* Copyright (C) 2001 Kyle A. Lucke (klucke@us.ibm.com), IBM Corp.
* Substantially cleaned up by:
* Copyright (C) 2003 David Gibson <dwg@au1.ibm.com>, IBM Corporation.
+ * Copyright (C) 2004-2005 Michael Ellerman, IBM Corporation.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* number of packets outstanding to a remote partition at a time.
*/
-#include <linux/config.h>
#include <linux/module.h>
-#include <linux/version.h>
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/ioport.h>
#include <linux/delay.h>
#include <linux/mm.h>
#include <linux/ethtool.h>
-#include <asm/iSeries/mf.h>
-#include <asm/iSeries/iSeries_pci.h>
+#include <linux/if_ether.h>
+
+#include <asm/abs_addr.h>
+#include <asm/iseries/mf.h>
#include <asm/uaccess.h>
-#include <asm/iSeries/HvLpConfig.h>
-#include <asm/iSeries/HvTypes.h>
-#include <asm/iSeries/HvLpEvent.h>
+#include <asm/iseries/hv_lp_config.h>
+#include <asm/iseries/hv_types.h>
+#include <asm/iseries/hv_lp_event.h>
#include <asm/iommu.h>
#include <asm/vio.h>
#undef DEBUG
-#include "iseries_veth.h"
-
MODULE_AUTHOR("Kyle Lucke <klucke@us.ibm.com>");
MODULE_DESCRIPTION("iSeries Virtual ethernet driver");
MODULE_LICENSE("GPL");
+#define VETH_EVENT_CAP (0)
+#define VETH_EVENT_FRAMES (1)
+#define VETH_EVENT_MONITOR (2)
+#define VETH_EVENT_FRAMES_ACK (3)
+
+#define VETH_MAX_ACKS_PER_MSG (20)
+#define VETH_MAX_FRAMES_PER_MSG (6)
+
+struct veth_frames_data {
+ u32 addr[VETH_MAX_FRAMES_PER_MSG];
+ u16 len[VETH_MAX_FRAMES_PER_MSG];
+ u32 eofmask;
+};
+#define VETH_EOF_SHIFT (32-VETH_MAX_FRAMES_PER_MSG)
+
+struct veth_frames_ack_data {
+ u16 token[VETH_MAX_ACKS_PER_MSG];
+};
+
+struct veth_cap_data {
+ u8 caps_version;
+ u8 rsvd1;
+ u16 num_buffers;
+ u16 ack_threshold;
+ u16 rsvd2;
+ u32 ack_timeout;
+ u32 rsvd3;
+ u64 rsvd4[3];
+};
+
+struct veth_lpevent {
+ struct HvLpEvent base_event;
+ union {
+ struct veth_cap_data caps_data;
+ struct veth_frames_data frames_data;
+ struct veth_frames_ack_data frames_ack_data;
+ } u;
+
+};
+
+#define DRV_NAME "iseries_veth"
+#define DRV_VERSION "2.0"
+
#define VETH_NUMBUFFERS (120)
#define VETH_ACKTIMEOUT (1000000) /* microseconds */
#define VETH_MAX_MCAST (12)
struct veth_msg {
struct veth_msg *next;
- struct VethFramesData data;
+ struct veth_frames_data data;
int token;
- unsigned long in_use;
+ int in_use;
struct sk_buff *skb;
struct device *dev;
};
struct work_struct statemachine_wq;
struct veth_msg *msgs;
int num_events;
- struct VethCapData local_caps;
+ struct veth_cap_data local_caps;
+ struct kobject kobject;
struct timer_list ack_timer;
+ struct timer_list reset_timer;
+ unsigned int reset_timeout;
+ unsigned long last_contact;
+ int outstanding_tx;
+
spinlock_t lock;
unsigned long state;
HvLpInstanceId src_inst;
HvLpInstanceId dst_inst;
- struct VethLpEvent cap_event, cap_ack_event;
+ struct veth_lpevent cap_event, cap_ack_event;
u16 pending_acks[VETH_MAX_ACKS_PER_MSG];
u32 num_pending_acks;
int num_ack_events;
- struct VethCapData remote_caps;
+ struct veth_cap_data remote_caps;
u32 ack_timeout;
- spinlock_t msg_stack_lock;
struct veth_msg *msg_stack_head;
};
u64 mac_addr;
HvLpIndexMap lpar_map;
- spinlock_t pending_gate;
- struct sk_buff *pending_skb;
- HvLpIndexMap pending_lpmask;
+ /* queue_lock protects the stopped_map and dev's queue. */
+ spinlock_t queue_lock;
+ HvLpIndexMap stopped_map;
+ /* mcast_gate protects promiscuous, num_mcast & mcast_addr. */
rwlock_t mcast_gate;
int promiscuous;
- int all_mcast;
int num_mcast;
u64 mcast_addr[VETH_MAX_MCAST];
+
+ struct kobject kobject;
};
static HvLpIndex this_lp;
static int veth_start_xmit(struct sk_buff *skb, struct net_device *dev);
static void veth_recycle_msg(struct veth_lpar_connection *, struct veth_msg *);
-static void veth_flush_pending(struct veth_lpar_connection *cnx);
-static void veth_receive(struct veth_lpar_connection *, struct VethLpEvent *);
-static void veth_timed_ack(unsigned long connectionPtr);
+static void veth_wake_queues(struct veth_lpar_connection *cnx);
+static void veth_stop_queues(struct veth_lpar_connection *cnx);
+static void veth_receive(struct veth_lpar_connection *, struct veth_lpevent *);
+static void veth_release_connection(struct kobject *kobject);
+static void veth_timed_ack(unsigned long ptr);
+static void veth_timed_reset(unsigned long ptr);
/*
* Utility functions
*/
#define veth_info(fmt, args...) \
- printk(KERN_INFO "iseries_veth: " fmt, ## args)
+ printk(KERN_INFO DRV_NAME ": " fmt, ## args)
#define veth_error(fmt, args...) \
- printk(KERN_ERR "iseries_veth: Error: " fmt, ## args)
+ printk(KERN_ERR DRV_NAME ": Error: " fmt, ## args)
#ifdef DEBUG
#define veth_debug(fmt, args...) \
- printk(KERN_DEBUG "iseries_veth: " fmt, ## args)
+ printk(KERN_DEBUG DRV_NAME ": " fmt, ## args)
#else
#define veth_debug(fmt, args...) do {} while (0)
#endif
+/* You must hold the connection's lock when you call this function. */
static inline void veth_stack_push(struct veth_lpar_connection *cnx,
struct veth_msg *msg)
{
- unsigned long flags;
-
- spin_lock_irqsave(&cnx->msg_stack_lock, flags);
msg->next = cnx->msg_stack_head;
cnx->msg_stack_head = msg;
- spin_unlock_irqrestore(&cnx->msg_stack_lock, flags);
}
+/* You must hold the connection's lock when you call this function. */
static inline struct veth_msg *veth_stack_pop(struct veth_lpar_connection *cnx)
{
- unsigned long flags;
struct veth_msg *msg;
- spin_lock_irqsave(&cnx->msg_stack_lock, flags);
msg = cnx->msg_stack_head;
if (msg)
cnx->msg_stack_head = cnx->msg_stack_head->next;
- spin_unlock_irqrestore(&cnx->msg_stack_lock, flags);
+
return msg;
}
+/* You must hold the connection's lock when you call this function. */
+static inline int veth_stack_is_empty(struct veth_lpar_connection *cnx)
+{
+ return cnx->msg_stack_head == NULL;
+}
+
static inline HvLpEvent_Rc
veth_signalevent(struct veth_lpar_connection *cnx, u16 subtype,
HvLpEvent_AckInd ackind, HvLpEvent_AckType acktype,
struct veth_allocation vc = { COMPLETION_INITIALIZER(vc.c), 0 };
mf_allocate_lp_events(rlp, HvLpEvent_Type_VirtualLan,
- sizeof(struct VethLpEvent), number,
+ sizeof(struct veth_lpevent), number,
&veth_complete_allocation, &vc);
wait_for_completion(&vc.c);
return vc.num;
}
+/*
+ * sysfs support
+ */
+
+struct veth_cnx_attribute {
+ struct attribute attr;
+ ssize_t (*show)(struct veth_lpar_connection *, char *buf);
+ ssize_t (*store)(struct veth_lpar_connection *, const char *buf);
+};
+
+static ssize_t veth_cnx_attribute_show(struct kobject *kobj,
+ struct attribute *attr, char *buf)
+{
+ struct veth_cnx_attribute *cnx_attr;
+ struct veth_lpar_connection *cnx;
+
+ cnx_attr = container_of(attr, struct veth_cnx_attribute, attr);
+ cnx = container_of(kobj, struct veth_lpar_connection, kobject);
+
+ if (!cnx_attr->show)
+ return -EIO;
+
+ return cnx_attr->show(cnx, buf);
+}
+
+#define CUSTOM_CNX_ATTR(_name, _format, _expression) \
+static ssize_t _name##_show(struct veth_lpar_connection *cnx, char *buf)\
+{ \
+ return sprintf(buf, _format, _expression); \
+} \
+struct veth_cnx_attribute veth_cnx_attr_##_name = __ATTR_RO(_name)
+
+#define SIMPLE_CNX_ATTR(_name) \
+ CUSTOM_CNX_ATTR(_name, "%lu\n", (unsigned long)cnx->_name)
+
+SIMPLE_CNX_ATTR(outstanding_tx);
+SIMPLE_CNX_ATTR(remote_lp);
+SIMPLE_CNX_ATTR(num_events);
+SIMPLE_CNX_ATTR(src_inst);
+SIMPLE_CNX_ATTR(dst_inst);
+SIMPLE_CNX_ATTR(num_pending_acks);
+SIMPLE_CNX_ATTR(num_ack_events);
+CUSTOM_CNX_ATTR(ack_timeout, "%d\n", jiffies_to_msecs(cnx->ack_timeout));
+CUSTOM_CNX_ATTR(reset_timeout, "%d\n", jiffies_to_msecs(cnx->reset_timeout));
+CUSTOM_CNX_ATTR(state, "0x%.4lX\n", cnx->state);
+CUSTOM_CNX_ATTR(last_contact, "%d\n", cnx->last_contact ?
+ jiffies_to_msecs(jiffies - cnx->last_contact) : 0);
+
+#define GET_CNX_ATTR(_name) (&veth_cnx_attr_##_name.attr)
+
+static struct attribute *veth_cnx_default_attrs[] = {
+ GET_CNX_ATTR(outstanding_tx),
+ GET_CNX_ATTR(remote_lp),
+ GET_CNX_ATTR(num_events),
+ GET_CNX_ATTR(reset_timeout),
+ GET_CNX_ATTR(last_contact),
+ GET_CNX_ATTR(state),
+ GET_CNX_ATTR(src_inst),
+ GET_CNX_ATTR(dst_inst),
+ GET_CNX_ATTR(num_pending_acks),
+ GET_CNX_ATTR(num_ack_events),
+ GET_CNX_ATTR(ack_timeout),
+ NULL
+};
+
+static struct sysfs_ops veth_cnx_sysfs_ops = {
+ .show = veth_cnx_attribute_show
+};
+
+static struct kobj_type veth_lpar_connection_ktype = {
+ .release = veth_release_connection,
+ .sysfs_ops = &veth_cnx_sysfs_ops,
+ .default_attrs = veth_cnx_default_attrs
+};
+
+struct veth_port_attribute {
+ struct attribute attr;
+ ssize_t (*show)(struct veth_port *, char *buf);
+ ssize_t (*store)(struct veth_port *, const char *buf);
+};
+
+static ssize_t veth_port_attribute_show(struct kobject *kobj,
+ struct attribute *attr, char *buf)
+{
+ struct veth_port_attribute *port_attr;
+ struct veth_port *port;
+
+ port_attr = container_of(attr, struct veth_port_attribute, attr);
+ port = container_of(kobj, struct veth_port, kobject);
+
+ if (!port_attr->show)
+ return -EIO;
+
+ return port_attr->show(port, buf);
+}
+
+#define CUSTOM_PORT_ATTR(_name, _format, _expression) \
+static ssize_t _name##_show(struct veth_port *port, char *buf) \
+{ \
+ return sprintf(buf, _format, _expression); \
+} \
+struct veth_port_attribute veth_port_attr_##_name = __ATTR_RO(_name)
+
+#define SIMPLE_PORT_ATTR(_name) \
+ CUSTOM_PORT_ATTR(_name, "%lu\n", (unsigned long)port->_name)
+
+SIMPLE_PORT_ATTR(promiscuous);
+SIMPLE_PORT_ATTR(num_mcast);
+CUSTOM_PORT_ATTR(lpar_map, "0x%X\n", port->lpar_map);
+CUSTOM_PORT_ATTR(stopped_map, "0x%X\n", port->stopped_map);
+CUSTOM_PORT_ATTR(mac_addr, "0x%lX\n", port->mac_addr);
+
+#define GET_PORT_ATTR(_name) (&veth_port_attr_##_name.attr)
+static struct attribute *veth_port_default_attrs[] = {
+ GET_PORT_ATTR(mac_addr),
+ GET_PORT_ATTR(lpar_map),
+ GET_PORT_ATTR(stopped_map),
+ GET_PORT_ATTR(promiscuous),
+ GET_PORT_ATTR(num_mcast),
+ NULL
+};
+
+static struct sysfs_ops veth_port_sysfs_ops = {
+ .show = veth_port_attribute_show
+};
+
+static struct kobj_type veth_port_ktype = {
+ .sysfs_ops = &veth_port_sysfs_ops,
+ .default_attrs = veth_port_default_attrs
+};
+
/*
* LPAR connection code
*/
}
static void veth_take_cap(struct veth_lpar_connection *cnx,
- struct VethLpEvent *event)
+ struct veth_lpevent *event)
{
unsigned long flags;
}
static void veth_take_cap_ack(struct veth_lpar_connection *cnx,
- struct VethLpEvent *event)
+ struct veth_lpevent *event)
{
unsigned long flags;
}
static void veth_take_monitor_ack(struct veth_lpar_connection *cnx,
- struct VethLpEvent *event)
+ struct veth_lpevent *event)
{
unsigned long flags;
spin_lock_irqsave(&cnx->lock, flags);
veth_debug("cnx %d: lost connection.\n", cnx->remote_lp);
- cnx->state |= VETH_STATE_RESET;
- veth_kick_statemachine(cnx);
+
+ /* Avoid kicking the statemachine once we're shutdown.
+ * It's unnecessary and it could break veth_stop_connection(). */
+
+ if (! (cnx->state & VETH_STATE_SHUTDOWN)) {
+ cnx->state |= VETH_STATE_RESET;
+ veth_kick_statemachine(cnx);
+ }
spin_unlock_irqrestore(&cnx->lock, flags);
}
-static void veth_handle_ack(struct VethLpEvent *event)
+static void veth_handle_ack(struct veth_lpevent *event)
{
HvLpIndex rlp = event->base_event.xTargetLp;
struct veth_lpar_connection *cnx = veth_cnx[rlp];
BUG_ON(! cnx);
switch (event->base_event.xSubtype) {
- case VethEventTypeCap:
+ case VETH_EVENT_CAP:
veth_take_cap_ack(cnx, event);
break;
- case VethEventTypeMonitor:
+ case VETH_EVENT_MONITOR:
veth_take_monitor_ack(cnx, event);
break;
default:
};
}
-static void veth_handle_int(struct VethLpEvent *event)
+static void veth_handle_int(struct veth_lpevent *event)
{
HvLpIndex rlp = event->base_event.xSourceLp;
struct veth_lpar_connection *cnx = veth_cnx[rlp];
unsigned long flags;
- int i;
+ int i, acked = 0;
BUG_ON(! cnx);
switch (event->base_event.xSubtype) {
- case VethEventTypeCap:
+ case VETH_EVENT_CAP:
veth_take_cap(cnx, event);
break;
- case VethEventTypeMonitor:
+ case VETH_EVENT_MONITOR:
/* do nothing... this'll hang out here til we're dead,
* and the hypervisor will return it for us. */
break;
- case VethEventTypeFramesAck:
+ case VETH_EVENT_FRAMES_ACK:
spin_lock_irqsave(&cnx->lock, flags);
+
for (i = 0; i < VETH_MAX_ACKS_PER_MSG; ++i) {
u16 msgnum = event->u.frames_ack_data.token[i];
- if (msgnum < VETH_NUMBUFFERS)
+ if (msgnum < VETH_NUMBUFFERS) {
veth_recycle_msg(cnx, cnx->msgs + msgnum);
+ cnx->outstanding_tx--;
+ acked++;
+ }
+ }
+
+ if (acked > 0) {
+ cnx->last_contact = jiffies;
+ veth_wake_queues(cnx);
}
+
spin_unlock_irqrestore(&cnx->lock, flags);
- veth_flush_pending(cnx);
break;
- case VethEventTypeFrames:
+ case VETH_EVENT_FRAMES:
veth_receive(cnx, event);
break;
default:
static void veth_handle_event(struct HvLpEvent *event, struct pt_regs *regs)
{
- struct VethLpEvent *veth_event = (struct VethLpEvent *)event;
+ struct veth_lpevent *veth_event = (struct veth_lpevent *)event;
- if (event->xFlags.xFunction == HvLpEvent_Function_Ack)
+ if (hvlpevent_is_ack(event))
veth_handle_ack(veth_event);
- else if (event->xFlags.xFunction == HvLpEvent_Function_Int)
+ else
veth_handle_int(veth_event);
}
static int veth_process_caps(struct veth_lpar_connection *cnx)
{
- struct VethCapData *remote_caps = &cnx->remote_caps;
+ struct veth_cap_data *remote_caps = &cnx->remote_caps;
int num_acks_needed;
/* Convert timer to jiffies */
restart:
if (cnx->state & VETH_STATE_RESET) {
- int i;
-
- del_timer(&cnx->ack_timer);
-
if (cnx->state & VETH_STATE_OPEN)
HvCallEvent_closeLpEventPath(cnx->remote_lp,
HvLpEvent_Type_VirtualLan);
- /* reset ack data */
+ /*
+ * Reset ack data. This prevents the ack_timer actually
+ * doing anything, even if it runs one more time when
+ * we drop the lock below.
+ */
memset(&cnx->pending_acks, 0xff, sizeof (cnx->pending_acks));
cnx->num_pending_acks = 0;
| VETH_STATE_SENTCAPACK | VETH_STATE_READY);
/* Clean up any leftover messages */
- if (cnx->msgs)
+ if (cnx->msgs) {
+ int i;
for (i = 0; i < VETH_NUMBUFFERS; ++i)
veth_recycle_msg(cnx, cnx->msgs + i);
+ }
+
+ cnx->outstanding_tx = 0;
+ veth_wake_queues(cnx);
+
+ /* Drop the lock so we can do stuff that might sleep or
+ * take other locks. */
spin_unlock_irq(&cnx->lock);
- veth_flush_pending(cnx);
+
+ del_timer_sync(&cnx->ack_timer);
+ del_timer_sync(&cnx->reset_timer);
+
spin_lock_irq(&cnx->lock);
+
if (cnx->state & VETH_STATE_RESET)
goto restart;
+
+ /* Hack, wait for the other end to reset itself. */
+ if (! (cnx->state & VETH_STATE_SHUTDOWN)) {
+ schedule_delayed_work(&cnx->statemachine_wq, 5 * HZ);
+ goto out;
+ }
}
if (cnx->state & VETH_STATE_SHUTDOWN)
if ( (cnx->state & VETH_STATE_OPEN)
&& !(cnx->state & VETH_STATE_SENTMON) ) {
- rc = veth_signalevent(cnx, VethEventTypeMonitor,
+ rc = veth_signalevent(cnx, VETH_EVENT_MONITOR,
HvLpEvent_AckInd_DoAck,
HvLpEvent_AckType_DeferredAck,
0, 0, 0, 0, 0, 0);
&& !(cnx->state & VETH_STATE_SENTCAPS)) {
u64 *rawcap = (u64 *)&cnx->local_caps;
- rc = veth_signalevent(cnx, VethEventTypeCap,
+ rc = veth_signalevent(cnx, VETH_EVENT_CAP,
HvLpEvent_AckInd_DoAck,
HvLpEvent_AckType_ImmediateAck,
0, rawcap[0], rawcap[1], rawcap[2],
if ((cnx->state & VETH_STATE_GOTCAPS)
&& !(cnx->state & VETH_STATE_SENTCAPACK)) {
- struct VethCapData *remote_caps = &cnx->remote_caps;
+ struct veth_cap_data *remote_caps = &cnx->remote_caps;
memcpy(remote_caps, &cnx->cap_event.u.caps_data,
sizeof(*remote_caps));
{
struct veth_lpar_connection *cnx;
struct veth_msg *msgs;
- int i;
+ int i, rc;
if ( (rlp == this_lp)
|| ! HvLpConfig_doLpsCommunicateOnVirtualLan(this_lp, rlp) )
cnx->remote_lp = rlp;
spin_lock_init(&cnx->lock);
INIT_WORK(&cnx->statemachine_wq, veth_statemachine, cnx);
+
init_timer(&cnx->ack_timer);
cnx->ack_timer.function = veth_timed_ack;
cnx->ack_timer.data = (unsigned long) cnx;
+
+ init_timer(&cnx->reset_timer);
+ cnx->reset_timer.function = veth_timed_reset;
+ cnx->reset_timer.data = (unsigned long) cnx;
+ cnx->reset_timeout = 5 * HZ * (VETH_ACKTIMEOUT / 1000000);
+
memset(&cnx->pending_acks, 0xff, sizeof (cnx->pending_acks));
veth_cnx[rlp] = cnx;
+ /* This gets us 1 reference, which is held on behalf of the driver
+ * infrastructure. It's released at module unload. */
+ kobject_init(&cnx->kobject);
+ cnx->kobject.ktype = &veth_lpar_connection_ktype;
+ rc = kobject_set_name(&cnx->kobject, "cnx%.2d", rlp);
+ if (rc != 0)
+ return rc;
+
msgs = kmalloc(VETH_NUMBUFFERS * sizeof(struct veth_msg), GFP_KERNEL);
if (! msgs) {
veth_error("Can't allocate buffers for LPAR %d.\n", rlp);
cnx->msgs = msgs;
memset(msgs, 0, VETH_NUMBUFFERS * sizeof(struct veth_msg));
- spin_lock_init(&cnx->msg_stack_lock);
for (i = 0; i < VETH_NUMBUFFERS; i++) {
msgs[i].token = i;
return 0;
}
-static void veth_stop_connection(u8 rlp)
+static void veth_stop_connection(struct veth_lpar_connection *cnx)
{
- struct veth_lpar_connection *cnx = veth_cnx[rlp];
-
- if (! cnx)
+ if (!cnx)
return;
spin_lock_irq(&cnx->lock);
veth_kick_statemachine(cnx);
spin_unlock_irq(&cnx->lock);
+ /* There's a slim chance the reset code has just queued the
+ * statemachine to run in five seconds. If so we need to cancel
+ * that and requeue the work to run now. */
+ if (cancel_delayed_work(&cnx->statemachine_wq)) {
+ spin_lock_irq(&cnx->lock);
+ veth_kick_statemachine(cnx);
+ spin_unlock_irq(&cnx->lock);
+ }
+
+ /* Wait for the state machine to run. */
flush_scheduled_work();
+}
- /* FIXME: not sure if this is necessary - will already have
- * been deleted by the state machine, just want to make sure
- * its not running any more */
- del_timer_sync(&cnx->ack_timer);
+static void veth_destroy_connection(struct veth_lpar_connection *cnx)
+{
+ if (!cnx)
+ return;
if (cnx->num_events > 0)
mf_deallocate_lp_events(cnx->remote_lp,
HvLpEvent_Type_VirtualLan,
cnx->num_ack_events,
NULL, NULL);
-}
-
-static void veth_destroy_connection(u8 rlp)
-{
- struct veth_lpar_connection *cnx = veth_cnx[rlp];
-
- if (! cnx)
- return;
kfree(cnx->msgs);
+ veth_cnx[cnx->remote_lp] = NULL;
kfree(cnx);
- veth_cnx[rlp] = NULL;
+}
+
+static void veth_release_connection(struct kobject *kobj)
+{
+ struct veth_lpar_connection *cnx;
+ cnx = container_of(kobj, struct veth_lpar_connection, kobject);
+ veth_stop_connection(cnx);
+ veth_destroy_connection(cnx);
}
/*
write_lock_irqsave(&port->mcast_gate, flags);
- if (dev->flags & IFF_PROMISC) { /* set promiscuous mode */
- printk(KERN_INFO "%s: Promiscuous mode enabled.\n",
- dev->name);
+ if ((dev->flags & IFF_PROMISC) || (dev->flags & IFF_ALLMULTI) ||
+ (dev->mc_count > VETH_MAX_MCAST)) {
port->promiscuous = 1;
- } else if ( (dev->flags & IFF_ALLMULTI)
- || (dev->mc_count > VETH_MAX_MCAST) ) {
- port->all_mcast = 1;
} else {
struct dev_mc_list *dmi = dev->mc_list;
int i;
+ port->promiscuous = 0;
+
/* Update table */
port->num_mcast = 0;
static void veth_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
{
- strncpy(info->driver, "veth", sizeof(info->driver) - 1);
+ strncpy(info->driver, DRV_NAME, sizeof(info->driver) - 1);
info->driver[sizeof(info->driver) - 1] = '\0';
- strncpy(info->version, "1.0", sizeof(info->version) - 1);
+ strncpy(info->version, DRV_VERSION, sizeof(info->version) - 1);
+ info->version[sizeof(info->version) - 1] = '\0';
}
static int veth_get_settings(struct net_device *dev, struct ethtool_cmd *ecmd)
return 1;
}
-static struct ethtool_ops ops = {
+static const struct ethtool_ops ops = {
.get_drvinfo = veth_get_drvinfo,
.get_settings = veth_get_settings,
.get_link = veth_get_link,
};
-static void veth_tx_timeout(struct net_device *dev)
-{
- struct veth_port *port = (struct veth_port *)dev->priv;
- struct net_device_stats *stats = &port->stats;
- unsigned long flags;
- int i;
-
- stats->tx_errors++;
-
- spin_lock_irqsave(&port->pending_gate, flags);
-
- if (!port->pending_lpmask) {
- spin_unlock_irqrestore(&port->pending_gate, flags);
- return;
- }
-
- printk(KERN_WARNING "%s: Tx timeout! Resetting lp connections: %08x\n",
- dev->name, port->pending_lpmask);
-
- for (i = 0; i < HVMAXARCHITECTEDLPS; i++) {
- struct veth_lpar_connection *cnx = veth_cnx[i];
-
- if (! (port->pending_lpmask & (1<<i)))
- continue;
-
- /* If we're pending on it, we must be connected to it,
- * so we should certainly have a structure for it. */
- BUG_ON(! cnx);
-
- /* Theoretically we could be kicking a connection
- * which doesn't deserve it, but in practice if we've
- * had a Tx timeout, the pending_lpmask will have
- * exactly one bit set - the connection causing the
- * problem. */
- spin_lock(&cnx->lock);
- cnx->state |= VETH_STATE_RESET;
- veth_kick_statemachine(cnx);
- spin_unlock(&cnx->lock);
- }
-
- spin_unlock_irqrestore(&port->pending_gate, flags);
-}
-
-static struct net_device * __init veth_probe_one(int vlan, struct device *vdev)
+static struct net_device * __init veth_probe_one(int vlan,
+ struct vio_dev *vio_dev)
{
struct net_device *dev;
struct veth_port *port;
+ struct device *vdev = &vio_dev->dev;
int i, rc;
+ const unsigned char *mac_addr;
+
+ mac_addr = vio_get_attribute(vio_dev, "local-mac-address", NULL);
+ if (mac_addr == NULL)
+ mac_addr = vio_get_attribute(vio_dev, "mac-address", NULL);
+ if (mac_addr == NULL) {
+ veth_error("Unable to fetch MAC address from device tree.\n");
+ return NULL;
+ }
dev = alloc_etherdev(sizeof (struct veth_port));
if (! dev) {
port = (struct veth_port *) dev->priv;
- spin_lock_init(&port->pending_gate);
+ spin_lock_init(&port->queue_lock);
rwlock_init(&port->mcast_gate);
+ port->stopped_map = 0;
for (i = 0; i < HVMAXARCHITECTEDLPS; i++) {
HvLpVirtualLanIndexMap map;
}
port->dev = vdev;
- dev->dev_addr[0] = 0x02;
- dev->dev_addr[1] = 0x01;
- dev->dev_addr[2] = 0xff;
- dev->dev_addr[3] = vlan;
- dev->dev_addr[4] = 0xff;
- dev->dev_addr[5] = this_lp;
+ memcpy(dev->dev_addr, mac_addr, ETH_ALEN);
dev->mtu = VETH_MAX_MTU;
- memcpy(&port->mac_addr, dev->dev_addr, 6);
+ memcpy(&port->mac_addr, mac_addr, ETH_ALEN);
dev->open = veth_open;
dev->hard_start_xmit = veth_start_xmit;
dev->set_multicast_list = veth_set_multicast_list;
SET_ETHTOOL_OPS(dev, &ops);
- dev->watchdog_timeo = 2 * (VETH_ACKTIMEOUT * HZ / 1000000);
- dev->tx_timeout = veth_tx_timeout;
-
SET_NETDEV_DEV(dev, vdev);
rc = register_netdev(dev);
return NULL;
}
+ kobject_init(&port->kobject);
+ port->kobject.parent = &dev->class_dev.kobj;
+ port->kobject.ktype = &veth_port_ktype;
+ kobject_set_name(&port->kobject, "veth_port");
+ if (0 != kobject_add(&port->kobject))
+ veth_error("Failed adding port for %s to sysfs.\n", dev->name);
+
veth_info("%s attached to iSeries vlan %d (LPAR map = 0x%.4X)\n",
dev->name, vlan, port->lpar_map);
struct veth_lpar_connection *cnx = veth_cnx[rlp];
struct veth_port *port = (struct veth_port *) dev->priv;
HvLpEvent_Rc rc;
- u32 dma_address, dma_length;
struct veth_msg *msg = NULL;
- int err = 0;
unsigned long flags;
- if (! cnx) {
- port->stats.tx_errors++;
- dev_kfree_skb(skb);
+ if (! cnx)
return 0;
- }
spin_lock_irqsave(&cnx->lock, flags);
if (! (cnx->state & VETH_STATE_READY))
- goto drop;
+ goto no_error;
- if ((skb->len - 14) > VETH_MAX_MTU)
+ if ((skb->len - ETH_HLEN) > VETH_MAX_MTU)
goto drop;
msg = veth_stack_pop(cnx);
-
- if (! msg) {
- err = 1;
+ if (! msg)
goto drop;
- }
- dma_length = skb->len;
- dma_address = dma_map_single(port->dev, skb->data,
- dma_length, DMA_TO_DEVICE);
+ msg->in_use = 1;
+ msg->skb = skb_get(skb);
+
+ msg->data.addr[0] = dma_map_single(port->dev, skb->data,
+ skb->len, DMA_TO_DEVICE);
- if (dma_mapping_error(dma_address))
+ if (dma_mapping_error(msg->data.addr[0]))
goto recycle_and_drop;
- /* Is it really necessary to check the length and address
- * fields of the first entry here? */
- msg->skb = skb;
msg->dev = port->dev;
- msg->data.addr[0] = dma_address;
- msg->data.len[0] = dma_length;
+ msg->data.len[0] = skb->len;
msg->data.eofmask = 1 << VETH_EOF_SHIFT;
- set_bit(0, &(msg->in_use));
- rc = veth_signaldata(cnx, VethEventTypeFrames, msg->token, &msg->data);
+
+ rc = veth_signaldata(cnx, VETH_EVENT_FRAMES, msg->token, &msg->data);
if (rc != HvLpEvent_Rc_Good)
goto recycle_and_drop;
+ /* If the timer's not already running, start it now. */
+ if (0 == cnx->outstanding_tx)
+ mod_timer(&cnx->reset_timer, jiffies + cnx->reset_timeout);
+
+ cnx->last_contact = jiffies;
+ cnx->outstanding_tx++;
+
+ if (veth_stack_is_empty(cnx))
+ veth_stop_queues(cnx);
+
+ no_error:
spin_unlock_irqrestore(&cnx->lock, flags);
return 0;
recycle_and_drop:
- msg->skb = NULL;
- /* need to set in use to make veth_recycle_msg in case this
- * was a mapping failure */
- set_bit(0, &msg->in_use);
veth_recycle_msg(cnx, msg);
drop:
- port->stats.tx_errors++;
- dev_kfree_skb(skb);
spin_unlock_irqrestore(&cnx->lock, flags);
- return err;
+ return 1;
}
-static HvLpIndexMap veth_transmit_to_many(struct sk_buff *skb,
+static void veth_transmit_to_many(struct sk_buff *skb,
HvLpIndexMap lpmask,
struct net_device *dev)
{
struct veth_port *port = (struct veth_port *) dev->priv;
- int i;
- int rc;
+ int i, success, error;
+
+ success = error = 0;
for (i = 0; i < HVMAXARCHITECTEDLPS; i++) {
if ((lpmask & (1 << i)) == 0)
continue;
- rc = veth_transmit_to_one(skb_get(skb), i, dev);
- if (! rc)
- lpmask &= ~(1<<i);
+ if (veth_transmit_to_one(skb, i, dev))
+ error = 1;
+ else
+ success = 1;
}
- if (! lpmask) {
+ if (error)
+ port->stats.tx_errors++;
+
+ if (success) {
port->stats.tx_packets++;
port->stats.tx_bytes += skb->len;
}
-
- return lpmask;
}
static int veth_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
unsigned char *frame = skb->data;
struct veth_port *port = (struct veth_port *) dev->priv;
- unsigned long flags;
HvLpIndexMap lpmask;
if (! (frame[0] & 0x01)) {
lpmask = port->lpar_map;
}
- spin_lock_irqsave(&port->pending_gate, flags);
-
- lpmask = veth_transmit_to_many(skb, lpmask, dev);
-
- dev->trans_start = jiffies;
+ veth_transmit_to_many(skb, lpmask, dev);
- if (! lpmask) {
- dev_kfree_skb(skb);
- } else {
- if (port->pending_skb) {
- veth_error("%s: TX while skb was pending!\n",
- dev->name);
- dev_kfree_skb(skb);
- spin_unlock_irqrestore(&port->pending_gate, flags);
- return 1;
- }
-
- port->pending_skb = skb;
- port->pending_lpmask = lpmask;
- netif_stop_queue(dev);
- }
-
- spin_unlock_irqrestore(&port->pending_gate, flags);
+ dev_kfree_skb(skb);
return 0;
}
+/* You must hold the connection's lock when you call this function. */
static void veth_recycle_msg(struct veth_lpar_connection *cnx,
struct veth_msg *msg)
{
u32 dma_address, dma_length;
- if (test_and_clear_bit(0, &msg->in_use)) {
+ if (msg->in_use) {
+ msg->in_use = 0;
dma_address = msg->data.addr[0];
dma_length = msg->data.len[0];
- dma_unmap_single(msg->dev, dma_address, dma_length,
- DMA_TO_DEVICE);
+ if (!dma_mapping_error(dma_address))
+ dma_unmap_single(msg->dev, dma_address, dma_length,
+ DMA_TO_DEVICE);
if (msg->skb) {
dev_kfree_skb_any(msg->skb);
}
}
-static void veth_flush_pending(struct veth_lpar_connection *cnx)
+static void veth_wake_queues(struct veth_lpar_connection *cnx)
{
int i;
+
for (i = 0; i < HVMAXARCHITECTEDVIRTUALLANS; i++) {
struct net_device *dev = veth_dev[i];
struct veth_port *port;
if (! (port->lpar_map & (1<<cnx->remote_lp)))
continue;
- spin_lock_irqsave(&port->pending_gate, flags);
- if (port->pending_skb) {
- port->pending_lpmask =
- veth_transmit_to_many(port->pending_skb,
- port->pending_lpmask,
- dev);
- if (! port->pending_lpmask) {
- dev_kfree_skb_any(port->pending_skb);
- port->pending_skb = NULL;
- netif_wake_queue(dev);
- }
+ spin_lock_irqsave(&port->queue_lock, flags);
+
+ port->stopped_map &= ~(1 << cnx->remote_lp);
+
+ if (0 == port->stopped_map && netif_queue_stopped(dev)) {
+ veth_debug("cnx %d: woke queue for %s.\n",
+ cnx->remote_lp, dev->name);
+ netif_wake_queue(dev);
}
- spin_unlock_irqrestore(&port->pending_gate, flags);
+ spin_unlock_irqrestore(&port->queue_lock, flags);
}
}
+static void veth_stop_queues(struct veth_lpar_connection *cnx)
+{
+ int i;
+
+ for (i = 0; i < HVMAXARCHITECTEDVIRTUALLANS; i++) {
+ struct net_device *dev = veth_dev[i];
+ struct veth_port *port;
+
+ if (! dev)
+ continue;
+
+ port = (struct veth_port *)dev->priv;
+
+ /* If this cnx is not on the vlan for this port, continue */
+ if (! (port->lpar_map & (1 << cnx->remote_lp)))
+ continue;
+
+ spin_lock(&port->queue_lock);
+
+ netif_stop_queue(dev);
+ port->stopped_map |= (1 << cnx->remote_lp);
+
+ veth_debug("cnx %d: stopped queue for %s, map = 0x%x.\n",
+ cnx->remote_lp, dev->name, port->stopped_map);
+
+ spin_unlock(&port->queue_lock);
+ }
+}
+
+static void veth_timed_reset(unsigned long ptr)
+{
+ struct veth_lpar_connection *cnx = (struct veth_lpar_connection *)ptr;
+ unsigned long trigger_time, flags;
+
+ /* FIXME is it possible this fires after veth_stop_connection()?
+ * That would reschedule the statemachine for 5 seconds and probably
+ * execute it after the module's been unloaded. Hmm. */
+
+ spin_lock_irqsave(&cnx->lock, flags);
+
+ if (cnx->outstanding_tx > 0) {
+ trigger_time = cnx->last_contact + cnx->reset_timeout;
+
+ if (trigger_time < jiffies) {
+ cnx->state |= VETH_STATE_RESET;
+ veth_kick_statemachine(cnx);
+ veth_error("%d packets not acked by LPAR %d within %d "
+ "seconds, resetting.\n",
+ cnx->outstanding_tx, cnx->remote_lp,
+ cnx->reset_timeout / HZ);
+ } else {
+ /* Reschedule the timer */
+ trigger_time = jiffies + cnx->reset_timeout;
+ mod_timer(&cnx->reset_timer, trigger_time);
+ }
+ }
+
+ spin_unlock_irqrestore(&cnx->lock, flags);
+}
+
/*
* Rx path
*/
if ( (mac_addr == port->mac_addr) || (mac_addr == 0xffffffffffff0000) )
return 1;
- if (! (((char *) &mac_addr)[0] & 0x01))
- return 0;
-
read_lock_irqsave(&port->mcast_gate, flags);
- if (port->promiscuous || port->all_mcast) {
+ if (port->promiscuous) {
wanted = 1;
goto out;
}
* it just at the granularity of iSeries real->absolute
* mapping? Indeed, given the way the allocator works, can we
* count on them being absolutely contiguous? */
- list[0].addr = ISERIES_HV_ADDR(p);
+ list[0].addr = iseries_hv_addr(p);
list[0].size = min(length,
PAGE_SIZE - ((unsigned long)p & ~PAGE_MASK));
done = list[0].size;
while (done < length) {
- list[i].addr = ISERIES_HV_ADDR(p + done);
+ list[i].addr = iseries_hv_addr(p + done);
list[i].size = min(length-done, PAGE_SIZE);
done += list[i].size;
i++;
{
HvLpEvent_Rc rc;
- rc = veth_signaldata(cnx, VethEventTypeFramesAck,
+ rc = veth_signaldata(cnx, VETH_EVENT_FRAMES_ACK,
0, &cnx->pending_acks);
if (rc != HvLpEvent_Rc_Good)
}
static void veth_receive(struct veth_lpar_connection *cnx,
- struct VethLpEvent *event)
+ struct veth_lpevent *event)
{
- struct VethFramesData *senddata = &event->u.frames_data;
+ struct veth_frames_data *senddata = &event->u.frames_data;
int startchunk = 0;
int nchunks;
unsigned long flags;
cnx->dst_inst,
HvLpDma_AddressType_RealAddress,
HvLpDma_AddressType_TceIndex,
- ISERIES_HV_ADDR(&local_list),
- ISERIES_HV_ADDR(&remote_list),
+ iseries_hv_addr(&local_list),
+ iseries_hv_addr(&remote_list),
length);
if (rc != HvLpDma_Rc_Good) {
dev_kfree_skb_irq(skb);
static int veth_remove(struct vio_dev *vdev)
{
- int i = vdev->unit_address;
+ struct veth_lpar_connection *cnx;
struct net_device *dev;
+ struct veth_port *port;
+ int i;
- dev = veth_dev[i];
- if (dev != NULL) {
- veth_dev[i] = NULL;
- unregister_netdev(dev);
- free_netdev(dev);
+ dev = veth_dev[vdev->unit_address];
+
+ if (! dev)
+ return 0;
+
+ port = netdev_priv(dev);
+
+ for (i = 0; i < HVMAXARCHITECTEDLPS; i++) {
+ cnx = veth_cnx[i];
+
+ if (cnx && (port->lpar_map & (1 << i))) {
+ /* Drop our reference to connections on our VLAN */
+ kobject_put(&cnx->kobject);
+ }
}
+
+ veth_dev[vdev->unit_address] = NULL;
+ kobject_del(&port->kobject);
+ kobject_put(&port->kobject);
+ unregister_netdev(dev);
+ free_netdev(dev);
+
return 0;
}
{
int i = vdev->unit_address;
struct net_device *dev;
+ struct veth_port *port;
- dev = veth_probe_one(i, &vdev->dev);
+ dev = veth_probe_one(i, vdev);
if (dev == NULL) {
veth_remove(vdev);
return 1;
}
veth_dev[i] = dev;
- /* Start the state machine on each connection, to commence
- * link negotiation */
- for (i = 0; i < HVMAXARCHITECTEDLPS; i++)
- if (veth_cnx[i])
- veth_kick_statemachine(veth_cnx[i]);
+ port = (struct veth_port*)netdev_priv(dev);
+
+ /* Start the state machine on each connection on this vlan. If we're
+ * the first dev to do so this will commence link negotiation */
+ for (i = 0; i < HVMAXARCHITECTEDLPS; i++) {
+ struct veth_lpar_connection *cnx;
+
+ if (! (port->lpar_map & (1 << i)))
+ continue;
+
+ cnx = veth_cnx[i];
+ if (!cnx)
+ continue;
+
+ kobject_get(&cnx->kobject);
+ veth_kick_statemachine(cnx);
+ }
return 0;
}
* support.
*/
static struct vio_device_id veth_device_table[] __devinitdata = {
- { "vlan", "" },
+ { "network", "IBM,iSeries-l-lan" },
{ "", "" }
};
MODULE_DEVICE_TABLE(vio, veth_device_table);
static struct vio_driver veth_driver = {
- .name = "iseries_veth",
.id_table = veth_device_table,
.probe = veth_probe,
- .remove = veth_remove
+ .remove = veth_remove,
+ .driver = {
+ .name = DRV_NAME,
+ .owner = THIS_MODULE,
+ }
};
/*
void __exit veth_module_cleanup(void)
{
int i;
+ struct veth_lpar_connection *cnx;
- /* Stop the queues first to stop any new packets being sent. */
- for (i = 0; i < HVMAXARCHITECTEDVIRTUALLANS; i++)
- if (veth_dev[i])
- netif_stop_queue(veth_dev[i]);
-
- /* Stop the connections before we unregister the driver. This
- * ensures there's no skbs lying around holding the device open. */
- for (i = 0; i < HVMAXARCHITECTEDLPS; ++i)
- veth_stop_connection(i);
-
+ /* Disconnect our "irq" to stop events coming from the Hypervisor. */
HvLpEvent_unregisterHandler(HvLpEvent_Type_VirtualLan);
- /* Hypervisor callbacks may have scheduled more work while we
- * were stoping connections. Now that we've disconnected from
- * the hypervisor make sure everything's finished. */
+ /* Make sure any work queued from Hypervisor callbacks is finished. */
flush_scheduled_work();
- vio_unregister_driver(&veth_driver);
+ for (i = 0; i < HVMAXARCHITECTEDLPS; ++i) {
+ cnx = veth_cnx[i];
- for (i = 0; i < HVMAXARCHITECTEDLPS; ++i)
- veth_destroy_connection(i);
+ if (!cnx)
+ continue;
+ /* Remove the connection from sysfs */
+ kobject_del(&cnx->kobject);
+ /* Drop the driver's reference to the connection */
+ kobject_put(&cnx->kobject);
+ }
+
+ /* Unregister the driver, which will close all the netdevs and stop
+ * the connections when they're no longer referenced. */
+ vio_unregister_driver(&veth_driver);
}
module_exit(veth_module_cleanup);
for (i = 0; i < HVMAXARCHITECTEDLPS; ++i) {
rc = veth_init_connection(i);
- if (rc != 0) {
- veth_module_cleanup();
- return rc;
- }
+ if (rc != 0)
+ goto error;
}
HvLpEvent_registerHandler(HvLpEvent_Type_VirtualLan,
&veth_handle_event);
- return vio_register_driver(&veth_driver);
+ rc = vio_register_driver(&veth_driver);
+ if (rc != 0)
+ goto error;
+
+ for (i = 0; i < HVMAXARCHITECTEDLPS; ++i) {
+ struct kobject *kobj;
+
+ if (!veth_cnx[i])
+ continue;
+
+ kobj = &veth_cnx[i]->kobject;
+ kobj->parent = &veth_driver.driver.kobj;
+ /* If the add failes, complain but otherwise continue */
+ if (0 != kobject_add(kobj))
+ veth_error("cnx %d: Failed adding to sysfs.\n", i);
+ }
+
+ return 0;
+
+error:
+ for (i = 0; i < HVMAXARCHITECTEDLPS; ++i) {
+ veth_destroy_connection(veth_cnx[i]);
+ }
+
+ return rc;
}
module_init(veth_module_init);
Code Review / linux-2.6.git / blobdiff