/* **************************************************************************** * Copyright IBM Corporation 1988, 1989 - All Rights Reserved * * Copyright Transarc Corporation 1989 - All Rights Reserved * * * * Permission to use, copy, modify, and distribute this software and its * * documentation for any purpose and without fee is hereby granted, * * provided that the above copyright notice appear in all copies and * * that both that copyright notice and this permission notice appear in * * supporting documentation, and that the name of IBM not be used in * * advertising or publicity pertaining to distribution of the software * * without specific, written prior permission. * * * * IBM DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL * * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL IBM * * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY * * DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER * * IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING * * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. * **************************************************************************** */ #include "../afs/param.h" #include "../rx/rx_kcommon.h" #include "../h/tcp-param.h" /* This must be loaded after proc.h to avoid macro collision with a variable*/ #include "../netinet/udp_var.h" #ifdef RXK_LISTENER_ENV /* osi_NetReceive * OS dependent part of kernel RX listener thread. * * Arguments: * so socket to receive on, typically rx_socket * from pointer to a sockaddr_in. * iov array of iovecs to fill in. * iovcnt how many iovecs there are. * lengthp IN/OUT in: total space available in iovecs. out: size of read. * * Return * 0 if successful * error code (such as EINTR) if not * * Environment * Note that the maximum number of iovecs is 2 + RX_MAXWVECS. This is * so we have a little space to look for packets larger than * rx_maxReceiveSize. */ int rxk_lastSocketError = 0; int rxk_nSocketErrors = 0; int rxk_nSignalsCleared = 0; int osi_NetReceive(osi_socket so, struct sockaddr_in *from, struct iovec *iov, int iovcnt, int *lengthp) { struct uio tuio; int code; struct mbuf *maddr = NULL; struct sockaddr_in *taddr; struct iovec tmpvec[RX_MAXWVECS+2]; #ifdef AFS_SGI65_ENV bhv_desc_t bhv; BHV_PDATA(&bhv) = (void*)so; #endif tuio.uio_iov = tmpvec; tuio.uio_iovcnt = iovcnt; tuio.uio_offset = 0; tuio.uio_segflg = AFS_UIOSYS; tuio.uio_fmode = 0; tuio.uio_resid = *lengthp; tuio.uio_pio = 0; tuio.uio_pbuf = 0; if (iovcnt > RX_MAXWVECS+2) { osi_Panic("Too many (%d) iovecs passed to osi_NetReceive\n", iovcnt); } bcopy((char*)iov, tmpvec, (RX_MAXWVECS+1) * sizeof(struct iovec)); #ifdef AFS_SGI65_ENV code = soreceive(&bhv, &maddr, &tuio, NULL, NULL); #else code = soreceive(so, &maddr, &tuio, NULL, NULL); #endif if (code) { #ifdef AFS_SGI65_ENV /* Clear the error before using the socket again. I've tried being nice * and blocking SIGKILL and SIGSTOP from the kernel, but they get * delivered anyway. So, time to be crude and just clear the signals * pending on this thread. */ if (code == EINTR) { uthread_t *ut = curuthread; int s; s = ut_lock(ut); sigemptyset(&ut->ut_sig); ut->ut_cursig = 0; thread_interrupt_clear(UT_TO_KT(ut), 1); ut_unlock(ut, s); rxk_nSignalsCleared++; } #endif /* Clear the error before using the socket again. */ so->so_error = 0; rxk_lastSocketError = code; rxk_nSocketErrors ++ ; if (maddr) m_freem(maddr); } else { *lengthp = *lengthp - tuio.uio_resid; if (maddr) { bcopy((char*)mtod(maddr, struct sockaddr_in *), (char*)from, sizeof(struct sockaddr_in)); m_freem(maddr); } else { return -1; } } return code; } #else /* RXK_LISTENER_ENV */ static struct protosw parent_proto; /* udp proto switch */ /* * RX input, fast timer and initialization routines. */ #ifdef AFS_SGI64_ENV static void rxk_input(struct mbuf *am, struct ifnet *aif, struct ipsec * spec) #else static void rxk_input(struct mbuf *am, struct ifnet *aif) #endif { void (*tproc)(); register unsigned short *tsp; int hdr; struct udphdr *tu; register struct ip *ti; struct udpiphdr *tvu; register int i; char *phandle; struct sockaddr_in taddr; int tlen; short port; int data_len; /* make sure we have base ip and udp headers in first mbuf */ if (am->m_off > MMAXOFF || am->m_len < 28) { am = m_pullup(am, 28); if (!am) return; } hdr = (mtod(am, struct ip *))->ip_hl; if (hdr > 5) { /* pull up more, the IP hdr is bigger than usual */ if (am->m_len < (8 + (hdr<<2))) { am = m_pullup(am, 8+(hdr<<2)); if (!am) return; } ti = mtod(am, struct ip *); /* recompute, since m_pullup allocates new mbuf */ tu = (struct udphdr *)(((char *)ti) + (hdr<<2)); /* skip ip hdr */ } else { ti = mtod(am, struct ip *); tu = (struct udphdr *)(((char *)ti) + 20); /* skip basic ip hdr */ } /* now read the port out */ port = tu->uh_dport; if (port) { for(tsp=rxk_ports, i=0; i 5) { ip_stripoptions(am, (struct mbuf *) 0); /* get rid of anything we don't need */ tu = (struct udphdr *)(((char *)ti) + 20); } /* * Make mbuf data length reflect UDP length. * If not enough data to reflect UDP length, drop. */ tvu = (struct udpiphdr *)ti; tlen = ntohs((u_short)tvu->ui_ulen); if ((int)ti->ip_len != tlen) { if (tlen > (int)ti->ip_len) { m_free(am); return; } m_adj(am, tlen - (int)ti->ip_len); } /* deliver packet to rx */ taddr.sin_family = AF_INET; /* compute source address */ taddr.sin_port = tu->uh_sport; taddr.sin_addr.s_addr = ti->ip_src.s_addr; /* handle the checksum. Note that this code damages the actual ip header (replacing it with the virtual one, which is the same size), so we must ensure we get everything out we need, first */ if ( tu->uh_sum != 0) { /* if the checksum is there, always check it. It's crazy not * to, unless you can really be sure that your * underlying network (and interfaces and drivers and * DMA hardware, etc!) is error-free. First, fill * in entire virtual ip header. */ tvu->ui_next = 0; tvu->ui_prev = 0; tvu->ui_x1 = 0; tvu->ui_len = tvu->ui_ulen; tlen = ntohs((unsigned short)(tvu->ui_ulen)); if ((!(am->m_flags & M_CKSUMMED)) && in_cksum(am, sizeof(struct ip) + tlen)){ /* checksum, including cksum field, doesn't come out 0, so this packet is bad */ m_freem(am); return; } } /* * 28 is IP (20) + UDP (8) header. ulen includes * udp header, and we *don't* tell RX about udp * header either. So, we remove those 8 as well. */ data_len = ntohs(tu->uh_ulen); data_len -= 8; if (!(*rxk_GetPacketProc)(&phandle, data_len)) { if (rx_mb_to_packet(am, m_freem, 28, data_len, phandle)) { /* XXX should just increment counter here.. */ printf("rx: truncated UDP packet\n"); rxi_FreePacket(phandle); } else (*rxk_PacketArrivalProc)(phandle, &taddr, rxk_portRocks[i], data_len); }else m_freem(am); return; } } } /* if we get here, try to deliver packet to udp */ if (tproc = parent_proto.pr_input) (*tproc)(am, aif); return; } /* * UDP fast timer to raise events for all but Solaris and NCR. * Called about 5 times per second (at unknown priority?). Must go to * splnet or obtain global lock before touching anything significant. */ static void rxk_fasttimo (void) { int (*tproc)(); struct clock temp; /* do rx fasttimo processing here */ rxevent_RaiseEvents(&temp); if (tproc = parent_proto.pr_fasttimo) (*tproc)(); } /* start intercepting basic calls */ void rxk_init(void) { register struct protosw *tpro, *last; if (rxk_initDone) return; last = inetdomain.dom_protoswNPROTOSW; for (tpro = inetdomain.dom_protosw; tpro < last; tpro++) { if (tpro->pr_protocol == IPPROTO_UDP) { bcopy(tpro, &parent_proto, sizeof(parent_proto)); tpro->pr_input = rxk_input; tpro->pr_fasttimo = rxk_fasttimo; rxk_initDone = 1; return; } } osi_Panic("inet:no udp"); } #endif /* RXK_LISTENER_ENV */ /* * RX IP address routines. */ static afs_uint32 myNetAddrs[ADDRSPERSITE]; static int myNetMTUs[ADDRSPERSITE]; static int myNetFlags[ADDRSPERSITE]; static int numMyNetAddrs = 0; /* This version doesn't even begin to handle iterative requests, but then * we don't yet use them anyway. Fix this when rxi_InitPeerParams is changed * to find a true maximum. */ static int rxi_MatchIfnet(struct hashbucket *h, caddr_t key, caddr_t arg1, caddr_t arg2) { afs_uint32 ppaddr = *(afs_uint32*)key; int match_value = *(int*)arg1; struct in_ifaddr *ifa = (struct in_ifaddr*)h; struct sockaddr_in *sin; if ((ppaddr & ifa->ia_netmask) == ifa->ia_net) { if ((ppaddr & ifa->ia_subnetmask) == ifa->ia_subnet) { sin=IA_SIN(ifa); if ( sin->sin_addr.s_addr == ppaddr) { /* ie, ME!!! */ match_value = 4; *(struct in_ifaddr**)arg2 = ifa; } if (match_value < 3) { *(struct in_ifaddr**)arg2 = ifa; match_value = 3; } } else { if (match_value < 2) { *(struct in_ifaddr**)arg2 = ifa; match_value = 2; } } } *(int*)arg1 = match_value; return 0; } struct ifnet * rxi_FindIfnet(addr, pifad) afs_uint32 addr; struct in_ifaddr **pifad; { afs_uint32 ppaddr; int match_value = 0; struct hashbucket *slop; if (numMyNetAddrs == 0) (void) rxi_GetIFInfo(); ppaddr = ntohl(addr); *pifad = (struct in_ifaddr*)&hashinfo_inaddr; slop = hash_enum(&hashinfo_inaddr, rxi_MatchIfnet, HTF_INET, (caddr_t)&ppaddr, (caddr_t)&match_value, (caddr_t)pifad); if (slop) return ((struct in_ifaddr*)slop)->ia_ifp; else return NULL; } static int rxi_EnumGetIfInfo(struct hashbucket *h, caddr_t key, caddr_t arg1, caddr_t arg2) { int different = *(int*)arg1; int i = *(int*)arg2; struct in_ifaddr *iap = (struct in_ifaddr*)h; struct ifnet *ifnp; afs_uint32 ifinaddr; afs_uint32 rxmtu; if (i>=ADDRSPERSITE) return 0; ifnp = iap->ia_ifp; rxmtu = (ifnp->if_mtu - RX_IPUDP_SIZE); ifinaddr = ntohl(iap->ia_addr.sin_addr.s_addr); if (myNetAddrs[i] != ifinaddr) { myNetAddrs[i] = ifinaddr; myNetMTUs[i] = rxmtu; different++; *(int*)arg1 = different; } rxmtu = rxmtu * rxi_nRecvFrags + ((rxi_nRecvFrags - 1) * UDP_HDR_SIZE); if ( ( ifinaddr != 0x7f000001 ) && (rxmtu > rx_maxReceiveSize) ) { rx_maxReceiveSize = MIN( RX_MAX_PACKET_SIZE, rxmtu); rx_maxReceiveSize = MIN( rx_maxReceiveSize, rx_maxReceiveSizeUser); } *(int*)arg2 = i + 1; return 0; } int rxi_GetIFInfo() { int i = 0; int different = 0; /* SGI 6.2 does not have a pointer from the ifnet to the list of * of addresses (if_addrlist). So it's more efficient to run the * in_ifaddr list and use the back pointers to the ifnet struct's. */ (void) hash_enum(&hashinfo_inaddr, rxi_EnumGetIfInfo, HTF_INET, NULL, (caddr_t)&different, (caddr_t)&i); rx_maxJumboRecvSize = RX_HEADER_SIZE + rxi_nDgramPackets * RX_JUMBOBUFFERSIZE + (rxi_nDgramPackets-1) * RX_JUMBOHEADERSIZE; rx_maxJumboRecvSize = MAX(rx_maxJumboRecvSize, rx_maxReceiveSize); return different; } /* osi_NetSend - from the now defunct afs_osinet.c */ #ifdef DEBUG #undef DEBUG #endif #ifdef MP #define _MP_NETLOCKS #endif #ifdef AFS_SGI65_ENV osi_NetSend(asocket, addr, dvec, nvec, asize, istack) register struct osi_socket *asocket; struct iovec *dvec; int nvec; register afs_int32 asize; struct sockaddr_in *addr; int istack; { int code; struct iovec tvecs[RX_MAXWVECS+1]; struct iovec *iovp; struct uio tuio; struct mbuf *to; int i; bhv_desc_t bhv; if (nvec > RX_MAXWVECS+1) { osi_Panic("osi_NetSend: %d: Too many iovecs.\n", nvec); } bcopy((char*)dvec, (char*)tvecs, nvec * sizeof(struct iovec)); tuio.uio_iov = tvecs; tuio.uio_iovcnt = nvec; tuio.uio_segflg = UIO_SYSSPACE; tuio.uio_offset = 0; tuio.uio_sigpipe = 0; tuio.uio_pio = 0; tuio.uio_pbuf = 0; tuio.uio_resid = 0; for (i=0, iovp = tvecs; iiov_len; to = m_get(M_WAIT, MT_SONAME); to->m_len = sizeof(struct sockaddr_in); bcopy((char*)addr, mtod(to, caddr_t), to->m_len); BHV_PDATA(&bhv) = (void*)asocket; code = sosend(&bhv, to, &tuio, 0, NULL); m_freem(to); return code; } #else /* AFS_SGI65_ENV */ int dummy_sblock(struct sockbuf *a, int b, struct socket *c, int *d, int e) { afs_warn("sblock was called before it was installed. Install proper afsd.\n"); } void dummy_sbunlock(struct sockbuf *a, int b, struct socket *c, int d) { afs_warn("sbunlock was called before it was installed. Install proper afsd.\n"); } int (*afs_sblockp)(struct sockbuf*, int, struct socket*, int*, int) = dummy_sblock; void (*afs_sbunlockp)(struct sockbuf*, int, struct socket*, int) = dummy_sbunlock; #define AFS_SBUNLOCK(SB, EV, SO, O) (*afs_sbunlockp)(SB, EV, SO, O) /* osi_NetSend - send asize bytes at adata from asocket to host at addr. * * Now, why do we allocate a new buffer when we could theoretically use the one * pointed to by adata? Because PRU_SEND returns after queueing the message, * not after sending it. If the sender changes the data after queueing it, * we'd see the already-queued data change. One attempt to fix this without * adding a copy would be to have this function wait until the datagram is * sent; however this doesn't work well. In particular, if a host is down, and * an ARP fails to that host, this packet will be queued until the ARP request * comes back, which could be hours later. We can't block in this routine that * long, since it prevents RPC timeouts from happening. */ /* XXX In the brave new world, steal the data bufs out of the rx_packet iovec, * and just queue those. XXX */ int osi_NetSend(asocket, addr, dvec, nvec, asize, istack) register struct socket *asocket; struct iovec *dvec; int nvec; register afs_int32 asize; struct sockaddr_in *addr; int istack; { register struct mbuf *tm, *um; register afs_int32 code; int s; struct mbuf *top = 0; register struct mbuf *m, **mp; int len; char *tdata; caddr_t tpa; int i,tl,rlen; NETSPL_DECL(s1) AFS_STATCNT(osi_NetSend); (*afs_sblockp)(&asocket->so_snd, NETEVENT_SODOWN, asocket, &s1, istack); s = splnet(); mp = ⊤ i = 0; tdata = dvec[i].iov_base; tl = dvec[i].iov_len; while (1) { if ((m = m_vget(M_DONTWAIT, MIN(asize, VCL_MAX), MT_DATA)) == NULL) { if (top) m_freem(top); splx(s); AFS_SBUNLOCK(&asocket->so_snd, NETEVENT_SODOWN, asocket, s1); return 1; } len = MIN(m->m_len, asize); m->m_len = 0; tpa = mtod(m, caddr_t); while (len) { rlen = MIN(len, tl); bcopy(tdata, tpa, rlen); asize -= rlen; len -= rlen; tpa += rlen; m->m_len += rlen; tdata += rlen; tl -= rlen; if (tl <= 0) { i++; if (i > nvec) { /* shouldn't come here! */ asize = 0; /* so we make progress toward completion */ break; } tdata = dvec[i].iov_base; tl = dvec[i].iov_len; } } *mp = m; mp = &m->m_next; if (asize <= 0) break; } tm = top; tm->m_act = (struct mbuf *) 0; /* setup mbuf corresponding to destination address */ um = m_get(M_DONTWAIT, MT_SONAME); if (!um) { if (top) m_freem(top); /* free mbuf chain */ /* if this were vfs40, we'd do sbunlock(asocket, &asocket->so_snd), but we don't do the locking at all for vfs40 systems */ splx(s); AFS_SBUNLOCK(&asocket->so_snd, NETEVENT_SODOWN, asocket, s1); return 1; } bcopy(addr, mtod(um, caddr_t), sizeof(*addr)); um->m_len = sizeof(*addr); /* note that udp_usrreq frees funny mbuf. We hold onto data, but mbuf * around it is gone. we free address ourselves. */ code = (*asocket->so_proto->pr_usrreq)(asocket, PRU_SEND, tm, um, 0); splx(s); m_free(um); AFS_SBUNLOCK(&asocket->so_snd, NETEVENT_SODOWN, asocket, s1); return code; } #endif /* AFS_SGI65_ENV */