Home | History | Annotate | Line # | Download | only in pci
if_bge.c revision 1.104
      1 /*	$NetBSD: if_bge.c,v 1.104 2006/03/04 19:20:50 thorpej Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 2001 Wind River Systems
      5  * Copyright (c) 1997, 1998, 1999, 2001
      6  *	Bill Paul <wpaul (at) windriver.com>.  All rights reserved.
      7  *
      8  * Redistribution and use in source and binary forms, with or without
      9  * modification, are permitted provided that the following conditions
     10  * are met:
     11  * 1. Redistributions of source code must retain the above copyright
     12  *    notice, this list of conditions and the following disclaimer.
     13  * 2. Redistributions in binary form must reproduce the above copyright
     14  *    notice, this list of conditions and the following disclaimer in the
     15  *    documentation and/or other materials provided with the distribution.
     16  * 3. All advertising materials mentioning features or use of this software
     17  *    must display the following acknowledgement:
     18  *	This product includes software developed by Bill Paul.
     19  * 4. Neither the name of the author nor the names of any co-contributors
     20  *    may be used to endorse or promote products derived from this software
     21  *    without specific prior written permission.
     22  *
     23  * THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS'' AND
     24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     26  * ARE DISCLAIMED.  IN NO EVENT SHALL Bill Paul OR THE VOICES IN HIS HEAD
     27  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     28  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     29  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     30  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     31  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     32  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
     33  * THE POSSIBILITY OF SUCH DAMAGE.
     34  *
     35  * $FreeBSD: if_bge.c,v 1.13 2002/04/04 06:01:31 wpaul Exp $
     36  */
     37 
     38 /*
     39  * Broadcom BCM570x family gigabit ethernet driver for NetBSD.
     40  *
     41  * NetBSD version by:
     42  *
     43  *	Frank van der Linden <fvdl (at) wasabisystems.com>
     44  *	Jason Thorpe <thorpej (at) wasabisystems.com>
     45  *	Jonathan Stone <jonathan (at) dsg.stanford.edu>
     46  *
     47  * Originally written for FreeBSD by Bill Paul <wpaul (at) windriver.com>
     48  * Senior Engineer, Wind River Systems
     49  */
     50 
     51 /*
     52  * The Broadcom BCM5700 is based on technology originally developed by
     53  * Alteon Networks as part of the Tigon I and Tigon II gigabit ethernet
     54  * MAC chips. The BCM5700, sometimes refered to as the Tigon III, has
     55  * two on-board MIPS R4000 CPUs and can have as much as 16MB of external
     56  * SSRAM. The BCM5700 supports TCP, UDP and IP checksum offload, jumbo
     57  * frames, highly configurable RX filtering, and 16 RX and TX queues
     58  * (which, along with RX filter rules, can be used for QOS applications).
     59  * Other features, such as TCP segmentation, may be available as part
     60  * of value-added firmware updates. Unlike the Tigon I and Tigon II,
     61  * firmware images can be stored in hardware and need not be compiled
     62  * into the driver.
     63  *
     64  * The BCM5700 supports the PCI v2.2 and PCI-X v1.0 standards, and will
     65  * function in a 32-bit/64-bit 33/66MHz bus, or a 64-bit/133MHz bus.
     66  *
     67  * The BCM5701 is a single-chip solution incorporating both the BCM5700
     68  * MAC and a BCM5401 10/100/1000 PHY. Unlike the BCM5700, the BCM5701
     69  * does not support external SSRAM.
     70  *
     71  * Broadcom also produces a variation of the BCM5700 under the "Altima"
     72  * brand name, which is functionally similar but lacks PCI-X support.
     73  *
     74  * Without external SSRAM, you can only have at most 4 TX rings,
     75  * and the use of the mini RX ring is disabled. This seems to imply
     76  * that these features are simply not available on the BCM5701. As a
     77  * result, this driver does not implement any support for the mini RX
     78  * ring.
     79  */
     80 
     81 #include <sys/cdefs.h>
     82 __KERNEL_RCSID(0, "$NetBSD: if_bge.c,v 1.104 2006/03/04 19:20:50 thorpej Exp $");
     83 
     84 #include "bpfilter.h"
     85 #include "vlan.h"
     86 
     87 #include <sys/param.h>
     88 #include <sys/systm.h>
     89 #include <sys/callout.h>
     90 #include <sys/sockio.h>
     91 #include <sys/mbuf.h>
     92 #include <sys/malloc.h>
     93 #include <sys/kernel.h>
     94 #include <sys/device.h>
     95 #include <sys/socket.h>
     96 #include <sys/sysctl.h>
     97 
     98 #include <net/if.h>
     99 #include <net/if_dl.h>
    100 #include <net/if_media.h>
    101 #include <net/if_ether.h>
    102 
    103 #ifdef INET
    104 #include <netinet/in.h>
    105 #include <netinet/in_systm.h>
    106 #include <netinet/in_var.h>
    107 #include <netinet/ip.h>
    108 #endif
    109 
    110 /* Headers for TCP  Segmentation Offload (TSO) */
    111 #include <netinet/in_systm.h>		/* n_time for <netinet/ip.h>... */
    112 #include <netinet/in.h>			/* ip_{src,dst}, for <netinet/ip.h> */
    113 #include <netinet/ip.h>			/* for struct ip */
    114 #include <netinet/tcp.h>		/* for struct tcphdr */
    115 
    116 
    117 #if NBPFILTER > 0
    118 #include <net/bpf.h>
    119 #endif
    120 
    121 #include <dev/pci/pcireg.h>
    122 #include <dev/pci/pcivar.h>
    123 #include <dev/pci/pcidevs.h>
    124 
    125 #include <dev/mii/mii.h>
    126 #include <dev/mii/miivar.h>
    127 #include <dev/mii/miidevs.h>
    128 #include <dev/mii/brgphyreg.h>
    129 
    130 #include <dev/pci/if_bgereg.h>
    131 
    132 #include <uvm/uvm_extern.h>
    133 
    134 #define ETHER_MIN_NOPAD (ETHER_MIN_LEN - ETHER_CRC_LEN) /* i.e., 60 */
    135 
    136 
    137 /*
    138  * Tunable thresholds for rx-side bge interrupt mitigation.
    139  */
    140 
    141 /*
    142  * The pairs of values below were obtained from empirical measurement
    143  * on bcm5700 rev B2; they ar designed to give roughly 1 receive
    144  * interrupt for every N packets received, where N is, approximately,
    145  * the second value (rx_max_bds) in each pair.  The values are chosen
    146  * such that moving from one pair to the succeeding pair was observed
    147  * to roughly halve interrupt rate under sustained input packet load.
    148  * The values were empirically chosen to avoid overflowing internal
    149  * limits on the  bcm5700: inreasing rx_ticks much beyond 600
    150  * results in internal wrapping and higher interrupt rates.
    151  * The limit of 46 frames was chosen to match NFS workloads.
    152  *
    153  * These values also work well on bcm5701, bcm5704C, and (less
    154  * tested) bcm5703.  On other chipsets, (including the Altima chip
    155  * family), the larger values may overflow internal chip limits,
    156  * leading to increasing interrupt rates rather than lower interrupt
    157  * rates.
    158  *
    159  * Applications using heavy interrupt mitigation (interrupting every
    160  * 32 or 46 frames) in both directions may need to increase the TCP
    161  * windowsize to above 131072 bytes (e.g., to 199608 bytes) to sustain
    162  * full link bandwidth, due to ACKs and window updates lingering
    163  * in the RX queue during the 30-to-40-frame interrupt-mitigation window.
    164  */
    165 static const struct bge_load_rx_thresh {
    166 	int rx_ticks;
    167 	int rx_max_bds; }
    168 bge_rx_threshes[] = {
    169 	{ 32,   2 },
    170 	{ 50,   4 },
    171 	{ 100,  8 },
    172 	{ 192, 16 },
    173 	{ 416, 32 },
    174 	{ 598, 46 }
    175 };
    176 #define NBGE_RX_THRESH (sizeof(bge_rx_threshes) / sizeof(bge_rx_threshes[0]))
    177 
    178 /* XXX patchable; should be sysctl'able */
    179 static int	bge_auto_thresh = 1;
    180 static int	bge_rx_thresh_lvl;
    181 
    182 static int	bge_rxthresh_nodenum;
    183 
    184 static int	bge_probe(device_t, cfdata_t, void *);
    185 static void	bge_attach(device_t, device_t, void *);
    186 static void	bge_powerhook(int, void *);
    187 static void	bge_release_resources(struct bge_softc *);
    188 static void	bge_txeof(struct bge_softc *);
    189 static void	bge_rxeof(struct bge_softc *);
    190 
    191 static void	bge_tick(void *);
    192 static void	bge_stats_update(struct bge_softc *);
    193 static int	bge_encap(struct bge_softc *, struct mbuf *, u_int32_t *);
    194 
    195 static int	bge_intr(void *);
    196 static void	bge_start(struct ifnet *);
    197 static int	bge_ioctl(struct ifnet *, u_long, caddr_t);
    198 static int	bge_init(struct ifnet *);
    199 static void	bge_stop(struct bge_softc *);
    200 static void	bge_watchdog(struct ifnet *);
    201 static void	bge_shutdown(void *);
    202 static int	bge_ifmedia_upd(struct ifnet *);
    203 static void	bge_ifmedia_sts(struct ifnet *, struct ifmediareq *);
    204 
    205 static void	bge_setmulti(struct bge_softc *);
    206 
    207 static void	bge_handle_events(struct bge_softc *);
    208 static int	bge_alloc_jumbo_mem(struct bge_softc *);
    209 #if 0 /* XXX */
    210 static void	bge_free_jumbo_mem(struct bge_softc *);
    211 #endif
    212 static void	*bge_jalloc(struct bge_softc *);
    213 static void	bge_jfree(struct mbuf *, caddr_t, size_t, void *);
    214 static int	bge_newbuf_std(struct bge_softc *, int, struct mbuf *,
    215 			       bus_dmamap_t);
    216 static int	bge_newbuf_jumbo(struct bge_softc *, int, struct mbuf *);
    217 static int	bge_init_rx_ring_std(struct bge_softc *);
    218 static void	bge_free_rx_ring_std(struct bge_softc *);
    219 static int	bge_init_rx_ring_jumbo(struct bge_softc *);
    220 static void	bge_free_rx_ring_jumbo(struct bge_softc *);
    221 static void	bge_free_tx_ring(struct bge_softc *);
    222 static int	bge_init_tx_ring(struct bge_softc *);
    223 
    224 static int	bge_chipinit(struct bge_softc *);
    225 static int	bge_blockinit(struct bge_softc *);
    226 static int	bge_setpowerstate(struct bge_softc *, int);
    227 
    228 static void	bge_reset(struct bge_softc *);
    229 
    230 #define BGE_DEBUG
    231 #ifdef BGE_DEBUG
    232 #define DPRINTF(x)	if (bgedebug) printf x
    233 #define DPRINTFN(n,x)	if (bgedebug >= (n)) printf x
    234 #define BGE_TSO_PRINTF(x)  do { if (bge_tso_debug) printf x ;} while (0)
    235 int	bgedebug = 0;
    236 int	bge_tso_debug = 0;
    237 #else
    238 #define DPRINTF(x)
    239 #define DPRINTFN(n,x)
    240 #define BGE_TSO_PRINTF(x)
    241 #endif
    242 
    243 #ifdef BGE_EVENT_COUNTERS
    244 #define	BGE_EVCNT_INCR(ev)	(ev).ev_count++
    245 #define	BGE_EVCNT_ADD(ev, val)	(ev).ev_count += (val)
    246 #define	BGE_EVCNT_UPD(ev, val)	(ev).ev_count = (val)
    247 #else
    248 #define	BGE_EVCNT_INCR(ev)	/* nothing */
    249 #define	BGE_EVCNT_ADD(ev, val)	/* nothing */
    250 #define	BGE_EVCNT_UPD(ev, val)	/* nothing */
    251 #endif
    252 
    253 /* Various chip quirks. */
    254 #define	BGE_QUIRK_LINK_STATE_BROKEN	0x00000001
    255 #define	BGE_QUIRK_CSUM_BROKEN		0x00000002
    256 #define	BGE_QUIRK_ONLY_PHY_1		0x00000004
    257 #define	BGE_QUIRK_5700_SMALLDMA		0x00000008
    258 #define	BGE_QUIRK_5700_PCIX_REG_BUG	0x00000010
    259 #define	BGE_QUIRK_PRODUCER_BUG		0x00000020
    260 #define	BGE_QUIRK_PCIX_DMA_ALIGN_BUG	0x00000040
    261 #define	BGE_QUIRK_5705_CORE		0x00000080
    262 #define	BGE_QUIRK_FEWER_MBUFS		0x00000100
    263 
    264 /*
    265  * XXX: how to handle variants based on 5750 and derivatives:
    266  * 5750 5751, 5721, possibly 5714, 5752, and 5708?, which
    267  * in general behave like a 5705, except with additional quirks.
    268  * This driver's current handling of the 5721 is wrong;
    269  * how we map ASIC revision to "quirks" needs more thought.
    270  * (defined here until the thought is done).
    271  */
    272 #define BGE_IS_5714_FAMILY(sc) \
    273 	(BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5714 ||	\
    274 	 BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5780 ||	\
    275 	 BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5715 )
    276 
    277 #define BGE_IS_5750_OR_BEYOND(sc)  \
    278 	(BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5750 || \
    279 	 BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5752 || \
    280 	 BGE_IS_5714_FAMILY(sc) )
    281 
    282 #define BGE_IS_5705_OR_BEYOND(sc)  \
    283 	( ((sc)->bge_quirks & BGE_QUIRK_5705_CORE) || \
    284 	  BGE_IS_5750_OR_BEYOND(sc) )
    285 
    286 
    287 /* following bugs are common to bcm5700 rev B, all flavours */
    288 #define BGE_QUIRK_5700_COMMON \
    289 	(BGE_QUIRK_5700_SMALLDMA|BGE_QUIRK_PRODUCER_BUG)
    290 
    291 CFATTACH_DECL(bge, sizeof(struct bge_softc),
    292     bge_probe, bge_attach, NULL, NULL);
    293 
    294 static u_int32_t
    295 bge_readmem_ind(struct bge_softc *sc, int off)
    296 {
    297 	struct pci_attach_args	*pa = &(sc->bge_pa);
    298 	pcireg_t val;
    299 
    300 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MEMWIN_BASEADDR, off);
    301 	val = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_MEMWIN_DATA);
    302 	return val;
    303 }
    304 
    305 static void
    306 bge_writemem_ind(struct bge_softc *sc, int off, int val)
    307 {
    308 	struct pci_attach_args	*pa = &(sc->bge_pa);
    309 
    310 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MEMWIN_BASEADDR, off);
    311 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MEMWIN_DATA, val);
    312 }
    313 
    314 #ifdef notdef
    315 static u_int32_t
    316 bge_readreg_ind(struct bge_softc *sc, int off)
    317 {
    318 	struct pci_attach_args	*pa = &(sc->bge_pa);
    319 
    320 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_REG_BASEADDR, off);
    321 	return(pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_REG_DATA));
    322 }
    323 #endif
    324 
    325 static void
    326 bge_writereg_ind(struct bge_softc *sc, int off, int val)
    327 {
    328 	struct pci_attach_args	*pa = &(sc->bge_pa);
    329 
    330 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_REG_BASEADDR, off);
    331 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_REG_DATA, val);
    332 }
    333 
    334 #ifdef notdef
    335 static u_int8_t
    336 bge_vpd_readbyte(struct bge_softc *sc, int addr)
    337 {
    338 	int i;
    339 	u_int32_t val;
    340 	struct pci_attach_args	*pa = &(sc->bge_pa);
    341 
    342 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_VPD_ADDR, addr);
    343 	for (i = 0; i < BGE_TIMEOUT * 10; i++) {
    344 		DELAY(10);
    345 		if (pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_VPD_ADDR) &
    346 		    BGE_VPD_FLAG)
    347 			break;
    348 	}
    349 
    350 	if (i == BGE_TIMEOUT) {
    351 		printf("%s: VPD read timed out\n", sc->bge_dev.dv_xname);
    352 		return(0);
    353 	}
    354 
    355 	val = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_VPD_DATA);
    356 
    357 	return((val >> ((addr % 4) * 8)) & 0xFF);
    358 }
    359 
    360 static void
    361 bge_vpd_read_res(struct bge_softc *sc, struct vpd_res *res, int addr)
    362 {
    363 	int i;
    364 	u_int8_t *ptr;
    365 
    366 	ptr = (u_int8_t *)res;
    367 	for (i = 0; i < sizeof(struct vpd_res); i++)
    368 		ptr[i] = bge_vpd_readbyte(sc, i + addr);
    369 }
    370 
    371 static void
    372 bge_vpd_read(struct bge_softc *sc)
    373 {
    374 	int pos = 0, i;
    375 	struct vpd_res res;
    376 
    377 	if (sc->bge_vpd_prodname != NULL)
    378 		free(sc->bge_vpd_prodname, M_DEVBUF);
    379 	if (sc->bge_vpd_readonly != NULL)
    380 		free(sc->bge_vpd_readonly, M_DEVBUF);
    381 	sc->bge_vpd_prodname = NULL;
    382 	sc->bge_vpd_readonly = NULL;
    383 
    384 	bge_vpd_read_res(sc, &res, pos);
    385 
    386 	if (res.vr_id != VPD_RES_ID) {
    387 		printf("%s: bad VPD resource id: expected %x got %x\n",
    388 			sc->bge_dev.dv_xname, VPD_RES_ID, res.vr_id);
    389 		return;
    390 	}
    391 
    392 	pos += sizeof(res);
    393 	sc->bge_vpd_prodname = malloc(res.vr_len + 1, M_DEVBUF, M_NOWAIT);
    394 	if (sc->bge_vpd_prodname == NULL)
    395 		panic("bge_vpd_read");
    396 	for (i = 0; i < res.vr_len; i++)
    397 		sc->bge_vpd_prodname[i] = bge_vpd_readbyte(sc, i + pos);
    398 	sc->bge_vpd_prodname[i] = '\0';
    399 	pos += i;
    400 
    401 	bge_vpd_read_res(sc, &res, pos);
    402 
    403 	if (res.vr_id != VPD_RES_READ) {
    404 		printf("%s: bad VPD resource id: expected %x got %x\n",
    405 		    sc->bge_dev.dv_xname, VPD_RES_READ, res.vr_id);
    406 		return;
    407 	}
    408 
    409 	pos += sizeof(res);
    410 	sc->bge_vpd_readonly = malloc(res.vr_len, M_DEVBUF, M_NOWAIT);
    411 	if (sc->bge_vpd_readonly == NULL)
    412 		panic("bge_vpd_read");
    413 	for (i = 0; i < res.vr_len + 1; i++)
    414 		sc->bge_vpd_readonly[i] = bge_vpd_readbyte(sc, i + pos);
    415 }
    416 #endif
    417 
    418 /*
    419  * Read a byte of data stored in the EEPROM at address 'addr.' The
    420  * BCM570x supports both the traditional bitbang interface and an
    421  * auto access interface for reading the EEPROM. We use the auto
    422  * access method.
    423  */
    424 static u_int8_t
    425 bge_eeprom_getbyte(struct bge_softc *sc, int addr, u_int8_t *dest)
    426 {
    427 	int i;
    428 	u_int32_t byte = 0;
    429 
    430 	/*
    431 	 * Enable use of auto EEPROM access so we can avoid
    432 	 * having to use the bitbang method.
    433 	 */
    434 	BGE_SETBIT(sc, BGE_MISC_LOCAL_CTL, BGE_MLC_AUTO_EEPROM);
    435 
    436 	/* Reset the EEPROM, load the clock period. */
    437 	CSR_WRITE_4(sc, BGE_EE_ADDR,
    438 	    BGE_EEADDR_RESET|BGE_EEHALFCLK(BGE_HALFCLK_384SCL));
    439 	DELAY(20);
    440 
    441 	/* Issue the read EEPROM command. */
    442 	CSR_WRITE_4(sc, BGE_EE_ADDR, BGE_EE_READCMD | addr);
    443 
    444 	/* Wait for completion */
    445 	for(i = 0; i < BGE_TIMEOUT * 10; i++) {
    446 		DELAY(10);
    447 		if (CSR_READ_4(sc, BGE_EE_ADDR) & BGE_EEADDR_DONE)
    448 			break;
    449 	}
    450 
    451 	if (i == BGE_TIMEOUT) {
    452 		printf("%s: eeprom read timed out\n", sc->bge_dev.dv_xname);
    453 		return(0);
    454 	}
    455 
    456 	/* Get result. */
    457 	byte = CSR_READ_4(sc, BGE_EE_DATA);
    458 
    459 	*dest = (byte >> ((addr % 4) * 8)) & 0xFF;
    460 
    461 	return(0);
    462 }
    463 
    464 /*
    465  * Read a sequence of bytes from the EEPROM.
    466  */
    467 static int
    468 bge_read_eeprom(struct bge_softc *sc, caddr_t dest, int off, int cnt)
    469 {
    470 	int err = 0, i;
    471 	u_int8_t byte = 0;
    472 
    473 	for (i = 0; i < cnt; i++) {
    474 		err = bge_eeprom_getbyte(sc, off + i, &byte);
    475 		if (err)
    476 			break;
    477 		*(dest + i) = byte;
    478 	}
    479 
    480 	return(err ? 1 : 0);
    481 }
    482 
    483 static int
    484 bge_miibus_readreg(device_t dev, int phy, int reg)
    485 {
    486 	struct bge_softc *sc = (struct bge_softc *)dev;
    487 	u_int32_t val;
    488 	u_int32_t saved_autopoll;
    489 	int i;
    490 
    491 	/*
    492 	 * Several chips with builtin PHYs will incorrectly answer to
    493 	 * other PHY instances than the builtin PHY at id 1.
    494 	 */
    495 	if (phy != 1 && (sc->bge_quirks & BGE_QUIRK_ONLY_PHY_1))
    496 		return(0);
    497 
    498 	/* Reading with autopolling on may trigger PCI errors */
    499 	saved_autopoll = CSR_READ_4(sc, BGE_MI_MODE);
    500 	if (saved_autopoll & BGE_MIMODE_AUTOPOLL) {
    501 		CSR_WRITE_4(sc, BGE_MI_MODE,
    502 		    saved_autopoll &~ BGE_MIMODE_AUTOPOLL);
    503 		DELAY(40);
    504 	}
    505 
    506 	CSR_WRITE_4(sc, BGE_MI_COMM, BGE_MICMD_READ|BGE_MICOMM_BUSY|
    507 	    BGE_MIPHY(phy)|BGE_MIREG(reg));
    508 
    509 	for (i = 0; i < BGE_TIMEOUT; i++) {
    510 		val = CSR_READ_4(sc, BGE_MI_COMM);
    511 		if (!(val & BGE_MICOMM_BUSY))
    512 			break;
    513 		delay(10);
    514 	}
    515 
    516 	if (i == BGE_TIMEOUT) {
    517 		printf("%s: PHY read timed out\n", sc->bge_dev.dv_xname);
    518 		val = 0;
    519 		goto done;
    520 	}
    521 
    522 	val = CSR_READ_4(sc, BGE_MI_COMM);
    523 
    524 done:
    525 	if (saved_autopoll & BGE_MIMODE_AUTOPOLL) {
    526 		CSR_WRITE_4(sc, BGE_MI_MODE, saved_autopoll);
    527 		DELAY(40);
    528 	}
    529 
    530 	if (val & BGE_MICOMM_READFAIL)
    531 		return(0);
    532 
    533 	return(val & 0xFFFF);
    534 }
    535 
    536 static void
    537 bge_miibus_writereg(device_t dev, int phy, int reg, int val)
    538 {
    539 	struct bge_softc *sc = (struct bge_softc *)dev;
    540 	u_int32_t saved_autopoll;
    541 	int i;
    542 
    543 	/* Touching the PHY while autopolling is on may trigger PCI errors */
    544 	saved_autopoll = CSR_READ_4(sc, BGE_MI_MODE);
    545 	if (saved_autopoll & BGE_MIMODE_AUTOPOLL) {
    546 		delay(40);
    547 		CSR_WRITE_4(sc, BGE_MI_MODE,
    548 		    saved_autopoll & (~BGE_MIMODE_AUTOPOLL));
    549 		delay(10); /* 40 usec is supposed to be adequate */
    550 	}
    551 
    552 	CSR_WRITE_4(sc, BGE_MI_COMM, BGE_MICMD_WRITE|BGE_MICOMM_BUSY|
    553 	    BGE_MIPHY(phy)|BGE_MIREG(reg)|val);
    554 
    555 	for (i = 0; i < BGE_TIMEOUT; i++) {
    556 		if (!(CSR_READ_4(sc, BGE_MI_COMM) & BGE_MICOMM_BUSY))
    557 			break;
    558 		delay(10);
    559 	}
    560 
    561 	if (saved_autopoll & BGE_MIMODE_AUTOPOLL) {
    562 		CSR_WRITE_4(sc, BGE_MI_MODE, saved_autopoll);
    563 		delay(40);
    564 	}
    565 
    566 	if (i == BGE_TIMEOUT) {
    567 		printf("%s: PHY read timed out\n", sc->bge_dev.dv_xname);
    568 	}
    569 }
    570 
    571 static void
    572 bge_miibus_statchg(device_t dev)
    573 {
    574 	struct bge_softc *sc = (struct bge_softc *)dev;
    575 	struct mii_data *mii = &sc->bge_mii;
    576 
    577 	/*
    578 	 * Get flow control negotiation result.
    579 	 */
    580 	if (IFM_SUBTYPE(mii->mii_media.ifm_cur->ifm_media) == IFM_AUTO &&
    581 	    (mii->mii_media_active & IFM_ETH_FMASK) != sc->bge_flowflags) {
    582 		sc->bge_flowflags = mii->mii_media_active & IFM_ETH_FMASK;
    583 		mii->mii_media_active &= ~IFM_ETH_FMASK;
    584 	}
    585 
    586 	BGE_CLRBIT(sc, BGE_MAC_MODE, BGE_MACMODE_PORTMODE);
    587 	if (IFM_SUBTYPE(mii->mii_media_active) == IFM_1000_T) {
    588 		BGE_SETBIT(sc, BGE_MAC_MODE, BGE_PORTMODE_GMII);
    589 	} else {
    590 		BGE_SETBIT(sc, BGE_MAC_MODE, BGE_PORTMODE_MII);
    591 	}
    592 
    593 	if ((mii->mii_media_active & IFM_GMASK) == IFM_FDX) {
    594 		BGE_CLRBIT(sc, BGE_MAC_MODE, BGE_MACMODE_HALF_DUPLEX);
    595 	} else {
    596 		BGE_SETBIT(sc, BGE_MAC_MODE, BGE_MACMODE_HALF_DUPLEX);
    597 	}
    598 
    599 	/*
    600 	 * 802.3x flow control
    601 	 */
    602 	if (sc->bge_flowflags & IFM_ETH_RXPAUSE) {
    603 		BGE_SETBIT(sc, BGE_RX_MODE, BGE_RXMODE_FLOWCTL_ENABLE);
    604 	} else {
    605 		BGE_CLRBIT(sc, BGE_RX_MODE, BGE_RXMODE_FLOWCTL_ENABLE);
    606 	}
    607 	if (sc->bge_flowflags & IFM_ETH_TXPAUSE) {
    608 		BGE_SETBIT(sc, BGE_TX_MODE, BGE_TXMODE_FLOWCTL_ENABLE);
    609 	} else {
    610 		BGE_CLRBIT(sc, BGE_TX_MODE, BGE_TXMODE_FLOWCTL_ENABLE);
    611 	}
    612 }
    613 
    614 /*
    615  * Update rx threshold levels to values in a particular slot
    616  * of the interrupt-mitigation table bge_rx_threshes.
    617  */
    618 static void
    619 bge_set_thresh(struct ifnet *ifp, int lvl)
    620 {
    621 	struct bge_softc *sc = ifp->if_softc;
    622 	int s;
    623 
    624 	/* For now, just save the new Rx-intr thresholds and record
    625 	 * that a threshold update is pending.  Updating the hardware
    626 	 * registers here (even at splhigh()) is observed to
    627 	 * occasionaly cause glitches where Rx-interrupts are not
    628 	 * honoured for up to 10 seconds. jonathan (at) NetBSD.org, 2003-04-05
    629 	 */
    630 	s = splnet();
    631 	sc->bge_rx_coal_ticks = bge_rx_threshes[lvl].rx_ticks;
    632 	sc->bge_rx_max_coal_bds = bge_rx_threshes[lvl].rx_max_bds;
    633 	sc->bge_pending_rxintr_change = 1;
    634 	splx(s);
    635 
    636 	 return;
    637 }
    638 
    639 
    640 /*
    641  * Update Rx thresholds of all bge devices
    642  */
    643 static void
    644 bge_update_all_threshes(int lvl)
    645 {
    646 	struct ifnet *ifp;
    647 	const char * const namebuf = "bge";
    648 	int namelen;
    649 
    650 	if (lvl < 0)
    651 		lvl = 0;
    652 	else if( lvl >= NBGE_RX_THRESH)
    653 		lvl = NBGE_RX_THRESH - 1;
    654 
    655 	namelen = strlen(namebuf);
    656 	/*
    657 	 * Now search all the interfaces for this name/number
    658 	 */
    659 	IFNET_FOREACH(ifp) {
    660 		if (strncmp(ifp->if_xname, namebuf, namelen) != 0)
    661 		      continue;
    662 		/* We got a match: update if doing auto-threshold-tuning */
    663 		if (bge_auto_thresh)
    664 			bge_set_thresh(ifp, lvl);
    665 	}
    666 }
    667 
    668 /*
    669  * Handle events that have triggered interrupts.
    670  */
    671 static void
    672 bge_handle_events(struct bge_softc *sc)
    673 {
    674 
    675 	return;
    676 }
    677 
    678 /*
    679  * Memory management for jumbo frames.
    680  */
    681 
    682 static int
    683 bge_alloc_jumbo_mem(struct bge_softc *sc)
    684 {
    685 	caddr_t			ptr, kva;
    686 	bus_dma_segment_t	seg;
    687 	int		i, rseg, state, error;
    688 	struct bge_jpool_entry   *entry;
    689 
    690 	state = error = 0;
    691 
    692 	/* Grab a big chunk o' storage. */
    693 	if (bus_dmamem_alloc(sc->bge_dmatag, BGE_JMEM, PAGE_SIZE, 0,
    694 	     &seg, 1, &rseg, BUS_DMA_NOWAIT)) {
    695 		printf("%s: can't alloc rx buffers\n", sc->bge_dev.dv_xname);
    696 		return ENOBUFS;
    697 	}
    698 
    699 	state = 1;
    700 	if (bus_dmamem_map(sc->bge_dmatag, &seg, rseg, BGE_JMEM, &kva,
    701 	    BUS_DMA_NOWAIT)) {
    702 		printf("%s: can't map DMA buffers (%d bytes)\n",
    703 		    sc->bge_dev.dv_xname, (int)BGE_JMEM);
    704 		error = ENOBUFS;
    705 		goto out;
    706 	}
    707 
    708 	state = 2;
    709 	if (bus_dmamap_create(sc->bge_dmatag, BGE_JMEM, 1, BGE_JMEM, 0,
    710 	    BUS_DMA_NOWAIT, &sc->bge_cdata.bge_rx_jumbo_map)) {
    711 		printf("%s: can't create DMA map\n", sc->bge_dev.dv_xname);
    712 		error = ENOBUFS;
    713 		goto out;
    714 	}
    715 
    716 	state = 3;
    717 	if (bus_dmamap_load(sc->bge_dmatag, sc->bge_cdata.bge_rx_jumbo_map,
    718 	    kva, BGE_JMEM, NULL, BUS_DMA_NOWAIT)) {
    719 		printf("%s: can't load DMA map\n", sc->bge_dev.dv_xname);
    720 		error = ENOBUFS;
    721 		goto out;
    722 	}
    723 
    724 	state = 4;
    725 	sc->bge_cdata.bge_jumbo_buf = (caddr_t)kva;
    726 	DPRINTFN(1,("bge_jumbo_buf = %p\n", sc->bge_cdata.bge_jumbo_buf));
    727 
    728 	SLIST_INIT(&sc->bge_jfree_listhead);
    729 	SLIST_INIT(&sc->bge_jinuse_listhead);
    730 
    731 	/*
    732 	 * Now divide it up into 9K pieces and save the addresses
    733 	 * in an array.
    734 	 */
    735 	ptr = sc->bge_cdata.bge_jumbo_buf;
    736 	for (i = 0; i < BGE_JSLOTS; i++) {
    737 		sc->bge_cdata.bge_jslots[i] = ptr;
    738 		ptr += BGE_JLEN;
    739 		entry = malloc(sizeof(struct bge_jpool_entry),
    740 		    M_DEVBUF, M_NOWAIT);
    741 		if (entry == NULL) {
    742 			printf("%s: no memory for jumbo buffer queue!\n",
    743 			    sc->bge_dev.dv_xname);
    744 			error = ENOBUFS;
    745 			goto out;
    746 		}
    747 		entry->slot = i;
    748 		SLIST_INSERT_HEAD(&sc->bge_jfree_listhead,
    749 				 entry, jpool_entries);
    750 	}
    751 out:
    752 	if (error != 0) {
    753 		switch (state) {
    754 		case 4:
    755 			bus_dmamap_unload(sc->bge_dmatag,
    756 			    sc->bge_cdata.bge_rx_jumbo_map);
    757 		case 3:
    758 			bus_dmamap_destroy(sc->bge_dmatag,
    759 			    sc->bge_cdata.bge_rx_jumbo_map);
    760 		case 2:
    761 			bus_dmamem_unmap(sc->bge_dmatag, kva, BGE_JMEM);
    762 		case 1:
    763 			bus_dmamem_free(sc->bge_dmatag, &seg, rseg);
    764 			break;
    765 		default:
    766 			break;
    767 		}
    768 	}
    769 
    770 	return error;
    771 }
    772 
    773 /*
    774  * Allocate a jumbo buffer.
    775  */
    776 static void *
    777 bge_jalloc(struct bge_softc *sc)
    778 {
    779 	struct bge_jpool_entry   *entry;
    780 
    781 	entry = SLIST_FIRST(&sc->bge_jfree_listhead);
    782 
    783 	if (entry == NULL) {
    784 		printf("%s: no free jumbo buffers\n", sc->bge_dev.dv_xname);
    785 		return(NULL);
    786 	}
    787 
    788 	SLIST_REMOVE_HEAD(&sc->bge_jfree_listhead, jpool_entries);
    789 	SLIST_INSERT_HEAD(&sc->bge_jinuse_listhead, entry, jpool_entries);
    790 	return(sc->bge_cdata.bge_jslots[entry->slot]);
    791 }
    792 
    793 /*
    794  * Release a jumbo buffer.
    795  */
    796 static void
    797 bge_jfree(struct mbuf *m, caddr_t buf, size_t size, void *arg)
    798 {
    799 	struct bge_jpool_entry *entry;
    800 	struct bge_softc *sc;
    801 	int i, s;
    802 
    803 	/* Extract the softc struct pointer. */
    804 	sc = (struct bge_softc *)arg;
    805 
    806 	if (sc == NULL)
    807 		panic("bge_jfree: can't find softc pointer!");
    808 
    809 	/* calculate the slot this buffer belongs to */
    810 
    811 	i = ((caddr_t)buf
    812 	     - (caddr_t)sc->bge_cdata.bge_jumbo_buf) / BGE_JLEN;
    813 
    814 	if ((i < 0) || (i >= BGE_JSLOTS))
    815 		panic("bge_jfree: asked to free buffer that we don't manage!");
    816 
    817 	s = splvm();
    818 	entry = SLIST_FIRST(&sc->bge_jinuse_listhead);
    819 	if (entry == NULL)
    820 		panic("bge_jfree: buffer not in use!");
    821 	entry->slot = i;
    822 	SLIST_REMOVE_HEAD(&sc->bge_jinuse_listhead, jpool_entries);
    823 	SLIST_INSERT_HEAD(&sc->bge_jfree_listhead, entry, jpool_entries);
    824 
    825 	if (__predict_true(m != NULL))
    826   		pool_cache_put(&mbpool_cache, m);
    827 	splx(s);
    828 }
    829 
    830 
    831 /*
    832  * Intialize a standard receive ring descriptor.
    833  */
    834 static int
    835 bge_newbuf_std(struct bge_softc *sc, int i, struct mbuf *m, bus_dmamap_t dmamap)
    836 {
    837 	struct mbuf		*m_new = NULL;
    838 	struct bge_rx_bd	*r;
    839 	int			error;
    840 
    841 	if (dmamap == NULL) {
    842 		error = bus_dmamap_create(sc->bge_dmatag, MCLBYTES, 1,
    843 		    MCLBYTES, 0, BUS_DMA_NOWAIT, &dmamap);
    844 		if (error != 0)
    845 			return error;
    846 	}
    847 
    848 	sc->bge_cdata.bge_rx_std_map[i] = dmamap;
    849 
    850 	if (m == NULL) {
    851 		MGETHDR(m_new, M_DONTWAIT, MT_DATA);
    852 		if (m_new == NULL) {
    853 			return(ENOBUFS);
    854 		}
    855 
    856 		MCLGET(m_new, M_DONTWAIT);
    857 		if (!(m_new->m_flags & M_EXT)) {
    858 			m_freem(m_new);
    859 			return(ENOBUFS);
    860 		}
    861 		m_new->m_len = m_new->m_pkthdr.len = MCLBYTES;
    862 		if (!sc->bge_rx_alignment_bug)
    863 		    m_adj(m_new, ETHER_ALIGN);
    864 
    865 		if (bus_dmamap_load_mbuf(sc->bge_dmatag, dmamap, m_new,
    866 		    BUS_DMA_READ|BUS_DMA_NOWAIT))
    867 			return(ENOBUFS);
    868 	} else {
    869 		m_new = m;
    870 		m_new->m_len = m_new->m_pkthdr.len = MCLBYTES;
    871 		m_new->m_data = m_new->m_ext.ext_buf;
    872 		if (!sc->bge_rx_alignment_bug)
    873 		    m_adj(m_new, ETHER_ALIGN);
    874 	}
    875 
    876 	sc->bge_cdata.bge_rx_std_chain[i] = m_new;
    877 	r = &sc->bge_rdata->bge_rx_std_ring[i];
    878 	bge_set_hostaddr(&r->bge_addr,
    879 	    dmamap->dm_segs[0].ds_addr);
    880 	r->bge_flags = BGE_RXBDFLAG_END;
    881 	r->bge_len = m_new->m_len;
    882 	r->bge_idx = i;
    883 
    884 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
    885 	    offsetof(struct bge_ring_data, bge_rx_std_ring) +
    886 		i * sizeof (struct bge_rx_bd),
    887 	    sizeof (struct bge_rx_bd),
    888 	    BUS_DMASYNC_PREWRITE|BUS_DMASYNC_PREREAD);
    889 
    890 	return(0);
    891 }
    892 
    893 /*
    894  * Initialize a jumbo receive ring descriptor. This allocates
    895  * a jumbo buffer from the pool managed internally by the driver.
    896  */
    897 static int
    898 bge_newbuf_jumbo(struct bge_softc *sc, int i, struct mbuf *m)
    899 {
    900 	struct mbuf *m_new = NULL;
    901 	struct bge_rx_bd *r;
    902 
    903 	if (m == NULL) {
    904 		caddr_t			buf = NULL;
    905 
    906 		/* Allocate the mbuf. */
    907 		MGETHDR(m_new, M_DONTWAIT, MT_DATA);
    908 		if (m_new == NULL) {
    909 			return(ENOBUFS);
    910 		}
    911 
    912 		/* Allocate the jumbo buffer */
    913 		buf = bge_jalloc(sc);
    914 		if (buf == NULL) {
    915 			m_freem(m_new);
    916 			printf("%s: jumbo allocation failed "
    917 			    "-- packet dropped!\n", sc->bge_dev.dv_xname);
    918 			return(ENOBUFS);
    919 		}
    920 
    921 		/* Attach the buffer to the mbuf. */
    922 		m_new->m_len = m_new->m_pkthdr.len = BGE_JUMBO_FRAMELEN;
    923 		MEXTADD(m_new, buf, BGE_JUMBO_FRAMELEN, M_DEVBUF,
    924 		    bge_jfree, sc);
    925 		m_new->m_flags |= M_EXT_RW;
    926 	} else {
    927 		m_new = m;
    928 		m_new->m_data = m_new->m_ext.ext_buf;
    929 		m_new->m_ext.ext_size = BGE_JUMBO_FRAMELEN;
    930 	}
    931 
    932 	if (!sc->bge_rx_alignment_bug)
    933 	    m_adj(m_new, ETHER_ALIGN);
    934 	/* Set up the descriptor. */
    935 	r = &sc->bge_rdata->bge_rx_jumbo_ring[i];
    936 	sc->bge_cdata.bge_rx_jumbo_chain[i] = m_new;
    937 	bge_set_hostaddr(&r->bge_addr, BGE_JUMBO_DMA_ADDR(sc, m_new));
    938 	r->bge_flags = BGE_RXBDFLAG_END|BGE_RXBDFLAG_JUMBO_RING;
    939 	r->bge_len = m_new->m_len;
    940 	r->bge_idx = i;
    941 
    942 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
    943 	    offsetof(struct bge_ring_data, bge_rx_jumbo_ring) +
    944 		i * sizeof (struct bge_rx_bd),
    945 	    sizeof (struct bge_rx_bd),
    946 	    BUS_DMASYNC_PREWRITE|BUS_DMASYNC_PREREAD);
    947 
    948 	return(0);
    949 }
    950 
    951 /*
    952  * The standard receive ring has 512 entries in it. At 2K per mbuf cluster,
    953  * that's 1MB or memory, which is a lot. For now, we fill only the first
    954  * 256 ring entries and hope that our CPU is fast enough to keep up with
    955  * the NIC.
    956  */
    957 static int
    958 bge_init_rx_ring_std(struct bge_softc *sc)
    959 {
    960 	int i;
    961 
    962 	if (sc->bge_flags & BGE_RXRING_VALID)
    963 		return 0;
    964 
    965 	for (i = 0; i < BGE_SSLOTS; i++) {
    966 		if (bge_newbuf_std(sc, i, NULL, 0) == ENOBUFS)
    967 			return(ENOBUFS);
    968 	}
    969 
    970 	sc->bge_std = i - 1;
    971 	CSR_WRITE_4(sc, BGE_MBX_RX_STD_PROD_LO, sc->bge_std);
    972 
    973 	sc->bge_flags |= BGE_RXRING_VALID;
    974 
    975 	return(0);
    976 }
    977 
    978 static void
    979 bge_free_rx_ring_std(struct bge_softc *sc)
    980 {
    981 	int i;
    982 
    983 	if (!(sc->bge_flags & BGE_RXRING_VALID))
    984 		return;
    985 
    986 	for (i = 0; i < BGE_STD_RX_RING_CNT; i++) {
    987 		if (sc->bge_cdata.bge_rx_std_chain[i] != NULL) {
    988 			m_freem(sc->bge_cdata.bge_rx_std_chain[i]);
    989 			sc->bge_cdata.bge_rx_std_chain[i] = NULL;
    990 			bus_dmamap_destroy(sc->bge_dmatag,
    991 			    sc->bge_cdata.bge_rx_std_map[i]);
    992 		}
    993 		memset((char *)&sc->bge_rdata->bge_rx_std_ring[i], 0,
    994 		    sizeof(struct bge_rx_bd));
    995 	}
    996 
    997 	sc->bge_flags &= ~BGE_RXRING_VALID;
    998 }
    999 
   1000 static int
   1001 bge_init_rx_ring_jumbo(struct bge_softc *sc)
   1002 {
   1003 	int i;
   1004 	volatile struct bge_rcb *rcb;
   1005 
   1006 	if (sc->bge_flags & BGE_JUMBO_RXRING_VALID)
   1007 		return 0;
   1008 
   1009 	for (i = 0; i < BGE_JUMBO_RX_RING_CNT; i++) {
   1010 		if (bge_newbuf_jumbo(sc, i, NULL) == ENOBUFS)
   1011 			return(ENOBUFS);
   1012 	};
   1013 
   1014 	sc->bge_jumbo = i - 1;
   1015 	sc->bge_flags |= BGE_JUMBO_RXRING_VALID;
   1016 
   1017 	rcb = &sc->bge_rdata->bge_info.bge_jumbo_rx_rcb;
   1018 	rcb->bge_maxlen_flags = 0;
   1019 	CSR_WRITE_4(sc, BGE_RX_JUMBO_RCB_MAXLEN_FLAGS, rcb->bge_maxlen_flags);
   1020 
   1021 	CSR_WRITE_4(sc, BGE_MBX_RX_JUMBO_PROD_LO, sc->bge_jumbo);
   1022 
   1023 	return(0);
   1024 }
   1025 
   1026 static void
   1027 bge_free_rx_ring_jumbo(struct bge_softc *sc)
   1028 {
   1029 	int i;
   1030 
   1031 	if (!(sc->bge_flags & BGE_JUMBO_RXRING_VALID))
   1032 		return;
   1033 
   1034 	for (i = 0; i < BGE_JUMBO_RX_RING_CNT; i++) {
   1035 		if (sc->bge_cdata.bge_rx_jumbo_chain[i] != NULL) {
   1036 			m_freem(sc->bge_cdata.bge_rx_jumbo_chain[i]);
   1037 			sc->bge_cdata.bge_rx_jumbo_chain[i] = NULL;
   1038 		}
   1039 		memset((char *)&sc->bge_rdata->bge_rx_jumbo_ring[i], 0,
   1040 		    sizeof(struct bge_rx_bd));
   1041 	}
   1042 
   1043 	sc->bge_flags &= ~BGE_JUMBO_RXRING_VALID;
   1044 }
   1045 
   1046 static void
   1047 bge_free_tx_ring(struct bge_softc *sc)
   1048 {
   1049 	int i, freed;
   1050 	struct txdmamap_pool_entry *dma;
   1051 
   1052 	if (!(sc->bge_flags & BGE_TXRING_VALID))
   1053 		return;
   1054 
   1055 	freed = 0;
   1056 
   1057 	for (i = 0; i < BGE_TX_RING_CNT; i++) {
   1058 		if (sc->bge_cdata.bge_tx_chain[i] != NULL) {
   1059 			freed++;
   1060 			m_freem(sc->bge_cdata.bge_tx_chain[i]);
   1061 			sc->bge_cdata.bge_tx_chain[i] = NULL;
   1062 			SLIST_INSERT_HEAD(&sc->txdma_list, sc->txdma[i],
   1063 					    link);
   1064 			sc->txdma[i] = 0;
   1065 		}
   1066 		memset((char *)&sc->bge_rdata->bge_tx_ring[i], 0,
   1067 		    sizeof(struct bge_tx_bd));
   1068 	}
   1069 
   1070 	while ((dma = SLIST_FIRST(&sc->txdma_list))) {
   1071 		SLIST_REMOVE_HEAD(&sc->txdma_list, link);
   1072 		bus_dmamap_destroy(sc->bge_dmatag, dma->dmamap);
   1073 		free(dma, M_DEVBUF);
   1074 	}
   1075 
   1076 	sc->bge_flags &= ~BGE_TXRING_VALID;
   1077 }
   1078 
   1079 static int
   1080 bge_init_tx_ring(struct bge_softc *sc)
   1081 {
   1082 	int i;
   1083 	bus_dmamap_t dmamap;
   1084 	struct txdmamap_pool_entry *dma;
   1085 
   1086 	if (sc->bge_flags & BGE_TXRING_VALID)
   1087 		return 0;
   1088 
   1089 	sc->bge_txcnt = 0;
   1090 	sc->bge_tx_saved_considx = 0;
   1091 
   1092 	/* Initialize transmit producer index for host-memory send ring. */
   1093 	sc->bge_tx_prodidx = 0;
   1094 	CSR_WRITE_4(sc, BGE_MBX_TX_HOST_PROD0_LO, sc->bge_tx_prodidx);
   1095 	if (sc->bge_quirks & BGE_QUIRK_PRODUCER_BUG)	/* 5700 b2 errata */
   1096 		CSR_WRITE_4(sc, BGE_MBX_TX_HOST_PROD0_LO, sc->bge_tx_prodidx);
   1097 
   1098 	/* NIC-memory send ring  not used; initialize to zero. */
   1099 	CSR_WRITE_4(sc, BGE_MBX_TX_NIC_PROD0_LO, 0);
   1100 	if (sc->bge_quirks & BGE_QUIRK_PRODUCER_BUG)	/* 5700 b2 errata */
   1101 		CSR_WRITE_4(sc, BGE_MBX_TX_HOST_PROD0_LO, 0);
   1102 
   1103 	SLIST_INIT(&sc->txdma_list);
   1104 	for (i = 0; i < BGE_RSLOTS; i++) {
   1105 		if (bus_dmamap_create(sc->bge_dmatag, BGE_TXDMA_MAX,
   1106 		    BGE_NTXSEG, ETHER_MAX_LEN_JUMBO, 0, BUS_DMA_NOWAIT,
   1107 		    &dmamap))
   1108 			return(ENOBUFS);
   1109 		if (dmamap == NULL)
   1110 			panic("dmamap NULL in bge_init_tx_ring");
   1111 		dma = malloc(sizeof(*dma), M_DEVBUF, M_NOWAIT);
   1112 		if (dma == NULL) {
   1113 			printf("%s: can't alloc txdmamap_pool_entry\n",
   1114 			    sc->bge_dev.dv_xname);
   1115 			bus_dmamap_destroy(sc->bge_dmatag, dmamap);
   1116 			return (ENOMEM);
   1117 		}
   1118 		dma->dmamap = dmamap;
   1119 		SLIST_INSERT_HEAD(&sc->txdma_list, dma, link);
   1120 	}
   1121 
   1122 	sc->bge_flags |= BGE_TXRING_VALID;
   1123 
   1124 	return(0);
   1125 }
   1126 
   1127 static void
   1128 bge_setmulti(struct bge_softc *sc)
   1129 {
   1130 	struct ethercom		*ac = &sc->ethercom;
   1131 	struct ifnet		*ifp = &ac->ec_if;
   1132 	struct ether_multi	*enm;
   1133 	struct ether_multistep  step;
   1134 	u_int32_t		hashes[4] = { 0, 0, 0, 0 };
   1135 	u_int32_t		h;
   1136 	int			i;
   1137 
   1138 	if (ifp->if_flags & IFF_PROMISC)
   1139 		goto allmulti;
   1140 
   1141 	/* Now program new ones. */
   1142 	ETHER_FIRST_MULTI(step, ac, enm);
   1143 	while (enm != NULL) {
   1144 		if (memcmp(enm->enm_addrlo, enm->enm_addrhi, ETHER_ADDR_LEN)) {
   1145 			/*
   1146 			 * We must listen to a range of multicast addresses.
   1147 			 * For now, just accept all multicasts, rather than
   1148 			 * trying to set only those filter bits needed to match
   1149 			 * the range.  (At this time, the only use of address
   1150 			 * ranges is for IP multicast routing, for which the
   1151 			 * range is big enough to require all bits set.)
   1152 			 */
   1153 			goto allmulti;
   1154 		}
   1155 
   1156 		h = ether_crc32_le(enm->enm_addrlo, ETHER_ADDR_LEN);
   1157 
   1158 		/* Just want the 7 least-significant bits. */
   1159 		h &= 0x7f;
   1160 
   1161 		hashes[(h & 0x60) >> 5] |= 1 << (h & 0x1F);
   1162 		ETHER_NEXT_MULTI(step, enm);
   1163 	}
   1164 
   1165 	ifp->if_flags &= ~IFF_ALLMULTI;
   1166 	goto setit;
   1167 
   1168  allmulti:
   1169 	ifp->if_flags |= IFF_ALLMULTI;
   1170 	hashes[0] = hashes[1] = hashes[2] = hashes[3] = 0xffffffff;
   1171 
   1172  setit:
   1173 	for (i = 0; i < 4; i++)
   1174 		CSR_WRITE_4(sc, BGE_MAR0 + (i * 4), hashes[i]);
   1175 }
   1176 
   1177 const int bge_swapbits[] = {
   1178 	0,
   1179 	BGE_MODECTL_BYTESWAP_DATA,
   1180 	BGE_MODECTL_WORDSWAP_DATA,
   1181 	BGE_MODECTL_BYTESWAP_NONFRAME,
   1182 	BGE_MODECTL_WORDSWAP_NONFRAME,
   1183 
   1184 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_WORDSWAP_DATA,
   1185 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_BYTESWAP_NONFRAME,
   1186 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_WORDSWAP_NONFRAME,
   1187 
   1188 	BGE_MODECTL_WORDSWAP_DATA|BGE_MODECTL_BYTESWAP_NONFRAME,
   1189 	BGE_MODECTL_WORDSWAP_DATA|BGE_MODECTL_WORDSWAP_NONFRAME,
   1190 
   1191 	BGE_MODECTL_BYTESWAP_NONFRAME|BGE_MODECTL_WORDSWAP_NONFRAME,
   1192 
   1193 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_WORDSWAP_DATA|
   1194 	    BGE_MODECTL_BYTESWAP_NONFRAME,
   1195 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_WORDSWAP_DATA|
   1196 	    BGE_MODECTL_WORDSWAP_NONFRAME,
   1197 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_BYTESWAP_NONFRAME|
   1198 	    BGE_MODECTL_WORDSWAP_NONFRAME,
   1199 	BGE_MODECTL_WORDSWAP_DATA|BGE_MODECTL_BYTESWAP_NONFRAME|
   1200 	    BGE_MODECTL_WORDSWAP_NONFRAME,
   1201 
   1202 	BGE_MODECTL_BYTESWAP_DATA|BGE_MODECTL_WORDSWAP_DATA|
   1203 	    BGE_MODECTL_BYTESWAP_NONFRAME|BGE_MODECTL_WORDSWAP_NONFRAME,
   1204 };
   1205 
   1206 int bge_swapindex = 0;
   1207 
   1208 /*
   1209  * Do endian, PCI and DMA initialization. Also check the on-board ROM
   1210  * self-test results.
   1211  */
   1212 static int
   1213 bge_chipinit(struct bge_softc *sc)
   1214 {
   1215 	u_int32_t		cachesize;
   1216 	int			i;
   1217 	u_int32_t		dma_rw_ctl;
   1218 	struct pci_attach_args	*pa = &(sc->bge_pa);
   1219 
   1220 
   1221 	/* Set endianness before we access any non-PCI registers. */
   1222 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MISC_CTL,
   1223 	    BGE_INIT);
   1224 
   1225 	/* Set power state to D0. */
   1226 	bge_setpowerstate(sc, 0);
   1227 
   1228 	/*
   1229 	 * Check the 'ROM failed' bit on the RX CPU to see if
   1230 	 * self-tests passed.
   1231 	 */
   1232 	if (CSR_READ_4(sc, BGE_RXCPU_MODE) & BGE_RXCPUMODE_ROMFAIL) {
   1233 		printf("%s: RX CPU self-diagnostics failed!\n",
   1234 		    sc->bge_dev.dv_xname);
   1235 		return(ENODEV);
   1236 	}
   1237 
   1238 	/* Clear the MAC control register */
   1239 	CSR_WRITE_4(sc, BGE_MAC_MODE, 0);
   1240 
   1241 	/*
   1242 	 * Clear the MAC statistics block in the NIC's
   1243 	 * internal memory.
   1244 	 */
   1245 	for (i = BGE_STATS_BLOCK;
   1246 	    i < BGE_STATS_BLOCK_END + 1; i += sizeof(u_int32_t))
   1247 		BGE_MEMWIN_WRITE(pa->pa_pc, pa->pa_tag, i, 0);
   1248 
   1249 	for (i = BGE_STATUS_BLOCK;
   1250 	    i < BGE_STATUS_BLOCK_END + 1; i += sizeof(u_int32_t))
   1251 		BGE_MEMWIN_WRITE(pa->pa_pc, pa->pa_tag, i, 0);
   1252 
   1253 	/* Set up the PCI DMA control register. */
   1254 	if (sc->bge_pcie) {
   1255 	  u_int32_t device_ctl;
   1256 
   1257 		/* From FreeBSD */
   1258 		DPRINTFN(4, ("(%s: PCI-Express DMA setting)\n",
   1259 		    sc->bge_dev.dv_xname));
   1260 		dma_rw_ctl = (BGE_PCI_READ_CMD | BGE_PCI_WRITE_CMD |
   1261 		    (0xf << BGE_PCIDMARWCTL_RD_WAT_SHIFT) |
   1262 		    (0x2 << BGE_PCIDMARWCTL_WR_WAT_SHIFT));
   1263 
   1264 		/* jonathan: alternative from Linux driver */
   1265 #define DMA_CTRL_WRITE_PCIE_H20MARK_128         0x00180000
   1266 #define DMA_CTRL_WRITE_PCIE_H20MARK_256         0x00380000
   1267 
   1268 		dma_rw_ctl =   0x76000000; /* XXX XXX XXX */;
   1269 		device_ctl = pci_conf_read(pa->pa_pc, pa->pa_tag,
   1270 					   BGE_PCI_CONF_DEV_CTRL);
   1271 		printf("%s: pcie mode=0x%x\n", sc->bge_dev.dv_xname, device_ctl);
   1272 
   1273 		if ((device_ctl & 0x00e0) && 0) {
   1274 			/*
   1275 			 * XXX jonathan (at) NetBSD.org:
   1276 			 * This clause is exactly what the Broadcom-supplied
   1277 			 * Linux does; but given overall register programming
   1278 			 * by if_bge(4), this larger DMA-write watermark
   1279 			 * value causes bcm5721 chips to totally wedge.
   1280 			 */
   1281 			dma_rw_ctl |= BGE_PCIDMA_RWCTL_PCIE_WRITE_WATRMARK_256;
   1282 		} else {
   1283 			dma_rw_ctl |= BGE_PCIDMA_RWCTL_PCIE_WRITE_WATRMARK_128;
   1284 		}
   1285 	} else if (pci_conf_read(pa->pa_pc, pa->pa_tag,BGE_PCI_PCISTATE) &
   1286 	    BGE_PCISTATE_PCI_BUSMODE) {
   1287 		/* Conventional PCI bus */
   1288 	  	DPRINTFN(4, ("(%s: PCI 2.2 DMA setting)\n", sc->bge_dev.dv_xname));
   1289 		dma_rw_ctl = (BGE_PCI_READ_CMD | BGE_PCI_WRITE_CMD |
   1290 		   (0x7 << BGE_PCIDMARWCTL_RD_WAT_SHIFT) |
   1291 		   (0x7 << BGE_PCIDMARWCTL_WR_WAT_SHIFT));
   1292 		if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1293 			dma_rw_ctl |= 0x0F;
   1294 		}
   1295 	} else {
   1296 	  	DPRINTFN(4, ("(:%s: PCI-X DMA setting)\n", sc->bge_dev.dv_xname));
   1297 		/* PCI-X bus */
   1298 		dma_rw_ctl = BGE_PCI_READ_CMD|BGE_PCI_WRITE_CMD |
   1299 		    (0x3 << BGE_PCIDMARWCTL_RD_WAT_SHIFT) |
   1300 		    (0x3 << BGE_PCIDMARWCTL_WR_WAT_SHIFT) |
   1301 		    (0x0F);
   1302 		/*
   1303 		 * 5703 and 5704 need ONEDMA_AT_ONCE as a workaround
   1304 		 * for hardware bugs, which means we should also clear
   1305 		 * the low-order MINDMA bits.  In addition, the 5704
   1306 		 * uses a different encoding of read/write watermarks.
   1307 		 */
   1308 		if (BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5704) {
   1309 			dma_rw_ctl = BGE_PCI_READ_CMD|BGE_PCI_WRITE_CMD |
   1310 			  /* should be 0x1f0000 */
   1311 			  (0x7 << BGE_PCIDMARWCTL_RD_WAT_SHIFT) |
   1312 			  (0x3 << BGE_PCIDMARWCTL_WR_WAT_SHIFT);
   1313 			dma_rw_ctl |= BGE_PCIDMARWCTL_ONEDMA_ATONCE;
   1314 		}
   1315 		else if (BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5703) {
   1316 			dma_rw_ctl &=  0xfffffff0;
   1317 			dma_rw_ctl |= BGE_PCIDMARWCTL_ONEDMA_ATONCE;
   1318 		}
   1319 		else if (BGE_IS_5714_FAMILY(sc)) {
   1320 			dma_rw_ctl = BGE_PCI_READ_CMD|BGE_PCI_WRITE_CMD;
   1321 			dma_rw_ctl &= ~BGE_PCIDMARWCTL_ONEDMA_ATONCE; /* XXX */
   1322 			/* XXX magic values, Broadcom-supplied Linux driver */
   1323 			if (BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5780)
   1324 				dma_rw_ctl |= (1 << 20) | (1 << 18) |
   1325 				  BGE_PCIDMARWCTL_ONEDMA_ATONCE;
   1326 			else
   1327 				dma_rw_ctl |= (1<<20) | (1<<18) | (1 << 15);
   1328 		}
   1329 	}
   1330 
   1331 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL, dma_rw_ctl);
   1332 
   1333 	/*
   1334 	 * Set up general mode register.
   1335 	 */
   1336 	CSR_WRITE_4(sc, BGE_MODE_CTL, BGE_DMA_SWAP_OPTIONS|
   1337 		    BGE_MODECTL_MAC_ATTN_INTR|BGE_MODECTL_HOST_SEND_BDS|
   1338 		    BGE_MODECTL_TX_NO_PHDR_CSUM|BGE_MODECTL_RX_NO_PHDR_CSUM);
   1339 
   1340 	/* Get cache line size. */
   1341 	cachesize = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_CACHESZ);
   1342 
   1343 	/*
   1344 	 * Avoid violating PCI spec on certain chip revs.
   1345 	 */
   1346 	if (pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_CMD) &
   1347 	    PCIM_CMD_MWIEN) {
   1348 		switch(cachesize) {
   1349 		case 1:
   1350 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1351 				   BGE_PCI_WRITE_BNDRY_16BYTES);
   1352 			break;
   1353 		case 2:
   1354 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1355 				   BGE_PCI_WRITE_BNDRY_32BYTES);
   1356 			break;
   1357 		case 4:
   1358 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1359 				   BGE_PCI_WRITE_BNDRY_64BYTES);
   1360 			break;
   1361 		case 8:
   1362 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1363 				   BGE_PCI_WRITE_BNDRY_128BYTES);
   1364 			break;
   1365 		case 16:
   1366 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1367 				   BGE_PCI_WRITE_BNDRY_256BYTES);
   1368 			break;
   1369 		case 32:
   1370 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1371 				   BGE_PCI_WRITE_BNDRY_512BYTES);
   1372 			break;
   1373 		case 64:
   1374 			PCI_SETBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_DMA_RW_CTL,
   1375 				   BGE_PCI_WRITE_BNDRY_1024BYTES);
   1376 			break;
   1377 		default:
   1378 		/* Disable PCI memory write and invalidate. */
   1379 #if 0
   1380 			if (bootverbose)
   1381 				printf("%s: cache line size %d not "
   1382 				    "supported; disabling PCI MWI\n",
   1383 				    sc->bge_dev.dv_xname, cachesize);
   1384 #endif
   1385 			PCI_CLRBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_CMD,
   1386 			    PCIM_CMD_MWIEN);
   1387 			break;
   1388 		}
   1389 	}
   1390 
   1391 	/*
   1392 	 * Disable memory write invalidate.  Apparently it is not supported
   1393 	 * properly by these devices.
   1394 	 */
   1395 	PCI_CLRBIT(pa->pa_pc, pa->pa_tag, BGE_PCI_CMD, PCIM_CMD_MWIEN);
   1396 
   1397 
   1398 #ifdef __brokenalpha__
   1399 	/*
   1400 	 * Must insure that we do not cross an 8K (bytes) boundary
   1401 	 * for DMA reads.  Our highest limit is 1K bytes.  This is a
   1402 	 * restriction on some ALPHA platforms with early revision
   1403 	 * 21174 PCI chipsets, such as the AlphaPC 164lx
   1404 	 */
   1405 	PCI_SETBIT(sc, BGE_PCI_DMA_RW_CTL, BGE_PCI_READ_BNDRY_1024, 4);
   1406 #endif
   1407 
   1408 	/* Set the timer prescaler (always 66MHz) */
   1409 	CSR_WRITE_4(sc, BGE_MISC_CFG, 65 << 1/*BGE_32BITTIME_66MHZ*/);
   1410 
   1411 	return(0);
   1412 }
   1413 
   1414 static int
   1415 bge_blockinit(struct bge_softc *sc)
   1416 {
   1417 	volatile struct bge_rcb		*rcb;
   1418 	bus_size_t		rcb_addr;
   1419 	int			i;
   1420 	struct ifnet		*ifp = &sc->ethercom.ec_if;
   1421 	bge_hostaddr		taddr;
   1422 
   1423 	/*
   1424 	 * Initialize the memory window pointer register so that
   1425 	 * we can access the first 32K of internal NIC RAM. This will
   1426 	 * allow us to set up the TX send ring RCBs and the RX return
   1427 	 * ring RCBs, plus other things which live in NIC memory.
   1428 	 */
   1429 
   1430 	pci_conf_write(sc->bge_pa.pa_pc, sc->bge_pa.pa_tag,
   1431 	    BGE_PCI_MEMWIN_BASEADDR, 0);
   1432 
   1433 	/* Configure mbuf memory pool */
   1434 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1435 		if (sc->bge_extram) {
   1436 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_BASEADDR,
   1437 			    BGE_EXT_SSRAM);
   1438 			if ((sc->bge_quirks & BGE_QUIRK_FEWER_MBUFS) != 0)
   1439 				CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_LEN, 0x10000);
   1440 			else
   1441 				CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_LEN, 0x18000);
   1442 		} else {
   1443 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_BASEADDR,
   1444 			    BGE_BUFFPOOL_1);
   1445 			if ((sc->bge_quirks & BGE_QUIRK_FEWER_MBUFS) != 0)
   1446 				CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_LEN, 0x10000);
   1447 			else
   1448 				CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_LEN, 0x18000);
   1449 		}
   1450 
   1451 		/* Configure DMA resource pool */
   1452 		CSR_WRITE_4(sc, BGE_BMAN_DMA_DESCPOOL_BASEADDR,
   1453 		    BGE_DMA_DESCRIPTORS);
   1454 		CSR_WRITE_4(sc, BGE_BMAN_DMA_DESCPOOL_LEN, 0x2000);
   1455 	}
   1456 
   1457 	/* Configure mbuf pool watermarks */
   1458 #ifdef ORIG_WPAUL_VALUES
   1459 	CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_READDMA_LOWAT, 24);
   1460 	CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_MACRX_LOWAT, 24);
   1461 	CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_HIWAT, 48);
   1462 #else
   1463 	/* new broadcom docs strongly recommend these: */
   1464 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1465 		if (ifp->if_mtu > ETHER_MAX_LEN) {
   1466 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_READDMA_LOWAT, 0x50);
   1467 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_MACRX_LOWAT, 0x20);
   1468 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_HIWAT, 0x60);
   1469 		} else {
   1470 			/* Values from Linux driver... */
   1471 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_READDMA_LOWAT, 304);
   1472 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_MACRX_LOWAT, 152);
   1473 			CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_HIWAT, 380);
   1474 		}
   1475 	} else {
   1476 		CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_READDMA_LOWAT, 0x0);
   1477 		CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_MACRX_LOWAT, 0x10);
   1478 		CSR_WRITE_4(sc, BGE_BMAN_MBUFPOOL_HIWAT, 0x60);
   1479 	}
   1480 #endif
   1481 
   1482 	/* Configure DMA resource watermarks */
   1483 	CSR_WRITE_4(sc, BGE_BMAN_DMA_DESCPOOL_LOWAT, 5);
   1484 	CSR_WRITE_4(sc, BGE_BMAN_DMA_DESCPOOL_HIWAT, 10);
   1485 
   1486 	/* Enable buffer manager */
   1487 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1488 		CSR_WRITE_4(sc, BGE_BMAN_MODE,
   1489 		    BGE_BMANMODE_ENABLE|BGE_BMANMODE_LOMBUF_ATTN);
   1490 
   1491 		/* Poll for buffer manager start indication */
   1492 		for (i = 0; i < BGE_TIMEOUT; i++) {
   1493 			if (CSR_READ_4(sc, BGE_BMAN_MODE) & BGE_BMANMODE_ENABLE)
   1494 				break;
   1495 			DELAY(10);
   1496 		}
   1497 
   1498 		if (i == BGE_TIMEOUT) {
   1499 			printf("%s: buffer manager failed to start\n",
   1500 			    sc->bge_dev.dv_xname);
   1501 			return(ENXIO);
   1502 		}
   1503 	}
   1504 
   1505 	/* Enable flow-through queues */
   1506 	CSR_WRITE_4(sc, BGE_FTQ_RESET, 0xFFFFFFFF);
   1507 	CSR_WRITE_4(sc, BGE_FTQ_RESET, 0);
   1508 
   1509 	/* Wait until queue initialization is complete */
   1510 	for (i = 0; i < BGE_TIMEOUT; i++) {
   1511 		if (CSR_READ_4(sc, BGE_FTQ_RESET) == 0)
   1512 			break;
   1513 		DELAY(10);
   1514 	}
   1515 
   1516 	if (i == BGE_TIMEOUT) {
   1517 		printf("%s: flow-through queue init failed\n",
   1518 		    sc->bge_dev.dv_xname);
   1519 		return(ENXIO);
   1520 	}
   1521 
   1522 	/* Initialize the standard RX ring control block */
   1523 	rcb = &sc->bge_rdata->bge_info.bge_std_rx_rcb;
   1524 	bge_set_hostaddr(&rcb->bge_hostaddr,
   1525 	    BGE_RING_DMA_ADDR(sc, bge_rx_std_ring));
   1526 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1527 		rcb->bge_maxlen_flags =
   1528 		    BGE_RCB_MAXLEN_FLAGS(BGE_MAX_FRAMELEN, 0);
   1529 	} else {
   1530 		rcb->bge_maxlen_flags = BGE_RCB_MAXLEN_FLAGS(512, 0);
   1531 	}
   1532 	if (sc->bge_extram)
   1533 		rcb->bge_nicaddr = BGE_EXT_STD_RX_RINGS;
   1534 	else
   1535 		rcb->bge_nicaddr = BGE_STD_RX_RINGS;
   1536 	CSR_WRITE_4(sc, BGE_RX_STD_RCB_HADDR_HI, rcb->bge_hostaddr.bge_addr_hi);
   1537 	CSR_WRITE_4(sc, BGE_RX_STD_RCB_HADDR_LO, rcb->bge_hostaddr.bge_addr_lo);
   1538 	CSR_WRITE_4(sc, BGE_RX_STD_RCB_MAXLEN_FLAGS, rcb->bge_maxlen_flags);
   1539 	CSR_WRITE_4(sc, BGE_RX_STD_RCB_NICADDR, rcb->bge_nicaddr);
   1540 
   1541 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1542 		sc->bge_return_ring_cnt = BGE_RETURN_RING_CNT;
   1543 	} else {
   1544 		sc->bge_return_ring_cnt = BGE_RETURN_RING_CNT_5705;
   1545 	}
   1546 
   1547 	/*
   1548 	 * Initialize the jumbo RX ring control block
   1549 	 * We set the 'ring disabled' bit in the flags
   1550 	 * field until we're actually ready to start
   1551 	 * using this ring (i.e. once we set the MTU
   1552 	 * high enough to require it).
   1553 	 */
   1554 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1555 		rcb = &sc->bge_rdata->bge_info.bge_jumbo_rx_rcb;
   1556 		bge_set_hostaddr(&rcb->bge_hostaddr,
   1557 		    BGE_RING_DMA_ADDR(sc, bge_rx_jumbo_ring));
   1558 		rcb->bge_maxlen_flags =
   1559 		    BGE_RCB_MAXLEN_FLAGS(BGE_MAX_FRAMELEN,
   1560 			BGE_RCB_FLAG_RING_DISABLED);
   1561 		if (sc->bge_extram)
   1562 			rcb->bge_nicaddr = BGE_EXT_JUMBO_RX_RINGS;
   1563 		else
   1564 			rcb->bge_nicaddr = BGE_JUMBO_RX_RINGS;
   1565 
   1566 		CSR_WRITE_4(sc, BGE_RX_JUMBO_RCB_HADDR_HI,
   1567 		    rcb->bge_hostaddr.bge_addr_hi);
   1568 		CSR_WRITE_4(sc, BGE_RX_JUMBO_RCB_HADDR_LO,
   1569 		    rcb->bge_hostaddr.bge_addr_lo);
   1570 		CSR_WRITE_4(sc, BGE_RX_JUMBO_RCB_MAXLEN_FLAGS,
   1571 		    rcb->bge_maxlen_flags);
   1572 		CSR_WRITE_4(sc, BGE_RX_JUMBO_RCB_NICADDR, rcb->bge_nicaddr);
   1573 
   1574 		/* Set up dummy disabled mini ring RCB */
   1575 		rcb = &sc->bge_rdata->bge_info.bge_mini_rx_rcb;
   1576 		rcb->bge_maxlen_flags = BGE_RCB_MAXLEN_FLAGS(0,
   1577 		    BGE_RCB_FLAG_RING_DISABLED);
   1578 		CSR_WRITE_4(sc, BGE_RX_MINI_RCB_MAXLEN_FLAGS,
   1579 		    rcb->bge_maxlen_flags);
   1580 
   1581 		bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   1582 		    offsetof(struct bge_ring_data, bge_info),
   1583 		    sizeof (struct bge_gib),
   1584 		    BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
   1585 	}
   1586 
   1587 	/*
   1588 	 * Set the BD ring replentish thresholds. The recommended
   1589 	 * values are 1/8th the number of descriptors allocated to
   1590 	 * each ring.
   1591 	 */
   1592 	CSR_WRITE_4(sc, BGE_RBDI_STD_REPL_THRESH, BGE_STD_RX_RING_CNT/8);
   1593 	CSR_WRITE_4(sc, BGE_RBDI_JUMBO_REPL_THRESH, BGE_JUMBO_RX_RING_CNT/8);
   1594 
   1595 	/*
   1596 	 * Disable all unused send rings by setting the 'ring disabled'
   1597 	 * bit in the flags field of all the TX send ring control blocks.
   1598 	 * These are located in NIC memory.
   1599 	 */
   1600 	rcb_addr = BGE_MEMWIN_START + BGE_SEND_RING_RCB;
   1601 	for (i = 0; i < BGE_TX_RINGS_EXTSSRAM_MAX; i++) {
   1602 		RCB_WRITE_4(sc, rcb_addr, bge_maxlen_flags,
   1603 		    BGE_RCB_MAXLEN_FLAGS(0,BGE_RCB_FLAG_RING_DISABLED));
   1604 		RCB_WRITE_4(sc, rcb_addr, bge_nicaddr, 0);
   1605 		rcb_addr += sizeof(struct bge_rcb);
   1606 	}
   1607 
   1608 	/* Configure TX RCB 0 (we use only the first ring) */
   1609 	rcb_addr = BGE_MEMWIN_START + BGE_SEND_RING_RCB;
   1610 	bge_set_hostaddr(&taddr, BGE_RING_DMA_ADDR(sc, bge_tx_ring));
   1611 	RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_hi, taddr.bge_addr_hi);
   1612 	RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_lo, taddr.bge_addr_lo);
   1613 	RCB_WRITE_4(sc, rcb_addr, bge_nicaddr,
   1614 		    BGE_NIC_TXRING_ADDR(0, BGE_TX_RING_CNT));
   1615 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1616 		RCB_WRITE_4(sc, rcb_addr, bge_maxlen_flags,
   1617 		    BGE_RCB_MAXLEN_FLAGS(BGE_TX_RING_CNT, 0));
   1618 	}
   1619 
   1620 	/* Disable all unused RX return rings */
   1621 	rcb_addr = BGE_MEMWIN_START + BGE_RX_RETURN_RING_RCB;
   1622 	for (i = 0; i < BGE_RX_RINGS_MAX; i++) {
   1623 		RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_hi, 0);
   1624 		RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_lo, 0);
   1625 		RCB_WRITE_4(sc, rcb_addr, bge_maxlen_flags,
   1626 			    BGE_RCB_MAXLEN_FLAGS(sc->bge_return_ring_cnt,
   1627                                      BGE_RCB_FLAG_RING_DISABLED));
   1628 		RCB_WRITE_4(sc, rcb_addr, bge_nicaddr, 0);
   1629 		CSR_WRITE_4(sc, BGE_MBX_RX_CONS0_LO +
   1630 		    (i * (sizeof(u_int64_t))), 0);
   1631 		rcb_addr += sizeof(struct bge_rcb);
   1632 	}
   1633 
   1634 	/* Initialize RX ring indexes */
   1635 	CSR_WRITE_4(sc, BGE_MBX_RX_STD_PROD_LO, 0);
   1636 	CSR_WRITE_4(sc, BGE_MBX_RX_JUMBO_PROD_LO, 0);
   1637 	CSR_WRITE_4(sc, BGE_MBX_RX_MINI_PROD_LO, 0);
   1638 
   1639 	/*
   1640 	 * Set up RX return ring 0
   1641 	 * Note that the NIC address for RX return rings is 0x00000000.
   1642 	 * The return rings live entirely within the host, so the
   1643 	 * nicaddr field in the RCB isn't used.
   1644 	 */
   1645 	rcb_addr = BGE_MEMWIN_START + BGE_RX_RETURN_RING_RCB;
   1646 	bge_set_hostaddr(&taddr, BGE_RING_DMA_ADDR(sc, bge_rx_return_ring));
   1647 	RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_hi, taddr.bge_addr_hi);
   1648 	RCB_WRITE_4(sc, rcb_addr, bge_hostaddr.bge_addr_lo, taddr.bge_addr_lo);
   1649 	RCB_WRITE_4(sc, rcb_addr, bge_nicaddr, 0x00000000);
   1650 	RCB_WRITE_4(sc, rcb_addr, bge_maxlen_flags,
   1651 	    BGE_RCB_MAXLEN_FLAGS(sc->bge_return_ring_cnt, 0));
   1652 
   1653 	/* Set random backoff seed for TX */
   1654 	CSR_WRITE_4(sc, BGE_TX_RANDOM_BACKOFF,
   1655 	    LLADDR(ifp->if_sadl)[0] + LLADDR(ifp->if_sadl)[1] +
   1656 	    LLADDR(ifp->if_sadl)[2] + LLADDR(ifp->if_sadl)[3] +
   1657 	    LLADDR(ifp->if_sadl)[4] + LLADDR(ifp->if_sadl)[5] +
   1658 	    BGE_TX_BACKOFF_SEED_MASK);
   1659 
   1660 	/* Set inter-packet gap */
   1661 	CSR_WRITE_4(sc, BGE_TX_LENGTHS, 0x2620);
   1662 
   1663 	/*
   1664 	 * Specify which ring to use for packets that don't match
   1665 	 * any RX rules.
   1666 	 */
   1667 	CSR_WRITE_4(sc, BGE_RX_RULES_CFG, 0x08);
   1668 
   1669 	/*
   1670 	 * Configure number of RX lists. One interrupt distribution
   1671 	 * list, sixteen active lists, one bad frames class.
   1672 	 */
   1673 	CSR_WRITE_4(sc, BGE_RXLP_CFG, 0x181);
   1674 
   1675 	/* Inialize RX list placement stats mask. */
   1676 	CSR_WRITE_4(sc, BGE_RXLP_STATS_ENABLE_MASK, 0x007FFFFF);
   1677 	CSR_WRITE_4(sc, BGE_RXLP_STATS_CTL, 0x1);
   1678 
   1679 	/* Disable host coalescing until we get it set up */
   1680 	CSR_WRITE_4(sc, BGE_HCC_MODE, 0x00000000);
   1681 
   1682 	/* Poll to make sure it's shut down. */
   1683 	for (i = 0; i < BGE_TIMEOUT; i++) {
   1684 		if (!(CSR_READ_4(sc, BGE_HCC_MODE) & BGE_HCCMODE_ENABLE))
   1685 			break;
   1686 		DELAY(10);
   1687 	}
   1688 
   1689 	if (i == BGE_TIMEOUT) {
   1690 		printf("%s: host coalescing engine failed to idle\n",
   1691 		    sc->bge_dev.dv_xname);
   1692 		return(ENXIO);
   1693 	}
   1694 
   1695 	/* Set up host coalescing defaults */
   1696 	CSR_WRITE_4(sc, BGE_HCC_RX_COAL_TICKS, sc->bge_rx_coal_ticks);
   1697 	CSR_WRITE_4(sc, BGE_HCC_TX_COAL_TICKS, sc->bge_tx_coal_ticks);
   1698 	CSR_WRITE_4(sc, BGE_HCC_RX_MAX_COAL_BDS, sc->bge_rx_max_coal_bds);
   1699 	CSR_WRITE_4(sc, BGE_HCC_TX_MAX_COAL_BDS, sc->bge_tx_max_coal_bds);
   1700 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1701 		CSR_WRITE_4(sc, BGE_HCC_RX_COAL_TICKS_INT, 0);
   1702 		CSR_WRITE_4(sc, BGE_HCC_TX_COAL_TICKS_INT, 0);
   1703 	}
   1704 	CSR_WRITE_4(sc, BGE_HCC_RX_MAX_COAL_BDS_INT, 0);
   1705 	CSR_WRITE_4(sc, BGE_HCC_TX_MAX_COAL_BDS_INT, 0);
   1706 
   1707 	/* Set up address of statistics block */
   1708 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1709 		bge_set_hostaddr(&taddr,
   1710 		    BGE_RING_DMA_ADDR(sc, bge_info.bge_stats));
   1711 		CSR_WRITE_4(sc, BGE_HCC_STATS_TICKS, sc->bge_stat_ticks);
   1712 		CSR_WRITE_4(sc, BGE_HCC_STATS_BASEADDR, BGE_STATS_BLOCK);
   1713 		CSR_WRITE_4(sc, BGE_HCC_STATS_ADDR_HI, taddr.bge_addr_hi);
   1714 		CSR_WRITE_4(sc, BGE_HCC_STATS_ADDR_LO, taddr.bge_addr_lo);
   1715 	}
   1716 
   1717 	/* Set up address of status block */
   1718 	bge_set_hostaddr(&taddr, BGE_RING_DMA_ADDR(sc, bge_status_block));
   1719 	CSR_WRITE_4(sc, BGE_HCC_STATUSBLK_BASEADDR, BGE_STATUS_BLOCK);
   1720 	CSR_WRITE_4(sc, BGE_HCC_STATUSBLK_ADDR_HI, taddr.bge_addr_hi);
   1721 	CSR_WRITE_4(sc, BGE_HCC_STATUSBLK_ADDR_LO, taddr.bge_addr_lo);
   1722 	sc->bge_rdata->bge_status_block.bge_idx[0].bge_rx_prod_idx = 0;
   1723 	sc->bge_rdata->bge_status_block.bge_idx[0].bge_tx_cons_idx = 0;
   1724 
   1725 	/* Turn on host coalescing state machine */
   1726 	CSR_WRITE_4(sc, BGE_HCC_MODE, BGE_HCCMODE_ENABLE);
   1727 
   1728 	/* Turn on RX BD completion state machine and enable attentions */
   1729 	CSR_WRITE_4(sc, BGE_RBDC_MODE,
   1730 	    BGE_RBDCMODE_ENABLE|BGE_RBDCMODE_ATTN);
   1731 
   1732 	/* Turn on RX list placement state machine */
   1733 	CSR_WRITE_4(sc, BGE_RXLP_MODE, BGE_RXLPMODE_ENABLE);
   1734 
   1735 	/* Turn on RX list selector state machine. */
   1736 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1737 		CSR_WRITE_4(sc, BGE_RXLS_MODE, BGE_RXLSMODE_ENABLE);
   1738 	}
   1739 
   1740 	/* Turn on DMA, clear stats */
   1741 	CSR_WRITE_4(sc, BGE_MAC_MODE, BGE_MACMODE_TXDMA_ENB|
   1742 	    BGE_MACMODE_RXDMA_ENB|BGE_MACMODE_RX_STATS_CLEAR|
   1743 	    BGE_MACMODE_TX_STATS_CLEAR|BGE_MACMODE_RX_STATS_ENB|
   1744 	    BGE_MACMODE_TX_STATS_ENB|BGE_MACMODE_FRMHDR_DMA_ENB|
   1745 	    (sc->bge_tbi ? BGE_PORTMODE_TBI : BGE_PORTMODE_MII));
   1746 
   1747 	/* Set misc. local control, enable interrupts on attentions */
   1748 	sc->bge_local_ctrl_reg = BGE_MLC_INTR_ONATTN | BGE_MLC_AUTO_EEPROM;
   1749 
   1750 #ifdef notdef
   1751 	/* Assert GPIO pins for PHY reset */
   1752 	BGE_SETBIT(sc, BGE_MISC_LOCAL_CTL, BGE_MLC_MISCIO_OUT0|
   1753 	    BGE_MLC_MISCIO_OUT1|BGE_MLC_MISCIO_OUT2);
   1754 	BGE_SETBIT(sc, BGE_MISC_LOCAL_CTL, BGE_MLC_MISCIO_OUTEN0|
   1755 	    BGE_MLC_MISCIO_OUTEN1|BGE_MLC_MISCIO_OUTEN2);
   1756 #endif
   1757 
   1758 #if defined(not_quite_yet)
   1759 	/* Linux driver enables enable gpio pin #1 on 5700s */
   1760 	if (sc->bge_chipid == BGE_CHIPID_BCM5700) {
   1761 		sc->bge_local_ctrl_reg |=
   1762 		  (BGE_MLC_MISCIO_OUT1|BGE_MLC_MISCIO_OUTEN1);
   1763 	}
   1764 #endif
   1765 	CSR_WRITE_4(sc, BGE_MISC_LOCAL_CTL, sc->bge_local_ctrl_reg);
   1766 
   1767 	/* Turn on DMA completion state machine */
   1768 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1769 		CSR_WRITE_4(sc, BGE_DMAC_MODE, BGE_DMACMODE_ENABLE);
   1770 	}
   1771 
   1772 	/* Turn on write DMA state machine */
   1773 	CSR_WRITE_4(sc, BGE_WDMA_MODE,
   1774 	    BGE_WDMAMODE_ENABLE|BGE_WDMAMODE_ALL_ATTNS);
   1775 
   1776 	/* Turn on read DMA state machine */
   1777 	{
   1778 		uint32_t dma_read_modebits;
   1779 
   1780 		dma_read_modebits =
   1781 		  BGE_RDMAMODE_ENABLE | BGE_RDMAMODE_ALL_ATTNS;
   1782 
   1783 		if (sc->bge_pcie && 0) {
   1784 			dma_read_modebits |= BGE_RDMA_MODE_FIFO_LONG_BURST;
   1785 		} else if ((sc->bge_quirks & BGE_QUIRK_5705_CORE)) {
   1786 			dma_read_modebits |= BGE_RDMA_MODE_FIFO_SIZE_128;
   1787 		}
   1788 
   1789 		/* XXX broadcom-supplied linux driver; undocumented */
   1790 		if (BGE_IS_5750_OR_BEYOND(sc)) {
   1791  			/*
   1792 			 * XXX: magic values.
   1793 			 * From Broadcom-supplied Linux driver;  apparently
   1794 			 * required to workaround a DMA bug affecting TSO
   1795 			 * on bcm575x/bcm5721?
   1796 			 */
   1797 			dma_read_modebits |= (1 << 27);
   1798 		}
   1799 		CSR_WRITE_4(sc, BGE_RDMA_MODE, dma_read_modebits);
   1800 	}
   1801 
   1802 	/* Turn on RX data completion state machine */
   1803 	CSR_WRITE_4(sc, BGE_RDC_MODE, BGE_RDCMODE_ENABLE);
   1804 
   1805 	/* Turn on RX BD initiator state machine */
   1806 	CSR_WRITE_4(sc, BGE_RBDI_MODE, BGE_RBDIMODE_ENABLE);
   1807 
   1808 	/* Turn on RX data and RX BD initiator state machine */
   1809 	CSR_WRITE_4(sc, BGE_RDBDI_MODE, BGE_RDBDIMODE_ENABLE);
   1810 
   1811 	/* Turn on Mbuf cluster free state machine */
   1812 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   1813 		CSR_WRITE_4(sc, BGE_MBCF_MODE, BGE_MBCFMODE_ENABLE);
   1814 	}
   1815 
   1816 	/* Turn on send BD completion state machine */
   1817 	CSR_WRITE_4(sc, BGE_SBDC_MODE, BGE_SBDCMODE_ENABLE);
   1818 
   1819 	/* Turn on send data completion state machine */
   1820 	CSR_WRITE_4(sc, BGE_SDC_MODE, BGE_SDCMODE_ENABLE);
   1821 
   1822 	/* Turn on send data initiator state machine */
   1823 	if (BGE_IS_5750_OR_BEYOND(sc)) {
   1824 		/* XXX: magic value from Linux driver */
   1825 		CSR_WRITE_4(sc, BGE_SDI_MODE, BGE_SDIMODE_ENABLE | 0x08);
   1826 	} else {
   1827 		CSR_WRITE_4(sc, BGE_SDI_MODE, BGE_SDIMODE_ENABLE);
   1828 	}
   1829 
   1830 	/* Turn on send BD initiator state machine */
   1831 	CSR_WRITE_4(sc, BGE_SBDI_MODE, BGE_SBDIMODE_ENABLE);
   1832 
   1833 	/* Turn on send BD selector state machine */
   1834 	CSR_WRITE_4(sc, BGE_SRS_MODE, BGE_SRSMODE_ENABLE);
   1835 
   1836 	CSR_WRITE_4(sc, BGE_SDI_STATS_ENABLE_MASK, 0x007FFFFF);
   1837 	CSR_WRITE_4(sc, BGE_SDI_STATS_CTL,
   1838 	    BGE_SDISTATSCTL_ENABLE|BGE_SDISTATSCTL_FASTER);
   1839 
   1840 	/* ack/clear link change events */
   1841 	CSR_WRITE_4(sc, BGE_MAC_STS, BGE_MACSTAT_SYNC_CHANGED|
   1842 	    BGE_MACSTAT_CFG_CHANGED);
   1843 	CSR_WRITE_4(sc, BGE_MI_STS, 0);
   1844 
   1845 	/* Enable PHY auto polling (for MII/GMII only) */
   1846 	if (sc->bge_tbi) {
   1847 		CSR_WRITE_4(sc, BGE_MI_STS, BGE_MISTS_LINK);
   1848  	} else {
   1849 		BGE_SETBIT(sc, BGE_MI_MODE, BGE_MIMODE_AUTOPOLL|10<<16);
   1850 		if (sc->bge_quirks & BGE_QUIRK_LINK_STATE_BROKEN)
   1851 			CSR_WRITE_4(sc, BGE_MAC_EVT_ENB,
   1852 			    BGE_EVTENB_MI_INTERRUPT);
   1853 	}
   1854 
   1855 	/* Enable link state change attentions. */
   1856 	BGE_SETBIT(sc, BGE_MAC_EVT_ENB, BGE_EVTENB_LINK_CHANGED);
   1857 
   1858 	return(0);
   1859 }
   1860 
   1861 static const struct bge_revision {
   1862 	uint32_t		br_chipid;
   1863 	uint32_t		br_quirks;
   1864 	const char		*br_name;
   1865 } bge_revisions[] = {
   1866 	{ BGE_CHIPID_BCM5700_A0,
   1867 	  BGE_QUIRK_LINK_STATE_BROKEN,
   1868 	  "BCM5700 A0" },
   1869 
   1870 	{ BGE_CHIPID_BCM5700_A1,
   1871 	  BGE_QUIRK_LINK_STATE_BROKEN,
   1872 	  "BCM5700 A1" },
   1873 
   1874 	{ BGE_CHIPID_BCM5700_B0,
   1875 	  BGE_QUIRK_LINK_STATE_BROKEN|BGE_QUIRK_CSUM_BROKEN|BGE_QUIRK_5700_COMMON,
   1876 	  "BCM5700 B0" },
   1877 
   1878 	{ BGE_CHIPID_BCM5700_B1,
   1879 	  BGE_QUIRK_LINK_STATE_BROKEN|BGE_QUIRK_5700_COMMON,
   1880 	  "BCM5700 B1" },
   1881 
   1882 	{ BGE_CHIPID_BCM5700_B2,
   1883 	  BGE_QUIRK_LINK_STATE_BROKEN|BGE_QUIRK_5700_COMMON,
   1884 	  "BCM5700 B2" },
   1885 
   1886 	/* This is treated like a BCM5700 Bx */
   1887 	{ BGE_CHIPID_BCM5700_ALTIMA,
   1888 	  BGE_QUIRK_LINK_STATE_BROKEN|BGE_QUIRK_5700_COMMON,
   1889 	  "BCM5700 Altima" },
   1890 
   1891 	{ BGE_CHIPID_BCM5700_C0,
   1892 	  0,
   1893 	  "BCM5700 C0" },
   1894 
   1895 	{ BGE_CHIPID_BCM5701_A0,
   1896 	  0, /*XXX really, just not known */
   1897 	  "BCM5701 A0" },
   1898 
   1899 	{ BGE_CHIPID_BCM5701_B0,
   1900 	  BGE_QUIRK_PCIX_DMA_ALIGN_BUG,
   1901 	  "BCM5701 B0" },
   1902 
   1903 	{ BGE_CHIPID_BCM5701_B2,
   1904 	  BGE_QUIRK_PCIX_DMA_ALIGN_BUG,
   1905 	  "BCM5701 B2" },
   1906 
   1907 	{ BGE_CHIPID_BCM5701_B5,
   1908 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_PCIX_DMA_ALIGN_BUG,
   1909 	  "BCM5701 B5" },
   1910 
   1911 	{ BGE_CHIPID_BCM5703_A0,
   1912 	  0,
   1913 	  "BCM5703 A0" },
   1914 
   1915 	{ BGE_CHIPID_BCM5703_A1,
   1916 	  0,
   1917 	  "BCM5703 A1" },
   1918 
   1919 	{ BGE_CHIPID_BCM5703_A2,
   1920 	  BGE_QUIRK_ONLY_PHY_1,
   1921 	  "BCM5703 A2" },
   1922 
   1923 	{ BGE_CHIPID_BCM5703_A3,
   1924 	  BGE_QUIRK_ONLY_PHY_1,
   1925 	  "BCM5703 A3" },
   1926 
   1927 	{ BGE_CHIPID_BCM5704_A0,
   1928   	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_FEWER_MBUFS,
   1929 	  "BCM5704 A0" },
   1930 
   1931 	{ BGE_CHIPID_BCM5704_A1,
   1932   	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_FEWER_MBUFS,
   1933 	  "BCM5704 A1" },
   1934 
   1935 	{ BGE_CHIPID_BCM5704_A2,
   1936   	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_FEWER_MBUFS,
   1937 	  "BCM5704 A2" },
   1938 
   1939 	{ BGE_CHIPID_BCM5704_A3,
   1940   	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_FEWER_MBUFS,
   1941 	  "BCM5704 A3" },
   1942 
   1943 	{ BGE_CHIPID_BCM5705_A0,
   1944 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1945 	  "BCM5705 A0" },
   1946 
   1947 	{ BGE_CHIPID_BCM5705_A1,
   1948 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1949 	  "BCM5705 A1" },
   1950 
   1951 	{ BGE_CHIPID_BCM5705_A2,
   1952 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1953 	  "BCM5705 A2" },
   1954 
   1955 	{ BGE_CHIPID_BCM5705_A3,
   1956 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1957 	  "BCM5705 A3" },
   1958 
   1959 	{ BGE_CHIPID_BCM5750_A0,
   1960 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1961 	  "BCM5750 A1" },
   1962 
   1963 	{ BGE_CHIPID_BCM5750_A1,
   1964 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1965 	  "BCM5750 A1" },
   1966 
   1967 	{ BGE_CHIPID_BCM5751_A1,
   1968 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1969 	  "BCM5751 A1" },
   1970 
   1971 	{ 0, 0, NULL }
   1972 };
   1973 
   1974 /*
   1975  * Some defaults for major revisions, so that newer steppings
   1976  * that we don't know about have a shot at working.
   1977  */
   1978 static const struct bge_revision bge_majorrevs[] = {
   1979 	{ BGE_ASICREV_BCM5700,
   1980 	  BGE_QUIRK_LINK_STATE_BROKEN,
   1981 	  "unknown BCM5700" },
   1982 
   1983 	{ BGE_ASICREV_BCM5701,
   1984 	  BGE_QUIRK_PCIX_DMA_ALIGN_BUG,
   1985 	  "unknown BCM5701" },
   1986 
   1987 	{ BGE_ASICREV_BCM5703,
   1988 	  0,
   1989 	  "unknown BCM5703" },
   1990 
   1991 	{ BGE_ASICREV_BCM5704,
   1992 	  BGE_QUIRK_ONLY_PHY_1,
   1993 	  "unknown BCM5704" },
   1994 
   1995 	{ BGE_ASICREV_BCM5705,
   1996 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   1997 	  "unknown BCM5705" },
   1998 
   1999 	{ BGE_ASICREV_BCM5750,
   2000 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   2001 	  "unknown BCM575x family" },
   2002 
   2003 	{ BGE_ASICREV_BCM5714,
   2004 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   2005 	  "unknown BCM5714" },
   2006 
   2007 	{ BGE_ASICREV_BCM5752,
   2008 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   2009 	  "unknown BCM5752 family" },
   2010 
   2011 
   2012 	{ BGE_ASICREV_BCM5715,
   2013 	  BGE_QUIRK_ONLY_PHY_1|BGE_QUIRK_5705_CORE,
   2014 	  "unknown BCM5715" },
   2015 
   2016 	{ 0,
   2017 	  0,
   2018 	  NULL }
   2019 };
   2020 
   2021 
   2022 static const struct bge_revision *
   2023 bge_lookup_rev(uint32_t chipid)
   2024 {
   2025 	const struct bge_revision *br;
   2026 
   2027 	for (br = bge_revisions; br->br_name != NULL; br++) {
   2028 		if (br->br_chipid == chipid)
   2029 			return (br);
   2030 	}
   2031 
   2032 	for (br = bge_majorrevs; br->br_name != NULL; br++) {
   2033 		if (br->br_chipid == BGE_ASICREV(chipid))
   2034 			return (br);
   2035 	}
   2036 
   2037 	return (NULL);
   2038 }
   2039 
   2040 static const struct bge_product {
   2041 	pci_vendor_id_t		bp_vendor;
   2042 	pci_product_id_t	bp_product;
   2043 	const char		*bp_name;
   2044 } bge_products[] = {
   2045 	/*
   2046 	 * The BCM5700 documentation seems to indicate that the hardware
   2047 	 * still has the Alteon vendor ID burned into it, though it
   2048 	 * should always be overridden by the value in the EEPROM.  We'll
   2049 	 * check for it anyway.
   2050 	 */
   2051 	{ PCI_VENDOR_ALTEON,
   2052 	  PCI_PRODUCT_ALTEON_BCM5700,
   2053 	  "Broadcom BCM5700 Gigabit Ethernet",
   2054 	  },
   2055 	{ PCI_VENDOR_ALTEON,
   2056 	  PCI_PRODUCT_ALTEON_BCM5701,
   2057 	  "Broadcom BCM5701 Gigabit Ethernet",
   2058 	  },
   2059 
   2060 	{ PCI_VENDOR_ALTIMA,
   2061 	  PCI_PRODUCT_ALTIMA_AC1000,
   2062 	  "Altima AC1000 Gigabit Ethernet",
   2063 	  },
   2064 	{ PCI_VENDOR_ALTIMA,
   2065 	  PCI_PRODUCT_ALTIMA_AC1001,
   2066 	  "Altima AC1001 Gigabit Ethernet",
   2067 	   },
   2068 	{ PCI_VENDOR_ALTIMA,
   2069 	  PCI_PRODUCT_ALTIMA_AC9100,
   2070 	  "Altima AC9100 Gigabit Ethernet",
   2071 	  },
   2072 
   2073 	{ PCI_VENDOR_BROADCOM,
   2074 	  PCI_PRODUCT_BROADCOM_BCM5700,
   2075 	  "Broadcom BCM5700 Gigabit Ethernet",
   2076 	  },
   2077 	{ PCI_VENDOR_BROADCOM,
   2078 	  PCI_PRODUCT_BROADCOM_BCM5701,
   2079 	  "Broadcom BCM5701 Gigabit Ethernet",
   2080 	  },
   2081 	{ PCI_VENDOR_BROADCOM,
   2082 	  PCI_PRODUCT_BROADCOM_BCM5702,
   2083 	  "Broadcom BCM5702 Gigabit Ethernet",
   2084 	  },
   2085 	{ PCI_VENDOR_BROADCOM,
   2086 	  PCI_PRODUCT_BROADCOM_BCM5702X,
   2087 	  "Broadcom BCM5702X Gigabit Ethernet" },
   2088 
   2089 	{ PCI_VENDOR_BROADCOM,
   2090 	  PCI_PRODUCT_BROADCOM_BCM5703,
   2091 	  "Broadcom BCM5703 Gigabit Ethernet",
   2092 	  },
   2093 	{ PCI_VENDOR_BROADCOM,
   2094 	  PCI_PRODUCT_BROADCOM_BCM5703X,
   2095 	  "Broadcom BCM5703X Gigabit Ethernet",
   2096 	  },
   2097 	{ PCI_VENDOR_BROADCOM,
   2098 	  PCI_PRODUCT_BROADCOM_BCM5703A3,
   2099 	  "Broadcom BCM5703A3 Gigabit Ethernet",
   2100 	  },
   2101 
   2102    	{ PCI_VENDOR_BROADCOM,
   2103 	  PCI_PRODUCT_BROADCOM_BCM5704C,
   2104 	  "Broadcom BCM5704C Dual Gigabit Ethernet",
   2105 	  },
   2106    	{ PCI_VENDOR_BROADCOM,
   2107 	  PCI_PRODUCT_BROADCOM_BCM5704S,
   2108 	  "Broadcom BCM5704S Dual Gigabit Ethernet",
   2109 	  },
   2110 
   2111    	{ PCI_VENDOR_BROADCOM,
   2112 	  PCI_PRODUCT_BROADCOM_BCM5705,
   2113 	  "Broadcom BCM5705 Gigabit Ethernet",
   2114 	  },
   2115    	{ PCI_VENDOR_BROADCOM,
   2116 	  PCI_PRODUCT_BROADCOM_BCM5705K,
   2117 	  "Broadcom BCM5705K Gigabit Ethernet",
   2118 	  },
   2119    	{ PCI_VENDOR_BROADCOM,
   2120 	  PCI_PRODUCT_BROADCOM_BCM5705_ALT,
   2121 	  "Broadcom BCM5705 Gigabit Ethernet",
   2122 	  },
   2123    	{ PCI_VENDOR_BROADCOM,
   2124 	  PCI_PRODUCT_BROADCOM_BCM5705M,
   2125 	  "Broadcom BCM5705M Gigabit Ethernet",
   2126 	  },
   2127 
   2128 	{ PCI_VENDOR_BROADCOM,
   2129 	  PCI_PRODUCT_BROADCOM_BCM5714,
   2130 	  "Broadcom BCM5714/5715 Gigabit Ethernet",
   2131 	  },
   2132 
   2133 	{ PCI_VENDOR_BROADCOM,
   2134 	  PCI_PRODUCT_BROADCOM_BCM5721,
   2135 	  "Broadcom BCM5721 Gigabit Ethernet",
   2136 	  },
   2137 
   2138 	{ PCI_VENDOR_BROADCOM,
   2139 	  PCI_PRODUCT_BROADCOM_BCM5750,
   2140 	  "Broadcom BCM5750 Gigabit Ethernet",
   2141 	  },
   2142 
   2143 	{ PCI_VENDOR_BROADCOM,
   2144 	  PCI_PRODUCT_BROADCOM_BCM5750M,
   2145 	  "Broadcom BCM5750M Gigabit Ethernet",
   2146 	  },
   2147 
   2148 	{ PCI_VENDOR_BROADCOM,
   2149 	  PCI_PRODUCT_BROADCOM_BCM5751,
   2150 	  "Broadcom BCM5751 Gigabit Ethernet",
   2151 	  },
   2152 
   2153 	{ PCI_VENDOR_BROADCOM,
   2154 	  PCI_PRODUCT_BROADCOM_BCM5751M,
   2155 	  "Broadcom BCM5751M Gigabit Ethernet",
   2156 	  },
   2157 
   2158 	{ PCI_VENDOR_BROADCOM,
   2159 	  PCI_PRODUCT_BROADCOM_BCM5752,
   2160 	  "Broadcom BCM5752 Gigabit Ethernet",
   2161 	  },
   2162 
   2163    	{ PCI_VENDOR_BROADCOM,
   2164 	  PCI_PRODUCT_BROADCOM_BCM5782,
   2165 	  "Broadcom BCM5782 Gigabit Ethernet",
   2166 	  },
   2167    	{ PCI_VENDOR_BROADCOM,
   2168 	  PCI_PRODUCT_BROADCOM_BCM5788,
   2169 	  "Broadcom BCM5788 Gigabit Ethernet",
   2170 	  },
   2171    	{ PCI_VENDOR_BROADCOM,
   2172 	  PCI_PRODUCT_BROADCOM_BCM5789,
   2173 	  "Broadcom BCM5789 Gigabit Ethernet",
   2174 	  },
   2175 
   2176    	{ PCI_VENDOR_BROADCOM,
   2177 	  PCI_PRODUCT_BROADCOM_BCM5901,
   2178 	  "Broadcom BCM5901 Fast Ethernet",
   2179 	  },
   2180    	{ PCI_VENDOR_BROADCOM,
   2181 	  PCI_PRODUCT_BROADCOM_BCM5901A2,
   2182 	  "Broadcom BCM5901A2 Fast Ethernet",
   2183 	  },
   2184 
   2185 	{ PCI_VENDOR_SCHNEIDERKOCH,
   2186 	  PCI_PRODUCT_SCHNEIDERKOCH_SK_9DX1,
   2187 	  "SysKonnect SK-9Dx1 Gigabit Ethernet",
   2188 	  },
   2189 
   2190 	{ PCI_VENDOR_3COM,
   2191 	  PCI_PRODUCT_3COM_3C996,
   2192 	  "3Com 3c996 Gigabit Ethernet",
   2193 	  },
   2194 
   2195 	{ 0,
   2196 	  0,
   2197 	  NULL },
   2198 };
   2199 
   2200 static const struct bge_product *
   2201 bge_lookup(const struct pci_attach_args *pa)
   2202 {
   2203 	const struct bge_product *bp;
   2204 
   2205 	for (bp = bge_products; bp->bp_name != NULL; bp++) {
   2206 		if (PCI_VENDOR(pa->pa_id) == bp->bp_vendor &&
   2207 		    PCI_PRODUCT(pa->pa_id) == bp->bp_product)
   2208 			return (bp);
   2209 	}
   2210 
   2211 	return (NULL);
   2212 }
   2213 
   2214 static int
   2215 bge_setpowerstate(struct bge_softc *sc, int powerlevel)
   2216 {
   2217 #ifdef NOTYET
   2218 	u_int32_t pm_ctl = 0;
   2219 
   2220 	/* XXX FIXME: make sure indirect accesses enabled? */
   2221 	pm_ctl = pci_conf_read(sc->bge_dev, BGE_PCI_MISC_CTL, 4);
   2222 	pm_ctl |= BGE_PCIMISCCTL_INDIRECT_ACCESS;
   2223 	pci_write_config(sc->bge_dev, BGE_PCI_MISC_CTL, pm_ctl, 4);
   2224 
   2225 	/* clear the PME_assert bit and power state bits, enable PME */
   2226 	pm_ctl = pci_conf_read(sc->bge_dev, BGE_PCI_PWRMGMT_CMD, 2);
   2227 	pm_ctl &= ~PCIM_PSTAT_DMASK;
   2228 	pm_ctl |= (1 << 8);
   2229 
   2230 	if (powerlevel == 0) {
   2231 		pm_ctl |= PCIM_PSTAT_D0;
   2232 		pci_write_config(sc->bge_dev, BGE_PCI_PWRMGMT_CMD,
   2233 		    pm_ctl, 2);
   2234 		DELAY(10000);
   2235 		CSR_WRITE_4(sc, BGE_MISC_LOCAL_CTL, sc->bge_local_ctrl_reg);
   2236 		DELAY(10000);
   2237 
   2238 #ifdef NOTYET
   2239 		/* XXX FIXME: write 0x02 to phy aux_Ctrl reg */
   2240 		bge_miibus_writereg(sc->bge_dev, 1, 0x18, 0x02);
   2241 #endif
   2242 		DELAY(40); DELAY(40); DELAY(40);
   2243 		DELAY(10000);	/* above not quite adequate on 5700 */
   2244 		return 0;
   2245 	}
   2246 
   2247 
   2248 	/*
   2249 	 * Entering ACPI power states D1-D3 is achieved by wiggling
   2250 	 * GMII gpio pins. Example code assumes all hardware vendors
   2251 	 * followed Broadom's sample pcb layout. Until we verify that
   2252 	 * for all supported OEM cards, states D1-D3 are  unsupported.
   2253 	 */
   2254 	printf("%s: power state %d unimplemented; check GPIO pins\n",
   2255 	       sc->bge_dev.dv_xname, powerlevel);
   2256 #endif
   2257 	return EOPNOTSUPP;
   2258 }
   2259 
   2260 
   2261 /*
   2262  * Probe for a Broadcom chip. Check the PCI vendor and device IDs
   2263  * against our list and return its name if we find a match. Note
   2264  * that since the Broadcom controller contains VPD support, we
   2265  * can get the device name string from the controller itself instead
   2266  * of the compiled-in string. This is a little slow, but it guarantees
   2267  * we'll always announce the right product name.
   2268  */
   2269 static int
   2270 bge_probe(device_t parent, cfdata_t match, void *aux)
   2271 {
   2272 	struct pci_attach_args *pa = (struct pci_attach_args *)aux;
   2273 
   2274 	if (bge_lookup(pa) != NULL)
   2275 		return (1);
   2276 
   2277 	return (0);
   2278 }
   2279 
   2280 static void
   2281 bge_attach(device_t parent, device_t self, void *aux)
   2282 {
   2283 	struct bge_softc	*sc = (struct bge_softc *)self;
   2284 	struct pci_attach_args	*pa = aux;
   2285 	const struct bge_product *bp;
   2286 	const struct bge_revision *br;
   2287 	pci_chipset_tag_t	pc = pa->pa_pc;
   2288 	pci_intr_handle_t	ih;
   2289 	const char		*intrstr = NULL;
   2290 	bus_dma_segment_t	seg;
   2291 	int			rseg;
   2292 	u_int32_t		hwcfg = 0;
   2293 	u_int32_t		mac_addr = 0;
   2294 	u_int32_t		command;
   2295 	struct ifnet		*ifp;
   2296 	caddr_t			kva;
   2297 	u_char			eaddr[ETHER_ADDR_LEN];
   2298 	pcireg_t		memtype;
   2299 	bus_addr_t		memaddr;
   2300 	bus_size_t		memsize;
   2301 	u_int32_t		pm_ctl;
   2302 
   2303 	bp = bge_lookup(pa);
   2304 	KASSERT(bp != NULL);
   2305 
   2306 	sc->bge_pa = *pa;
   2307 
   2308 	aprint_naive(": Ethernet controller\n");
   2309 	aprint_normal(": %s\n", bp->bp_name);
   2310 
   2311 	/*
   2312 	 * Map control/status registers.
   2313 	 */
   2314 	DPRINTFN(5, ("Map control/status regs\n"));
   2315 	command = pci_conf_read(pc, pa->pa_tag, PCI_COMMAND_STATUS_REG);
   2316 	command |= PCI_COMMAND_MEM_ENABLE | PCI_COMMAND_MASTER_ENABLE;
   2317 	pci_conf_write(pc, pa->pa_tag, PCI_COMMAND_STATUS_REG, command);
   2318 	command = pci_conf_read(pc, pa->pa_tag, PCI_COMMAND_STATUS_REG);
   2319 
   2320 	if (!(command & PCI_COMMAND_MEM_ENABLE)) {
   2321 		aprint_error("%s: failed to enable memory mapping!\n",
   2322 		    sc->bge_dev.dv_xname);
   2323 		return;
   2324 	}
   2325 
   2326 	DPRINTFN(5, ("pci_mem_find\n"));
   2327 	memtype = pci_mapreg_type(pa->pa_pc, pa->pa_tag, BGE_PCI_BAR0);
   2328  	switch (memtype) {
   2329 	case PCI_MAPREG_TYPE_MEM | PCI_MAPREG_MEM_TYPE_32BIT:
   2330 	case PCI_MAPREG_TYPE_MEM | PCI_MAPREG_MEM_TYPE_64BIT:
   2331 		if (pci_mapreg_map(pa, BGE_PCI_BAR0,
   2332 		    memtype, 0, &sc->bge_btag, &sc->bge_bhandle,
   2333 		    &memaddr, &memsize) == 0)
   2334 			break;
   2335 	default:
   2336 		aprint_error("%s: can't find mem space\n",
   2337 		    sc->bge_dev.dv_xname);
   2338 		return;
   2339 	}
   2340 
   2341 	DPRINTFN(5, ("pci_intr_map\n"));
   2342 	if (pci_intr_map(pa, &ih)) {
   2343 		aprint_error("%s: couldn't map interrupt\n",
   2344 		    sc->bge_dev.dv_xname);
   2345 		return;
   2346 	}
   2347 
   2348 	DPRINTFN(5, ("pci_intr_string\n"));
   2349 	intrstr = pci_intr_string(pc, ih);
   2350 
   2351 	DPRINTFN(5, ("pci_intr_establish\n"));
   2352 	sc->bge_intrhand = pci_intr_establish(pc, ih, IPL_NET, bge_intr, sc);
   2353 
   2354 	if (sc->bge_intrhand == NULL) {
   2355 		aprint_error("%s: couldn't establish interrupt",
   2356 		    sc->bge_dev.dv_xname);
   2357 		if (intrstr != NULL)
   2358 			aprint_normal(" at %s", intrstr);
   2359 		aprint_normal("\n");
   2360 		return;
   2361 	}
   2362 	aprint_normal("%s: interrupting at %s\n",
   2363 	    sc->bge_dev.dv_xname, intrstr);
   2364 
   2365 	/*
   2366 	 * Kludge for 5700 Bx bug: a hardware bug (PCIX byte enable?)
   2367 	 * can clobber the chip's PCI config-space power control registers,
   2368 	 * leaving the card in D3 powersave state.
   2369 	 * We do not have memory-mapped registers in this state,
   2370 	 * so force device into D0 state before starting initialization.
   2371 	 */
   2372 	pm_ctl = pci_conf_read(pc, pa->pa_tag, BGE_PCI_PWRMGMT_CMD);
   2373 	pm_ctl &= ~(PCI_PWR_D0|PCI_PWR_D1|PCI_PWR_D2|PCI_PWR_D3);
   2374 	pm_ctl |= (1 << 8) | PCI_PWR_D0 ; /* D0 state */
   2375 	pci_conf_write(pc, pa->pa_tag, BGE_PCI_PWRMGMT_CMD, pm_ctl);
   2376 	DELAY(1000);	/* 27 usec is allegedly sufficent */
   2377 
   2378 	/*
   2379 	 * Save ASIC rev.  Look up any quirks associated with this
   2380 	 * ASIC.
   2381 	 */
   2382 	sc->bge_chipid =
   2383 	    pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_MISC_CTL) &
   2384 	    BGE_PCIMISCCTL_ASICREV;
   2385 
   2386 	/*
   2387 	 * Detect PCI-Express devices
   2388 	 * XXX: guessed from Linux/FreeBSD; no documentation
   2389 	 */
   2390 	if (BGE_ASICREV(sc->bge_chipid) == BGE_ASICREV_BCM5750 &&
   2391 	    pci_get_capability(pa->pa_pc, pa->pa_tag, PCI_CAP_PCIEXPRESS,
   2392 	    NULL, NULL) != 0)
   2393 		sc->bge_pcie = 1;
   2394 	else
   2395 		sc->bge_pcie = 0;
   2396 
   2397 	/* Try to reset the chip. */
   2398 	DPRINTFN(5, ("bge_reset\n"));
   2399 	bge_reset(sc);
   2400 
   2401 	if (bge_chipinit(sc)) {
   2402 		aprint_error("%s: chip initialization failed\n",
   2403 		    sc->bge_dev.dv_xname);
   2404 		bge_release_resources(sc);
   2405 		return;
   2406 	}
   2407 
   2408 	/*
   2409 	 * Get station address from the EEPROM.
   2410 	 */
   2411 	mac_addr = bge_readmem_ind(sc, 0x0c14);
   2412 	if ((mac_addr >> 16) == 0x484b) {
   2413 		eaddr[0] = (u_char)(mac_addr >> 8);
   2414 		eaddr[1] = (u_char)(mac_addr >> 0);
   2415 		mac_addr = bge_readmem_ind(sc, 0x0c18);
   2416 		eaddr[2] = (u_char)(mac_addr >> 24);
   2417 		eaddr[3] = (u_char)(mac_addr >> 16);
   2418 		eaddr[4] = (u_char)(mac_addr >> 8);
   2419 		eaddr[5] = (u_char)(mac_addr >> 0);
   2420 	} else if (bge_read_eeprom(sc, (caddr_t)eaddr,
   2421 	    BGE_EE_MAC_OFFSET + 2, ETHER_ADDR_LEN)) {
   2422 		aprint_error("%s: failed to read station address\n",
   2423 		    sc->bge_dev.dv_xname);
   2424 		bge_release_resources(sc);
   2425 		return;
   2426 	}
   2427 
   2428 	br = bge_lookup_rev(sc->bge_chipid);
   2429 	aprint_normal("%s: ", sc->bge_dev.dv_xname);
   2430 
   2431 	if (br == NULL) {
   2432 		aprint_normal("unknown ASIC (0x%04x)", sc->bge_chipid >> 16);
   2433 		sc->bge_quirks = 0;
   2434 	} else {
   2435 		aprint_normal("ASIC %s (0x%04x)",
   2436 		    br->br_name, sc->bge_chipid >> 16);
   2437 		sc->bge_quirks |= br->br_quirks;
   2438 	}
   2439 	aprint_normal(", Ethernet address %s\n", ether_sprintf(eaddr));
   2440 
   2441 	/* Allocate the general information block and ring buffers. */
   2442 	if (pci_dma64_available(pa))
   2443 		sc->bge_dmatag = pa->pa_dmat64;
   2444 	else
   2445 		sc->bge_dmatag = pa->pa_dmat;
   2446 	DPRINTFN(5, ("bus_dmamem_alloc\n"));
   2447 	if (bus_dmamem_alloc(sc->bge_dmatag, sizeof(struct bge_ring_data),
   2448 			     PAGE_SIZE, 0, &seg, 1, &rseg, BUS_DMA_NOWAIT)) {
   2449 		aprint_error("%s: can't alloc rx buffers\n",
   2450 		    sc->bge_dev.dv_xname);
   2451 		return;
   2452 	}
   2453 	DPRINTFN(5, ("bus_dmamem_map\n"));
   2454 	if (bus_dmamem_map(sc->bge_dmatag, &seg, rseg,
   2455 			   sizeof(struct bge_ring_data), &kva,
   2456 			   BUS_DMA_NOWAIT)) {
   2457 		aprint_error("%s: can't map DMA buffers (%d bytes)\n",
   2458 		    sc->bge_dev.dv_xname, (int)sizeof(struct bge_ring_data));
   2459 		bus_dmamem_free(sc->bge_dmatag, &seg, rseg);
   2460 		return;
   2461 	}
   2462 	DPRINTFN(5, ("bus_dmamem_create\n"));
   2463 	if (bus_dmamap_create(sc->bge_dmatag, sizeof(struct bge_ring_data), 1,
   2464 	    sizeof(struct bge_ring_data), 0,
   2465 	    BUS_DMA_NOWAIT, &sc->bge_ring_map)) {
   2466 		aprint_error("%s: can't create DMA map\n",
   2467 		    sc->bge_dev.dv_xname);
   2468 		bus_dmamem_unmap(sc->bge_dmatag, kva,
   2469 				 sizeof(struct bge_ring_data));
   2470 		bus_dmamem_free(sc->bge_dmatag, &seg, rseg);
   2471 		return;
   2472 	}
   2473 	DPRINTFN(5, ("bus_dmamem_load\n"));
   2474 	if (bus_dmamap_load(sc->bge_dmatag, sc->bge_ring_map, kva,
   2475 			    sizeof(struct bge_ring_data), NULL,
   2476 			    BUS_DMA_NOWAIT)) {
   2477 		bus_dmamap_destroy(sc->bge_dmatag, sc->bge_ring_map);
   2478 		bus_dmamem_unmap(sc->bge_dmatag, kva,
   2479 				 sizeof(struct bge_ring_data));
   2480 		bus_dmamem_free(sc->bge_dmatag, &seg, rseg);
   2481 		return;
   2482 	}
   2483 
   2484 	DPRINTFN(5, ("bzero\n"));
   2485 	sc->bge_rdata = (struct bge_ring_data *)kva;
   2486 
   2487 	memset(sc->bge_rdata, 0, sizeof(struct bge_ring_data));
   2488 
   2489 	/* Try to allocate memory for jumbo buffers. */
   2490 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   2491 		if (bge_alloc_jumbo_mem(sc)) {
   2492 			aprint_error("%s: jumbo buffer allocation failed\n",
   2493 			    sc->bge_dev.dv_xname);
   2494 		} else
   2495 			sc->ethercom.ec_capabilities |= ETHERCAP_JUMBO_MTU;
   2496 	}
   2497 
   2498 	/* Set default tuneable values. */
   2499 	sc->bge_stat_ticks = BGE_TICKS_PER_SEC;
   2500 	sc->bge_rx_coal_ticks = 150;
   2501 	sc->bge_rx_max_coal_bds = 64;
   2502 #ifdef ORIG_WPAUL_VALUES
   2503 	sc->bge_tx_coal_ticks = 150;
   2504 	sc->bge_tx_max_coal_bds = 128;
   2505 #else
   2506 	sc->bge_tx_coal_ticks = 300;
   2507 	sc->bge_tx_max_coal_bds = 400;
   2508 #endif
   2509 	if (sc->bge_quirks & BGE_QUIRK_5705_CORE) {
   2510 		sc->bge_tx_coal_ticks = (12 * 5);
   2511 		sc->bge_rx_max_coal_bds = (12 * 5);
   2512 			aprint_error("%s: setting short Tx thresholds\n",
   2513 			    sc->bge_dev.dv_xname);
   2514 	}
   2515 
   2516 	/* Set up ifnet structure */
   2517 	ifp = &sc->ethercom.ec_if;
   2518 	ifp->if_softc = sc;
   2519 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
   2520 	ifp->if_ioctl = bge_ioctl;
   2521 	ifp->if_start = bge_start;
   2522 	ifp->if_init = bge_init;
   2523 	ifp->if_watchdog = bge_watchdog;
   2524 	IFQ_SET_MAXLEN(&ifp->if_snd, max(BGE_TX_RING_CNT - 1, IFQ_MAXLEN));
   2525 	IFQ_SET_READY(&ifp->if_snd);
   2526 	DPRINTFN(5, ("bcopy\n"));
   2527 	strcpy(ifp->if_xname, sc->bge_dev.dv_xname);
   2528 
   2529 	if ((sc->bge_quirks & BGE_QUIRK_CSUM_BROKEN) == 0)
   2530 		sc->ethercom.ec_if.if_capabilities |=
   2531 		    IFCAP_CSUM_IPv4_Tx | IFCAP_CSUM_IPv4_Rx |
   2532 		    IFCAP_CSUM_TCPv4_Tx | IFCAP_CSUM_TCPv4_Rx |
   2533 		    IFCAP_CSUM_UDPv4_Tx | IFCAP_CSUM_UDPv4_Rx;
   2534 	sc->ethercom.ec_capabilities |=
   2535 	    ETHERCAP_VLAN_HWTAGGING | ETHERCAP_VLAN_MTU;
   2536 
   2537 	if (sc->bge_pcie)
   2538 		sc->ethercom.ec_if.if_capabilities |= IFCAP_TSOv4;
   2539 
   2540 	/*
   2541 	 * Do MII setup.
   2542 	 */
   2543 	DPRINTFN(5, ("mii setup\n"));
   2544 	sc->bge_mii.mii_ifp = ifp;
   2545 	sc->bge_mii.mii_readreg = bge_miibus_readreg;
   2546 	sc->bge_mii.mii_writereg = bge_miibus_writereg;
   2547 	sc->bge_mii.mii_statchg = bge_miibus_statchg;
   2548 
   2549 	/*
   2550 	 * Figure out what sort of media we have by checking the
   2551 	 * hardware config word in the first 32k of NIC internal memory,
   2552 	 * or fall back to the config word in the EEPROM. Note: on some BCM5700
   2553 	 * cards, this value appears to be unset. If that's the
   2554 	 * case, we have to rely on identifying the NIC by its PCI
   2555 	 * subsystem ID, as we do below for the SysKonnect SK-9D41.
   2556 	 */
   2557 	if (bge_readmem_ind(sc, BGE_SOFTWARE_GENCOMM_SIG) == BGE_MAGIC_NUMBER) {
   2558 		hwcfg = bge_readmem_ind(sc, BGE_SOFTWARE_GENCOMM_NICCFG);
   2559 	} else {
   2560 		bge_read_eeprom(sc, (caddr_t)&hwcfg,
   2561 		    BGE_EE_HWCFG_OFFSET, sizeof(hwcfg));
   2562 		hwcfg = be32toh(hwcfg);
   2563 	}
   2564 	if ((hwcfg & BGE_HWCFG_MEDIA) == BGE_MEDIA_FIBER)
   2565 		sc->bge_tbi = 1;
   2566 
   2567 	/* The SysKonnect SK-9D41 is a 1000baseSX card. */
   2568 	if ((pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_SUBSYS) >> 16) ==
   2569 	    SK_SUBSYSID_9D41)
   2570 		sc->bge_tbi = 1;
   2571 
   2572 	if (sc->bge_tbi) {
   2573 		ifmedia_init(&sc->bge_ifmedia, IFM_IMASK, bge_ifmedia_upd,
   2574 		    bge_ifmedia_sts);
   2575 		ifmedia_add(&sc->bge_ifmedia, IFM_ETHER|IFM_1000_SX, 0, NULL);
   2576 		ifmedia_add(&sc->bge_ifmedia, IFM_ETHER|IFM_1000_SX|IFM_FDX,
   2577 			    0, NULL);
   2578 		ifmedia_add(&sc->bge_ifmedia, IFM_ETHER|IFM_AUTO, 0, NULL);
   2579 		ifmedia_set(&sc->bge_ifmedia, IFM_ETHER|IFM_AUTO);
   2580 	} else {
   2581 		/*
   2582 		 * Do transceiver setup.
   2583 		 */
   2584 		ifmedia_init(&sc->bge_mii.mii_media, 0, bge_ifmedia_upd,
   2585 			     bge_ifmedia_sts);
   2586 		mii_attach(&sc->bge_dev, &sc->bge_mii, 0xffffffff,
   2587 			   MII_PHY_ANY, MII_OFFSET_ANY,
   2588 			   MIIF_FORCEANEG|MIIF_DOPAUSE);
   2589 
   2590 		if (LIST_FIRST(&sc->bge_mii.mii_phys) == NULL) {
   2591 			printf("%s: no PHY found!\n", sc->bge_dev.dv_xname);
   2592 			ifmedia_add(&sc->bge_mii.mii_media,
   2593 				    IFM_ETHER|IFM_MANUAL, 0, NULL);
   2594 			ifmedia_set(&sc->bge_mii.mii_media,
   2595 				    IFM_ETHER|IFM_MANUAL);
   2596 		} else
   2597 			ifmedia_set(&sc->bge_mii.mii_media,
   2598 				    IFM_ETHER|IFM_AUTO);
   2599 	}
   2600 
   2601 	/*
   2602 	 * When using the BCM5701 in PCI-X mode, data corruption has
   2603 	 * been observed in the first few bytes of some received packets.
   2604 	 * Aligning the packet buffer in memory eliminates the corruption.
   2605 	 * Unfortunately, this misaligns the packet payloads.  On platforms
   2606 	 * which do not support unaligned accesses, we will realign the
   2607 	 * payloads by copying the received packets.
   2608 	 */
   2609 	if (sc->bge_quirks & BGE_QUIRK_PCIX_DMA_ALIGN_BUG) {
   2610 		/* If in PCI-X mode, work around the alignment bug. */
   2611 		if ((pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_PCISTATE) &
   2612                     (BGE_PCISTATE_PCI_BUSMODE | BGE_PCISTATE_PCI_BUSSPEED)) ==
   2613                          BGE_PCISTATE_PCI_BUSSPEED)
   2614 		sc->bge_rx_alignment_bug = 1;
   2615         }
   2616 
   2617 	/*
   2618 	 * Call MI attach routine.
   2619 	 */
   2620 	DPRINTFN(5, ("if_attach\n"));
   2621 	if_attach(ifp);
   2622 	DPRINTFN(5, ("ether_ifattach\n"));
   2623 	ether_ifattach(ifp, eaddr);
   2624 #ifdef BGE_EVENT_COUNTERS
   2625 	/*
   2626 	 * Attach event counters.
   2627 	 */
   2628 	evcnt_attach_dynamic(&sc->bge_ev_intr, EVCNT_TYPE_INTR,
   2629 	    NULL, sc->bge_dev.dv_xname, "intr");
   2630 	evcnt_attach_dynamic(&sc->bge_ev_tx_xoff, EVCNT_TYPE_MISC,
   2631 	    NULL, sc->bge_dev.dv_xname, "tx_xoff");
   2632 	evcnt_attach_dynamic(&sc->bge_ev_tx_xon, EVCNT_TYPE_MISC,
   2633 	    NULL, sc->bge_dev.dv_xname, "tx_xon");
   2634 	evcnt_attach_dynamic(&sc->bge_ev_rx_xoff, EVCNT_TYPE_MISC,
   2635 	    NULL, sc->bge_dev.dv_xname, "rx_xoff");
   2636 	evcnt_attach_dynamic(&sc->bge_ev_rx_xon, EVCNT_TYPE_MISC,
   2637 	    NULL, sc->bge_dev.dv_xname, "rx_xon");
   2638 	evcnt_attach_dynamic(&sc->bge_ev_rx_macctl, EVCNT_TYPE_MISC,
   2639 	    NULL, sc->bge_dev.dv_xname, "rx_macctl");
   2640 	evcnt_attach_dynamic(&sc->bge_ev_xoffentered, EVCNT_TYPE_MISC,
   2641 	    NULL, sc->bge_dev.dv_xname, "xoffentered");
   2642 #endif /* BGE_EVENT_COUNTERS */
   2643 	DPRINTFN(5, ("callout_init\n"));
   2644 	callout_init(&sc->bge_timeout);
   2645 
   2646 	sc->bge_powerhook = powerhook_establish(bge_powerhook, sc);
   2647 	if (sc->bge_powerhook == NULL)
   2648 		printf("%s: WARNING: unable to establish PCI power hook\n",
   2649 		    sc->bge_dev.dv_xname);
   2650 }
   2651 
   2652 static void
   2653 bge_release_resources(struct bge_softc *sc)
   2654 {
   2655 	if (sc->bge_vpd_prodname != NULL)
   2656 		free(sc->bge_vpd_prodname, M_DEVBUF);
   2657 
   2658 	if (sc->bge_vpd_readonly != NULL)
   2659 		free(sc->bge_vpd_readonly, M_DEVBUF);
   2660 }
   2661 
   2662 static void
   2663 bge_reset(struct bge_softc *sc)
   2664 {
   2665 	struct pci_attach_args *pa = &sc->bge_pa;
   2666 	u_int32_t cachesize, command, pcistate, new_pcistate;
   2667 	int i, val;
   2668 
   2669 	/* Save some important PCI state. */
   2670 	cachesize = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_CACHESZ);
   2671 	command = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_CMD);
   2672 	pcistate = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_PCISTATE);
   2673 
   2674 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MISC_CTL,
   2675 	    BGE_PCIMISCCTL_INDIRECT_ACCESS|BGE_PCIMISCCTL_MASK_PCI_INTR|
   2676 	    BGE_HIF_SWAP_OPTIONS|BGE_PCIMISCCTL_PCISTATE_RW);
   2677 
   2678 	val = BGE_MISCCFG_RESET_CORE_CLOCKS | (65<<1);
   2679 	/*
   2680 	 * XXX: from FreeBSD/Linux; no documentation
   2681 	 */
   2682 	if (sc->bge_pcie) {
   2683 		if (CSR_READ_4(sc, BGE_PCIE_CTL1) == 0x60)
   2684 			CSR_WRITE_4(sc, BGE_PCIE_CTL1, 0x20);
   2685 		if (sc->bge_chipid != BGE_CHIPID_BCM5750_A0) {
   2686 			/* No idea what that actually means */
   2687 			CSR_WRITE_4(sc, BGE_MISC_CFG, 1 << 29);
   2688 			val |= (1<<29);
   2689 		}
   2690 	}
   2691 
   2692 	/* Issue global reset */
   2693 	bge_writereg_ind(sc, BGE_MISC_CFG, val);
   2694 
   2695 	DELAY(1000);
   2696 
   2697 	/*
   2698 	 * XXX: from FreeBSD/Linux; no documentation
   2699 	 */
   2700 	if (sc->bge_pcie) {
   2701 		if (sc->bge_chipid == BGE_CHIPID_BCM5750_A0) {
   2702 			pcireg_t reg;
   2703 
   2704 			DELAY(500000);
   2705 			/* XXX: Magic Numbers */
   2706 			reg = pci_conf_read(pa->pa_pc, pa->pa_tag, BGE_PCI_UNKNOWN0);
   2707 			pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_UNKNOWN0,
   2708 			    reg | (1 << 15));
   2709 		}
   2710 		/*
   2711 		 * XXX: Magic Numbers.
   2712 		 * Sets maximal PCI-e payload and clears any PCI-e errors.
   2713 		 * Should be replaced with references to PCI config-space
   2714 		 * capability block for PCI-Express.
   2715 		 */
   2716 		pci_conf_write(pa->pa_pc, pa->pa_tag,
   2717 		    BGE_PCI_CONF_DEV_CTRL, 0xf5000);
   2718 
   2719 	}
   2720 
   2721 	/* Reset some of the PCI state that got zapped by reset */
   2722 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_MISC_CTL,
   2723 	    BGE_PCIMISCCTL_INDIRECT_ACCESS|BGE_PCIMISCCTL_MASK_PCI_INTR|
   2724 	    BGE_HIF_SWAP_OPTIONS|BGE_PCIMISCCTL_PCISTATE_RW);
   2725 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_CMD, command);
   2726 	pci_conf_write(pa->pa_pc, pa->pa_tag, BGE_PCI_CACHESZ, cachesize);
   2727 	bge_writereg_ind(sc, BGE_MISC_CFG, (65 << 1));
   2728 
   2729 	/* Enable memory arbiter. */
   2730 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   2731 		uint32_t marbmode = 0;
   2732 		if (BGE_IS_5714_FAMILY(sc)) {
   2733 			marbmode = CSR_READ_4(sc, BGE_MARB_MODE);
   2734 		}
   2735  		CSR_WRITE_4(sc, BGE_MARB_MODE, BGE_MARBMODE_ENABLE | marbmode);
   2736 	}
   2737 
   2738 	/*
   2739 	 * Prevent PXE restart: write a magic number to the
   2740 	 * general communications memory at 0xB50.
   2741 	 */
   2742 	bge_writemem_ind(sc, BGE_SOFTWARE_GENCOMM, BGE_MAGIC_NUMBER);
   2743 
   2744 	/*
   2745 	 * Poll the value location we just wrote until
   2746 	 * we see the 1's complement of the magic number.
   2747 	 * This indicates that the firmware initialization
   2748 	 * is complete.
   2749 	 */
   2750 	for (i = 0; i < BGE_TIMEOUT; i++) {
   2751 		val = bge_readmem_ind(sc, BGE_SOFTWARE_GENCOMM);
   2752 		if (val == ~BGE_MAGIC_NUMBER)
   2753 			break;
   2754 		DELAY(1000);
   2755 	}
   2756 
   2757 	if (i >= BGE_TIMEOUT) {
   2758 		printf("%s: firmware handshake timed out, val = %x\n",
   2759 		    sc->bge_dev.dv_xname, val);
   2760 		/*
   2761 		 * XXX: occasionally fired on bcm5721, but without
   2762 		 * apparent harm.  For now, keep going if we timeout
   2763 		 * against PCI-E devices.
   2764 		 */
   2765 		 if (!sc->bge_pcie)
   2766 		  return;
   2767 	}
   2768 
   2769 	/*
   2770 	 * XXX Wait for the value of the PCISTATE register to
   2771 	 * return to its original pre-reset state. This is a
   2772 	 * fairly good indicator of reset completion. If we don't
   2773 	 * wait for the reset to fully complete, trying to read
   2774 	 * from the device's non-PCI registers may yield garbage
   2775 	 * results.
   2776 	 */
   2777 	for (i = 0; i < BGE_TIMEOUT; i++) {
   2778 		new_pcistate = pci_conf_read(pa->pa_pc, pa->pa_tag,
   2779 		    BGE_PCI_PCISTATE);
   2780 		if ((new_pcistate & ~BGE_PCISTATE_RESERVED) ==
   2781 		    (pcistate & ~BGE_PCISTATE_RESERVED))
   2782 			break;
   2783 		DELAY(10);
   2784 	}
   2785 	if ((new_pcistate & ~BGE_PCISTATE_RESERVED) !=
   2786 	    (pcistate & ~BGE_PCISTATE_RESERVED)) {
   2787 		printf("%s: pcistate failed to revert\n",
   2788 		    sc->bge_dev.dv_xname);
   2789 	}
   2790 
   2791 	/* XXX: from FreeBSD/Linux; no documentation */
   2792 	if (sc->bge_pcie && sc->bge_chipid != BGE_CHIPID_BCM5750_A0)
   2793 		CSR_WRITE_4(sc, BGE_PCIE_CTL0, CSR_READ_4(sc, BGE_PCIE_CTL0) | (1<<25));
   2794 
   2795 	/* Enable memory arbiter. */
   2796 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   2797 		uint32_t marbmode = 0;
   2798 		if (BGE_IS_5714_FAMILY(sc)) {
   2799 			marbmode = CSR_READ_4(sc, BGE_MARB_MODE);
   2800 		}
   2801  		CSR_WRITE_4(sc, BGE_MARB_MODE, BGE_MARBMODE_ENABLE | marbmode);
   2802 	}
   2803 
   2804 	/* Fix up byte swapping */
   2805 	CSR_WRITE_4(sc, BGE_MODE_CTL, BGE_DMA_SWAP_OPTIONS);
   2806 
   2807 	CSR_WRITE_4(sc, BGE_MAC_MODE, 0);
   2808 
   2809 	DELAY(10000);
   2810 }
   2811 
   2812 /*
   2813  * Frame reception handling. This is called if there's a frame
   2814  * on the receive return list.
   2815  *
   2816  * Note: we have to be able to handle two possibilities here:
   2817  * 1) the frame is from the jumbo recieve ring
   2818  * 2) the frame is from the standard receive ring
   2819  */
   2820 
   2821 static void
   2822 bge_rxeof(struct bge_softc *sc)
   2823 {
   2824 	struct ifnet *ifp;
   2825 	int stdcnt = 0, jumbocnt = 0;
   2826 	bus_dmamap_t dmamap;
   2827 	bus_addr_t offset, toff;
   2828 	bus_size_t tlen;
   2829 	int tosync;
   2830 
   2831 	ifp = &sc->ethercom.ec_if;
   2832 
   2833 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2834 	    offsetof(struct bge_ring_data, bge_status_block),
   2835 	    sizeof (struct bge_status_block),
   2836 	    BUS_DMASYNC_POSTREAD);
   2837 
   2838 	offset = offsetof(struct bge_ring_data, bge_rx_return_ring);
   2839 	tosync = sc->bge_rdata->bge_status_block.bge_idx[0].bge_rx_prod_idx -
   2840 	    sc->bge_rx_saved_considx;
   2841 
   2842 	toff = offset + (sc->bge_rx_saved_considx * sizeof (struct bge_rx_bd));
   2843 
   2844 	if (tosync < 0) {
   2845 		tlen = (sc->bge_return_ring_cnt - sc->bge_rx_saved_considx) *
   2846 		    sizeof (struct bge_rx_bd);
   2847 		bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2848 		    toff, tlen, BUS_DMASYNC_POSTREAD);
   2849 		tosync = -tosync;
   2850 	}
   2851 
   2852 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2853 	    offset, tosync * sizeof (struct bge_rx_bd),
   2854 	    BUS_DMASYNC_POSTREAD);
   2855 
   2856 	while(sc->bge_rx_saved_considx !=
   2857 	    sc->bge_rdata->bge_status_block.bge_idx[0].bge_rx_prod_idx) {
   2858 		struct bge_rx_bd	*cur_rx;
   2859 		u_int32_t		rxidx;
   2860 		struct mbuf		*m = NULL;
   2861 
   2862 		cur_rx = &sc->bge_rdata->
   2863 			bge_rx_return_ring[sc->bge_rx_saved_considx];
   2864 
   2865 		rxidx = cur_rx->bge_idx;
   2866 		BGE_INC(sc->bge_rx_saved_considx, sc->bge_return_ring_cnt);
   2867 
   2868 		if (cur_rx->bge_flags & BGE_RXBDFLAG_JUMBO_RING) {
   2869 			BGE_INC(sc->bge_jumbo, BGE_JUMBO_RX_RING_CNT);
   2870 			m = sc->bge_cdata.bge_rx_jumbo_chain[rxidx];
   2871 			sc->bge_cdata.bge_rx_jumbo_chain[rxidx] = NULL;
   2872 			jumbocnt++;
   2873 			if (cur_rx->bge_flags & BGE_RXBDFLAG_ERROR) {
   2874 				ifp->if_ierrors++;
   2875 				bge_newbuf_jumbo(sc, sc->bge_jumbo, m);
   2876 				continue;
   2877 			}
   2878 			if (bge_newbuf_jumbo(sc, sc->bge_jumbo,
   2879 					     NULL)== ENOBUFS) {
   2880 				ifp->if_ierrors++;
   2881 				bge_newbuf_jumbo(sc, sc->bge_jumbo, m);
   2882 				continue;
   2883 			}
   2884 		} else {
   2885 			BGE_INC(sc->bge_std, BGE_STD_RX_RING_CNT);
   2886 			m = sc->bge_cdata.bge_rx_std_chain[rxidx];
   2887 			sc->bge_cdata.bge_rx_std_chain[rxidx] = NULL;
   2888 			stdcnt++;
   2889 			dmamap = sc->bge_cdata.bge_rx_std_map[rxidx];
   2890 			sc->bge_cdata.bge_rx_std_map[rxidx] = 0;
   2891 			if (cur_rx->bge_flags & BGE_RXBDFLAG_ERROR) {
   2892 				ifp->if_ierrors++;
   2893 				bge_newbuf_std(sc, sc->bge_std, m, dmamap);
   2894 				continue;
   2895 			}
   2896 			if (bge_newbuf_std(sc, sc->bge_std,
   2897 			    NULL, dmamap) == ENOBUFS) {
   2898 				ifp->if_ierrors++;
   2899 				bge_newbuf_std(sc, sc->bge_std, m, dmamap);
   2900 				continue;
   2901 			}
   2902 		}
   2903 
   2904 		ifp->if_ipackets++;
   2905 #ifndef __NO_STRICT_ALIGNMENT
   2906                 /*
   2907                  * XXX: if the 5701 PCIX-Rx-DMA workaround is in effect,
   2908                  * the Rx buffer has the layer-2 header unaligned.
   2909                  * If our CPU requires alignment, re-align by copying.
   2910                  */
   2911 		if (sc->bge_rx_alignment_bug) {
   2912 			memmove(mtod(m, caddr_t) + ETHER_ALIGN, m->m_data,
   2913                                 cur_rx->bge_len);
   2914 			m->m_data += ETHER_ALIGN;
   2915 		}
   2916 #endif
   2917 
   2918 		m->m_pkthdr.len = m->m_len = cur_rx->bge_len - ETHER_CRC_LEN;
   2919 		m->m_pkthdr.rcvif = ifp;
   2920 
   2921 #if NBPFILTER > 0
   2922 		/*
   2923 		 * Handle BPF listeners. Let the BPF user see the packet.
   2924 		 */
   2925 		if (ifp->if_bpf)
   2926 			bpf_mtap(ifp->if_bpf, m);
   2927 #endif
   2928 
   2929 		m->m_pkthdr.csum_flags = M_CSUM_IPv4;
   2930 
   2931 		if ((cur_rx->bge_ip_csum ^ 0xffff) != 0)
   2932 			m->m_pkthdr.csum_flags |= M_CSUM_IPv4_BAD;
   2933 		/*
   2934 		 * Rx transport checksum-offload may also
   2935 		 * have bugs with packets which, when transmitted,
   2936 		 * were `runts' requiring padding.
   2937 		 */
   2938 		if (cur_rx->bge_flags & BGE_RXBDFLAG_TCP_UDP_CSUM &&
   2939 		    (/* (sc->_bge_quirks & BGE_QUIRK_SHORT_CKSUM_BUG) == 0 ||*/
   2940 		     m->m_pkthdr.len >= ETHER_MIN_NOPAD)) {
   2941 			m->m_pkthdr.csum_data =
   2942 			    cur_rx->bge_tcp_udp_csum;
   2943 			m->m_pkthdr.csum_flags |=
   2944 			    (M_CSUM_TCPv4|M_CSUM_UDPv4|
   2945 			     M_CSUM_DATA|M_CSUM_NO_PSEUDOHDR);
   2946 		}
   2947 
   2948 		/*
   2949 		 * If we received a packet with a vlan tag, pass it
   2950 		 * to vlan_input() instead of ether_input().
   2951 		 */
   2952 		if (cur_rx->bge_flags & BGE_RXBDFLAG_VLAN_TAG)
   2953 			VLAN_INPUT_TAG(ifp, m, cur_rx->bge_vlan_tag, continue);
   2954 
   2955 		(*ifp->if_input)(ifp, m);
   2956 	}
   2957 
   2958 	CSR_WRITE_4(sc, BGE_MBX_RX_CONS0_LO, sc->bge_rx_saved_considx);
   2959 	if (stdcnt)
   2960 		CSR_WRITE_4(sc, BGE_MBX_RX_STD_PROD_LO, sc->bge_std);
   2961 	if (jumbocnt)
   2962 		CSR_WRITE_4(sc, BGE_MBX_RX_JUMBO_PROD_LO, sc->bge_jumbo);
   2963 }
   2964 
   2965 static void
   2966 bge_txeof(struct bge_softc *sc)
   2967 {
   2968 	struct bge_tx_bd *cur_tx = NULL;
   2969 	struct ifnet *ifp;
   2970 	struct txdmamap_pool_entry *dma;
   2971 	bus_addr_t offset, toff;
   2972 	bus_size_t tlen;
   2973 	int tosync;
   2974 	struct mbuf *m;
   2975 
   2976 	ifp = &sc->ethercom.ec_if;
   2977 
   2978 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2979 	    offsetof(struct bge_ring_data, bge_status_block),
   2980 	    sizeof (struct bge_status_block),
   2981 	    BUS_DMASYNC_POSTREAD);
   2982 
   2983 	offset = offsetof(struct bge_ring_data, bge_tx_ring);
   2984 	tosync = sc->bge_rdata->bge_status_block.bge_idx[0].bge_tx_cons_idx -
   2985 	    sc->bge_tx_saved_considx;
   2986 
   2987 	toff = offset + (sc->bge_tx_saved_considx * sizeof (struct bge_tx_bd));
   2988 
   2989 	if (tosync < 0) {
   2990 		tlen = (BGE_TX_RING_CNT - sc->bge_tx_saved_considx) *
   2991 		    sizeof (struct bge_tx_bd);
   2992 		bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2993 		    toff, tlen, BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
   2994 		tosync = -tosync;
   2995 	}
   2996 
   2997 	bus_dmamap_sync(sc->bge_dmatag, sc->bge_ring_map,
   2998 	    offset, tosync * sizeof (struct bge_tx_bd),
   2999 	    BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
   3000 
   3001 	/*
   3002 	 * Go through our tx ring and free mbufs for those
   3003 	 * frames that have been sent.
   3004 	 */
   3005 	while (sc->bge_tx_saved_considx !=
   3006 	    sc->bge_rdata->bge_status_block.bge_idx[0].bge_tx_cons_idx) {
   3007 		u_int32_t		idx = 0;
   3008 
   3009 		idx = sc->bge_tx_saved_considx;
   3010 		cur_tx = &sc->bge_rdata->bge_tx_ring[idx];
   3011 		if (cur_tx->bge_flags & BGE_TXBDFLAG_END)
   3012 			ifp->if_opackets++;
   3013 		m = sc->bge_cdata.bge_tx_chain[idx];
   3014 		if (m != NULL) {
   3015 			sc->bge_cdata.bge_tx_chain[idx] = NULL;
   3016 			dma = sc->txdma[idx];
   3017 			bus_dmamap_sync(sc->bge_dmatag, dma->dmamap, 0,
   3018 			    dma->dmamap->dm_mapsize, BUS_DMASYNC_POSTWRITE);
   3019 			bus_dmamap_unload(sc->bge_dmatag, dma->dmamap);
   3020 			SLIST_INSERT_HEAD(&sc->txdma_list, dma, link);
   3021 			sc->txdma[idx] = NULL;
   3022 
   3023 			m_freem(m);
   3024 		}
   3025 		sc->bge_txcnt--;
   3026 		BGE_INC(sc->bge_tx_saved_considx, BGE_TX_RING_CNT);
   3027 		ifp->if_timer = 0;
   3028 	}
   3029 
   3030 	if (cur_tx != NULL)
   3031 		ifp->if_flags &= ~IFF_OACTIVE;
   3032 }
   3033 
   3034 static int
   3035 bge_intr(void *xsc)
   3036 {
   3037 	struct bge_softc *sc;
   3038 	struct ifnet *ifp;
   3039 
   3040 	sc = xsc;
   3041 	ifp = &sc->ethercom.ec_if;
   3042 
   3043 #ifdef notdef
   3044 	/* Avoid this for now -- checking this register is expensive. */
   3045 	/* Make sure this is really our interrupt. */
   3046 	if (!(CSR_READ_4(sc, BGE_MISC_LOCAL_CTL) & BGE_MLC_INTR_STATE))
   3047 		return (0);
   3048 #endif
   3049 	/* Ack interrupt and stop others from occuring. */
   3050 	CSR_WRITE_4(sc, BGE_MBX_IRQ0_LO, 1);
   3051 
   3052 	BGE_EVCNT_INCR(sc->bge_ev_intr);
   3053 
   3054 	/*
   3055 	 * Process link state changes.
   3056 	 * Grrr. The link status word in the status block does
   3057 	 * not work correctly on the BCM5700 rev AX and BX chips,
   3058 	 * according to all available information. Hence, we have
   3059 	 * to enable MII interrupts in order to properly obtain
   3060 	 * async link changes. Unfortunately, this also means that
   3061 	 * we have to read the MAC status register to detect link
   3062 	 * changes, thereby adding an additional register access to
   3063 	 * the interrupt handler.
   3064 	 */
   3065 
   3066 	if (sc->bge_quirks & BGE_QUIRK_LINK_STATE_BROKEN) {
   3067 		u_int32_t		status;
   3068 
   3069 		status = CSR_READ_4(sc, BGE_MAC_STS);
   3070 		if (status & BGE_MACSTAT_MI_INTERRUPT) {
   3071 			sc->bge_link = 0;
   3072 			callout_stop(&sc->bge_timeout);
   3073 			bge_tick(sc);
   3074 			/* Clear the interrupt */
   3075 			CSR_WRITE_4(sc, BGE_MAC_EVT_ENB,
   3076 			    BGE_EVTENB_MI_INTERRUPT);
   3077 			bge_miibus_readreg(&sc->bge_dev, 1, BRGPHY_MII_ISR);
   3078 			bge_miibus_writereg(&sc->bge_dev, 1, BRGPHY_MII_IMR,
   3079 			    BRGPHY_INTRS);
   3080 		}
   3081 	} else {
   3082 		if (sc->bge_rdata->bge_status_block.bge_status &
   3083 		    BGE_STATFLAG_LINKSTATE_CHANGED) {
   3084 			sc->bge_link = 0;
   3085 			callout_stop(&sc->bge_timeout);
   3086 			bge_tick(sc);
   3087 			/* Clear the interrupt */
   3088 			CSR_WRITE_4(sc, BGE_MAC_STS, BGE_MACSTAT_SYNC_CHANGED|
   3089 			    BGE_MACSTAT_CFG_CHANGED|BGE_MACSTAT_MI_COMPLETE|
   3090 			    BGE_MACSTAT_LINK_CHANGED);
   3091 		}
   3092 	}
   3093 
   3094 	if (ifp->if_flags & IFF_RUNNING) {
   3095 		/* Check RX return ring producer/consumer */
   3096 		bge_rxeof(sc);
   3097 
   3098 		/* Check TX ring producer/consumer */
   3099 		bge_txeof(sc);
   3100 	}
   3101 
   3102 	if (sc->bge_pending_rxintr_change) {
   3103 		uint32_t rx_ticks = sc->bge_rx_coal_ticks;
   3104 		uint32_t rx_bds = sc->bge_rx_max_coal_bds;
   3105 		uint32_t junk;
   3106 
   3107 		CSR_WRITE_4(sc, BGE_HCC_RX_COAL_TICKS, rx_ticks);
   3108 		DELAY(10);
   3109 		junk = CSR_READ_4(sc, BGE_HCC_RX_COAL_TICKS);
   3110 
   3111 		CSR_WRITE_4(sc, BGE_HCC_RX_MAX_COAL_BDS, rx_bds);
   3112 		DELAY(10);
   3113 		junk = CSR_READ_4(sc, BGE_HCC_RX_MAX_COAL_BDS);
   3114 
   3115 		sc->bge_pending_rxintr_change = 0;
   3116 	}
   3117 	bge_handle_events(sc);
   3118 
   3119 	/* Re-enable interrupts. */
   3120 	CSR_WRITE_4(sc, BGE_MBX_IRQ0_LO, 0);
   3121 
   3122 	if (ifp->if_flags & IFF_RUNNING && !IFQ_IS_EMPTY(&ifp->if_snd))
   3123 		bge_start(ifp);
   3124 
   3125 	return (1);
   3126 }
   3127 
   3128 static void
   3129 bge_tick(void *xsc)
   3130 {
   3131 	struct bge_softc *sc = xsc;
   3132 	struct mii_data *mii = &sc->bge_mii;
   3133 	struct ifmedia *ifm = NULL;
   3134 	struct ifnet *ifp = &sc->ethercom.ec_if;
   3135 	int s;
   3136 
   3137 	s = splnet();
   3138 
   3139 	bge_stats_update(sc);
   3140 	callout_reset(&sc->bge_timeout, hz, bge_tick, sc);
   3141 	if (sc->bge_link) {
   3142 		splx(s);
   3143 		return;
   3144 	}
   3145 
   3146 	if (sc->bge_tbi) {
   3147 		ifm = &sc->bge_ifmedia;
   3148 		if (CSR_READ_4(sc, BGE_MAC_STS) &
   3149 		    BGE_MACSTAT_TBI_PCS_SYNCHED) {
   3150 			sc->bge_link++;
   3151 			CSR_WRITE_4(sc, BGE_MAC_STS, 0xFFFFFFFF);
   3152 			if (!IFQ_IS_EMPTY(&ifp->if_snd))
   3153 				bge_start(ifp);
   3154 		}
   3155 		splx(s);
   3156 		return;
   3157 	}
   3158 
   3159 	mii_tick(mii);
   3160 
   3161 	if (!sc->bge_link && mii->mii_media_status & IFM_ACTIVE &&
   3162 	    IFM_SUBTYPE(mii->mii_media_active) != IFM_NONE) {
   3163 		sc->bge_link++;
   3164 		if (!IFQ_IS_EMPTY(&ifp->if_snd))
   3165 			bge_start(ifp);
   3166 	}
   3167 
   3168 	splx(s);
   3169 }
   3170 
   3171 static void
   3172 bge_stats_update(struct bge_softc *sc)
   3173 {
   3174 	struct ifnet *ifp = &sc->ethercom.ec_if;
   3175 	bus_size_t stats = BGE_MEMWIN_START + BGE_STATS_BLOCK;
   3176 	bus_size_t rstats = BGE_RX_STATS;
   3177 
   3178 #define READ_RSTAT(sc, stats, stat) \
   3179 	  CSR_READ_4(sc, stats + offsetof(struct bge_mac_stats_regs, stat))
   3180 
   3181 	if (sc->bge_quirks & BGE_QUIRK_5705_CORE) {
   3182 		ifp->if_collisions +=
   3183 		    READ_RSTAT(sc, rstats, dot3StatsSingleCollisionFrames) +
   3184 		    READ_RSTAT(sc, rstats, dot3StatsMultipleCollisionFrames) +
   3185 		    READ_RSTAT(sc, rstats, dot3StatsExcessiveCollisions) +
   3186 		    READ_RSTAT(sc, rstats, dot3StatsLateCollisions);
   3187 
   3188 		BGE_EVCNT_ADD(sc->bge_ev_tx_xoff,
   3189 			      READ_RSTAT(sc, rstats, outXoffSent));
   3190 		BGE_EVCNT_ADD(sc->bge_ev_tx_xon,
   3191 			      READ_RSTAT(sc, rstats, outXonSent));
   3192 		BGE_EVCNT_ADD(sc->bge_ev_rx_xoff,
   3193 			      READ_RSTAT(sc, rstats, xoffPauseFramesReceived));
   3194 		BGE_EVCNT_ADD(sc->bge_ev_rx_xon,
   3195 			      READ_RSTAT(sc, rstats, xonPauseFramesReceived));
   3196 		BGE_EVCNT_ADD(sc->bge_ev_rx_macctl,
   3197 			      READ_RSTAT(sc, rstats, macControlFramesReceived));
   3198 		BGE_EVCNT_ADD(sc->bge_ev_xoffentered,
   3199 			      READ_RSTAT(sc, rstats, xoffStateEntered));
   3200 		return;
   3201 	}
   3202 
   3203 #undef READ_RSTAT
   3204 #define READ_STAT(sc, stats, stat) \
   3205 	  CSR_READ_4(sc, stats + offsetof(struct bge_stats, stat))
   3206 
   3207 	ifp->if_collisions +=
   3208 	  (READ_STAT(sc, stats, dot3StatsSingleCollisionFrames.bge_addr_lo) +
   3209 	   READ_STAT(sc, stats, dot3StatsMultipleCollisionFrames.bge_addr_lo) +
   3210 	   READ_STAT(sc, stats, dot3StatsExcessiveCollisions.bge_addr_lo) +
   3211 	   READ_STAT(sc, stats, dot3StatsLateCollisions.bge_addr_lo)) -
   3212 	  ifp->if_collisions;
   3213 
   3214 	BGE_EVCNT_UPD(sc->bge_ev_tx_xoff,
   3215 		      READ_STAT(sc, stats, outXoffSent.bge_addr_lo));
   3216 	BGE_EVCNT_UPD(sc->bge_ev_tx_xon,
   3217 		      READ_STAT(sc, stats, outXonSent.bge_addr_lo));
   3218 	BGE_EVCNT_UPD(sc->bge_ev_rx_xoff,
   3219 		      READ_STAT(sc, stats,
   3220 		      		xoffPauseFramesReceived.bge_addr_lo));
   3221 	BGE_EVCNT_UPD(sc->bge_ev_rx_xon,
   3222 		      READ_STAT(sc, stats, xonPauseFramesReceived.bge_addr_lo));
   3223 	BGE_EVCNT_UPD(sc->bge_ev_rx_macctl,
   3224 		      READ_STAT(sc, stats,
   3225 		      		macControlFramesReceived.bge_addr_lo));
   3226 	BGE_EVCNT_UPD(sc->bge_ev_xoffentered,
   3227 		      READ_STAT(sc, stats, xoffStateEntered.bge_addr_lo));
   3228 
   3229 #undef READ_STAT
   3230 
   3231 #ifdef notdef
   3232 	ifp->if_collisions +=
   3233 	   (sc->bge_rdata->bge_info.bge_stats.dot3StatsSingleCollisionFrames +
   3234 	   sc->bge_rdata->bge_info.bge_stats.dot3StatsMultipleCollisionFrames +
   3235 	   sc->bge_rdata->bge_info.bge_stats.dot3StatsExcessiveCollisions +
   3236 	   sc->bge_rdata->bge_info.bge_stats.dot3StatsLateCollisions) -
   3237 	   ifp->if_collisions;
   3238 #endif
   3239 }
   3240 
   3241 /*
   3242  * Pad outbound frame to ETHER_MIN_NOPAD for an unusual reason.
   3243  * The bge hardware will pad out Tx runts to ETHER_MIN_NOPAD,
   3244  * but when such padded frames employ the  bge IP/TCP checksum offload,
   3245  * the hardware checksum assist gives incorrect results (possibly
   3246  * from incorporating its own padding into the UDP/TCP checksum; who knows).
   3247  * If we pad such runts with zeros, the onboard checksum comes out correct.
   3248  */
   3249 static inline int
   3250 bge_cksum_pad(struct mbuf *pkt)
   3251 {
   3252 	struct mbuf *last = NULL;
   3253 	int padlen;
   3254 
   3255 	padlen = ETHER_MIN_NOPAD - pkt->m_pkthdr.len;
   3256 
   3257 	/* if there's only the packet-header and we can pad there, use it. */
   3258 	if (pkt->m_pkthdr.len == pkt->m_len &&
   3259 	    !M_READONLY(pkt) && M_TRAILINGSPACE(pkt) >= padlen) {
   3260 		last = pkt;
   3261 	} else {
   3262 		/*
   3263 		 * Walk packet chain to find last mbuf. We will either
   3264 		 * pad there, or append a new mbuf and pad it
   3265 		 * (thus perhaps avoiding the bcm5700 dma-min bug).
   3266 		 */
   3267 		for (last = pkt; last->m_next != NULL; last = last->m_next) {
   3268 	      	       (void) 0; /* do nothing*/
   3269 		}
   3270 
   3271 		/* `last' now points to last in chain. */
   3272 		if (!M_READONLY(last) && M_TRAILINGSPACE(last) >= padlen) {
   3273 			(void) 0; /* we can pad here, in-place. */
   3274 		} else {
   3275 			/* Allocate new empty mbuf, pad it. Compact later. */
   3276 			struct mbuf *n;
   3277 			MGET(n, M_DONTWAIT, MT_DATA);
   3278 			n->m_len = 0;
   3279 			last->m_next = n;
   3280 			last = n;
   3281 		}
   3282 	}
   3283 
   3284 #ifdef DEBUG
   3285 	  /*KASSERT(M_WRITABLE(last), ("to-pad mbuf not writeable\n"));*/
   3286 	  KASSERT(M_TRAILINGSPACE(last) >= padlen /*, ("insufficient space to pad\n")*/ );
   3287 #endif
   3288 	/* Now zero the pad area, to avoid the bge cksum-assist bug */
   3289 	memset(mtod(last, caddr_t) + last->m_len, 0, padlen);
   3290 	last->m_len += padlen;
   3291 	pkt->m_pkthdr.len += padlen;
   3292 	return 0;
   3293 }
   3294 
   3295 /*
   3296  * Compact outbound packets to avoid bug with DMA segments less than 8 bytes.
   3297  */
   3298 static inline int
   3299 bge_compact_dma_runt(struct mbuf *pkt)
   3300 {
   3301 	struct mbuf	*m, *prev;
   3302 	int 		totlen, prevlen;
   3303 
   3304 	prev = NULL;
   3305 	totlen = 0;
   3306 	prevlen = -1;
   3307 
   3308 	for (m = pkt; m != NULL; prev = m,m = m->m_next) {
   3309 		int mlen = m->m_len;
   3310 		int shortfall = 8 - mlen ;
   3311 
   3312 		totlen += mlen;
   3313 		if (mlen == 0) {
   3314 			continue;
   3315 		}
   3316 		if (mlen >= 8)
   3317 			continue;
   3318 
   3319 		/* If we get here, mbuf data is too small for DMA engine.
   3320 		 * Try to fix by shuffling data to prev or next in chain.
   3321 		 * If that fails, do a compacting deep-copy of the whole chain.
   3322 		 */
   3323 
   3324 		/* Internal frag. If fits in prev, copy it there. */
   3325 		if (prev && !M_READONLY(prev) &&
   3326 		      M_TRAILINGSPACE(prev) >= m->m_len) {
   3327 		  	bcopy(m->m_data,
   3328 			      prev->m_data+prev->m_len,
   3329 			      mlen);
   3330 			prev->m_len += mlen;
   3331 			m->m_len = 0;
   3332 			/* XXX stitch chain */
   3333 			prev->m_next = m_free(m);
   3334 			m = prev;
   3335 			continue;
   3336 		}
   3337 		else if (m->m_next != NULL && !M_READONLY(m) &&
   3338 			     M_TRAILINGSPACE(m) >= shortfall &&
   3339 			     m->m_next->m_len >= (8 + shortfall)) {
   3340 		    /* m is writable and have enough data in next, pull up. */
   3341 
   3342 		  	bcopy(m->m_next->m_data,
   3343 			      m->m_data+m->m_len,
   3344 			      shortfall);
   3345 			m->m_len += shortfall;
   3346 			m->m_next->m_len -= shortfall;
   3347 			m->m_next->m_data += shortfall;
   3348 		}
   3349 		else if (m->m_next == NULL || 1) {
   3350 		  	/* Got a runt at the very end of the packet.
   3351 			 * borrow data from the tail of the preceding mbuf and
   3352 			 * update its length in-place. (The original data is still
   3353 			 * valid, so we can do this even if prev is not writable.)
   3354 			 */
   3355 
   3356 			/* if we'd make prev a runt, just move all of its data. */
   3357 #ifdef DEBUG
   3358 			KASSERT(prev != NULL /*, ("runt but null PREV")*/);
   3359 			KASSERT(prev->m_len >= 8 /*, ("runt prev")*/);
   3360 #endif
   3361 			if ((prev->m_len - shortfall) < 8)
   3362 				shortfall = prev->m_len;
   3363 
   3364 #ifdef notyet	/* just do the safe slow thing for now */
   3365 			if (!M_READONLY(m)) {
   3366 				if (M_LEADINGSPACE(m) < shorfall) {
   3367 					void *m_dat;
   3368 					m_dat = (m->m_flags & M_PKTHDR) ?
   3369 					  m->m_pktdat : m->dat;
   3370 					memmove(m_dat, mtod(m, void*), m->m_len);
   3371 					m->m_data = m_dat;
   3372 				    }
   3373 			} else
   3374 #endif	/* just do the safe slow thing */
   3375 			{
   3376 				struct mbuf * n = NULL;
   3377 				int newprevlen = prev->m_len - shortfall;
   3378 
   3379 				MGET(n, M_NOWAIT, MT_DATA);
   3380 				if (n == NULL)
   3381 				   return ENOBUFS;
   3382 				KASSERT(m->m_len + shortfall < MLEN
   3383 					/*,
   3384 					  ("runt %d +prev %d too big\n", m->m_len, shortfall)*/);
   3385 
   3386 				/* first copy the data we're stealing from prev */
   3387 				bcopy(prev->m_data + newprevlen, n->m_data, shortfall);
   3388 
   3389 				/* update prev->m_len accordingly */
   3390 				prev->m_len -= shortfall;
   3391 
   3392 				/* copy data from runt m */
   3393 				bcopy(m->m_data, n->m_data + shortfall, m->m_len);
   3394 
   3395 				/* n holds what we stole from prev, plus m */
   3396 				n->m_len = shortfall + m->m_len;
   3397 
   3398 				/* stitch n into chain and free m */
   3399 				n->m_next = m->m_next;
   3400 				prev->m_next = n;
   3401 				/* KASSERT(m->m_next == NULL); */
   3402 				m->m_next = NULL;
   3403 				m_free(m);
   3404 				m = n;	/* for continuing loop */
   3405 			}
   3406 		}
   3407 		prevlen = m->m_len;
   3408 	}
   3409 	return 0;
   3410 }
   3411 
   3412 /*
   3413  * Encapsulate an mbuf chain in the tx ring  by coupling the mbuf data
   3414  * pointers to descriptors.
   3415  */
   3416 static int
   3417 bge_encap(struct bge_softc *sc, struct mbuf *m_head, u_int32_t *txidx)
   3418 {
   3419 	struct bge_tx_bd	*f = NULL;
   3420 	u_int32_t		frag, cur, cnt = 0;
   3421 	u_int16_t		csum_flags = 0;
   3422 	u_int16_t		txbd_tso_flags = 0;
   3423 	struct txdmamap_pool_entry *dma;
   3424 	bus_dmamap_t dmamap;
   3425 	int			i = 0;
   3426 	struct m_tag		*mtag;
   3427 	int			use_tso, maxsegsize, error;
   3428 
   3429 	cur = frag = *txidx;
   3430 
   3431 	if (m_head->m_pkthdr.csum_flags) {
   3432 		if (m_head->m_pkthdr.csum_flags & M_CSUM_IPv4)
   3433 			csum_flags |= BGE_TXBDFLAG_IP_CSUM;
   3434 		if (m_head->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4))
   3435 			csum_flags |= BGE_TXBDFLAG_TCP_UDP_CSUM;
   3436 	}
   3437 
   3438 	/*
   3439 	 * If we were asked to do an outboard checksum, and the NIC
   3440 	 * has the bug where it sometimes adds in the Ethernet padding,
   3441 	 * explicitly pad with zeros so the cksum will be correct either way.
   3442 	 * (For now, do this for all chip versions, until newer
   3443 	 * are confirmed to not require the workaround.)
   3444 	 */
   3445 	if ((csum_flags & BGE_TXBDFLAG_TCP_UDP_CSUM) == 0 ||
   3446 #ifdef notyet
   3447 	    (sc->bge_quirks & BGE_QUIRK_SHORT_CKSUM_BUG) == 0 ||
   3448 #endif
   3449 	    m_head->m_pkthdr.len >= ETHER_MIN_NOPAD)
   3450 		goto check_dma_bug;
   3451 
   3452 	if (bge_cksum_pad(m_head) != 0) {
   3453 	    return ENOBUFS;
   3454 	}
   3455 
   3456 check_dma_bug:
   3457 	if (!(sc->bge_quirks & BGE_QUIRK_5700_SMALLDMA))
   3458 		goto doit;
   3459 	/*
   3460 	 * bcm5700 Revision B silicon cannot handle DMA descriptors with
   3461 	 * less than eight bytes.  If we encounter a teeny mbuf
   3462 	 * at the end of a chain, we can pad.  Otherwise, copy.
   3463 	 */
   3464 	if (bge_compact_dma_runt(m_head) != 0)
   3465 		return ENOBUFS;
   3466 
   3467 doit:
   3468 	dma = SLIST_FIRST(&sc->txdma_list);
   3469 	if (dma == NULL)
   3470 		return ENOBUFS;
   3471 	dmamap = dma->dmamap;
   3472 
   3473 	/*
   3474 	 * Set up any necessary TSO state before we start packing...
   3475 	 */
   3476 	use_tso = (m_head->m_pkthdr.csum_flags & M_CSUM_TSOv4) != 0;
   3477 	if (!use_tso) {
   3478 		maxsegsize = 0;
   3479 	} else {	/* TSO setup */
   3480 		unsigned  mss;
   3481 		struct ether_header *eh;
   3482 		unsigned ip_tcp_hlen, iptcp_opt_words, tcp_seg_flags, offset;
   3483 		struct mbuf * m0 = m_head;
   3484 		struct ip *ip;
   3485 		struct tcphdr *th;
   3486 		int iphl, hlen;
   3487 
   3488 		/*
   3489 		 * XXX It would be nice if the mbuf pkthdr had offset
   3490 		 * fields for the protocol headers.
   3491 		 */
   3492 
   3493 		eh = mtod(m0, struct ether_header *);
   3494 		switch (htons(eh->ether_type)) {
   3495 		case ETHERTYPE_IP:
   3496 			offset = ETHER_HDR_LEN;
   3497 			break;
   3498 
   3499 		case ETHERTYPE_VLAN:
   3500 			offset = ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN;
   3501 			break;
   3502 
   3503 		default:
   3504 			/*
   3505 			 * Don't support this protocol or encapsulation.
   3506 			 */
   3507 			return (ENOBUFS);
   3508 		}
   3509 
   3510 		/*
   3511 		 * TCP/IP headers are in the first mbuf; we can do
   3512 		 * this the easy way.
   3513 		 */
   3514 		iphl = M_CSUM_DATA_IPv4_IPHL(m0->m_pkthdr.csum_data);
   3515 		hlen = iphl + offset;
   3516 		if (__predict_false(m0->m_len <
   3517 				    (hlen + sizeof(struct tcphdr)))) {
   3518 
   3519 			  printf("TSO: hard case m0->m_len == %d <"
   3520 				 " ip/tcp hlen %zd, not handled yet\n",
   3521 				 m0->m_len, hlen+ sizeof(struct tcphdr));
   3522 #ifdef NOTYET
   3523 			/*
   3524 			 * XXX jonathan (at) NetBSD.org: untested.
   3525 			 * how to force  this branch to be taken?
   3526 			 */
   3527 			BGE_EVCNT_INCR(&sc->sc_ev_txtsopain);
   3528 
   3529 			m_copydata(m0, offset, sizeof(ip), &ip);
   3530 			m_copydata(m0, hlen, sizeof(th), &th);
   3531 
   3532 			ip.ip_len = 0;
   3533 
   3534 			m_copyback(m0, hlen + offsetof(struct ip, ip_len),
   3535 			    sizeof(ip.ip_len), &ip.ip_len);
   3536 
   3537 			th.th_sum = in_cksum_phdr(ip.ip_src.s_addr,
   3538 			    ip.ip_dst.s_addr, htons(IPPROTO_TCP));
   3539 
   3540 			m_copyback(m0, hlen + offsetof(struct tcphdr, th_sum),
   3541 			    sizeof(th.th_sum), &th.th_sum);
   3542 
   3543 			hlen += th.th_off << 2;
   3544 			iptcp_opt_words	= hlen;
   3545 #else
   3546 			/*
   3547 			 * if_wm "hard" case not yet supported, can we not
   3548 			 * mandate it out of existence?
   3549 			 */
   3550 			(void) ip; (void)th; (void) ip_tcp_hlen;
   3551 
   3552 			return ENOBUFS;
   3553 #endif
   3554 		} else {
   3555 			ip = (struct ip *) (mtod(m0, caddr_t) + offset);
   3556 			th = (struct tcphdr *) (mtod(m0, caddr_t) + hlen);
   3557 			ip_tcp_hlen = iphl +  (th->th_off << 2);
   3558 
   3559 			/* Total IP/TCP options, in 32-bit words */
   3560 			iptcp_opt_words = (ip_tcp_hlen
   3561 					   - sizeof(struct tcphdr)
   3562 					   - sizeof(struct ip)) >> 2;
   3563 		}
   3564 		if (BGE_IS_5750_OR_BEYOND(sc)) {
   3565 			th->th_sum = 0;
   3566 			csum_flags &= ~(BGE_TXBDFLAG_TCP_UDP_CSUM);
   3567 		} else {
   3568 			/*
   3569 			 * XXX jonathan (at) NetBSD.org: 5705 untested.
   3570 			 * Requires TSO firmware patch for 5701/5703/5704.
   3571 			 */
   3572 			th->th_sum = in_cksum_phdr(ip->ip_src.s_addr,
   3573 			    ip->ip_dst.s_addr, htons(IPPROTO_TCP));
   3574 		}
   3575 
   3576 		mss = m_head->m_pkthdr.segsz;
   3577 		txbd_tso_flags |=
   3578 		    BGE_TXBDFLAG_CPU_PRE_DMA |
   3579 		    BGE_TXBDFLAG_CPU_POST_DMA;
   3580 
   3581 		/*
   3582 		 * Our NIC TSO-assist assumes TSO has standard, optionless
   3583 		 * IPv4 and TCP headers, which total 40 bytes. By default,
   3584 		 * the NIC copies 40 bytes of IP/TCP header from the
   3585 		 * supplied header into the IP/TCP header portion of
   3586 		 * each post-TSO-segment. If the supplied packet has IP or
   3587 		 * TCP options, we need to tell the NIC to copy those extra
   3588 		 * bytes into each  post-TSO header, in addition to the normal
   3589 		 * 40-byte IP/TCP header (and to leave space accordingly).
   3590 		 * Unfortunately, the driver encoding of option length
   3591 		 * varies across different ASIC families.
   3592 		 */
   3593 		tcp_seg_flags = 0;
   3594 		if (iptcp_opt_words) {
   3595 			if ( BGE_IS_5705_OR_BEYOND(sc)) {
   3596 				tcp_seg_flags =
   3597 					iptcp_opt_words << 11;
   3598 			} else {
   3599 				txbd_tso_flags |=
   3600 					iptcp_opt_words << 12;
   3601 			}
   3602 		}
   3603 		maxsegsize = mss | tcp_seg_flags;
   3604 		ip->ip_len = htons(mss + ip_tcp_hlen);
   3605 
   3606 	}	/* TSO setup */
   3607 
   3608 	/*
   3609 	 * Start packing the mbufs in this chain into
   3610 	 * the fragment pointers. Stop when we run out
   3611 	 * of fragments or hit the end of the mbuf chain.
   3612 	 */
   3613 	error = bus_dmamap_load_mbuf(sc->bge_dmatag, dmamap, m_head,
   3614 	    BUS_DMA_NOWAIT);
   3615 	if (error) {
   3616 		return(ENOBUFS);
   3617 	}
   3618 
   3619 	mtag = sc->ethercom.ec_nvlans ?
   3620 	    m_tag_find(m_head, PACKET_TAG_VLAN, NULL) : NULL;
   3621 
   3622 
   3623 	/* Iterate over dmap-map fragments. */
   3624 	for (i = 0; i < dmamap->dm_nsegs; i++) {
   3625 		f = &sc->bge_rdata->bge_tx_ring[frag];
   3626 		if (sc->bge_cdata.bge_tx_chain[frag] != NULL)
   3627 			break;
   3628 
   3629 		bge_set_hostaddr(&f->bge_addr, dmamap->dm_segs[i].ds_addr);
   3630 		f->bge_len = dmamap->dm_segs[i].ds_len;
   3631 
   3632 		/*
   3633 		 * For 5751 and follow-ons, for TSO we must turn
   3634 		 * off checksum-assist flag in the tx-descr, and
   3635 		 * supply the ASIC-revision-specific encoding
   3636 		 * of TSO flags and segsize.
   3637 		 */
   3638 		if (use_tso) {
   3639 			if (BGE_IS_5750_OR_BEYOND(sc) || i == 0) {
   3640 				f->bge_rsvd = maxsegsize;
   3641 				f->bge_flags = csum_flags | txbd_tso_flags;
   3642 			} else {
   3643 				f->bge_rsvd = 0;
   3644 				f->bge_flags =
   3645 				  (csum_flags | txbd_tso_flags) & 0x0fff;
   3646 			}
   3647 		} else {
   3648 			f->bge_rsvd = 0;
   3649 			f->bge_flags = csum_flags;
   3650 		}
   3651 
   3652 		if (mtag != NULL) {
   3653 			f->bge_flags |= BGE_TXBDFLAG_VLAN_TAG;
   3654 			f->bge_vlan_tag = VLAN_TAG_VALUE(mtag);
   3655 		} else {
   3656 			f->bge_vlan_tag = 0;
   3657 		}
   3658 		/*
   3659 		 * Sanity check: avoid coming within 16 descriptors
   3660 		 * of the end of the ring.
   3661 		 */
   3662 		if ((BGE_TX_RING_CNT - (sc->bge_txcnt + cnt)) < 16) {
   3663 			BGE_TSO_PRINTF(("%s: "
   3664 			    " dmamap_load_mbuf too close to ring wrap\n",
   3665 			    sc->bge_dev.dv_xname));
   3666 			return(ENOBUFS);
   3667 		}
   3668 		cur = frag;
   3669 		BGE_INC(frag, BGE_TX_RING_CNT);
   3670 		cnt++;
   3671 	}
   3672 
   3673 	if (i < dmamap->dm_nsegs) {
   3674 		BGE_TSO_PRINTF(("%s: reached %d < dm_nsegs %d\n",
   3675 		    sc->bge_dev.dv_xname, i, dmamap->dm_nsegs));
   3676 		return ENOBUFS;
   3677 	}
   3678 
   3679 	bus_dmamap_sync(sc->bge_dmatag, dmamap, 0, dmamap->dm_mapsize,
   3680 	    BUS_DMASYNC_PREWRITE);
   3681 
   3682 	if (frag == sc->bge_tx_saved_considx) {
   3683 		BGE_TSO_PRINTF(("%s: frag %d = wrapped id %d?\n",
   3684 		    sc->bge_dev.dv_xname, frag, sc->bge_tx_saved_considx));
   3685 
   3686 		return(ENOBUFS);
   3687 	}
   3688 
   3689 	sc->bge_rdata->bge_tx_ring[cur].bge_flags |= BGE_TXBDFLAG_END;
   3690 	sc->bge_cdata.bge_tx_chain[cur] = m_head;
   3691 	SLIST_REMOVE_HEAD(&sc->txdma_list, link);
   3692 	sc->txdma[cur] = dma;
   3693 	sc->bge_txcnt += cnt;
   3694 
   3695 	*txidx = frag;
   3696 
   3697 	return(0);
   3698 }
   3699 
   3700 /*
   3701  * Main transmit routine. To avoid having to do mbuf copies, we put pointers
   3702  * to the mbuf data regions directly in the transmit descriptors.
   3703  */
   3704 static void
   3705 bge_start(struct ifnet *ifp)
   3706 {
   3707 	struct bge_softc *sc;
   3708 	struct mbuf *m_head = NULL;
   3709 	u_int32_t prodidx;
   3710 	int pkts = 0;
   3711 
   3712 	sc = ifp->if_softc;
   3713 
   3714 	if (!sc->bge_link && ifp->if_snd.ifq_len < 10)
   3715 		return;
   3716 
   3717 	prodidx = sc->bge_tx_prodidx;
   3718 
   3719 	while(sc->bge_cdata.bge_tx_chain[prodidx] == NULL) {
   3720 		IFQ_POLL(&ifp->if_snd, m_head);
   3721 		if (m_head == NULL)
   3722 			break;
   3723 
   3724 #if 0
   3725 		/*
   3726 		 * XXX
   3727 		 * safety overkill.  If this is a fragmented packet chain
   3728 		 * with delayed TCP/UDP checksums, then only encapsulate
   3729 		 * it if we have enough descriptors to handle the entire
   3730 		 * chain at once.
   3731 		 * (paranoia -- may not actually be needed)
   3732 		 */
   3733 		if (m_head->m_flags & M_FIRSTFRAG &&
   3734 		    m_head->m_pkthdr.csum_flags & (CSUM_DELAY_DATA)) {
   3735 			if ((BGE_TX_RING_CNT - sc->bge_txcnt) <
   3736 			    M_CSUM_DATA_IPv4_OFFSET(m_head->m_pkthdr.csum_data) + 16) {
   3737 				ifp->if_flags |= IFF_OACTIVE;
   3738 				break;
   3739 			}
   3740 		}
   3741 #endif
   3742 
   3743 		/*
   3744 		 * Pack the data into the transmit ring. If we
   3745 		 * don't have room, set the OACTIVE flag and wait
   3746 		 * for the NIC to drain the ring.
   3747 		 */
   3748 		if (bge_encap(sc, m_head, &prodidx)) {
   3749 			printf("bge: failed on len %d?\n", m_head->m_pkthdr.len);
   3750 			ifp->if_flags |= IFF_OACTIVE;
   3751 			break;
   3752 		}
   3753 
   3754 		/* now we are committed to transmit the packet */
   3755 		IFQ_DEQUEUE(&ifp->if_snd, m_head);
   3756 		pkts++;
   3757 
   3758 #if NBPFILTER > 0
   3759 		/*
   3760 		 * If there's a BPF listener, bounce a copy of this frame
   3761 		 * to him.
   3762 		 */
   3763 		if (ifp->if_bpf)
   3764 			bpf_mtap(ifp->if_bpf, m_head);
   3765 #endif
   3766 	}
   3767 	if (pkts == 0)
   3768 		return;
   3769 
   3770 	/* Transmit */
   3771 	CSR_WRITE_4(sc, BGE_MBX_TX_HOST_PROD0_LO, prodidx);
   3772 	if (sc->bge_quirks & BGE_QUIRK_PRODUCER_BUG)	/* 5700 b2 errata */
   3773 		CSR_WRITE_4(sc, BGE_MBX_TX_HOST_PROD0_LO, prodidx);
   3774 
   3775 	sc->bge_tx_prodidx = prodidx;
   3776 
   3777 	/*
   3778 	 * Set a timeout in case the chip goes out to lunch.
   3779 	 */
   3780 	ifp->if_timer = 5;
   3781 }
   3782 
   3783 static int
   3784 bge_init(struct ifnet *ifp)
   3785 {
   3786 	struct bge_softc *sc = ifp->if_softc;
   3787 	u_int16_t *m;
   3788 	int s, error;
   3789 
   3790 	s = splnet();
   3791 
   3792 	ifp = &sc->ethercom.ec_if;
   3793 
   3794 	/* Cancel pending I/O and flush buffers. */
   3795 	bge_stop(sc);
   3796 	bge_reset(sc);
   3797 	bge_chipinit(sc);
   3798 
   3799 	/*
   3800 	 * Init the various state machines, ring
   3801 	 * control blocks and firmware.
   3802 	 */
   3803 	error = bge_blockinit(sc);
   3804 	if (error != 0) {
   3805 		printf("%s: initialization error %d\n", sc->bge_dev.dv_xname,
   3806 		    error);
   3807 		splx(s);
   3808 		return error;
   3809 	}
   3810 
   3811 	ifp = &sc->ethercom.ec_if;
   3812 
   3813 	/* Specify MTU. */
   3814 	CSR_WRITE_4(sc, BGE_RX_MTU, ifp->if_mtu +
   3815 	    ETHER_HDR_LEN + ETHER_CRC_LEN + ETHER_VLAN_ENCAP_LEN);
   3816 
   3817 	/* Load our MAC address. */
   3818 	m = (u_int16_t *)&(LLADDR(ifp->if_sadl)[0]);
   3819 	CSR_WRITE_4(sc, BGE_MAC_ADDR1_LO, htons(m[0]));
   3820 	CSR_WRITE_4(sc, BGE_MAC_ADDR1_HI, (htons(m[1]) << 16) | htons(m[2]));
   3821 
   3822 	/* Enable or disable promiscuous mode as needed. */
   3823 	if (ifp->if_flags & IFF_PROMISC) {
   3824 		BGE_SETBIT(sc, BGE_RX_MODE, BGE_RXMODE_RX_PROMISC);
   3825 	} else {
   3826 		BGE_CLRBIT(sc, BGE_RX_MODE, BGE_RXMODE_RX_PROMISC);
   3827 	}
   3828 
   3829 	/* Program multicast filter. */
   3830 	bge_setmulti(sc);
   3831 
   3832 	/* Init RX ring. */
   3833 	bge_init_rx_ring_std(sc);
   3834 
   3835 	/* Init jumbo RX ring. */
   3836 	if (ifp->if_mtu > (ETHERMTU + ETHER_HDR_LEN + ETHER_CRC_LEN))
   3837 		bge_init_rx_ring_jumbo(sc);
   3838 
   3839 	/* Init our RX return ring index */
   3840 	sc->bge_rx_saved_considx = 0;
   3841 
   3842 	/* Init TX ring. */
   3843 	bge_init_tx_ring(sc);
   3844 
   3845 	/* Turn on transmitter */
   3846 	BGE_SETBIT(sc, BGE_TX_MODE, BGE_TXMODE_ENABLE);
   3847 
   3848 	/* Turn on receiver */
   3849 	BGE_SETBIT(sc, BGE_RX_MODE, BGE_RXMODE_ENABLE);
   3850 
   3851 	CSR_WRITE_4(sc, BGE_MAX_RX_FRAME_LOWAT, 2);
   3852 
   3853 	/* Tell firmware we're alive. */
   3854 	BGE_SETBIT(sc, BGE_MODE_CTL, BGE_MODECTL_STACKUP);
   3855 
   3856 	/* Enable host interrupts. */
   3857 	BGE_SETBIT(sc, BGE_PCI_MISC_CTL, BGE_PCIMISCCTL_CLEAR_INTA);
   3858 	BGE_CLRBIT(sc, BGE_PCI_MISC_CTL, BGE_PCIMISCCTL_MASK_PCI_INTR);
   3859 	CSR_WRITE_4(sc, BGE_MBX_IRQ0_LO, 0);
   3860 
   3861 	bge_ifmedia_upd(ifp);
   3862 
   3863 	ifp->if_flags |= IFF_RUNNING;
   3864 	ifp->if_flags &= ~IFF_OACTIVE;
   3865 
   3866 	splx(s);
   3867 
   3868 	callout_reset(&sc->bge_timeout, hz, bge_tick, sc);
   3869 
   3870 	return 0;
   3871 }
   3872 
   3873 /*
   3874  * Set media options.
   3875  */
   3876 static int
   3877 bge_ifmedia_upd(struct ifnet *ifp)
   3878 {
   3879 	struct bge_softc *sc = ifp->if_softc;
   3880 	struct mii_data *mii = &sc->bge_mii;
   3881 	struct ifmedia *ifm = &sc->bge_ifmedia;
   3882 
   3883 	/* If this is a 1000baseX NIC, enable the TBI port. */
   3884 	if (sc->bge_tbi) {
   3885 		if (IFM_TYPE(ifm->ifm_media) != IFM_ETHER)
   3886 			return(EINVAL);
   3887 		switch(IFM_SUBTYPE(ifm->ifm_media)) {
   3888 		case IFM_AUTO:
   3889 			break;
   3890 		case IFM_1000_SX:
   3891 			if ((ifm->ifm_media & IFM_GMASK) == IFM_FDX) {
   3892 				BGE_CLRBIT(sc, BGE_MAC_MODE,
   3893 				    BGE_MACMODE_HALF_DUPLEX);
   3894 			} else {
   3895 				BGE_SETBIT(sc, BGE_MAC_MODE,
   3896 				    BGE_MACMODE_HALF_DUPLEX);
   3897 			}
   3898 			break;
   3899 		default:
   3900 			return(EINVAL);
   3901 		}
   3902 		/* XXX 802.3x flow control for 1000BASE-SX */
   3903 		return(0);
   3904 	}
   3905 
   3906 	sc->bge_link = 0;
   3907 	mii_mediachg(mii);
   3908 
   3909 	return(0);
   3910 }
   3911 
   3912 /*
   3913  * Report current media status.
   3914  */
   3915 static void
   3916 bge_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr)
   3917 {
   3918 	struct bge_softc *sc = ifp->if_softc;
   3919 	struct mii_data *mii = &sc->bge_mii;
   3920 
   3921 	if (sc->bge_tbi) {
   3922 		ifmr->ifm_status = IFM_AVALID;
   3923 		ifmr->ifm_active = IFM_ETHER;
   3924 		if (CSR_READ_4(sc, BGE_MAC_STS) &
   3925 		    BGE_MACSTAT_TBI_PCS_SYNCHED)
   3926 			ifmr->ifm_status |= IFM_ACTIVE;
   3927 		ifmr->ifm_active |= IFM_1000_SX;
   3928 		if (CSR_READ_4(sc, BGE_MAC_MODE) & BGE_MACMODE_HALF_DUPLEX)
   3929 			ifmr->ifm_active |= IFM_HDX;
   3930 		else
   3931 			ifmr->ifm_active |= IFM_FDX;
   3932 		return;
   3933 	}
   3934 
   3935 	mii_pollstat(mii);
   3936 	ifmr->ifm_status = mii->mii_media_status;
   3937 	ifmr->ifm_active = (mii->mii_media_active & ~IFM_ETH_FMASK) |
   3938 	    sc->bge_flowflags;
   3939 }
   3940 
   3941 static int
   3942 bge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
   3943 {
   3944 	struct bge_softc *sc = ifp->if_softc;
   3945 	struct ifreq *ifr = (struct ifreq *) data;
   3946 	int s, error = 0;
   3947 	struct mii_data *mii;
   3948 
   3949 	s = splnet();
   3950 
   3951 	switch(command) {
   3952 	case SIOCSIFFLAGS:
   3953 		if (ifp->if_flags & IFF_UP) {
   3954 			/*
   3955 			 * If only the state of the PROMISC flag changed,
   3956 			 * then just use the 'set promisc mode' command
   3957 			 * instead of reinitializing the entire NIC. Doing
   3958 			 * a full re-init means reloading the firmware and
   3959 			 * waiting for it to start up, which may take a
   3960 			 * second or two.
   3961 			 */
   3962 			if (ifp->if_flags & IFF_RUNNING &&
   3963 			    ifp->if_flags & IFF_PROMISC &&
   3964 			    !(sc->bge_if_flags & IFF_PROMISC)) {
   3965 				BGE_SETBIT(sc, BGE_RX_MODE,
   3966 				    BGE_RXMODE_RX_PROMISC);
   3967 			} else if (ifp->if_flags & IFF_RUNNING &&
   3968 			    !(ifp->if_flags & IFF_PROMISC) &&
   3969 			    sc->bge_if_flags & IFF_PROMISC) {
   3970 				BGE_CLRBIT(sc, BGE_RX_MODE,
   3971 				    BGE_RXMODE_RX_PROMISC);
   3972 			} else if (!(sc->bge_if_flags & IFF_UP))
   3973 				bge_init(ifp);
   3974 		} else {
   3975 			if (ifp->if_flags & IFF_RUNNING) {
   3976 				bge_stop(sc);
   3977 			}
   3978 		}
   3979 		sc->bge_if_flags = ifp->if_flags;
   3980 		error = 0;
   3981 		break;
   3982 	case SIOCSIFMEDIA:
   3983 		/* XXX Flow control is not supported for 1000BASE-SX */
   3984 		if (sc->bge_tbi) {
   3985 			ifr->ifr_media &= ~IFM_ETH_FMASK;
   3986 			sc->bge_flowflags = 0;
   3987 		}
   3988 
   3989 		/* Flow control requires full-duplex mode. */
   3990 		if (IFM_SUBTYPE(ifr->ifr_media) == IFM_AUTO ||
   3991 		    (ifr->ifr_media & IFM_FDX) == 0) {
   3992 		    	ifr->ifr_media &= ~IFM_ETH_FMASK;
   3993 		}
   3994 		if (IFM_SUBTYPE(ifr->ifr_media) != IFM_AUTO) {
   3995 			if ((ifr->ifr_media & IFM_ETH_FMASK) == IFM_FLOW) {
   3996 				/* We an do both TXPAUSE and RXPAUSE. */
   3997 				ifr->ifr_media |=
   3998 				    IFM_ETH_TXPAUSE | IFM_ETH_RXPAUSE;
   3999 			}
   4000 			sc->bge_flowflags = ifr->ifr_media & IFM_ETH_FMASK;
   4001 		}
   4002 		/* FALLTHROUGH */
   4003 	case SIOCGIFMEDIA:
   4004 		if (sc->bge_tbi) {
   4005 			error = ifmedia_ioctl(ifp, ifr, &sc->bge_ifmedia,
   4006 			    command);
   4007 		} else {
   4008 			mii = &sc->bge_mii;
   4009 			error = ifmedia_ioctl(ifp, ifr, &mii->mii_media,
   4010 			    command);
   4011 		}
   4012 		break;
   4013 	default:
   4014 		error = ether_ioctl(ifp, command, data);
   4015 		if (error == ENETRESET) {
   4016 			if (ifp->if_flags & IFF_RUNNING)
   4017 				bge_setmulti(sc);
   4018 			error = 0;
   4019 		}
   4020 		break;
   4021 	}
   4022 
   4023 	splx(s);
   4024 
   4025 	return(error);
   4026 }
   4027 
   4028 static void
   4029 bge_watchdog(struct ifnet *ifp)
   4030 {
   4031 	struct bge_softc *sc;
   4032 
   4033 	sc = ifp->if_softc;
   4034 
   4035 	printf("%s: watchdog timeout -- resetting\n", sc->bge_dev.dv_xname);
   4036 
   4037 	ifp->if_flags &= ~IFF_RUNNING;
   4038 	bge_init(ifp);
   4039 
   4040 	ifp->if_oerrors++;
   4041 }
   4042 
   4043 static void
   4044 bge_stop_block(struct bge_softc *sc, bus_addr_t reg, uint32_t bit)
   4045 {
   4046 	int i;
   4047 
   4048 	BGE_CLRBIT(sc, reg, bit);
   4049 
   4050 	for (i = 0; i < BGE_TIMEOUT; i++) {
   4051 		if ((CSR_READ_4(sc, reg) & bit) == 0)
   4052 			return;
   4053 		delay(100);
   4054 		if (sc->bge_pcie)
   4055 		  DELAY(1000);
   4056 	}
   4057 
   4058 	printf("%s: block failed to stop: reg 0x%lx, bit 0x%08x\n",
   4059 	    sc->bge_dev.dv_xname, (u_long) reg, bit);
   4060 }
   4061 
   4062 /*
   4063  * Stop the adapter and free any mbufs allocated to the
   4064  * RX and TX lists.
   4065  */
   4066 static void
   4067 bge_stop(struct bge_softc *sc)
   4068 {
   4069 	struct ifnet *ifp = &sc->ethercom.ec_if;
   4070 
   4071 	callout_stop(&sc->bge_timeout);
   4072 
   4073 	/*
   4074 	 * Disable all of the receiver blocks
   4075 	 */
   4076 	bge_stop_block(sc, BGE_RX_MODE, BGE_RXMODE_ENABLE);
   4077 	bge_stop_block(sc, BGE_RBDI_MODE, BGE_RBDIMODE_ENABLE);
   4078 	bge_stop_block(sc, BGE_RXLP_MODE, BGE_RXLPMODE_ENABLE);
   4079 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   4080 		bge_stop_block(sc, BGE_RXLS_MODE, BGE_RXLSMODE_ENABLE);
   4081 	}
   4082 	bge_stop_block(sc, BGE_RDBDI_MODE, BGE_RBDIMODE_ENABLE);
   4083 	bge_stop_block(sc, BGE_RDC_MODE, BGE_RDCMODE_ENABLE);
   4084 	bge_stop_block(sc, BGE_RBDC_MODE, BGE_RBDCMODE_ENABLE);
   4085 
   4086 	/*
   4087 	 * Disable all of the transmit blocks
   4088 	 */
   4089 	bge_stop_block(sc, BGE_SRS_MODE, BGE_SRSMODE_ENABLE);
   4090 	bge_stop_block(sc, BGE_SBDI_MODE, BGE_SBDIMODE_ENABLE);
   4091 	bge_stop_block(sc, BGE_SDI_MODE, BGE_SDIMODE_ENABLE);
   4092 	bge_stop_block(sc, BGE_RDMA_MODE, BGE_RDMAMODE_ENABLE);
   4093 	bge_stop_block(sc, BGE_SDC_MODE, BGE_SDCMODE_ENABLE);
   4094 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   4095 		bge_stop_block(sc, BGE_DMAC_MODE, BGE_DMACMODE_ENABLE);
   4096 	}
   4097 	bge_stop_block(sc, BGE_SBDC_MODE, BGE_SBDCMODE_ENABLE);
   4098 
   4099 	/*
   4100 	 * Shut down all of the memory managers and related
   4101 	 * state machines.
   4102 	 */
   4103 	bge_stop_block(sc, BGE_HCC_MODE, BGE_HCCMODE_ENABLE);
   4104 	bge_stop_block(sc, BGE_WDMA_MODE, BGE_WDMAMODE_ENABLE);
   4105 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   4106 		bge_stop_block(sc, BGE_MBCF_MODE, BGE_MBCFMODE_ENABLE);
   4107 	}
   4108 
   4109 	CSR_WRITE_4(sc, BGE_FTQ_RESET, 0xFFFFFFFF);
   4110 	CSR_WRITE_4(sc, BGE_FTQ_RESET, 0);
   4111 
   4112 	if ((sc->bge_quirks & BGE_QUIRK_5705_CORE) == 0) {
   4113 		bge_stop_block(sc, BGE_BMAN_MODE, BGE_BMANMODE_ENABLE);
   4114 		bge_stop_block(sc, BGE_MARB_MODE, BGE_MARBMODE_ENABLE);
   4115 	}
   4116 
   4117 	/* Disable host interrupts. */
   4118 	BGE_SETBIT(sc, BGE_PCI_MISC_CTL, BGE_PCIMISCCTL_MASK_PCI_INTR);
   4119 	CSR_WRITE_4(sc, BGE_MBX_IRQ0_LO, 1);
   4120 
   4121 	/*
   4122 	 * Tell firmware we're shutting down.
   4123 	 */
   4124 	BGE_CLRBIT(sc, BGE_MODE_CTL, BGE_MODECTL_STACKUP);
   4125 
   4126 	/* Free the RX lists. */
   4127 	bge_free_rx_ring_std(sc);
   4128 
   4129 	/* Free jumbo RX list. */
   4130 	bge_free_rx_ring_jumbo(sc);
   4131 
   4132 	/* Free TX buffers. */
   4133 	bge_free_tx_ring(sc);
   4134 
   4135 	/*
   4136 	 * Isolate/power down the PHY.
   4137 	 */
   4138 	if (!sc->bge_tbi)
   4139 		mii_down(&sc->bge_mii);
   4140 
   4141 	sc->bge_link = 0;
   4142 
   4143 	sc->bge_tx_saved_considx = BGE_TXCONS_UNSET;
   4144 
   4145 	ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE);
   4146 }
   4147 
   4148 /*
   4149  * Stop all chip I/O so that the kernel's probe routines don't
   4150  * get confused by errant DMAs when rebooting.
   4151  */
   4152 static void
   4153 bge_shutdown(void *xsc)
   4154 {
   4155 	struct bge_softc *sc = (struct bge_softc *)xsc;
   4156 
   4157 	bge_stop(sc);
   4158 	bge_reset(sc);
   4159 }
   4160 
   4161 
   4162 static int
   4163 sysctl_bge_verify(SYSCTLFN_ARGS)
   4164 {
   4165 	int error, t;
   4166 	struct sysctlnode node;
   4167 
   4168 	node = *rnode;
   4169 	t = *(int*)rnode->sysctl_data;
   4170 	node.sysctl_data = &t;
   4171 	error = sysctl_lookup(SYSCTLFN_CALL(&node));
   4172 	if (error || newp == NULL)
   4173 		return (error);
   4174 
   4175 #if 0
   4176 	DPRINTF2(("%s: t = %d, nodenum = %d, rnodenum = %d\n", __func__, t,
   4177 	    node.sysctl_num, rnode->sysctl_num));
   4178 #endif
   4179 
   4180 	if (node.sysctl_num == bge_rxthresh_nodenum) {
   4181 		if (t < 0 || t >= NBGE_RX_THRESH)
   4182 			return (EINVAL);
   4183 		bge_update_all_threshes(t);
   4184 	} else
   4185 		return (EINVAL);
   4186 
   4187 	*(int*)rnode->sysctl_data = t;
   4188 
   4189 	return (0);
   4190 }
   4191 
   4192 /*
   4193  * Set up sysctl(3) MIB, hw.bge.*.
   4194  *
   4195  * TBD condition SYSCTL_PERMANENT on being an LKM or not
   4196  */
   4197 SYSCTL_SETUP(sysctl_bge, "sysctl bge subtree setup")
   4198 {
   4199 	int rc, bge_root_num;
   4200 	const struct sysctlnode *node;
   4201 
   4202 	if ((rc = sysctl_createv(clog, 0, NULL, NULL,
   4203 	    CTLFLAG_PERMANENT, CTLTYPE_NODE, "hw", NULL,
   4204 	    NULL, 0, NULL, 0, CTL_HW, CTL_EOL)) != 0) {
   4205 		goto err;
   4206 	}
   4207 
   4208 	if ((rc = sysctl_createv(clog, 0, NULL, &node,
   4209 	    CTLFLAG_PERMANENT, CTLTYPE_NODE, "bge",
   4210 	    SYSCTL_DESCR("BGE interface controls"),
   4211 	    NULL, 0, NULL, 0, CTL_HW, CTL_CREATE, CTL_EOL)) != 0) {
   4212 		goto err;
   4213 	}
   4214 
   4215 	bge_root_num = node->sysctl_num;
   4216 
   4217 	/* BGE Rx interrupt mitigation level */
   4218 	if ((rc = sysctl_createv(clog, 0, NULL, &node,
   4219 	    CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   4220 	    CTLTYPE_INT, "rx_lvl",
   4221 	    SYSCTL_DESCR("BGE receive interrupt mitigation level"),
   4222 	    sysctl_bge_verify, 0,
   4223 	    &bge_rx_thresh_lvl,
   4224 	    0, CTL_HW, bge_root_num, CTL_CREATE,
   4225 	    CTL_EOL)) != 0) {
   4226 		goto err;
   4227 	}
   4228 
   4229 	bge_rxthresh_nodenum = node->sysctl_num;
   4230 
   4231 	return;
   4232 
   4233 err:
   4234 	printf("%s: sysctl_createv failed (rc = %d)\n", __func__, rc);
   4235 }
   4236 
   4237 static void
   4238 bge_powerhook(int why, void *hdl)
   4239 {
   4240 	struct bge_softc *sc = (struct bge_softc *)hdl;
   4241 	struct ifnet *ifp = &sc->ethercom.ec_if;
   4242 	struct pci_attach_args *pa = &(sc->bge_pa);
   4243 	pci_chipset_tag_t pc = pa->pa_pc;
   4244 	pcitag_t tag = pa->pa_tag;
   4245 
   4246 	switch (why) {
   4247 	case PWR_SOFTSUSPEND:
   4248 	case PWR_SOFTSTANDBY:
   4249 		bge_shutdown(sc);
   4250 		break;
   4251 	case PWR_SOFTRESUME:
   4252 		if (ifp->if_flags & IFF_UP) {
   4253 			ifp->if_flags &= ~IFF_RUNNING;
   4254 			bge_init(ifp);
   4255 		}
   4256 		break;
   4257 	case PWR_SUSPEND:
   4258 	case PWR_STANDBY:
   4259 		pci_conf_capture(pc, tag, &sc->bge_pciconf);
   4260 		break;
   4261 	case PWR_RESUME:
   4262 		pci_conf_restore(pc, tag, &sc->bge_pciconf);
   4263 		break;
   4264 	}
   4265 
   4266 	return;
   4267 }
   4268