Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 1424 insertions, 0 deletions

diff --git a/drivers/net/ethernet/i825xx/lib82596.c b/drivers/net/ethernet/i825xx/lib82596.c
new file mode 100644
index 000000000..ca2fb303f
--- /dev/null
+++ b/drivers/net/ethernet/i825xx/lib82596.c
@@ -0,0 +1,1424 @@
+/* lasi_82596.c -- driver for the intel 82596 ethernet controller, as
+   munged into HPPA boxen .
+
+   This driver is based upon 82596.c, original credits are below...
+   but there were too many hoops which HP wants jumped through to
+   keep this code in there in a sane manner.
+
+   3 primary sources of the mess --
+   1) hppa needs *lots* of cacheline flushing to keep this kind of
+   MMIO running.
+
+   2) The 82596 needs to see all of its pointers as their physical
+   address.  Thus virt_to_bus/bus_to_virt are *everywhere*.
+
+   3) The implementation HP is using seems to be significantly pickier
+   about when and how the command and RX units are started.  some
+   command ordering was changed.
+
+   Examination of the mach driver leads one to believe that there
+   might be a saner way to pull this off...  anyone who feels like a
+   full rewrite can be my guest.
+
+   Split 02/13/2000 Sam Creasey (sammy@oh.verio.com)
+
+   02/01/2000  Initial modifications for parisc by Helge Deller (deller@gmx.de)
+   03/02/2000  changes for better/correct(?) cache-flushing (deller)
+*/
+
+/* 82596.c: A generic 82596 ethernet driver for linux. */
+/*
+   Based on Apricot.c
+   Written 1994 by Mark Evans.
+   This driver is for the Apricot 82596 bus-master interface
+
+   Modularised 12/94 Mark Evans
+
+
+   Modified to support the 82596 ethernet chips on 680x0 VME boards.
+   by Richard Hirst <richard@sleepie.demon.co.uk>
+   Renamed to be 82596.c
+
+   980825:  Changed to receive directly in to sk_buffs which are
+   allocated at open() time.  Eliminates copy on incoming frames
+   (small ones are still copied).  Shared data now held in a
+   non-cached page, so we can run on 68060 in copyback mode.
+
+   TBD:
+   * look at deferring rx frames rather than discarding (as per tulip)
+   * handle tx ring full as per tulip
+   * performance test to tune rx_copybreak
+
+   Most of my modifications relate to the braindead big-endian
+   implementation by Intel.  When the i596 is operating in
+   'big-endian' mode, it thinks a 32 bit value of 0x12345678
+   should be stored as 0x56781234.  This is a real pain, when
+   you have linked lists which are shared by the 680x0 and the
+   i596.
+
+   Driver skeleton
+   Written 1993 by Donald Becker.
+   Copyright 1993 United States Government as represented by the Director,
+   National Security Agency. This software may only be used and distributed
+   according to the terms of the GNU General Public License as modified by SRC,
+   incorporated herein by reference.
+
+   The author may be reached as becker@scyld.com, or C/O
+   Scyld Computing Corporation, 410 Severn Ave., Suite 210, Annapolis MD 21403
+
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/ioport.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/types.h>
+#include <linux/bitops.h>
+#include <linux/dma-mapping.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/gfp.h>
+
+/* DEBUG flags
+ */
+
+#define DEB_INIT	0x0001
+#define DEB_PROBE	0x0002
+#define DEB_SERIOUS	0x0004
+#define DEB_ERRORS	0x0008
+#define DEB_MULTI	0x0010
+#define DEB_TDR		0x0020
+#define DEB_OPEN	0x0040
+#define DEB_RESET	0x0080
+#define DEB_ADDCMD	0x0100
+#define DEB_STATUS	0x0200
+#define DEB_STARTTX	0x0400
+#define DEB_RXADDR	0x0800
+#define DEB_TXADDR	0x1000
+#define DEB_RXFRAME	0x2000
+#define DEB_INTS	0x4000
+#define DEB_STRUCT	0x8000
+#define DEB_ANY		0xffff
+
+
+#define DEB(x, y)	if (i596_debug & (x)) { y; }
+
+
+/*
+ * The MPU_PORT command allows direct access to the 82596. With PORT access
+ * the following commands are available (p5-18). The 32-bit port command
+ * must be word-swapped with the most significant word written first.
+ * This only applies to VME boards.
+ */
+#define PORT_RESET		0x00	/* reset 82596 */
+#define PORT_SELFTEST		0x01	/* selftest */
+#define PORT_ALTSCP		0x02	/* alternate SCB address */
+#define PORT_ALTDUMP		0x03	/* Alternate DUMP address */
+
+static int i596_debug = (DEB_SERIOUS|DEB_PROBE);
+
+/* Copy frames shorter than rx_copybreak, otherwise pass on up in
+ * a full sized sk_buff.  Value of 100 stolen from tulip.c (!alpha).
+ */
+static int rx_copybreak = 100;
+
+#define PKT_BUF_SZ	1536
+#define MAX_MC_CNT	64
+
+#define ISCP_BUSY	0x0001
+
+#define I596_NULL ((u32)0xffffffff)
+
+#define CMD_EOL		0x8000	/* The last command of the list, stop. */
+#define CMD_SUSP	0x4000	/* Suspend after doing cmd. */
+#define CMD_INTR	0x2000	/* Interrupt after doing cmd. */
+
+#define CMD_FLEX	0x0008	/* Enable flexible memory model */
+
+enum commands {
+	CmdNOp = 0, CmdSASetup = 1, CmdConfigure = 2, CmdMulticastList = 3,
+	CmdTx = 4, CmdTDR = 5, CmdDump = 6, CmdDiagnose = 7
+};
+
+#define STAT_C		0x8000	/* Set to 0 after execution */
+#define STAT_B		0x4000	/* Command being executed */
+#define STAT_OK		0x2000	/* Command executed ok */
+#define STAT_A		0x1000	/* Command aborted */
+
+#define	 CUC_START	0x0100
+#define	 CUC_RESUME	0x0200
+#define	 CUC_SUSPEND    0x0300
+#define	 CUC_ABORT	0x0400
+#define	 RX_START	0x0010
+#define	 RX_RESUME	0x0020
+#define	 RX_SUSPEND	0x0030
+#define	 RX_ABORT	0x0040
+
+#define TX_TIMEOUT	(HZ/20)
+
+
+struct i596_reg {
+	unsigned short porthi;
+	unsigned short portlo;
+	u32            ca;
+};
+
+#define EOF		0x8000
+#define SIZE_MASK	0x3fff
+
+struct i596_tbd {
+	unsigned short size;
+	unsigned short pad;
+	u32            next;
+	u32            data;
+	u32 cache_pad[5];		/* Total 32 bytes... */
+};
+
+/* The command structure has two 'next' pointers; v_next is the address of
+ * the next command as seen by the CPU, b_next is the address of the next
+ * command as seen by the 82596.  The b_next pointer, as used by the 82596
+ * always references the status field of the next command, rather than the
+ * v_next field, because the 82596 is unaware of v_next.  It may seem more
+ * logical to put v_next at the end of the structure, but we cannot do that
+ * because the 82596 expects other fields to be there, depending on command
+ * type.
+ */
+
+struct i596_cmd {
+	struct i596_cmd *v_next;	/* Address from CPUs viewpoint */
+	unsigned short status;
+	unsigned short command;
+	u32            b_next;	/* Address from i596 viewpoint */
+};
+
+struct tx_cmd {
+	struct i596_cmd cmd;
+	u32            tbd;
+	unsigned short size;
+	unsigned short pad;
+	struct sk_buff *skb;		/* So we can free it after tx */
+	dma_addr_t dma_addr;
+#ifdef __LP64__
+	u32 cache_pad[6];		/* Total 64 bytes... */
+#else
+	u32 cache_pad[1];		/* Total 32 bytes... */
+#endif
+};
+
+struct tdr_cmd {
+	struct i596_cmd cmd;
+	unsigned short status;
+	unsigned short pad;
+};
+
+struct mc_cmd {
+	struct i596_cmd cmd;
+	short mc_cnt;
+	char mc_addrs[MAX_MC_CNT*6];
+};
+
+struct sa_cmd {
+	struct i596_cmd cmd;
+	char eth_addr[8];
+};
+
+struct cf_cmd {
+	struct i596_cmd cmd;
+	char i596_config[16];
+};
+
+struct i596_rfd {
+	unsigned short stat;
+	unsigned short cmd;
+	u32            b_next;	/* Address from i596 viewpoint */
+	u32            rbd;
+	unsigned short count;
+	unsigned short size;
+	struct i596_rfd *v_next;	/* Address from CPUs viewpoint */
+	struct i596_rfd *v_prev;
+#ifndef __LP64__
+	u32 cache_pad[2];		/* Total 32 bytes... */
+#endif
+};
+
+struct i596_rbd {
+	/* hardware data */
+	unsigned short count;
+	unsigned short zero1;
+	u32            b_next;
+	u32            b_data;		/* Address from i596 viewpoint */
+	unsigned short size;
+	unsigned short zero2;
+	/* driver data */
+	struct sk_buff *skb;
+	struct i596_rbd *v_next;
+	u32            b_addr;		/* This rbd addr from i596 view */
+	unsigned char *v_data;		/* Address from CPUs viewpoint */
+					/* Total 32 bytes... */
+#ifdef __LP64__
+    u32 cache_pad[4];
+#endif
+};
+
+/* These values as chosen so struct i596_dma fits in one page... */
+
+#define TX_RING_SIZE 32
+#define RX_RING_SIZE 16
+
+struct i596_scb {
+	unsigned short status;
+	unsigned short command;
+	u32           cmd;
+	u32           rfd;
+	u32           crc_err;
+	u32           align_err;
+	u32           resource_err;
+	u32           over_err;
+	u32           rcvdt_err;
+	u32           short_err;
+	unsigned short t_on;
+	unsigned short t_off;
+};
+
+struct i596_iscp {
+	u32 stat;
+	u32 scb;
+};
+
+struct i596_scp {
+	u32 sysbus;
+	u32 pad;
+	u32 iscp;
+};
+
+struct i596_dma {
+	struct i596_scp scp		        __attribute__((aligned(32)));
+	volatile struct i596_iscp iscp		__attribute__((aligned(32)));
+	volatile struct i596_scb scb		__attribute__((aligned(32)));
+	struct sa_cmd sa_cmd			__attribute__((aligned(32)));
+	struct cf_cmd cf_cmd			__attribute__((aligned(32)));
+	struct tdr_cmd tdr_cmd			__attribute__((aligned(32)));
+	struct mc_cmd mc_cmd			__attribute__((aligned(32)));
+	struct i596_rfd rfds[RX_RING_SIZE]	__attribute__((aligned(32)));
+	struct i596_rbd rbds[RX_RING_SIZE]	__attribute__((aligned(32)));
+	struct tx_cmd tx_cmds[TX_RING_SIZE]	__attribute__((aligned(32)));
+	struct i596_tbd tbds[TX_RING_SIZE]	__attribute__((aligned(32)));
+};
+
+struct i596_private {
+	struct i596_dma *dma;
+	u32    stat;
+	int last_restart;
+	struct i596_rfd *rfd_head;
+	struct i596_rbd *rbd_head;
+	struct i596_cmd *cmd_tail;
+	struct i596_cmd *cmd_head;
+	int cmd_backlog;
+	u32    last_cmd;
+	int next_tx_cmd;
+	int options;
+	spinlock_t lock;       /* serialize access to chip */
+	dma_addr_t dma_addr;
+	void __iomem *mpu_port;
+	void __iomem *ca;
+};
+
+static const char init_setup[] =
+{
+	0x8E,		/* length, prefetch on */
+	0xC8,		/* fifo to 8, monitor off */
+	0x80,		/* don't save bad frames */
+	0x2E,		/* No source address insertion, 8 byte preamble */
+	0x00,		/* priority and backoff defaults */
+	0x60,		/* interframe spacing */
+	0x00,		/* slot time LSB */
+	0xf2,		/* slot time and retries */
+	0x00,		/* promiscuous mode */
+	0x00,		/* collision detect */
+	0x40,		/* minimum frame length */
+	0xff,
+	0x00,
+	0x7f /*  *multi IA */ };
+
+static int i596_open(struct net_device *dev);
+static netdev_tx_t i596_start_xmit(struct sk_buff *skb, struct net_device *dev);
+static irqreturn_t i596_interrupt(int irq, void *dev_id);
+static int i596_close(struct net_device *dev);
+static void i596_add_cmd(struct net_device *dev, struct i596_cmd *cmd);
+static void i596_tx_timeout (struct net_device *dev, unsigned int txqueue);
+static void print_eth(unsigned char *buf, char *str);
+static void set_multicast_list(struct net_device *dev);
+static inline void ca(struct net_device *dev);
+static void mpu_port(struct net_device *dev, int c, dma_addr_t x);
+
+static int rx_ring_size = RX_RING_SIZE;
+static int ticks_limit = 100;
+static int max_cmd_backlog = TX_RING_SIZE-1;
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void i596_poll_controller(struct net_device *dev);
+#endif
+
+static inline dma_addr_t virt_to_dma(struct i596_private *lp, volatile void *v)
+{
+	return lp->dma_addr + ((unsigned long)v - (unsigned long)lp->dma);
+}
+
+#ifdef NONCOHERENT_DMA
+static inline void dma_sync_dev(struct net_device *ndev, volatile void *addr,
+		size_t len)
+{
+	dma_sync_single_for_device(ndev->dev.parent,
+			virt_to_dma(netdev_priv(ndev), addr), len,
+			DMA_BIDIRECTIONAL);
+}
+
+static inline void dma_sync_cpu(struct net_device *ndev, volatile void *addr,
+		size_t len)
+{
+	dma_sync_single_for_cpu(ndev->dev.parent,
+			virt_to_dma(netdev_priv(ndev), addr), len,
+			DMA_BIDIRECTIONAL);
+}
+#else
+static inline void dma_sync_dev(struct net_device *ndev, volatile void *addr,
+		size_t len)
+{
+}
+static inline void dma_sync_cpu(struct net_device *ndev, volatile void *addr,
+		size_t len)
+{
+}
+#endif /* NONCOHERENT_DMA */
+
+static inline int wait_istat(struct net_device *dev, struct i596_dma *dma, int delcnt, char *str)
+{
+	dma_sync_cpu(dev, &(dma->iscp), sizeof(struct i596_iscp));
+	while (--delcnt && dma->iscp.stat) {
+		udelay(10);
+		dma_sync_cpu(dev, &(dma->iscp), sizeof(struct i596_iscp));
+	}
+	if (!delcnt) {
+		printk(KERN_ERR "%s: %s, iscp.stat %04x, didn't clear\n",
+		     dev->name, str, SWAP16(dma->iscp.stat));
+		return -1;
+	} else
+		return 0;
+}
+
+
+static inline int wait_cmd(struct net_device *dev, struct i596_dma *dma, int delcnt, char *str)
+{
+	dma_sync_cpu(dev, &(dma->scb), sizeof(struct i596_scb));
+	while (--delcnt && dma->scb.command) {
+		udelay(10);
+		dma_sync_cpu(dev, &(dma->scb), sizeof(struct i596_scb));
+	}
+	if (!delcnt) {
+		printk(KERN_ERR "%s: %s, status %4.4x, cmd %4.4x.\n",
+		       dev->name, str,
+		       SWAP16(dma->scb.status),
+		       SWAP16(dma->scb.command));
+		return -1;
+	} else
+		return 0;
+}
+
+
+static void i596_display_data(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	struct i596_cmd *cmd;
+	struct i596_rfd *rfd;
+	struct i596_rbd *rbd;
+
+	printk(KERN_DEBUG "lp and scp at %p, .sysbus = %08x, .iscp = %08x\n",
+	       &dma->scp, dma->scp.sysbus, SWAP32(dma->scp.iscp));
+	printk(KERN_DEBUG "iscp at %p, iscp.stat = %08x, .scb = %08x\n",
+	       &dma->iscp, SWAP32(dma->iscp.stat), SWAP32(dma->iscp.scb));
+	printk(KERN_DEBUG "scb at %p, scb.status = %04x, .command = %04x,"
+		" .cmd = %08x, .rfd = %08x\n",
+	       &dma->scb, SWAP16(dma->scb.status), SWAP16(dma->scb.command),
+		SWAP16(dma->scb.cmd), SWAP32(dma->scb.rfd));
+	printk(KERN_DEBUG "   errors: crc %x, align %x, resource %x,"
+	       " over %x, rcvdt %x, short %x\n",
+	       SWAP32(dma->scb.crc_err), SWAP32(dma->scb.align_err),
+	       SWAP32(dma->scb.resource_err), SWAP32(dma->scb.over_err),
+	       SWAP32(dma->scb.rcvdt_err), SWAP32(dma->scb.short_err));
+	cmd = lp->cmd_head;
+	while (cmd != NULL) {
+		printk(KERN_DEBUG
+		       "cmd at %p, .status = %04x, .command = %04x,"
+		       " .b_next = %08x\n",
+		       cmd, SWAP16(cmd->status), SWAP16(cmd->command),
+		       SWAP32(cmd->b_next));
+		cmd = cmd->v_next;
+	}
+	rfd = lp->rfd_head;
+	printk(KERN_DEBUG "rfd_head = %p\n", rfd);
+	do {
+		printk(KERN_DEBUG
+		       "   %p .stat %04x, .cmd %04x, b_next %08x, rbd %08x,"
+		       " count %04x\n",
+		       rfd, SWAP16(rfd->stat), SWAP16(rfd->cmd),
+		       SWAP32(rfd->b_next), SWAP32(rfd->rbd),
+		       SWAP16(rfd->count));
+		rfd = rfd->v_next;
+	} while (rfd != lp->rfd_head);
+	rbd = lp->rbd_head;
+	printk(KERN_DEBUG "rbd_head = %p\n", rbd);
+	do {
+		printk(KERN_DEBUG
+		       "   %p .count %04x, b_next %08x, b_data %08x,"
+		       " size %04x\n",
+			rbd, SWAP16(rbd->count), SWAP32(rbd->b_next),
+		       SWAP32(rbd->b_data), SWAP16(rbd->size));
+		rbd = rbd->v_next;
+	} while (rbd != lp->rbd_head);
+	dma_sync_cpu(dev, dma, sizeof(struct i596_dma));
+}
+
+static inline int init_rx_bufs(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	int i;
+	struct i596_rfd *rfd;
+	struct i596_rbd *rbd;
+
+	/* First build the Receive Buffer Descriptor List */
+
+	for (i = 0, rbd = dma->rbds; i < rx_ring_size; i++, rbd++) {
+		dma_addr_t dma_addr;
+		struct sk_buff *skb;
+
+		skb = netdev_alloc_skb_ip_align(dev, PKT_BUF_SZ);
+		if (skb == NULL)
+			return -1;
+		dma_addr = dma_map_single(dev->dev.parent, skb->data,
+					  PKT_BUF_SZ, DMA_FROM_DEVICE);
+		rbd->v_next = rbd+1;
+		rbd->b_next = SWAP32(virt_to_dma(lp, rbd+1));
+		rbd->b_addr = SWAP32(virt_to_dma(lp, rbd));
+		rbd->skb = skb;
+		rbd->v_data = skb->data;
+		rbd->b_data = SWAP32(dma_addr);
+		rbd->size = SWAP16(PKT_BUF_SZ);
+	}
+	lp->rbd_head = dma->rbds;
+	rbd = dma->rbds + rx_ring_size - 1;
+	rbd->v_next = dma->rbds;
+	rbd->b_next = SWAP32(virt_to_dma(lp, dma->rbds));
+
+	/* Now build the Receive Frame Descriptor List */
+
+	for (i = 0, rfd = dma->rfds; i < rx_ring_size; i++, rfd++) {
+		rfd->rbd = I596_NULL;
+		rfd->v_next = rfd+1;
+		rfd->v_prev = rfd-1;
+		rfd->b_next = SWAP32(virt_to_dma(lp, rfd+1));
+		rfd->cmd = SWAP16(CMD_FLEX);
+	}
+	lp->rfd_head = dma->rfds;
+	dma->scb.rfd = SWAP32(virt_to_dma(lp, dma->rfds));
+	rfd = dma->rfds;
+	rfd->rbd = SWAP32(virt_to_dma(lp, lp->rbd_head));
+	rfd->v_prev = dma->rfds + rx_ring_size - 1;
+	rfd = dma->rfds + rx_ring_size - 1;
+	rfd->v_next = dma->rfds;
+	rfd->b_next = SWAP32(virt_to_dma(lp, dma->rfds));
+	rfd->cmd = SWAP16(CMD_EOL|CMD_FLEX);
+
+	dma_sync_dev(dev, dma, sizeof(struct i596_dma));
+	return 0;
+}
+
+static inline void remove_rx_bufs(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_rbd *rbd;
+	int i;
+
+	for (i = 0, rbd = lp->dma->rbds; i < rx_ring_size; i++, rbd++) {
+		if (rbd->skb == NULL)
+			break;
+		dma_unmap_single(dev->dev.parent,
+				 (dma_addr_t)SWAP32(rbd->b_data),
+				 PKT_BUF_SZ, DMA_FROM_DEVICE);
+		dev_kfree_skb(rbd->skb);
+	}
+}
+
+
+static void rebuild_rx_bufs(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	int i;
+
+	/* Ensure rx frame/buffer descriptors are tidy */
+
+	for (i = 0; i < rx_ring_size; i++) {
+		dma->rfds[i].rbd = I596_NULL;
+		dma->rfds[i].cmd = SWAP16(CMD_FLEX);
+	}
+	dma->rfds[rx_ring_size-1].cmd = SWAP16(CMD_EOL|CMD_FLEX);
+	lp->rfd_head = dma->rfds;
+	dma->scb.rfd = SWAP32(virt_to_dma(lp, dma->rfds));
+	lp->rbd_head = dma->rbds;
+	dma->rfds[0].rbd = SWAP32(virt_to_dma(lp, dma->rbds));
+
+	dma_sync_dev(dev, dma, sizeof(struct i596_dma));
+}
+
+
+static int init_i596_mem(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	unsigned long flags;
+
+	mpu_port(dev, PORT_RESET, 0);
+	udelay(100);			/* Wait 100us - seems to help */
+
+	/* change the scp address */
+
+	lp->last_cmd = jiffies;
+
+	dma->scp.sysbus = SYSBUS;
+	dma->scp.iscp = SWAP32(virt_to_dma(lp, &(dma->iscp)));
+	dma->iscp.scb = SWAP32(virt_to_dma(lp, &(dma->scb)));
+	dma->iscp.stat = SWAP32(ISCP_BUSY);
+	lp->cmd_backlog = 0;
+
+	lp->cmd_head = NULL;
+	dma->scb.cmd = I596_NULL;
+
+	DEB(DEB_INIT, printk(KERN_DEBUG "%s: starting i82596.\n", dev->name));
+
+	dma_sync_dev(dev, &(dma->scp), sizeof(struct i596_scp));
+	dma_sync_dev(dev, &(dma->iscp), sizeof(struct i596_iscp));
+	dma_sync_dev(dev, &(dma->scb), sizeof(struct i596_scb));
+
+	mpu_port(dev, PORT_ALTSCP, virt_to_dma(lp, &dma->scp));
+	ca(dev);
+	if (wait_istat(dev, dma, 1000, "initialization timed out"))
+		goto failed;
+	DEB(DEB_INIT, printk(KERN_DEBUG
+			     "%s: i82596 initialization successful\n",
+			     dev->name));
+
+	if (request_irq(dev->irq, i596_interrupt, 0, "i82596", dev)) {
+		printk(KERN_ERR "%s: IRQ %d not free\n", dev->name, dev->irq);
+		goto failed;
+	}
+
+	/* Ensure rx frame/buffer descriptors are tidy */
+	rebuild_rx_bufs(dev);
+
+	dma->scb.command = 0;
+	dma_sync_dev(dev, &(dma->scb), sizeof(struct i596_scb));
+
+	DEB(DEB_INIT, printk(KERN_DEBUG
+			     "%s: queuing CmdConfigure\n", dev->name));
+	memcpy(dma->cf_cmd.i596_config, init_setup, 14);
+	dma->cf_cmd.cmd.command = SWAP16(CmdConfigure);
+	dma_sync_dev(dev, &(dma->cf_cmd), sizeof(struct cf_cmd));
+	i596_add_cmd(dev, &dma->cf_cmd.cmd);
+
+	DEB(DEB_INIT, printk(KERN_DEBUG "%s: queuing CmdSASetup\n", dev->name));
+	memcpy(dma->sa_cmd.eth_addr, dev->dev_addr, ETH_ALEN);
+	dma->sa_cmd.cmd.command = SWAP16(CmdSASetup);
+	dma_sync_dev(dev, &(dma->sa_cmd), sizeof(struct sa_cmd));
+	i596_add_cmd(dev, &dma->sa_cmd.cmd);
+
+	DEB(DEB_INIT, printk(KERN_DEBUG "%s: queuing CmdTDR\n", dev->name));
+	dma->tdr_cmd.cmd.command = SWAP16(CmdTDR);
+	dma_sync_dev(dev, &(dma->tdr_cmd), sizeof(struct tdr_cmd));
+	i596_add_cmd(dev, &dma->tdr_cmd.cmd);
+
+	spin_lock_irqsave (&lp->lock, flags);
+
+	if (wait_cmd(dev, dma, 1000, "timed out waiting to issue RX_START")) {
+		spin_unlock_irqrestore (&lp->lock, flags);
+		goto failed_free_irq;
+	}
+	DEB(DEB_INIT, printk(KERN_DEBUG "%s: Issuing RX_START\n", dev->name));
+	dma->scb.command = SWAP16(RX_START);
+	dma->scb.rfd = SWAP32(virt_to_dma(lp, dma->rfds));
+	dma_sync_dev(dev, &(dma->scb), sizeof(struct i596_scb));
+
+	ca(dev);
+
+	spin_unlock_irqrestore (&lp->lock, flags);
+	if (wait_cmd(dev, dma, 1000, "RX_START not processed"))
+		goto failed_free_irq;
+	DEB(DEB_INIT, printk(KERN_DEBUG
+			     "%s: Receive unit started OK\n", dev->name));
+	return 0;
+
+failed_free_irq:
+	free_irq(dev->irq, dev);
+failed:
+	printk(KERN_ERR "%s: Failed to initialise 82596\n", dev->name);
+	mpu_port(dev, PORT_RESET, 0);
+	return -1;
+}
+
+
+static inline int i596_rx(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_rfd *rfd;
+	struct i596_rbd *rbd;
+	int frames = 0;
+
+	DEB(DEB_RXFRAME, printk(KERN_DEBUG
+				"i596_rx(), rfd_head %p, rbd_head %p\n",
+				lp->rfd_head, lp->rbd_head));
+
+
+	rfd = lp->rfd_head;		/* Ref next frame to check */
+
+	dma_sync_cpu(dev, rfd, sizeof(struct i596_rfd));
+	while (rfd->stat & SWAP16(STAT_C)) {	/* Loop while complete frames */
+		if (rfd->rbd == I596_NULL)
+			rbd = NULL;
+		else if (rfd->rbd == lp->rbd_head->b_addr) {
+			rbd = lp->rbd_head;
+			dma_sync_cpu(dev, rbd, sizeof(struct i596_rbd));
+		} else {
+			printk(KERN_ERR "%s: rbd chain broken!\n", dev->name);
+			/* XXX Now what? */
+			rbd = NULL;
+		}
+		DEB(DEB_RXFRAME, printk(KERN_DEBUG
+				      "  rfd %p, rfd.rbd %08x, rfd.stat %04x\n",
+				      rfd, rfd->rbd, rfd->stat));
+
+		if (rbd != NULL && (rfd->stat & SWAP16(STAT_OK))) {
+			/* a good frame */
+			int pkt_len = SWAP16(rbd->count) & 0x3fff;
+			struct sk_buff *skb = rbd->skb;
+			int rx_in_place = 0;
+
+			DEB(DEB_RXADDR, print_eth(rbd->v_data, "received"));
+			frames++;
+
+			/* Check if the packet is long enough to just accept
+			 * without copying to a properly sized skbuff.
+			 */
+
+			if (pkt_len > rx_copybreak) {
+				struct sk_buff *newskb;
+				dma_addr_t dma_addr;
+
+				dma_unmap_single(dev->dev.parent,
+						 (dma_addr_t)SWAP32(rbd->b_data),
+						 PKT_BUF_SZ, DMA_FROM_DEVICE);
+				/* Get fresh skbuff to replace filled one. */
+				newskb = netdev_alloc_skb_ip_align(dev,
+								   PKT_BUF_SZ);
+				if (newskb == NULL) {
+					skb = NULL;	/* drop pkt */
+					goto memory_squeeze;
+				}
+
+				/* Pass up the skb already on the Rx ring. */
+				skb_put(skb, pkt_len);
+				rx_in_place = 1;
+				rbd->skb = newskb;
+				dma_addr = dma_map_single(dev->dev.parent,
+							  newskb->data,
+							  PKT_BUF_SZ,
+							  DMA_FROM_DEVICE);
+				rbd->v_data = newskb->data;
+				rbd->b_data = SWAP32(dma_addr);
+				dma_sync_dev(dev, rbd, sizeof(struct i596_rbd));
+			} else {
+				skb = netdev_alloc_skb_ip_align(dev, pkt_len);
+			}
+memory_squeeze:
+			if (skb == NULL) {
+				/* XXX tulip.c can defer packets here!! */
+				dev->stats.rx_dropped++;
+			} else {
+				if (!rx_in_place) {
+					/* 16 byte align the data fields */
+					dma_sync_single_for_cpu(dev->dev.parent,
+								(dma_addr_t)SWAP32(rbd->b_data),
+								PKT_BUF_SZ, DMA_FROM_DEVICE);
+					skb_put_data(skb, rbd->v_data,
+						     pkt_len);
+					dma_sync_single_for_device(dev->dev.parent,
+								   (dma_addr_t)SWAP32(rbd->b_data),
+								   PKT_BUF_SZ, DMA_FROM_DEVICE);
+				}
+				skb->len = pkt_len;
+				skb->protocol = eth_type_trans(skb, dev);
+				netif_rx(skb);
+				dev->stats.rx_packets++;
+				dev->stats.rx_bytes += pkt_len;
+			}
+		} else {
+			DEB(DEB_ERRORS, printk(KERN_DEBUG
+					       "%s: Error, rfd.stat = 0x%04x\n",
+					       dev->name, rfd->stat));
+			dev->stats.rx_errors++;
+			if (rfd->stat & SWAP16(0x0100))
+				dev->stats.collisions++;
+			if (rfd->stat & SWAP16(0x8000))
+				dev->stats.rx_length_errors++;
+			if (rfd->stat & SWAP16(0x0001))
+				dev->stats.rx_over_errors++;
+			if (rfd->stat & SWAP16(0x0002))
+				dev->stats.rx_fifo_errors++;
+			if (rfd->stat & SWAP16(0x0004))
+				dev->stats.rx_frame_errors++;
+			if (rfd->stat & SWAP16(0x0008))
+				dev->stats.rx_crc_errors++;
+			if (rfd->stat & SWAP16(0x0010))
+				dev->stats.rx_length_errors++;
+		}
+
+		/* Clear the buffer descriptor count and EOF + F flags */
+
+		if (rbd != NULL && (rbd->count & SWAP16(0x4000))) {
+			rbd->count = 0;
+			lp->rbd_head = rbd->v_next;
+			dma_sync_dev(dev, rbd, sizeof(struct i596_rbd));
+		}
+
+		/* Tidy the frame descriptor, marking it as end of list */
+
+		rfd->rbd = I596_NULL;
+		rfd->stat = 0;
+		rfd->cmd = SWAP16(CMD_EOL|CMD_FLEX);
+		rfd->count = 0;
+
+		/* Update record of next frame descriptor to process */
+
+		lp->dma->scb.rfd = rfd->b_next;
+		lp->rfd_head = rfd->v_next;
+		dma_sync_dev(dev, rfd, sizeof(struct i596_rfd));
+
+		/* Remove end-of-list from old end descriptor */
+
+		rfd->v_prev->cmd = SWAP16(CMD_FLEX);
+		dma_sync_dev(dev, rfd->v_prev, sizeof(struct i596_rfd));
+		rfd = lp->rfd_head;
+		dma_sync_cpu(dev, rfd, sizeof(struct i596_rfd));
+	}
+
+	DEB(DEB_RXFRAME, printk(KERN_DEBUG "frames %d\n", frames));
+
+	return 0;
+}
+
+
+static inline void i596_cleanup_cmd(struct net_device *dev, struct i596_private *lp)
+{
+	struct i596_cmd *ptr;
+
+	while (lp->cmd_head != NULL) {
+		ptr = lp->cmd_head;
+		lp->cmd_head = ptr->v_next;
+		lp->cmd_backlog--;
+
+		switch (SWAP16(ptr->command) & 0x7) {
+		case CmdTx:
+			{
+				struct tx_cmd *tx_cmd = (struct tx_cmd *) ptr;
+				struct sk_buff *skb = tx_cmd->skb;
+				dma_unmap_single(dev->dev.parent,
+						 tx_cmd->dma_addr,
+						 skb->len, DMA_TO_DEVICE);
+
+				dev_kfree_skb(skb);
+
+				dev->stats.tx_errors++;
+				dev->stats.tx_aborted_errors++;
+
+				ptr->v_next = NULL;
+				ptr->b_next = I596_NULL;
+				tx_cmd->cmd.command = 0;  /* Mark as free */
+				break;
+			}
+		default:
+			ptr->v_next = NULL;
+			ptr->b_next = I596_NULL;
+		}
+		dma_sync_dev(dev, ptr, sizeof(struct i596_cmd));
+	}
+
+	wait_cmd(dev, lp->dma, 100, "i596_cleanup_cmd timed out");
+	lp->dma->scb.cmd = I596_NULL;
+	dma_sync_dev(dev, &(lp->dma->scb), sizeof(struct i596_scb));
+}
+
+
+static inline void i596_reset(struct net_device *dev, struct i596_private *lp)
+{
+	unsigned long flags;
+
+	DEB(DEB_RESET, printk(KERN_DEBUG "i596_reset\n"));
+
+	spin_lock_irqsave (&lp->lock, flags);
+
+	wait_cmd(dev, lp->dma, 100, "i596_reset timed out");
+
+	netif_stop_queue(dev);
+
+	/* FIXME: this command might cause an lpmc */
+	lp->dma->scb.command = SWAP16(CUC_ABORT | RX_ABORT);
+	dma_sync_dev(dev, &(lp->dma->scb), sizeof(struct i596_scb));
+	ca(dev);
+
+	/* wait for shutdown */
+	wait_cmd(dev, lp->dma, 1000, "i596_reset 2 timed out");
+	spin_unlock_irqrestore (&lp->lock, flags);
+
+	i596_cleanup_cmd(dev, lp);
+	i596_rx(dev);
+
+	netif_start_queue(dev);
+	init_i596_mem(dev);
+}
+
+
+static void i596_add_cmd(struct net_device *dev, struct i596_cmd *cmd)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	unsigned long flags;
+
+	DEB(DEB_ADDCMD, printk(KERN_DEBUG "i596_add_cmd cmd_head %p\n",
+			       lp->cmd_head));
+
+	cmd->status = 0;
+	cmd->command |= SWAP16(CMD_EOL | CMD_INTR);
+	cmd->v_next = NULL;
+	cmd->b_next = I596_NULL;
+	dma_sync_dev(dev, cmd, sizeof(struct i596_cmd));
+
+	spin_lock_irqsave (&lp->lock, flags);
+
+	if (lp->cmd_head != NULL) {
+		lp->cmd_tail->v_next = cmd;
+		lp->cmd_tail->b_next = SWAP32(virt_to_dma(lp, &cmd->status));
+		dma_sync_dev(dev, lp->cmd_tail, sizeof(struct i596_cmd));
+	} else {
+		lp->cmd_head = cmd;
+		wait_cmd(dev, dma, 100, "i596_add_cmd timed out");
+		dma->scb.cmd = SWAP32(virt_to_dma(lp, &cmd->status));
+		dma->scb.command = SWAP16(CUC_START);
+		dma_sync_dev(dev, &(dma->scb), sizeof(struct i596_scb));
+		ca(dev);
+	}
+	lp->cmd_tail = cmd;
+	lp->cmd_backlog++;
+
+	spin_unlock_irqrestore (&lp->lock, flags);
+
+	if (lp->cmd_backlog > max_cmd_backlog) {
+		unsigned long tickssofar = jiffies - lp->last_cmd;
+
+		if (tickssofar < ticks_limit)
+			return;
+
+		printk(KERN_ERR
+		       "%s: command unit timed out, status resetting.\n",
+		       dev->name);
+#if 1
+		i596_reset(dev, lp);
+#endif
+	}
+}
+
+static int i596_open(struct net_device *dev)
+{
+	DEB(DEB_OPEN, printk(KERN_DEBUG
+			     "%s: i596_open() irq %d.\n", dev->name, dev->irq));
+
+	if (init_rx_bufs(dev)) {
+		printk(KERN_ERR "%s: Failed to init rx bufs\n", dev->name);
+		return -EAGAIN;
+	}
+	if (init_i596_mem(dev)) {
+		printk(KERN_ERR "%s: Failed to init memory\n", dev->name);
+		goto out_remove_rx_bufs;
+	}
+	netif_start_queue(dev);
+
+	return 0;
+
+out_remove_rx_bufs:
+	remove_rx_bufs(dev);
+	return -EAGAIN;
+}
+
+static void i596_tx_timeout (struct net_device *dev, unsigned int txqueue)
+{
+	struct i596_private *lp = netdev_priv(dev);
+
+	/* Transmitter timeout, serious problems. */
+	DEB(DEB_ERRORS, printk(KERN_DEBUG
+			       "%s: transmit timed out, status resetting.\n",
+			       dev->name));
+
+	dev->stats.tx_errors++;
+
+	/* Try to restart the adaptor */
+	if (lp->last_restart == dev->stats.tx_packets) {
+		DEB(DEB_ERRORS, printk(KERN_DEBUG "Resetting board.\n"));
+		/* Shutdown and restart */
+		i596_reset (dev, lp);
+	} else {
+		/* Issue a channel attention signal */
+		DEB(DEB_ERRORS, printk(KERN_DEBUG "Kicking board.\n"));
+		lp->dma->scb.command = SWAP16(CUC_START | RX_START);
+		dma_sync_dev(dev, &(lp->dma->scb), sizeof(struct i596_scb));
+		ca (dev);
+		lp->last_restart = dev->stats.tx_packets;
+	}
+
+	netif_trans_update(dev); /* prevent tx timeout */
+	netif_wake_queue (dev);
+}
+
+
+static netdev_tx_t i596_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct tx_cmd *tx_cmd;
+	struct i596_tbd *tbd;
+	short length = skb->len;
+
+	DEB(DEB_STARTTX, printk(KERN_DEBUG
+				"%s: i596_start_xmit(%x,%p) called\n",
+				dev->name, skb->len, skb->data));
+
+	if (length < ETH_ZLEN) {
+		if (skb_padto(skb, ETH_ZLEN))
+			return NETDEV_TX_OK;
+		length = ETH_ZLEN;
+	}
+
+	netif_stop_queue(dev);
+
+	tx_cmd = lp->dma->tx_cmds + lp->next_tx_cmd;
+	tbd = lp->dma->tbds + lp->next_tx_cmd;
+
+	if (tx_cmd->cmd.command) {
+		DEB(DEB_ERRORS, printk(KERN_DEBUG
+				       "%s: xmit ring full, dropping packet.\n",
+				       dev->name));
+		dev->stats.tx_dropped++;
+
+		dev_kfree_skb_any(skb);
+	} else {
+		if (++lp->next_tx_cmd == TX_RING_SIZE)
+			lp->next_tx_cmd = 0;
+		tx_cmd->tbd = SWAP32(virt_to_dma(lp, tbd));
+		tbd->next = I596_NULL;
+
+		tx_cmd->cmd.command = SWAP16(CMD_FLEX | CmdTx);
+		tx_cmd->skb = skb;
+
+		tx_cmd->pad = 0;
+		tx_cmd->size = 0;
+		tbd->pad = 0;
+		tbd->size = SWAP16(EOF | length);
+
+		tx_cmd->dma_addr = dma_map_single(dev->dev.parent, skb->data,
+						  skb->len, DMA_TO_DEVICE);
+		tbd->data = SWAP32(tx_cmd->dma_addr);
+
+		DEB(DEB_TXADDR, print_eth(skb->data, "tx-queued"));
+		dma_sync_dev(dev, tx_cmd, sizeof(struct tx_cmd));
+		dma_sync_dev(dev, tbd, sizeof(struct i596_tbd));
+		i596_add_cmd(dev, &tx_cmd->cmd);
+
+		dev->stats.tx_packets++;
+		dev->stats.tx_bytes += length;
+	}
+
+	netif_start_queue(dev);
+
+	return NETDEV_TX_OK;
+}
+
+static void print_eth(unsigned char *add, char *str)
+{
+	printk(KERN_DEBUG "i596 0x%p, %pM --> %pM %02X%02X, %s\n",
+	       add, add + 6, add, add[12], add[13], str);
+}
+static const struct net_device_ops i596_netdev_ops = {
+	.ndo_open		= i596_open,
+	.ndo_stop		= i596_close,
+	.ndo_start_xmit		= i596_start_xmit,
+	.ndo_set_rx_mode	= set_multicast_list,
+	.ndo_tx_timeout		= i596_tx_timeout,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_set_mac_address	= eth_mac_addr,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+	.ndo_poll_controller	= i596_poll_controller,
+#endif
+};
+
+static int i82596_probe(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	int ret;
+
+	/* This lot is ensure things have been cache line aligned. */
+	BUILD_BUG_ON(sizeof(struct i596_rfd) != 32);
+	BUILD_BUG_ON(sizeof(struct i596_rbd) &  31);
+	BUILD_BUG_ON(sizeof(struct tx_cmd)   &  31);
+	BUILD_BUG_ON(sizeof(struct i596_tbd) != 32);
+#ifndef __LP64__
+	BUILD_BUG_ON(sizeof(struct i596_dma) > 4096);
+#endif
+
+	if (!dev->base_addr || !dev->irq)
+		return -ENODEV;
+
+	dev->netdev_ops = &i596_netdev_ops;
+	dev->watchdog_timeo = TX_TIMEOUT;
+
+	memset(lp->dma, 0, sizeof(struct i596_dma));
+	lp->dma->scb.command = 0;
+	lp->dma->scb.cmd = I596_NULL;
+	lp->dma->scb.rfd = I596_NULL;
+	spin_lock_init(&lp->lock);
+
+	dma_sync_dev(dev, lp->dma, sizeof(struct i596_dma));
+
+	ret = register_netdev(dev);
+	if (ret)
+		return ret;
+
+	DEB(DEB_PROBE, printk(KERN_INFO "%s: 82596 at %#3lx, %pM IRQ %d.\n",
+			      dev->name, dev->base_addr, dev->dev_addr,
+			      dev->irq));
+	DEB(DEB_INIT, printk(KERN_INFO
+			     "%s: dma at 0x%p (%d bytes), lp->scb at 0x%p\n",
+			     dev->name, lp->dma, (int)sizeof(struct i596_dma),
+			     &lp->dma->scb));
+
+	return 0;
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void i596_poll_controller(struct net_device *dev)
+{
+	disable_irq(dev->irq);
+	i596_interrupt(dev->irq, dev);
+	enable_irq(dev->irq);
+}
+#endif
+
+static irqreturn_t i596_interrupt(int irq, void *dev_id)
+{
+	struct net_device *dev = dev_id;
+	struct i596_private *lp;
+	struct i596_dma *dma;
+	unsigned short status, ack_cmd = 0;
+
+	lp = netdev_priv(dev);
+	dma = lp->dma;
+
+	spin_lock (&lp->lock);
+
+	wait_cmd(dev, dma, 100, "i596 interrupt, timeout");
+	status = SWAP16(dma->scb.status);
+
+	DEB(DEB_INTS, printk(KERN_DEBUG
+			     "%s: i596 interrupt, IRQ %d, status %4.4x.\n",
+			dev->name, dev->irq, status));
+
+	ack_cmd = status & 0xf000;
+
+	if (!ack_cmd) {
+		DEB(DEB_ERRORS, printk(KERN_DEBUG
+				       "%s: interrupt with no events\n",
+				       dev->name));
+		spin_unlock (&lp->lock);
+		return IRQ_NONE;
+	}
+
+	if ((status & 0x8000) || (status & 0x2000)) {
+		struct i596_cmd *ptr;
+
+		if ((status & 0x8000))
+			DEB(DEB_INTS,
+			    printk(KERN_DEBUG
+				   "%s: i596 interrupt completed command.\n",
+				   dev->name));
+		if ((status & 0x2000))
+			DEB(DEB_INTS,
+			    printk(KERN_DEBUG
+				   "%s: i596 interrupt command unit inactive %x.\n",
+				   dev->name, status & 0x0700));
+
+		while (lp->cmd_head != NULL) {
+			dma_sync_cpu(dev, lp->cmd_head, sizeof(struct i596_cmd));
+			if (!(lp->cmd_head->status & SWAP16(STAT_C)))
+				break;
+
+			ptr = lp->cmd_head;
+
+			DEB(DEB_STATUS,
+			    printk(KERN_DEBUG
+				   "cmd_head->status = %04x, ->command = %04x\n",
+				   SWAP16(lp->cmd_head->status),
+				   SWAP16(lp->cmd_head->command)));
+			lp->cmd_head = ptr->v_next;
+			lp->cmd_backlog--;
+
+			switch (SWAP16(ptr->command) & 0x7) {
+			case CmdTx:
+			    {
+				struct tx_cmd *tx_cmd = (struct tx_cmd *) ptr;
+				struct sk_buff *skb = tx_cmd->skb;
+
+				if (ptr->status & SWAP16(STAT_OK)) {
+					DEB(DEB_TXADDR,
+					    print_eth(skb->data, "tx-done"));
+				} else {
+					dev->stats.tx_errors++;
+					if (ptr->status & SWAP16(0x0020))
+						dev->stats.collisions++;
+					if (!(ptr->status & SWAP16(0x0040)))
+						dev->stats.tx_heartbeat_errors++;
+					if (ptr->status & SWAP16(0x0400))
+						dev->stats.tx_carrier_errors++;
+					if (ptr->status & SWAP16(0x0800))
+						dev->stats.collisions++;
+					if (ptr->status & SWAP16(0x1000))
+						dev->stats.tx_aborted_errors++;
+				}
+				dma_unmap_single(dev->dev.parent,
+						 tx_cmd->dma_addr,
+						 skb->len, DMA_TO_DEVICE);
+				dev_consume_skb_irq(skb);
+
+				tx_cmd->cmd.command = 0; /* Mark free */
+				break;
+			    }
+			case CmdTDR:
+			    {
+				unsigned short status = SWAP16(((struct tdr_cmd *)ptr)->status);
+
+				if (status & 0x8000) {
+					DEB(DEB_ANY,
+					    printk(KERN_DEBUG "%s: link ok.\n",
+						   dev->name));
+				} else {
+					if (status & 0x4000)
+						printk(KERN_ERR
+						       "%s: Transceiver problem.\n",
+						       dev->name);
+					if (status & 0x2000)
+						printk(KERN_ERR
+						       "%s: Termination problem.\n",
+						       dev->name);
+					if (status & 0x1000)
+						printk(KERN_ERR
+						       "%s: Short circuit.\n",
+						       dev->name);
+
+					DEB(DEB_TDR,
+					    printk(KERN_DEBUG "%s: Time %d.\n",
+						   dev->name, status & 0x07ff));
+				}
+				break;
+			    }
+			case CmdConfigure:
+				/*
+				 * Zap command so set_multicast_list() know
+				 * it is free
+				 */
+				ptr->command = 0;
+				break;
+			}
+			ptr->v_next = NULL;
+			ptr->b_next = I596_NULL;
+			dma_sync_dev(dev, ptr, sizeof(struct i596_cmd));
+			lp->last_cmd = jiffies;
+		}
+
+		/* This mess is arranging that only the last of any outstanding
+		 * commands has the interrupt bit set.  Should probably really
+		 * only add to the cmd queue when the CU is stopped.
+		 */
+		ptr = lp->cmd_head;
+		while ((ptr != NULL) && (ptr != lp->cmd_tail)) {
+			struct i596_cmd *prev = ptr;
+
+			ptr->command &= SWAP16(0x1fff);
+			ptr = ptr->v_next;
+			dma_sync_dev(dev, prev, sizeof(struct i596_cmd));
+		}
+
+		if (lp->cmd_head != NULL)
+			ack_cmd |= CUC_START;
+		dma->scb.cmd = SWAP32(virt_to_dma(lp, &lp->cmd_head->status));
+		dma_sync_dev(dev, &dma->scb, sizeof(struct i596_scb));
+	}
+	if ((status & 0x1000) || (status & 0x4000)) {
+		if ((status & 0x4000))
+			DEB(DEB_INTS,
+			    printk(KERN_DEBUG
+				   "%s: i596 interrupt received a frame.\n",
+				   dev->name));
+		i596_rx(dev);
+		/* Only RX_START if stopped - RGH 07-07-96 */
+		if (status & 0x1000) {
+			if (netif_running(dev)) {
+				DEB(DEB_ERRORS,
+				    printk(KERN_DEBUG
+					   "%s: i596 interrupt receive unit inactive, status 0x%x\n",
+					   dev->name, status));
+				ack_cmd |= RX_START;
+				dev->stats.rx_errors++;
+				dev->stats.rx_fifo_errors++;
+				rebuild_rx_bufs(dev);
+			}
+		}
+	}
+	wait_cmd(dev, dma, 100, "i596 interrupt, timeout");
+	dma->scb.command = SWAP16(ack_cmd);
+	dma_sync_dev(dev, &dma->scb, sizeof(struct i596_scb));
+
+	/* DANGER: I suspect that some kind of interrupt
+	 acknowledgement aside from acking the 82596 might be needed
+	 here...  but it's running acceptably without */
+
+	ca(dev);
+
+	wait_cmd(dev, dma, 100, "i596 interrupt, exit timeout");
+	DEB(DEB_INTS, printk(KERN_DEBUG "%s: exiting interrupt.\n", dev->name));
+
+	spin_unlock (&lp->lock);
+	return IRQ_HANDLED;
+}
+
+static int i596_close(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	unsigned long flags;
+
+	netif_stop_queue(dev);
+
+	DEB(DEB_INIT,
+	    printk(KERN_DEBUG
+		   "%s: Shutting down ethercard, status was %4.4x.\n",
+		   dev->name, SWAP16(lp->dma->scb.status)));
+
+	spin_lock_irqsave(&lp->lock, flags);
+
+	wait_cmd(dev, lp->dma, 100, "close1 timed out");
+	lp->dma->scb.command = SWAP16(CUC_ABORT | RX_ABORT);
+	dma_sync_dev(dev, &lp->dma->scb, sizeof(struct i596_scb));
+
+	ca(dev);
+
+	wait_cmd(dev, lp->dma, 100, "close2 timed out");
+	spin_unlock_irqrestore(&lp->lock, flags);
+	DEB(DEB_STRUCT, i596_display_data(dev));
+	i596_cleanup_cmd(dev, lp);
+
+	free_irq(dev->irq, dev);
+	remove_rx_bufs(dev);
+
+	return 0;
+}
+
+/*
+ *    Set or clear the multicast filter for this adaptor.
+ */
+
+static void set_multicast_list(struct net_device *dev)
+{
+	struct i596_private *lp = netdev_priv(dev);
+	struct i596_dma *dma = lp->dma;
+	int config = 0, cnt;
+
+	DEB(DEB_MULTI,
+	    printk(KERN_DEBUG
+		   "%s: set multicast list, %d entries, promisc %s, allmulti %s\n",
+		   dev->name, netdev_mc_count(dev),
+		   dev->flags & IFF_PROMISC ? "ON" : "OFF",
+		   dev->flags & IFF_ALLMULTI ? "ON" : "OFF"));
+
+	if ((dev->flags & IFF_PROMISC) &&
+	    !(dma->cf_cmd.i596_config[8] & 0x01)) {
+		dma->cf_cmd.i596_config[8] |= 0x01;
+		config = 1;
+	}
+	if (!(dev->flags & IFF_PROMISC) &&
+	    (dma->cf_cmd.i596_config[8] & 0x01)) {
+		dma->cf_cmd.i596_config[8] &= ~0x01;
+		config = 1;
+	}
+	if ((dev->flags & IFF_ALLMULTI) &&
+	    (dma->cf_cmd.i596_config[11] & 0x20)) {
+		dma->cf_cmd.i596_config[11] &= ~0x20;
+		config = 1;
+	}
+	if (!(dev->flags & IFF_ALLMULTI) &&
+	    !(dma->cf_cmd.i596_config[11] & 0x20)) {
+		dma->cf_cmd.i596_config[11] |= 0x20;
+		config = 1;
+	}
+	if (config) {
+		if (dma->cf_cmd.cmd.command)
+			printk(KERN_INFO
+			       "%s: config change request already queued\n",
+			       dev->name);
+		else {
+			dma->cf_cmd.cmd.command = SWAP16(CmdConfigure);
+			dma_sync_dev(dev, &dma->cf_cmd, sizeof(struct cf_cmd));
+			i596_add_cmd(dev, &dma->cf_cmd.cmd);
+		}
+	}
+
+	cnt = netdev_mc_count(dev);
+	if (cnt > MAX_MC_CNT) {
+		cnt = MAX_MC_CNT;
+		printk(KERN_NOTICE "%s: Only %d multicast addresses supported",
+			dev->name, cnt);
+	}
+
+	if (!netdev_mc_empty(dev)) {
+		struct netdev_hw_addr *ha;
+		unsigned char *cp;
+		struct mc_cmd *cmd;
+
+		cmd = &dma->mc_cmd;
+		cmd->cmd.command = SWAP16(CmdMulticastList);
+		cmd->mc_cnt = SWAP16(netdev_mc_count(dev) * 6);
+		cp = cmd->mc_addrs;
+		netdev_for_each_mc_addr(ha, dev) {
+			if (!cnt--)
+				break;
+			memcpy(cp, ha->addr, ETH_ALEN);
+			if (i596_debug > 1)
+				DEB(DEB_MULTI,
+				    printk(KERN_DEBUG
+					   "%s: Adding address %pM\n",
+					   dev->name, cp));
+			cp += ETH_ALEN;
+		}
+		dma_sync_dev(dev, &dma->mc_cmd, sizeof(struct mc_cmd));
+		i596_add_cmd(dev, &cmd->cmd);
+	}
+}