae96 --- /dev/null +++ b/sys/net/if_arp.h @@ -0,0 +1,87 @@ +/* $NetBSD: if_arp.h,v 1.8 1995/03/08 02:56:52 cgd Exp $ */ + +/* + * Copyright (c) 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_arp.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * Address Resolution Protocol. + * + * See RFC 826 for protocol description. ARP packets are variable + * in size; the arphdr structure defines the fixed-length portion. + * Protocol type values are the same as those for 10 Mb/s Ethernet. + * It is followed by the variable-sized fields ar_sha, arp_spa, + * arp_tha and arp_tpa in that order, according to the lengths + * specified. Field names used correspond to RFC 826. + */ +struct arphdr { + u_int16_t ar_hrd; /* format of hardware address */ +#define ARPHRD_ETHER 1 /* ethernet hardware format */ +#define ARPHRD_FRELAY 15 /* frame relay hardware format */ + u_int16_t ar_pro; /* format of protocol address */ + u_int8_t ar_hln; /* length of hardware address */ + u_int8_t ar_pln; /* length of protocol address */ + u_int16_t ar_op; /* one of: */ +#define ARPOP_REQUEST 1 /* request to resolve address */ +#define ARPOP_REPLY 2 /* response to previous request */ +#define ARPOP_REVREQUEST 3 /* request protocol address given hardware */ +#define ARPOP_REVREPLY 4 /* response giving protocol address */ +#define ARPOP_INVREQUEST 8 /* request to identify peer */ +#define ARPOP_INVREPLY 9 /* response identifying peer */ +/* + * The remaining fields are variable in size, + * according to the sizes above. + */ +#ifdef COMMENT_ONLY + u_int8_t ar_sha[]; /* sender hardware address */ + u_int8_t ar_spa[]; /* sender protocol address */ + u_int8_t ar_tha[]; /* target hardware address */ + u_int8_t ar_tpa[]; /* target protocol address */ +#endif +}; + +/* + * ARP ioctl request + */ +struct arpreq { + struct sockaddr arp_pa; /* protocol address */ + struct sockaddr arp_ha; /* hardware address */ + int arp_flags; /* flags */ +}; +/* arp_flags and at_flags field values */ +#define ATF_INUSE 0x01 /* entry in use */ +#define ATF_COM 0x02 /* completed entry (enaddr valid) */ +#define ATF_PERM 0x04 /* permanent entry */ +#define ATF_PUBL 0x08 /* publish entry (respond for other host) */ +#define ATF_USETRAILERS 0x10 /* has requested trailers */ diff --git a/sys/net/if_dl.h b/sys/net/if_dl.h new file mode 100644 index 00000000000..0536717a9a4 --- /dev/null +++ b/sys/net/if_dl.h @@ -0,0 +1,82 @@ +/* $NetBSD: if_dl.h,v 1.8 1995/03/26 20:30:13 jtc Exp $ */ + +/* + * Copyright (c) 1990, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_dl.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * A Link-Level Sockaddr may specify the interface in one of two + * ways: either by means of a system-provided index number (computed + * anew and possibly differently on every reboot), or by a human-readable + * string such as "il0" (for managerial convenience). + * + * Census taking actions, such as something akin to SIOCGCONF would return + * both the index and the human name. + * + * High volume transactions (such as giving a link-level ``from'' address + * in a recvfrom or recvmsg call) may be likely only to provide the indexed + * form, (which requires fewer copy operations and less space). + * + * The form and interpretation of the link-level address is purely a matter + * of convention between the device driver and its consumers; however, it is + * expected that all drivers for an interface of a given if_type will agree. + */ + +/* + * Structure of a Link-Level sockaddr: + */ +struct sockaddr_dl { + u_char sdl_len; /* Total length of sockaddr */ + u_char sdl_family; /* AF_DLI */ + u_int16_t sdl_index; /* if != 0, system given index for interface */ + u_char sdl_type; /* interface type */ + u_char sdl_nlen; /* interface name length, no trailing 0 reqd. */ + u_char sdl_alen; /* link level address length */ + u_char sdl_slen; /* link layer selector length */ + char sdl_data[12]; /* minimum work area, can be larger; + contains both if name and ll address */ +}; + +#define LLADDR(s) ((caddr_t)((s)->sdl_data + (s)->sdl_nlen)) + +#ifndef _KERNEL + +#include <sys/cdefs.h> + +__BEGIN_DECLS +void link_addr __P((const char *, struct sockaddr_dl *)); +char *link_ntoa __P((const struct sockaddr_dl *)); +__END_DECLS + +#endif /* !_KERNEL */ diff --git a/sys/net/if_ethersubr.c b/sys/net/if_ethersubr.c new file mode 100644 index 00000000000..d777c51669f --- /dev/null +++ b/sys/net/if_ethersubr.c @@ -0,0 +1,680 @@ +/* $NetBSD: if_ethersubr.c,v 1.15 1995/09/29 03:37:43 phil Exp $ */ + +/* + * Copyright (c) 1982, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_ethersubr.c 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/kernel.h> +#include <sys/malloc.h> +#include <sys/mbuf.h> +#include <sys/protosw.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/errno.h> +#include <sys/syslog.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/netisr.h> +#include <net/route.h> +#include <net/if_llc.h> +#include <net/if_dl.h> +#include <net/if_types.h> + +#include <netinet/in.h> +#ifdef INET +#include <netinet/in_var.h> +#endif +#include <netinet/if_ether.h> + +#ifdef NS +#include <netns/ns.h> +#include <netns/ns_if.h> +#endif + +#ifdef ISO +#include <netiso/argo_debug.h> +#include <netiso/iso.h> +#include <netiso/iso_var.h> +#include <netiso/iso_snpac.h> +#endif + +#ifdef LLC +#include <netccitt/dll.h> +#include <netccitt/llc_var.h> +#endif + +#if defined(LLC) && defined(CCITT) +extern struct ifqueue pkintrq; +#endif + +u_char etherbroadcastaddr[6] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }; +#define senderr(e) { error = (e); goto bad;} + +/* + * Ethernet output routine. + * Encapsulate a packet of type family for the local net. + * Use trailer local net encapsulation if enough data in first + * packet leaves a multiple of 512 bytes of data in remainder. + * Assumes that ifp is actually pointer to arpcom structure. + */ +int +ether_output(ifp, m0, dst, rt0) + register struct ifnet *ifp; + struct mbuf *m0; + struct sockaddr *dst; + struct rtentry *rt0; +{ + u_int16_t etype; + int s, error = 0; + u_char edst[6]; + register struct mbuf *m = m0; + register struct rtentry *rt; + struct mbuf *mcopy = (struct mbuf *)0; + register struct ether_header *eh; + struct arpcom *ac = (struct arpcom *)ifp; + + if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING)) + senderr(ENETDOWN); + ifp->if_lastchange = time; + if (rt = rt0) { + if ((rt->rt_flags & RTF_UP) == 0) { + if (rt0 = rt = rtalloc1(dst, 1)) + rt->rt_refcnt--; + else + senderr(EHOSTUNREACH); + } + if (rt->rt_flags & RTF_GATEWAY) { + if (rt->rt_gwroute == 0) + goto lookup; + if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) { + rtfree(rt); rt = rt0; + lookup: rt->rt_gwroute = rtalloc1(rt->rt_gateway, 1); + if ((rt = rt->rt_gwroute) == 0) + senderr(EHOSTUNREACH); + } + } + if (rt->rt_flags & RTF_REJECT) + if (rt->rt_rmx.rmx_expire == 0 || + time.tv_sec < rt->rt_rmx.rmx_expire) + senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH); + } + switch (dst->sa_family) { + +#ifdef INET + case AF_INET: + if (!arpresolve(ac, rt, m, dst, edst)) + return (0); /* if not yet resolved */ + /* If broadcasting on a simplex interface, loopback a copy */ + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX)) + mcopy = m_copy(m, 0, (int)M_COPYALL); + etype = ETHERTYPE_IP; + break; +#endif +#ifdef NS + case AF_NS: + etype = ETHERTYPE_NS; + bcopy((caddr_t)&(((struct sockaddr_ns *)dst)->sns_addr.x_host), + (caddr_t)edst, sizeof (edst)); + if (!bcmp((caddr_t)edst, (caddr_t)&ns_thishost, sizeof(edst))) + return (looutput(ifp, m, dst, rt)); + /* If broadcasting on a simplex interface, loopback a copy */ + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX)) + mcopy = m_copy(m, 0, (int)M_COPYALL); + break; +#endif +#ifdef ISO + case AF_ISO: { + int snpalen; + struct llc *l; + register struct sockaddr_dl *sdl; + + if (rt && (sdl = (struct sockaddr_dl *)rt->rt_gateway) && + sdl->sdl_family == AF_LINK && sdl->sdl_alen > 0) { + bcopy(LLADDR(sdl), (caddr_t)edst, sizeof(edst)); + } else if (error = + iso_snparesolve(ifp, (struct sockaddr_iso *)dst, + (char *)edst, &snpalen)) + goto bad; /* Not Resolved */ + /* If broadcasting on a simplex interface, loopback a copy */ + if (*edst & 1) + m->m_flags |= (M_BCAST|M_MCAST); + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX) && + (mcopy = m_copy(m, 0, (int)M_COPYALL))) { + M_PREPEND(mcopy, sizeof (*eh), M_DONTWAIT); + if (mcopy) { + eh = mtod(mcopy, struct ether_header *); + bcopy((caddr_t)edst, + (caddr_t)eh->ether_dhost, sizeof (edst)); + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)eh->ether_shost, sizeof (edst)); + } + } + M_PREPEND(m, 3, M_DONTWAIT); + if (m == NULL) + return (0); + etype = m->m_pkthdr.len; + l = mtod(m, struct llc *); + l->llc_dsap = l->llc_ssap = LLC_ISO_LSAP; + l->llc_control = LLC_UI; + IFDEBUG(D_ETHER) + int i; + printf("unoutput: sending pkt to: "); + for (i=0; i<6; i++) + printf("%x ", edst[i] & 0xff); + printf("\n"); + ENDDEBUG + } break; +#endif /* ISO */ +#ifdef LLC +/* case AF_NSAP: */ + case AF_CCITT: { + register struct sockaddr_dl *sdl = + (struct sockaddr_dl *) rt -> rt_gateway; + + if (sdl && sdl->sdl_family == AF_LINK + && sdl->sdl_alen > 0) { + bcopy(LLADDR(sdl), (char *)edst, + sizeof(edst)); + } else goto bad; /* Not a link interface ? Funny ... */ + if ((ifp->if_flags & IFF_SIMPLEX) && (*edst & 1) && + (mcopy = m_copy(m, 0, (int)M_COPYALL))) { + M_PREPEND(mcopy, sizeof (*eh), M_DONTWAIT); + if (mcopy) { + eh = mtod(mcopy, struct ether_header *); + bcopy((caddr_t)edst, + (caddr_t)eh->ether_dhost, sizeof (edst)); + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)eh->ether_shost, sizeof (edst)); + } + } + etype = m->m_pkthdr.len; +#ifdef LLC_DEBUG + { + int i; + register struct llc *l = mtod(m, struct llc *); + + printf("ether_output: sending LLC2 pkt to: "); + for (i=0; i<6; i++) + printf("%x ", edst[i] & 0xff); + printf(" len 0x%x dsap 0x%x ssap 0x%x control 0x%x\n", + etype & 0xff, l->llc_dsap & 0xff, l->llc_ssap &0xff, + l->llc_control & 0xff); + + } +#endif /* LLC_DEBUG */ + } break; +#endif /* LLC */ + + case AF_UNSPEC: + eh = (struct ether_header *)dst->sa_data; + bcopy((caddr_t)eh->ether_dhost, (caddr_t)edst, sizeof (edst)); + /* AF_UNSPEC doesn't swap the byte order of the ether_type. */ + etype = ntohs(eh->ether_type); + break; + + default: + printf("%s%d: can't handle af%d\n", ifp->if_name, ifp->if_unit, + dst->sa_family); + senderr(EAFNOSUPPORT); + } + + + if (mcopy) + (void) looutput(ifp, mcopy, dst, rt); + /* + * Add local net header. If no space in first mbuf, + * allocate another. + */ + M_PREPEND(m, sizeof (struct ether_header), M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + eh = mtod(m, struct ether_header *); + etype = htons(etype); + bcopy((caddr_t)&etype,(caddr_t)&eh->ether_type, + sizeof(eh->ether_type)); + bcopy((caddr_t)edst, (caddr_t)eh->ether_dhost, sizeof (edst)); + bcopy((caddr_t)ac->ac_enaddr, (caddr_t)eh->ether_shost, + sizeof(eh->ether_shost)); + s = splimp(); + /* + * Queue message on interface, and start output if interface + * not yet active. + */ + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + splx(s); + senderr(ENOBUFS); + } + ifp->if_obytes += m->m_pkthdr.len; + IF_ENQUEUE(&ifp->if_snd, m); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + splx(s); + if (m->m_flags & M_MCAST) + ifp->if_omcasts++; + return (error); + +bad: + if (m) + m_freem(m); + return (error); +} + +/* + * Process a received Ethernet packet; + * the packet is in the mbuf chain m without + * the ether header, which is provided separately. + */ +void +ether_input(ifp, eh, m) + struct ifnet *ifp; + register struct ether_header *eh; + struct mbuf *m; +{ + register struct ifqueue *inq; + register struct llc *l; + u_int16_t etype; + struct arpcom *ac = (struct arpcom *)ifp; + int s; + + if ((ifp->if_flags & IFF_UP) == 0) { + m_freem(m); + return; + } + ifp->if_lastchange = time; + ifp->if_ibytes += m->m_pkthdr.len + sizeof (*eh); + if (eh->ether_dhost[0] & 1) { + if (bcmp((caddr_t)etherbroadcastaddr, (caddr_t)eh->ether_dhost, + sizeof(etherbroadcastaddr)) == 0) + m->m_flags |= M_BCAST; + else + m->m_flags |= M_MCAST; + } + if (m->m_flags & (M_BCAST|M_MCAST)) + ifp->if_imcasts++; + + etype = ntohs(eh->ether_type); + switch (etype) { +#ifdef INET + case ETHERTYPE_IP: + schednetisr(NETISR_IP); + inq = &ipintrq; + break; + + case ETHERTYPE_ARP: + schednetisr(NETISR_ARP); + inq = &arpintrq; + break; + + case ETHERTYPE_REVARP: + revarpinput(m); /* XXX queue? */ + return; +#endif +#ifdef NS + case ETHERTYPE_NS: + schednetisr(NETISR_NS); + inq = &nsintrq; + break; + +#endif + default: +#if defined (ISO) || defined (LLC) + if (etype > ETHERMTU) + goto dropanyway; + l = mtod(m, struct llc *); + switch (l->llc_dsap) { +#ifdef ISO + case LLC_ISO_LSAP: + switch (l->llc_control) { + case LLC_UI: + /* LLC_UI_P forbidden in class 1 service */ + if ((l->llc_dsap == LLC_ISO_LSAP) && + (l->llc_ssap == LLC_ISO_LSAP)) { + /* LSAP for ISO */ + if (m->m_pkthdr.len > etype) + m_adj(m, etype - m->m_pkthdr.len); + m->m_data += 3; /* XXX */ + m->m_len -= 3; /* XXX */ + m->m_pkthdr.len -= 3; /* XXX */ + M_PREPEND(m, sizeof *eh, M_DONTWAIT); + if (m == 0) + return; + *mtod(m, struct ether_header *) = *eh; + IFDEBUG(D_ETHER) + printf("clnp packet"); + ENDDEBUG + schednetisr(NETISR_ISO); + inq = &clnlintrq; + break; + } + goto dropanyway; + + case LLC_XID: + case LLC_XID_P: + if(m->m_len < 6) + goto dropanyway; + l->llc_window = 0; + l->llc_fid = 9; + l->llc_class = 1; + l->llc_dsap = l->llc_ssap = 0; + /* Fall through to */ + case LLC_TEST: + case LLC_TEST_P: + { + struct sockaddr sa; + register struct ether_header *eh2; + int i; + u_char c = l->llc_dsap; + + l->llc_dsap = l->llc_ssap; + l->llc_ssap = c; + if (m->m_flags & (M_BCAST | M_MCAST)) + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)eh->ether_dhost, 6); + sa.sa_family = AF_UNSPEC; + sa.sa_len = sizeof(sa); + eh2 = (struct ether_header *)sa.sa_data; + for (i = 0; i < 6; i++) { + eh2->ether_shost[i] = c = eh->ether_dhost[i]; + eh2->ether_dhost[i] = + eh->ether_dhost[i] = eh->ether_shost[i]; + eh->ether_shost[i] = c; + } + ifp->if_output(ifp, m, &sa, NULL); + return; + } + default: + m_freem(m); + return; + } + break; +#endif /* ISO */ +#ifdef LLC + case LLC_X25_LSAP: + { + if (m->m_pkthdr.len > etype) + m_adj(m, etype - m->m_pkthdr.len); + M_PREPEND(m, sizeof(struct sdl_hdr) , M_DONTWAIT); + if (m == 0) + return; + if ( !sdl_sethdrif(ifp, eh->ether_shost, LLC_X25_LSAP, + eh->ether_dhost, LLC_X25_LSAP, 6, + mtod(m, struct sdl_hdr *))) + panic("ETHER cons addr failure"); + mtod(m, struct sdl_hdr *)->sdlhdr_len = etype; +#ifdef LLC_DEBUG + printf("llc packet\n"); +#endif /* LLC_DEBUG */ + schednetisr(NETISR_CCITT); + inq = &llcintrq; + break; + } +#endif /* LLC */ + dropanyway: + default: + m_freem(m); + return; + } +#else /* ISO || LLC */ + m_freem(m); + return; +#endif /* ISO || LLC */ + } + + s = splimp(); + if (IF_QFULL(inq)) { + IF_DROP(inq); + m_freem(m); + } else + IF_ENQUEUE(inq, m); + splx(s); +} + +/* + * Convert Ethernet address to printable (loggable) representation. + */ +static char digits[] = "0123456789abcdef"; +char * +ether_sprintf(ap) + register u_char *ap; +{ + register i; + static char etherbuf[18]; + register char *cp = etherbuf; + + for (i = 0; i < 6; i++) { + *cp++ = digits[*ap >> 4]; + *cp++ = digits[*ap++ & 0xf]; + *cp++ = ':'; + } + *--cp = 0; + return (etherbuf); +} + +/* + * Perform common duties while attaching to interface list + */ +void +ether_ifattach(ifp) + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + register struct sockaddr_dl *sdl; + + ifp->if_type = IFT_ETHER; + ifp->if_addrlen = 6; + ifp->if_hdrlen = 14; + ifp->if_mtu = ETHERMTU; + ifp->if_output = ether_output; + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; + ifa = ifa->ifa_list.tqe_next) + if ((sdl = (struct sockaddr_dl *)ifa->ifa_addr) && + sdl->sdl_family == AF_LINK) { + sdl->sdl_type = IFT_ETHER; + sdl->sdl_alen = ifp->if_addrlen; + bcopy((caddr_t)((struct arpcom *)ifp)->ac_enaddr, + LLADDR(sdl), ifp->if_addrlen); + break; + } + LIST_INIT(&((struct arpcom *)ifp)->ac_multiaddrs); +} + +u_char ether_ipmulticast_min[6] = { 0x01, 0x00, 0x5e, 0x00, 0x00, 0x00 }; +u_char ether_ipmulticast_max[6] = { 0x01, 0x00, 0x5e, 0x7f, 0xff, 0xff }; +/* + * Add an Ethernet multicast address or range of addresses to the list for a + * given interface. + */ +int +ether_addmulti(ifr, ac) + struct ifreq *ifr; + register struct arpcom *ac; +{ + register struct ether_multi *enm; + struct sockaddr_in *sin; + u_char addrlo[6]; + u_char addrhi[6]; + int s = splimp(); + + switch (ifr->ifr_addr.sa_family) { + + case AF_UNSPEC: + bcopy(ifr->ifr_addr.sa_data, addrlo, 6); + bcopy(addrlo, addrhi, 6); + break; + +#ifdef INET + case AF_INET: + sin = (struct sockaddr_in *)&(ifr->ifr_addr); + if (sin->sin_addr.s_addr == INADDR_ANY) { + /* + * An IP address of INADDR_ANY means listen to all + * of the Ethernet multicast addresses used for IP. + * (This is for the sake of IP multicast routers.) + */ + bcopy(ether_ipmulticast_min, addrlo, 6); + bcopy(ether_ipmulticast_max, addrhi, 6); + } + else { + ETHER_MAP_IP_MULTICAST(&sin->sin_addr, addrlo); + bcopy(addrlo, addrhi, 6); + } + break; +#endif + + default: + splx(s); + return (EAFNOSUPPORT); + } + + /* + * Verify that we have valid Ethernet multicast addresses. + */ + if ((addrlo[0] & 0x01) != 1 || (addrhi[0] & 0x01) != 1) { + splx(s); + return (EINVAL); + } + /* + * See if the address range is already in the list. + */ + ETHER_LOOKUP_MULTI(addrlo, addrhi, ac, enm); + if (enm != NULL) { + /* + * Found it; just increment the reference count. + */ + ++enm->enm_refcount; + splx(s); + return (0); + } + /* + * New address or range; malloc a new multicast record + * and link it into the interface's multicast list. + */ + enm = (struct ether_multi *)malloc(sizeof(*enm), M_IFMADDR, M_NOWAIT); + if (enm == NULL) { + splx(s); + return (ENOBUFS); + } + bcopy(addrlo, enm->enm_addrlo, 6); + bcopy(addrhi, enm->enm_addrhi, 6); + enm->enm_ac = ac; + enm->enm_refcount = 1; + LIST_INSERT_HEAD(&ac->ac_multiaddrs, enm, enm_list); + ac->ac_multicnt++; + splx(s); + /* + * Return ENETRESET to inform the driver that the list has changed + * and its reception filter should be adjusted accordingly. + */ + return (ENETRESET); +} + +/* + * Delete a multicast address record. + */ +int +ether_delmulti(ifr, ac) + struct ifreq *ifr; + register struct arpcom *ac; +{ + register struct ether_multi *enm; + register struct ether_multi **p; + struct sockaddr_in *sin; + u_char addrlo[6]; + u_char addrhi[6]; + int s = splimp(); + + switch (ifr->ifr_addr.sa_family) { + + case AF_UNSPEC: + bcopy(ifr->ifr_addr.sa_data, addrlo, 6); + bcopy(addrlo, addrhi, 6); + break; + +#ifdef INET + case AF_INET: + sin = (struct sockaddr_in *)&(ifr->ifr_addr); + if (sin->sin_addr.s_addr == INADDR_ANY) { + /* + * An IP address of INADDR_ANY means stop listening + * to the range of Ethernet multicast addresses used + * for IP. + */ + bcopy(ether_ipmulticast_min, addrlo, 6); + bcopy(ether_ipmulticast_max, addrhi, 6); + } + else { + ETHER_MAP_IP_MULTICAST(&sin->sin_addr, addrlo); + bcopy(addrlo, addrhi, 6); + } + break; +#endif + + default: + splx(s); + return (EAFNOSUPPORT); + } + + /* + * Look up the address in our list. + */ + ETHER_LOOKUP_MULTI(addrlo, addrhi, ac, enm); + if (enm == NULL) { + splx(s); + return (ENXIO); + } + if (--enm->enm_refcount != 0) { + /* + * Still some claims to this record. + */ + splx(s); + return (0); + } + /* + * No remaining claims to this record; unlink and free it. + */ + LIST_REMOVE(enm, enm_list); + free(enm, M_IFMADDR); + ac->ac_multicnt--; + splx(s); + /* + * Return ENETRESET to inform the driver that the list has changed + * and its reception filter should be adjusted accordingly. + */ + return (ENETRESET); +} diff --git a/sys/net/if_fddi.h b/sys/net/if_fddi.h new file mode 100644 index 00000000000..a61e4fc466c --- /dev/null +++ b/sys/net/if_fddi.h @@ -0,0 +1,85 @@ +/* $NetBSD: if_fddi.h,v 1.2 1995/08/19 04:35:28 cgd Exp $ */ + +/* + * 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_fddi.h 8.1 (Berkeley) 6/10/93 + */ + +#ifndef _NETINET_IF_FDDI_H_ +#define _NETINET_IF_FDDI_H_ + +/* + * Structure of an 100Mb/s FDDI header. + */ +struct fddi_header { + u_char fddi_fc; + u_char fddi_dhost[6]; + u_char fddi_shost[6]; +}; + +#define FDDIMTU 4470 +#define FDDIMIN 3 + +#define FDDIFC_C 0x80 /* 0b10000000 */ +#define FDDIFC_L 0x40 /* 0b01000000 */ +#define FDDIFC_F 0x30 /* 0b00110000 */ +#define FDDIFC_Z 0x0F /* 0b00001111 */ + +#define FDDIFC_LLC_ASYNC 0x50 +#define FDDIFC_LLC_PRIO0 0 +#define FDDIFC_LLC_PRIO1 1 +#define FDDIFC_LLC_PRIO2 2 +#define FDDIFC_LLC_PRIO3 3 +#define FDDIFC_LLC_PRIO4 4 +#define FDDIFC_LLC_PRIO5 5 +#define FDDIFC_LLC_PRIO6 6 +#define FDDIFC_LLC_PRIO7 7 +#define FDDIFC_LLC_SYNC 0xd0 +#define FDDIFC_SMT 0x40 + +#if defined(_KERNEL) +#define fddibroadcastaddr etherbroadcastaddr +#define fddi_ipmulticast_min ether_ipmulticast_min +#define fddi_ipmulticast_max ether_ipmulticast_max +#define fddi_addmulti ether_addmulti +#define fddi_delmulti ether_delmulti +#define fddi_sprintf ether_sprintf + +void fddi_ifattach __P((struct ifnet *)); +void fddi_input __P((struct ifnet *, struct fddi_header *, struct mbuf *)); +int fddi_output __P((struct ifnet *, + struct mbuf *, struct sockaddr *, struct rtentry *)); + +#endif + +#endif diff --git a/sys/net/if_fddisubr.c b/sys/net/if_fddisubr.c new file mode 100644 index 00000000000..4da82a76661 --- /dev/null +++ b/sys/net/if_fddisubr.c @@ -0,0 +1,580 @@ +/* $NetBSD: if_fddisubr.c,v 1.2 1995/08/19 04:35:29 cgd Exp $ */ + +/* + * Copyright (c) 1995 + * Matt Thomas. All rights reserved. + * Copyright (c) 1982, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_fddisubr.c 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/kernel.h> +#include <sys/malloc.h> +#include <sys/mbuf.h> +#include <sys/protosw.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/errno.h> +#include <sys/syslog.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/netisr.h> +#include <net/route.h> +#include <net/if_llc.h> +#include <net/if_dl.h> +#include <net/if_types.h> + +#ifdef INET +#include <netinet/in.h> +#include <netinet/in_var.h> +#endif +#include <netinet/if_ether.h> +#if defined(__FreeBSD__) +#include <netinet/if_fddi.h> +#else +#include <net/if_fddi.h> +#endif + +#ifdef NS +#include <netns/ns.h> +#include <netns/ns_if.h> +#endif + +#ifdef DECNET +#include <netdnet/dn.h> +#endif + +#ifdef ISO +#include <netiso/argo_debug.h> +#include <netiso/iso.h> +#include <netiso/iso_var.h> +#include <netiso/iso_snpac.h> +#endif + +#include "bpfilter.h" + +#ifdef LLC +#include <netccitt/dll.h> +#include <netccitt/llc_var.h> +#endif + +#if defined(LLC) && defined(CCITT) +extern struct ifqueue pkintrq; +#endif + +#define senderr(e) { error = (e); goto bad;} + +/* + * This really should be defined in if_llc.h but in case it isn't. + */ +#ifndef llc_snap +#define llc_snap llc_un.type_snap +#endif + +#if defined(__bsdi__) || defined(__NetBSD__) +#define RTALLOC1(a, b) rtalloc1(a, b) +#define ARPRESOLVE(a, b, c, d, e, f) arpresolve(a, b, c, d, e) +#define TYPEHTONS(t) (t) +#elif defined(__FreeBSD__) +#define RTALLOC1(a, b) rtalloc1(a, b, 0UL) +#define ARPRESOLVE(a, b, c, d, e, f) arpresolve(a, b, c, d, e, f) +#define TYPEHTONS(t) (htons(t)) +#endif +/* + * FDDI output routine. + * Encapsulate a packet of type family for the local net. + * Use trailer local net encapsulation if enough data in first + * packet leaves a multiple of 512 bytes of data in remainder. + * Assumes that ifp is actually pointer to arpcom structure. + */ +int +fddi_output(ifp, m0, dst, rt0) + register struct ifnet *ifp; + struct mbuf *m0; + struct sockaddr *dst; + struct rtentry *rt0; +{ + short type; + int s, error = 0; + u_char edst[6]; + register struct mbuf *m = m0; + register struct rtentry *rt; + struct mbuf *mcopy = (struct mbuf *)0; + register struct fddi_header *fh; + struct arpcom *ac = (struct arpcom *)ifp; + + if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING)) + senderr(ENETDOWN); + ifp->if_lastchange = time; + if (rt = rt0) { + if ((rt->rt_flags & RTF_UP) == 0) { + if (rt0 = rt = RTALLOC1(dst, 1)) + rt->rt_refcnt--; + else + senderr(EHOSTUNREACH); + } + if (rt->rt_flags & RTF_GATEWAY) { + if (rt->rt_gwroute == 0) + goto lookup; + if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) { + rtfree(rt); rt = rt0; + lookup: rt->rt_gwroute = RTALLOC1(rt->rt_gateway, 1); + if ((rt = rt->rt_gwroute) == 0) + senderr(EHOSTUNREACH); + } + } + if (rt->rt_flags & RTF_REJECT) + if (rt->rt_rmx.rmx_expire == 0 || + time.tv_sec < rt->rt_rmx.rmx_expire) + senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH); + } + switch (dst->sa_family) { + +#ifdef INET + case AF_INET: + if (!ARPRESOLVE(ac, rt, m, dst, edst, rt0)) + return (0); /* if not yet resolved */ + /* If broadcasting on a simplex interface, loopback a copy */ + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX)) + mcopy = m_copy(m, 0, (int)M_COPYALL); + type = htons(ETHERTYPE_IP); + break; +#endif +#ifdef NS + case AF_NS: + type = htons(ETHERTYPE_NS); + bcopy((caddr_t)&(((struct sockaddr_ns *)dst)->sns_addr.x_host), + (caddr_t)edst, sizeof (edst)); + if (!bcmp((caddr_t)edst, (caddr_t)&ns_thishost, sizeof(edst))) + return (looutput(ifp, m, dst, rt)); + /* If broadcasting on a simplex interface, loopback a copy */ + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX)) + mcopy = m_copy(m, 0, (int)M_COPYALL); + break; +#endif +#ifdef ISO + case AF_ISO: { + int snpalen; + struct llc *l; + register struct sockaddr_dl *sdl; + + if (rt && (sdl = (struct sockaddr_dl *)rt->rt_gateway) && + sdl->sdl_family == AF_LINK && sdl->sdl_alen > 0) { + bcopy(LLADDR(sdl), (caddr_t)edst, sizeof(edst)); + } else if (error = + iso_snparesolve(ifp, (struct sockaddr_iso *)dst, + (char *)edst, &snpalen)) + goto bad; /* Not Resolved */ + /* If broadcasting on a simplex interface, loopback a copy */ + if (*edst & 1) + m->m_flags |= (M_BCAST|M_MCAST); + if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX) && + (mcopy = m_copy(m, 0, (int)M_COPYALL))) { + M_PREPEND(mcopy, sizeof (*fh), M_DONTWAIT); + if (mcopy) { + fh = mtod(mcopy, struct fddi_header *); + bcopy((caddr_t)edst, + (caddr_t)fh->fddi_dhost, sizeof (edst)); + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)fh->fddi_shost, sizeof (edst)); + } + } + M_PREPEND(m, 3, M_DONTWAIT); + if (m == NULL) + return (0); + type = 0; + l = mtod(m, struct llc *); + l->llc_dsap = l->llc_ssap = LLC_ISO_LSAP; + l->llc_control = LLC_UI; + IFDEBUG(D_ETHER) + int i; + printf("unoutput: sending pkt to: "); + for (i=0; i<6; i++) + printf("%x ", edst[i] & 0xff); + printf("\n"); + ENDDEBUG + } break; +#endif /* ISO */ +#ifdef LLC +/* case AF_NSAP: */ + case AF_CCITT: { + register struct sockaddr_dl *sdl = + (struct sockaddr_dl *) rt -> rt_gateway; + + if (sdl && sdl->sdl_family == AF_LINK + && sdl->sdl_alen > 0) { + bcopy(LLADDR(sdl), (char *)edst, + sizeof(edst)); + } else goto bad; /* Not a link interface ? Funny ... */ + if ((ifp->if_flags & IFF_SIMPLEX) && (*edst & 1) && + (mcopy = m_copy(m, 0, (int)M_COPYALL))) { + M_PREPEND(mcopy, sizeof (*fh), M_DONTWAIT); + if (mcopy) { + fh = mtod(mcopy, struct fddi_header *); + bcopy((caddr_t)edst, + (caddr_t)fh->fddi_dhost, sizeof (edst)); + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)fh->fddi_shost, sizeof (edst)); + fh->fddi_fc = FDDIFC_LLC_ASYNC|FDDIFC_LLC_PRIO4; + } + } + type = 0; +#ifdef LLC_DEBUG + { + int i; + register struct llc *l = mtod(m, struct llc *); + + printf("fddi_output: sending LLC2 pkt to: "); + for (i=0; i<6; i++) + printf("%x ", edst[i] & 0xff); + printf(" len 0x%x dsap 0x%x ssap 0x%x control 0x%x\n", + type & 0xff, l->llc_dsap & 0xff, l->llc_ssap &0xff, + l->llc_control & 0xff); + + } +#endif /* LLC_DEBUG */ + } break; +#endif /* LLC */ + + case AF_UNSPEC: + { + struct ether_header *eh; + eh = (struct ether_header *)dst->sa_data; + bcopy((caddr_t)eh->ether_dhost, (caddr_t)edst, sizeof (edst)); + if (*edst & 1) + m->m_flags |= (M_BCAST|M_MCAST); + type = TYPEHTONS(eh->ether_type); + break; + } + +#if NBPFILTER > 0 + case AF_IMPLINK: + { + fh = mtod(m, struct fddi_header *); + error = EPROTONOSUPPORT; + switch (fh->fddi_fc & (FDDIFC_C|FDDIFC_L|FDDIFC_F)) { + case FDDIFC_LLC_ASYNC: { + /* legal priorities are 0 through 7 */ + if ((fh->fddi_fc & FDDIFC_Z) > 7) + goto bad; + break; + } + case FDDIFC_LLC_SYNC: { + /* FDDIFC_Z bits reserved, must be zero */ + if (fh->fddi_fc & FDDIFC_Z) + goto bad; + break; + } + case FDDIFC_SMT: { + /* FDDIFC_Z bits must be non zero */ + if ((fh->fddi_fc & FDDIFC_Z) == 0) + goto bad; + break; + } + default: { + /* anything else is too dangerous */ + goto bad; + } + } + error = 0; + if (fh->fddi_dhost[0] & 1) + m->m_flags |= (M_BCAST|M_MCAST); + goto queue_it; + } +#endif + default: + printf("%s%d: can't handle af%d\n", ifp->if_name, ifp->if_unit, + dst->sa_family); + senderr(EAFNOSUPPORT); + } + + + if (mcopy) + (void) looutput(ifp, mcopy, dst, rt); + if (type != 0) { + register struct llc *l; + M_PREPEND(m, sizeof (struct llc), M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + l = mtod(m, struct llc *); + l->llc_control = LLC_UI; + l->llc_dsap = l->llc_ssap = LLC_SNAP_LSAP; + l->llc_snap.org_code[0] = l->llc_snap.org_code[1] = l->llc_snap.org_code[2] = 0; + bcopy((caddr_t) &type, (caddr_t) &l->llc_snap.ether_type, + sizeof(u_short)); + } + /* + * Add local net header. If no space in first mbuf, + * allocate another. + */ + M_PREPEND(m, sizeof (struct fddi_header), M_DONTWAIT); + if (m == 0) + senderr(ENOBUFS); + fh = mtod(m, struct fddi_header *); + fh->fddi_fc = FDDIFC_LLC_ASYNC|FDDIFC_LLC_PRIO4; + bcopy((caddr_t)edst, (caddr_t)fh->fddi_dhost, sizeof (edst)); + queue_it: + bcopy((caddr_t)ac->ac_enaddr, (caddr_t)fh->fddi_shost, + sizeof(fh->fddi_shost)); + s = splimp(); + /* + * Queue message on interface, and start output if interface + * not yet active. + */ + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + splx(s); + senderr(ENOBUFS); + } + ifp->if_obytes += m->m_pkthdr.len; + IF_ENQUEUE(&ifp->if_snd, m); + if ((ifp->if_flags & IFF_OACTIVE) == 0) + (*ifp->if_start)(ifp); + splx(s); + if (m->m_flags & M_MCAST) + ifp->if_omcasts++; + return (error); + +bad: + if (m) + m_freem(m); + return (error); +} + +/* + * Process a received FDDI packet; + * the packet is in the mbuf chain m without + * the fddi header, which is provided separately. + */ +void +fddi_input(ifp, fh, m) + struct ifnet *ifp; + register struct fddi_header *fh; + struct mbuf *m; +{ + register struct ifqueue *inq; + register struct llc *l; + struct arpcom *ac = (struct arpcom *)ifp; + int s; + + if ((ifp->if_flags & IFF_UP) == 0) { + m_freem(m); + return; + } + ifp->if_lastchange = time; + ifp->if_ibytes += m->m_pkthdr.len + sizeof (*fh); + if (bcmp((caddr_t)fddibroadcastaddr, (caddr_t)fh->fddi_dhost, + sizeof(fddibroadcastaddr)) == 0) + m->m_flags |= M_BCAST; + else if (fh->fddi_dhost[0] & 1) + m->m_flags |= M_MCAST; + if (m->m_flags & (M_BCAST|M_MCAST)) + ifp->if_imcasts++; + + l = mtod(m, struct llc *); + switch (l->llc_dsap) { +#if defined(INET) || defined(NS) || defined(DECNET) + case LLC_SNAP_LSAP: + { + unsigned fddi_type; + if (l->llc_control != LLC_UI || l->llc_ssap != LLC_SNAP_LSAP) + goto dropanyway; + if (l->llc_snap.org_code[0] != 0 || l->llc_snap.org_code[1] != 0|| l->llc_snap.org_code[2] != 0) + goto dropanyway; + fddi_type = ntohs(l->llc_snap.ether_type); + m_adj(m, 8); + switch (fddi_type) { +#ifdef INET + case ETHERTYPE_IP: + schednetisr(NETISR_IP); + inq = &ipintrq; + break; + + case ETHERTYPE_ARP: + schednetisr(NETISR_ARP); + inq = &arpintrq; + break; +#endif +#ifdef NS + case ETHERTYPE_NS: + schednetisr(NETISR_NS); + inq = &nsintrq; + break; +#endif +#ifdef DECNET + case ETHERTYPE_DECENT: + schednetisr(NETISR_DECNET); + inq = &decnetintrq; + break; +#endif + default: + /* printf("fddi_input: unknown protocol 0x%x\n", fddi_type); */ + ifp->if_noproto++; + goto dropanyway; + } + break; + } +#endif /* INET || NS */ +#ifdef ISO + case LLC_ISO_LSAP: + switch (l->llc_control) { + case LLC_UI: + /* LLC_UI_P forbidden in class 1 service */ + if ((l->llc_dsap == LLC_ISO_LSAP) && + (l->llc_ssap == LLC_ISO_LSAP)) { + /* LSAP for ISO */ + m->m_data += 3; /* XXX */ + m->m_len -= 3; /* XXX */ + m->m_pkthdr.len -= 3; /* XXX */ + M_PREPEND(m, sizeof *fh, M_DONTWAIT); + if (m == 0) + return; + *mtod(m, struct fddi_header *) = *fh; + IFDEBUG(D_ETHER) + printf("clnp packet"); + ENDDEBUG + schednetisr(NETISR_ISO); + inq = &clnlintrq; + break; + } + goto dropanyway; + + case LLC_XID: + case LLC_XID_P: + if(m->m_len < 6) + goto dropanyway; + l->llc_window = 0; + l->llc_fid = 9; + l->llc_class = 1; + l->llc_dsap = l->llc_ssap = 0; + /* Fall through to */ + case LLC_TEST: + case LLC_TEST_P: + { + struct sockaddr sa; + register struct ether_header *eh; + int i; + u_char c = l->llc_dsap; + + l->llc_dsap = l->llc_ssap; + l->llc_ssap = c; + if (m->m_flags & (M_BCAST | M_MCAST)) + bcopy((caddr_t)ac->ac_enaddr, + (caddr_t)fh->fddi_dhost, 6); + sa.sa_family = AF_UNSPEC; + sa.sa_len = sizeof(sa); + eh = (struct ether_header *)sa.sa_data; + for (i = 0; i < 6; i++) { + eh->ether_shost[i] = c = fh->fddi_dhost[i]; + eh->ether_dhost[i] = + eh->ether_dhost[i] = fh->fddi_shost[i]; + eh->ether_shost[i] = c; + } + eh->ether_type = 0; + ifp->if_output(ifp, m, &sa, NULL); + return; + } + default: + m_freem(m); + return; + } + break; +#endif /* ISO */ +#ifdef LLC + case LLC_X25_LSAP: + { + M_PREPEND(m, sizeof(struct sdl_hdr) , M_DONTWAIT); + if (m == 0) + return; + if ( !sdl_sethdrif(ifp, fh->fddi_shost, LLC_X25_LSAP, + fh->fddi_dhost, LLC_X25_LSAP, 6, + mtod(m, struct sdl_hdr *))) + panic("ETHER cons addr failure"); + mtod(m, struct sdl_hdr *)->sdlhdr_len = m->m_pkthdr.len - sizeof(struct sdl_hdr); +#ifdef LLC_DEBUG + printf("llc packet\n"); +#endif /* LLC_DEBUG */ + schednetisr(NETISR_CCITT); + inq = &llcintrq; + break; + } +#endif /* LLC */ + + default: + /* printf("fddi_input: unknown dsap 0x%x\n", l->llc_dsap); */ + ifp->if_noproto++; + dropanyway: + m_freem(m); + return; + } + + s = splimp(); + if (IF_QFULL(inq)) { + IF_DROP(inq); + m_freem(m); + } else + IF_ENQUEUE(inq, m); + splx(s); +} +/* + * Perform common duties while attaching to interface list + */ +void +fddi_ifattach(ifp) + register struct ifnet *ifp; +{ + register struct ifaddr *ifa; + register struct sockaddr_dl *sdl; + + ifp->if_type = IFT_FDDI; + ifp->if_addrlen = 6; + ifp->if_hdrlen = 21; + ifp->if_mtu = FDDIMTU; +#ifdef __NetBSD__ + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; + ifa = ifa->ifa_list.tqe_next) +#else + for (ifa = ifp->if_addrlist; ifa; ifa = ifa->ifa_next) +#endif + if ((sdl = (struct sockaddr_dl *)ifa->ifa_addr) && + sdl->sdl_family == AF_LINK) { + sdl->sdl_type = IFT_FDDI; + sdl->sdl_alen = ifp->if_addrlen; + bcopy((caddr_t)((struct arpcom *)ifp)->ac_enaddr, + LLADDR(sdl), ifp->if_addrlen); + break; + } +} diff --git a/sys/net/if_llc.h b/sys/net/if_llc.h new file mode 100644 index 00000000000..150c4249260 --- /dev/null +++ b/sys/net/if_llc.h @@ -0,0 +1,141 @@ +/* $NetBSD: if_llc.h,v 1.6 1995/03/08 02:56:57 cgd Exp $ */ + +/* + * Copyright (c) 1988, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_llc.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * IEEE 802.2 Link Level Control headers, for use in conjunction with + * 802.{3,4,5} media access control methods. + * + * Headers here do not use bit fields due to shortcommings in many + * compilers. + */ + +struct llc { + u_int8_t llc_dsap; + u_int8_t llc_ssap; + union { + struct { + u_int8_t control; + u_int8_t format_id; + u_int8_t class; + u_int8_t window_x2; + } type_u; + struct { + u_int8_t num_snd_x2; + u_int8_t num_rcv_x2; + } type_i; + struct { + u_int8_t control; + u_int8_t num_rcv_x2; + } type_s; + struct { + u_int8_t control; + struct frmrinfo { + u_int8_t rej_pdu_0; + u_int8_t rej_pdu_1; + u_int8_t frmr_control; + u_int8_t frmr_control_ext; + u_int8_t frmr_cause; + } frmrinfo; + } type_frmr; + struct { + u_int8_t control; + u_int8_t org_code[3]; + u_int16_t ether_type; + } type_snap; + struct { + u_int8_t control; + u_int8_t control_ext; + } type_raw; + } llc_un; +}; +#define llc_control llc_un.type_u.control +#define llc_control_ext llc_un.type_raw.control_ext +#define llc_fid llc_un.type_u.format_id +#define llc_class llc_un.type_u.class +#define llc_window llc_un.type_u.window_x2 +#define llc_frmrinfo llc_un.type_frmr.frmrinfo +#define llc_frmr_pdu0 llc_un.type_frmr.frmrinfo.rej_pdu0 +#define llc_frmr_pdu1 llc_un.type_frmr.frmrinfo.rej_pdu1 +#define llc_frmr_control llc_un.type_frmr.frmrinfo.frmr_control +#define llc_frmr_control_ext llc_un.type_frmr.frmrinfo.frmr_control_ext +#define llc_frmr_cause llc_un.type_frmr.frmrinfo.frmr_control_ext + +/* + * Don't use sizeof(struct llc_un) for LLC header sizes + */ +#define LLC_ISFRAMELEN 4 +#define LLC_UFRAMELEN 3 +#define LLC_FRMRLEN 7 + +/* + * Unnumbered LLC format commands + */ +#define LLC_UI 0x3 +#define LLC_UI_P 0x13 +#define LLC_DISC 0x43 +#define LLC_DISC_P 0x53 +#define LLC_UA 0x63 +#define LLC_UA_P 0x73 +#define LLC_TEST 0xe3 +#define LLC_TEST_P 0xf3 +#define LLC_FRMR 0x87 +#define LLC_FRMR_P 0x97 +#define LLC_DM 0x0f +#define LLC_DM_P 0x1f +#define LLC_XID 0xaf +#define LLC_XID_P 0xbf +#define LLC_SABME 0x6f +#define LLC_SABME_P 0x7f + +/* + * Supervisory LLC commands + */ +#define LLC_RR 0x01 +#define LLC_RNR 0x05 +#define LLC_REJ 0x09 + +/* + * Info format - dummy only + */ +#define LLC_INFO 0x00 + +/* + * ISO PDTR 10178 contains among others + */ +#define LLC_X25_LSAP 0x7e +#define LLC_SNAP_LSAP 0xaa +#define LLC_ISO_LSAP 0xfe diff --git a/sys/net/if_loop.c b/sys/net/if_loop.c new file mode 100644 index 00000000000..5c3db2e0059 --- /dev/null +++ b/sys/net/if_loop.c @@ -0,0 +1,255 @@ +/* $NetBSD: if_loop.c,v 1.14 1995/07/23 16:33:08 mycroft Exp $ */ + +/* + * 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_loop.c 8.1 (Berkeley) 6/10/93 + */ + +/* + * Loopback interface driver for protocol testing and timing. + */ + +#include "bpfilter.h" +#include "loop.h" + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/kernel.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/errno.h> +#include <sys/ioctl.h> +#include <sys/time.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/if_types.h> +#include <net/netisr.h> +#include <net/route.h> + +#ifdef INET +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/in_var.h> +#include <netinet/ip.h> +#endif + +#ifdef NS +#include <netns/ns.h> +#include <netns/ns_if.h> +#endif + +#ifdef ISO +#include <netiso/iso.h> +#include <netiso/iso_var.h> +#endif + +#if NBPFILTER > 0 +#include <net/bpf.h> +#endif + +#define LOMTU (32768) + +struct ifnet loif[NLOOP]; + +void +loopattach(n) + int n; +{ + register int i; + register struct ifnet *ifp; + + for (i = 0; i < NLOOP; i++) { + ifp = &loif[i]; + ifp->if_unit = i; + ifp->if_name = "lo"; + ifp->if_mtu = LOMTU; + ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST; + ifp->if_ioctl = loioctl; + ifp->if_output = looutput; + ifp->if_type = IFT_LOOP; + ifp->if_hdrlen = 0; + ifp->if_addrlen = 0; + if_attach(ifp); +#if NBPFILTER > 0 + bpfattach(&ifp->if_bpf, ifp, DLT_NULL, sizeof(u_int)); +#endif + } +} + +int +looutput(ifp, m, dst, rt) + struct ifnet *ifp; + register struct mbuf *m; + struct sockaddr *dst; + register struct rtentry *rt; +{ + int s, isr; + register struct ifqueue *ifq = 0; + + if ((m->m_flags & M_PKTHDR) == 0) + panic("looutput: no header mbuf"); + ifp->if_lastchange = time; +#if NBPFILTER > 0 + if (ifp->if_bpf) { + /* + * We need to prepend the address family as + * a four byte field. Cons up a dummy header + * to pacify bpf. This is safe because bpf + * will only read from the mbuf (i.e., it won't + * try to free it or keep a pointer to it). + */ + struct mbuf m0; + u_int af = dst->sa_family; + + m0.m_next = m; + m0.m_len = 4; + m0.m_data = (char *)⁡ + + bpf_mtap(ifp->if_bpf, &m0); + } +#endif + m->m_pkthdr.rcvif = ifp; + + if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) { + m_freem(m); + return (rt->rt_flags & RTF_BLACKHOLE ? 0 : + rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); + } + ifp->if_opackets++; + ifp->if_obytes += m->m_pkthdr.len; + switch (dst->sa_family) { + +#ifdef INET + case AF_INET: + ifq = &ipintrq; + isr = NETISR_IP; + break; +#endif +#ifdef NS + case AF_NS: + ifq = &nsintrq; + isr = NETISR_NS; + break; +#endif +#ifdef ISO + case AF_ISO: + ifq = &clnlintrq; + isr = NETISR_ISO; + break; +#endif + default: + printf("lo%d: can't handle af%d\n", ifp->if_unit, + dst->sa_family); + m_freem(m); + return (EAFNOSUPPORT); + } + s = splimp(); + if (IF_QFULL(ifq)) { + IF_DROP(ifq); + m_freem(m); + splx(s); + return (ENOBUFS); + } + IF_ENQUEUE(ifq, m); + schednetisr(isr); + ifp->if_ipackets++; + ifp->if_ibytes += m->m_pkthdr.len; + splx(s); + return (0); +} + +/* ARGSUSED */ +void +lortrequest(cmd, rt, sa) + int cmd; + struct rtentry *rt; + struct sockaddr *sa; +{ + + if (rt) + rt->rt_rmx.rmx_mtu = LOMTU; +} + +/* + * Process an ioctl request. + */ +/* ARGSUSED */ +int +loioctl(ifp, cmd, data) + register struct ifnet *ifp; + u_long cmd; + caddr_t data; +{ + register struct ifaddr *ifa; + register struct ifreq *ifr; + register int error = 0; + + switch (cmd) { + + case SIOCSIFADDR: + ifp->if_flags |= IFF_UP; + ifa = (struct ifaddr *)data; + if (ifa != 0 && ifa->ifa_addr->sa_family == AF_ISO) + ifa->ifa_rtrequest = lortrequest; + /* + * Everything else is done at a higher level. + */ + break; + + case SIOCADDMULTI: + case SIOCDELMULTI: + ifr = (struct ifreq *)data; + if (ifr == 0) { + error = EAFNOSUPPORT; /* XXX */ + break; + } + switch (ifr->ifr_addr.sa_family) { + +#ifdef INET + case AF_INET: + break; +#endif + + default: + error = EAFNOSUPPORT; + break; + } + break; + + default: + error = EINVAL; + } + return (error); +} diff --git a/sys/net/if_ppp.c b/sys/net/if_ppp.c new file mode 100644 index 00000000000..8eb431dee11 --- /dev/null +++ b/sys/net/if_ppp.c @@ -0,0 +1,1370 @@ +/* $NetBSD: if_ppp.c,v 1.24 1995/10/05 05:55:09 mycroft Exp $ */ + +/* + * if_ppp.c - Point-to-Point Protocol (PPP) Asynchronous driver. + * + * Copyright (c) 1989 Carnegie Mellon University. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by Carnegie Mellon University. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Drew D. Perkins + * Carnegie Mellon University + * 4910 Forbes Ave. + * Pittsburgh, PA 15213 + * (412) 268-8576 + * ddp@andrew.cmu.edu + * + * Based on: + * @(#)if_sl.c 7.6.1.2 (Berkeley) 2/15/89 + * + * Copyright (c) 1987 Regents of the University of California. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by the University of California, Berkeley. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Serial Line interface + * + * Rick Adams + * Center for Seismic Studies + * 1300 N 17th Street, Suite 1450 + * Arlington, Virginia 22209 + * (703)276-7900 + * rick@seismo.ARPA + * seismo!rick + * + * Pounded on heavily by Chris Torek (chris@mimsy.umd.edu, umcp-cs!chris). + * Converted to 4.3BSD Beta by Chris Torek. + * Other changes made at Berkeley, based in part on code by Kirk Smith. + * + * Converted to 4.3BSD+ 386BSD by Brad Parker (brad@cayman.com) + * Added VJ tcp header compression; more unified ioctls + * + * Extensively modified by Paul Mackerras (paulus@cs.anu.edu.au). + * Cleaned up a lot of the mbuf-related code to fix bugs that + * caused system crashes and packet corruption. Changed pppstart + * so that it doesn't just give up with a collision if the whole + * packet doesn't fit in the output ring buffer. + * + * Added priority queueing for interactive IP packets, following + * the model of if_sl.c, plus hooks for bpf. + * Paul Mackerras (paulus@cs.anu.edu.au). + */ + +/* from if_sl.c,v 1.11 84/10/04 12:54:47 rick Exp */ +/* from NetBSD: if_ppp.c,v 1.15.2.2 1994/07/28 05:17:58 cgd Exp */ + +#include "ppp.h" +#if NPPP > 0 + +#define VJC +#define PPP_COMPRESS + +#include <sys/param.h> +#include <sys/proc.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/kernel.h> + +#ifdef i386 +#include <machine/psl.h> +#endif + +#include <net/if.h> +#include <net/if_types.h> +#include <net/netisr.h> +#include <net/route.h> + +#if INET +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/in_var.h> +#include <netinet/ip.h> +#endif + +#include "bpfilter.h" +#if NBPFILTER > 0 +#include <sys/time.h> +#include <net/bpf.h> +#endif + +#ifdef VJC +#include <net/slcompress.h> +#endif + +#include <net/ppp_defs.h> +#include <net/if_ppp.h> +#include <net/if_pppvar.h> +#include <machine/cpu.h> + +#ifdef PPP_COMPRESS +#define PACKETPTR struct mbuf * +#include <net/ppp-comp.h> +#endif + +void pppattach __P((void)); +int pppioctl __P((struct ppp_softc *sc, u_long cmd, caddr_t data, int flag, + struct proc *)); +int pppoutput __P((struct ifnet *ifp, struct mbuf *m0, + struct sockaddr *dst, struct rtentry *rtp)); +int pppsioctl __P((struct ifnet *ifp, u_long cmd, caddr_t data)); +void pppintr __P((void)); + +static void ppp_requeue __P((struct ppp_softc *)); +static void ppp_outpkt __P((struct ppp_softc *)); +static int ppp_ccp __P((struct ppp_softc *, struct mbuf *m, int rcvd)); +static void ppp_ccp_closed __P((struct ppp_softc *)); +static void ppp_inproc __P((struct ppp_softc *, struct mbuf *)); +static void pppdumpm __P((struct mbuf *m0)); + +/* + * Some useful mbuf macros not in mbuf.h. + */ +#define M_IS_CLUSTER(m) ((m)->m_flags & M_EXT) + +#define M_DATASTART(m) \ + (M_IS_CLUSTER(m) ? (m)->m_ext.ext_buf : \ + (m)->m_flags & M_PKTHDR ? (m)->m_pktdat : (m)->m_dat) + +#define M_DATASIZE(m) \ + (M_IS_CLUSTER(m) ? (m)->m_ext.ext_size : \ + (m)->m_flags & M_PKTHDR ? MHLEN: MLEN) + +/* + * We steal two bits in the mbuf m_flags, to mark high-priority packets + * for output, and received packets following lost/corrupted packets. + */ +#define M_HIGHPRI 0x2000 /* output packet for sc_fastq */ +#define M_ERRMARK 0x4000 /* steal a bit in mbuf m_flags */ + + +#ifdef PPP_COMPRESS +/* + * List of compressors we know about. + * We leave some space so maybe we can modload compressors. + */ + +extern struct compressor ppp_bsd_compress; + +struct compressor *ppp_compressors[8] = { +#if DO_BSD_COMPRESS + &ppp_bsd_compress, +#endif + NULL +}; +#endif /* PPP_COMPRESS */ + +/* + * Called from boot code to establish ppp interfaces. + */ +void +pppattach() +{ + register struct ppp_softc *sc; + register int i = 0; + + for (sc = ppp_softc; i < NPPP; sc++) { + sc->sc_if.if_name = "ppp"; + sc->sc_if.if_unit = i++; + sc->sc_if.if_mtu = PPP_MTU; + sc->sc_if.if_flags = IFF_POINTOPOINT; + sc->sc_if.if_type = IFT_PPP; + sc->sc_if.if_hdrlen = PPP_HDRLEN; + sc->sc_if.if_ioctl = pppsioctl; + sc->sc_if.if_output = pppoutput; + sc->sc_if.if_snd.ifq_maxlen = IFQ_MAXLEN; + sc->sc_inq.ifq_maxlen = IFQ_MAXLEN; + sc->sc_fastq.ifq_maxlen = IFQ_MAXLEN; + sc->sc_rawq.ifq_maxlen = IFQ_MAXLEN; + if_attach(&sc->sc_if); +#if NBPFILTER > 0 + bpfattach(&sc->sc_bpf, &sc->sc_if, DLT_PPP, PPP_HDRLEN); +#endif + } + +#ifdef i386 + /* + * XXX kludge to fix the bug in the i386 interrupt handling code, + * where software interrupts could be taken while hardware + * interrupts were blocked. + */ + if ((imask[IPL_TTY] & SIR_NETMASK) == 0) { + imask[IPL_TTY] |= SIR_NETMASK; + intr_calculatemasks(); + } +#endif +} + +/* + * Allocate a ppp interface unit and initialize it. + */ +struct ppp_softc * +pppalloc(pid) + pid_t pid; +{ + int nppp, i; + struct ppp_softc *sc; + + for (nppp = 0, sc = ppp_softc; nppp < NPPP; nppp++, sc++) + if (sc->sc_xfer == pid) { + sc->sc_xfer = 0; + return sc; + } + for (nppp = 0, sc = ppp_softc; nppp < NPPP; nppp++, sc++) + if (sc->sc_devp == NULL) + break; + if (nppp >= NPPP) + return NULL; + + sc->sc_flags = 0; + sc->sc_mru = PPP_MRU; + sc->sc_relinq = NULL; +#ifdef VJC + sl_compress_init(&sc->sc_comp, -1); +#endif +#ifdef PPP_COMPRESS + sc->sc_xc_state = NULL; + sc->sc_rc_state = NULL; +#endif /* PPP_COMPRESS */ + for (i = 0; i < NUM_NP; ++i) + sc->sc_npmode[i] = NPMODE_ERROR; + sc->sc_npqueue = NULL; + sc->sc_npqtail = &sc->sc_npqueue; + sc->sc_last_sent = sc->sc_last_recv = time.tv_sec; + + return sc; +} + +/* + * Deallocate a ppp unit. Must be called at splsoftnet or higher. + */ +void +pppdealloc(sc) + struct ppp_softc *sc; +{ + struct mbuf *m; + + if_down(&sc->sc_if); + sc->sc_if.if_flags &= ~(IFF_UP|IFF_RUNNING); + sc->sc_devp = NULL; + sc->sc_xfer = 0; + for (;;) { + IF_DEQUEUE(&sc->sc_rawq, m); + if (m == NULL) + break; + m_freem(m); + } + for (;;) { + IF_DEQUEUE(&sc->sc_inq, m); + if (m == NULL) + break; + m_freem(m); + } + for (;;) { + IF_DEQUEUE(&sc->sc_fastq, m); + if (m == NULL) + break; + m_freem(m); + } + while ((m = sc->sc_npqueue) != NULL) { + sc->sc_npqueue = m->m_nextpkt; + m_freem(m); + } + if (sc->sc_togo != NULL) { + m_freem(sc->sc_togo); + sc->sc_togo = NULL; + } +#ifdef PPP_COMPRESS + ppp_ccp_closed(sc); + sc->sc_xc_state = NULL; + sc->sc_rc_state = NULL; +#endif /* PPP_COMPRESS */ +} + +/* + * Ioctl routine for generic ppp devices. + */ +int +pppioctl(sc, cmd, data, flag, p) + struct ppp_softc *sc; + u_long cmd; + caddr_t data; + int flag; + struct proc *p; +{ + int s, error, flags, mru, nb, npx; + struct ppp_option_data *odp; + struct compressor **cp; + struct npioctl *npi; + time_t t; +#ifdef PPP_COMPRESS + u_char ccp_option[CCP_MAX_OPTION_LENGTH]; +#endif + + switch (cmd) { + case FIONREAD: + *(int *)data = sc->sc_inq.ifq_len; + break; + + case PPPIOCGUNIT: + *(int *)data = sc->sc_if.if_unit; + break; + + case PPPIOCGFLAGS: + *(u_int *)data = sc->sc_flags; + break; + + case PPPIOCSFLAGS: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + flags = *(int *)data & SC_MASK; + s = splsoftnet(); +#ifdef PPP_COMPRESS + if (sc->sc_flags & SC_CCP_OPEN && !(flags & SC_CCP_OPEN)) + ppp_ccp_closed(sc); +#endif + splhigh(); + sc->sc_flags = (sc->sc_flags & ~SC_MASK) | flags; + splx(s); + break; + + case PPPIOCSMRU: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + mru = *(int *)data; + if (mru >= PPP_MRU && mru <= PPP_MAXMRU) + sc->sc_mru = mru; + break; + + case PPPIOCGMRU: + *(int *)data = sc->sc_mru; + break; + +#ifdef VJC + case PPPIOCSMAXCID: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + s = splsoftnet(); + sl_compress_init(&sc->sc_comp, *(int *)data); + splx(s); + break; +#endif + + case PPPIOCXFERUNIT: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + sc->sc_xfer = p->p_pid; + break; + +#ifdef PPP_COMPRESS + case PPPIOCSCOMPRESS: + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + odp = (struct ppp_option_data *) data; + nb = odp->length; + if (nb > sizeof(ccp_option)) + nb = sizeof(ccp_option); + if (error = copyin(odp->ptr, ccp_option, nb)) + return (error); + if (ccp_option[1] < 2) /* preliminary check on the length byte */ + return (EINVAL); + for (cp = ppp_compressors; *cp != NULL; ++cp) + if ((*cp)->compress_proto == ccp_option[0]) { + /* + * Found a handler for the protocol - try to allocate + * a compressor or decompressor. + */ + error = 0; + if (odp->transmit) { + s = splsoftnet(); + if (sc->sc_xc_state != NULL) + (*sc->sc_xcomp->comp_free)(sc->sc_xc_state); + sc->sc_xcomp = *cp; + sc->sc_xc_state = (*cp)->comp_alloc(ccp_option, nb); + if (sc->sc_xc_state == NULL) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: comp_alloc failed\n", + sc->sc_if.if_unit); + error = ENOBUFS; + } + splhigh(); + sc->sc_flags &= ~SC_COMP_RUN; + splx(s); + } else { + s = splsoftnet(); + if (sc->sc_rc_state != NULL) + (*sc->sc_rcomp->decomp_free)(sc->sc_rc_state); + sc->sc_rcomp = *cp; + sc->sc_rc_state = (*cp)->decomp_alloc(ccp_option, nb); + if (sc->sc_rc_state == NULL) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: decomp_alloc failed\n", + sc->sc_if.if_unit); + error = ENOBUFS; + } + splhigh(); + sc->sc_flags &= ~SC_DECOMP_RUN; + splx(s); + } + return (error); + } + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: no compressor for [%x %x %x], %x\n", + sc->sc_if.if_unit, ccp_option[0], ccp_option[1], + ccp_option[2], nb); + return (EINVAL); /* no handler found */ +#endif /* PPP_COMPRESS */ + + case PPPIOCGNPMODE: + case PPPIOCSNPMODE: + npi = (struct npioctl *) data; + switch (npi->protocol) { + case PPP_IP: + npx = NP_IP; + break; + default: + return EINVAL; + } + if (cmd == PPPIOCGNPMODE) { + npi->mode = sc->sc_npmode[npx]; + } else { + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + if (npi->mode != sc->sc_npmode[npx]) { + s = splsoftnet(); + sc->sc_npmode[npx] = npi->mode; + if (npi->mode != NPMODE_QUEUE) { + ppp_requeue(sc); + (*sc->sc_start)(sc); + } + splx(s); + } + } + break; + + case PPPIOCGIDLE: + s = splsoftnet(); + t = time.tv_sec; + ((struct ppp_idle *)data)->xmit_idle = t - sc->sc_last_sent; + ((struct ppp_idle *)data)->recv_idle = t - sc->sc_last_recv; + splx(s); + break; + + default: + return (-1); + } + return (0); +} + +/* + * Process an ioctl request to the ppp network interface. + */ +int +pppsioctl(ifp, cmd, data) + register struct ifnet *ifp; + u_long cmd; + caddr_t data; +{ + struct proc *p = curproc; /* XXX */ + register struct ppp_softc *sc = &ppp_softc[ifp->if_unit]; + register struct ifaddr *ifa = (struct ifaddr *)data; + register struct ifreq *ifr = (struct ifreq *)data; + struct ppp_stats *psp; +#ifdef PPP_COMPRESS + struct ppp_comp_stats *pcp; +#endif + int s = splimp(), error = 0; + + switch (cmd) { + case SIOCSIFFLAGS: + if ((ifp->if_flags & IFF_RUNNING) == 0) + ifp->if_flags &= ~IFF_UP; + break; + + case SIOCSIFADDR: + if (ifa->ifa_addr->sa_family != AF_INET) + error = EAFNOSUPPORT; + break; + + case SIOCSIFDSTADDR: + if (ifa->ifa_addr->sa_family != AF_INET) + error = EAFNOSUPPORT; + break; + + case SIOCSIFMTU: + if (error = suser(p->p_ucred, &p->p_acflag)) + break; + sc->sc_if.if_mtu = ifr->ifr_mtu; + break; + + case SIOCGIFMTU: + ifr->ifr_mtu = sc->sc_if.if_mtu; + break; + + case SIOCGPPPSTATS: + psp = &((struct ifpppstatsreq *) data)->stats; + bzero(psp, sizeof(*psp)); + psp->p.ppp_ibytes = sc->sc_bytesrcvd; + psp->p.ppp_ipackets = ifp->if_ipackets; + psp->p.ppp_ierrors = ifp->if_ierrors; + psp->p.ppp_obytes = sc->sc_bytessent; + psp->p.ppp_opackets = ifp->if_opackets; + psp->p.ppp_oerrors = ifp->if_oerrors; +#ifdef VJC + psp->vj.vjs_packets = sc->sc_comp.sls_packets; + psp->vj.vjs_compressed = sc->sc_comp.sls_compressed; + psp->vj.vjs_searches = sc->sc_comp.sls_searches; + psp->vj.vjs_misses = sc->sc_comp.sls_misses; + psp->vj.vjs_uncompressedin = sc->sc_comp.sls_uncompressedin; + psp->vj.vjs_compressedin = sc->sc_comp.sls_compressedin; + psp->vj.vjs_errorin = sc->sc_comp.sls_errorin; + psp->vj.vjs_tossed = sc->sc_comp.sls_tossed; +#endif /* VJC */ + break; + +#ifdef PPP_COMPRESS + case SIOCGPPPCSTATS: + pcp = &((struct ifpppcstatsreq *) data)->stats; + bzero(pcp, sizeof(*pcp)); + if (sc->sc_xc_state != NULL) + (*sc->sc_xcomp->comp_stat)(sc->sc_xc_state, &pcp->c); + if (sc->sc_rc_state != NULL) + (*sc->sc_rcomp->decomp_stat)(sc->sc_rc_state, &pcp->d); + break; +#endif /* PPP_COMPRESS */ + + default: + error = EINVAL; + } + splx(s); + return (error); +} + +/* + * Queue a packet. Start transmission if not active. + * Packet is placed in Information field of PPP frame. + */ +int +pppoutput(ifp, m0, dst, rtp) + struct ifnet *ifp; + struct mbuf *m0; + struct sockaddr *dst; + struct rtentry *rtp; +{ + register struct ppp_softc *sc = &ppp_softc[ifp->if_unit]; + struct ppp_header *ph; + int protocol, address, control; + u_char *cp; + int s, error; + struct ip *ip; + struct ifqueue *ifq; + enum NPmode mode; + + if (sc->sc_devp == NULL || (ifp->if_flags & IFF_RUNNING) == 0 + || (ifp->if_flags & IFF_UP) == 0 && dst->sa_family != AF_UNSPEC) { + error = ENETDOWN; /* sort of */ + goto bad; + } + + /* + * Compute PPP header. + */ + m0->m_flags &= ~M_HIGHPRI; + switch (dst->sa_family) { +#ifdef INET + case AF_INET: + address = PPP_ALLSTATIONS; + control = PPP_UI; + protocol = PPP_IP; + mode = sc->sc_npmode[NP_IP]; + + /* + * If this packet has the "low delay" bit set in the IP header, + * put it on the fastq instead. + */ + ip = mtod(m0, struct ip *); + if (ip->ip_tos & IPTOS_LOWDELAY) + m0->m_flags |= M_HIGHPRI; + break; +#endif + case AF_UNSPEC: + address = PPP_ADDRESS(dst->sa_data); + control = PPP_CONTROL(dst->sa_data); + protocol = PPP_PROTOCOL(dst->sa_data); + mode = NPMODE_PASS; + break; + default: + printf("ppp%d: af%d not supported\n", ifp->if_unit, dst->sa_family); + error = EAFNOSUPPORT; + goto bad; + } + + /* + * Drop this packet, or return an error, if necessary. + */ + if (mode == NPMODE_ERROR) { + error = ENETDOWN; + goto bad; + } + if (mode == NPMODE_DROP) { + error = 0; + goto bad; + } + + /* + * Add PPP header. If no space in first mbuf, allocate another. + * (This assumes M_LEADINGSPACE is always 0 for a cluster mbuf.) + */ + if (M_LEADINGSPACE(m0) < PPP_HDRLEN) { + m0 = m_prepend(m0, PPP_HDRLEN, M_DONTWAIT); + if (m0 == 0) { + error = ENOBUFS; + goto bad; + } + m0->m_len = 0; + } else + m0->m_data -= PPP_HDRLEN; + + cp = mtod(m0, u_char *); + *cp++ = address; + *cp++ = control; + *cp++ = protocol >> 8; + *cp++ = protocol & 0xff; + m0->m_len += PPP_HDRLEN; + + if (sc->sc_flags & SC_LOG_OUTPKT) { + printf("ppp%d output: ", ifp->if_unit); + pppdumpm(m0); + } + +#if NBPFILTER > 0 + /* + * See if bpf wants to look at the packet. + */ + if (sc->sc_bpf) + bpf_mtap(sc->sc_bpf, m0); +#endif + + /* + * Put the packet on the appropriate queue. + */ + s = splsoftnet(); + if (mode == NPMODE_QUEUE) { + /* XXX we should limit the number of packets on this queue */ + *sc->sc_npqtail = m0; + m0->m_nextpkt = NULL; + sc->sc_npqtail = &m0->m_nextpkt; + } else { + ifq = (m0->m_flags & M_HIGHPRI)? &sc->sc_fastq: &ifp->if_snd; + if (IF_QFULL(ifq)) { + IF_DROP(ifq); + splx(s); + sc->sc_if.if_oerrors++; + error = ENOBUFS; + goto bad; + } + IF_ENQUEUE(ifq, m0); + (*sc->sc_start)(sc); + } + ifp->if_lastchange = time; + + splx(s); + return (0); + +bad: + m_freem(m0); + return (error); +} + +/* + * After a change in the NPmode for some NP, move packets from the + * npqueue to the send queue or the fast queue as appropriate. + * Should be called at splsoftnet. + */ +static void +ppp_requeue(sc) + struct ppp_softc *sc; +{ + struct mbuf *m, **mpp; + struct ifqueue *ifq; + enum NPmode mode; + + for (mpp = &sc->sc_npqueue; (m = *mpp) != NULL; ) { + switch (PPP_PROTOCOL(mtod(m, u_char *))) { + case PPP_IP: + mode = sc->sc_npmode[NP_IP]; + break; + default: + mode = NPMODE_PASS; + } + + switch (mode) { + case NPMODE_PASS: + /* + * This packet can now go on one of the queues to be sent. + */ + *mpp = m->m_nextpkt; + m->m_nextpkt = NULL; + ifq = (m->m_flags & M_HIGHPRI)? &sc->sc_fastq: &sc->sc_if.if_snd; + if (IF_QFULL(ifq)) { + IF_DROP(ifq); + sc->sc_if.if_oerrors++; + } else + IF_ENQUEUE(ifq, m); + break; + + case NPMODE_DROP: + case NPMODE_ERROR: + *mpp = m->m_nextpkt; + m_freem(m); + break; + + case NPMODE_QUEUE: + mpp = &m->m_nextpkt; + break; + } + } + sc->sc_npqtail = mpp; +} + +/* + * Get a packet to send. This procedure is intended to be called at + * spltty or splimp, so it takes little time. If there isn't a packet + * waiting to go out, it schedules a software interrupt to prepare a + * new packet; the device start routine gets called again when a + * packet is ready. + */ +struct mbuf * +ppp_dequeue(sc) + struct ppp_softc *sc; +{ + struct mbuf *m; + int s = splhigh(); + + m = sc->sc_togo; + if (m) { + /* + * Had a packet waiting - send it. + */ + sc->sc_togo = NULL; + sc->sc_flags |= SC_TBUSY; + splx(s); + return m; + } + /* + * Remember we wanted a packet and schedule a software interrupt. + */ + sc->sc_flags &= ~SC_TBUSY; + schednetisr(NETISR_PPP); + splx(s); + return NULL; +} + +/* + * Software interrupt routine, called at splsoftnet. + */ +void +pppintr() +{ + struct ppp_softc *sc; + int i, s; + struct mbuf *m; + + sc = ppp_softc; + for (i = 0; i < NPPP; ++i, ++sc) { + if (!(sc->sc_flags & SC_TBUSY) && sc->sc_togo == NULL + && (sc->sc_if.if_snd.ifq_head || sc->sc_fastq.ifq_head)) + ppp_outpkt(sc); + for (;;) { + s = splhigh(); + IF_DEQUEUE(&sc->sc_rawq, m); + splx(s); + if (m == NULL) + break; + ppp_inproc(sc, m); + } + } +} + +/* + * Grab another packet off a queue and apply VJ compression, + * packet compression, address/control and/or protocol compression + * if enabled. Should be called at splsoftnet. + */ +static void +ppp_outpkt(sc) + struct ppp_softc *sc; +{ + int s; + struct mbuf *m, *mp; + u_char *cp; + int address, control, protocol; + enum NPmode mode; + + /* + * Grab a packet to send: first try the fast queue, then the + * normal queue. + */ + IF_DEQUEUE(&sc->sc_fastq, m); + if (m == NULL) + IF_DEQUEUE(&sc->sc_if.if_snd, m); + if (m == NULL) + return; + + /* + * Extract the ppp header of the new packet. + * The ppp header will be in one mbuf. + */ + cp = mtod(m, u_char *); + address = PPP_ADDRESS(cp); + control = PPP_CONTROL(cp); + protocol = PPP_PROTOCOL(cp); + + switch (protocol) { + case PPP_IP: + /* + * Update the time we sent the most recent packet. + */ + sc->sc_last_sent = time.tv_sec; + +#ifdef VJC + /* + * If the packet is a TCP/IP packet, see if we can compress it. + */ + if (sc->sc_flags & SC_COMP_TCP) { + struct ip *ip; + int type; + + mp = m; + ip = (struct ip *) (cp + PPP_HDRLEN); + if (mp->m_len <= PPP_HDRLEN) { + mp = mp->m_next; + if (mp == NULL) + break; + ip = mtod(mp, struct ip *); + } + /* this code assumes the IP/TCP header is in one non-shared mbuf */ + if (ip->ip_p == IPPROTO_TCP) { + type = sl_compress_tcp(mp, ip, &sc->sc_comp, + !(sc->sc_flags & SC_NO_TCP_CCID)); + switch (type) { + case TYPE_UNCOMPRESSED_TCP: + protocol = PPP_VJC_UNCOMP; + break; + case TYPE_COMPRESSED_TCP: + protocol = PPP_VJC_COMP; + cp = mtod(m, u_char *); + cp[0] = address; /* header has moved */ + cp[1] = control; + cp[2] = 0; + break; + } + cp[3] = protocol; /* update protocol in PPP header */ + } + } +#endif /* VJC */ + break; + +#ifdef PPP_COMPRESS + case PPP_CCP: + ppp_ccp(sc, m, 0); + break; +#endif /* PPP_COMPRESS */ + } + +#ifdef PPP_COMPRESS + if (protocol != PPP_LCP && protocol != PPP_CCP + && sc->sc_xc_state && (sc->sc_flags & SC_COMP_RUN)) { + struct mbuf *mcomp; + int slen, clen; + + slen = 0; + for (mp = m; mp != NULL; mp = mp->m_next) + slen += mp->m_len; + clen = (*sc->sc_xcomp->compress) + (sc->sc_xc_state, &mcomp, m, slen, + (sc->sc_flags & SC_CCP_UP? sc->sc_if.if_mtu: 0)); + if (mcomp != NULL) { + m_freem(m); + m = mcomp; + cp = mtod(m, u_char *); + protocol = cp[3]; + } + } +#endif /* PPP_COMPRESS */ + + /* + * Compress the address/control and protocol, if possible. + */ + if (sc->sc_flags & SC_COMP_AC && address == PPP_ALLSTATIONS && + control == PPP_UI && protocol != PPP_ALLSTATIONS && + protocol != PPP_LCP) { + /* can compress address/control */ + m->m_data += 2; + m->m_len -= 2; + } + if (sc->sc_flags & SC_COMP_PROT && protocol < 0xFF) { + /* can compress protocol */ + if (mtod(m, u_char *) == cp) { + cp[2] = cp[1]; /* move address/control up */ + cp[1] = cp[0]; + } + ++m->m_data; + --m->m_len; + } + + sc->sc_togo = m; + (*sc->sc_start)(sc); +} + +#ifdef PPP_COMPRESS +/* + * Handle a CCP packet. `rcvd' is 1 if the packet was received, + * 0 if it is about to be transmitted. + */ +static int +ppp_ccp(sc, m, rcvd) + struct ppp_softc *sc; + struct mbuf *m; + int rcvd; +{ + u_char *dp, *ep; + struct mbuf *mp; + int slen, s; + + /* + * Get a pointer to the data after the PPP header. + */ + if (m->m_len <= PPP_HDRLEN) { + mp = m->m_next; + if (mp == NULL) + return; + dp = (mp != NULL)? mtod(mp, u_char *): NULL; + } else { + mp = m; + dp = mtod(mp, u_char *) + PPP_HDRLEN; + } + + ep = mtod(mp, u_char *) + mp->m_len; + if (dp + CCP_HDRLEN > ep) + return; + slen = CCP_LENGTH(dp); + if (dp + slen > ep) { + if (sc->sc_flags & SC_DEBUG) + printf("if_ppp/ccp: not enough data in mbuf (%x+%x > %x+%x)\n", + dp, slen, mtod(mp, u_char *), mp->m_len); + return; + } + + switch (CCP_CODE(dp)) { + case CCP_CONFREQ: + case CCP_TERMREQ: + case CCP_TERMACK: + /* CCP must be going down - disable compression */ + if (sc->sc_flags & SC_CCP_UP) { + s = splhigh(); + sc->sc_flags &= ~(SC_CCP_UP | SC_COMP_RUN | SC_DECOMP_RUN); + splx(s); + } + break; + + case CCP_CONFACK: + if (sc->sc_flags & SC_CCP_OPEN && !(sc->sc_flags & SC_CCP_UP) + && slen >= CCP_HDRLEN + CCP_OPT_MINLEN + && slen >= CCP_OPT_LENGTH(dp + CCP_HDRLEN) + CCP_HDRLEN) { + if (!rcvd) { + /* we're agreeing to send compressed packets. */ + if (sc->sc_xc_state != NULL + && (*sc->sc_xcomp->comp_init) + (sc->sc_xc_state, dp + CCP_HDRLEN, slen - CCP_HDRLEN, + sc->sc_if.if_unit, 0, sc->sc_flags & SC_DEBUG)) { + s = splhigh(); + sc->sc_flags |= SC_COMP_RUN; + splx(s); + } + } else { + /* peer is agreeing to send compressed packets. */ + if (sc->sc_rc_state != NULL + && (*sc->sc_rcomp->decomp_init) + (sc->sc_rc_state, dp + CCP_HDRLEN, slen - CCP_HDRLEN, + sc->sc_if.if_unit, 0, sc->sc_mru, + sc->sc_flags & SC_DEBUG)) { + s = splhigh(); + sc->sc_flags |= SC_DECOMP_RUN; + sc->sc_flags &= ~(SC_DC_ERROR | SC_DC_FERROR); + splx(s); + } + } + } + break; + + case CCP_RESETACK: + if (sc->sc_flags & SC_CCP_UP) { + if (!rcvd) { + if (sc->sc_xc_state && (sc->sc_flags & SC_COMP_RUN)) + (*sc->sc_xcomp->comp_reset)(sc->sc_xc_state); + } else { + if (sc->sc_rc_state && (sc->sc_flags & SC_DECOMP_RUN)) { + (*sc->sc_rcomp->decomp_reset)(sc->sc_rc_state); + s = splhigh(); + sc->sc_flags &= ~SC_DC_ERROR; + splx(s); + } + } + } + break; + } +} + +/* + * CCP is down; free (de)compressor state if necessary. + */ +static void +ppp_ccp_closed(sc) + struct ppp_softc *sc; +{ + if (sc->sc_xc_state) { + (*sc->sc_xcomp->comp_free)(sc->sc_xc_state); + sc->sc_xc_state = NULL; + } + if (sc->sc_rc_state) { + (*sc->sc_rcomp->decomp_free)(sc->sc_rc_state); + sc->sc_rc_state = NULL; + } +} +#endif /* PPP_COMPRESS */ + +/* + * PPP packet input routine. + * The caller has checked and removed the FCS and has inserted + * the address/control bytes and the protocol high byte if they + * were omitted. + */ +void +ppppktin(sc, m, lost) + struct ppp_softc *sc; + struct mbuf *m; + int lost; +{ + int s = splhigh(); + + if (lost) + m->m_flags |= M_ERRMARK; + IF_ENQUEUE(&sc->sc_rawq, m); + schednetisr(NETISR_PPP); + splx(s); +} + +/* + * Process a received PPP packet, doing decompression as necessary. + * Should be called at splsoftnet. + */ +#define COMPTYPE(proto) ((proto) == PPP_VJC_COMP? TYPE_COMPRESSED_TCP: \ + TYPE_UNCOMPRESSED_TCP) + +static void +ppp_inproc(sc, m) + struct ppp_softc *sc; + struct mbuf *m; +{ + struct ifnet *ifp = &sc->sc_if; + struct ifqueue *inq; + int s, ilen, xlen, proto, rv; + u_char *cp, adrs, ctrl; + struct mbuf *mp, *dmp; + u_char *iphdr; + u_int hlen; + + ifp->if_ipackets++; + ifp->if_lastchange = time; + + if (sc->sc_flags & SC_LOG_INPKT) { + printf("ppp%d: got %d bytes\n", ifp->if_unit, ilen); + pppdumpm(m); + } + + cp = mtod(m, u_char *); + adrs = PPP_ADDRESS(cp); + ctrl = PPP_CONTROL(cp); + proto = PPP_PROTOCOL(cp); + + if (m->m_flags & M_ERRMARK) { + m->m_flags &= ~M_ERRMARK; + s = splhigh(); + sc->sc_flags |= SC_VJ_RESET; + splx(s); + } + +#ifdef PPP_COMPRESS + /* + * Decompress this packet if necessary, update the receiver's + * dictionary, or take appropriate action on a CCP packet. + */ + if (proto == PPP_COMP && sc->sc_rc_state && (sc->sc_flags & SC_DECOMP_RUN) + && !(sc->sc_flags & SC_DC_ERROR) && !(sc->sc_flags & SC_DC_FERROR)) { + /* decompress this packet */ + rv = (*sc->sc_rcomp->decompress)(sc->sc_rc_state, m, &dmp); + if (rv == DECOMP_OK) { + m_freem(m); + if (dmp == NULL) { + /* no error, but no decompressed packet produced */ + return; + } + m = dmp; + cp = mtod(m, u_char *); + proto = PPP_PROTOCOL(cp); + + } else { + /* + * An error has occurred in decompression. + * Pass the compressed packet up to pppd, which may take + * CCP down or issue a Reset-Req. + */ + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: decompress failed %d\n", ifp->if_unit, rv); + s = splhigh(); + sc->sc_flags |= SC_VJ_RESET; + if (rv == DECOMP_ERROR) + sc->sc_flags |= SC_DC_ERROR; + else + sc->sc_flags |= SC_DC_FERROR; + splx(s); + } + + } else { + if (sc->sc_rc_state && (sc->sc_flags & SC_DECOMP_RUN)) { + (*sc->sc_rcomp->incomp)(sc->sc_rc_state, m); + } + if (proto == PPP_CCP) { + ppp_ccp(sc, m, 1); + } + } +#endif + + ilen = 0; + for (mp = m; mp != NULL; mp = mp->m_next) + ilen += mp->m_len; + +#ifdef VJC + if (sc->sc_flags & SC_VJ_RESET) { + /* + * If we've missed a packet, we must toss subsequent compressed + * packets which don't have an explicit connection ID. + */ + sl_uncompress_tcp(NULL, 0, TYPE_ERROR, &sc->sc_comp); + s = splhigh(); + sc->sc_flags &= ~SC_VJ_RESET; + splx(s); + } + + /* + * See if we have a VJ-compressed packet to uncompress. + */ + if (proto == PPP_VJC_COMP) { + if (sc->sc_flags & SC_REJ_COMP_TCP) + goto bad; + + xlen = sl_uncompress_tcp_core(cp + PPP_HDRLEN, m->m_len - PPP_HDRLEN, + ilen - PPP_HDRLEN, TYPE_COMPRESSED_TCP, + &sc->sc_comp, &iphdr, &hlen); + + if (xlen <= 0) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: VJ uncompress failed on type comp\n", + ifp->if_unit); + goto bad; + } + + /* Copy the PPP and IP headers into a new mbuf. */ + MGETHDR(mp, M_DONTWAIT, MT_DATA); + if (mp == NULL) + goto bad; + mp->m_len = 0; + mp->m_next = NULL; + if (hlen + PPP_HDRLEN > MHLEN) { + MCLGET(mp, M_DONTWAIT); + if (M_TRAILINGSPACE(mp) < hlen + PPP_HDRLEN) { + m_freem(mp); + goto bad; /* lose if big headers and no clusters */ + } + } + cp = mtod(mp, u_char *); + cp[0] = adrs; + cp[1] = ctrl; + cp[2] = 0; + cp[3] = PPP_IP; + proto = PPP_IP; + bcopy(iphdr, cp + PPP_HDRLEN, hlen); + mp->m_len = hlen + PPP_HDRLEN; + + /* + * Trim the PPP and VJ headers off the old mbuf + * and stick the new and old mbufs together. + */ + m->m_data += PPP_HDRLEN + xlen; + m->m_len -= PPP_HDRLEN + xlen; + if (m->m_len <= M_TRAILINGSPACE(mp)) { + bcopy(mtod(m, u_char *), mtod(mp, u_char *) + mp->m_len, m->m_len); + mp->m_len += m->m_len; + MFREE(m, mp->m_next); + } else + mp->m_next = m; + m = mp; + ilen += hlen - xlen; + + } else if (proto == PPP_VJC_UNCOMP) { + if (sc->sc_flags & SC_REJ_COMP_TCP) + goto bad; + + xlen = sl_uncompress_tcp_core(cp + PPP_HDRLEN, m->m_len - PPP_HDRLEN, + ilen - PPP_HDRLEN, TYPE_UNCOMPRESSED_TCP, + &sc->sc_comp, &iphdr, &hlen); + + if (xlen < 0) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: VJ uncompress failed on type uncomp\n", + ifp->if_unit); + goto bad; + } + + proto = PPP_IP; + cp[3] = PPP_IP; + } +#endif /* VJC */ + + /* + * If the packet will fit in a header mbuf, don't waste a + * whole cluster on it. + */ + if (ilen <= MHLEN && M_IS_CLUSTER(m)) { + MGETHDR(mp, M_DONTWAIT, MT_DATA); + if (mp != NULL) { + m_copydata(m, 0, ilen, mtod(mp, caddr_t)); + m_freem(m); + m = mp; + m->m_len = ilen; + } + } + m->m_pkthdr.len = ilen; + m->m_pkthdr.rcvif = ifp; + +#if NBPFILTER > 0 + /* See if bpf wants to look at the packet. */ + if (sc->sc_bpf) + bpf_mtap(sc->sc_bpf, m); +#endif + + rv = 0; + switch (proto) { +#ifdef INET + case PPP_IP: + /* + * IP packet - take off the ppp header and pass it up to IP. + */ + if ((ifp->if_flags & IFF_UP) == 0 + || sc->sc_npmode[NP_IP] != NPMODE_PASS) { + /* interface is down - drop the packet. */ + m_freem(m); + return; + } + m->m_pkthdr.len -= PPP_HDRLEN; + m->m_data += PPP_HDRLEN; + m->m_len -= PPP_HDRLEN; + schednetisr(NETISR_IP); + inq = &ipintrq; + sc->sc_last_recv = time.tv_sec; /* update time of last pkt rcvd */ + break; +#endif + + default: + /* + * Some other protocol - place on input queue for read(). + */ + inq = &sc->sc_inq; + rv = 1; + break; + } + + /* + * Put the packet on the appropriate input queue. + */ + s = splhigh(); + if (IF_QFULL(inq)) { + IF_DROP(inq); + splx(s); + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: input queue full\n", ifp->if_unit); + ifp->if_iqdrops++; + goto bad; + } + IF_ENQUEUE(inq, m); + splx(s); + + if (rv) + (*sc->sc_ctlp)(sc); + + return; + + bad: + m_freem(m); + sc->sc_if.if_ierrors++; +} + +#define MAX_DUMP_BYTES 128 + +static void +pppdumpm(m0) + struct mbuf *m0; +{ + char buf[3*MAX_DUMP_BYTES+4]; + char *bp = buf; + struct mbuf *m; + static char digits[] = "0123456789abcdef"; + + for (m = m0; m; m = m->m_next) { + int l = m->m_len; + u_char *rptr = (u_char *)m->m_data; + + while (l--) { + if (bp > buf + sizeof(buf) - 4) + goto done; + *bp++ = digits[*rptr >> 4]; /* convert byte to ascii hex */ + *bp++ = digits[*rptr++ & 0xf]; + } + + if (m->m_next) { + if (bp > buf + sizeof(buf) - 3) + goto done; + *bp++ = '|'; + } else + *bp++ = ' '; + } +done: + if (m) + *bp++ = '>'; + *bp = 0; + printf("%s\n", buf); +} + +#endif /* NPPP > 0 */ diff --git a/sys/net/if_ppp.h b/sys/net/if_ppp.h new file mode 100644 index 00000000000..aa918868352 --- /dev/null +++ b/sys/net/if_ppp.h @@ -0,0 +1,130 @@ +/* $NetBSD: if_ppp.h,v 1.9 1995/07/04 06:28:22 paulus Exp $ */ + +/* + * if_ppp.h - Point-to-Point Protocol definitions. + * + * Copyright (c) 1989 Carnegie Mellon University. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by Carnegie Mellon University. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + */ + +#ifndef _IF_PPP_H_ +#define _IF_PPP_H_ + +/* + * Packet sizes + */ +#define PPP_MTU 1500 /* Default MTU (size of Info field) */ +#define PPP_MAXMRU 65000 /* Largest MRU we allow */ + +/* + * Bit definitions for flags. + */ +#define SC_COMP_PROT 0x00000001 /* protocol compression (output) */ +#define SC_COMP_AC 0x00000002 /* header compression (output) */ +#define SC_COMP_TCP 0x00000004 /* TCP (VJ) compression (output) */ +#define SC_NO_TCP_CCID 0x00000008 /* disable VJ connection-id comp. */ +#define SC_REJ_COMP_AC 0x00000010 /* reject adrs/ctrl comp. on input */ +#define SC_REJ_COMP_TCP 0x00000020 /* reject TCP (VJ) comp. on input */ +#define SC_CCP_OPEN 0x00000040 /* Look at CCP packets */ +#define SC_CCP_UP 0x00000080 /* May send/recv compressed packets */ +#define SC_DEBUG 0x00010000 /* enable debug messages */ +#define SC_LOG_INPKT 0x00020000 /* log contents of good pkts recvd */ +#define SC_LOG_OUTPKT 0x00040000 /* log contents of pkts sent */ +#define SC_LOG_RAWIN 0x00080000 /* log all chars received */ +#define SC_LOG_FLUSH 0x00100000 /* log all chars flushed */ +#define SC_RCV_B7_0 0x01000000 /* have rcvd char with bit 7 = 0 */ +#define SC_RCV_B7_1 0x02000000 /* have rcvd char with bit 7 = 1 */ +#define SC_RCV_EVNP 0x04000000 /* have rcvd char with even parity */ +#define SC_RCV_ODDP 0x08000000 /* have rcvd char with odd parity */ +#define SC_MASK 0x0fff00ff /* bits that user can change */ + +/* + * State bits in sc_flags, not changeable by user. + */ +#define SC_TIMEOUT 0x00000400 /* timeout is currently pending */ +#define SC_VJ_RESET 0x00000800 /* need to reset VJ decomp */ +#define SC_COMP_RUN 0x00001000 /* compressor has been inited */ +#define SC_DECOMP_RUN 0x00002000 /* decompressor has been inited */ +#define SC_DC_ERROR 0x00004000 /* non-fatal decomp error detected */ +#define SC_DC_FERROR 0x00008000 /* fatal decomp error detected */ +#define SC_TBUSY 0x10000000 /* xmitter doesn't need a packet yet */ +#define SC_PKTLOST 0x20000000 /* have lost or dropped a packet */ +#define SC_FLUSH 0x40000000 /* flush input until next PPP_FLAG */ +#define SC_ESCAPED 0x80000000 /* saw a PPP_ESCAPE */ + +/* + * Ioctl definitions. + */ + +struct npioctl { + int protocol; /* PPP procotol, e.g. PPP_IP */ + enum NPmode mode; +}; + +/* Structure describing a CCP configuration option, for PPPIOCSCOMPRESS */ +struct ppp_option_data { + u_char *ptr; + u_int length; + int transmit; +}; + +struct ifpppstatsreq { + char ifr_name[IFNAMSIZ]; + struct ppp_stats stats; +}; + +struct ifpppcstatsreq { + char ifr_name[IFNAMSIZ]; + struct ppp_comp_stats stats; +}; + +/* + * Ioctl definitions. + */ + +#define PPPIOCGFLAGS _IOR('t', 90, int) /* get configuration flags */ +#define PPPIOCSFLAGS _IOW('t', 89, int) /* set configuration flags */ +#define PPPIOCGASYNCMAP _IOR('t', 88, int) /* get async map */ +#define PPPIOCSASYNCMAP _IOW('t', 87, int) /* set async map */ +#define PPPIOCGUNIT _IOR('t', 86, int) /* get ppp unit number */ +#define PPPIOCGRASYNCMAP _IOR('t', 85, int) /* get receive async map */ +#define PPPIOCSRASYNCMAP _IOW('t', 84, int) /* set receive async map */ +#define PPPIOCGMRU _IOR('t', 83, int) /* get max receive unit */ +#define PPPIOCSMRU _IOW('t', 82, int) /* set max receive unit */ +#define PPPIOCSMAXCID _IOW('t', 81, int) /* set VJ max slot ID */ +#define PPPIOCGXASYNCMAP _IOR('t', 80, ext_accm) /* get extended ACCM */ +#define PPPIOCSXASYNCMAP _IOW('t', 79, ext_accm) /* set extended ACCM */ +#define PPPIOCXFERUNIT _IO('t', 78) /* transfer PPP unit */ +#define PPPIOCSCOMPRESS _IOW('t', 77, struct ppp_option_data) +#define PPPIOCGNPMODE _IOWR('t', 76, struct npioctl) /* get NP mode */ +#define PPPIOCSNPMODE _IOW('t', 75, struct npioctl) /* set NP mode */ +#define PPPIOCGIDLE _IOR('t', 74, struct ppp_idle) /* get idle time */ + +/* PPPIOC[GS]MTU are alternatives to SIOC[GS]IFMTU, used under Ultrix */ +#define PPPIOCGMTU _IOR('t', 73, int) /* get interface MTU */ +#define PPPIOCSMTU _IOW('t', 72, int) /* set interface MTU */ + +/* + * These two are interface ioctls so that pppstats can do them on + * a socket without having to open the serial device. + */ +#define SIOCGPPPSTATS _IOWR('i', 123, struct ifpppstatsreq) +#define SIOCGPPPCSTATS _IOWR('i', 122, struct ifpppcstatsreq) + +#if !defined(ifr_mtu) +#define ifr_mtu ifr_ifru.ifru_metric +#endif + +#endif /* _IF_PPP_H_ */ diff --git a/sys/net/if_pppvar.h b/sys/net/if_pppvar.h new file mode 100644 index 00000000000..721bd7f7a2d --- /dev/null +++ b/sys/net/if_pppvar.h @@ -0,0 +1,104 @@ +/* $NetBSD: if_pppvar.h,v 1.2 1995/07/04 15:30:39 briggs Exp $ */ +/* + * if_pppvar.h - private structures and declarations for PPP. + * + * Copyright (c) 1994 The Australian National University. + * All rights reserved. + * + * Permission to use, copy, modify, and distribute this software and its + * documentation is hereby granted, provided that the above copyright + * notice appears in all copies. This software is provided without any + * warranty, express or implied. The Australian National University + * makes no representations about the suitability of this software for + * any purpose. + * + * IN NO EVENT SHALL THE AUSTRALIAN NATIONAL UNIVERSITY BE LIABLE TO ANY + * PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF + * THE AUSTRALIAN NATIONAL UNIVERSITY HAVE BEEN ADVISED OF THE POSSIBILITY + * OF SUCH DAMAGE. + * + * THE AUSTRALIAN NATIONAL UNIVERSITY SPECIFICALLY DISCLAIMS ANY WARRANTIES, + * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY + * AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS + * ON AN "AS IS" BASIS, AND THE AUSTRALIAN NATIONAL UNIVERSITY HAS NO + * OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, + * OR MODIFICATIONS. + * + * Copyright (c) 1989 Carnegie Mellon University. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by Carnegie Mellon University. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + */ + +/* + * Supported network protocols. These values are used for + * indexing sc_npmode. + */ +#define NP_IP 0 /* Internet Protocol */ +#define NUM_NP 1 /* Number of NPs. */ + +/* + * Structure describing each ppp unit. + */ +struct ppp_softc { + struct ifnet sc_if; /* network-visible interface */ + u_int sc_flags; /* control/status bits; see if_ppp.h */ + void *sc_devp; /* pointer to device-dep structure */ + void (*sc_start) __P((struct ppp_softc *)); /* start output proc */ + void (*sc_ctlp) __P((struct ppp_softc *)); /* rcvd control pkt */ + void (*sc_relinq) __P((struct ppp_softc *)); /* relinquish ifunit */ + u_int16_t sc_mru; /* max receive unit */ + pid_t sc_xfer; /* used in transferring unit */ + struct ifqueue sc_rawq; /* received packets */ + struct ifqueue sc_inq; /* queue of input packets for daemon */ + struct ifqueue sc_fastq; /* interactive output packet q */ + struct mbuf *sc_togo; /* output packet ready to go */ + struct mbuf *sc_npqueue; /* output packets not to be sent yet */ + struct mbuf **sc_npqtail; /* ptr to last next ptr in npqueue */ +#ifdef VJC + struct slcompress sc_comp; /* vjc control buffer */ +#endif + u_int sc_bytessent; /* count of octets sent */ + u_int sc_bytesrcvd; /* count of octets received */ + caddr_t sc_bpf; /* hook for BPF */ + enum NPmode sc_npmode[NUM_NP]; /* what to do with each NP */ + struct compressor *sc_xcomp; /* transmit compressor */ + void *sc_xc_state; /* transmit compressor state */ + struct compressor *sc_rcomp; /* receive decompressor */ + void *sc_rc_state; /* receive decompressor state */ + time_t sc_last_sent; /* time (secs) last NP pkt sent */ + time_t sc_last_recv; /* time (secs) last NP pkt rcvd */ + + /* Device-dependent part for async lines. */ + ext_accm sc_asyncmap; /* async control character map */ + u_int32_t sc_rasyncmap; /* receive async control char map */ + struct mbuf *sc_outm; /* mbuf chain currently being output */ + struct mbuf *sc_m; /* pointer to input mbuf chain */ + struct mbuf *sc_mc; /* pointer to current input mbuf */ + char *sc_mp; /* ptr to next char in input mbuf */ + u_int16_t sc_ilen; /* length of input packet so far */ + u_int16_t sc_fcs; /* FCS so far (input) */ + u_int16_t sc_outfcs; /* FCS so far for output packet */ + u_char sc_rawin[16]; /* chars as received */ + int sc_rawin_count; /* # in sc_rawin */ +}; + +struct ppp_softc ppp_softc[NPPP]; + +struct ppp_softc *pppalloc __P((pid_t pid)); +void pppdealloc __P((struct ppp_softc *sc)); +int pppioctl __P((struct ppp_softc *sc, u_long cmd, caddr_t data, + int flag, struct proc *p)); +void ppppktin __P((struct ppp_softc *sc, struct mbuf *m, int lost)); +struct mbuf *ppp_dequeue __P((struct ppp_softc *sc)); diff --git a/sys/net/if_sl.c b/sys/net/if_sl.c new file mode 100644 index 00000000000..d2b768c41df --- /dev/null +++ b/sys/net/if_sl.c @@ -0,0 +1,908 @@ +/* $NetBSD: if_sl.c,v 1.37 1995/08/12 23:59:22 mycroft Exp $ */ + +/* + * Copyright (c) 1987, 1989, 1992, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_sl.c 8.6 (Berkeley) 2/1/94 + */ + +/* + * Serial Line interface + * + * Rick Adams + * Center for Seismic Studies + * 1300 N 17th Street, Suite 1450 + * Arlington, Virginia 22209 + * (703)276-7900 + * rick@seismo.ARPA + * seismo!rick + * + * Pounded on heavily by Chris Torek (chris@mimsy.umd.edu, umcp-cs!chris). + * N.B.: this belongs in netinet, not net, the way it stands now. + * Should have a link-layer type designation, but wouldn't be + * backwards-compatible. + * + * Converted to 4.3BSD Beta by Chris Torek. + * Other changes made at Berkeley, based in part on code by Kirk Smith. + * W. Jolitz added slip abort. + * + * Hacked almost beyond recognition by Van Jacobson (van@helios.ee.lbl.gov). + * Added priority queuing for "interactive" traffic; hooks for TCP + * header compression; ICMP filtering (at 2400 baud, some cretin + * pinging you can use up all your bandwidth). Made low clist behavior + * more robust and slightly less likely to hang serial line. + * Sped up a bunch of things. + * + * Note that splimp() is used throughout to block both (tty) input + * interrupts and network activity; thus, splimp must be >= spltty. + */ + +#include "sl.h" +#if NSL > 0 + +#include "bpfilter.h" + +#include <sys/param.h> +#include <sys/proc.h> +#include <sys/mbuf.h> +#include <sys/buf.h> +#include <sys/dkstat.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/file.h> +#include <sys/tty.h> +#include <sys/kernel.h> +#include <sys/conf.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/if_types.h> +#include <net/netisr.h> +#include <net/route.h> + +#if INET +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/in_var.h> +#include <netinet/ip.h> +#else +Huh? Slip without inet? +#endif + +#include <net/slcompress.h> +#include <net/if_slvar.h> +#include <net/slip.h> + +#if NBPFILTER > 0 +#include <sys/time.h> +#include <net/bpf.h> +#endif + +/* + * SLMAX is a hard limit on input packet size. To simplify the code + * and improve performance, we require that packets fit in an mbuf + * cluster, and if we get a compressed packet, there's enough extra + * room to expand the header into a max length tcp/ip header (128 + * bytes). So, SLMAX can be at most + * MCLBYTES - 128 + * + * SLMTU is a hard limit on output packet size. To insure good + * interactive response, SLMTU wants to be the smallest size that + * amortizes the header cost. (Remember that even with + * type-of-service queuing, we have to wait for any in-progress + * packet to finish. I.e., we wait, on the average, 1/2 * mtu / + * cps, where cps is the line speed in characters per second. + * E.g., 533ms wait for a 1024 byte MTU on a 9600 baud line. The + * average compressed header size is 6-8 bytes so any MTU > 90 + * bytes will give us 90% of the line bandwidth. A 100ms wait is + * tolerable (500ms is not), so want an MTU around 296. (Since TCP + * will send 256 byte segments (to allow for 40 byte headers), the + * typical packet size on the wire will be around 260 bytes). In + * 4.3tahoe+ systems, we can set an MTU in a route so we do that & + * leave the interface MTU relatively high (so we don't IP fragment + * when acting as a gateway to someone using a stupid MTU). + * + * Similar considerations apply to SLIP_HIWAT: It's the amount of + * data that will be queued 'downstream' of us (i.e., in clists + * waiting to be picked up by the tty output interrupt). If we + * queue a lot of data downstream, it's immune to our t.o.s. queuing. + * E.g., if SLIP_HIWAT is 1024, the interactive traffic in mixed + * telnet/ftp will see a 1 sec wait, independent of the mtu (the + * wait is dependent on the ftp window size but that's typically + * 1k - 4k). So, we want SLIP_HIWAT just big enough to amortize + * the cost (in idle time on the wire) of the tty driver running + * off the end of its clists & having to call back slstart for a + * new packet. For a tty interface with any buffering at all, this + * cost will be zero. Even with a totally brain dead interface (like + * the one on a typical workstation), the cost will be <= 1 character + * time. So, setting SLIP_HIWAT to ~100 guarantees that we'll lose + * at most 1% while maintaining good interactive response. + */ +#if NBPFILTER > 0 +#define BUFOFFSET (128+sizeof(struct ifnet **)+SLIP_HDRLEN) +#else +#define BUFOFFSET (128+sizeof(struct ifnet **)) +#endif +#define SLMAX (MCLBYTES - BUFOFFSET) +#define SLBUFSIZE (SLMAX + BUFOFFSET) +#ifndef SLMTU +#define SLMTU 296 +#endif +#if (SLMTU < 3) +Huh? SLMTU way too small. +#endif +#define SLIP_HIWAT roundup(50,CBSIZE) +#ifndef NetBSD /* XXX - cgd */ +#define CLISTRESERVE 1024 /* Can't let clists get too low */ +#endif /* !NetBSD */ + +/* + * SLIP ABORT ESCAPE MECHANISM: + * (inspired by HAYES modem escape arrangement) + * 1sec escape 1sec escape 1sec escape { 1sec escape 1sec escape } + * within window time signals a "soft" exit from slip mode by remote end + * if the IFF_DEBUG flag is on. + */ +#define ABT_ESC '\033' /* can't be t_intr - distant host must know it*/ +#define ABT_IDLE 1 /* in seconds - idle before an escape */ +#define ABT_COUNT 3 /* count of escapes for abort */ +#define ABT_WINDOW (ABT_COUNT*2+2) /* in seconds - time to count */ + +struct sl_softc sl_softc[NSL]; + +#define FRAME_END 0xc0 /* Frame End */ +#define FRAME_ESCAPE 0xdb /* Frame Esc */ +#define TRANS_FRAME_END 0xdc /* transposed frame end */ +#define TRANS_FRAME_ESCAPE 0xdd /* transposed frame esc */ + +static int slinit __P((struct sl_softc *)); +static struct mbuf *sl_btom __P((struct sl_softc *, int)); + +/* + * Called from boot code to establish sl interfaces. + */ +void +slattach() +{ + register struct sl_softc *sc; + register int i = 0; + + for (sc = sl_softc; i < NSL; sc++) { + sc->sc_if.if_name = "sl"; + sc->sc_if.if_unit = i++; + sc->sc_if.if_mtu = SLMTU; + sc->sc_if.if_flags = + IFF_POINTOPOINT | SC_AUTOCOMP | IFF_MULTICAST; + sc->sc_if.if_type = IFT_SLIP; + sc->sc_if.if_ioctl = slioctl; + sc->sc_if.if_output = sloutput; + sc->sc_if.if_snd.ifq_maxlen = 50; + sc->sc_fastq.ifq_maxlen = 32; + if_attach(&sc->sc_if); +#if NBPFILTER > 0 + bpfattach(&sc->sc_bpf, &sc->sc_if, DLT_SLIP, SLIP_HDRLEN); +#endif + } +} + +static int +slinit(sc) + register struct sl_softc *sc; +{ + register caddr_t p; + + if (sc->sc_ep == (u_char *) 0) { + MCLALLOC(p, M_WAIT); + if (p) + sc->sc_ep = (u_char *)p + SLBUFSIZE; + else { + printf("sl%d: can't allocate buffer\n", sc - sl_softc); + sc->sc_if.if_flags &= ~IFF_UP; + return (0); + } + } + sc->sc_buf = sc->sc_ep - SLMAX; + sc->sc_mp = sc->sc_buf; + sl_compress_init(&sc->sc_comp, -1); + return (1); +} + +/* + * Line specific open routine. + * Attach the given tty to the first available sl unit. + */ +/* ARGSUSED */ +int +slopen(dev, tp) + dev_t dev; + register struct tty *tp; +{ + struct proc *p = curproc; /* XXX */ + register struct sl_softc *sc; + register int nsl; + int error; +#ifdef NetBSD + int s; +#endif + + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + + if (tp->t_line == SLIPDISC) + return (0); + + for (nsl = NSL, sc = sl_softc; --nsl >= 0; sc++) + if (sc->sc_ttyp == NULL) { + if (slinit(sc) == 0) + return (ENOBUFS); + tp->t_sc = (caddr_t)sc; + sc->sc_ttyp = tp; + sc->sc_if.if_baudrate = tp->t_ospeed; + ttyflush(tp, FREAD | FWRITE); +#ifdef NetBSD + /* + * make sure tty output queue is large enough + * to hold a full-sized packet (including frame + * end, and a possible extra frame end). full-sized + * packet occupies a max of 2*SLMTU bytes (because + * of possible escapes), and add two on for frame + * ends. + */ + s = spltty(); + if (tp->t_outq.c_cn < 2*SLMTU+2) { + sc->sc_oldbufsize = tp->t_outq.c_cn; + sc->sc_oldbufquot = tp->t_outq.c_cq != 0; + + clfree(&tp->t_outq); + error = clalloc(&tp->t_outq, 3*SLMTU, 0); + if (error) { + splx(s); + return(error); + } + } else + sc->sc_oldbufsize = sc->sc_oldbufquot = 0; + splx(s); +#endif /* NetBSD */ + return (0); + } + return (ENXIO); +} + +/* + * Line specific close routine. + * Detach the tty from the sl unit. + */ +void +slclose(tp) + struct tty *tp; +{ + register struct sl_softc *sc; + int s; + + ttywflush(tp); + s = splimp(); /* actually, max(spltty, splsoftnet) */ + tp->t_line = 0; + sc = (struct sl_softc *)tp->t_sc; + if (sc != NULL) { + if_down(&sc->sc_if); + sc->sc_ttyp = NULL; + tp->t_sc = NULL; + MCLFREE((caddr_t)(sc->sc_ep - SLBUFSIZE)); + sc->sc_ep = 0; + sc->sc_mp = 0; + sc->sc_buf = 0; + } +#ifdef NetBSD + /* if necessary, install a new outq buffer of the appropriate size */ + if (sc->sc_oldbufsize != 0) { + clfree(&tp->t_outq); + clalloc(&tp->t_outq, sc->sc_oldbufsize, sc->sc_oldbufquot); + } +#endif + splx(s); +} + +/* + * Line specific (tty) ioctl routine. + * Provide a way to get the sl unit number. + */ +/* ARGSUSED */ +int +sltioctl(tp, cmd, data, flag) + struct tty *tp; + u_long cmd; + caddr_t data; + int flag; +{ + struct sl_softc *sc = (struct sl_softc *)tp->t_sc; + + switch (cmd) { + case SLIOCGUNIT: + *(int *)data = sc->sc_if.if_unit; + break; + + default: + return (-1); + } + return (0); +} + +/* + * Queue a packet. Start transmission if not active. + * Compression happens in slstart; if we do it here, IP TOS + * will cause us to not compress "background" packets, because + * ordering gets trashed. It can be done for all packets in slstart. + */ +int +sloutput(ifp, m, dst, rtp) + struct ifnet *ifp; + register struct mbuf *m; + struct sockaddr *dst; + struct rtentry *rtp; +{ + register struct sl_softc *sc = &sl_softc[ifp->if_unit]; + register struct ip *ip; + register struct ifqueue *ifq; + int s; + + /* + * `Cannot happen' (see slioctl). Someday we will extend + * the line protocol to support other address families. + */ + if (dst->sa_family != AF_INET) { + printf("sl%d: af%d not supported\n", sc->sc_if.if_unit, + dst->sa_family); + m_freem(m); + sc->sc_if.if_noproto++; + return (EAFNOSUPPORT); + } + + if (sc->sc_ttyp == NULL) { + m_freem(m); + return (ENETDOWN); /* sort of */ + } + if ((sc->sc_ttyp->t_state & TS_CARR_ON) == 0 && + (sc->sc_ttyp->t_cflag & CLOCAL) == 0) { + m_freem(m); + return (EHOSTUNREACH); + } + ifq = &sc->sc_if.if_snd; + ip = mtod(m, struct ip *); + if (sc->sc_if.if_flags & SC_NOICMP && ip->ip_p == IPPROTO_ICMP) { + m_freem(m); + return (ENETRESET); /* XXX ? */ + } + if (ip->ip_tos & IPTOS_LOWDELAY) + ifq = &sc->sc_fastq; + s = splimp(); + if (sc->sc_oqlen && sc->sc_ttyp->t_outq.c_cc == sc->sc_oqlen) { + struct timeval tv; + + /* if output's been stalled for too long, and restart */ + timersub(&time, &sc->sc_if.if_lastchange, &tv); + if (tv.tv_sec > 0) { + sc->sc_otimeout++; + slstart(sc->sc_ttyp); + } + } + if (IF_QFULL(ifq)) { + IF_DROP(ifq); + m_freem(m); + splx(s); + sc->sc_if.if_oerrors++; + return (ENOBUFS); + } + IF_ENQUEUE(ifq, m); + sc->sc_if.if_lastchange = time; + if ((sc->sc_oqlen = sc->sc_ttyp->t_outq.c_cc) == 0) + slstart(sc->sc_ttyp); + splx(s); + return (0); +} + +/* + * Start output on interface. Get another datagram + * to send from the interface queue and map it to + * the interface before starting output. + */ +void +slstart(tp) + register struct tty *tp; +{ + register struct sl_softc *sc = (struct sl_softc *)tp->t_sc; + register struct mbuf *m; + register u_char *cp; + register struct ip *ip; + int s; + struct mbuf *m2; +#if NBPFILTER > 0 + u_char bpfbuf[SLMTU + SLIP_HDRLEN]; + register int len; +#endif +#ifndef NetBSD /* XXX - cgd */ + extern int cfreecount; +#endif + + for (;;) { + /* + * If there is more in the output queue, just send it now. + * We are being called in lieu of ttstart and must do what + * it would. + */ + if (tp->t_outq.c_cc != 0) { + (*tp->t_oproc)(tp); + if (tp->t_outq.c_cc > SLIP_HIWAT) + return; + } + /* + * This happens briefly when the line shuts down. + */ + if (sc == NULL) + return; + +#ifdef NetBSD /* XXX - cgd */ + /* + * Do not remove the packet from the IP queue if it + * doesn't look like the packet will fit into the + * current serial output queue, with a packet full of + * escapes this could be as bad as SLMTU*2+2. + */ + if (tp->t_outq.c_cn - tp->t_outq.c_cc < 2*SLMTU+2) + return; +#endif /* NetBSD */ + + /* + * Get a packet and send it to the interface. + */ + s = splimp(); + IF_DEQUEUE(&sc->sc_fastq, m); + if (m) + sc->sc_if.if_omcasts++; /* XXX */ + else + IF_DEQUEUE(&sc->sc_if.if_snd, m); + splx(s); + if (m == NULL) + return; + + /* + * We do the header compression here rather than in sloutput + * because the packets will be out of order if we are using TOS + * queueing, and the connection id compression will get + * munged when this happens. + */ +#if NBPFILTER > 0 + if (sc->sc_bpf) { + /* + * We need to save the TCP/IP header before it's + * compressed. To avoid complicated code, we just + * copy the entire packet into a stack buffer (since + * this is a serial line, packets should be short + * and/or the copy should be negligible cost compared + * to the packet transmission time). + */ + register struct mbuf *m1 = m; + register u_char *cp = bpfbuf + SLIP_HDRLEN; + + len = 0; + do { + register int mlen = m1->m_len; + + bcopy(mtod(m1, caddr_t), cp, mlen); + cp += mlen; + len += mlen; + } while (m1 = m1->m_next); + } +#endif + if ((ip = mtod(m, struct ip *))->ip_p == IPPROTO_TCP) { + if (sc->sc_if.if_flags & SC_COMPRESS) + *mtod(m, u_char *) |= sl_compress_tcp(m, ip, + &sc->sc_comp, 1); + } +#if NBPFILTER > 0 + if (sc->sc_bpf) { + /* + * Put the SLIP pseudo-"link header" in place. The + * compressed header is now at the beginning of the + * mbuf. + */ + bpfbuf[SLX_DIR] = SLIPDIR_OUT; + bcopy(mtod(m, caddr_t), &bpfbuf[SLX_CHDR], CHDR_LEN); + bpf_tap(sc->sc_bpf, bpfbuf, len + SLIP_HDRLEN); + } +#endif + sc->sc_if.if_lastchange = time; + +#ifndef NetBSD /* XXX - cgd */ + /* + * If system is getting low on clists, just flush our + * output queue (if the stuff was important, it'll get + * retransmitted). + */ + if (cfreecount < CLISTRESERVE + SLMTU) { + m_freem(m); + sc->sc_if.if_collisions++; + continue; + } +#endif /* !NetBSD */ + /* + * The extra FRAME_END will start up a new packet, and thus + * will flush any accumulated garbage. We do this whenever + * the line may have been idle for some time. + */ + if (tp->t_outq.c_cc == 0) { + ++sc->sc_if.if_obytes; + (void) putc(FRAME_END, &tp->t_outq); + } + + while (m) { + register u_char *ep; + + cp = mtod(m, u_char *); ep = cp + m->m_len; + while (cp < ep) { + /* + * Find out how many bytes in the string we can + * handle without doing something special. + */ + register u_char *bp = cp; + + while (cp < ep) { + switch (*cp++) { + case FRAME_ESCAPE: + case FRAME_END: + --cp; + goto out; + } + } + out: + if (cp > bp) { + /* + * Put n characters at once + * into the tty output queue. + */ +#ifdef NetBSD /* XXX - cgd */ + if (b_to_q((u_char *)bp, cp - bp, +#else + if (b_to_q((char *)bp, cp - bp, +#endif + &tp->t_outq)) + break; + sc->sc_if.if_obytes += cp - bp; + } + /* + * If there are characters left in the mbuf, + * the first one must be special.. + * Put it out in a different form. + */ + if (cp < ep) { + if (putc(FRAME_ESCAPE, &tp->t_outq)) + break; + if (putc(*cp++ == FRAME_ESCAPE ? + TRANS_FRAME_ESCAPE : TRANS_FRAME_END, + &tp->t_outq)) { + (void) unputc(&tp->t_outq); + break; + } + sc->sc_if.if_obytes += 2; + } + } + MFREE(m, m2); + m = m2; + } + + if (putc(FRAME_END, &tp->t_outq)) { + /* + * Not enough room. Remove a char to make room + * and end the packet normally. + * If you get many collisions (more than one or two + * a day) you probably do not have enough clists + * and you should increase "nclist" in param.c. + */ + (void) unputc(&tp->t_outq); + (void) putc(FRAME_END, &tp->t_outq); + sc->sc_if.if_collisions++; + } else { + ++sc->sc_if.if_obytes; + sc->sc_if.if_opackets++; + } + } +} + +/* + * Copy data buffer to mbuf chain; add ifnet pointer. + */ +static struct mbuf * +sl_btom(sc, len) + register struct sl_softc *sc; + register int len; +{ + register struct mbuf *m; + + MGETHDR(m, M_DONTWAIT, MT_DATA); + if (m == NULL) + return (NULL); + + /* + * If we have more than MHLEN bytes, it's cheaper to + * queue the cluster we just filled & allocate a new one + * for the input buffer. Otherwise, fill the mbuf we + * allocated above. Note that code in the input routine + * guarantees that packet will fit in a cluster. + */ + if (len >= MHLEN) { + MCLGET(m, M_DONTWAIT); + if ((m->m_flags & M_EXT) == 0) { + /* + * we couldn't get a cluster - if memory's this + * low, it's time to start dropping packets. + */ + (void) m_free(m); + return (NULL); + } + sc->sc_ep = mtod(m, u_char *) + SLBUFSIZE; + m->m_data = (caddr_t)sc->sc_buf; + m->m_ext.ext_buf = (caddr_t)((long)sc->sc_buf &~ MCLOFSET); + } else + bcopy((caddr_t)sc->sc_buf, mtod(m, caddr_t), len); + + m->m_len = len; + m->m_pkthdr.len = len; + m->m_pkthdr.rcvif = &sc->sc_if; + return (m); +} + +/* + * tty interface receiver interrupt. + */ +void +slinput(c, tp) + register int c; + register struct tty *tp; +{ + register struct sl_softc *sc; + register struct mbuf *m; + register int len; + int s; +#if NBPFILTER > 0 + u_char chdr[CHDR_LEN]; +#endif + + tk_nin++; + sc = (struct sl_softc *)tp->t_sc; + if (sc == NULL) + return; + if (c & TTY_ERRORMASK || ((tp->t_state & TS_CARR_ON) == 0 && + (tp->t_cflag & CLOCAL) == 0)) { + sc->sc_flags |= SC_ERROR; + return; + } + c &= TTY_CHARMASK; + + ++sc->sc_if.if_ibytes; + + if (sc->sc_if.if_flags & IFF_DEBUG) { + if (c == ABT_ESC) { + /* + * If we have a previous abort, see whether + * this one is within the time limit. + */ + if (sc->sc_abortcount && + time.tv_sec >= sc->sc_starttime + ABT_WINDOW) + sc->sc_abortcount = 0; + /* + * If we see an abort after "idle" time, count it; + * record when the first abort escape arrived. + */ + if (time.tv_sec >= sc->sc_lasttime + ABT_IDLE) { + if (++sc->sc_abortcount == 1) + sc->sc_starttime = time.tv_sec; + if (sc->sc_abortcount >= ABT_COUNT) { + slclose(tp); + return; + } + } + } else + sc->sc_abortcount = 0; + sc->sc_lasttime = time.tv_sec; + } + + switch (c) { + + case TRANS_FRAME_ESCAPE: + if (sc->sc_escape) + c = FRAME_ESCAPE; + break; + + case TRANS_FRAME_END: + if (sc->sc_escape) + c = FRAME_END; + break; + + case FRAME_ESCAPE: + sc->sc_escape = 1; + return; + + case FRAME_END: + if(sc->sc_flags & SC_ERROR) { + sc->sc_flags &= ~SC_ERROR; + goto newpack; + } + len = sc->sc_mp - sc->sc_buf; + if (len < 3) + /* less than min length packet - ignore */ + goto newpack; + +#if NBPFILTER > 0 + if (sc->sc_bpf) { + /* + * Save the compressed header, so we + * can tack it on later. Note that we + * will end up copying garbage in some + * cases but this is okay. We remember + * where the buffer started so we can + * compute the new header length. + */ + bcopy(sc->sc_buf, chdr, CHDR_LEN); + } +#endif + + if ((c = (*sc->sc_buf & 0xf0)) != (IPVERSION << 4)) { + if (c & 0x80) + c = TYPE_COMPRESSED_TCP; + else if (c == TYPE_UNCOMPRESSED_TCP) + *sc->sc_buf &= 0x4f; /* XXX */ + /* + * We've got something that's not an IP packet. + * If compression is enabled, try to decompress it. + * Otherwise, if `auto-enable' compression is on and + * it's a reasonable packet, decompress it and then + * enable compression. Otherwise, drop it. + */ + if (sc->sc_if.if_flags & SC_COMPRESS) { + len = sl_uncompress_tcp(&sc->sc_buf, len, + (u_int)c, &sc->sc_comp); + if (len <= 0) + goto error; + } else if ((sc->sc_if.if_flags & SC_AUTOCOMP) && + c == TYPE_UNCOMPRESSED_TCP && len >= 40) { + len = sl_uncompress_tcp(&sc->sc_buf, len, + (u_int)c, &sc->sc_comp); + if (len <= 0) + goto error; + sc->sc_if.if_flags |= SC_COMPRESS; + } else + goto error; + } +#if NBPFILTER > 0 + if (sc->sc_bpf) { + /* + * Put the SLIP pseudo-"link header" in place. + * We couldn't do this any earlier since + * decompression probably moved the buffer + * pointer. Then, invoke BPF. + */ + register u_char *hp = sc->sc_buf - SLIP_HDRLEN; + + hp[SLX_DIR] = SLIPDIR_IN; + bcopy(chdr, &hp[SLX_CHDR], CHDR_LEN); + bpf_tap(sc->sc_bpf, hp, len + SLIP_HDRLEN); + } +#endif + m = sl_btom(sc, len); + if (m == NULL) + goto error; + + sc->sc_if.if_ipackets++; + sc->sc_if.if_lastchange = time; + s = splimp(); + if (IF_QFULL(&ipintrq)) { + IF_DROP(&ipintrq); + sc->sc_if.if_ierrors++; + sc->sc_if.if_iqdrops++; + m_freem(m); + } else { + IF_ENQUEUE(&ipintrq, m); + schednetisr(NETISR_IP); + } + splx(s); + goto newpack; + } + if (sc->sc_mp < sc->sc_ep) { + *sc->sc_mp++ = c; + sc->sc_escape = 0; + return; + } + + /* can't put lower; would miss an extra frame */ + sc->sc_flags |= SC_ERROR; + +error: + sc->sc_if.if_ierrors++; +newpack: + sc->sc_mp = sc->sc_buf = sc->sc_ep - SLMAX; + sc->sc_escape = 0; +} + +/* + * Process an ioctl request. + */ +int +slioctl(ifp, cmd, data) + register struct ifnet *ifp; + u_long cmd; + caddr_t data; +{ + register struct ifaddr *ifa = (struct ifaddr *)data; + register struct ifreq *ifr; + register int s = splimp(), error = 0; + + switch (cmd) { + + case SIOCSIFADDR: + if (ifa->ifa_addr->sa_family == AF_INET) + ifp->if_flags |= IFF_UP; + else + error = EAFNOSUPPORT; + break; + + case SIOCSIFDSTADDR: + if (ifa->ifa_addr->sa_family != AF_INET) + error = EAFNOSUPPORT; + break; + + case SIOCADDMULTI: + case SIOCDELMULTI: + ifr = (struct ifreq *)data; + if (ifr == 0) { + error = EAFNOSUPPORT; /* XXX */ + break; + } + switch (ifr->ifr_addr.sa_family) { + +#ifdef INET + case AF_INET: + break; +#endif + + default: + error = EAFNOSUPPORT; + break; + } + break; + + default: + error = EINVAL; + } + splx(s); + return (error); +} +#endif diff --git a/sys/net/if_slvar.h b/sys/net/if_slvar.h new file mode 100644 index 00000000000..17cc730c162 --- /dev/null +++ b/sys/net/if_slvar.h @@ -0,0 +1,86 @@ +/* $NetBSD: if_slvar.h,v 1.15 1995/03/26 20:30:14 jtc Exp $ */ + +/*- + * Copyright (c) 1991, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_slvar.h 8.3 (Berkeley) 2/1/94 + */ + +/* + * Definitions for SLIP interface data structures + * + * (This exists so programs like slstats can get at the definition + * of sl_softc.) + */ +struct sl_softc { + struct ifnet sc_if; /* network-visible interface */ + struct ifqueue sc_fastq; /* interactive output queue */ + struct tty *sc_ttyp; /* pointer to tty structure */ + u_char *sc_mp; /* pointer to next available buf char */ + u_char *sc_ep; /* pointer to last available buf char */ + u_char *sc_buf; /* input buffer */ + u_int sc_flags; /* see below */ + u_int sc_escape; /* =1 if last char input was FRAME_ESCAPE */ + long sc_lasttime; /* last time a char arrived */ + long sc_abortcount; /* number of abort esacpe chars */ + long sc_starttime; /* time of first abort in window */ + long sc_oqlen; /* previous output queue size */ + long sc_otimeout; /* number of times output's stalled */ +#ifdef NetBSD + int sc_oldbufsize; /* previous output buffer size */ + int sc_oldbufquot; /* previous output buffer quoting */ +#endif +#ifdef INET /* XXX */ + struct slcompress sc_comp; /* tcp compression data */ +#endif + caddr_t sc_bpf; /* BPF data */ +}; + +/* internal flags */ +#define SC_ERROR 0x0001 /* had an input error */ + +/* visible flags */ +#define SC_COMPRESS IFF_LINK0 /* compress TCP traffic */ +#define SC_NOICMP IFF_LINK1 /* supress ICMP traffic */ +#define SC_AUTOCOMP IFF_LINK2 /* auto-enable TCP compression */ + +#ifdef _KERNEL +void slattach __P((void)); +void slclose __P((struct tty *)); +void slinput __P((int, struct tty *)); +int slioctl __P((struct ifnet *, u_long, caddr_t)); +int slopen __P((dev_t, struct tty *)); +int sloutput __P((struct ifnet *, + struct mbuf *, struct sockaddr *, struct rtentry *)); +void slstart __P((struct tty *)); +int sltioctl __P((struct tty *, u_long, caddr_t, int)); +#endif /* _KERNEL */ diff --git a/sys/net/if_tun.c b/sys/net/if_tun.c new file mode 100644 index 00000000000..dcd59335bcc --- /dev/null +++ b/sys/net/if_tun.c @@ -0,0 +1,558 @@ +/* $NetBSD: if_tun.c,v 1.18 1995/06/13 05:59:37 mycroft Exp $ */ + +/* + * Copyright (c) 1988, Julian Onions <jpo@cs.nott.ac.uk> + * Nottingham University 1987. + * + * This source may be freely distributed, however I would be interested + * in any changes that are made. + * + * This driver takes packets off the IP i/f and hands them up to a + * user process to have it's wicked way with. This driver has it's + * roots in a similar driver written by Phil Cockcroft (formerly) at + * UCL. This driver is based much more on read/write/select mode of + * operation though. + */ + +#include "tun.h" +#if NTUN > 0 + +#include <sys/param.h> +#include <sys/proc.h> +#include <sys/systm.h> +#include <sys/mbuf.h> +#include <sys/buf.h> +#include <sys/protosw.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/errno.h> +#include <sys/syslog.h> +#include <sys/select.h> +#include <sys/file.h> + +#include <machine/cpu.h> + +#include <net/if.h> +#include <net/netisr.h> +#include <net/route.h> + +#ifdef INET +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/in_var.h> +#include <netinet/ip.h> +#include <netinet/if_ether.h> +#endif + +#ifdef NS +#include <netns/ns.h> +#include <netns/ns_if.h> +#endif + +#include "bpfilter.h" +#if NBPFILTER > 0 +#include <sys/time.h> +#include <net/bpf.h> +#endif + +#include <net/if_tun.h> + +#define TUNDEBUG if (tundebug) printf +int tundebug = 0; + +struct tun_softc tunctl[NTUN]; +extern int ifqmaxlen; + +int tunopen __P((dev_t, int, int, struct proc *)); +int tunclose __P((dev_t, int)); +int tunoutput __P((struct ifnet *, struct mbuf *, struct sockaddr *, + struct rtentry *rt)); +int tunread __P((dev_t, struct uio *)); +int tunwrite __P((dev_t, struct uio *)); +int tuncioctl __P((dev_t, u_long, caddr_t, int, struct proc *)); +int tunioctl __P((struct ifnet *, u_long, caddr_t)); +int tunselect __P((dev_t, int)); +void tunattach __P((int)); + +static int tuninit __P((int)); + +void +tunattach(unused) + int unused; +{ + register int i; + struct ifnet *ifp; + struct sockaddr_in *sin; + + for (i = 0; i < NTUN; i++) { + tunctl[i].tun_flags = TUN_INITED; + + ifp = &tunctl[i].tun_if; + ifp->if_unit = i; + ifp->if_name = "tun"; + ifp->if_mtu = TUNMTU; + ifp->if_ioctl = tunioctl; + ifp->if_output = tunoutput; + ifp->if_flags = IFF_POINTOPOINT; + ifp->if_snd.ifq_maxlen = ifqmaxlen; + ifp->if_collisions = 0; + ifp->if_ierrors = 0; + ifp->if_oerrors = 0; + ifp->if_ipackets = 0; + ifp->if_opackets = 0; + if_attach(ifp); +#if NBPFILTER > 0 + bpfattach(&tunctl[i].tun_bpf, ifp, DLT_NULL, sizeof(u_int32_t)); +#endif + } +} + +/* + * tunnel open - must be superuser & the device must be + * configured in + */ +int +tunopen(dev, flag, mode, p) + dev_t dev; + int flag, mode; + struct proc *p; +{ + struct ifnet *ifp; + struct tun_softc *tp; + register int unit, error; + + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + + if ((unit = minor(dev)) >= NTUN) + return (ENXIO); + tp = &tunctl[unit]; + if (tp->tun_flags & TUN_OPEN) + return ENXIO; + ifp = &tp->tun_if; + tp->tun_flags |= TUN_OPEN; + TUNDEBUG("%s%d: open\n", ifp->if_name, ifp->if_unit); + return (0); +} + +/* + * tunclose - close the device - mark i/f down & delete + * routing info + */ +int +tunclose(dev, flag) + dev_t dev; + int flag; +{ + register int unit = minor(dev), s; + struct tun_softc *tp = &tunctl[unit]; + struct ifnet *ifp = &tp->tun_if; + struct mbuf *m; + + tp->tun_flags &= ~TUN_OPEN; + + /* + * junk all pending output + */ + do { + s = splimp(); + IF_DEQUEUE(&ifp->if_snd, m); + splx(s); + if (m) + m_freem(m); + } while (m); + + if (ifp->if_flags & IFF_UP) { + s = splimp(); + if_down(ifp); + if (ifp->if_flags & IFF_RUNNING) { + /* find internet addresses and delete routes */ + register struct ifaddr *ifa; + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; + ifa = ifa->ifa_list.tqe_next) { + if (ifa->ifa_addr->sa_family == AF_INET) { + rtinit(ifa, (int)RTM_DELETE, + tp->tun_flags & TUN_DSTADDR ? RTF_HOST : 0); + } + } + } + splx(s); + } + tp->tun_pgrp = 0; + selwakeup(&tp->tun_rsel); + + TUNDEBUG ("%s%d: closed\n", ifp->if_name, ifp->if_unit); + return (0); +} + +static int +tuninit(unit) + int unit; +{ + struct tun_softc *tp = &tunctl[unit]; + struct ifnet *ifp = &tp->tun_if; + register struct ifaddr *ifa; + + TUNDEBUG("%s%d: tuninit\n", ifp->if_name, ifp->if_unit); + + ifp->if_flags |= IFF_UP | IFF_RUNNING; + + for (ifa = ifp->if_addrlist.tqh_first; ifa != 0; + ifa = ifa->ifa_list.tqe_next) { + if (ifa->ifa_addr->sa_family == AF_INET) { + struct sockaddr_in *sin; + + sin = satosin(ifa->ifa_addr); + if (sin && sin->sin_addr.s_addr) + tp->tun_flags |= TUN_IASET; + + sin = satosin(ifa->ifa_dstaddr); + if (sin && sin->sin_addr.s_addr) + tp->tun_flags |= TUN_DSTADDR; + } + } + + return 0; +} + +/* + * Process an ioctl request. + */ +int +tunioctl(ifp, cmd, data) + struct ifnet *ifp; + u_long cmd; + caddr_t data; +{ + struct tun_softc *tp = &tunctl[ifp->if_unit]; + int error = 0, s; + + s = splimp(); + switch(cmd) { + case SIOCSIFADDR: + tuninit(ifp->if_unit); + TUNDEBUG("%s%d: address set\n", + ifp->if_name, ifp->if_unit); + break; + case SIOCSIFDSTADDR: + tuninit(ifp->if_unit); + TUNDEBUG("%s%d: destination address set\n", + ifp->if_name, ifp->if_unit); + break; + default: + error = EINVAL; + } + splx(s); + return (error); +} + +/* + * tunoutput - queue packets from higher level ready to put out. + */ +int +tunoutput(ifp, m0, dst, rt) + struct ifnet *ifp; + struct mbuf *m0; + struct sockaddr *dst; + struct rtentry *rt; +{ + struct tun_softc *tp = &tunctl[ifp->if_unit]; + struct proc *p; + int s; + + TUNDEBUG ("%s%d: tunoutput\n", ifp->if_name, ifp->if_unit); + + if ((tp->tun_flags & TUN_READY) != TUN_READY) { + TUNDEBUG ("%s%d: not ready 0%o\n", ifp->if_name, + ifp->if_unit, tp->tun_flags); + m_freem (m0); + return EHOSTDOWN; + } + +#if NBPFILTER > 0 + if (tp->tun_bpf) { + /* + * We need to prepend the address family as + * a four byte field. Cons up a dummy header + * to pacify bpf. This is safe because bpf + * will only read from the mbuf (i.e., it won't + * try to free it or keep a pointer to it). + */ + struct mbuf m; + u_int32_t af = dst->sa_family; + + m.m_next = m0; + m.m_len = sizeof(af); + m.m_data = (char *)⁡ + + bpf_mtap(tp->tun_bpf, &m); + } +#endif + + switch(dst->sa_family) { +#ifdef INET + case AF_INET: + s = splimp(); + if (IF_QFULL(&ifp->if_snd)) { + IF_DROP(&ifp->if_snd); + m_freem(m0); + splx(s); + ifp->if_collisions++; + return (ENOBUFS); + } + IF_ENQUEUE(&ifp->if_snd, m0); + splx(s); + ifp->if_opackets++; + break; +#endif + default: + m_freem(m0); + return EAFNOSUPPORT; + } + + if (tp->tun_flags & TUN_RWAIT) { + tp->tun_flags &= ~TUN_RWAIT; + wakeup((caddr_t)tp); + } + if (tp->tun_flags & TUN_ASYNC && tp->tun_pgrp) { + if (tp->tun_pgrp > 0) + gsignal(tp->tun_pgrp, SIGIO); + else if (p = pfind(-tp->tun_pgrp)) + psignal(p, SIGIO); + } + selwakeup(&tp->tun_rsel); + return 0; +} + +/* + * the cdevsw interface is now pretty minimal. + */ +int +tuncioctl(dev, cmd, data, flag, p) + dev_t dev; + u_long cmd; + caddr_t data; + int flag; + struct proc *p; +{ + int unit = minor(dev), s; + struct tun_softc *tp = &tunctl[unit]; + + switch (cmd) { + case TUNSDEBUG: + tundebug = *(int *)data; + break; + case TUNGDEBUG: + *(int *)data = tundebug; + break; + case FIONBIO: + if (*(int *)data) + tp->tun_flags |= TUN_NBIO; + else + tp->tun_flags &= ~TUN_NBIO; + break; + case FIOASYNC: + if (*(int *)data) + tp->tun_flags |= TUN_ASYNC; + else + tp->tun_flags &= ~TUN_ASYNC; + break; + case FIONREAD: + s = splimp(); + if (tp->tun_if.if_snd.ifq_head) + *(int *)data = tp->tun_if.if_snd.ifq_head->m_len; + else + *(int *)data = 0; + splx(s); + break; + case TIOCSPGRP: + tp->tun_pgrp = *(int *)data; + break; + case TIOCGPGRP: + *(int *)data = tp->tun_pgrp; + break; + default: + return (ENOTTY); + } + return (0); +} + +/* + * The cdevsw read interface - reads a packet at a time, or at + * least as much of a packet as can be read. + */ +int +tunread(dev, uio) + dev_t dev; + struct uio *uio; +{ + int unit = minor(dev); + struct tun_softc *tp = &tunctl[unit]; + struct ifnet *ifp = &tp->tun_if; + struct mbuf *m, *m0; + int error=0, len, s; + + TUNDEBUG ("%s%d: read\n", ifp->if_name, ifp->if_unit); + if ((tp->tun_flags & TUN_READY) != TUN_READY) { + TUNDEBUG ("%s%d: not ready 0%o\n", ifp->if_name, + ifp->if_unit, tp->tun_flags); + return EHOSTDOWN; + } + + tp->tun_flags &= ~TUN_RWAIT; + + s = splimp(); + do { + IF_DEQUEUE(&ifp->if_snd, m0); + if (m0 == 0) { + if (tp->tun_flags & TUN_NBIO) { + splx(s); + return EWOULDBLOCK; + } + tp->tun_flags |= TUN_RWAIT; + tsleep((caddr_t)tp, PZERO + 1, "tunread", 0); + } + } while (m0 == 0); + splx(s); + + while (m0 && uio->uio_resid > 0 && error == 0) { + len = min(uio->uio_resid, m0->m_len); + if (len == 0) + break; + error = uiomove(mtod(m0, caddr_t), len, uio); + MFREE(m0, m); + m0 = m; + } + + if (m0) { + TUNDEBUG("Dropping mbuf\n"); + m_freem(m0); + } + return error; +} + +/* + * the cdevsw write interface - an atomic write is a packet - or else! + */ +int +tunwrite(dev, uio) + dev_t dev; + struct uio *uio; +{ + int unit = minor (dev); + struct ifnet *ifp = &tunctl[unit].tun_if; + struct mbuf *top, **mp, *m; + int error=0, s, tlen, mlen; + + TUNDEBUG("%s%d: tunwrite\n", ifp->if_name, ifp->if_unit); + + if (uio->uio_resid < 0 || uio->uio_resid > TUNMTU) { + TUNDEBUG("%s%d: len=%d!\n", ifp->if_name, ifp->if_unit, + uio->uio_resid); + return EIO; + } + tlen = uio->uio_resid; + + /* get a header mbuf */ + MGETHDR(m, M_DONTWAIT, MT_DATA); + if (m == NULL) + return ENOBUFS; + mlen = MHLEN; + + top = 0; + mp = ⊤ + while (error == 0 && uio->uio_resid > 0) { + m->m_len = min(mlen, uio->uio_resid); + error = uiomove(mtod (m, caddr_t), m->m_len, uio); + *mp = m; + mp = &m->m_next; + if (uio->uio_resid > 0) { + MGET (m, M_DONTWAIT, MT_DATA); + if (m == 0) { + error = ENOBUFS; + break; + } + mlen = MLEN; + } + } + if (error) { + if (top) + m_freem (top); + return error; + } + + top->m_pkthdr.len = tlen; + top->m_pkthdr.rcvif = ifp; + +#if NBPFILTER > 0 + if (tunctl[unit].tun_bpf) { + /* + * We need to prepend the address family as + * a four byte field. Cons up a dummy header + * to pacify bpf. This is safe because bpf + * will only read from the mbuf (i.e., it won't + * try to free it or keep a pointer to it). + */ + struct mbuf m; + u_int32_t af = AF_INET; + + m.m_next = top; + m.m_len = sizeof(af); + m.m_data = (char *)⁡ + + bpf_mtap(tunctl[unit].tun_bpf, &m); + } +#endif + + s = splimp(); + if (IF_QFULL (&ipintrq)) { + IF_DROP(&ipintrq); + splx(s); + ifp->if_collisions++; + m_freem(top); + return ENOBUFS; + } + IF_ENQUEUE(&ipintrq, top); + splx(s); + ifp->if_ipackets++; + schednetisr(NETISR_IP); + return error; +} + +/* + * tunselect - the select interface, this is only useful on reads + * really. The write detect always returns true, write never blocks + * anyway, it either accepts the packet or drops it. + */ +int +tunselect(dev, rw) + dev_t dev; + int rw; +{ + int unit = minor(dev), s; + struct tun_softc *tp = &tunctl[unit]; + struct ifnet *ifp = &tp->tun_if; + + s = splimp(); + TUNDEBUG("%s%d: tunselect\n", ifp->if_name, ifp->if_unit); + + switch (rw) { + case FREAD: + if (ifp->if_snd.ifq_len > 0) { + splx(s); + TUNDEBUG("%s%d: tunselect q=%d\n", ifp->if_name, + ifp->if_unit, ifp->if_snd.ifq_len); + return 1; + } + selrecord(curproc, &tp->tun_rsel); + break; + case FWRITE: + splx(s); + return 1; + } + splx(s); + TUNDEBUG("%s%d: tunselect waiting\n", ifp->if_name, ifp->if_unit); + return 0; +} + +#endif /* NTUN */ diff --git a/sys/net/if_tun.h b/sys/net/if_tun.h new file mode 100644 index 00000000000..116c10d60b7 --- /dev/null +++ b/sys/net/if_tun.h @@ -0,0 +1,51 @@ +/* $NetBSD: if_tun.h,v 1.5 1994/06/29 06:36:27 cgd Exp $ */ + +/* + * Copyright (c) 1988, Julian Onions <jpo@cs.nott.ac.uk> + * Nottingham University 1987. + * + * This source may be freely distributed, however I would be interested + * in any changes that are made. + * + * This driver takes packets off the IP i/f and hands them up to a + * user process to have it's wicked way with. This driver has it's + * roots in a similar driver written by Phil Cockcroft (formerly) at + * UCL. This driver is based much more on read/write/select mode of + * operation though. + * + * : $Header: if_tnreg.h,v 1.1.2.1 1992/07/16 22:39:16 friedl Exp + */ + +#ifndef _NET_IF_TUN_H_ +#define _NET_IF_TUN_H_ + +struct tun_softc { + u_short tun_flags; /* misc flags */ +#define TUN_OPEN 0x0001 +#define TUN_INITED 0x0002 +#define TUN_RCOLL 0x0004 +#define TUN_IASET 0x0008 +#define TUN_DSTADDR 0x0010 +#define TUN_RWAIT 0x0040 +#define TUN_ASYNC 0x0080 +#define TUN_NBIO 0x0100 + +#define TUN_READY (TUN_OPEN | TUN_INITED | TUN_IASET) + + struct ifnet tun_if; /* the interface */ + int tun_pgrp; /* the process group - if any */ + struct selinfo tun_rsel; /* read select */ + struct selinfo tun_wsel; /* write select (not used) */ +#if NBPFILTER > 0 + caddr_t tun_bpf; +#endif +}; + +/* Maximum packet size */ +#define TUNMTU 1500 + +/* ioctl's for get/set debug */ +#define TUNSDEBUG _IOW('t', 90, int) +#define TUNGDEBUG _IOR('t', 89, int) + +#endif /* !_NET_IF_TUN_H_ */ diff --git a/sys/net/if_types.h b/sys/net/if_types.h new file mode 100644 index 00000000000..fc3e749edac --- /dev/null +++ b/sys/net/if_types.h @@ -0,0 +1,97 @@ +/* $NetBSD: if_types.h,v 1.7 1995/02/27 09:10:24 glass Exp $ */ + +/* + * Copyright (c) 1989, 1993, 1994 + * 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)if_types.h 8.2 (Berkeley) 4/20/94 + */ + +/* + * Interface types for benefit of parsing media address headers. + * This list is derived from the SNMP list of ifTypes, currently + * documented in RFC1573. + */ + +#define IFT_OTHER 0x1 /* none of the following */ +#define IFT_1822 0x2 /* old-style arpanet imp */ +#define IFT_HDH1822 0x3 /* HDH arpanet imp */ +#define IFT_X25DDN 0x4 /* x25 to imp */ +#define IFT_X25 0x5 /* PDN X25 interface (RFC877) */ +#define IFT_ETHER 0x6 /* Ethernet CSMACD */ +#define IFT_ISO88023 0x7 /* CMSA CD */ +#define IFT_ISO88024 0x8 /* Token Bus */ +#define IFT_ISO88025 0x9 /* Token Ring */ +#define IFT_ISO88026 0xa /* MAN */ +#define IFT_STARLAN 0xb +#define IFT_P10 0xc /* Proteon 10MBit ring */ +#define IFT_P80 0xd /* Proteon 80MBit ring */ +#define IFT_HY 0xe /* Hyperchannel */ +#define IFT_FDDI 0xf +#define IFT_LAPB 0x10 +#define IFT_SDLC 0x11 +#define IFT_T1 0x12 +#define IFT_CEPT 0x13 /* E1 - european T1 */ +#define IFT_ISDNBASIC 0x14 +#define IFT_ISDNPRIMARY 0x15 +#define IFT_PTPSERIAL 0x16 /* Proprietary PTP serial */ +#define IFT_PPP 0x17 /* RFC 1331 */ +#define IFT_LOOP 0x18 /* loopback */ +#define IFT_EON 0x19 /* ISO over IP */ +#define IFT_XETHER 0x1a /* obsolete 3MB experimental ethernet */ +#define IFT_NSIP 0x1b /* XNS over IP */ +#define IFT_SLIP 0x1c /* IP over generic TTY */ +#define IFT_ULTRA 0x1d /* Ultra Technologies */ +#define IFT_DS3 0x1e /* Generic T3 */ +#define IFT_SIP 0x1f /* SMDS */ +#define IFT_FRELAY 0x20 /* Frame Relay DTE only */ +#define IFT_RS232 0x21 +#define IFT_PARA 0x22 /* parallel-port */ +#define IFT_ARCNET 0x23 +#define IFT_ARCNETPLUS 0x24 +#define IFT_ATM 0x25 /* ATM cells */ +#define IFT_MIOX25 0x26 +#define IFT_SONET 0x27 /* SONET or SDH */ +#define IFT_X25PLE 0x28 +#define IFT_ISO88022LLC 0x29 +#define IFT_LOCALTALK 0x2a +#define IFT_SMDSDXI 0x2b +#define IFT_FRELAYDCE 0x2c /* Frame Relay DCE */ +#define IFT_V35 0x2d +#define IFT_HSSI 0x2e +#define IFT_HIPPI 0x2f +#define IFT_MODEM 0x30 /* Generic Modem */ +#define IFT_AAL5 0x31 /* AAL5 over ATM */ +#define IFT_SONETPATH 0x32 +#define IFT_SONETVT 0x33 +#define IFT_SMDSICIP 0x34 /* SMDS InterCarrier Interface */ +#define IFT_PROPVIRTUAL 0x35 /* Proprietary Virtual/internal */ +#define IFT_PROPMUX 0x36 /* Proprietary Multiplexing */ diff --git a/sys/net/netisr.h b/sys/net/netisr.h new file mode 100644 index 00000000000..22e9a81a6cc --- /dev/null +++ b/sys/net/netisr.h @@ -0,0 +1,70 @@ +/* $NetBSD: netisr.h,v 1.12 1995/08/12 23:59:24 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 1986, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)netisr.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * The networking code runs off software interrupts. + * + * You can switch into the network by doing splsoftnet() and return by splx(). + * The software interrupt level for the network is higher than the software + * level for the clock (so you can enter the network in routines called + * at timeout time). + * + * The routine to request a network software interrupt, setsoftnet(), + * is defined in the machine-specific include files. + */ + +/* + * Each ``pup-level-1'' input queue has a bit in a ``netisr'' status + * word which is used to de-multiplex a single software + * interrupt used for scheduling the network code to calls + * on the lowest level routine of each protocol. + */ +#define NETISR_IP 2 /* same as AF_INET */ +#define NETISR_IMP 3 /* same as AF_IMPLINK */ +#define NETISR_NS 6 /* same as AF_NS */ +#define NETISR_ISO 7 /* same as AF_ISO */ +#define NETISR_CCITT 10 /* same as AF_CCITT */ +#define NETISR_ARP 18 /* same as AF_LINK */ +#define NETISR_PPP 26 /* for PPP processing */ + +#define schednetisr(anisr) { netisr |= 1<<(anisr); setsoftnet(); } + +#ifndef LOCORE +#ifdef _KERNEL +int netisr; /* scheduling bits for network */ +#endif +#endif diff --git a/sys/net/ppp-comp.h b/sys/net/ppp-comp.h new file mode 100644 index 00000000000..b3b2681ef70 --- /dev/null +++ b/sys/net/ppp-comp.h @@ -0,0 +1,138 @@ +/* $NetBSD: ppp-comp.h,v 1.1 1995/07/04 06:28:24 paulus Exp $ */ + +/* + * ppp-comp.h - Definitions for doing PPP packet compression. + * + * Copyright (c) 1994 The Australian National University. + * All rights reserved. + * + * Permission to use, copy, modify, and distribute this software and its + * documentation is hereby granted, provided that the above copyright + * notice appears in all copies. This software is provided without any + * warranty, express or implied. The Australian National University + * makes no representations about the suitability of this software for + * any purpose. + * + * IN NO EVENT SHALL THE AUSTRALIAN NATIONAL UNIVERSITY BE LIABLE TO ANY + * PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF + * THE AUSTRALIAN NATIONAL UNIVERSITY HAVE BEEN ADVISED OF THE POSSIBILITY + * OF SUCH DAMAGE. + * + * THE AUSTRALIAN NATIONAL UNIVERSITY SPECIFICALLY DISCLAIMS ANY WARRANTIES, + * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY + * AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS + * ON AN "AS IS" BASIS, AND THE AUSTRALIAN NATIONAL UNIVERSITY HAS NO + * OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, + * OR MODIFICATIONS. + * + * $Id: ppp-comp.h,v 1.1 1995/10/18 08:53:07 deraadt Exp $ + */ + +#ifndef _NET_PPP_COMP_H +#define _NET_PPP_COMP_H + +/* + * The following symbols control whether we include code for + * various compression methods. + */ +#ifndef DO_BSD_COMPRESS +#define DO_BSD_COMPRESS 1 /* by default, include BSD-Compress */ +#endif + +/* + * Structure giving methods for compression/decompression. + */ +#ifdef PACKETPTR +struct compressor { + int compress_proto; /* CCP compression protocol number */ + + /* Allocate space for a compressor (transmit side) */ + void *(*comp_alloc) __P((u_char *options, int opt_len)); + /* Free space used by a compressor */ + void (*comp_free) __P((void *state)); + /* Initialize a compressor */ + int (*comp_init) __P((void *state, u_char *options, int opt_len, + int unit, int hdrlen, int debug)); + /* Reset a compressor */ + void (*comp_reset) __P((void *state)); + /* Compress a packet */ + int (*compress) __P((void *state, PACKETPTR *mret, + PACKETPTR mp, int orig_len, int max_len)); + /* Return compression statistics */ + void (*comp_stat) __P((void *state, struct compstat *stats)); + + /* Allocate space for a decompressor (receive side) */ + void *(*decomp_alloc) __P((u_char *options, int opt_len)); + /* Free space used by a decompressor */ + void (*decomp_free) __P((void *state)); + /* Initialize a decompressor */ + int (*decomp_init) __P((void *state, u_char *options, int opt_len, + int unit, int hdrlen, int mru, int debug)); + /* Reset a decompressor */ + void (*decomp_reset) __P((void *state)); + /* Decompress a packet. */ + int (*decompress) __P((void *state, PACKETPTR mp, + PACKETPTR *dmpp)); + /* Update state for an incompressible packet received */ + void (*incomp) __P((void *state, PACKETPTR mp)); + /* Return decompression statistics */ + void (*decomp_stat) __P((void *state, struct compstat *stats)); +}; +#endif /* PACKETPTR */ + +/* + * Return values for decompress routine. + * We need to make these distinctions so that we can disable certain + * useful functionality, namely sending a CCP reset-request as a result + * of an error detected after decompression. This is to avoid infringing + * a patent held by Motorola. + * Don't you just lurve software patents. + */ +#define DECOMP_OK 0 /* everything went OK */ +#define DECOMP_ERROR 1 /* error detected before decomp. */ +#define DECOMP_FATALERROR 2 /* error detected after decomp. */ + +/* + * CCP codes. + */ +#define CCP_CONFREQ 1 +#define CCP_CONFACK 2 +#define CCP_TERMREQ 5 +#define CCP_TERMACK 6 +#define CCP_RESETREQ 14 +#define CCP_RESETACK 15 + +/* + * Max # bytes for a CCP option + */ +#define CCP_MAX_OPTION_LENGTH 32 + +/* + * Parts of a CCP packet. + */ +#define CCP_CODE(dp) ((dp)[0]) +#define CCP_ID(dp) ((dp)[1]) +#define CCP_LENGTH(dp) (((dp)[2] << 8) + (dp)[3]) +#define CCP_HDRLEN 4 + +#define CCP_OPT_CODE(dp) ((dp)[0]) +#define CCP_OPT_LENGTH(dp) ((dp)[1]) +#define CCP_OPT_MINLEN 2 + +/* + * Definitions for BSD-Compress. + */ +#define CI_BSD_COMPRESS 21 /* config. option for BSD-Compress */ +#define CILEN_BSD_COMPRESS 3 /* length of config. option */ + +/* Macros for handling the 3rd byte of the BSD-Compress config option. */ +#define BSD_NBITS(x) ((x) & 0x1F) /* number of bits requested */ +#define BSD_VERSION(x) ((x) >> 5) /* version of option format */ +#define BSD_CURRENT_VERSION 1 /* current version number */ +#define BSD_MAKE_OPT(v, n) (((v) << 5) | (n)) + +#define BSD_MIN_BITS 9 /* smallest code size supported */ +#define BSD_MAX_BITS 15 /* largest code size supported */ + +#endif /* _NET_PPP_COMP_H */ diff --git a/sys/net/ppp_defs.h b/sys/net/ppp_defs.h new file mode 100644 index 00000000000..41c63642dd5 --- /dev/null +++ b/sys/net/ppp_defs.h @@ -0,0 +1,160 @@ +/* $NetBSD: ppp_defs.h,v 1.1 1995/07/04 06:28:26 paulus Exp $ */ + +/* + * ppp_defs.h - PPP definitions. + * + * Copyright (c) 1994 The Australian National University. + * All rights reserved. + * + * Permission to use, copy, modify, and distribute this software and its + * documentation is hereby granted, provided that the above copyright + * notice appears in all copies. This software is provided without any + * warranty, express or implied. The Australian National University + * makes no representations about the suitability of this software for + * any purpose. + * + * IN NO EVENT SHALL THE AUSTRALIAN NATIONAL UNIVERSITY BE LIABLE TO ANY + * PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES + * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF + * THE AUSTRALIAN NATIONAL UNIVERSITY HAVE BEEN ADVISED OF THE POSSIBILITY + * OF SUCH DAMAGE. + * + * THE AUSTRALIAN NATIONAL UNIVERSITY SPECIFICALLY DISCLAIMS ANY WARRANTIES, + * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY + * AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS + * ON AN "AS IS" BASIS, AND THE AUSTRALIAN NATIONAL UNIVERSITY HAS NO + * OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, + * OR MODIFICATIONS. + */ + +#ifndef _PPP_DEFS_H_ +#define _PPP_DEFS_H_ + +/* + * The basic PPP frame. + */ +#define PPP_HDRLEN 4 /* octets for standard ppp header */ +#define PPP_FCSLEN 2 /* octets for FCS */ +#define PPP_MRU 1500 /* default MRU = max length of info field */ + +#define PPP_ADDRESS(p) (((u_char *)(p))[0]) +#define PPP_CONTROL(p) (((u_char *)(p))[1]) +#define PPP_PROTOCOL(p) ((((u_char *)(p))[2] << 8) + ((u_char *)(p))[3]) + +/* + * Significant octet values. + */ +#define PPP_ALLSTATIONS 0xff /* All-Stations broadcast address */ +#define PPP_UI 0x03 /* Unnumbered Information */ +#define PPP_FLAG 0x7e /* Flag Sequence */ +#define PPP_ESCAPE 0x7d /* Asynchronous Control Escape */ +#define PPP_TRANS 0x20 /* Asynchronous transparency modifier */ + +/* + * Protocol field values. + */ +#define PPP_IP 0x21 /* Internet Protocol */ +#define PPP_VJC_COMP 0x2d /* VJ compressed TCP */ +#define PPP_VJC_UNCOMP 0x2f /* VJ uncompressed TCP */ +#define PPP_COMP 0xfd /* compressed packet */ +#define PPP_IPCP 0x8021 /* IP Control Protocol */ +#define PPP_CCP 0x80fd /* Compression Control Protocol */ +#define PPP_LCP 0xc021 /* Link Control Protocol */ +#define PPP_PAP 0xc023 /* Password Authentication Protocol */ +#define PPP_LQR 0xc025 /* Link Quality Report protocol */ +#define PPP_CHAP 0xc223 /* Cryptographic Handshake Auth. Protocol */ + +/* + * Values for FCS calculations. + */ +#define PPP_INITFCS 0xffff /* Initial FCS value */ +#define PPP_GOODFCS 0xf0b8 /* Good final FCS value */ +#define PPP_FCS(fcs, c) (((fcs) >> 8) ^ fcstab[((fcs) ^ (c)) & 0xff]) + +/* + * A 32-bit unsigned integral type. + */ +#ifndef __BIT_TYPES_DEFINED__ +#ifdef UINT32_T +typedef UINT32_T u_int32_t; +#else +typedef unsigned int u_int32_t; +#endif +#endif + +/* + * Extended asyncmap - allows any character to be escaped. + */ +typedef u_int32_t ext_accm[8]; + +/* + * What to do with network protocol (NP) packets. + */ +enum NPmode { + NPMODE_PASS, /* pass the packet through */ + NPMODE_DROP, /* silently drop the packet */ + NPMODE_ERROR, /* return an error */ + NPMODE_QUEUE /* save it up for later. */ +}; + +/* + * Statistics. + */ +struct pppstat { + u_int ppp_ibytes; /* bytes received */ + u_int ppp_ipackets; /* packets received */ + u_int ppp_ierrors; /* receive errors */ + u_int ppp_obytes; /* bytes sent */ + u_int ppp_opackets; /* packets sent */ + u_int ppp_oerrors; /* transmit errors */ +}; + +struct vjstat { + u_int vjs_packets; /* outbound packets */ + u_int vjs_compressed; /* outbound compressed packets */ + u_int vjs_searches; /* searches for connection state */ + u_int vjs_misses; /* times couldn't find conn. state */ + u_int vjs_uncompressedin; /* inbound uncompressed packets */ + u_int vjs_compressedin; /* inbound compressed packets */ + u_int vjs_errorin; /* inbound unknown type packets */ + u_int vjs_tossed; /* inbound packets tossed because of error */ +}; + +struct ppp_stats { + struct pppstat p; /* basic PPP statistics */ + struct vjstat vj; /* VJ header compression statistics */ +}; + +struct compstat { + u_int unc_bytes; /* total uncompressed bytes */ + u_int unc_packets; /* total uncompressed packets */ + u_int comp_bytes; /* compressed bytes */ + u_int comp_packets; /* compressed packets */ + u_int inc_bytes; /* incompressible bytes */ + u_int inc_packets; /* incompressible packets */ + u_int ratio; /* recent compression ratio << 8 */ +}; + +struct ppp_comp_stats { + struct compstat c; /* packet compression statistics */ + struct compstat d; /* packet decompression statistics */ +}; + +/* + * The following structure records the time in seconds since + * the last NP packet was sent or received. + */ +struct ppp_idle { + time_t xmit_idle; /* time since last NP packet sent */ + time_t recv_idle; /* time since last NP packet received */ +}; + +#ifndef __P +#ifdef __STDC__ +#define __P(x) x +#else +#define __P(x) () +#endif +#endif + +#endif /* _PPP_DEFS_H_ */ diff --git a/sys/net/ppp_tty.c b/sys/net/ppp_tty.c new file mode 100644 index 00000000000..c484a5e22e7 --- /dev/null +++ b/sys/net/ppp_tty.c @@ -0,0 +1,1025 @@ +/* $NetBSD: ppp_tty.c,v 1.3 1995/10/05 05:55:14 mycroft Exp $ */ + +/* + * ppp_tty.c - Point-to-Point Protocol (PPP) driver for asynchronous + * tty devices. + * + * Copyright (c) 1989 Carnegie Mellon University. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by Carnegie Mellon University. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Drew D. Perkins + * Carnegie Mellon University + * 4910 Forbes Ave. + * Pittsburgh, PA 15213 + * (412) 268-8576 + * ddp@andrew.cmu.edu + * + * Based on: + * @(#)if_sl.c 7.6.1.2 (Berkeley) 2/15/89 + * + * Copyright (c) 1987 Regents of the University of California. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by the University of California, Berkeley. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Serial Line interface + * + * Rick Adams + * Center for Seismic Studies + * 1300 N 17th Street, Suite 1450 + * Arlington, Virginia 22209 + * (703)276-7900 + * rick@seismo.ARPA + * seismo!rick + * + * Pounded on heavily by Chris Torek (chris@mimsy.umd.edu, umcp-cs!chris). + * Converted to 4.3BSD Beta by Chris Torek. + * Other changes made at Berkeley, based in part on code by Kirk Smith. + * + * Converted to 4.3BSD+ 386BSD by Brad Parker (brad@cayman.com) + * Added VJ tcp header compression; more unified ioctls + * + * Extensively modified by Paul Mackerras (paulus@cs.anu.edu.au). + * Cleaned up a lot of the mbuf-related code to fix bugs that + * caused system crashes and packet corruption. Changed pppstart + * so that it doesn't just give up with a collision if the whole + * packet doesn't fit in the output ring buffer. + * + * Added priority queueing for interactive IP packets, following + * the model of if_sl.c, plus hooks for bpf. + * Paul Mackerras (paulus@cs.anu.edu.au). + */ + +/* from if_sl.c,v 1.11 84/10/04 12:54:47 rick Exp */ +/* from NetBSD: if_ppp.c,v 1.15.2.2 1994/07/28 05:17:58 cgd Exp */ + +#include "ppp.h" +#if NPPP > 0 + +#define VJC +#define PPP_COMPRESS + +#include <sys/param.h> +#include <sys/proc.h> +#include <sys/mbuf.h> +#include <sys/dkstat.h> +#include <sys/socket.h> +#include <sys/ioctl.h> +#include <sys/file.h> +#include <sys/tty.h> +#include <sys/kernel.h> +#include <sys/conf.h> +#include <sys/vnode.h> + +#include <net/if.h> +#include <net/if_types.h> + +#ifdef VJC +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/ip.h> +#include <net/slcompress.h> +#endif + +#include <net/ppp_defs.h> +#include <net/if_ppp.h> +#include <net/if_pppvar.h> + +int pppopen __P((dev_t dev, struct tty *tp)); +int pppclose __P((struct tty *tp, int flag)); +int pppread __P((struct tty *tp, struct uio *uio, int flag)); +int pppwrite __P((struct tty *tp, struct uio *uio, int flag)); +int ppptioctl __P((struct tty *tp, u_long cmd, caddr_t data, int flag, + struct proc *)); +int pppinput __P((int c, struct tty *tp)); +int pppstart __P((struct tty *tp)); + +static u_int16_t pppfcs __P((u_int16_t fcs, u_char *cp, int len)); +static void pppasyncstart __P((struct ppp_softc *)); +static void pppasyncctlp __P((struct ppp_softc *)); +static void pppasyncrelinq __P((struct ppp_softc *)); +static void ppp_timeout __P((void *)); +static void pppgetm __P((struct ppp_softc *sc)); +static void pppdumpb __P((u_char *b, int l)); +static void ppplogchar __P((struct ppp_softc *, int)); + +/* + * Some useful mbuf macros not in mbuf.h. + */ +#define M_IS_CLUSTER(m) ((m)->m_flags & M_EXT) + +#define M_DATASTART(m) \ + (M_IS_CLUSTER(m) ? (m)->m_ext.ext_buf : \ + (m)->m_flags & M_PKTHDR ? (m)->m_pktdat : (m)->m_dat) + +#define M_DATASIZE(m) \ + (M_IS_CLUSTER(m) ? (m)->m_ext.ext_size : \ + (m)->m_flags & M_PKTHDR ? MHLEN: MLEN) + +/* + * Does c need to be escaped? + */ +#define ESCAPE_P(c) (sc->sc_asyncmap[(c) >> 5] & (1 << ((c) & 0x1F))) + +/* + * Procedures for using an async tty interface for PPP. + */ + +/* This is a NetBSD-1.0 or later kernel. */ +#define CCOUNT(q) ((q)->c_cc) + +#define PPP_HIWAT 400 /* Don't start a new packet if HIWAT on que */ + +/* + * Line specific open routine for async tty devices. + * Attach the given tty to the first available ppp unit. + * Called from device open routine or ttioctl. + */ +/* ARGSUSED */ +int +pppopen(dev, tp) + dev_t dev; + register struct tty *tp; +{ + struct proc *p = curproc; /* XXX */ + register struct ppp_softc *sc; + int error, s, i; + + if (error = suser(p->p_ucred, &p->p_acflag)) + return (error); + + s = spltty(); + + if (tp->t_line == PPPDISC) { + sc = (struct ppp_softc *) tp->t_sc; + if (sc != NULL && sc->sc_devp == (void *) tp) { + splx(s); + return (0); + } + } + + if ((sc = pppalloc(p->p_pid)) == NULL) { + splx(s); + return ENXIO; + } + + if (sc->sc_relinq) + (*sc->sc_relinq)(sc); /* get previous owner to relinquish the unit */ + + sc->sc_ilen = 0; + sc->sc_m = NULL; + bzero(sc->sc_asyncmap, sizeof(sc->sc_asyncmap)); + sc->sc_asyncmap[0] = 0xffffffff; + sc->sc_asyncmap[3] = 0x60000000; + sc->sc_rasyncmap = 0; + sc->sc_devp = (void *) tp; + sc->sc_start = pppasyncstart; + sc->sc_ctlp = pppasyncctlp; + sc->sc_relinq = pppasyncrelinq; + sc->sc_outm = NULL; + pppgetm(sc); + sc->sc_if.if_flags |= IFF_RUNNING; + sc->sc_if.if_baudrate = tp->t_ospeed; + + tp->t_sc = (caddr_t) sc; + ttyflush(tp, FREAD | FWRITE); + + splx(s); + return (0); +} + +/* + * Line specific close routine, called from device close routine + * and from ttioctl. + * Detach the tty from the ppp unit. + * Mimics part of ttyclose(). + */ +int +pppclose(tp, flag) + struct tty *tp; + int flag; +{ + register struct ppp_softc *sc; + struct mbuf *m; + int s; + + s = spltty(); + ttywflush(tp); + tp->t_line = 0; + sc = (struct ppp_softc *) tp->t_sc; + if (sc != NULL) { + tp->t_sc = NULL; + if (tp == (struct tty *) sc->sc_devp) { + pppasyncrelinq(sc); + pppdealloc(sc); + } + } + splx(s); + return 0; +} + +/* + * Relinquish the interface unit to another device. + */ +static void +pppasyncrelinq(sc) + struct ppp_softc *sc; +{ + int s; + + s = spltty(); + if (sc->sc_outm) { + m_freem(sc->sc_outm); + sc->sc_outm = NULL; + } + if (sc->sc_m) { + m_freem(sc->sc_m); + sc->sc_m = NULL; + } + if (sc->sc_flags & SC_TIMEOUT) { + untimeout(ppp_timeout, (void *) sc); + sc->sc_flags &= ~SC_TIMEOUT; + } + splx(s); +} + +/* + * Line specific (tty) read routine. + */ +int +pppread(tp, uio, flag) + register struct tty *tp; + struct uio *uio; + int flag; +{ + register struct ppp_softc *sc = (struct ppp_softc *)tp->t_sc; + struct mbuf *m, *m0; + register int s; + int error = 0; + + if (sc == NULL) + return 0; + /* + * Loop waiting for input, checking that nothing disasterous + * happens in the meantime. + */ + s = spltty(); + for (;;) { + if (tp != (struct tty *) sc->sc_devp || tp->t_line != PPPDISC) { + splx(s); + return 0; + } + if (sc->sc_inq.ifq_head != NULL) + break; + if ((tp->t_state & TS_CARR_ON) == 0 && (tp->t_cflag & CLOCAL) == 0 + && (tp->t_state & TS_ISOPEN)) { + splx(s); + return 0; /* end of file */ + } + if (tp->t_state & TS_ASYNC || flag & IO_NDELAY) { + splx(s); + return (EWOULDBLOCK); + } + error = ttysleep(tp, (caddr_t)&tp->t_rawq, TTIPRI|PCATCH, ttyin, 0); + if (error) { + splx(s); + return error; + } + } + + /* Pull place-holder byte out of canonical queue */ + getc(&tp->t_canq); + + /* Get the packet from the input queue */ + IF_DEQUEUE(&sc->sc_inq, m0); + splx(s); + + for (m = m0; m && uio->uio_resid; m = m->m_next) + if (error = uiomove(mtod(m, u_char *), m->m_len, uio)) + break; + m_freem(m0); + return (error); +} + +/* + * Line specific (tty) write routine. + */ +int +pppwrite(tp, uio, flag) + register struct tty *tp; + struct uio *uio; + int flag; +{ + register struct ppp_softc *sc = (struct ppp_softc *)tp->t_sc; + struct mbuf *m, *m0, **mp; + struct sockaddr dst; + int len, error; + + if ((tp->t_state & TS_CARR_ON) == 0 && (tp->t_cflag & CLOCAL) == 0) + return 0; /* wrote 0 bytes */ + if (tp->t_line != PPPDISC) + return (EINVAL); + if (sc == NULL || tp != (struct tty *) sc->sc_devp) + return EIO; + if (uio->uio_resid > sc->sc_if.if_mtu + PPP_HDRLEN || + uio->uio_resid < PPP_HDRLEN) + return (EMSGSIZE); + for (mp = &m0; uio->uio_resid; mp = &m->m_next) { + MGET(m, M_WAIT, MT_DATA); + if ((*mp = m) == NULL) { + m_freem(m0); + return (ENOBUFS); + } + m->m_len = 0; + if (uio->uio_resid >= MCLBYTES / 2) + MCLGET(m, M_DONTWAIT); + len = M_TRAILINGSPACE(m); + if (len > uio->uio_resid) + len = uio->uio_resid; + if (error = uiomove(mtod(m, u_char *), len, uio)) { + m_freem(m0); + return (error); + } + m->m_len = len; + } + dst.sa_family = AF_UNSPEC; + bcopy(mtod(m0, u_char *), dst.sa_data, PPP_HDRLEN); + m0->m_data += PPP_HDRLEN; + m0->m_len -= PPP_HDRLEN; + return (pppoutput(&sc->sc_if, m0, &dst, (struct rtentry *)0)); +} + +/* + * Line specific (tty) ioctl routine. + * This discipline requires that tty device drivers call + * the line specific l_ioctl routine from their ioctl routines. + */ +/* ARGSUSED */ +int +ppptioctl(tp, cmd, data, flag, p) + struct tty *tp; + u_long cmd; + caddr_t data; + int flag; + struct proc *p; +{ + struct ppp_softc *sc = (struct ppp_softc *) tp->t_sc; + int error, s; + + if (sc == NULL || tp != (struct tty *) sc->sc_devp) + return -1; + + error = 0; + switch (cmd) { + case PPPIOCSASYNCMAP: + if (error = suser(p->p_ucred, &p->p_acflag)) + break; + sc->sc_asyncmap[0] = *(u_int *)data; + break; + + case PPPIOCGASYNCMAP: + *(u_int *)data = sc->sc_asyncmap[0]; + break; + + case PPPIOCSRASYNCMAP: + if (error = suser(p->p_ucred, &p->p_acflag)) + break; + sc->sc_rasyncmap = *(u_int *)data; + break; + + case PPPIOCGRASYNCMAP: + *(u_int *)data = sc->sc_rasyncmap; + break; + + case PPPIOCSXASYNCMAP: + if (error = suser(p->p_ucred, &p->p_acflag)) + break; + s = spltty(); + bcopy(data, sc->sc_asyncmap, sizeof(sc->sc_asyncmap)); + sc->sc_asyncmap[1] = 0; /* mustn't escape 0x20 - 0x3f */ + sc->sc_asyncmap[2] &= ~0x40000000; /* mustn't escape 0x5e */ + sc->sc_asyncmap[3] |= 0x60000000; /* must escape 0x7d, 0x7e */ + splx(s); + break; + + case PPPIOCGXASYNCMAP: + bcopy(sc->sc_asyncmap, data, sizeof(sc->sc_asyncmap)); + break; + + default: + error = pppioctl(sc, cmd, data, flag, p); + if (error == 0 && cmd == PPPIOCSMRU) + pppgetm(sc); + } + + return error; +} + +/* + * FCS lookup table as calculated by genfcstab. + */ +static u_int16_t fcstab[256] = { + 0x0000, 0x1189, 0x2312, 0x329b, 0x4624, 0x57ad, 0x6536, 0x74bf, + 0x8c48, 0x9dc1, 0xaf5a, 0xbed3, 0xca6c, 0xdbe5, 0xe97e, 0xf8f7, + 0x1081, 0x0108, 0x3393, 0x221a, 0x56a5, 0x472c, 0x75b7, 0x643e, + 0x9cc9, 0x8d40, 0xbfdb, 0xae52, 0xdaed, 0xcb64, 0xf9ff, 0xe876, + 0x2102, 0x308b, 0x0210, 0x1399, 0x6726, 0x76af, 0x4434, 0x55bd, + 0xad4a, 0xbcc3, 0x8e58, 0x9fd1, 0xeb6e, 0xfae7, 0xc87c, 0xd9f5, + 0x3183, 0x200a, 0x1291, 0x0318, 0x77a7, 0x662e, 0x54b5, 0x453c, + 0xbdcb, 0xac42, 0x9ed9, 0x8f50, 0xfbef, 0xea66, 0xd8fd, 0xc974, + 0x4204, 0x538d, 0x6116, 0x709f, 0x0420, 0x15a9, 0x2732, 0x36bb, + 0xce4c, 0xdfc5, 0xed5e, 0xfcd7, 0x8868, 0x99e1, 0xab7a, 0xbaf3, + 0x5285, 0x430c, 0x7197, 0x601e, 0x14a1, 0x0528, 0x37b3, 0x263a, + 0xdecd, 0xcf44, 0xfddf, 0xec56, 0x98e9, 0x8960, 0xbbfb, 0xaa72, + 0x6306, 0x728f, 0x4014, 0x519d, 0x2522, 0x34ab, 0x0630, 0x17b9, + 0xef4e, 0xfec7, 0xcc5c, 0xddd5, 0xa96a, 0xb8e3, 0x8a78, 0x9bf1, + 0x7387, 0x620e, 0x5095, 0x411c, 0x35a3, 0x242a, 0x16b1, 0x0738, + 0xffcf, 0xee46, 0xdcdd, 0xcd54, 0xb9eb, 0xa862, 0x9af9, 0x8b70, + 0x8408, 0x9581, 0xa71a, 0xb693, 0xc22c, 0xd3a5, 0xe13e, 0xf0b7, + 0x0840, 0x19c9, 0x2b52, 0x3adb, 0x4e64, 0x5fed, 0x6d76, 0x7cff, + 0x9489, 0x8500, 0xb79b, 0xa612, 0xd2ad, 0xc324, 0xf1bf, 0xe036, + 0x18c1, 0x0948, 0x3bd3, 0x2a5a, 0x5ee5, 0x4f6c, 0x7df7, 0x6c7e, + 0xa50a, 0xb483, 0x8618, 0x9791, 0xe32e, 0xf2a7, 0xc03c, 0xd1b5, + 0x2942, 0x38cb, 0x0a50, 0x1bd9, 0x6f66, 0x7eef, 0x4c74, 0x5dfd, + 0xb58b, 0xa402, 0x9699, 0x8710, 0xf3af, 0xe226, 0xd0bd, 0xc134, + 0x39c3, 0x284a, 0x1ad1, 0x0b58, 0x7fe7, 0x6e6e, 0x5cf5, 0x4d7c, + 0xc60c, 0xd785, 0xe51e, 0xf497, 0x8028, 0x91a1, 0xa33a, 0xb2b3, + 0x4a44, 0x5bcd, 0x6956, 0x78df, 0x0c60, 0x1de9, 0x2f72, 0x3efb, + 0xd68d, 0xc704, 0xf59f, 0xe416, 0x90a9, 0x8120, 0xb3bb, 0xa232, + 0x5ac5, 0x4b4c, 0x79d7, 0x685e, 0x1ce1, 0x0d68, 0x3ff3, 0x2e7a, + 0xe70e, 0xf687, 0xc41c, 0xd595, 0xa12a, 0xb0a3, 0x8238, 0x93b1, + 0x6b46, 0x7acf, 0x4854, 0x59dd, 0x2d62, 0x3ceb, 0x0e70, 0x1ff9, + 0xf78f, 0xe606, 0xd49d, 0xc514, 0xb1ab, 0xa022, 0x92b9, 0x8330, + 0x7bc7, 0x6a4e, 0x58d5, 0x495c, 0x3de3, 0x2c6a, 0x1ef1, 0x0f78 +}; + +/* + * Calculate a new FCS given the current FCS and the new data. + */ +static u_int16_t +pppfcs(fcs, cp, len) + register u_int16_t fcs; + register u_char *cp; + register int len; +{ + while (len--) + fcs = PPP_FCS(fcs, *cp++); + return (fcs); +} + +/* + * This gets called from pppoutput when a new packet is + * put on a queue, at splsoftnet. + */ +static void +pppasyncstart(sc) + register struct ppp_softc *sc; +{ + register struct tty *tp = (struct tty *) sc->sc_devp; + int s; + + s = spltty(); + pppstart(tp); + splx(s); +} + +/* + * This gets called when a received packet is placed on + * the inq, at splsoftnet. + */ +static void +pppasyncctlp(sc) + struct ppp_softc *sc; +{ + struct tty *tp; + int s; + + /* Put a placeholder byte in canq for ttselect()/ttnread(). */ + s = spltty(); + tp = (struct tty *) sc->sc_devp; + putc(0, &tp->t_canq); + ttwakeup(tp); + splx(s); +} + +/* + * Start output on async tty interface. Get another datagram + * to send from the interface queue and start sending it. + * Called at spltty or higher. + */ +int +pppstart(tp) + register struct tty *tp; +{ + register struct ppp_softc *sc = (struct ppp_softc *) tp->t_sc; + register struct mbuf *m; + register int len; + register u_char *start, *stop, *cp; + int n, s, ndone, done, idle; + struct mbuf *m2; + + if ((tp->t_state & TS_CARR_ON) == 0 && (tp->t_cflag & CLOCAL) == 0 + || sc == NULL || tp != (struct tty *) sc->sc_devp) { + if (tp->t_oproc != NULL) + (*tp->t_oproc)(tp); + return 0; + } + + idle = 0; + while (CCOUNT(&tp->t_outq) < PPP_HIWAT) { + /* + * See if we have an existing packet partly sent. + * If not, get a new packet and start sending it. + */ + m = sc->sc_outm; + if (m == NULL) { + /* + * Get another packet to be sent. + */ + m = ppp_dequeue(sc); + if (m == NULL) { + idle = 1; + break; + } + + /* + * The extra PPP_FLAG will start up a new packet, and thus + * will flush any accumulated garbage. We do this whenever + * the line may have been idle for some time. + */ + if (CCOUNT(&tp->t_outq) == 0) { + ++sc->sc_bytessent; + (void) putc(PPP_FLAG, &tp->t_outq); + } + + /* Calculate the FCS for the first mbuf's worth. */ + sc->sc_outfcs = pppfcs(PPP_INITFCS, mtod(m, u_char *), m->m_len); + sc->sc_if.if_lastchange = time; + } + + for (;;) { + start = mtod(m, u_char *); + len = m->m_len; + stop = start + len; + while (len > 0) { + /* + * Find out how many bytes in the string we can + * handle without doing something special. + */ + for (cp = start; cp < stop; cp++) + if (ESCAPE_P(*cp)) + break; + n = cp - start; + if (n) { + /* NetBSD (0.9 or later), 4.3-Reno or similar. */ + ndone = n - b_to_q(start, n, &tp->t_outq); + len -= ndone; + start += ndone; + sc->sc_bytessent += ndone; + + if (ndone < n) + break; /* packet doesn't fit */ + } + /* + * If there are characters left in the mbuf, + * the first one must be special.. + * Put it out in a different form. + */ + if (len) { + if (putc(PPP_ESCAPE, &tp->t_outq)) + break; + if (putc(*start ^ PPP_TRANS, &tp->t_outq)) { + (void) unputc(&tp->t_outq); + break; + } + sc->sc_bytessent += 2; + start++; + len--; + } + } + + /* + * If we didn't empty this mbuf, remember where we're up to. + * If we emptied the last mbuf, try to add the FCS and closing + * flag, and if we can't, leave sc_outm pointing to m, but with + * m->m_len == 0, to remind us to output the FCS and flag later. + */ + done = len == 0; + if (done && m->m_next == NULL) { + u_char *p, *q; + int c; + u_char endseq[8]; + + /* + * We may have to escape the bytes in the FCS. + */ + p = endseq; + c = ~sc->sc_outfcs & 0xFF; + if (ESCAPE_P(c)) { + *p++ = PPP_ESCAPE; + *p++ = c ^ PPP_TRANS; + } else + *p++ = c; + c = (~sc->sc_outfcs >> 8) & 0xFF; + if (ESCAPE_P(c)) { + *p++ = PPP_ESCAPE; + *p++ = c ^ PPP_TRANS; + } else + *p++ = c; + *p++ = PPP_FLAG; + + /* + * Try to output the FCS and flag. If the bytes + * don't all fit, back out. + */ + for (q = endseq; q < p; ++q) + if (putc(*q, &tp->t_outq)) { + done = 0; + for (; q > endseq; --q) + unputc(&tp->t_outq); + break; + } + sc->sc_bytessent += q - endseq; + } + + if (!done) { + /* remember where we got to */ + m->m_data = start; + m->m_len = len; + break; + } + + /* Finished with this mbuf; free it and move on. */ + MFREE(m, m2); + m = m2; + if (m == NULL) { + /* Finished a packet */ + sc->sc_if.if_opackets++; + sc->sc_if.if_obytes = sc->sc_bytessent; + break; + } + sc->sc_outfcs = pppfcs(sc->sc_outfcs, mtod(m, u_char *), m->m_len); + } + + /* + * Here we have either finished a packet (m == NULL) + * or filled up the output queue (m != NULL). + */ + sc->sc_outm = m; + if (m) + break; + } + + /* + * If there is stuff in the output queue, send it now. + * We are being called in lieu of ttstart and must do what it would. + */ + if (tp->t_oproc != NULL) + (*tp->t_oproc)(tp); + + /* + * This timeout is needed for operation on a pseudo-tty, + * because the pty code doesn't call pppstart after it has + * drained the t_outq. + */ + if (!idle && (sc->sc_flags & SC_TIMEOUT) == 0) { + timeout(ppp_timeout, (void *) sc, 1); + sc->sc_flags |= SC_TIMEOUT; + } + + return 0; +} + +/* + * Timeout routine - try to start some more output. + */ +static void +ppp_timeout(x) + void *x; +{ + struct ppp_softc *sc = (struct ppp_softc *) x; + struct tty *tp = (struct tty *) sc->sc_devp; + int s; + + s = spltty(); + sc->sc_flags &= ~SC_TIMEOUT; + pppstart(tp); + splx(s); +} + +/* + * Allocate enough mbuf to handle current MRU. + */ +static void +pppgetm(sc) + register struct ppp_softc *sc; +{ + struct mbuf *m, **mp; + int len; + int s; + + s = spltty(); + mp = &sc->sc_m; + for (len = sc->sc_mru + PPP_HDRLEN + PPP_FCSLEN; len > 0; ){ + if ((m = *mp) == NULL) { + MGETHDR(m, M_DONTWAIT, MT_DATA); + if (m == NULL) + break; + *mp = m; + MCLGET(m, M_DONTWAIT); + } + len -= M_DATASIZE(m); + mp = &m->m_next; + } + splx(s); +} + +/* + * tty interface receiver interrupt. + */ +static unsigned paritytab[8] = { + 0x96696996, 0x69969669, 0x69969669, 0x96696996, + 0x69969669, 0x96696996, 0x96696996, 0x69969669 +}; + +int +pppinput(c, tp) + int c; + register struct tty *tp; +{ + register struct ppp_softc *sc; + struct mbuf *m; + int ilen, s; + + sc = (struct ppp_softc *) tp->t_sc; + if (sc == NULL || tp != (struct tty *) sc->sc_devp) + return 0; + + s = spltty(); /* should be unnecessary */ + ++tk_nin; + ++sc->sc_bytesrcvd; + + if (c & TTY_FE) { + /* framing error or overrun on this char - abort packet */ + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: bad char %x\n", sc->sc_if.if_unit, c); + goto flush; + } + + c &= 0xff; + + if (c & 0x80) + sc->sc_flags |= SC_RCV_B7_1; + else + sc->sc_flags |= SC_RCV_B7_0; + if (paritytab[c >> 5] & (1 << (c & 0x1F))) + sc->sc_flags |= SC_RCV_ODDP; + else + sc->sc_flags |= SC_RCV_EVNP; + + if (sc->sc_flags & SC_LOG_RAWIN) + ppplogchar(sc, c); + + if (c == PPP_FLAG) { + ilen = sc->sc_ilen; + sc->sc_ilen = 0; + sc->sc_if.if_ibytes = sc->sc_bytesrcvd; + + if (sc->sc_rawin_count > 0) + ppplogchar(sc, -1); + + /* + * If SC_ESCAPED is set, then we've seen the packet + * abort sequence "}~". + */ + if (sc->sc_flags & (SC_FLUSH | SC_ESCAPED) + || ilen > 0 && sc->sc_fcs != PPP_GOODFCS) { + sc->sc_flags |= SC_PKTLOST; /* note the dropped packet */ + if ((sc->sc_flags & (SC_FLUSH | SC_ESCAPED)) == 0){ + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: bad fcs %x\n", sc->sc_if.if_unit, + sc->sc_fcs); + sc->sc_if.if_ierrors++; + } else + sc->sc_flags &= ~(SC_FLUSH | SC_ESCAPED); + splx(s); + return 0; + } + + if (ilen < PPP_HDRLEN + PPP_FCSLEN) { + if (ilen) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: too short (%d)\n", sc->sc_if.if_unit, ilen); + sc->sc_if.if_ierrors++; + sc->sc_flags |= SC_PKTLOST; + } + splx(s); + return 0; + } + + /* + * Remove FCS trailer. Somewhat painful... + */ + ilen -= 2; + if (--sc->sc_mc->m_len == 0) { + for (m = sc->sc_m; m->m_next != sc->sc_mc; m = m->m_next) + ; + sc->sc_mc = m; + } + sc->sc_mc->m_len--; + + /* excise this mbuf chain */ + m = sc->sc_m; + sc->sc_m = sc->sc_mc->m_next; + sc->sc_mc->m_next = NULL; + + ppppktin(sc, m, sc->sc_flags & SC_PKTLOST); + sc->sc_flags &= ~SC_PKTLOST; + + pppgetm(sc); + splx(s); + return 0; + } + + if (sc->sc_flags & SC_FLUSH) { + if (sc->sc_flags & SC_LOG_FLUSH) + ppplogchar(sc, c); + splx(s); + return 0; + } + + if (c < 0x20 && (sc->sc_rasyncmap & (1 << c))) { + splx(s); + return 0; + } + + if (sc->sc_flags & SC_ESCAPED) { + sc->sc_flags &= ~SC_ESCAPED; + c ^= PPP_TRANS; + } else if (c == PPP_ESCAPE) { + sc->sc_flags |= SC_ESCAPED; + splx(s); + return 0; + } + + /* + * Initialize buffer on first octet received. + * First octet could be address or protocol (when compressing + * address/control). + * Second octet is control. + * Third octet is first or second (when compressing protocol) + * octet of protocol. + * Fourth octet is second octet of protocol. + */ + if (sc->sc_ilen == 0) { + /* reset the first input mbuf */ + if (sc->sc_m == NULL) { + pppgetm(sc); + if (sc->sc_m == NULL) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: no input mbufs!\n", sc->sc_if.if_unit); + goto flush; + } + } + m = sc->sc_m; + m->m_len = 0; + m->m_data = M_DATASTART(sc->sc_m); + sc->sc_mc = m; + sc->sc_mp = mtod(m, char *); + sc->sc_fcs = PPP_INITFCS; + if (c != PPP_ALLSTATIONS) { + if (sc->sc_flags & SC_REJ_COMP_AC) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: garbage received: 0x%x (need 0xFF)\n", + sc->sc_if.if_unit, c); + goto flush; + } + *sc->sc_mp++ = PPP_ALLSTATIONS; + *sc->sc_mp++ = PPP_UI; + sc->sc_ilen += 2; + m->m_len += 2; + } + } + if (sc->sc_ilen == 1 && c != PPP_UI) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: missing UI (0x3), got 0x%x\n", + sc->sc_if.if_unit, c); + goto flush; + } + if (sc->sc_ilen == 2 && (c & 1) == 1) { + /* a compressed protocol */ + *sc->sc_mp++ = 0; + sc->sc_ilen++; + sc->sc_mc->m_len++; + } + if (sc->sc_ilen == 3 && (c & 1) == 0) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: bad protocol %x\n", sc->sc_if.if_unit, + (sc->sc_mp[-1] << 8) + c); + goto flush; + } + + /* packet beyond configured mru? */ + if (++sc->sc_ilen > sc->sc_mru + PPP_HDRLEN + PPP_FCSLEN) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: packet too big\n", sc->sc_if.if_unit); + goto flush; + } + + /* is this mbuf full? */ + m = sc->sc_mc; + if (M_TRAILINGSPACE(m) <= 0) { + if (m->m_next == NULL) { + pppgetm(sc); + if (m->m_next == NULL) { + if (sc->sc_flags & SC_DEBUG) + printf("ppp%d: too few input mbufs!\n", sc->sc_if.if_unit); + goto flush; + } + } + sc->sc_mc = m = m->m_next; + m->m_len = 0; + m->m_data = M_DATASTART(m); + sc->sc_mp = mtod(m, char *); + } + + ++m->m_len; + *sc->sc_mp++ = c; + sc->sc_fcs = PPP_FCS(sc->sc_fcs, c); + splx(s); + return 0; + + flush: + if (!(sc->sc_flags & SC_FLUSH)) { + sc->sc_if.if_ierrors++; + sc->sc_flags |= SC_FLUSH; + if (sc->sc_flags & SC_LOG_FLUSH) + ppplogchar(sc, c); + } + splx(s); + return 0; +} + +#define MAX_DUMP_BYTES 128 + +static void +ppplogchar(sc, c) + struct ppp_softc *sc; + int c; +{ + if (c >= 0) + sc->sc_rawin[sc->sc_rawin_count++] = c; + if (sc->sc_rawin_count >= sizeof(sc->sc_rawin) + || c < 0 && sc->sc_rawin_count > 0) { + printf("ppp%d input: ", sc->sc_if.if_unit); + pppdumpb(sc->sc_rawin, sc->sc_rawin_count); + sc->sc_rawin_count = 0; + } +} + +static void +pppdumpb(b, l) + u_char *b; + int l; +{ + char buf[3*MAX_DUMP_BYTES+4]; + char *bp = buf; + static char digits[] = "0123456789abcdef"; + + while (l--) { + if (bp >= buf + sizeof(buf) - 3) { + *bp++ = '>'; + break; + } + *bp++ = digits[*b >> 4]; /* convert byte to ascii hex */ + *bp++ = digits[*b++ & 0xf]; + *bp++ = ' '; + } + + *bp = 0; + printf("%s\n", buf); +} + +#endif /* NPPP > 0 */ diff --git a/sys/net/radix.c b/sys/net/radix.c new file mode 100644 index 00000000000..ea7cb1828dd --- /dev/null +++ b/sys/net/radix.c @@ -0,0 +1,893 @@ +/* $NetBSD: radix.c,v 1.9 1995/05/17 15:50:06 mycroft Exp $ */ + +/* + * Copyright (c) 1988, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)radix.c 8.4 (Berkeley) 11/2/94 + */ + +/* + * Routines to build and maintain radix trees for routing lookups. + */ +#include <sys/param.h> +#ifdef _KERNEL +#include <sys/systm.h> +#include <sys/malloc.h> +#define M_DONTWAIT M_NOWAIT +#include <sys/domain.h> +#else +#include <stdlib.h> +#endif +#include <sys/syslog.h> +#include <net/radix.h> + +int max_keylen; +struct radix_mask *rn_mkfreelist; +struct radix_node_head *mask_rnhead; +static char *addmask_key; +static char normal_chars[] = {0, 0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc, 0xfe, -1}; +static char *rn_zeros, *rn_ones; + +#define rn_masktop (mask_rnhead->rnh_treetop) +#undef Bcmp +#define Bcmp(a, b, l) (l == 0 ? 0 : bcmp((caddr_t)(a), (caddr_t)(b), (u_long)l)) +/* + * The data structure for the keys is a radix tree with one way + * branching removed. The index rn_b at an internal node n represents a bit + * position to be tested. The tree is arranged so that all descendants + * of a node n have keys whose bits all agree up to position rn_b - 1. + * (We say the index of n is rn_b.) + * + * There is at least one descendant which has a one bit at position rn_b, + * and at least one with a zero there. + * + * A route is determined by a pair of key and mask. We require that the + * bit-wise logical and of the key and mask to be the key. + * We define the index of a route to associated with the mask to be + * the first bit number in the mask where 0 occurs (with bit number 0 + * representing the highest order bit). + * + * We say a mask is normal if every bit is 0, past the index of the mask. + * If a node n has a descendant (k, m) with index(m) == index(n) == rn_b, + * and m is a normal mask, then the route applies to every descendant of n. + * If the index(m) < rn_b, this implies the trailing last few bits of k + * before bit b are all 0, (and hence consequently true of every descendant + * of n), so the route applies to all descendants of the node as well. + * + * Similar logic shows that a non-normal mask m such that + * index(m) <= index(n) could potentially apply to many children of n. + * Thus, for each non-host route, we attach its mask to a list at an internal + * node as high in the tree as we can go. + * + * The present version of the code makes use of normal routes in short- + * circuiting an explict mask and compare operation when testing whether + * a key satisfies a normal route, and also in remembering the unique leaf + * that governs a subtree. + */ + +struct radix_node * +rn_search(v_arg, head) + void *v_arg; + struct radix_node *head; +{ + register struct radix_node *x; + register caddr_t v; + + for (x = head, v = v_arg; x->rn_b >= 0;) { + if (x->rn_bmask & v[x->rn_off]) + x = x->rn_r; + else + x = x->rn_l; + } + return (x); +}; + +struct radix_node * +rn_search_m(v_arg, head, m_arg) + struct radix_node *head; + void *v_arg, *m_arg; +{ + register struct radix_node *x; + register caddr_t v = v_arg, m = m_arg; + + for (x = head; x->rn_b >= 0;) { + if ((x->rn_bmask & m[x->rn_off]) && + (x->rn_bmask & v[x->rn_off])) + x = x->rn_r; + else + x = x->rn_l; + } + return x; +}; + +int +rn_refines(m_arg, n_arg) + void *m_arg, *n_arg; +{ + register caddr_t m = m_arg, n = n_arg; + register caddr_t lim, lim2 = lim = n + *(u_char *)n; + int longer = (*(u_char *)n++) - (int)(*(u_char *)m++); + int masks_are_equal = 1; + + if (longer > 0) + lim -= longer; + while (n < lim) { + if (*n & ~(*m)) + return 0; + if (*n++ != *m++) + masks_are_equal = 0; + } + while (n < lim2) + if (*n++) + return 0; + if (masks_are_equal && (longer < 0)) + for (lim2 = m - longer; m < lim2; ) + if (*m++) + return 1; + return (!masks_are_equal); +} + +struct radix_node * +rn_lookup(v_arg, m_arg, head) + void *v_arg, *m_arg; + struct radix_node_head *head; +{ + register struct radix_node *x; + caddr_t netmask = 0; + + if (m_arg) { + if ((x = rn_addmask(m_arg, 1, head->rnh_treetop->rn_off)) == 0) + return (0); + netmask = x->rn_key; + } + x = rn_match(v_arg, head); + if (x && netmask) { + while (x && x->rn_mask != netmask) + x = x->rn_dupedkey; + } + return x; +} + +static +rn_satsifies_leaf(trial, leaf, skip) + char *trial; + register struct radix_node *leaf; + int skip; +{ + register char *cp = trial, *cp2 = leaf->rn_key, *cp3 = leaf->rn_mask; + char *cplim; + int length = min(*(u_char *)cp, *(u_char *)cp2); + + if (cp3 == 0) + cp3 = rn_ones; + else + length = min(length, *(u_char *)cp3); + cplim = cp + length; cp3 += skip; cp2 += skip; + for (cp += skip; cp < cplim; cp++, cp2++, cp3++) + if ((*cp ^ *cp2) & *cp3) + return 0; + return 1; +} + +struct radix_node * +rn_match(v_arg, head) + void *v_arg; + struct radix_node_head *head; +{ + caddr_t v = v_arg; + register struct radix_node *t = head->rnh_treetop, *x; + register caddr_t cp = v, cp2; + caddr_t cplim; + struct radix_node *saved_t, *top = t; + int off = t->rn_off, vlen = *(u_char *)cp, matched_off; + register int test, b, rn_b; + + /* + * Open code rn_search(v, top) to avoid overhead of extra + * subroutine call. + */ + for (; t->rn_b >= 0; ) { + if (t->rn_bmask & cp[t->rn_off]) + t = t->rn_r; + else + t = t->rn_l; + } + /* + * See if we match exactly as a host destination + * or at least learn how many bits match, for normal mask finesse. + * + * It doesn't hurt us to limit how many bytes to check + * to the length of the mask, since if it matches we had a genuine + * match and the leaf we have is the most specific one anyway; + * if it didn't match with a shorter length it would fail + * with a long one. This wins big for class B&C netmasks which + * are probably the most common case... + */ + if (t->rn_mask) + vlen = *(u_char *)t->rn_mask; + cp += off; cp2 = t->rn_key + off; cplim = v + vlen; + for (; cp < cplim; cp++, cp2++) + if (*cp != *cp2) + goto on1; + /* + * This extra grot is in case we are explicitly asked + * to look up the default. Ugh! + */ + if ((t->rn_flags & RNF_ROOT) && t->rn_dupedkey) + t = t->rn_dupedkey; + return t; +on1: + test = (*cp ^ *cp2) & 0xff; /* find first bit that differs */ + for (b = 7; (test >>= 1) > 0;) + b--; + matched_off = cp - v; + b += matched_off << 3; + rn_b = -1 - b; + /* + * If there is a host route in a duped-key chain, it will be first. + */ + if ((saved_t = t)->rn_mask == 0) + t = t->rn_dupedkey; + for (; t; t = t->rn_dupedkey) + /* + * Even if we don't match exactly as a host, + * we may match if the leaf we wound up at is + * a route to a net. + */ + if (t->rn_flags & RNF_NORMAL) { + if (rn_b <= t->rn_b) + return t; + } else if (rn_satsifies_leaf(v, t, matched_off)) + return t; + t = saved_t; + /* start searching up the tree */ + do { + register struct radix_mask *m; + t = t->rn_p; + if (m = t->rn_mklist) { + /* + * If non-contiguous masks ever become important + * we can restore the masking and open coding of + * the search and satisfaction test and put the + * calculation of "off" back before the "do". + */ + do { + if (m->rm_flags & RNF_NORMAL) { + if (rn_b <= m->rm_b) + return (m->rm_leaf); + } else { + off = min(t->rn_off, matched_off); + x = rn_search_m(v, t, m->rm_mask); + while (x && x->rn_mask != m->rm_mask) + x = x->rn_dupedkey; + if (x && rn_satsifies_leaf(v, x, off)) + return x; + } + } while (m = m->rm_mklist); + } + } while (t != top); + return 0; +}; + +#ifdef RN_DEBUG +int rn_nodenum; +struct radix_node *rn_clist; +int rn_saveinfo; +int rn_debug = 1; +#endif + +struct radix_node * +rn_newpair(v, b, nodes) + void *v; + int b; + struct radix_node nodes[2]; +{ + register struct radix_node *tt = nodes, *t = tt + 1; + t->rn_b = b; t->rn_bmask = 0x80 >> (b & 7); + t->rn_l = tt; t->rn_off = b >> 3; + tt->rn_b = -1; tt->rn_key = (caddr_t)v; tt->rn_p = t; + tt->rn_flags = t->rn_flags = RNF_ACTIVE; +#ifdef RN_DEBUG + tt->rn_info = rn_nodenum++; t->rn_info = rn_nodenum++; + tt->rn_twin = t; tt->rn_ybro = rn_clist; rn_clist = tt; +#endif + return t; +} + +struct radix_node * +rn_insert(v_arg, head, dupentry, nodes) + void *v_arg; + struct radix_node_head *head; + int *dupentry; + struct radix_node nodes[2]; +{ + caddr_t v = v_arg; + struct radix_node *top = head->rnh_treetop; + int head_off = top->rn_off, vlen = (int)*((u_char *)v); + register struct radix_node *t = rn_search(v_arg, top); + register caddr_t cp = v + head_off; + register int b; + struct radix_node *tt; + /* + * Find first bit at which v and t->rn_key differ + */ + { + register caddr_t cp2 = t->rn_key + head_off; + register int cmp_res; + caddr_t cplim = v + vlen; + + while (cp < cplim) + if (*cp2++ != *cp++) + goto on1; + *dupentry = 1; + return t; +on1: + *dupentry = 0; + cmp_res = (cp[-1] ^ cp2[-1]) & 0xff; + for (b = (cp - v) << 3; cmp_res; b--) + cmp_res >>= 1; + } + { + register struct radix_node *p, *x = top; + cp = v; + do { + p = x; + if (cp[x->rn_off] & x->rn_bmask) + x = x->rn_r; + else x = x->rn_l; + } while (b > (unsigned) x->rn_b); /* x->rn_b < b && x->rn_b >= 0 */ +#ifdef RN_DEBUG + if (rn_debug) + log(LOG_DEBUG, "rn_insert: Going In:\n"), traverse(p); +#endif + t = rn_newpair(v_arg, b, nodes); tt = t->rn_l; + if ((cp[p->rn_off] & p->rn_bmask) == 0) + p->rn_l = t; + else + p->rn_r = t; + x->rn_p = t; t->rn_p = p; /* frees x, p as temp vars below */ + if ((cp[t->rn_off] & t->rn_bmask) == 0) { + t->rn_r = x; + } else { + t->rn_r = tt; t->rn_l = x; + } +#ifdef RN_DEBUG + if (rn_debug) + log(LOG_DEBUG, "rn_insert: Coming Out:\n"), traverse(p); +#endif + } + return (tt); +} + +struct radix_node * +rn_addmask(n_arg, search, skip) + int search, skip; + void *n_arg; +{ + caddr_t netmask = (caddr_t)n_arg; + register struct radix_node *x; + register caddr_t cp, cplim; + register int b = 0, mlen, j; + int maskduplicated, m0, isnormal; + struct radix_node *saved_x; + static int last_zeroed = 0; + + if ((mlen = *(u_char *)netmask) > max_keylen) + mlen = max_keylen; + if (skip == 0) + skip = 1; + if (mlen <= skip) + return (mask_rnhead->rnh_nodes); + if (skip > 1) + Bcopy(rn_ones + 1, addmask_key + 1, skip - 1); + if ((m0 = mlen) > skip) + Bcopy(netmask + skip, addmask_key + skip, mlen - skip); + /* + * Trim trailing zeroes. + */ + for (cp = addmask_key + mlen; (cp > addmask_key) && cp[-1] == 0;) + cp--; + mlen = cp - addmask_key; + if (mlen <= skip) { + if (m0 >= last_zeroed) + last_zeroed = mlen; + return (mask_rnhead->rnh_nodes); + } + if (m0 < last_zeroed) + Bzero(addmask_key + m0, last_zeroed - m0); + *addmask_key = last_zeroed = mlen; + x = rn_search(addmask_key, rn_masktop); + if (Bcmp(addmask_key, x->rn_key, mlen) != 0) + x = 0; + if (x || search) + return (x); + R_Malloc(x, struct radix_node *, max_keylen + 2 * sizeof (*x)); + if ((saved_x = x) == 0) + return (0); + Bzero(x, max_keylen + 2 * sizeof (*x)); + netmask = cp = (caddr_t)(x + 2); + Bcopy(addmask_key, cp, mlen); + x = rn_insert(cp, mask_rnhead, &maskduplicated, x); + if (maskduplicated) { + log(LOG_ERR, "rn_addmask: mask impossibly already in tree"); + Free(saved_x); + return (x); + } + /* + * Calculate index of mask, and check for normalcy. + */ + cplim = netmask + mlen; isnormal = 1; + for (cp = netmask + skip; (cp < cplim) && *(u_char *)cp == 0xff;) + cp++; + if (cp != cplim) { + for (j = 0x80; (j & *cp) != 0; j >>= 1) + b++; + if (*cp != normal_chars[b] || cp != (cplim - 1)) + isnormal = 0; + } + b += (cp - netmask) << 3; + x->rn_b = -1 - b; + if (isnormal) + x->rn_flags |= RNF_NORMAL; + return (x); +} + +static int /* XXX: arbitrary ordering for non-contiguous masks */ +rn_lexobetter(m_arg, n_arg) + void *m_arg, *n_arg; +{ + register u_char *mp = m_arg, *np = n_arg, *lim; + + if (*mp > *np) + return 1; /* not really, but need to check longer one first */ + if (*mp == *np) + for (lim = mp + *mp; mp < lim;) + if (*mp++ > *np++) + return 1; + return 0; +} + +static struct radix_mask * +rn_new_radix_mask(tt, next) + register struct radix_node *tt; + register struct radix_mask *next; +{ + register struct radix_mask *m; + + MKGet(m); + if (m == 0) { + log(LOG_ERR, "Mask for route not entered\n"); + return (0); + } + Bzero(m, sizeof *m); + m->rm_b = tt->rn_b; + m->rm_flags = tt->rn_flags; + if (tt->rn_flags & RNF_NORMAL) + m->rm_leaf = tt; + else + m->rm_mask = tt->rn_mask; + m->rm_mklist = next; + tt->rn_mklist = m; + return m; +} + +struct radix_node * +rn_addroute(v_arg, n_arg, head, treenodes) + void *v_arg, *n_arg; + struct radix_node_head *head; + struct radix_node treenodes[2]; +{ + caddr_t v = (caddr_t)v_arg, netmask = (caddr_t)n_arg; + register struct radix_node *t, *x, *tt; + struct radix_node *saved_tt, *top = head->rnh_treetop; + short b = 0, b_leaf; + int keyduplicated; + caddr_t mmask; + struct radix_mask *m, **mp; + + /* + * In dealing with non-contiguous masks, there may be + * many different routes which have the same mask. + * We will find it useful to have a unique pointer to + * the mask to speed avoiding duplicate references at + * nodes and possibly save time in calculating indices. + */ + if (netmask) { + if ((x = rn_addmask(netmask, 0, top->rn_off)) == 0) + return (0); + b_leaf = x->rn_b; + b = -1 - x->rn_b; + netmask = x->rn_key; + } + /* + * Deal with duplicated keys: attach node to previous instance + */ + saved_tt = tt = rn_insert(v, head, &keyduplicated, treenodes); + if (keyduplicated) { + for (t = tt; tt; t = tt, tt = tt->rn_dupedkey) { + if (tt->rn_mask == netmask) + return (0); + if (netmask == 0 || + (tt->rn_mask && + ((b_leaf < tt->rn_b) || /* index(netmask) > node */ + rn_refines(netmask, tt->rn_mask) || + rn_lexobetter(netmask, tt->rn_mask)))) + break; + } + /* + * If the mask is not duplicated, we wouldn't + * find it among possible duplicate key entries + * anyway, so the above test doesn't hurt. + * + * We sort the masks for a duplicated key the same way as + * in a masklist -- most specific to least specific. + * This may require the unfortunate nuisance of relocating + * the head of the list. + */ + if (tt == saved_tt) { + struct radix_node *xx = x; + /* link in at head of list */ + (tt = treenodes)->rn_dupedkey = t; + tt->rn_flags = t->rn_flags; + tt->rn_p = x = t->rn_p; + if (x->rn_l == t) x->rn_l = tt; else x->rn_r = tt; + saved_tt = tt; x = xx; + } else { + (tt = treenodes)->rn_dupedkey = t->rn_dupedkey; + t->rn_dupedkey = tt; + } +#ifdef RN_DEBUG + t=tt+1; tt->rn_info = rn_nodenum++; t->rn_info = rn_nodenum++; + tt->rn_twin = t; tt->rn_ybro = rn_clist; rn_clist = tt; +#endif + tt->rn_key = (caddr_t) v; + tt->rn_b = -1; + tt->rn_flags = RNF_ACTIVE; + } + /* + * Put mask in tree. + */ + if (netmask) { + tt->rn_mask = netmask; + tt->rn_b = x->rn_b; + tt->rn_flags |= x->rn_flags & RNF_NORMAL; + } + t = saved_tt->rn_p; + if (keyduplicated) + goto on2; + b_leaf = -1 - t->rn_b; + if (t->rn_r == saved_tt) x = t->rn_l; else x = t->rn_r; + /* Promote general routes from below */ + if (x->rn_b < 0) { + for (mp = &t->rn_mklist; x; x = x->rn_dupedkey) + if (x->rn_mask && (x->rn_b >= b_leaf) && x->rn_mklist == 0) { + if (*mp = m = rn_new_radix_mask(x, 0)) + mp = &m->rm_mklist; + } + } else if (x->rn_mklist) { + /* + * Skip over masks whose index is > that of new node + */ + for (mp = &x->rn_mklist; m = *mp; mp = &m->rm_mklist) + if (m->rm_b >= b_leaf) + break; + t->rn_mklist = m; *mp = 0; + } +on2: + /* Add new route to highest possible ancestor's list */ + if ((netmask == 0) || (b > t->rn_b )) + return tt; /* can't lift at all */ + b_leaf = tt->rn_b; + do { + x = t; + t = t->rn_p; + } while (b <= t->rn_b && x != top); + /* + * Search through routes associated with node to + * insert new route according to index. + * Need same criteria as when sorting dupedkeys to avoid + * double loop on deletion. + */ + for (mp = &x->rn_mklist; m = *mp; mp = &m->rm_mklist) { + if (m->rm_b < b_leaf) + continue; + if (m->rm_b > b_leaf) + break; + if (m->rm_flags & RNF_NORMAL) { + mmask = m->rm_leaf->rn_mask; + if (tt->rn_flags & RNF_NORMAL) { + log(LOG_ERR, + "Non-unique normal route, mask not entered"); + return tt; + } + } else + mmask = m->rm_mask; + if (mmask == netmask) { + m->rm_refs++; + tt->rn_mklist = m; + return tt; + } + if (rn_refines(netmask, mmask) || rn_lexobetter(netmask, mmask)) + break; + } + *mp = rn_new_radix_mask(tt, *mp); + return tt; +} + +struct radix_node * +rn_delete(v_arg, netmask_arg, head) + void *v_arg, *netmask_arg; + struct radix_node_head *head; +{ + register struct radix_node *t, *p, *x, *tt; + struct radix_mask *m, *saved_m, **mp; + struct radix_node *dupedkey, *saved_tt, *top; + caddr_t v, netmask; + int b, head_off, vlen; + + v = v_arg; + netmask = netmask_arg; + x = head->rnh_treetop; + tt = rn_search(v, x); + head_off = x->rn_off; + vlen = *(u_char *)v; + saved_tt = tt; + top = x; + if (tt == 0 || + Bcmp(v + head_off, tt->rn_key + head_off, vlen - head_off)) + return (0); + /* + * Delete our route from mask lists. + */ + if (netmask) { + if ((x = rn_addmask(netmask, 1, head_off)) == 0) + return (0); + netmask = x->rn_key; + while (tt->rn_mask != netmask) + if ((tt = tt->rn_dupedkey) == 0) + return (0); + } + if (tt->rn_mask == 0 || (saved_m = m = tt->rn_mklist) == 0) + goto on1; + if (tt->rn_flags & RNF_NORMAL) { + if (m->rm_leaf != tt || m->rm_refs > 0) { + log(LOG_ERR, "rn_delete: inconsistent annotation\n"); + return 0; /* dangling ref could cause disaster */ + } + } else { + if (m->rm_mask != tt->rn_mask) { + log(LOG_ERR, "rn_delete: inconsistent annotation\n"); + goto on1; + } + if (--m->rm_refs >= 0) + goto on1; + } + b = -1 - tt->rn_b; + t = saved_tt->rn_p; + if (b > t->rn_b) + goto on1; /* Wasn't lifted at all */ + do { + x = t; + t = t->rn_p; + } while (b <= t->rn_b && x != top); + for (mp = &x->rn_mklist; m = *mp; mp = &m->rm_mklist) + if (m == saved_m) { + *mp = m->rm_mklist; + MKFree(m); + break; + } + if (m == 0) { + log(LOG_ERR, "rn_delete: couldn't find our annotation\n"); + if (tt->rn_flags & RNF_NORMAL) + return (0); /* Dangling ref to us */ + } +on1: + /* + * Eliminate us from tree + */ + if (tt->rn_flags & RNF_ROOT) + return (0); +#ifdef RN_DEBUG + /* Get us out of the creation list */ + for (t = rn_clist; t && t->rn_ybro != tt; t = t->rn_ybro) {} + if (t) t->rn_ybro = tt->rn_ybro; +#endif + t = tt->rn_p; + if (dupedkey = saved_tt->rn_dupedkey) { + if (tt == saved_tt) { + x = dupedkey; x->rn_p = t; + if (t->rn_l == tt) t->rn_l = x; else t->rn_r = x; + } else { + for (x = p = saved_tt; p && p->rn_dupedkey != tt;) + p = p->rn_dupedkey; + if (p) p->rn_dupedkey = tt->rn_dupedkey; + else log(LOG_ERR, "rn_delete: couldn't find us\n"); + } + t = tt + 1; + if (t->rn_flags & RNF_ACTIVE) { +#ifndef RN_DEBUG + *++x = *t; p = t->rn_p; +#else + b = t->rn_info; *++x = *t; t->rn_info = b; p = t->rn_p; +#endif + if (p->rn_l == t) p->rn_l = x; else p->rn_r = x; + x->rn_l->rn_p = x; x->rn_r->rn_p = x; + } + goto out; + } + if (t->rn_l == tt) x = t->rn_r; else x = t->rn_l; + p = t->rn_p; + if (p->rn_r == t) p->rn_r = x; else p->rn_l = x; + x->rn_p = p; + /* + * Demote routes attached to us. + */ + if (t->rn_mklist) { + if (x->rn_b >= 0) { + for (mp = &x->rn_mklist; m = *mp;) + mp = &m->rm_mklist; + *mp = t->rn_mklist; + } else { + /* If there are any key,mask pairs in a sibling + duped-key chain, some subset will appear sorted + in the same order attached to our mklist */ + for (m = t->rn_mklist; m && x; x = x->rn_dupedkey) + if (m == x->rn_mklist) { + struct radix_mask *mm = m->rm_mklist; + x->rn_mklist = 0; + if (--(m->rm_refs) < 0) + MKFree(m); + m = mm; + } + if (m) + log(LOG_ERR, "%s %x at %x\n", + "rn_delete: Orphaned Mask", m, x); + } + } + /* + * We may be holding an active internal node in the tree. + */ + x = tt + 1; + if (t != x) { +#ifndef RN_DEBUG + *t = *x; +#else + b = t->rn_info; *t = *x; t->rn_info = b; +#endif + t->rn_l->rn_p = t; t->rn_r->rn_p = t; + p = x->rn_p; + if (p->rn_l == x) p->rn_l = t; else p->rn_r = t; + } +out: + tt->rn_flags &= ~RNF_ACTIVE; + tt[1].rn_flags &= ~RNF_ACTIVE; + return (tt); +} + +int +rn_walktree(h, f, w) + struct radix_node_head *h; + register int (*f)(); + void *w; +{ + int error; + struct radix_node *base, *next; + register struct radix_node *rn = h->rnh_treetop; + /* + * This gets complicated because we may delete the node + * while applying the function f to it, so we need to calculate + * the successor node in advance. + */ + /* First time through node, go left */ + while (rn->rn_b >= 0) + rn = rn->rn_l; + for (;;) { + base = rn; + /* If at right child go back up, otherwise, go right */ + while (rn->rn_p->rn_r == rn && (rn->rn_flags & RNF_ROOT) == 0) + rn = rn->rn_p; + /* Find the next *leaf* since next node might vanish, too */ + for (rn = rn->rn_p->rn_r; rn->rn_b >= 0;) + rn = rn->rn_l; + next = rn; + /* Process leaves */ + while (rn = base) { + base = rn->rn_dupedkey; + if (!(rn->rn_flags & RNF_ROOT) && (error = (*f)(rn, w))) + return (error); + } + rn = next; + if (rn->rn_flags & RNF_ROOT) + return (0); + } + /* NOTREACHED */ +} + +int +rn_inithead(head, off) + void **head; + int off; +{ + register struct radix_node_head *rnh; + register struct radix_node *t, *tt, *ttt; + if (*head) + return (1); + R_Malloc(rnh, struct radix_node_head *, sizeof (*rnh)); + if (rnh == 0) + return (0); + Bzero(rnh, sizeof (*rnh)); + *head = rnh; + t = rn_newpair(rn_zeros, off, rnh->rnh_nodes); + ttt = rnh->rnh_nodes + 2; + t->rn_r = ttt; + t->rn_p = t; + tt = t->rn_l; + tt->rn_flags = t->rn_flags = RNF_ROOT | RNF_ACTIVE; + tt->rn_b = -1 - off; + *ttt = *tt; + ttt->rn_key = rn_ones; + rnh->rnh_addaddr = rn_addroute; + rnh->rnh_deladdr = rn_delete; + rnh->rnh_matchaddr = rn_match; + rnh->rnh_lookup = rn_lookup; + rnh->rnh_walktree = rn_walktree; + rnh->rnh_treetop = t; + return (1); +} + +void +rn_init() +{ + char *cp, *cplim; +#ifdef _KERNEL + struct domain *dom; + + for (dom = domains; dom; dom = dom->dom_next) + if (dom->dom_maxrtkey > max_keylen) + max_keylen = dom->dom_maxrtkey; +#endif + if (max_keylen == 0) { + log(LOG_ERR, + "rn_init: radix functions require max_keylen be set\n"); + return; + } + R_Malloc(rn_zeros, char *, 3 * max_keylen); + if (rn_zeros == NULL) + panic("rn_init"); + Bzero(rn_zeros, 3 * max_keylen); + rn_ones = cp = rn_zeros + max_keylen; + addmask_key = cplim = rn_ones + max_keylen; + while (cp < cplim) + *cp++ = -1; + if (rn_inithead((void **)&mask_rnhead, 0) == 0) + panic("rn_init 2"); +} diff --git a/sys/net/radix.h b/sys/net/radix.h new file mode 100644 index 00000000000..86d1d2a37f3 --- /dev/null +++ b/sys/net/radix.h @@ -0,0 +1,165 @@ +/* $NetBSD: radix.h,v 1.7 1995/05/17 15:50:08 mycroft Exp $ */ + +/* + * Copyright (c) 1988, 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)radix.h 8.2 (Berkeley) 10/31/94 + */ + +#ifndef _NET_RADIX_H_ +#define _NET_RADIX_H_ + +/* + * Radix search tree node layout. + */ + +struct radix_node { + struct radix_mask *rn_mklist; /* list of masks contained in subtree */ + struct radix_node *rn_p; /* parent */ + short rn_b; /* bit offset; -1-index(netmask) */ + char rn_bmask; /* node: mask for bit test*/ + u_char rn_flags; /* enumerated next */ +#define RNF_NORMAL 1 /* leaf contains normal route */ +#define RNF_ROOT 2 /* leaf is root leaf for tree */ +#define RNF_ACTIVE 4 /* This node is alive (for rtfree) */ + union { + struct { /* leaf only data: */ + caddr_t rn_Key; /* object of search */ + caddr_t rn_Mask; /* netmask, if present */ + struct radix_node *rn_Dupedkey; + } rn_leaf; + struct { /* node only data: */ + int rn_Off; /* where to start compare */ + struct radix_node *rn_L;/* progeny */ + struct radix_node *rn_R;/* progeny */ + } rn_node; + } rn_u; +#ifdef RN_DEBUG + int rn_info; + struct radix_node *rn_twin; + struct radix_node *rn_ybro; +#endif +}; + +#define rn_dupedkey rn_u.rn_leaf.rn_Dupedkey +#define rn_key rn_u.rn_leaf.rn_Key +#define rn_mask rn_u.rn_leaf.rn_Mask +#define rn_off rn_u.rn_node.rn_Off +#define rn_l rn_u.rn_node.rn_L +#define rn_r rn_u.rn_node.rn_R + +/* + * Annotations to tree concerning potential routes applying to subtrees. + */ + +extern struct radix_mask { + short rm_b; /* bit offset; -1-index(netmask) */ + char rm_unused; /* cf. rn_bmask */ + u_char rm_flags; /* cf. rn_flags */ + struct radix_mask *rm_mklist; /* more masks to try */ + union { + caddr_t rmu_mask; /* the mask */ + struct radix_node *rmu_leaf; /* for normal routes */ + } rm_rmu; + int rm_refs; /* # of references to this struct */ +} *rn_mkfreelist; + +#define rm_mask rm_rmu.rmu_mask +#define rm_leaf rm_rmu.rmu_leaf /* extra field would make 32 bytes */ + +#define MKGet(m) {\ + if (rn_mkfreelist) {\ + m = rn_mkfreelist; \ + rn_mkfreelist = (m)->rm_mklist; \ + } else \ + R_Malloc(m, struct radix_mask *, sizeof (*(m))); }\ + +#define MKFree(m) { (m)->rm_mklist = rn_mkfreelist; rn_mkfreelist = (m);} + +struct radix_node_head { + struct radix_node *rnh_treetop; + int rnh_addrsize; /* permit, but not require fixed keys */ + int rnh_pktsize; /* permit, but not require fixed keys */ + struct radix_node *(*rnh_addaddr) /* add based on sockaddr */ + __P((void *v, void *mask, + struct radix_node_head *head, struct radix_node nodes[])); + struct radix_node *(*rnh_addpkt) /* add based on packet hdr */ + __P((void *v, void *mask, + struct radix_node_head *head, struct radix_node nodes[])); + struct radix_node *(*rnh_deladdr) /* remove based on sockaddr */ + __P((void *v, void *mask, struct radix_node_head *head)); + struct radix_node *(*rnh_delpkt) /* remove based on packet hdr */ + __P((void *v, void *mask, struct radix_node_head *head)); + struct radix_node *(*rnh_matchaddr) /* locate based on sockaddr */ + __P((void *v, struct radix_node_head *head)); + struct radix_node *(*rnh_lookup) /* locate based on sockaddr */ + __P((void *v, void *mask, struct radix_node_head *head)); + struct radix_node *(*rnh_matchpkt) /* locate based on packet hdr */ + __P((void *v, struct radix_node_head *head)); + int (*rnh_walktree) /* traverse tree */ + __P((struct radix_node_head *head, int (*f)(), void *w)); + struct radix_node rnh_nodes[3]; /* empty tree for common case */ +}; + + +#ifndef _KERNEL +#define Bcmp(a, b, n) bcmp(((char *)(a)), ((char *)(b)), (n)) +#define Bcopy(a, b, n) bcopy(((char *)(a)), ((char *)(b)), (unsigned)(n)) +#define Bzero(p, n) bzero((char *)(p), (int)(n)); +#define R_Malloc(p, t, n) (p = (t) malloc((unsigned int)(n))) +#define Free(p) free((char *)p); +#else +#define Bcmp(a, b, n) bcmp(((caddr_t)(a)), ((caddr_t)(b)), (unsigned)(n)) +#define Bcopy(a, b, n) bcopy(((caddr_t)(a)), ((caddr_t)(b)), (unsigned)(n)) +#define Bzero(p, n) bzero((caddr_t)(p), (unsigned)(n)); +#define R_Malloc(p, t, n) (p = (t) malloc((unsigned long)(n), M_RTABLE, M_DONTWAIT)) +#define Free(p) free((caddr_t)p, M_RTABLE); + +void rn_init __P((void)); +int rn_inithead __P((void **, int)); +int rn_refines __P((void *, void *)); +int rn_walktree __P((struct radix_node_head *, int (*)(), void *)); +struct radix_node + *rn_addmask __P((void *, int, int)), + *rn_addroute __P((void *, void *, struct radix_node_head *, + struct radix_node [2])), + *rn_delete __P((void *, void *, struct radix_node_head *)), + *rn_insert __P((void *, struct radix_node_head *, int *, + struct radix_node [2])), + *rn_lookup __P((void *, void *, struct radix_node_head *)), + *rn_match __P((void *, struct radix_node_head *)), + *rn_newpair __P((void *, int, struct radix_node[2])), + *rn_search __P((void *, struct radix_node *)), + *rn_search_m __P((void *, struct radix_node *, void *)); +#endif /* !_KERNEL */ + +#endif /* !_NET_RADIX_H_ */ diff --git a/sys/net/raw_cb.c b/sys/net/raw_cb.c new file mode 100644 index 00000000000..a8e1fc36ce0 --- /dev/null +++ b/sys/net/raw_cb.c @@ -0,0 +1,146 @@ +/* $NetBSD: raw_cb.c,v 1.8 1995/06/12 00:46:53 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)raw_cb.c 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/socketvar.h> +#include <sys/domain.h> +#include <sys/protosw.h> +#include <sys/errno.h> + +#include <net/if.h> +#include <net/route.h> +#include <net/raw_cb.h> +#include <netinet/in.h> + +/* + * Routines to manage the raw protocol control blocks. + * + * TODO: + * hash lookups by protocol family/protocol + address family + * take care of unique address problems per AF? + * redo address binding to allow wildcards + */ + +u_long raw_sendspace = RAWSNDQ; +u_long raw_recvspace = RAWRCVQ; + +/* + * Allocate a control block and a nominal amount + * of buffer space for the socket. + */ +int +raw_attach(so, proto) + register struct socket *so; + int proto; +{ + register struct rawcb *rp = sotorawcb(so); + int error; + + /* + * It is assumed that raw_attach is called + * after space has been allocated for the + * rawcb. + */ + if (rp == 0) + return (ENOBUFS); + if (error = soreserve(so, raw_sendspace, raw_recvspace)) + return (error); + rp->rcb_socket = so; + rp->rcb_proto.sp_family = so->so_proto->pr_domain->dom_family; + rp->rcb_proto.sp_protocol = proto; + LIST_INSERT_HEAD(&rawcb, rp, rcb_list); + return (0); +} + +/* + * Detach the raw connection block and discard + * socket resources. + */ +void +raw_detach(rp) + register struct rawcb *rp; +{ + struct socket *so = rp->rcb_socket; + + so->so_pcb = 0; + sofree(so); + LIST_REMOVE(rp, rcb_list); +#ifdef notdef + if (rp->rcb_laddr) + m_freem(dtom(rp->rcb_laddr)); + rp->rcb_laddr = 0; +#endif + free((caddr_t)(rp), M_PCB); +} + +/* + * Disconnect and possibly release resources. + */ +void +raw_disconnect(rp) + struct rawcb *rp; +{ + +#ifdef notdef + if (rp->rcb_faddr) + m_freem(dtom(rp->rcb_faddr)); + rp->rcb_faddr = 0; +#endif + if (rp->rcb_socket->so_state & SS_NOFDREF) + raw_detach(rp); +} + +#ifdef notdef +int +raw_bind(so, nam) + register struct socket *so; + struct mbuf *nam; +{ + struct sockaddr *addr = mtod(nam, struct sockaddr *); + register struct rawcb *rp; + + if (ifnet == 0) + return (EADDRNOTAVAIL); + rp = sotorawcb(so); + nam = m_copym(nam, 0, M_COPYALL, M_WAITOK); + rp->rcb_laddr = mtod(nam, struct sockaddr *); + return (0); +} +#endif diff --git a/sys/net/raw_cb.h b/sys/net/raw_cb.h new file mode 100644 index 00000000000..34c5a72ca9a --- /dev/null +++ b/sys/net/raw_cb.h @@ -0,0 +1,70 @@ +/* $NetBSD: raw_cb.h,v 1.8 1995/06/12 00:46:54 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)raw_cb.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * Raw protocol interface control block. Used + * to tie a socket to the generic raw interface. + */ +struct rawcb { + LIST_ENTRY(rawcb) rcb_list; /* doubly linked list */ + struct socket *rcb_socket; /* back pointer to socket */ + struct sockaddr *rcb_faddr; /* destination address */ + struct sockaddr *rcb_laddr; /* socket's address */ + struct sockproto rcb_proto; /* protocol family, protocol */ +}; + +#define sotorawcb(so) ((struct rawcb *)(so)->so_pcb) + +/* + * Nominal space allocated to a raw socket. + */ +#define RAWSNDQ 8192 +#define RAWRCVQ 8192 + +#ifdef _KERNEL +LIST_HEAD(, rawcb) rawcb; /* head of list */ + +int raw_attach __P((struct socket *, int)); +void raw_ctlinput __P((int, struct sockaddr *)); +void raw_detach __P((struct rawcb *)); +void raw_disconnect __P((struct rawcb *)); +void raw_init __P((void)); +void raw_input __P((struct mbuf *, + struct sockproto *, struct sockaddr *, struct sockaddr *)); +int raw_usrreq __P((struct socket *, + int, struct mbuf *, struct mbuf *, struct mbuf *)); +#endif /* _KERNEL */ diff --git a/sys/net/raw_usrreq.c b/sys/net/raw_usrreq.c new file mode 100644 index 00000000000..04f5e19a0e4 --- /dev/null +++ b/sys/net/raw_usrreq.c @@ -0,0 +1,308 @@ +/* $NetBSD: raw_usrreq.c,v 1.10 1995/06/12 00:46:55 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)raw_usrreq.c 8.1 (Berkeley) 6/10/93 + */ + +#include <sys/param.h> +#include <sys/mbuf.h> +#include <sys/domain.h> +#include <sys/protosw.h> +#include <sys/socket.h> +#include <sys/socketvar.h> +#include <sys/errno.h> + +#include <net/if.h> +#include <net/route.h> +#include <net/netisr.h> +#include <net/raw_cb.h> + +/* + * Initialize raw connection block q. + */ +void +raw_init() +{ + + LIST_INIT(&rawcb); +} + + +/* + * Raw protocol input routine. Find the socket + * associated with the packet(s) and move them over. If + * nothing exists for this packet, drop it. + */ +/* + * Raw protocol interface. + */ +void +raw_input(m0, proto, src, dst) + struct mbuf *m0; + register struct sockproto *proto; + struct sockaddr *src, *dst; +{ + register struct rawcb *rp; + register struct mbuf *m = m0; + register int sockets = 0; + struct socket *last; + + last = 0; + for (rp = rawcb.lh_first; rp != 0; rp = rp->rcb_list.le_next) { + if (rp->rcb_proto.sp_family != proto->sp_family) + continue; + if (rp->rcb_proto.sp_protocol && + rp->rcb_proto.sp_protocol != proto->sp_protocol) + continue; + /* + * We assume the lower level routines have + * placed the address in a canonical format + * suitable for a structure comparison. + * + * Note that if the lengths are not the same + * the comparison will fail at the first byte. + */ +#define equal(a1, a2) \ + (bcmp((caddr_t)(a1), (caddr_t)(a2), a1->sa_len) == 0) + if (rp->rcb_laddr && !equal(rp->rcb_laddr, dst)) + continue; + if (rp->rcb_faddr && !equal(rp->rcb_faddr, src)) + continue; + if (last) { + struct mbuf *n; + if (n = m_copy(m, 0, (int)M_COPYALL)) { + if (sbappendaddr(&last->so_rcv, src, + n, (struct mbuf *)0) == 0) + /* should notify about lost packet */ + m_freem(n); + else { + sorwakeup(last); + sockets++; + } + } + } + last = rp->rcb_socket; + } + if (last) { + if (sbappendaddr(&last->so_rcv, src, + m, (struct mbuf *)0) == 0) + m_freem(m); + else { + sorwakeup(last); + sockets++; + } + } else + m_freem(m); +} + +/*ARGSUSED*/ +void +raw_ctlinput(cmd, arg) + int cmd; + struct sockaddr *arg; +{ + + if (cmd < 0 || cmd > PRC_NCMDS) + return; + /* INCOMPLETE */ +} + +/*ARGSUSED*/ +int +raw_usrreq(so, req, m, nam, control) + struct socket *so; + int req; + struct mbuf *m, *nam, *control; +{ + register struct rawcb *rp = sotorawcb(so); + register int error = 0; + int len; + + if (req == PRU_CONTROL) + return (EOPNOTSUPP); + if (control && control->m_len) { + error = EOPNOTSUPP; + goto release; + } + if (rp == 0) { + error = EINVAL; + goto release; + } + switch (req) { + + /* + * Allocate a raw control block and fill in the + * necessary info to allow packets to be routed to + * the appropriate raw interface routine. + */ + case PRU_ATTACH: + if ((so->so_state & SS_PRIV) == 0) { + error = EACCES; + break; + } + error = raw_attach(so, (int)(long)nam); + break; + + /* + * Destroy state just before socket deallocation. + * Flush data or not depending on the options. + */ + case PRU_DETACH: + if (rp == 0) { + error = ENOTCONN; + break; + } + raw_detach(rp); + break; + +#ifdef notdef + /* + * If a socket isn't bound to a single address, + * the raw input routine will hand it anything + * within that protocol family (assuming there's + * nothing else around it should go to). + */ + case PRU_CONNECT: + if (rp->rcb_faddr) { + error = EISCONN; + break; + } + nam = m_copym(nam, 0, M_COPYALL, M_WAIT); + rp->rcb_faddr = mtod(nam, struct sockaddr *); + soisconnected(so); + break; + + case PRU_BIND: + if (rp->rcb_laddr) { + error = EINVAL; /* XXX */ + break; + } + error = raw_bind(so, nam); + break; +#endif + + case PRU_CONNECT2: + error = EOPNOTSUPP; + goto release; + + case PRU_DISCONNECT: + if (rp->rcb_faddr == 0) { + error = ENOTCONN; + break; + } + raw_disconnect(rp); + soisdisconnected(so); + break; + + /* + * Mark the connection as being incapable of further input. + */ + case PRU_SHUTDOWN: + socantsendmore(so); + break; + + /* + * Ship a packet out. The appropriate raw output + * routine handles any massaging necessary. + */ + case PRU_SEND: + if (nam) { + if (rp->rcb_faddr) { + error = EISCONN; + break; + } + rp->rcb_faddr = mtod(nam, struct sockaddr *); + } else if (rp->rcb_faddr == 0) { + error = ENOTCONN; + break; + } + error = (*so->so_proto->pr_output)(m, so); + m = NULL; + if (nam) + rp->rcb_faddr = 0; + break; + + case PRU_ABORT: + raw_disconnect(rp); + sofree(so); + soisdisconnected(so); + break; + + case PRU_SENSE: + /* + * stat: don't bother with a blocksize. + */ + return (0); + + /* + * Not supported. + */ + case PRU_RCVOOB: + case PRU_RCVD: + return(EOPNOTSUPP); + + case PRU_LISTEN: + case PRU_ACCEPT: + case PRU_SENDOOB: + error = EOPNOTSUPP; + break; + + case PRU_SOCKADDR: + if (rp->rcb_laddr == 0) { + error = EINVAL; + break; + } + len = rp->rcb_laddr->sa_len; + bcopy((caddr_t)rp->rcb_laddr, mtod(nam, caddr_t), (unsigned)len); + nam->m_len = len; + break; + + case PRU_PEERADDR: + if (rp->rcb_faddr == 0) { + error = ENOTCONN; + break; + } + len = rp->rcb_faddr->sa_len; + bcopy((caddr_t)rp->rcb_faddr, mtod(nam, caddr_t), (unsigned)len); + nam->m_len = len; + break; + + default: + panic("raw_usrreq"); + } +release: + if (m != NULL) + m_freem(m); + return (error); +} diff --git a/sys/net/route.c b/sys/net/route.c new file mode 100644 index 00000000000..80a0cda6493 --- /dev/null +++ b/sys/net/route.c @@ -0,0 +1,539 @@ +/* $NetBSD: route.c,v 1.13 1995/08/12 23:59:25 mycroft Exp $ */ + +/* + * Copyright (c) 1980, 1986, 1991, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)route.c 8.2 (Berkeley) 11/15/93 + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/proc.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/socketvar.h> +#include <sys/domain.h> +#include <sys/protosw.h> +#include <sys/ioctl.h> + +#include <net/if.h> +#include <net/route.h> +#include <net/raw_cb.h> + +#include <netinet/in.h> +#include <netinet/in_var.h> + +#ifdef NS +#include <netns/ns.h> +#endif + +#define SA(p) ((struct sockaddr *)(p)) + +int rttrash; /* routes not in table but not freed */ +struct sockaddr wildcard; /* zero valued cookie for wildcard searches */ + +void +rtable_init(table) + void **table; +{ + struct domain *dom; + for (dom = domains; dom; dom = dom->dom_next) + if (dom->dom_rtattach) + dom->dom_rtattach(&table[dom->dom_family], + dom->dom_rtoffset); +} + +void +route_init() +{ + rn_init(); /* initialize all zeroes, all ones, mask table */ + rtable_init((void **)rt_tables); +} + +/* + * Packet routing routines. + */ +void +rtalloc(ro) + register struct route *ro; +{ + if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP)) + return; /* XXX */ + ro->ro_rt = rtalloc1(&ro->ro_dst, 1); +} + +struct rtentry * +rtalloc1(dst, report) + register struct sockaddr *dst; + int report; +{ + register struct radix_node_head *rnh = rt_tables[dst->sa_family]; + register struct rtentry *rt; + register struct radix_node *rn; + struct rtentry *newrt = 0; + struct rt_addrinfo info; + int s = splsoftnet(), err = 0, msgtype = RTM_MISS; + + if (rnh && (rn = rnh->rnh_matchaddr((caddr_t)dst, rnh)) && + ((rn->rn_flags & RNF_ROOT) == 0)) { + newrt = rt = (struct rtentry *)rn; + if (report && (rt->rt_flags & RTF_CLONING)) { + err = rtrequest(RTM_RESOLVE, dst, SA(0), + SA(0), 0, &newrt); + if (err) { + newrt = rt; + rt->rt_refcnt++; + goto miss; + } + if ((rt = newrt) && (rt->rt_flags & RTF_XRESOLVE)) { + msgtype = RTM_RESOLVE; + goto miss; + } + } else + rt->rt_refcnt++; + } else { + rtstat.rts_unreach++; + miss: if (report) { + bzero((caddr_t)&info, sizeof(info)); + info.rti_info[RTAX_DST] = dst; + rt_missmsg(msgtype, &info, 0, err); + } + } + splx(s); + return (newrt); +} + +void +rtfree(rt) + register struct rtentry *rt; +{ + register struct ifaddr *ifa; + + if (rt == 0) + panic("rtfree"); + rt->rt_refcnt--; + if (rt->rt_refcnt <= 0 && (rt->rt_flags & RTF_UP) == 0) { + if (rt->rt_nodes->rn_flags & (RNF_ACTIVE | RNF_ROOT)) + panic ("rtfree 2"); + rttrash--; + if (rt->rt_refcnt < 0) { + printf("rtfree: %x not freed (neg refs)\n", rt); + return; + } + ifa = rt->rt_ifa; + IFAFREE(ifa); + Free(rt_key(rt)); + Free(rt); + } +} + +void +ifafree(ifa) + register struct ifaddr *ifa; +{ + if (ifa == NULL) + panic("ifafree"); + if (ifa->ifa_refcnt == 0) + free(ifa, M_IFADDR); + else + ifa->ifa_refcnt--; +} + +/* + * Force a routing table entry to the specified + * destination to go through the given gateway. + * Normally called as a result of a routing redirect + * message from the network layer. + * + * N.B.: must be called at splsoftnet + */ +int +rtredirect(dst, gateway, netmask, flags, src, rtp) + struct sockaddr *dst, *gateway, *netmask, *src; + int flags; + struct rtentry **rtp; +{ + register struct rtentry *rt; + int error = 0; + short *stat = 0; + struct rt_addrinfo info; + struct ifaddr *ifa; + + /* verify the gateway is directly reachable */ + if ((ifa = ifa_ifwithnet(gateway)) == 0) { + error = ENETUNREACH; + goto out; + } + rt = rtalloc1(dst, 0); + /* + * If the redirect isn't from our current router for this dst, + * it's either old or wrong. If it redirects us to ourselves, + * we have a routing loop, perhaps as a result of an interface + * going down recently. + */ +#define equal(a1, a2) (bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0) + if (!(flags & RTF_DONE) && rt && + (!equal(src, rt->rt_gateway) || rt->rt_ifa != ifa)) + error = EINVAL; + else if (ifa_ifwithaddr(gateway)) + error = EHOSTUNREACH; + if (error) + goto done; + /* + * Create a new entry if we just got back a wildcard entry + * or the the lookup failed. This is necessary for hosts + * which use routing redirects generated by smart gateways + * to dynamically build the routing tables. + */ + if ((rt == 0) || (rt_mask(rt) && rt_mask(rt)->sa_len < 2)) + goto create; + /* + * Don't listen to the redirect if it's + * for a route to an interface. + */ + if (rt->rt_flags & RTF_GATEWAY) { + if (((rt->rt_flags & RTF_HOST) == 0) && (flags & RTF_HOST)) { + /* + * Changing from route to net => route to host. + * Create new route, rather than smashing route to net. + */ + create: + flags |= RTF_GATEWAY | RTF_DYNAMIC; + error = rtrequest((int)RTM_ADD, dst, gateway, + netmask, flags, + (struct rtentry **)0); + stat = &rtstat.rts_dynamic; + } else { + /* + * Smash the current notion of the gateway to + * this destination. Should check about netmask!!! + */ + rt->rt_flags |= RTF_MODIFIED; + flags |= RTF_MODIFIED; + stat = &rtstat.rts_newgateway; + rt_setgate(rt, rt_key(rt), gateway); + } + } else + error = EHOSTUNREACH; +done: + if (rt) { + if (rtp && !error) + *rtp = rt; + else + rtfree(rt); + } +out: + if (error) + rtstat.rts_badredirect++; + else if (stat != NULL) + (*stat)++; + bzero((caddr_t)&info, sizeof(info)); + info.rti_info[RTAX_DST] = dst; + info.rti_info[RTAX_GATEWAY] = gateway; + info.rti_info[RTAX_NETMASK] = netmask; + info.rti_info[RTAX_AUTHOR] = src; + rt_missmsg(RTM_REDIRECT, &info, flags, error); +} + +/* +* Routing table ioctl interface. +*/ +int +rtioctl(req, data, p) + u_long req; + caddr_t data; + struct proc *p; +{ + return (EOPNOTSUPP); +} + +struct ifaddr * +ifa_ifwithroute(flags, dst, gateway) + int flags; + struct sockaddr *dst, *gateway; +{ + register struct ifaddr *ifa; + if ((flags & RTF_GATEWAY) == 0) { + /* + * If we are adding a route to an interface, + * and the interface is a pt to pt link + * we should search for the destination + * as our clue to the interface. Otherwise + * we can use the local address. + */ + ifa = 0; + if (flags & RTF_HOST) + ifa = ifa_ifwithdstaddr(dst); + if (ifa == 0) + ifa = ifa_ifwithaddr(gateway); + } else { + /* + * If we are adding a route to a remote net + * or host, the gateway may still be on the + * other end of a pt to pt link. + */ + ifa = ifa_ifwithdstaddr(gateway); + } + if (ifa == 0) + ifa = ifa_ifwithnet(gateway); + if (ifa == 0) { + struct rtentry *rt = rtalloc1(dst, 0); + if (rt == 0) + return (0); + rt->rt_refcnt--; + if ((ifa = rt->rt_ifa) == 0) + return (0); + } + if (ifa->ifa_addr->sa_family != dst->sa_family) { + struct ifaddr *oifa = ifa; + ifa = ifaof_ifpforaddr(dst, ifa->ifa_ifp); + if (ifa == 0) + ifa = oifa; + } + return (ifa); +} + +#define ROUNDUP(a) (a>0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) + +int +rtrequest(req, dst, gateway, netmask, flags, ret_nrt) + int req, flags; + struct sockaddr *dst, *gateway, *netmask; + struct rtentry **ret_nrt; +{ + int s = splsoftnet(); int error = 0; + register struct rtentry *rt; + register struct radix_node *rn; + register struct radix_node_head *rnh; + struct ifaddr *ifa; + struct sockaddr *ndst; +#define senderr(x) { error = x ; goto bad; } + + if ((rnh = rt_tables[dst->sa_family]) == 0) + senderr(ESRCH); + if (flags & RTF_HOST) + netmask = 0; + switch (req) { + case RTM_DELETE: + if ((rn = rnh->rnh_deladdr(dst, netmask, rnh)) == 0) + senderr(ESRCH); + if (rn->rn_flags & (RNF_ACTIVE | RNF_ROOT)) + panic ("rtrequest delete"); + rt = (struct rtentry *)rn; + rt->rt_flags &= ~RTF_UP; + if (rt->rt_gwroute) { + rt = rt->rt_gwroute; RTFREE(rt); + (rt = (struct rtentry *)rn)->rt_gwroute = 0; + } + if ((ifa = rt->rt_ifa) && ifa->ifa_rtrequest) + ifa->ifa_rtrequest(RTM_DELETE, rt, SA(0)); + rttrash++; + if (ret_nrt) + *ret_nrt = rt; + else if (rt->rt_refcnt <= 0) { + rt->rt_refcnt++; + rtfree(rt); + } + break; + + case RTM_RESOLVE: + if (ret_nrt == 0 || (rt = *ret_nrt) == 0) + senderr(EINVAL); + ifa = rt->rt_ifa; + flags = rt->rt_flags & ~RTF_CLONING; + gateway = rt->rt_gateway; + if ((netmask = rt->rt_genmask) == 0) + flags |= RTF_HOST; + goto makeroute; + + case RTM_ADD: + if ((ifa = ifa_ifwithroute(flags, dst, gateway)) == 0) + senderr(ENETUNREACH); + makeroute: + R_Malloc(rt, struct rtentry *, sizeof(*rt)); + if (rt == 0) + senderr(ENOBUFS); + Bzero(rt, sizeof(*rt)); + rt->rt_flags = RTF_UP | flags; + if (rt_setgate(rt, dst, gateway)) { + Free(rt); + senderr(ENOBUFS); + } + ndst = rt_key(rt); + if (netmask) { + rt_maskedcopy(dst, ndst, netmask); + } else + Bcopy(dst, ndst, dst->sa_len); + rn = rnh->rnh_addaddr((caddr_t)ndst, (caddr_t)netmask, + rnh, rt->rt_nodes); + if (rn == 0) { + if (rt->rt_gwroute) + rtfree(rt->rt_gwroute); + Free(rt_key(rt)); + Free(rt); + senderr(EEXIST); + } + ifa->ifa_refcnt++; + rt->rt_ifa = ifa; + rt->rt_ifp = ifa->ifa_ifp; + if (req == RTM_RESOLVE) + rt->rt_rmx = (*ret_nrt)->rt_rmx; /* copy metrics */ + if (ifa->ifa_rtrequest) + ifa->ifa_rtrequest(req, rt, SA(ret_nrt ? *ret_nrt : 0)); + if (ret_nrt) { + *ret_nrt = rt; + rt->rt_refcnt++; + } + break; + } +bad: + splx(s); + return (error); +} + +int +rt_setgate(rt0, dst, gate) + struct rtentry *rt0; + struct sockaddr *dst, *gate; +{ + caddr_t new, old; + int dlen = ROUNDUP(dst->sa_len), glen = ROUNDUP(gate->sa_len); + register struct rtentry *rt = rt0; + + if (rt->rt_gateway == 0 || glen > ROUNDUP(rt->rt_gateway->sa_len)) { + old = (caddr_t)rt_key(rt); + R_Malloc(new, caddr_t, dlen + glen); + if (new == 0) + return 1; + rt->rt_nodes->rn_key = new; + } else { + new = rt->rt_nodes->rn_key; + old = 0; + } + Bcopy(gate, (rt->rt_gateway = (struct sockaddr *)(new + dlen)), glen); + if (old) { + Bcopy(dst, new, dlen); + Free(old); + } + if (rt->rt_gwroute) { + rt = rt->rt_gwroute; RTFREE(rt); + rt = rt0; rt->rt_gwroute = 0; + } + if (rt->rt_flags & RTF_GATEWAY) { + rt->rt_gwroute = rtalloc1(gate, 1); + } + return 0; +} + +void +rt_maskedcopy(src, dst, netmask) + struct sockaddr *src, *dst, *netmask; +{ + register u_char *cp1 = (u_char *)src; + register u_char *cp2 = (u_char *)dst; + register u_char *cp3 = (u_char *)netmask; + u_char *cplim = cp2 + *cp3; + u_char *cplim2 = cp2 + *cp1; + + *cp2++ = *cp1++; *cp2++ = *cp1++; /* copies sa_len & sa_family */ + cp3 += 2; + if (cplim > cplim2) + cplim = cplim2; + while (cp2 < cplim) + *cp2++ = *cp1++ & *cp3++; + if (cp2 < cplim2) + bzero((caddr_t)cp2, (unsigned)(cplim2 - cp2)); +} + +/* + * Set up a routing table entry, normally + * for an interface. + */ +int +rtinit(ifa, cmd, flags) + register struct ifaddr *ifa; + int cmd, flags; +{ + register struct rtentry *rt; + register struct sockaddr *dst; + register struct sockaddr *deldst; + struct mbuf *m = 0; + struct rtentry *nrt = 0; + int error; + + dst = flags & RTF_HOST ? ifa->ifa_dstaddr : ifa->ifa_addr; + if (cmd == RTM_DELETE) { + if ((flags & RTF_HOST) == 0 && ifa->ifa_netmask) { + m = m_get(M_WAIT, MT_SONAME); + deldst = mtod(m, struct sockaddr *); + rt_maskedcopy(dst, deldst, ifa->ifa_netmask); + dst = deldst; + } + if (rt = rtalloc1(dst, 0)) { + rt->rt_refcnt--; + if (rt->rt_ifa != ifa) { + if (m) + (void) m_free(m); + return (flags & RTF_HOST ? EHOSTUNREACH + : ENETUNREACH); + } + } + } + error = rtrequest(cmd, dst, ifa->ifa_addr, ifa->ifa_netmask, + flags | ifa->ifa_flags, &nrt); + if (m) + (void) m_free(m); + if (cmd == RTM_DELETE && error == 0 && (rt = nrt)) { + rt_newaddrmsg(cmd, ifa, error, nrt); + if (rt->rt_refcnt <= 0) { + rt->rt_refcnt++; + rtfree(rt); + } + } + if (cmd == RTM_ADD && error == 0 && (rt = nrt)) { + rt->rt_refcnt--; + if (rt->rt_ifa != ifa) { + printf("rtinit: wrong ifa (%x) was (%x)\n", ifa, + rt->rt_ifa); + if (rt->rt_ifa->ifa_rtrequest) + rt->rt_ifa->ifa_rtrequest(RTM_DELETE, rt, SA(0)); + IFAFREE(rt->rt_ifa); + rt->rt_ifa = ifa; + rt->rt_ifp = ifa->ifa_ifp; + ifa->ifa_refcnt++; + if (ifa->ifa_rtrequest) + ifa->ifa_rtrequest(RTM_ADD, rt, SA(0)); + } + rt_newaddrmsg(cmd, ifa, error, nrt); + } + return (error); +} diff --git a/sys/net/route.h b/sys/net/route.h new file mode 100644 index 00000000000..0fc39fe4d3b --- /dev/null +++ b/sys/net/route.h @@ -0,0 +1,264 @@ +/* $NetBSD: route.h,v 1.8 1995/03/26 20:30:19 jtc Exp $ */ + +/* + * Copyright (c) 1980, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)route.h 8.3 (Berkeley) 4/19/94 + */ + +/* + * Kernel resident routing tables. + * + * The routing tables are initialized when interface addresses + * are set by making entries for all directly connected interfaces. + */ + +/* + * A route consists of a destination address and a reference + * to a routing entry. These are often held by protocols + * in their control blocks, e.g. inpcb. + */ +struct route { + struct rtentry *ro_rt; + struct sockaddr ro_dst; +}; + +/* + * These numbers are used by reliable protocols for determining + * retransmission behavior and are included in the routing structure. + */ +struct rt_metrics { + u_long rmx_locks; /* Kernel must leave these values alone */ + u_long rmx_mtu; /* MTU for this path */ + u_long rmx_hopcount; /* max hops expected */ + u_long rmx_expire; /* lifetime for route, e.g. redirect */ + u_long rmx_recvpipe; /* inbound delay-bandwith product */ + u_long rmx_sendpipe; /* outbound delay-bandwith product */ + u_long rmx_ssthresh; /* outbound gateway buffer limit */ + u_long rmx_rtt; /* estimated round trip time */ + u_long rmx_rttvar; /* estimated rtt variance */ + u_long rmx_pksent; /* packets sent using this route */ +}; + +/* + * rmx_rtt and rmx_rttvar are stored as microseconds; + * RTTTOPRHZ(rtt) converts to a value suitable for use + * by a protocol slowtimo counter. + */ +#define RTM_RTTUNIT 1000000 /* units for rtt, rttvar, as units per sec */ +#define RTTTOPRHZ(r) ((r) / (RTM_RTTUNIT / PR_SLOWHZ)) + +/* + * We distinguish between routes to hosts and routes to networks, + * preferring the former if available. For each route we infer + * the interface to use from the gateway address supplied when + * the route was entered. Routes that forward packets through + * gateways are marked so that the output routines know to address the + * gateway rather than the ultimate destination. + */ +#ifndef RNF_NORMAL +#include <net/radix.h> +#endif +struct rtentry { + struct radix_node rt_nodes[2]; /* tree glue, and other values */ +#define rt_key(r) ((struct sockaddr *)((r)->rt_nodes->rn_key)) +#define rt_mask(r) ((struct sockaddr *)((r)->rt_nodes->rn_mask)) + struct sockaddr *rt_gateway; /* value */ + short rt_flags; /* up/down?, host/net */ + short rt_refcnt; /* # held references */ + u_long rt_use; /* raw # packets forwarded */ + struct ifnet *rt_ifp; /* the answer: interface to use */ + struct ifaddr *rt_ifa; /* the answer: interface to use */ + struct sockaddr *rt_genmask; /* for generation of cloned routes */ + caddr_t rt_llinfo; /* pointer to link level info cache */ + struct rt_metrics rt_rmx; /* metrics used by rx'ing protocols */ + struct rtentry *rt_gwroute; /* implied entry for gatewayed routes */ +}; + +/* + * Following structure necessary for 4.3 compatibility; + * We should eventually move it to a compat file. + */ +struct ortentry { + u_int32_t rt_hash; /* to speed lookups */ + struct sockaddr rt_dst; /* key */ + struct sockaddr rt_gateway; /* value */ + int16_t rt_flags; /* up/down?, host/net */ + int16_t rt_refcnt; /* # held references */ + u_int32_t rt_use; /* raw # packets forwarded */ + struct ifnet *rt_ifp; /* the answer: interface to use */ +}; + +#define RTF_UP 0x1 /* route usable */ +#define RTF_GATEWAY 0x2 /* destination is a gateway */ +#define RTF_HOST 0x4 /* host entry (net otherwise) */ +#define RTF_REJECT 0x8 /* host or net unreachable */ +#define RTF_DYNAMIC 0x10 /* created dynamically (by redirect) */ +#define RTF_MODIFIED 0x20 /* modified dynamically (by redirect) */ +#define RTF_DONE 0x40 /* message confirmed */ +#define RTF_MASK 0x80 /* subnet mask present */ +#define RTF_CLONING 0x100 /* generate new routes on use */ +#define RTF_XRESOLVE 0x200 /* external daemon resolves name */ +#define RTF_LLINFO 0x400 /* generated by ARP or ESIS */ +#define RTF_STATIC 0x800 /* manually added */ +#define RTF_BLACKHOLE 0x1000 /* just discard pkts (during updates) */ +#define RTF_PROTO2 0x4000 /* protocol specific routing flag */ +#define RTF_PROTO1 0x8000 /* protocol specific routing flag */ + + +/* + * Routing statistics. + */ +struct rtstat { + short rts_badredirect; /* bogus redirect calls */ + short rts_dynamic; /* routes created by redirects */ + short rts_newgateway; /* routes modified by redirects */ + short rts_unreach; /* lookups which failed */ + short rts_wildcard; /* lookups satisfied by a wildcard */ +}; +/* + * Structures for routing messages. + */ +struct rt_msghdr { + u_short rtm_msglen; /* to skip over non-understood messages */ + u_char rtm_version; /* future binary compatibility */ + u_char rtm_type; /* message type */ + u_short rtm_index; /* index for associated ifp */ + int rtm_flags; /* flags, incl. kern & message, e.g. DONE */ + int rtm_addrs; /* bitmask identifying sockaddrs in msg */ + pid_t rtm_pid; /* identify sender */ + int rtm_seq; /* for sender to identify action */ + int rtm_errno; /* why failed */ + int rtm_use; /* from rtentry */ + u_long rtm_inits; /* which metrics we are initializing */ + struct rt_metrics rtm_rmx; /* metrics themselves */ +}; + +#define RTM_VERSION 3 /* Up the ante and ignore older versions */ + +#define RTM_ADD 0x1 /* Add Route */ +#define RTM_DELETE 0x2 /* Delete Route */ +#define RTM_CHANGE 0x3 /* Change Metrics or flags */ +#define RTM_GET 0x4 /* Report Metrics */ +#define RTM_LOSING 0x5 /* Kernel Suspects Partitioning */ +#define RTM_REDIRECT 0x6 /* Told to use different route */ +#define RTM_MISS 0x7 /* Lookup failed on this address */ +#define RTM_LOCK 0x8 /* fix specified metrics */ +#define RTM_OLDADD 0x9 /* caused by SIOCADDRT */ +#define RTM_OLDDEL 0xa /* caused by SIOCDELRT */ +#define RTM_RESOLVE 0xb /* req to resolve dst to LL addr */ +#define RTM_NEWADDR 0xc /* address being added to iface */ +#define RTM_DELADDR 0xd /* address being removed from iface */ +#define RTM_IFINFO 0xe /* iface going up/down etc. */ + +#define RTV_MTU 0x1 /* init or lock _mtu */ +#define RTV_HOPCOUNT 0x2 /* init or lock _hopcount */ +#define RTV_EXPIRE 0x4 /* init or lock _hopcount */ +#define RTV_RPIPE 0x8 /* init or lock _recvpipe */ +#define RTV_SPIPE 0x10 /* init or lock _sendpipe */ +#define RTV_SSTHRESH 0x20 /* init or lock _ssthresh */ +#define RTV_RTT 0x40 /* init or lock _rtt */ +#define RTV_RTTVAR 0x80 /* init or lock _rttvar */ + +/* + * Bitmask values for rtm_addr. + */ +#define RTA_DST 0x1 /* destination sockaddr present */ +#define RTA_GATEWAY 0x2 /* gateway sockaddr present */ +#define RTA_NETMASK 0x4 /* netmask sockaddr present */ +#define RTA_GENMASK 0x8 /* cloning mask sockaddr present */ +#define RTA_IFP 0x10 /* interface name sockaddr present */ +#define RTA_IFA 0x20 /* interface addr sockaddr present */ +#define RTA_AUTHOR 0x40 /* sockaddr for author of redirect */ +#define RTA_BRD 0x80 /* for NEWADDR, broadcast or p-p dest addr */ + +/* + * Index offsets for sockaddr array for alternate internal encoding. + */ +#define RTAX_DST 0 /* destination sockaddr present */ +#define RTAX_GATEWAY 1 /* gateway sockaddr present */ +#define RTAX_NETMASK 2 /* netmask sockaddr present */ +#define RTAX_GENMASK 3 /* cloning mask sockaddr present */ +#define RTAX_IFP 4 /* interface name sockaddr present */ +#define RTAX_IFA 5 /* interface addr sockaddr present */ +#define RTAX_AUTHOR 6 /* sockaddr for author of redirect */ +#define RTAX_BRD 7 /* for NEWADDR, broadcast or p-p dest addr */ +#define RTAX_MAX 8 /* size of array to allocate */ + +struct rt_addrinfo { + int rti_addrs; + struct sockaddr *rti_info[RTAX_MAX]; +}; + +struct route_cb { + int ip_count; + int ns_count; + int iso_count; + int any_count; +}; + +#ifdef _KERNEL +#define RTFREE(rt) \ + if ((rt)->rt_refcnt <= 1) \ + rtfree(rt); \ + else \ + (rt)->rt_refcnt--; + +struct route_cb route_cb; +struct rtstat rtstat; +struct radix_node_head *rt_tables[AF_MAX+1]; + +struct socket; +void route_init __P((void)); +int route_output __P((struct mbuf *, struct socket *)); +int route_usrreq __P((struct socket *, + int, struct mbuf *, struct mbuf *, struct mbuf *)); +void rt_ifmsg __P((struct ifnet *)); +void rt_maskedcopy __P((struct sockaddr *, + struct sockaddr *, struct sockaddr *)); +void rt_missmsg __P((int, struct rt_addrinfo *, int, int)); +void rt_newaddrmsg __P((int, struct ifaddr *, int, struct rtentry *)); +int rt_setgate __P((struct rtentry *, + struct sockaddr *, struct sockaddr *)); +void rt_setmetrics __P((u_long, struct rt_metrics *, struct rt_metrics *)); +void rtable_init __P((void **)); +void rtalloc __P((struct route *)); +struct rtentry * + rtalloc1 __P((struct sockaddr *, int)); +void rtfree __P((struct rtentry *)); +int rtinit __P((struct ifaddr *, int, int)); +int rtioctl __P((u_long, caddr_t, struct proc *)); +int rtredirect __P((struct sockaddr *, struct sockaddr *, + struct sockaddr *, int, struct sockaddr *, struct rtentry **)); +int rtrequest __P((int, struct sockaddr *, + struct sockaddr *, struct sockaddr *, int, struct rtentry **)); +#endif /* _KERNEL */ diff --git a/sys/net/rtsock.c b/sys/net/rtsock.c new file mode 100644 index 00000000000..79e6921cb5e --- /dev/null +++ b/sys/net/rtsock.c @@ -0,0 +1,839 @@ +/* $NetBSD: rtsock.c,v 1.16 1995/08/19 07:48:14 cgd Exp $ */ + +/* + * Copyright (c) 1988, 1991, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)rtsock.c 8.6 (Berkeley) 2/11/95 + */ + +#include <sys/param.h> +#include <sys/systm.h> +#include <sys/proc.h> +#include <sys/mbuf.h> +#include <sys/socket.h> +#include <sys/socketvar.h> +#include <sys/domain.h> +#include <sys/protosw.h> + +#include <net/if.h> +#include <net/route.h> +#include <net/raw_cb.h> + +struct sockaddr route_dst = { 2, PF_ROUTE, }; +struct sockaddr route_src = { 2, PF_ROUTE, }; +struct sockproto route_proto = { PF_ROUTE, }; + +struct walkarg { + int w_op, w_arg, w_given, w_needed, w_tmemsize; + caddr_t w_where, w_tmem; +}; + +static struct mbuf * + rt_msg1 __P((int, struct rt_addrinfo *)); +static int rt_msg2 __P((int, + struct rt_addrinfo *, caddr_t, struct walkarg *)); +static void rt_xaddrs __P((caddr_t, caddr_t, struct rt_addrinfo *)); + +/* Sleazy use of local variables throughout file, warning!!!! */ +#define dst info.rti_info[RTAX_DST] +#define gate info.rti_info[RTAX_GATEWAY] +#define netmask info.rti_info[RTAX_NETMASK] +#define genmask info.rti_info[RTAX_GENMASK] +#define ifpaddr info.rti_info[RTAX_IFP] +#define ifaaddr info.rti_info[RTAX_IFA] +#define brdaddr info.rti_info[RTAX_BRD] + +/*ARGSUSED*/ +int +route_usrreq(so, req, m, nam, control) + register struct socket *so; + int req; + struct mbuf *m, *nam, *control; +{ + register int error = 0; + register struct rawcb *rp = sotorawcb(so); + int s; + + if (req == PRU_ATTACH) { + MALLOC(rp, struct rawcb *, sizeof(*rp), M_PCB, M_WAITOK); + if (so->so_pcb = rp) + bzero((caddr_t)so->so_pcb, sizeof(*rp)); + + } + if (req == PRU_DETACH && rp) { + int af = rp->rcb_proto.sp_protocol; + if (af == AF_INET) + route_cb.ip_count--; + else if (af == AF_NS) + route_cb.ns_count--; + else if (af == AF_ISO) + route_cb.iso_count--; + route_cb.any_count--; + } + s = splsoftnet(); + error = raw_usrreq(so, req, m, nam, control); + rp = sotorawcb(so); + if (req == PRU_ATTACH && rp) { + int af = rp->rcb_proto.sp_protocol; + if (error) { + free((caddr_t)rp, M_PCB); + splx(s); + return (error); + } + if (af == AF_INET) + route_cb.ip_count++; + else if (af == AF_NS) + route_cb.ns_count++; + else if (af == AF_ISO) + route_cb.iso_count++; + rp->rcb_faddr = &route_src; + route_cb.any_count++; + soisconnected(so); + so->so_options |= SO_USELOOPBACK; + } + splx(s); + return (error); +} + +/*ARGSUSED*/ +int +route_output(m, so) + register struct mbuf *m; + struct socket *so; +{ + register struct rt_msghdr *rtm = 0; + register struct rtentry *rt = 0; + struct rtentry *saved_nrt = 0; + struct radix_node_head *rnh; + struct rt_addrinfo info; + int len, error = 0; + struct ifnet *ifp = 0; + struct ifaddr *ifa = 0; + +#define senderr(e) { error = e; goto flush;} + if (m == 0 || ((m->m_len < sizeof(int32_t)) && + (m = m_pullup(m, sizeof(int32_t))) == 0)) + return (ENOBUFS); + if ((m->m_flags & M_PKTHDR) == 0) + panic("route_output"); + len = m->m_pkthdr.len; + if (len < sizeof(*rtm) || + len != mtod(m, struct rt_msghdr *)->rtm_msglen) { + dst = 0; + senderr(EINVAL); + } + R_Malloc(rtm, struct rt_msghdr *, len); + if (rtm == 0) { + dst = 0; + senderr(ENOBUFS); + } + m_copydata(m, 0, len, (caddr_t)rtm); + if (rtm->rtm_version != RTM_VERSION) { + dst = 0; + senderr(EPROTONOSUPPORT); + } + rtm->rtm_pid = curproc->p_pid; + info.rti_addrs = rtm->rtm_addrs; + rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info); + if (dst == 0) + senderr(EINVAL); + if (genmask) { + struct radix_node *t; + t = rn_addmask((caddr_t)genmask, 0, 1); + if (t && Bcmp(genmask, t->rn_key, *(u_char *)genmask) == 0) + genmask = (struct sockaddr *)(t->rn_key); + else + senderr(ENOBUFS); + } + switch (rtm->rtm_type) { + + case RTM_ADD: + if (gate == 0) + senderr(EINVAL); + error = rtrequest(RTM_ADD, dst, gate, netmask, + rtm->rtm_flags, &saved_nrt); + if (error == 0 && saved_nrt) { + rt_setmetrics(rtm->rtm_inits, + &rtm->rtm_rmx, &saved_nrt->rt_rmx); + saved_nrt->rt_refcnt--; + saved_nrt->rt_genmask = genmask; + } + break; + + case RTM_DELETE: + error = rtrequest(RTM_DELETE, dst, gate, netmask, + rtm->rtm_flags, &saved_nrt); + if (error == 0) { + (rt = saved_nrt)->rt_refcnt++; + goto report; + } + break; + + case RTM_GET: + case RTM_CHANGE: + case RTM_LOCK: + if ((rnh = rt_tables[dst->sa_family]) == 0) { + senderr(EAFNOSUPPORT); + } else if (rt = (struct rtentry *) + rnh->rnh_lookup(dst, netmask, rnh)) + rt->rt_refcnt++; + else + senderr(ESRCH); + switch(rtm->rtm_type) { + + case RTM_GET: + report: + dst = rt_key(rt); + gate = rt->rt_gateway; + netmask = rt_mask(rt); + genmask = rt->rt_genmask; + if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { + if (ifp = rt->rt_ifp) { + ifpaddr = ifp->if_addrlist.tqh_first->ifa_addr; + ifaaddr = rt->rt_ifa->ifa_addr; + if (ifp->if_flags & IFF_POINTOPOINT) + brdaddr = rt->rt_ifa->ifa_dstaddr; + else + brdaddr = 0; + rtm->rtm_index = ifp->if_index; + } else { + ifpaddr = 0; + ifaaddr = 0; + } + } + len = rt_msg2(rtm->rtm_type, &info, (caddr_t)0, + (struct walkarg *)0); + if (len > rtm->rtm_msglen) { + struct rt_msghdr *new_rtm; + R_Malloc(new_rtm, struct rt_msghdr *, len); + if (new_rtm == 0) + senderr(ENOBUFS); + Bcopy(rtm, new_rtm, rtm->rtm_msglen); + Free(rtm); rtm = new_rtm; + } + (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, + (struct walkarg *)0); + rtm->rtm_flags = rt->rt_flags; + rtm->rtm_rmx = rt->rt_rmx; + rtm->rtm_addrs = info.rti_addrs; + break; + + case RTM_CHANGE: + if (gate && rt_setgate(rt, rt_key(rt), gate)) + senderr(EDQUOT); + /* new gateway could require new ifaddr, ifp; + flags may also be different; ifp may be specified + by ll sockaddr when protocol address is ambiguous */ + if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) && + (ifp = ifa->ifa_ifp)) + ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate, + ifp); + else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) || + (ifa = ifa_ifwithroute(rt->rt_flags, + rt_key(rt), gate))) + ifp = ifa->ifa_ifp; + if (ifa) { + register struct ifaddr *oifa = rt->rt_ifa; + if (oifa != ifa) { + if (oifa && oifa->ifa_rtrequest) + oifa->ifa_rtrequest(RTM_DELETE, + rt, gate); + IFAFREE(rt->rt_ifa); + rt->rt_ifa = ifa; + ifa->ifa_refcnt++; + rt->rt_ifp = ifp; + } + } + rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, + &rt->rt_rmx); + if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) + rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, gate); + if (genmask) + rt->rt_genmask = genmask; + /* + * Fall into + */ + case RTM_LOCK: + rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); + rt->rt_rmx.rmx_locks |= + (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); + break; + } + break; + + default: + senderr(EOPNOTSUPP); + } + +flush: + if (rtm) { + if (error) + rtm->rtm_errno = error; + else + rtm->rtm_flags |= RTF_DONE; + } + if (rt) + rtfree(rt); + { + register struct rawcb *rp = 0; + /* + * Check to see if we don't want our own messages. + */ + if ((so->so_options & SO_USELOOPBACK) == 0) { + if (route_cb.any_count <= 1) { + if (rtm) + Free(rtm); + m_freem(m); + return (error); + } + /* There is another listener, so construct message */ + rp = sotorawcb(so); + } + if (rtm) { + m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); + Free(rtm); + } + if (rp) + rp->rcb_proto.sp_family = 0; /* Avoid us */ + if (dst) + route_proto.sp_protocol = dst->sa_family; + raw_input(m, &route_proto, &route_src, &route_dst); + if (rp) + rp->rcb_proto.sp_family = PF_ROUTE; + } + return (error); +} + +void +rt_setmetrics(which, in, out) + u_long which; + register struct rt_metrics *in, *out; +{ +#define metric(f, e) if (which & (f)) out->e = in->e; + metric(RTV_RPIPE, rmx_recvpipe); + metric(RTV_SPIPE, rmx_sendpipe); + metric(RTV_SSTHRESH, rmx_ssthresh); + metric(RTV_RTT, rmx_rtt); + metric(RTV_RTTVAR, rmx_rttvar); + metric(RTV_HOPCOUNT, rmx_hopcount); + metric(RTV_MTU, rmx_mtu); + metric(RTV_EXPIRE, rmx_expire); +#undef metric +} + +#define ROUNDUP(a) \ + ((a) > 0 ? (1 + (((a) - 1) | (sizeof(int32_t) - 1))) : sizeof(int32_t)) +#define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) + +static void +rt_xaddrs(cp, cplim, rtinfo) + register caddr_t cp, cplim; + register struct rt_addrinfo *rtinfo; +{ + register struct sockaddr *sa; + register int i; + + bzero(rtinfo->rti_info, sizeof(rtinfo->rti_info)); + for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { + if ((rtinfo->rti_addrs & (1 << i)) == 0) + continue; + rtinfo->rti_info[i] = sa = (struct sockaddr *)cp; + ADVANCE(cp, sa); + } +} + +/* + * Copy data from a buffer back into the indicated mbuf chain, + * starting "off" bytes from the beginning, extending the mbuf + * chain if necessary. + */ +void +m_copyback(m0, off, len, cp) + struct mbuf *m0; + register int off; + register int len; + caddr_t cp; +{ + register int mlen; + register struct mbuf *m = m0, *n; + int totlen = 0; + + if (m0 == 0) + return; + while (off > (mlen = m->m_len)) { + off -= mlen; + totlen += mlen; + if (m->m_next == 0) { + n = m_getclr(M_DONTWAIT, m->m_type); + if (n == 0) + goto out; + n->m_len = min(MLEN, len + off); + m->m_next = n; + } + m = m->m_next; + } + while (len > 0) { + mlen = min (m->m_len - off, len); + bcopy(cp, off + mtod(m, caddr_t), (unsigned)mlen); + cp += mlen; + len -= mlen; + mlen += off; + off = 0; + totlen += mlen; + if (len == 0) + break; + if (m->m_next == 0) { + n = m_get(M_DONTWAIT, m->m_type); + if (n == 0) + break; + n->m_len = min(MLEN, len); + m->m_next = n; + } + m = m->m_next; + } +out: if (((m = m0)->m_flags & M_PKTHDR) && (m->m_pkthdr.len < totlen)) + m->m_pkthdr.len = totlen; +} + +static struct mbuf * +rt_msg1(type, rtinfo) + int type; + register struct rt_addrinfo *rtinfo; +{ + register struct rt_msghdr *rtm; + register struct mbuf *m; + register int i; + register struct sockaddr *sa; + int len, dlen; + + m = m_gethdr(M_DONTWAIT, MT_DATA); + if (m == 0) + return (m); + switch (type) { + + case RTM_DELADDR: + case RTM_NEWADDR: + len = sizeof(struct ifa_msghdr); + break; + + case RTM_IFINFO: + len = sizeof(struct if_msghdr); + break; + + default: + len = sizeof(struct rt_msghdr); + } + if (len > MHLEN) + panic("rt_msg1"); + m->m_pkthdr.len = m->m_len = len; + m->m_pkthdr.rcvif = 0; + rtm = mtod(m, struct rt_msghdr *); + bzero((caddr_t)rtm, len); + for (i = 0; i < RTAX_MAX; i++) { + if ((sa = rtinfo->rti_info[i]) == NULL) + continue; + rtinfo->rti_addrs |= (1 << i); + dlen = ROUNDUP(sa->sa_len); + m_copyback(m, len, dlen, (caddr_t)sa); + len += dlen; + } + if (m->m_pkthdr.len != len) { + m_freem(m); + return (NULL); + } + rtm->rtm_msglen = len; + rtm->rtm_version = RTM_VERSION; + rtm->rtm_type = type; + return (m); +} + +static int +rt_msg2(type, rtinfo, cp, w) + int type; + register struct rt_addrinfo *rtinfo; + caddr_t cp; + struct walkarg *w; +{ + register int i; + int len, dlen, second_time = 0; + caddr_t cp0; + + rtinfo->rti_addrs = 0; +again: + switch (type) { + + case RTM_DELADDR: + case RTM_NEWADDR: + len = sizeof(struct ifa_msghdr); + break; + + case RTM_IFINFO: + len = sizeof(struct if_msghdr); + break; + + default: + len = sizeof(struct rt_msghdr); + } + if (cp0 = cp) + cp += len; + for (i = 0; i < RTAX_MAX; i++) { + register struct sockaddr *sa; + + if ((sa = rtinfo->rti_info[i]) == 0) + continue; + rtinfo->rti_addrs |= (1 << i); + dlen = ROUNDUP(sa->sa_len); + if (cp) { + bcopy((caddr_t)sa, cp, (unsigned)dlen); + cp += dlen; + } + len += dlen; + } + if (cp == 0 && w != NULL && !second_time) { + register struct walkarg *rw = w; + + rw->w_needed += len; + if (rw->w_needed <= 0 && rw->w_where) { + if (rw->w_tmemsize < len) { + if (rw->w_tmem) + free(rw->w_tmem, M_RTABLE); + if (rw->w_tmem = (caddr_t) + malloc(len, M_RTABLE, M_NOWAIT)) + rw->w_tmemsize = len; + } + if (rw->w_tmem) { + cp = rw->w_tmem; + second_time = 1; + goto again; + } else + rw->w_where = 0; + } + } + if (cp) { + register struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; + + rtm->rtm_version = RTM_VERSION; + rtm->rtm_type = type; + rtm->rtm_msglen = len; + } + return (len); +} + +/* + * This routine is called to generate a message from the routing + * socket indicating that a redirect has occured, a routing lookup + * has failed, or that a protocol has detected timeouts to a particular + * destination. + */ +void +rt_missmsg(type, rtinfo, flags, error) + int type, flags, error; + register struct rt_addrinfo *rtinfo; +{ + register struct rt_msghdr *rtm; + register struct mbuf *m; + struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; + + if (route_cb.any_count == 0) + return; + m = rt_msg1(type, rtinfo); + if (m == 0) + return; + rtm = mtod(m, struct rt_msghdr *); + rtm->rtm_flags = RTF_DONE | flags; + rtm->rtm_errno = error; + rtm->rtm_addrs = rtinfo->rti_addrs; + route_proto.sp_protocol = sa ? sa->sa_family : 0; + raw_input(m, &route_proto, &route_src, &route_dst); +} + +/* + * This routine is called to generate a message from the routing + * socket indicating that the status of a network interface has changed. + */ +void +rt_ifmsg(ifp) + register struct ifnet *ifp; +{ + register struct if_msghdr *ifm; + struct mbuf *m; + struct rt_addrinfo info; + + if (route_cb.any_count == 0) + return; + bzero((caddr_t)&info, sizeof(info)); + m = rt_msg1(RTM_IFINFO, &info); + if (m == 0) + return; + ifm = mtod(m, struct if_msghdr *); + ifm->ifm_index = ifp->if_index; + ifm->ifm_flags = ifp->if_flags; + ifm->ifm_data = ifp->if_data; + ifm->ifm_addrs = 0; + route_proto.sp_protocol = 0; + raw_input(m, &route_proto, &route_src, &route_dst); +} + +/* + * This is called to generate messages from the routing socket + * indicating a network interface has had addresses associated with it. + * if we ever reverse the logic and replace messages TO the routing + * socket indicate a request to configure interfaces, then it will + * be unnecessary as the routing socket will automatically generate + * copies of it. + */ +void +rt_newaddrmsg(cmd, ifa, error, rt) + int cmd, error; + register struct ifaddr *ifa; + register struct rtentry *rt; +{ + struct rt_addrinfo info; + struct sockaddr *sa; + int pass; + struct mbuf *m; + struct ifnet *ifp = ifa->ifa_ifp; + + if (route_cb.any_count == 0) + return; + for (pass = 1; pass < 3; pass++) { + bzero((caddr_t)&info, sizeof(info)); + if ((cmd == RTM_ADD && pass == 1) || + (cmd == RTM_DELETE && pass == 2)) { + register struct ifa_msghdr *ifam; + int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; + + ifaaddr = sa = ifa->ifa_addr; + ifpaddr = ifp->if_addrlist.tqh_first->ifa_addr; + netmask = ifa->ifa_netmask; + brdaddr = ifa->ifa_dstaddr; + if ((m = rt_msg1(ncmd, &info)) == NULL) + continue; + ifam = mtod(m, struct ifa_msghdr *); + ifam->ifam_index = ifp->if_index; + ifam->ifam_metric = ifa->ifa_metric; + ifam->ifam_flags = ifa->ifa_flags; + ifam->ifam_addrs = info.rti_addrs; + } + if ((cmd == RTM_ADD && pass == 2) || + (cmd == RTM_DELETE && pass == 1)) { + register struct rt_msghdr *rtm; + + if (rt == 0) + continue; + netmask = rt_mask(rt); + dst = sa = rt_key(rt); + gate = rt->rt_gateway; + if ((m = rt_msg1(cmd, &info)) == NULL) + continue; + rtm = mtod(m, struct rt_msghdr *); + rtm->rtm_index = ifp->if_index; + rtm->rtm_flags |= rt->rt_flags; + rtm->rtm_errno = error; + rtm->rtm_addrs = info.rti_addrs; + } + route_proto.sp_protocol = sa ? sa->sa_family : 0; + raw_input(m, &route_proto, &route_src, &route_dst); + } +} + +/* + * This is used in dumping the kernel table via sysctl(). + */ +int +sysctl_dumpentry(rn, w) + struct radix_node *rn; + register struct walkarg *w; +{ + register struct rtentry *rt = (struct rtentry *)rn; + int error = 0, size; + struct rt_addrinfo info; + + if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) + return 0; + bzero((caddr_t)&info, sizeof(info)); + dst = rt_key(rt); + gate = rt->rt_gateway; + netmask = rt_mask(rt); + genmask = rt->rt_genmask; + if (rt->rt_ifp) { + ifpaddr = rt->rt_ifp->if_addrlist.tqh_first->ifa_addr; + ifaaddr = rt->rt_ifa->ifa_addr; + if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) + brdaddr = rt->rt_ifa->ifa_dstaddr; + } + size = rt_msg2(RTM_GET, &info, 0, w); + if (w->w_where && w->w_tmem) { + register struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; + + rtm->rtm_flags = rt->rt_flags; + rtm->rtm_use = rt->rt_use; + rtm->rtm_rmx = rt->rt_rmx; + rtm->rtm_index = rt->rt_ifp->if_index; + rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; + rtm->rtm_addrs = info.rti_addrs; + if (error = copyout((caddr_t)rtm, w->w_where, size)) + w->w_where = NULL; + else + w->w_where += size; + } + return (error); +} + +int +sysctl_iflist(af, w) + int af; + register struct walkarg *w; +{ + register struct ifnet *ifp; + register struct ifaddr *ifa; + struct rt_addrinfo info; + int len, error = 0; + + bzero((caddr_t)&info, sizeof(info)); + for (ifp = ifnet.tqh_first; ifp != 0; ifp = ifp->if_list.tqe_next) { + if (w->w_arg && w->w_arg != ifp->if_index) + continue; + ifa = ifp->if_addrlist.tqh_first; + ifpaddr = ifa->ifa_addr; + len = rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w); + ifpaddr = 0; + if (w->w_where && w->w_tmem) { + register struct if_msghdr *ifm; + + ifm = (struct if_msghdr *)w->w_tmem; + ifm->ifm_index = ifp->if_index; + ifm->ifm_flags = ifp->if_flags; + ifm->ifm_data = ifp->if_data; + ifm->ifm_addrs = info.rti_addrs; + if (error = copyout((caddr_t)ifm, w->w_where, len)) + return (error); + w->w_where += len; + } + while (ifa = ifa->ifa_list.tqe_next) { + if (af && af != ifa->ifa_addr->sa_family) + continue; + ifaaddr = ifa->ifa_addr; + netmask = ifa->ifa_netmask; + brdaddr = ifa->ifa_dstaddr; + len = rt_msg2(RTM_NEWADDR, &info, 0, w); + if (w->w_where && w->w_tmem) { + register struct ifa_msghdr *ifam; + + ifam = (struct ifa_msghdr *)w->w_tmem; + ifam->ifam_index = ifa->ifa_ifp->if_index; + ifam->ifam_flags = ifa->ifa_flags; + ifam->ifam_metric = ifa->ifa_metric; + ifam->ifam_addrs = info.rti_addrs; + if (error = copyout(w->w_tmem, w->w_where, len)) + return (error); + w->w_where += len; + } + } + ifaaddr = netmask = brdaddr = 0; + } + return (0); +} + +int +sysctl_rtable(name, namelen, where, given, new, newlen) + int *name; + int namelen; + caddr_t where; + size_t *given; + caddr_t *new; + size_t newlen; +{ + register struct radix_node_head *rnh; + int i, s, error = EINVAL; + u_char af; + struct walkarg w; + + if (new) + return (EPERM); + if (namelen != 3) + return (EINVAL); + af = name[0]; + Bzero(&w, sizeof(w)); + w.w_where = where; + w.w_given = *given; + w.w_needed = 0 - w.w_given; + w.w_op = name[1]; + w.w_arg = name[2]; + + s = splsoftnet(); + switch (w.w_op) { + + case NET_RT_DUMP: + case NET_RT_FLAGS: + for (i = 1; i <= AF_MAX; i++) + if ((rnh = rt_tables[i]) && (af == 0 || af == i) && + (error = rnh->rnh_walktree(rnh, + sysctl_dumpentry, &w))) + break; + break; + + case NET_RT_IFLIST: + error = sysctl_iflist(af, &w); + } + splx(s); + if (w.w_tmem) + free(w.w_tmem, M_RTABLE); + w.w_needed += w.w_given; + if (where) { + *given = w.w_where - where; + if (*given < w.w_needed) + return (ENOMEM); + } else { + *given = (11 * w.w_needed) / 10; + } + return (error); +} + +/* + * Definitions of protocols supported in the ROUTE domain. + */ + +extern struct domain routedomain; /* or at least forward */ + +struct protosw routesw[] = { +{ SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, + raw_input, route_output, raw_ctlinput, 0, + route_usrreq, + raw_init, 0, 0, 0, + sysctl_rtable, +} +}; + +struct domain routedomain = + { PF_ROUTE, "route", route_init, 0, 0, + routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; diff --git a/sys/net/slcompress.c b/sys/net/slcompress.c new file mode 100644 index 00000000000..1e2e2e37dc6 --- /dev/null +++ b/sys/net/slcompress.c @@ -0,0 +1,583 @@ +/* $NetBSD: slcompress.c,v 1.12 1995/07/04 06:28:28 paulus Exp $ */ + +/* + * Copyright (c) 1989, 1993, 1994 + * 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)slcompress.c 8.2 (Berkeley) 4/16/94 + */ + +/* + * Routines to compress and uncompess tcp packets (for transmission + * over low speed serial lines. + * + * Van Jacobson (van@helios.ee.lbl.gov), Dec 31, 1989: + * - Initial distribution. + */ + +#include <sys/param.h> +#include <sys/mbuf.h> + +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/ip.h> +#include <netinet/tcp.h> + +#include <net/slcompress.h> + +#ifndef SL_NO_STATS +#define INCR(counter) ++comp->counter; +#else +#define INCR(counter) +#endif + +#define BCMP(p1, p2, n) bcmp((char *)(p1), (char *)(p2), (int)(n)) +#define BCOPY(p1, p2, n) bcopy((char *)(p1), (char *)(p2), (int)(n)) +#ifndef _KERNEL +#define ovbcopy bcopy +#endif + +void +sl_compress_init(comp, max_state) + struct slcompress *comp; + int max_state; +{ + register u_int i; + register struct cstate *tstate = comp->tstate; + + if (max_state == -1) + max_state = MAX_STATES - 1; + bzero((char *)comp, sizeof(*comp)); + for (i = max_state; i > 0; --i) { + tstate[i].cs_id = i; + tstate[i].cs_next = &tstate[i - 1]; + } + tstate[0].cs_next = &tstate[max_state]; + tstate[0].cs_id = 0; + comp->last_cs = &tstate[0]; + comp->last_recv = 255; + comp->last_xmit = 255; + comp->flags = SLF_TOSS; +} + + +/* ENCODE encodes a number that is known to be non-zero. ENCODEZ + * checks for zero (since zero has to be encoded in the long, 3 byte + * form). + */ +#define ENCODE(n) { \ + if ((u_int16_t)(n) >= 256) { \ + *cp++ = 0; \ + cp[1] = (n); \ + cp[0] = (n) >> 8; \ + cp += 2; \ + } else { \ + *cp++ = (n); \ + } \ +} +#define ENCODEZ(n) { \ + if ((u_int16_t)(n) >= 256 || (u_int16_t)(n) == 0) { \ + *cp++ = 0; \ + cp[1] = (n); \ + cp[0] = (n) >> 8; \ + cp += 2; \ + } else { \ + *cp++ = (n); \ + } \ +} + +#define DECODEL(f) { \ + if (*cp == 0) {\ + (f) = htonl(ntohl(f) + ((cp[1] << 8) | cp[2])); \ + cp += 3; \ + } else { \ + (f) = htonl(ntohl(f) + (u_int32_t)*cp++); \ + } \ +} + +#define DECODES(f) { \ + if (*cp == 0) {\ + (f) = htons(ntohs(f) + ((cp[1] << 8) | cp[2])); \ + cp += 3; \ + } else { \ + (f) = htons(ntohs(f) + (u_int32_t)*cp++); \ + } \ +} + +#define DECODEU(f) { \ + if (*cp == 0) {\ + (f) = htons((cp[1] << 8) | cp[2]); \ + cp += 3; \ + } else { \ + (f) = htons((u_int32_t)*cp++); \ + } \ +} + +u_int +sl_compress_tcp(m, ip, comp, compress_cid) + struct mbuf *m; + register struct ip *ip; + struct slcompress *comp; + int compress_cid; +{ + register struct cstate *cs = comp->last_cs->cs_next; + register u_int hlen = ip->ip_hl; + register struct tcphdr *oth; + register struct tcphdr *th; + register u_int deltaS, deltaA; + register u_int changes = 0; + u_char new_seq[16]; + register u_char *cp = new_seq; + + /* + * Bail if this is an IP fragment or if the TCP packet isn't + * `compressible' (i.e., ACK isn't set or some other control bit is + * set). (We assume that the caller has already made sure the + * packet is IP proto TCP). + */ + if ((ip->ip_off & htons(0x3fff)) || m->m_len < 40) + return (TYPE_IP); + + th = (struct tcphdr *)&((int32_t *)ip)[hlen]; + if ((th->th_flags & (TH_SYN|TH_FIN|TH_RST|TH_ACK)) != TH_ACK) + return (TYPE_IP); + /* + * Packet is compressible -- we're going to send either a + * COMPRESSED_TCP or UNCOMPRESSED_TCP packet. Either way we need + * to locate (or create) the connection state. Special case the + * most recently used connection since it's most likely to be used + * again & we don't have to do any reordering if it's used. + */ + INCR(sls_packets) + if (ip->ip_src.s_addr != cs->cs_ip.ip_src.s_addr || + ip->ip_dst.s_addr != cs->cs_ip.ip_dst.s_addr || + *(int32_t *)th != ((int32_t *)&cs->cs_ip)[cs->cs_ip.ip_hl]) { + /* + * Wasn't the first -- search for it. + * + * States are kept in a circularly linked list with + * last_cs pointing to the end of the list. The + * list is kept in lru order by moving a state to the + * head of the list whenever it is referenced. Since + * the list is short and, empirically, the connection + * we want is almost always near the front, we locate + * states via linear search. If we don't find a state + * for the datagram, the oldest state is (re-)used. + */ + register struct cstate *lcs; + register struct cstate *lastcs = comp->last_cs; + + do { + lcs = cs; cs = cs->cs_next; + INCR(sls_searches) + if (ip->ip_src.s_addr == cs->cs_ip.ip_src.s_addr + && ip->ip_dst.s_addr == cs->cs_ip.ip_dst.s_addr + && *(int32_t *)th == + ((int32_t *)&cs->cs_ip)[cs->cs_ip.ip_hl]) + goto found; + } while (cs != lastcs); + + /* + * Didn't find it -- re-use oldest cstate. Send an + * uncompressed packet that tells the other side what + * connection number we're using for this conversation. + * Note that since the state list is circular, the oldest + * state points to the newest and we only need to set + * last_cs to update the lru linkage. + */ + INCR(sls_misses) + comp->last_cs = lcs; + hlen += th->th_off; + hlen <<= 2; + goto uncompressed; + + found: + /* + * Found it -- move to the front on the connection list. + */ + if (cs == lastcs) + comp->last_cs = lcs; + else { + lcs->cs_next = cs->cs_next; + cs->cs_next = lastcs->cs_next; + lastcs->cs_next = cs; + } + } + + /* + * Make sure that only what we expect to change changed. The first + * line of the `if' checks the IP protocol version, header length & + * type of service. The 2nd line checks the "Don't fragment" bit. + * The 3rd line checks the time-to-live and protocol (the protocol + * check is unnecessary but costless). The 4th line checks the TCP + * header length. The 5th line checks IP options, if any. The 6th + * line checks TCP options, if any. If any of these things are + * different between the previous & current datagram, we send the + * current datagram `uncompressed'. + */ + oth = (struct tcphdr *)&((int32_t *)&cs->cs_ip)[hlen]; + deltaS = hlen; + hlen += th->th_off; + hlen <<= 2; + + if (((u_int16_t *)ip)[0] != ((u_int16_t *)&cs->cs_ip)[0] || + ((u_int16_t *)ip)[3] != ((u_int16_t *)&cs->cs_ip)[3] || + ((u_int16_t *)ip)[4] != ((u_int16_t *)&cs->cs_ip)[4] || + th->th_off != oth->th_off || + (deltaS > 5 && + BCMP(ip + 1, &cs->cs_ip + 1, (deltaS - 5) << 2)) || + (th->th_off > 5 && + BCMP(th + 1, oth + 1, (th->th_off - 5) << 2))) + goto uncompressed; + + /* + * Figure out which of the changing fields changed. The + * receiver expects changes in the order: urgent, window, + * ack, seq (the order minimizes the number of temporaries + * needed in this section of code). + */ + if (th->th_flags & TH_URG) { + deltaS = ntohs(th->th_urp); + ENCODEZ(deltaS); + changes |= NEW_U; + } else if (th->th_urp != oth->th_urp) + /* argh! URG not set but urp changed -- a sensible + * implementation should never do this but RFC793 + * doesn't prohibit the change so we have to deal + * with it. */ + goto uncompressed; + + if (deltaS = (u_int16_t)(ntohs(th->th_win) - ntohs(oth->th_win))) { + ENCODE(deltaS); + changes |= NEW_W; + } + + if (deltaA = ntohl(th->th_ack) - ntohl(oth->th_ack)) { + if (deltaA > 0xffff) + goto uncompressed; + ENCODE(deltaA); + changes |= NEW_A; + } + + if (deltaS = ntohl(th->th_seq) - ntohl(oth->th_seq)) { + if (deltaS > 0xffff) + goto uncompressed; + ENCODE(deltaS); + changes |= NEW_S; + } + + switch(changes) { + + case 0: + /* + * Nothing changed. If this packet contains data and the + * last one didn't, this is probably a data packet following + * an ack (normal on an interactive connection) and we send + * it compressed. Otherwise it's probably a retransmit, + * retransmitted ack or window probe. Send it uncompressed + * in case the other side missed the compressed version. + */ + if (ip->ip_len != cs->cs_ip.ip_len && + ntohs(cs->cs_ip.ip_len) == hlen) + break; + + /* (fall through) */ + + case SPECIAL_I: + case SPECIAL_D: + /* + * actual changes match one of our special case encodings -- + * send packet uncompressed. + */ + goto uncompressed; + + case NEW_S|NEW_A: + if (deltaS == deltaA && + deltaS == ntohs(cs->cs_ip.ip_len) - hlen) { + /* special case for echoed terminal traffic */ + changes = SPECIAL_I; + cp = new_seq; + } + break; + + case NEW_S: + if (deltaS == ntohs(cs->cs_ip.ip_len) - hlen) { + /* special case for data xfer */ + changes = SPECIAL_D; + cp = new_seq; + } + break; + } + + deltaS = ntohs(ip->ip_id) - ntohs(cs->cs_ip.ip_id); + if (deltaS != 1) { + ENCODEZ(deltaS); + changes |= NEW_I; + } + if (th->th_flags & TH_PUSH) + changes |= TCP_PUSH_BIT; + /* + * Grab the cksum before we overwrite it below. Then update our + * state with this packet's header. + */ + deltaA = ntohs(th->th_sum); + BCOPY(ip, &cs->cs_ip, hlen); + + /* + * We want to use the original packet as our compressed packet. + * (cp - new_seq) is the number of bytes we need for compressed + * sequence numbers. In addition we need one byte for the change + * mask, one for the connection id and two for the tcp checksum. + * So, (cp - new_seq) + 4 bytes of header are needed. hlen is how + * many bytes of the original packet to toss so subtract the two to + * get the new packet size. + */ + deltaS = cp - new_seq; + cp = (u_char *)ip; + if (compress_cid == 0 || comp->last_xmit != cs->cs_id) { + comp->last_xmit = cs->cs_id; + hlen -= deltaS + 4; + cp += hlen; + *cp++ = changes | NEW_C; + *cp++ = cs->cs_id; + } else { + hlen -= deltaS + 3; + cp += hlen; + *cp++ = changes; + } + m->m_len -= hlen; + m->m_data += hlen; + *cp++ = deltaA >> 8; + *cp++ = deltaA; + BCOPY(new_seq, cp, deltaS); + INCR(sls_compressed) + return (TYPE_COMPRESSED_TCP); + + /* + * Update connection state cs & send uncompressed packet ('uncompressed' + * means a regular ip/tcp packet but with the 'conversation id' we hope + * to use on future compressed packets in the protocol field). + */ +uncompressed: + BCOPY(ip, &cs->cs_ip, hlen); + ip->ip_p = cs->cs_id; + comp->last_xmit = cs->cs_id; + return (TYPE_UNCOMPRESSED_TCP); +} + + +int +sl_uncompress_tcp(bufp, len, type, comp) + u_char **bufp; + int len; + u_int type; + struct slcompress *comp; +{ + u_char *hdr, *cp; + int hlen, vjlen; + + cp = bufp? *bufp: NULL; + vjlen = sl_uncompress_tcp_core(cp, len, len, type, comp, &hdr, &hlen); + if (vjlen < 0) + return (0); /* error */ + if (vjlen == 0) + return (len); /* was uncompressed already */ + + cp += vjlen; + len -= vjlen; + + /* + * At this point, cp points to the first byte of data in the + * packet. If we're not aligned on a 4-byte boundary, copy the + * data down so the ip & tcp headers will be aligned. Then back up + * cp by the tcp/ip header length to make room for the reconstructed + * header (we assume the packet we were handed has enough space to + * prepend 128 bytes of header). + */ + if ((int)cp & 3) { + if (len > 0) + (void) ovbcopy(cp, (caddr_t)((int)cp &~ 3), len); + cp = (u_char *)((int)cp &~ 3); + } + cp -= hlen; + len += hlen; + BCOPY(hdr, cp, hlen); + + *bufp = cp; + return (len); +} + +/* + * Uncompress a packet of total length total_len. The first buflen + * bytes are at buf; this must include the entire (compressed or + * uncompressed) TCP/IP header. This procedure returns the length + * of the VJ header, with a pointer to the uncompressed IP header + * in *hdrp and its length in *hlenp. + */ +int +sl_uncompress_tcp_core(buf, buflen, total_len, type, comp, hdrp, hlenp) + u_char *buf; + int buflen, total_len; + u_int type; + struct slcompress *comp; + u_char **hdrp; + u_int *hlenp; +{ + register u_char *cp; + register u_int hlen, changes; + register struct tcphdr *th; + register struct cstate *cs; + register struct ip *ip; + register u_int16_t *bp; + register u_int vjlen; + + switch (type) { + + case TYPE_UNCOMPRESSED_TCP: + ip = (struct ip *) buf; + if (ip->ip_p >= MAX_STATES) + goto bad; + cs = &comp->rstate[comp->last_recv = ip->ip_p]; + comp->flags &=~ SLF_TOSS; + ip->ip_p = IPPROTO_TCP; + hlen = ip->ip_hl; + hlen += ((struct tcphdr *)&((int32_t *)ip)[hlen])->th_off; + hlen <<= 2; + BCOPY(ip, &cs->cs_ip, hlen); + cs->cs_hlen = hlen; + INCR(sls_uncompressedin) + *hdrp = (u_char *) &cs->cs_ip; + *hlenp = hlen; + return (0); + + default: + goto bad; + + case TYPE_COMPRESSED_TCP: + break; + } + /* We've got a compressed packet. */ + INCR(sls_compressedin) + cp = buf; + changes = *cp++; + if (changes & NEW_C) { + /* Make sure the state index is in range, then grab the state. + * If we have a good state index, clear the 'discard' flag. */ + if (*cp >= MAX_STATES) + goto bad; + + comp->flags &=~ SLF_TOSS; + comp->last_recv = *cp++; + } else { + /* this packet has an implicit state index. If we've + * had a line error since the last time we got an + * explicit state index, we have to toss the packet. */ + if (comp->flags & SLF_TOSS) { + INCR(sls_tossed) + return (-1); + } + } + cs = &comp->rstate[comp->last_recv]; + hlen = cs->cs_ip.ip_hl << 2; + th = (struct tcphdr *)&((u_char *)&cs->cs_ip)[hlen]; + th->th_sum = htons((*cp << 8) | cp[1]); + cp += 2; + if (changes & TCP_PUSH_BIT) + th->th_flags |= TH_PUSH; + else + th->th_flags &=~ TH_PUSH; + + switch (changes & SPECIALS_MASK) { + case SPECIAL_I: + { + register u_int i = ntohs(cs->cs_ip.ip_len) - cs->cs_hlen; + th->th_ack = htonl(ntohl(th->th_ack) + i); + th->th_seq = htonl(ntohl(th->th_seq) + i); + } + break; + + case SPECIAL_D: + th->th_seq = htonl(ntohl(th->th_seq) + ntohs(cs->cs_ip.ip_len) + - cs->cs_hlen); + break; + + default: + if (changes & NEW_U) { + th->th_flags |= TH_URG; + DECODEU(th->th_urp) + } else + th->th_flags &=~ TH_URG; + if (changes & NEW_W) + DECODES(th->th_win) + if (changes & NEW_A) + DECODEL(th->th_ack) + if (changes & NEW_S) + DECODEL(th->th_seq) + break; + } + if (changes & NEW_I) { + DECODES(cs->cs_ip.ip_id) + } else + cs->cs_ip.ip_id = htons(ntohs(cs->cs_ip.ip_id) + 1); + + /* + * At this point, cp points to the first byte of data in the + * packet. Fill in the IP total length and update the IP + * header checksum. + */ + vjlen = cp - buf; + buflen -= vjlen; + if (buflen < 0) + /* we must have dropped some characters (crc should detect + * this but the old slip framing won't) */ + goto bad; + + total_len += cs->cs_hlen - vjlen; + cs->cs_ip.ip_len = htons(total_len); + + /* recompute the ip header checksum */ + bp = (u_int16_t *) &cs->cs_ip; + cs->cs_ip.ip_sum = 0; + for (changes = 0; hlen > 0; hlen -= 2) + changes += *bp++; + changes = (changes & 0xffff) + (changes >> 16); + changes = (changes & 0xffff) + (changes >> 16); + cs->cs_ip.ip_sum = ~ changes; + + *hdrp = (u_char *) &cs->cs_ip; + *hlenp = cs->cs_hlen; + return vjlen; + +bad: + comp->flags |= SLF_TOSS; + INCR(sls_errorin) + return (-1); +} diff --git a/sys/net/slcompress.h b/sys/net/slcompress.h new file mode 100644 index 00000000000..34f3ee7d521 --- /dev/null +++ b/sys/net/slcompress.h @@ -0,0 +1,162 @@ +/* $NetBSD: slcompress.h,v 1.9 1995/07/04 06:28:29 paulus Exp $ */ + +/* + * Copyright (c) 1989, 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)slcompress.h 8.1 (Berkeley) 6/10/93 + */ + +/* + * Definitions for tcp compression routines. + * + * Van Jacobson (van@helios.ee.lbl.gov), Dec 31, 1989: + * - Initial distribution. + */ + +#define MAX_STATES 16 /* must be > 2 and < 256 */ +#define MAX_HDR MLEN /* XXX 4bsd-ism: should really be 128 */ + +/* + * Compressed packet format: + * + * The first octet contains the packet type (top 3 bits), TCP + * 'push' bit, and flags that indicate which of the 4 TCP sequence + * numbers have changed (bottom 5 bits). The next octet is a + * conversation number that associates a saved IP/TCP header with + * the compressed packet. The next two octets are the TCP checksum + * from the original datagram. The next 0 to 15 octets are + * sequence number changes, one change per bit set in the header + * (there may be no changes and there are two special cases where + * the receiver implicitly knows what changed -- see below). + * + * There are 5 numbers which can change (they are always inserted + * in the following order): TCP urgent pointer, window, + * acknowlegement, sequence number and IP ID. (The urgent pointer + * is different from the others in that its value is sent, not the + * change in value.) Since typical use of SLIP links is biased + * toward small packets (see comments on MTU/MSS below), changes + * use a variable length coding with one octet for numbers in the + * range 1 - 255 and 3 octets (0, MSB, LSB) for numbers in the + * range 256 - 65535 or 0. (If the change in sequence number or + * ack is more than 65535, an uncompressed packet is sent.) + */ + +/* + * Packet types (must not conflict with IP protocol version) + * + * The top nibble of the first octet is the packet type. There are + * three possible types: IP (not proto TCP or tcp with one of the + * control flags set); uncompressed TCP (a normal IP/TCP packet but + * with the 8-bit protocol field replaced by an 8-bit connection id -- + * this type of packet syncs the sender & receiver); and compressed + * TCP (described above). + * + * LSB of 4-bit field is TCP "PUSH" bit (a worthless anachronism) and + * is logically part of the 4-bit "changes" field that follows. Top + * three bits are actual packet type. For backward compatibility + * and in the interest of conserving bits, numbers are chosen so the + * IP protocol version number (4) which normally appears in this nibble + * means "IP packet". + */ + +/* packet types */ +#define TYPE_IP 0x40 +#define TYPE_UNCOMPRESSED_TCP 0x70 +#define TYPE_COMPRESSED_TCP 0x80 +#define TYPE_ERROR 0x00 + +/* Bits in first octet of compressed packet */ +#define NEW_C 0x40 /* flag bits for what changed in a packet */ +#define NEW_I 0x20 +#define NEW_S 0x08 +#define NEW_A 0x04 +#define NEW_W 0x02 +#define NEW_U 0x01 + +/* reserved, special-case values of above */ +#define SPECIAL_I (NEW_S|NEW_W|NEW_U) /* echoed interactive traffic */ +#define SPECIAL_D (NEW_S|NEW_A|NEW_W|NEW_U) /* unidirectional data */ +#define SPECIALS_MASK (NEW_S|NEW_A|NEW_W|NEW_U) + +#define TCP_PUSH_BIT 0x10 + + +/* + * "state" data for each active tcp conversation on the wire. This is + * basically a copy of the entire IP/TCP header from the last packet + * we saw from the conversation together with a small identifier + * the transmit & receive ends of the line use to locate saved header. + */ +struct cstate { + struct cstate *cs_next; /* next most recently used cstate (xmit only) */ + u_int16_t cs_hlen; /* size of hdr (receive only) */ + u_char cs_id; /* connection # associated with this state */ + u_char cs_filler; + union { + char csu_hdr[MAX_HDR]; + struct ip csu_ip; /* ip/tcp hdr from most recent packet */ + } slcs_u; +}; +#define cs_ip slcs_u.csu_ip +#define cs_hdr slcs_u.csu_hdr + +/* + * all the state data for one serial line (we need one of these + * per line). + */ +struct slcompress { + struct cstate *last_cs; /* most recently used tstate */ + u_char last_recv; /* last rcvd conn. id */ + u_char last_xmit; /* last sent conn. id */ + u_int16_t flags; +#ifndef SL_NO_STATS + int sls_packets; /* outbound packets */ + int sls_compressed; /* outbound compressed packets */ + int sls_searches; /* searches for connection state */ + int sls_misses; /* times couldn't find conn. state */ + int sls_uncompressedin; /* inbound uncompressed packets */ + int sls_compressedin; /* inbound compressed packets */ + int sls_errorin; /* inbound unknown type packets */ + int sls_tossed; /* inbound packets tossed because of error */ +#endif + struct cstate tstate[MAX_STATES]; /* xmit connection states */ + struct cstate rstate[MAX_STATES]; /* receive connection states */ +}; +/* flag values */ +#define SLF_TOSS 1 /* tossing rcvd frames because of input err */ + +void sl_compress_init __P((struct slcompress *, int)); +u_int sl_compress_tcp __P((struct mbuf *, + struct ip *, struct slcompress *, int)); +int sl_uncompress_tcp __P((u_char **, int, u_int, struct slcompress *)); +int sl_uncompress_tcp_core __P((u_char *, int, int, u_int, + struct slcompress *, u_char **, u_int *)); diff --git a/sys/net/slip.h b/sys/net/slip.h new file mode 100644 index 00000000000..dac2d10dda9 --- /dev/null +++ b/sys/net/slip.h @@ -0,0 +1,53 @@ +/* $NetBSD: slip.h,v 1.6 1994/06/29 06:36:53 cgd Exp $ */ + +/*- + * Copyright (c) 1994 + * 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 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. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * This product includes software developed by the University of + * California, Berkeley and its contributors. + * 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 DISCLAIMED. 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. + * + * @(#)slip.h 8.1 (Berkeley) 2/12/94 + */ + +/* Ioctls operating on SLIP ttys. */ +#define SLIOCGUNIT _IOR('t', 88, int) /* get slip unit number */ + +/* + * Definitions of the pseudo-link-level header attached to slip + * packets grabbed by the packet filter (bpf) traffic monitor. + */ +#define SLIP_HDRLEN 16 /* BPF SLIP header length */ + +/* Offsets into BPF SLIP header. */ +#define SLX_DIR 0 /* direction; see below */ +#define SLX_CHDR 1 /* compressed header data */ +#define CHDR_LEN 15 /* length of compressed header data */ + +#define SLIPDIR_IN 0 /* incoming */ +#define SLIPDIR_OUT 1 /* outgoing */ |