/* $OpenBSD: if_pn.c,v 1.6 1999/06/29 02:28:22 jason Exp $ */ /* * Copyright (c) 1997, 1998 * Bill Paul . 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 Bill Paul. * 4. Neither the name of the author nor the names of any co-contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY Bill Paul 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 Bill Paul OR THE VOICES IN HIS HEAD * 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. * * $FreeBSD: if_pn.c,v 1.21 1999/05/28 18:43:10 wpaul Exp $ */ /* * 82c168/82c169 PNIC fast ethernet PCI NIC driver * * Supports various network adapters based on the Lite-On PNIC * PCI network controller chip including the LinkSys LNE100TX. * * Written by Bill Paul * Electrical Engineering Department * Columbia University, New York City */ /* * The PNIC chip is a DEC tulip clone. This driver uses much of the * same code from the driver for the Winbond chip (which is also a * tulip clone) except for the MII, EEPROM and filter programming. * * Technically we could merge support for this chip into the 'de' * driver, but it's such a mess that I'm afraid to go near it. * * The PNIC appears to support both an external MII and an internal * transceiver. I think most 100Mbps implementations use a PHY attached * the the MII. The LinkSys board that I have uses a Myson MTD972 * 100BaseTX PHY. */ #include "bpfilter.h" #include #include #include #include #include #include #include #ifdef __FreeBSD__ #include #include #include #include #include #if NBPFILTER > 0 #include #endif #include /* for vtophys */ #include /* for vtophys */ #include /* for DELAY */ #include #include #include #include #include #endif /* FreeBSD */ #ifdef __OpenBSD__ #include /* for vtophys */ #include /* for vtophys */ #include #include #include #include #include #if NBPFILTER > 0 #include #endif #ifdef INET #include #include #include #include #include #endif #include #include #include #endif #define PN_USEIOSPACE /* #define PN_BACKGROUND_AUTONEG */ #define PN_RX_BUG_WAR #ifdef __FreeBSD__ #include #else #include #endif #if !defined(lint) && defined(__FreeBSD__) static const char rcsid[] = "$FreeBSD: if_pn.c,v 1.21 1999/05/28 18:43:10 wpaul Exp $"; #endif #if defined(__FreeBSD__) /* * Various supported device vendors/types and their names. */ static struct pn_type pn_devs[] = { { PN_VENDORID, PN_DEVICEID_PNIC, "82c168/82c169 PNIC 10/100BaseTX" }, { 0, 0, NULL } }; #endif /* * Various supported PHY vendors/types and their names. Note that * this driver will work with pretty much any MII-compliant PHY, * so failure to positively identify the chip is not a fatal error. */ static struct pn_type pn_phys[] = { { TI_PHY_VENDORID, TI_PHY_10BT, "" }, { TI_PHY_VENDORID, TI_PHY_100VGPMI, "" }, { NS_PHY_VENDORID, NS_PHY_83840A, ""}, { LEVEL1_PHY_VENDORID, LEVEL1_PHY_LXT970, "" }, { INTEL_PHY_VENDORID, INTEL_PHY_82555, "" }, { SEEQ_PHY_VENDORID, SEEQ_PHY_80220, "" }, { 0, 0, "" } }; #if defined(__FreeBSD__) static unsigned long pn_count = 0; static const char *pn_probe __P((pcici_t, pcidi_t)); static void pn_attach __P((pcici_t, int)); static void pn_intr __P((void *)); static void pn_shutdown __P((int, void *)); #elif defined(__OpenBSD__) static int pn_probe __P((struct device *, void *, void *)); static void pn_attach __P((struct device *, struct device *, void *)); static int pn_intr __P((void *)); static void pn_shutdown __P((void *)); #endif static int pn_newbuf __P((struct pn_softc *, struct pn_chain_onefrag *)); static int pn_encap __P((struct pn_softc *, struct pn_chain *, struct mbuf *)); #ifdef PN_RX_BUG_WAR static void pn_rx_bug_war __P((struct pn_softc *, struct pn_chain_onefrag *)); #endif static void pn_rxeof __P((struct pn_softc *)); static void pn_rxeoc __P((struct pn_softc *)); static void pn_txeof __P((struct pn_softc *)); static void pn_txeoc __P((struct pn_softc *)); static void pn_start __P((struct ifnet *)); static int pn_ioctl __P((struct ifnet *, u_long, caddr_t)); static void pn_init __P((void *)); static void pn_stop __P((struct pn_softc *)); static void pn_watchdog __P((struct ifnet *)); static int pn_ifmedia_upd __P((struct ifnet *)); static void pn_ifmedia_sts __P((struct ifnet *, struct ifmediareq *)); static void pn_eeprom_getword __P((struct pn_softc *, u_int8_t, u_int16_t *)); static void pn_read_eeprom __P((struct pn_softc *, caddr_t, int, int, int)); static u_int16_t pn_phy_readreg __P((struct pn_softc *, int)); static void pn_phy_writereg __P((struct pn_softc *, u_int16_t, u_int16_t)); static void pn_autoneg_xmit __P((struct pn_softc *)); static void pn_autoneg_mii __P((struct pn_softc *, int, int)); static void pn_setmode_mii __P((struct pn_softc *, int)); static void pn_getmode_mii __P((struct pn_softc *)); static void pn_autoneg __P((struct pn_softc *, int, int)); static void pn_setmode __P((struct pn_softc *, int)); static void pn_setcfg __P((struct pn_softc *, u_int32_t)); static u_int32_t pn_calchash __P((u_int8_t *)); static void pn_setfilt __P((struct pn_softc *)); static void pn_reset __P((struct pn_softc *)); static int pn_list_rx_init __P((struct pn_softc *)); static int pn_list_tx_init __P((struct pn_softc *)); #define PN_SETBIT(sc, reg, x) \ CSR_WRITE_4(sc, reg, \ CSR_READ_4(sc, reg) | (x)) #define PN_CLRBIT(sc, reg, x) \ CSR_WRITE_4(sc, reg, \ CSR_READ_4(sc, reg) & ~(x)) /* * Read a word of data stored in the EEPROM at address 'addr.' */ static void pn_eeprom_getword(sc, addr, dest) struct pn_softc *sc; u_int8_t addr; u_int16_t *dest; { register int i; u_int32_t r; CSR_WRITE_4(sc, PN_SIOCTL, PN_EE_READ|addr); for (i = 0; i < PN_TIMEOUT; i++) { DELAY(1); r = CSR_READ_4(sc, PN_SIO); if (!(r & PN_SIO_BUSY)) { *dest = (u_int16_t)(r & 0x0000FFFF); return; } } return; } /* * Read a sequence of words from the EEPROM. */ static void pn_read_eeprom(sc, dest, off, cnt, swap) struct pn_softc *sc; caddr_t dest; int off; int cnt; int swap; { int i; u_int16_t word = 0, *ptr; for (i = 0; i < cnt; i++) { pn_eeprom_getword(sc, off + i, &word); ptr = (u_int16_t *)(dest + (i * 2)); if (swap) *ptr = ntohs(word); else *ptr = word; } return; } static u_int16_t pn_phy_readreg(sc, reg) struct pn_softc *sc; int reg; { int i; u_int32_t rval; CSR_WRITE_4(sc, PN_MII, PN_MII_READ | (sc->pn_phy_addr << 23) | (reg << 18)); for (i = 0; i < PN_TIMEOUT; i++) { DELAY(1); rval = CSR_READ_4(sc, PN_MII); if (!(rval & PN_MII_BUSY)) { if ((u_int16_t)(rval & 0x0000FFFF) == 0xFFFF) return(0); else return((u_int16_t)(rval & 0x0000FFFF)); } } return(0); } static void pn_phy_writereg(sc, reg, data) struct pn_softc *sc; u_int16_t reg; u_int16_t data; { int i; CSR_WRITE_4(sc, PN_MII, PN_MII_WRITE | (sc->pn_phy_addr << 23) | (reg << 18) | data); for (i = 0; i < PN_TIMEOUT; i++) { if (!(CSR_READ_4(sc, PN_MII) & PN_MII_BUSY)) break; } return; } #define PN_POLY 0xEDB88320 #define PN_BITS 9 static u_int32_t pn_calchash(addr) u_int8_t *addr; { u_int32_t idx, bit, data, crc; /* Compute CRC for the address value. */ crc = 0xFFFFFFFF; /* initial value */ for (idx = 0; idx < 6; idx++) { for (data = *addr++, bit = 0; bit < 8; bit++, data >>= 1) crc = (crc >> 1) ^ (((crc ^ data) & 1) ? PN_POLY : 0); } return (crc & ((1 << PN_BITS) - 1)); } /* * Initiate an autonegotiation session. */ static void pn_autoneg_xmit(sc) struct pn_softc *sc; { u_int16_t phy_sts; pn_phy_writereg(sc, PHY_BMCR, PHY_BMCR_RESET); DELAY(500); while(pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_RESET); phy_sts = pn_phy_readreg(sc, PHY_BMCR); phy_sts |= PHY_BMCR_AUTONEGENBL|PHY_BMCR_AUTONEGRSTR; pn_phy_writereg(sc, PHY_BMCR, phy_sts); return; } /* * Invoke autonegotiation on a PHY. */ static void pn_autoneg_mii(sc, flag, verbose) struct pn_softc *sc; int flag; int verbose; { u_int16_t phy_sts = 0, media, advert, ability; struct ifnet *ifp; struct ifmedia *ifm; ifm = &sc->ifmedia; ifp = &sc->arpcom.ac_if; ifm->ifm_media = IFM_ETHER | IFM_AUTO; /* * The 100baseT4 PHY on the 3c905-T4 has the 'autoneg supported' * bit cleared in the status register, but has the 'autoneg enabled' * bit set in the control register. This is a contradiction, and * I'm not sure how to handle it. If you want to force an attempt * to autoneg for 100baseT4 PHYs, #define FORCE_AUTONEG_TFOUR * and see what happens. */ #ifndef FORCE_AUTONEG_TFOUR /* * First, see if autoneg is supported. If not, there's * no point in continuing. */ phy_sts = pn_phy_readreg(sc, PHY_BMSR); if (!(phy_sts & PHY_BMSR_CANAUTONEG)) { if (verbose) printf("pn%d: autonegotiation not supported\n", sc->pn_unit); ifm->ifm_media = IFM_ETHER|IFM_10_T|IFM_HDX; return; } #endif switch (flag) { case PN_FLAG_FORCEDELAY: /* * XXX Never use this option anywhere but in the probe * routine: making the kernel stop dead in its tracks * for three whole seconds after we've gone multi-user * is really bad manners. */ pn_autoneg_xmit(sc); DELAY(5000000); break; case PN_FLAG_SCHEDDELAY: /* * Wait for the transmitter to go idle before starting * an autoneg session, otherwise pn_start() may clobber * our timeout, and we don't want to allow transmission * during an autoneg session since that can screw it up. */ if (sc->pn_cdata.pn_tx_head != NULL) { sc->pn_want_auto = 1; return; } pn_autoneg_xmit(sc); ifp->if_timer = 5; sc->pn_autoneg = 1; sc->pn_want_auto = 0; return; break; case PN_FLAG_DELAYTIMEO: ifp->if_timer = 0; sc->pn_autoneg = 0; break; default: printf("pn%d: invalid autoneg flag: %d\n", sc->pn_unit, flag); return; } if (pn_phy_readreg(sc, PHY_BMSR) & PHY_BMSR_AUTONEGCOMP) { if (verbose) printf("pn%d: autoneg complete, ", sc->pn_unit); phy_sts = pn_phy_readreg(sc, PHY_BMSR); } else { if (verbose) printf("pn%d: autoneg not complete, ", sc->pn_unit); } media = pn_phy_readreg(sc, PHY_BMCR); /* Link is good. Report modes and set duplex mode. */ if (pn_phy_readreg(sc, PHY_BMSR) & PHY_BMSR_LINKSTAT) { if (verbose) printf("link status good "); advert = pn_phy_readreg(sc, PHY_ANAR); ability = pn_phy_readreg(sc, PHY_LPAR); if (advert & PHY_ANAR_100BT4 && ability & PHY_ANAR_100BT4) { ifm->ifm_media = IFM_ETHER|IFM_100_T4; media |= PHY_BMCR_SPEEDSEL; media &= ~PHY_BMCR_DUPLEX; printf("(100baseT4)\n"); } else if (advert & PHY_ANAR_100BTXFULL && ability & PHY_ANAR_100BTXFULL) { ifm->ifm_media = IFM_ETHER|IFM_100_TX|IFM_FDX; media |= PHY_BMCR_SPEEDSEL; media |= PHY_BMCR_DUPLEX; printf("(full-duplex, 100Mbps)\n"); } else if (advert & PHY_ANAR_100BTXHALF && ability & PHY_ANAR_100BTXHALF) { ifm->ifm_media = IFM_ETHER|IFM_100_TX|IFM_HDX; media |= PHY_BMCR_SPEEDSEL; media &= ~PHY_BMCR_DUPLEX; printf("(half-duplex, 100Mbps)\n"); } else if (advert & PHY_ANAR_10BTFULL && ability & PHY_ANAR_10BTFULL) { ifm->ifm_media = IFM_ETHER|IFM_10_T|IFM_FDX; media &= ~PHY_BMCR_SPEEDSEL; media |= PHY_BMCR_DUPLEX; printf("(full-duplex, 10Mbps)\n"); } else if (advert & PHY_ANAR_10BTHALF && ability & PHY_ANAR_10BTHALF) { ifm->ifm_media = IFM_ETHER|IFM_10_T|IFM_HDX; media &= ~PHY_BMCR_SPEEDSEL; media &= ~PHY_BMCR_DUPLEX; printf("(half-duplex, 10Mbps)\n"); } media &= ~PHY_BMCR_AUTONEGENBL; /* Set ASIC's duplex mode to match the PHY. */ pn_setcfg(sc, ifm->ifm_media); pn_phy_writereg(sc, PHY_BMCR, media); } else { if (verbose) printf("no carrier\n"); } pn_init(sc); if (sc->pn_tx_pend) { sc->pn_autoneg = 0; sc->pn_tx_pend = 0; pn_start(ifp); } return; } static void pn_getmode_mii(sc) struct pn_softc *sc; { u_int16_t bmsr; struct ifnet *ifp; ifp = &sc->arpcom.ac_if; bmsr = pn_phy_readreg(sc, PHY_BMSR); /* fallback */ sc->ifmedia.ifm_media = IFM_ETHER|IFM_10_T|IFM_HDX; if (bmsr & PHY_BMSR_10BTHALF) { ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_HDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T, 0, NULL); } if (bmsr & PHY_BMSR_10BTFULL) { ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_FDX, 0, NULL); sc->ifmedia.ifm_media = IFM_ETHER|IFM_10_T|IFM_FDX; } if (bmsr & PHY_BMSR_100BTXHALF) { ifp->if_baudrate = 100000000; ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_HDX, 0, NULL); sc->ifmedia.ifm_media = IFM_ETHER|IFM_100_TX|IFM_HDX; } if (bmsr & PHY_BMSR_100BTXFULL) { ifp->if_baudrate = 100000000; ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_FDX, 0, NULL); sc->ifmedia.ifm_media = IFM_ETHER|IFM_100_TX|IFM_FDX; } /* Some also support 100BaseT4. */ if (bmsr & PHY_BMSR_100BT4) { ifp->if_baudrate = 100000000; ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_T4, 0, NULL); sc->ifmedia.ifm_media = IFM_ETHER|IFM_100_T4; #ifdef FORCE_AUTONEG_TFOUR ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_AUTO, 0 NULL): sc->ifmedia.ifm_media = IFM_ETHER|IFM_AUTO; #endif } if (bmsr & PHY_BMSR_CANAUTONEG) { ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_AUTO, 0, NULL); sc->ifmedia.ifm_media = IFM_ETHER|IFM_AUTO; } return; } static void pn_autoneg(sc, flag, verbose) struct pn_softc *sc; int flag; int verbose; { u_int32_t nway = 0, ability; struct ifnet *ifp; struct ifmedia *ifm; ifm = &sc->ifmedia; ifp = &sc->arpcom.ac_if; ifm->ifm_media = IFM_ETHER | IFM_AUTO; switch (flag) { case PN_FLAG_FORCEDELAY: /* * XXX Never use this option anywhere but in the probe * routine: making the kernel stop dead in its tracks * for three whole seconds after we've gone multi-user * is really bad manners. */ CSR_WRITE_4(sc, PN_GEN, PN_GEN_MUSTBEONE|PN_GEN_100TX_LOOP); PN_CLRBIT(sc, PN_NWAY, PN_NWAY_AUTONEGRSTR); PN_SETBIT(sc, PN_NWAY, PN_NWAY_AUTOENB); DELAY(5000000); break; case PN_FLAG_SCHEDDELAY: /* * Wait for the transmitter to go idle before starting * an autoneg session, otherwise pn_start() may clobber * our timeout, and we don't want to allow transmission * during an autoneg session since that can screw it up. */ if (sc->pn_cdata.pn_tx_head != NULL) { sc->pn_want_auto = 1; return; } CSR_WRITE_4(sc, PN_GEN, PN_GEN_MUSTBEONE|PN_GEN_100TX_LOOP); PN_CLRBIT(sc, PN_NWAY, PN_NWAY_AUTONEGRSTR); PN_SETBIT(sc, PN_NWAY, PN_NWAY_AUTOENB); ifp->if_timer = 5; sc->pn_autoneg = 1; sc->pn_want_auto = 0; return; break; case PN_FLAG_DELAYTIMEO: ifp->if_timer = 0; sc->pn_autoneg = 0; break; default: printf("pn%d: invalid autoneg flag: %d\n", sc->pn_unit, flag); return; } if (CSR_READ_4(sc, PN_NWAY) & PN_NWAY_LPAR) { if (verbose) printf("pn%d: autoneg complete, ", sc->pn_unit); } else { if (verbose) printf("pn%d: autoneg not complete, ", sc->pn_unit); } /* Link is good. Report the modes and set the duplex mode. */ if (CSR_READ_4(sc, PN_ISR) & PN_ISR_LINKPASS) { if (verbose) printf("link status good."); ability = CSR_READ_4(sc, PN_NWAY); if (ability & PN_NWAY_LPAR100T4) { ifm->ifm_media = IFM_ETHER|IFM_100_T4; nway = PN_NWAY_MODE_100T4; printf("(100baseT4)\n"); } else if (ability & PN_NWAY_LPAR100FULL) { ifm->ifm_media = IFM_ETHER|IFM_100_TX|IFM_FDX; nway = PN_NWAY_MODE_100FD; printf("(full-duplex, 100Mbps)\n"); } else if (ability & PN_NWAY_LPAR100HALF) { ifm->ifm_media = IFM_ETHER|IFM_100_TX|IFM_HDX; nway = PN_NWAY_MODE_100HD; printf("(half-duplex, 100Mbps)\n"); } else if (ability & PN_NWAY_LPAR10FULL) { ifm->ifm_media = IFM_ETHER|IFM_10_T|IFM_FDX; nway = PN_NWAY_MODE_10FD; printf("(full-duplex, 10Mbps)\n"); } else if (ability & PN_NWAY_LPAR10HALF) { ifm->ifm_media = IFM_ETHER|IFM_10_T|IFM_HDX; nway = PN_NWAY_MODE_10HD; printf("(half-duplex, 10Mbps)\n"); } /* Set ASIC's duplex mode to match the PHY. */ pn_setcfg(sc, ifm->ifm_media); CSR_WRITE_4(sc, PN_NWAY, nway); } else { if (verbose) printf("no carrier\n"); } pn_init(sc); if (sc->pn_tx_pend) { sc->pn_autoneg = 0; sc->pn_tx_pend = 0; pn_start(ifp); } return; } static void pn_setmode(sc, media) struct pn_softc *sc; int media; { struct ifnet *ifp; ifp = &sc->arpcom.ac_if; /* * If an autoneg session is in progress, stop it. */ if (sc->pn_autoneg) { printf("pn%d: canceling autoneg session\n", sc->pn_unit); ifp->if_timer = sc->pn_autoneg = sc->pn_want_auto = 0; PN_CLRBIT(sc, PN_NWAY, PN_NWAY_AUTONEGRSTR); } printf("pn%d: selecting NWAY, ", sc->pn_unit); if (IFM_SUBTYPE(media) == IFM_100_T4) { printf("100Mbps/T4, half-duplex\n"); } if (IFM_SUBTYPE(media) == IFM_100_TX) { printf("100Mbps, "); } if (IFM_SUBTYPE(media) == IFM_10_T) { printf("10Mbps, "); } if ((media & IFM_GMASK) == IFM_FDX) { printf("full duplex\n"); } else { printf("half duplex\n"); } pn_setcfg(sc, media); return; } static void pn_setmode_mii(sc, media) struct pn_softc *sc; int media; { u_int16_t bmcr; struct ifnet *ifp; ifp = &sc->arpcom.ac_if; /* * If an autoneg session is in progress, stop it. */ if (sc->pn_autoneg) { printf("pn%d: canceling autoneg session\n", sc->pn_unit); ifp->if_timer = sc->pn_autoneg = sc->pn_want_auto = 0; bmcr = pn_phy_readreg(sc, PHY_BMCR); bmcr &= ~PHY_BMCR_AUTONEGENBL; pn_phy_writereg(sc, PHY_BMCR, bmcr); } printf("pn%d: selecting MII, ", sc->pn_unit); bmcr = pn_phy_readreg(sc, PHY_BMCR); bmcr &= ~(PHY_BMCR_AUTONEGENBL|PHY_BMCR_SPEEDSEL| PHY_BMCR_DUPLEX|PHY_BMCR_LOOPBK); if (IFM_SUBTYPE(media) == IFM_100_T4) { printf("100Mbps/T4, half-duplex\n"); bmcr |= PHY_BMCR_SPEEDSEL; bmcr &= ~PHY_BMCR_DUPLEX; } if (IFM_SUBTYPE(media) == IFM_100_TX) { printf("100Mbps, "); bmcr |= PHY_BMCR_SPEEDSEL; } if (IFM_SUBTYPE(media) == IFM_10_T) { printf("10Mbps, "); bmcr &= ~PHY_BMCR_SPEEDSEL; } if ((media & IFM_GMASK) == IFM_FDX) { printf("full duplex\n"); bmcr |= PHY_BMCR_DUPLEX; } else { printf("half duplex\n"); bmcr &= ~PHY_BMCR_DUPLEX; } pn_setcfg(sc, media); pn_phy_writereg(sc, PHY_BMCR, bmcr); return; } /* * Programming the receiver filter on the tulip/PNIC is gross. You * have to construct a special setup frame and download it to the * chip via the transmit DMA engine. This routine is also somewhat * gross, as the setup frame is sent synchronously rather than putting * on the transmit queue. The transmitter has to be stopped, then we * can download the frame and wait for the 'owned' bit to clear. * * We always program the chip using 'hash perfect' mode, i.e. one perfect * address (our node address) and a 512-bit hash filter for multicast * frames. We also sneak the broadcast address into the hash filter since * we need that too. */ void pn_setfilt(sc) struct pn_softc *sc; { struct pn_desc *sframe; u_int32_t h, *sp; #ifdef __FreeBSD__ struct ifmultiaddr *ifma; #else struct arpcom *ac = &sc->arpcom; struct ether_multi *enm; struct ether_multistep step; #endif struct ifnet *ifp; int i; ifp = &sc->arpcom.ac_if; PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON); PN_SETBIT(sc, PN_ISR, PN_ISR_TX_IDLE); sframe = &sc->pn_cdata.pn_sframe; sp = (u_int32_t *)&sc->pn_cdata.pn_sbuf; bzero((char *)sp, PN_SFRAME_LEN); sframe->pn_status = PN_TXSTAT_OWN; sframe->pn_next = vtophys(&sc->pn_ldata->pn_tx_list[0]); sframe->pn_data = vtophys(&sc->pn_cdata.pn_sbuf); sframe->pn_ctl = PN_SFRAME_LEN | PN_TXCTL_TLINK | PN_TXCTL_SETUP | PN_FILTER_HASHPERF; /* If we want promiscuous mode, set the allframes bit. */ if (ifp->if_flags & IFF_PROMISC) PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_RX_PROMISC); else PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_RX_PROMISC); if (ifp->if_flags & IFF_ALLMULTI) PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_RX_ALLMULTI); #ifdef __FreeBSD__ for (ifma = ifp->if_multiaddrs.lh_first; ifma != NULL; ifma = ifma->ifma_link.le_next) { if (ifma->ifma_addr->sa_family != AF_LINK) continue; h = pn_calchash(LLADDR((struct sockaddr_dl *)ifma->ifma_addr)); sp[h >> 4] |= 1 << (h & 0xF); } #endif #ifdef __OpenBSD__ ETHER_FIRST_MULTI(step, ac, enm); while (enm != NULL) { h = pn_calchash(enm->enm_addrlo); sp[h >> 4] |= 1 << (h & 0xf); ETHER_NEXT_MULTI(step, enm); } #endif if (ifp->if_flags & IFF_BROADCAST) { h = pn_calchash(etherbroadcastaddr); sp[h >> 4] |= 1 << (h & 0xF); } sp[39] = ((u_int16_t *)sc->arpcom.ac_enaddr)[0]; sp[40] = ((u_int16_t *)sc->arpcom.ac_enaddr)[1]; sp[41] = ((u_int16_t *)sc->arpcom.ac_enaddr)[2]; CSR_WRITE_4(sc, PN_TXADDR, vtophys(sframe)); PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON); CSR_WRITE_4(sc, PN_TXSTART, 0xFFFFFFFF); /* * Wait for chip to clear the 'own' bit. */ for (i = 0; i < PN_TIMEOUT; i++) { DELAY(10); if (sframe->pn_status != PN_TXSTAT_OWN) break; } if (i == PN_TIMEOUT) printf("pn%d: failed to send setup frame\n", sc->pn_unit); PN_SETBIT(sc, PN_ISR, PN_ISR_TX_NOBUF|PN_ISR_TX_IDLE); return; } /* * In order to fiddle with the * 'full-duplex' and '100Mbps' bits in the netconfig register, we * first have to put the transmit and/or receive logic in the idle state. */ static void pn_setcfg(sc, media) struct pn_softc *sc; u_int32_t media; { int i, restart = 0; if (CSR_READ_4(sc, PN_NETCFG) & (PN_NETCFG_TX_ON|PN_NETCFG_RX_ON)) { restart = 1; PN_CLRBIT(sc, PN_NETCFG, (PN_NETCFG_TX_ON|PN_NETCFG_RX_ON)); for (i = 0; i < PN_TIMEOUT; i++) { DELAY(10); if ((CSR_READ_4(sc, PN_ISR) & PN_ISR_TX_IDLE) && (CSR_READ_4(sc, PN_ISR) & PN_ISR_RX_IDLE)) break; } if (i == PN_TIMEOUT) printf("pn%d: failed to force tx and " "rx to idle state\n", sc->pn_unit); } if (IFM_SUBTYPE(media) == IFM_100_TX) { PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_SPEEDSEL); if (sc->pn_pinfo == NULL) { CSR_WRITE_4(sc, PN_GEN, PN_GEN_MUSTBEONE| PN_GEN_SPEEDSEL|PN_GEN_100TX_LOOP); PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_PCS| PN_NETCFG_SCRAMBLER|PN_NETCFG_MIIENB); PN_SETBIT(sc, PN_NWAY, PN_NWAY_SPEEDSEL); } } else { PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_SPEEDSEL); if (sc->pn_pinfo == NULL) { CSR_WRITE_4(sc, PN_GEN, PN_GEN_MUSTBEONE|PN_GEN_100TX_LOOP); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_PCS| PN_NETCFG_SCRAMBLER|PN_NETCFG_MIIENB); PN_CLRBIT(sc, PN_NWAY, PN_NWAY_SPEEDSEL); } } if ((media & IFM_GMASK) == IFM_FDX) { PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_FULLDUPLEX); if (sc->pn_pinfo == NULL) PN_SETBIT(sc, PN_NWAY, PN_NWAY_DUPLEX); } else { PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_FULLDUPLEX); if (sc->pn_pinfo == NULL) PN_CLRBIT(sc, PN_NWAY, PN_NWAY_DUPLEX); } if (restart) PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON|PN_NETCFG_RX_ON); return; } static void pn_reset(sc) struct pn_softc *sc; { register int i; PN_SETBIT(sc, PN_BUSCTL, PN_BUSCTL_RESET); for (i = 0; i < PN_TIMEOUT; i++) { DELAY(10); if (!(CSR_READ_4(sc, PN_BUSCTL) & PN_BUSCTL_RESET)) break; } if (i == PN_TIMEOUT) printf("pn%d: reset never completed!\n", sc->pn_unit); /* Wait a little while for the chip to get its brains in order. */ DELAY(1000); return; } #ifdef __FreeBSD__ /* * Probe for a Lite-On PNIC chip. Check the PCI vendor and device * IDs against our list and return a device name if we find a match. */ static const char * pn_probe(config_id, device_id) pcici_t config_id; pcidi_t device_id; { struct pn_type *t; u_int32_t rev; t = pn_devs; while(t->pn_name != NULL) { if ((device_id & 0xFFFF) == t->pn_vid && ((device_id >> 16) & 0xFFFF) == t->pn_did) { if (t->pn_did == PN_DEVICEID_PNIC) { rev = pci_conf_read(config_id, PN_PCI_REVISION) & 0xFF; switch (rev & PN_REVMASK) { case PN_REVID_82C168: return (t->pn_name); break; case PN_REVID_82C169: t++; return (t->pn_name); default: printf("uknown PNIC rev: %x\n", rev); break; } } return(t->pn_name); } t++; } return(NULL); } /* * Attach the interface. Allocate softc structures, do ifmedia * setup and ethernet/BPF attach. */ static void pn_attach(config_id, unit) pcici_t config_id; int unit; { int s, i; #ifndef PN_USEIOSPACE vm_offset_t pbase, vbase; #endif u_char eaddr[ETHER_ADDR_LEN]; u_int32_t command; struct pn_softc *sc; struct ifnet *ifp; int media = IFM_ETHER|IFM_100_TX|IFM_FDX; unsigned int round; caddr_t roundptr; struct pn_type *p; u_int16_t phy_vid, phy_did, phy_sts; #ifdef PN_RX_BUG_WAR u_int32_t revision = 0; #endif s = splimp(); sc = malloc(sizeof(struct pn_softc), M_DEVBUF, M_NOWAIT); if (sc == NULL) { printf("pn%d: no memory for softc struct!\n", unit); return; } bzero(sc, sizeof(struct pn_softc)); /* * Handle power management nonsense. */ command = pci_conf_read(config_id, PN_PCI_CAPID) & 0x000000FF; if (command == 0x01) { command = pci_conf_read(config_id, PN_PCI_PWRMGMTCTRL); if (command & PN_PSTATE_MASK) { u_int32_t iobase, membase, irq; /* Save important PCI config data. */ iobase = pci_conf_read(config_id, PN_PCI_LOIO); membase = pci_conf_read(config_id, PN_PCI_LOMEM); irq = pci_conf_read(config_id, PN_PCI_INTLINE); /* Reset the power state. */ printf("pn%d: chip is in D%d power mode " "-- setting to D0\n", unit, command & PN_PSTATE_MASK); command &= 0xFFFFFFFC; pci_conf_write(config_id, PN_PCI_PWRMGMTCTRL, command); /* Restore PCI config data. */ pci_conf_write(config_id, PN_PCI_LOIO, iobase); pci_conf_write(config_id, PN_PCI_LOMEM, membase); pci_conf_write(config_id, PN_PCI_INTLINE, irq); } } /* * Map control/status registers. */ command = pci_conf_read(config_id, PCI_COMMAND_STATUS_REG); command |= (PCIM_CMD_PORTEN|PCIM_CMD_MEMEN|PCIM_CMD_BUSMASTEREN); pci_conf_write(config_id, PCI_COMMAND_STATUS_REG, command); command = pci_conf_read(config_id, PCI_COMMAND_STATUS_REG); #ifdef PN_USEIOSPACE if (!(command & PCIM_CMD_PORTEN)) { printf("pn%d: failed to enable I/O ports!\n", unit); free(sc, M_DEVBUF); goto fail; } if (!pci_map_port(config_id, PN_PCI_LOIO, (u_short *)&(sc->pn_bhandle))) { printf ("pn%d: couldn't map ports\n", unit); goto fail; } #ifdef __i386__ sc->pn_btag = I386_BUS_SPACE_IO; #endif #ifdef __alpha__ sc->pn_btag = ALPHA_BUS_SPACE_IO; #endif #else if (!(command & PCIM_CMD_MEMEN)) { printf("pn%d: failed to enable memory mapping!\n", unit); goto fail; } if (!pci_map_mem(config_id, PN_PCI_LOMEM, &vbase, &pbase)) { printf ("pn%d: couldn't map memory\n", unit); goto fail; } sc->pn_bhandle = vbase; #ifdef __i386__ sc->pn_btag = I386_BUS_SPACE_MEM; #endif #ifdef __alpha__ sc->pn_btag = ALPHA_BUS_SPACE_MEM; #endif #endif /* Allocate interrupt */ if (!pci_map_int(config_id, pn_intr, sc, &net_imask)) { printf("pn%d: couldn't map interrupt\n", unit); goto fail; } sc->pn_cachesize = pci_conf_read(config_id, PN_PCI_CACHELEN) & 0xFF; /* Reset the adapter. */ pn_reset(sc); /* * Get station address from the EEPROM. */ pn_read_eeprom(sc, (caddr_t)&eaddr, 0, 3, 1); /* * A PNIC chip was detected. Inform the world. */ printf("pn%d: Ethernet address: %6D\n", unit, eaddr, ":"); sc->pn_unit = unit; bcopy(eaddr, (char *)&sc->arpcom.ac_enaddr, ETHER_ADDR_LEN); sc->pn_ldata_ptr = malloc(sizeof(struct pn_list_data) + 8, M_DEVBUF, M_NOWAIT); if (sc->pn_ldata_ptr == NULL) { free(sc, M_DEVBUF); printf("pn%d: no memory for list buffers!\n", unit); goto fail; } sc->pn_ldata = (struct pn_list_data *)sc->pn_ldata_ptr; round = (unsigned int)sc->pn_ldata_ptr & 0xF; roundptr = sc->pn_ldata_ptr; for (i = 0; i < 8; i++) { if (round % 8) { round++; roundptr++; } else break; } sc->pn_ldata = (struct pn_list_data *)roundptr; bzero(sc->pn_ldata, sizeof(struct pn_list_data)); #ifdef PN_RX_BUG_WAR revision = pci_conf_read(config_id, PN_PCI_REVISION) & 0x000000FF; if (revision == PN_169B_REV || revision == PN_169_REV || (revision & 0xF0) == PN_168_REV) { sc->pn_rx_war = 1; sc->pn_rx_buf = malloc(PN_RXLEN * 5, M_DEVBUF, M_NOWAIT); if (sc->pn_rx_buf == NULL) { printf("pn%d: no memory for workaround buffer\n", unit); goto fail; } } else { sc->pn_rx_war = 0; } #endif ifp = &sc->arpcom.ac_if; ifp->if_softc = sc; ifp->if_unit = unit; ifp->if_name = "pn"; ifp->if_mtu = ETHERMTU; ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; ifp->if_ioctl = pn_ioctl; ifp->if_output = ether_output; ifp->if_start = pn_start; ifp->if_watchdog = pn_watchdog; ifp->if_init = pn_init; ifp->if_baudrate = 10000000; ifp->if_snd.ifq_maxlen = PN_TX_LIST_CNT - 1; ifmedia_init(&sc->ifmedia, 0, pn_ifmedia_upd, pn_ifmedia_sts); if (bootverbose) printf("pn%d: probing for a PHY\n", sc->pn_unit); for (i = PN_PHYADDR_MIN; i < PN_PHYADDR_MAX + 1; i++) { sc->pn_phy_addr = i; pn_phy_writereg(sc, PHY_BMCR, PHY_BMCR_RESET); DELAY(500); while(pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_RESET); if ((phy_sts = pn_phy_readreg(sc, PHY_BMSR))) break; } if (phy_sts) { phy_vid = pn_phy_readreg(sc, PHY_VENID); phy_did = pn_phy_readreg(sc, PHY_DEVID); p = pn_phys; while(p->pn_vid) { if (phy_vid == p->pn_vid && (phy_did | 0x000F) == p->pn_did) { sc->pn_pinfo = p; break; } p++; } if (sc->pn_pinfo == NULL) sc->pn_pinfo = &pn_phys[PHY_UNKNOWN]; pn_getmode_mii(sc); pn_autoneg_mii(sc, PN_FLAG_FORCEDELAY, 1); } else { ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_HDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_FDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_HDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_FDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_T4, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_AUTO, 0, NULL); pn_autoneg(sc, PN_FLAG_FORCEDELAY, 1); } media = sc->ifmedia.ifm_media; pn_stop(sc); ifmedia_set(&sc->ifmedia, media); /* * Call MI attach routines. */ if_attach(ifp); ether_ifattach(ifp); #if NBPFILTER > 0 bpfattach(ifp, DLT_EN10MB, sizeof(struct ether_header)); #endif at_shutdown(pn_shutdown, sc, SHUTDOWN_POST_SYNC); fail: splx(s); return; } #endif /* FreeBSD */ /* * Initialize the transmit descriptors. */ static int pn_list_tx_init(sc) struct pn_softc *sc; { struct pn_chain_data *cd; struct pn_list_data *ld; int i; cd = &sc->pn_cdata; ld = sc->pn_ldata; for (i = 0; i < PN_TX_LIST_CNT; i++) { cd->pn_tx_chain[i].pn_ptr = &ld->pn_tx_list[i]; if (i == (PN_TX_LIST_CNT - 1)) cd->pn_tx_chain[i].pn_nextdesc = &cd->pn_tx_chain[0]; else cd->pn_tx_chain[i].pn_nextdesc = &cd->pn_tx_chain[i + 1]; } cd->pn_tx_free = &cd->pn_tx_chain[0]; cd->pn_tx_tail = cd->pn_tx_head = NULL; return(0); } /* * Initialize the RX descriptors and allocate mbufs for them. Note that * we arrange the descriptors in a closed ring, so that the last descriptor * points back to the first. */ static int pn_list_rx_init(sc) struct pn_softc *sc; { struct pn_chain_data *cd; struct pn_list_data *ld; int i; cd = &sc->pn_cdata; ld = sc->pn_ldata; for (i = 0; i < PN_RX_LIST_CNT; i++) { cd->pn_rx_chain[i].pn_ptr = (struct pn_desc *)&ld->pn_rx_list[i]; if (pn_newbuf(sc, &cd->pn_rx_chain[i]) == ENOBUFS) return(ENOBUFS); if (i == (PN_RX_LIST_CNT - 1)) { cd->pn_rx_chain[i].pn_nextdesc = &cd->pn_rx_chain[0]; ld->pn_rx_list[i].pn_next = vtophys(&ld->pn_rx_list[0]); } else { cd->pn_rx_chain[i].pn_nextdesc = &cd->pn_rx_chain[i + 1]; ld->pn_rx_list[i].pn_next = vtophys(&ld->pn_rx_list[i + 1]); } } cd->pn_rx_head = &cd->pn_rx_chain[0]; return(0); } /* * Initialize an RX descriptor and attach an MBUF cluster. * Note: the length fields are only 11 bits wide, which means the * largest size we can specify is 2047. This is important because * MCLBYTES is 2048, so we have to subtract one otherwise we'll * overflow the field and make a mess. */ static int pn_newbuf(sc, c) struct pn_softc *sc; struct pn_chain_onefrag *c; { struct mbuf *m_new = NULL; MGETHDR(m_new, M_DONTWAIT, MT_DATA); if (m_new == NULL) { printf("pn%d: no memory for rx list -- packet dropped!\n", sc->pn_unit); return(ENOBUFS); } MCLGET(m_new, M_DONTWAIT); if (!(m_new->m_flags & M_EXT)) { printf("pn%d: no memory for rx list -- packet dropped!\n", sc->pn_unit); m_freem(m_new); return(ENOBUFS); } /* * Zero the buffer. This is part of the workaround for the * promiscuous mode bug in the revision 33 PNIC chips. */ bzero((char *)mtod(m_new, char *), MCLBYTES); m_new->m_len = m_new->m_pkthdr.len = MCLBYTES; c->pn_mbuf = m_new; c->pn_ptr->pn_status = PN_RXSTAT; c->pn_ptr->pn_data = vtophys(mtod(m_new, caddr_t)); c->pn_ptr->pn_ctl = PN_RXCTL_RLINK | PN_RXLEN; return(0); } #ifdef PN_RX_BUG_WAR /* * Grrrrr. * The PNIC chip has a terrible bug in it that manifests itself during * periods of heavy activity. The exact mode of failure if difficult to * pinpoint: sometimes it only happens in promiscuous mode, sometimes it * will happen on slow machines. The bug is that sometimes instead of * uploading one complete frame during reception, it uploads what looks * like the entire contents of its FIFO memory. The frame we want is at * the end of the whole mess, but we never know exactly how much data has * been uploaded, so salvaging the frame is hard. * * There is only one way to do it reliably, and it's disgusting. * Here's what we know: * * - We know there will always be somewhere between one and three extra * descriptors uploaded. * * - We know the desired received frame will always be at the end of the * total data upload. * * - We know the size of the desired received frame because it will be * provided in the length field of the status word in the last descriptor. * * Here's what we do: * * - When we allocate buffers for the receive ring, we bzero() them. * This means that we know that the buffer contents should be all * zeros, except for data uploaded by the chip. * * - We also force the PNIC chip to upload frames that include the * ethernet CRC at the end. * * - We gather all of the bogus frame data into a single buffer. * * - We then position a pointer at the end of this buffer and scan * backwards until we encounter the first non-zero byte of data. * This is the end of the received frame. We know we will encounter * some data at the end of the frame because the CRC will always be * there, so even if the sender transmits a packet of all zeros, * we won't be fooled. * * - We know the size of the actual received frame, so we subtract * that value from the current pointer location. This brings us * to the start of the actual received packet. * * - We copy this into an mbuf and pass it on, along with the actual * frame length. * * The performance hit is tremendous, but it beats dropping frames all * the time. */ #define PN_WHOLEFRAME (PN_RXSTAT_FIRSTFRAG|PN_RXSTAT_LASTFRAG) static void pn_rx_bug_war(sc, cur_rx) struct pn_softc *sc; struct pn_chain_onefrag *cur_rx; { struct pn_chain_onefrag *c; unsigned char *ptr; int total_len; u_int32_t rxstat = 0; c = sc->pn_rx_bug_save; ptr = sc->pn_rx_buf; bzero(ptr, sizeof(PN_RXLEN * 5)); /* Copy all the bytes from the bogus buffers. */ while ((c->pn_ptr->pn_status & PN_WHOLEFRAME) != PN_WHOLEFRAME) { rxstat = c->pn_ptr->pn_status; m_copydata(c->pn_mbuf, 0, PN_RXLEN, ptr); ptr += PN_RXLEN - 2; /* round down to 32-bit boundary */ if (c == cur_rx) break; if (rxstat & PN_RXSTAT_LASTFRAG) break; c->pn_ptr->pn_status = PN_RXSTAT; c->pn_ptr->pn_ctl = PN_RXCTL_RLINK | PN_RXLEN; bzero((char *)mtod(c->pn_mbuf, char *), MCLBYTES); c = c->pn_nextdesc; } /* Find the length of the actual receive frame. */ total_len = PN_RXBYTES(rxstat); /* Scan backwards until we hit a non-zero byte. */ while(*ptr == 0x00) ptr--; /* Round off. */ #if defined(__i386__) if ((u_int32_t)(ptr) & 0x3) ptr -= 1; #elif defined(__alpha__) if ((u_int64_t)(ptr) & 0x3) ptr -= 1; #else # error "unsupported architecture" #endif /* Now find the start of the frame. */ ptr -= total_len; if (ptr < sc->pn_rx_buf) ptr = sc->pn_rx_buf; /* * Now copy the salvaged frame to the last mbuf and fake up * the status word to make it look like a successful * frame reception. */ m_copyback(cur_rx->pn_mbuf, 0, total_len, ptr); cur_rx->pn_mbuf->m_len = c->pn_mbuf->m_pkthdr.len = MCLBYTES; cur_rx->pn_ptr->pn_status |= PN_RXSTAT_FIRSTFRAG; return; } #endif /* * A frame has been uploaded: pass the resulting mbuf chain up to * the higher level protocols. */ static void pn_rxeof(sc) struct pn_softc *sc; { struct ether_header *eh; struct mbuf *m; struct ifnet *ifp; struct pn_chain_onefrag *cur_rx; int total_len = 0; u_int32_t rxstat; ifp = &sc->arpcom.ac_if; while(!((rxstat = sc->pn_cdata.pn_rx_head->pn_ptr->pn_status) & PN_RXSTAT_OWN)) { #ifdef __alpha__ struct mbuf *m0 = NULL; #endif cur_rx = sc->pn_cdata.pn_rx_head; sc->pn_cdata.pn_rx_head = cur_rx->pn_nextdesc; #ifdef PN_RX_BUG_WAR /* * XXX The PNIC has a nasty receiver bug that manifests * under certain conditions (sometimes only in promiscuous * mode, sometimes only on slow machines even when not in * promiscuous mode). We have to keep an eye out for the * failure condition and employ a workaround to recover * any mangled frames. */ if (sc->pn_rx_war) { if ((rxstat & PN_WHOLEFRAME) != PN_WHOLEFRAME) { if (rxstat & PN_RXSTAT_FIRSTFRAG) sc->pn_rx_bug_save = cur_rx; if ((rxstat & PN_RXSTAT_LASTFRAG) == 0) continue; pn_rx_bug_war(sc, cur_rx); rxstat = cur_rx->pn_ptr->pn_status; } } #endif /* * If an error occurs, update stats, clear the * status word and leave the mbuf cluster in place: * it should simply get re-used next time this descriptor * comes up in the ring. */ if (rxstat & PN_RXSTAT_RXERR) { ifp->if_ierrors++; if (rxstat & PN_RXSTAT_COLLSEEN) ifp->if_collisions++; cur_rx->pn_ptr->pn_status = PN_RXSTAT; cur_rx->pn_ptr->pn_ctl = PN_RXCTL_RLINK | PN_RXLEN; bzero((char *)mtod(cur_rx->pn_mbuf, char *), MCLBYTES); continue; } /* No errors; receive the packet. */ m = cur_rx->pn_mbuf; total_len = PN_RXBYTES(cur_rx->pn_ptr->pn_status); /* Trim off the CRC. */ total_len -= ETHER_CRC_LEN; /* * Try to conjure up a new mbuf cluster. If that * fails, it means we have an out of memory condition and * should leave the buffer in place and continue. This will * result in a lost packet, but there's little else we * can do in this situation. */ if (pn_newbuf(sc, cur_rx) == ENOBUFS) { ifp->if_ierrors++; cur_rx->pn_ptr->pn_status = PN_RXSTAT; cur_rx->pn_ptr->pn_ctl = PN_RXCTL_RLINK | PN_RXLEN; bzero((char *)mtod(cur_rx->pn_mbuf, char *), MCLBYTES); continue; } #ifdef __alpha__ /* * Grrrr! On the alpha platform, the start of the * packet data must be longword aligned so that ip_input() * doesn't perform any unaligned accesses when it tries * to fiddle with the IP header. But the PNIC is stupid * and wants RX buffers to start on longword boundaries. * So we can't just shift the DMA address over a few * bytes to alter the payload alignment. Instead, we * have to chop out ethernet and IP header parts of * the packet and place then in a separate mbuf with * the alignment fixed up properly. * * As if this chip wasn't broken enough already. */ MGETHDR(m0, M_DONTWAIT, MT_DATA); if (m0 == NULL) { ifp->if_ierrors++; cur_rx->pn_ptr->pn_status = PN_RXSTAT; cur_rx->pn_ptr->pn_ctl = PN_RXCTL_RLINK | PN_RXLEN; bzero((char *)mtod(cur_rx->pn_mbuf, char *), MCLBYTES); continue; } m0->m_data += 2; if (total_len <= (MHLEN - 2)) { bcopy(mtod(m, caddr_t), mtod(m0, caddr_t), total_len); m_freem(m); m = m0; m->m_pkthdr.len = m->m_len = total_len; } else { bcopy(mtod(m, caddr_t), mtod(m0, caddr_t), (MHLEN - 2)); m->m_len = total_len - (MHLEN - 2); m->m_data += (MHLEN - 2); m0->m_next = m; m0->m_len = (MHLEN - 2); m = m0; m->m_pkthdr.len = total_len; } #else m->m_pkthdr.len = m->m_len = total_len; #endif ifp->if_ipackets++; eh = mtod(m, struct ether_header *); m->m_pkthdr.rcvif = ifp; #if NBPFILTER > 0 /* * Handle BPF listeners. Let the BPF user see the packet. */ if (ifp->if_bpf) { #ifdef __FreeBSD__ bpf_mtap(ifp, m); #else bpf_mtap(ifp->if_bpf, m); #endif } #endif /* Remove header from mbuf and pass it on. */ m_adj(m, sizeof(struct ether_header)); ether_input(ifp, eh, m); } return; } void pn_rxeoc(sc) struct pn_softc *sc; { pn_rxeof(sc); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_RX_ON); CSR_WRITE_4(sc, PN_RXADDR, vtophys(sc->pn_cdata.pn_rx_head->pn_ptr)); PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_RX_ON); CSR_WRITE_4(sc, PN_RXSTART, 0xFFFFFFFF); return; } /* * A frame was downloaded to the chip. It's safe for us to clean up * the list buffers. */ static void pn_txeof(sc) struct pn_softc *sc; { struct pn_chain *cur_tx; struct ifnet *ifp; ifp = &sc->arpcom.ac_if; /* Clear the timeout timer. */ ifp->if_timer = 0; if (sc->pn_cdata.pn_tx_head == NULL) return; /* * Go through our tx list and free mbufs for those * frames that have been transmitted. */ while(sc->pn_cdata.pn_tx_head->pn_mbuf != NULL) { u_int32_t txstat; cur_tx = sc->pn_cdata.pn_tx_head; txstat = PN_TXSTATUS(cur_tx); if (txstat & PN_TXSTAT_OWN) break; if (txstat & PN_TXSTAT_ERRSUM) { ifp->if_oerrors++; if (txstat & PN_TXSTAT_EXCESSCOLL) ifp->if_collisions++; if (txstat & PN_TXSTAT_LATECOLL) ifp->if_collisions++; } ifp->if_collisions += (txstat & PN_TXSTAT_COLLCNT) >> 3; ifp->if_opackets++; m_freem(cur_tx->pn_mbuf); cur_tx->pn_mbuf = NULL; if (sc->pn_cdata.pn_tx_head == sc->pn_cdata.pn_tx_tail) { sc->pn_cdata.pn_tx_head = NULL; sc->pn_cdata.pn_tx_tail = NULL; break; } sc->pn_cdata.pn_tx_head = cur_tx->pn_nextdesc; } return; } /* * TX 'end of channel' interrupt handler. */ static void pn_txeoc(sc) struct pn_softc *sc; { struct ifnet *ifp; ifp = &sc->arpcom.ac_if; ifp->if_timer = 0; if (sc->pn_cdata.pn_tx_head == NULL) { ifp->if_flags &= ~IFF_OACTIVE; sc->pn_cdata.pn_tx_tail = NULL; if (sc->pn_want_auto) { if (sc->pn_pinfo == NULL) pn_autoneg(sc, PN_FLAG_SCHEDDELAY, 1); else pn_autoneg_mii(sc, PN_FLAG_SCHEDDELAY, 1); } } return; } #ifdef __OpenBSD__ static int pn_intr(arg) #else static void pn_intr(arg) #endif void *arg; { struct pn_softc *sc; struct ifnet *ifp; u_int32_t status; #ifdef __OpenBSD__ int claimed = 0; #endif sc = arg; ifp = &sc->arpcom.ac_if; /* Supress unwanted interrupts. */ if (!(ifp->if_flags & IFF_UP)) { pn_stop(sc); #ifdef __OpenBSD__ return claimed; #else return; #endif } /* Disable interrupts. */ CSR_WRITE_4(sc, PN_IMR, 0x00000000); for (;;) { status = CSR_READ_4(sc, PN_ISR); if (status) CSR_WRITE_4(sc, PN_ISR, status); if ((status & PN_INTRS) == 0) break; #ifdef __OpenBSD__ claimed = 1; #endif if (status & PN_ISR_RX_OK) pn_rxeof(sc); if ((status & PN_ISR_RX_WATCHDOG) || (status & PN_ISR_RX_IDLE) || (status & PN_ISR_RX_NOBUF)) pn_rxeoc(sc); if (status & PN_ISR_TX_OK) pn_txeof(sc); if (status & PN_ISR_TX_NOBUF) pn_txeoc(sc); if (status & PN_ISR_TX_IDLE) { pn_txeof(sc); if (sc->pn_cdata.pn_tx_head != NULL) { PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON); CSR_WRITE_4(sc, PN_TXSTART, 0xFFFFFFFF); } } if (status & PN_ISR_TX_UNDERRUN) { ifp->if_oerrors++; pn_txeof(sc); if (sc->pn_cdata.pn_tx_head != NULL) { PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON); CSR_WRITE_4(sc, PN_TXSTART, 0xFFFFFFFF); } } if (status & PN_ISR_BUS_ERR) { pn_reset(sc); pn_init(sc); } } /* Re-enable interrupts. */ CSR_WRITE_4(sc, PN_IMR, PN_INTRS); if (ifp->if_snd.ifq_head != NULL) { pn_start(ifp); } #ifdef __OpenBSD__ return claimed; #else return; #endif } /* * Encapsulate an mbuf chain in a descriptor by coupling the mbuf data * pointers to the fragment pointers. */ static int pn_encap(sc, c, m_head) struct pn_softc *sc; struct pn_chain *c; struct mbuf *m_head; { int frag = 0; struct pn_desc *f = NULL; int total_len; struct mbuf *m; /* * Start packing the mbufs in this chain into * the fragment pointers. Stop when we run out * of fragments or hit the end of the mbuf chain. */ m = m_head; total_len = 0; for (m = m_head, frag = 0; m != NULL; m = m->m_next) { if (m->m_len != 0) { if (frag == PN_MAXFRAGS) break; total_len += m->m_len; f = &c->pn_ptr->pn_frag[frag]; f->pn_ctl = PN_TXCTL_TLINK | m->m_len; if (frag == 0) { f->pn_ctl |= PN_TXCTL_FIRSTFRAG; f->pn_status = 0; } else f->pn_status = PN_TXSTAT_OWN; f->pn_data = vtophys(mtod(m, vm_offset_t)); f->pn_next = vtophys(&c->pn_ptr->pn_frag[frag + 1]); frag++; } } /* * Handle special case: we used up all 16 fragments, * but we have more mbufs left in the chain. Copy the * data into an mbuf cluster. Note that we don't * bother clearing the values in the other fragment * pointers/counters; it wouldn't gain us anything, * and would waste cycles. */ if (m != NULL) { struct mbuf *m_new = NULL; MGETHDR(m_new, M_DONTWAIT, MT_DATA); if (m_new == NULL) { printf("pn%d: no memory for tx list", sc->pn_unit); return(1); } if (m_head->m_pkthdr.len > MHLEN) { MCLGET(m_new, M_DONTWAIT); if (!(m_new->m_flags & M_EXT)) { m_freem(m_new); printf("pn%d: no memory for tx list", sc->pn_unit); return(1); } } m_copydata(m_head, 0, m_head->m_pkthdr.len, mtod(m_new, caddr_t)); m_new->m_pkthdr.len = m_new->m_len = m_head->m_pkthdr.len; m_freem(m_head); m_head = m_new; f = &c->pn_ptr->pn_frag[0]; f->pn_data = vtophys(mtod(m_new, caddr_t)); f->pn_ctl = total_len = m_new->m_len; f->pn_ctl |= PN_TXCTL_TLINK|PN_TXCTL_FIRSTFRAG; frag = 1; } c->pn_mbuf = m_head; c->pn_lastdesc = frag - 1; PN_TXCTL(c) |= PN_TXCTL_LASTFRAG|PN_TXCTL_FINT; PN_TXNEXT(c) = vtophys(&c->pn_nextdesc->pn_ptr->pn_frag[0]); return(0); } /* * Main transmit routine. To avoid having to do mbuf copies, we put pointers * to the mbuf data regions directly in the transmit lists. We also save a * copy of the pointers since the transmit list fragment pointers are * physical addresses. */ static void pn_start(ifp) struct ifnet *ifp; { struct pn_softc *sc; struct mbuf *m_head = NULL; struct pn_chain *cur_tx = NULL, *start_tx; sc = ifp->if_softc; if (sc->pn_autoneg) { sc->pn_tx_pend = 1; return; } /* * Check for an available queue slot. If there are none, * punt. */ if (sc->pn_cdata.pn_tx_free->pn_mbuf != NULL) { ifp->if_flags |= IFF_OACTIVE; return; } start_tx = sc->pn_cdata.pn_tx_free; while(sc->pn_cdata.pn_tx_free->pn_mbuf == NULL) { IF_DEQUEUE(&ifp->if_snd, m_head); if (m_head == NULL) break; /* Pick a descriptor off the free list. */ cur_tx = sc->pn_cdata.pn_tx_free; sc->pn_cdata.pn_tx_free = cur_tx->pn_nextdesc; /* Pack the data into the descriptor. */ pn_encap(sc, cur_tx, m_head); if (cur_tx != start_tx) PN_TXOWN(cur_tx) = PN_TXSTAT_OWN; #if NBPFILTER > 0 /* * If there's a BPF listener, bounce a copy of this frame * to him. */ if (ifp->if_bpf) #ifdef __FreeBSD__ bpf_mtap(ifp, cur_tx->pn_mbuf); #else bpf_mtap(ifp->if_bpf, cur_tx->pn_mbuf); #endif #endif PN_TXOWN(cur_tx) = PN_TXSTAT_OWN; CSR_WRITE_4(sc, PN_TXSTART, 0xFFFFFFFF); } /* * If there are no packets queued, bail. */ if (cur_tx == NULL) return; sc->pn_cdata.pn_tx_tail = cur_tx; if (sc->pn_cdata.pn_tx_head == NULL) sc->pn_cdata.pn_tx_head = start_tx; /* * Set a timeout in case the chip goes out to lunch. */ ifp->if_timer = 5; return; } static void pn_init(xsc) void *xsc; { struct pn_softc *sc = xsc; struct ifnet *ifp = &sc->arpcom.ac_if; u_int16_t phy_bmcr = 0; int s; if (sc->pn_autoneg) return; s = splimp(); if (sc->pn_pinfo != NULL) phy_bmcr = pn_phy_readreg(sc, PHY_BMCR); /* * Cancel pending I/O and free all RX/TX buffers. */ pn_stop(sc); pn_reset(sc); /* * Set cache alignment and burst length. */ CSR_WRITE_4(sc, PN_BUSCTL, PN_BUSCTL_MUSTBEONE|PN_BUSCTL_ARBITRATION); PN_SETBIT(sc, PN_BUSCTL, PN_BURSTLEN_16LONG); switch(sc->pn_cachesize) { case 32: PN_SETBIT(sc, PN_BUSCTL, PN_CACHEALIGN_32LONG); break; case 16: PN_SETBIT(sc, PN_BUSCTL, PN_CACHEALIGN_16LONG); break; case 8: PN_SETBIT(sc, PN_BUSCTL, PN_CACHEALIGN_8LONG); break; case 0: default: PN_SETBIT(sc, PN_BUSCTL, PN_CACHEALIGN_NONE); break; } PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_TX_IMMEDIATE); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_NO_RXCRC); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_HEARTBEAT); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_STORENFWD); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_TX_BACKOFF); PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_TX_THRESH); PN_SETBIT(sc, PN_NETCFG, PN_TXTHRESH_72BYTES); if (sc->pn_pinfo == NULL) { PN_CLRBIT(sc, PN_NETCFG, PN_NETCFG_MIIENB); PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_BACKOFF); } else { PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_MIIENB); PN_SETBIT(sc, PN_ENDEC, PN_ENDEC_JABBERDIS); } pn_setcfg(sc, sc->ifmedia.ifm_media); /* Init circular RX list. */ if (pn_list_rx_init(sc) == ENOBUFS) { printf("pn%d: initialization failed: no " "memory for rx buffers\n", sc->pn_unit); pn_stop(sc); (void)splx(s); return; } /* * Init tx descriptors. */ pn_list_tx_init(sc); /* * Load the address of the RX list. */ CSR_WRITE_4(sc, PN_RXADDR, vtophys(sc->pn_cdata.pn_rx_head->pn_ptr)); /* * Load the RX/multicast filter. */ pn_setfilt(sc); /* * Enable interrupts. */ CSR_WRITE_4(sc, PN_IMR, PN_INTRS); CSR_WRITE_4(sc, PN_ISR, 0xFFFFFFFF); /* Enable receiver and transmitter. */ PN_SETBIT(sc, PN_NETCFG, PN_NETCFG_TX_ON|PN_NETCFG_RX_ON); CSR_WRITE_4(sc, PN_RXSTART, 0xFFFFFFFF); /* Restore state of BMCR */ if (sc->pn_pinfo != NULL) pn_phy_writereg(sc, PHY_BMCR, phy_bmcr); ifp->if_flags |= IFF_RUNNING; ifp->if_flags &= ~IFF_OACTIVE; (void)splx(s); return; } /* * Set media options. */ static int pn_ifmedia_upd(ifp) struct ifnet *ifp; { struct pn_softc *sc; struct ifmedia *ifm; sc = ifp->if_softc; ifm = &sc->ifmedia; if (IFM_TYPE(ifm->ifm_media) != IFM_ETHER) return(EINVAL); if (IFM_SUBTYPE(ifm->ifm_media) == IFM_AUTO) { if (sc->pn_pinfo == NULL) pn_autoneg(sc, PN_FLAG_SCHEDDELAY, 1); else pn_autoneg_mii(sc, PN_FLAG_SCHEDDELAY, 1); } else { if (sc->pn_pinfo == NULL) pn_setmode(sc, ifm->ifm_media); else pn_setmode_mii(sc, ifm->ifm_media); } return(0); } /* * Report current media status. */ static void pn_ifmedia_sts(ifp, ifmr) struct ifnet *ifp; struct ifmediareq *ifmr; { struct pn_softc *sc; u_int16_t advert = 0, ability = 0; sc = ifp->if_softc; ifmr->ifm_active = IFM_ETHER; if (sc->pn_pinfo == NULL) { if (CSR_READ_4(sc, PN_NETCFG) & PN_NETCFG_SPEEDSEL) ifmr->ifm_active = IFM_ETHER|IFM_10_T; else ifmr->ifm_active = IFM_ETHER|IFM_100_TX; if (CSR_READ_4(sc, PN_NETCFG) & PN_NETCFG_FULLDUPLEX) ifmr->ifm_active |= IFM_FDX; else ifmr->ifm_active |= IFM_HDX; return; } if (!(pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_AUTONEGENBL)) { if (pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_SPEEDSEL) ifmr->ifm_active = IFM_ETHER|IFM_100_TX; else ifmr->ifm_active = IFM_ETHER|IFM_10_T; if (pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_DUPLEX) ifmr->ifm_active |= IFM_FDX; else ifmr->ifm_active |= IFM_HDX; return; } ability = pn_phy_readreg(sc, PHY_LPAR); advert = pn_phy_readreg(sc, PHY_ANAR); if (advert & PHY_ANAR_100BT4 && ability & PHY_ANAR_100BT4) { ifmr->ifm_active = IFM_ETHER|IFM_100_T4; } else if (advert & PHY_ANAR_100BTXFULL && ability & PHY_ANAR_100BTXFULL) { ifmr->ifm_active = IFM_ETHER|IFM_100_TX|IFM_FDX; } else if (advert & PHY_ANAR_100BTXHALF && ability & PHY_ANAR_100BTXHALF) { ifmr->ifm_active = IFM_ETHER|IFM_100_TX|IFM_HDX; } else if (advert & PHY_ANAR_10BTFULL && ability & PHY_ANAR_10BTFULL) { ifmr->ifm_active = IFM_ETHER|IFM_10_T|IFM_FDX; } else if (advert & PHY_ANAR_10BTHALF && ability & PHY_ANAR_10BTHALF) { ifmr->ifm_active = IFM_ETHER|IFM_10_T|IFM_HDX; } return; } static int pn_ioctl(ifp, command, data) struct ifnet *ifp; u_long command; caddr_t data; { struct pn_softc *sc = ifp->if_softc; struct ifreq *ifr = (struct ifreq *) data; int s, error = 0; #ifdef __OpenBSD__ struct ifaddr *ifa = (struct ifaddr *) data; #endif s = splimp(); #ifdef __OpenBSD__ if ((error = ether_ioctl(ifp, &sc->arpcom, command, data)) > 0) { splx(s); return error; } #endif switch(command) { #ifdef __FreeBSD__ case SIOCSIFADDR: case SIOCGIFADDR: case SIOCSIFMTU: error = ether_ioctl(ifp, command, data); break; #else case SIOCSIFADDR: ifp->if_flags |= IFF_UP; switch (ifa->ifa_addr->sa_family) { #ifdef INET case AF_INET: pn_init(sc); arp_ifinit(&sc->arpcom, ifa); break; #endif /* INET */ default: pn_init(sc); break; } break; #endif case SIOCSIFFLAGS: if (ifp->if_flags & IFF_UP) { pn_init(sc); } else { if (ifp->if_flags & IFF_RUNNING) pn_stop(sc); } error = 0; break; case SIOCADDMULTI: case SIOCDELMULTI: pn_init(sc); error = 0; break; case SIOCGIFMEDIA: case SIOCSIFMEDIA: error = ifmedia_ioctl(ifp, ifr, &sc->ifmedia, command); break; default: error = EINVAL; break; } (void)splx(s); return(error); } static void pn_watchdog(ifp) struct ifnet *ifp; { struct pn_softc *sc; sc = ifp->if_softc; if (sc->pn_autoneg) { if (sc->pn_pinfo == NULL) pn_autoneg(sc, PN_FLAG_DELAYTIMEO, 1); else pn_autoneg_mii(sc, PN_FLAG_DELAYTIMEO, 1); return; } ifp->if_oerrors++; printf("pn%d: watchdog timeout\n", sc->pn_unit); if (!(pn_phy_readreg(sc, PHY_BMSR) & PHY_BMSR_LINKSTAT)) printf("pn%d: no carrier - transceiver cable problem?\n", sc->pn_unit); pn_stop(sc); pn_reset(sc); pn_init(sc); if (ifp->if_snd.ifq_head != NULL) pn_start(ifp); return; } /* * Stop the adapter and free any mbufs allocated to the * RX and TX lists. */ static void pn_stop(sc) struct pn_softc *sc; { register int i; struct ifnet *ifp; ifp = &sc->arpcom.ac_if; ifp->if_timer = 0; PN_CLRBIT(sc, PN_NETCFG, (PN_NETCFG_RX_ON|PN_NETCFG_TX_ON)); CSR_WRITE_4(sc, PN_IMR, 0x00000000); CSR_WRITE_4(sc, PN_TXADDR, 0x00000000); CSR_WRITE_4(sc, PN_RXADDR, 0x00000000); /* * Free data in the RX lists. */ for (i = 0; i < PN_RX_LIST_CNT; i++) { if (sc->pn_cdata.pn_rx_chain[i].pn_mbuf != NULL) { m_freem(sc->pn_cdata.pn_rx_chain[i].pn_mbuf); sc->pn_cdata.pn_rx_chain[i].pn_mbuf = NULL; } } bzero((char *)&sc->pn_ldata->pn_rx_list, sizeof(sc->pn_ldata->pn_rx_list)); /* * Free the TX list buffers. */ for (i = 0; i < PN_TX_LIST_CNT; i++) { if (sc->pn_cdata.pn_tx_chain[i].pn_mbuf != NULL) { m_freem(sc->pn_cdata.pn_tx_chain[i].pn_mbuf); sc->pn_cdata.pn_tx_chain[i].pn_mbuf = NULL; } } bzero((char *)&sc->pn_ldata->pn_tx_list, sizeof(sc->pn_ldata->pn_tx_list)); ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE); return; } #ifdef __FreeBSD__ /* * Stop all chip I/O so that the kernel's probe routines don't * get confused by errant DMAs when rebooting. */ static void pn_shutdown(howto, arg) int howto; void *arg; { struct pn_softc *sc = (struct pn_softc *)arg; pn_stop(sc); return; } static struct pci_device pn_device = { "pn", pn_probe, pn_attach, &pn_count, NULL }; COMPAT_PCI_DRIVER(pn, pn_device); #endif /* __FreeBSD__ */ #ifdef __OpenBSD__ static int pn_probe(parent, match, aux) struct device *parent; void *match; void *aux; { struct pci_attach_args *pa = (struct pci_attach_args *)aux; if (PCI_VENDOR(pa->pa_id) == PCI_VENDOR_LITEON) { switch (PCI_PRODUCT(pa->pa_id)) { case PCI_PRODUCT_LITEON_PNIC: return (1); } } return (0); } static void pn_attach(parent, self, aux) struct device *parent, *self; void *aux; { struct pn_softc *sc = (struct pn_softc *)self; struct pci_attach_args *pa = aux; pci_chipset_tag_t pc = pa->pa_pc; pci_intr_handle_t ih; const char *intrstr = NULL; struct ifnet *ifp = &sc->arpcom.ac_if; bus_addr_t iobase; bus_size_t iosize; u_int32_t command; u_int16_t phy_did, phy_vid, phy_sts; struct pn_type *p = NULL; int s, i, media = IFM_ETHER|IFM_100_TX|IFM_FDX; u_int round; caddr_t roundptr; #ifdef PN_RX_BUG_WAR u_int32_t revision = 0; #endif s = splimp(); sc->pn_unit = sc->sc_dev.dv_unit; command = pci_conf_read(pa->pa_pc, pa->pa_tag, PCI_COMMAND_STATUS_REG); #ifdef PN_USEIOSPACE if (!(command & PCI_COMMAND_IO_ENABLE)) { printf(": failed to enable i/o ports\n"); goto fail; } /* * Map control/status registers. */ if (pci_io_find(pc, pa->pa_tag, PN_PCI_LOIO, &iobase, &iosize)) { printf(": can't find i/o space\n"); goto fail; } if (bus_space_map(pa->pa_iot, iobase, iosize, 0, &sc->pn_bhandle)) { printf(": can't map i/o space\n"); goto fail; } sc->pn_btag = pa->pa_iot; #else if (!(command & PCI_COMMAND_MEM_ENABLE)) { printf(": failed to enable memory mapping\n"); goto fail; } if (pci_mem_find(pc, pa->pa_tag, PN_PCI_LOMEM, &iobase, &iosize, NULL)){ printf(": can't find mem space\n"); goto fail; } if (bus_space_map(pa->pa_memt, iobase, iosize, 0, &sc->pn_bhandle)) { printf(": can't map mem space\n"); goto fail; } sc->pn_btag = pa->pa_memt; #endif /* * Allocate our interrupt. */ if (pci_intr_map(pc, pa->pa_intrtag, pa->pa_intrpin, pa->pa_intrline, &ih)) { printf(": couldn't map interrupt\n"); goto fail; } intrstr = pci_intr_string(pc, ih); sc->sc_ih = pci_intr_establish(pc, ih, IPL_NET, pn_intr, sc, self->dv_xname); if (sc->sc_ih == NULL) { printf(": couldn't establish interrupt"); if (intrstr != NULL) printf(" at %s", intrstr); printf("\n"); goto fail; } printf(": %s", intrstr); sc->pn_cachesize = pci_conf_read(pa->pa_pc, pa->pa_tag, PN_PCI_CACHELEN) & 0xFF; pn_reset(sc); pn_read_eeprom(sc, (caddr_t)&sc->arpcom.ac_enaddr, 0, 3, 1); printf(" address %s\n", ether_sprintf(sc->arpcom.ac_enaddr)); sc->pn_ldata_ptr = malloc(sizeof(struct pn_list_data) + 8, M_DEVBUF, M_NOWAIT); if (sc->pn_ldata_ptr == NULL) { printf("%s: no memory for list buffers\n", sc->sc_dev.dv_xname); goto fail; } sc->pn_ldata = (struct pn_list_data *)sc->pn_ldata_ptr; #if defined(__i386__) round = (u_int32_t)sc->pn_ldata_ptr & 0xf; #elif defined(__alpha__) round = (u_int64_t)sc->pn_ldata_ptr & 0xf; #endif roundptr = sc->pn_ldata_ptr; for (i = 0; i < 8; i++) { if (round % 8) { round++; roundptr++; } else break; } sc->pn_ldata = (struct pn_list_data *)roundptr; bzero(sc->pn_ldata, sizeof(struct pn_list_data)); revision = pci_conf_read(pa->pa_pc, pa->pa_tag, PN_PCI_REVISION) & 0xFF; if (revision == PN_169B_REV || revision == PN_169_REV || (revision & 0xF0) == PN_168_REV) { sc->pn_rx_war = 1; sc->pn_rx_buf = malloc(PN_RXLEN * 5, M_DEVBUF, M_NOWAIT); if (sc->pn_rx_buf == NULL) { printf("%s: no memory for workaround buffer\n", sc->sc_dev.dv_xname); goto fail; } } else sc->pn_rx_war = 0; ifp = &sc->arpcom.ac_if; ifp->if_softc = sc; ifp->if_mtu = ETHERMTU; ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; ifp->if_ioctl = pn_ioctl; ifp->if_output = ether_output; ifp->if_start = pn_start; ifp->if_watchdog = pn_watchdog; ifp->if_baudrate = 10000000; ifp->if_snd.ifq_maxlen = PN_TX_LIST_CNT - 1; bcopy(sc->sc_dev.dv_xname, ifp->if_xname, IFNAMSIZ); ifmedia_init(&sc->ifmedia, 0, pn_ifmedia_upd, pn_ifmedia_sts); for (i = PN_PHYADDR_MIN; i < PN_PHYADDR_MAX + 1; i++) { sc->pn_phy_addr = i; pn_phy_writereg(sc, PHY_BMCR, PHY_BMCR_RESET); DELAY(500); while (pn_phy_readreg(sc, PHY_BMCR) & PHY_BMCR_RESET); if ((phy_sts = pn_phy_readreg(sc, PHY_BMSR))) break; } if (phy_sts) { phy_vid = pn_phy_readreg(sc, PHY_VENID); phy_did = pn_phy_readreg(sc, PHY_DEVID); p = pn_phys; while (p->pn_vid) { if (phy_vid == p->pn_vid && (phy_did | 0xf) == p->pn_vid) { sc->pn_pinfo = p; break; } p++; } if (sc->pn_pinfo == NULL) sc->pn_pinfo = &pn_phys[PHY_UNKNOWN]; pn_getmode_mii(sc); pn_autoneg_mii(sc, PN_FLAG_FORCEDELAY, 1); } else { ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_HDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_10_T|IFM_FDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_HDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_TX|IFM_FDX, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_100_T4, 0, NULL); ifmedia_add(&sc->ifmedia, IFM_ETHER|IFM_AUTO, 0, NULL); pn_autoneg(sc, PN_FLAG_FORCEDELAY, 1); } media = sc->ifmedia.ifm_media; pn_stop(sc); ifmedia_set(&sc->ifmedia, media); if_attach(ifp); ether_ifattach(ifp); #if NBPFILTER > 0 bpfattach(&sc->arpcom.ac_if.if_bpf, ifp, DLT_EN10MB, sizeof(struct ether_header)); #endif shutdownhook_establish(pn_shutdown, sc); fail: splx(s); } static void pn_shutdown(v) void *v; { struct pn_softc *sc = (struct pn_softc *)v; pn_stop(sc); } struct cfattach pn_ca = { sizeof(struct pn_softc), pn_probe, pn_attach }; struct cfdriver pn_cd = { 0, "pn", DV_IFNET }; #endif