linux-zen-server/drivers/scsi/fnic/fnic_fcs.c

1398 lines
38 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright 2008 Cisco Systems, Inc. All rights reserved.
* Copyright 2007 Nuova Systems, Inc. All rights reserved.
*/
#include <linux/errno.h>
#include <linux/pci.h>
#include <linux/slab.h>
#include <linux/skbuff.h>
#include <linux/interrupt.h>
#include <linux/spinlock.h>
#include <linux/if_ether.h>
#include <linux/if_vlan.h>
#include <linux/workqueue.h>
#include <scsi/fc/fc_fip.h>
#include <scsi/fc/fc_els.h>
#include <scsi/fc/fc_fcoe.h>
#include <scsi/fc_frame.h>
#include <scsi/libfc.h>
#include "fnic_io.h"
#include "fnic.h"
#include "fnic_fip.h"
#include "cq_enet_desc.h"
#include "cq_exch_desc.h"
static u8 fcoe_all_fcfs[ETH_ALEN] = FIP_ALL_FCF_MACS;
struct workqueue_struct *fnic_fip_queue;
struct workqueue_struct *fnic_event_queue;
static void fnic_set_eth_mode(struct fnic *);
static void fnic_fcoe_send_vlan_req(struct fnic *fnic);
static void fnic_fcoe_start_fcf_disc(struct fnic *fnic);
static void fnic_fcoe_process_vlan_resp(struct fnic *fnic, struct sk_buff *);
static int fnic_fcoe_vlan_check(struct fnic *fnic, u16 flag);
static int fnic_fcoe_handle_fip_frame(struct fnic *fnic, struct sk_buff *skb);
void fnic_handle_link(struct work_struct *work)
{
struct fnic *fnic = container_of(work, struct fnic, link_work);
unsigned long flags;
int old_link_status;
u32 old_link_down_cnt;
u64 old_port_speed, new_port_speed;
spin_lock_irqsave(&fnic->fnic_lock, flags);
fnic->link_events = 1; /* less work to just set everytime*/
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
old_link_down_cnt = fnic->link_down_cnt;
old_link_status = fnic->link_status;
old_port_speed = atomic64_read(
&fnic->fnic_stats.misc_stats.current_port_speed);
fnic->link_status = vnic_dev_link_status(fnic->vdev);
fnic->link_down_cnt = vnic_dev_link_down_cnt(fnic->vdev);
new_port_speed = vnic_dev_port_speed(fnic->vdev);
atomic64_set(&fnic->fnic_stats.misc_stats.current_port_speed,
new_port_speed);
if (old_port_speed != new_port_speed)
FNIC_MAIN_DBG(KERN_INFO, fnic->lport->host,
"Current vnic speed set to : %llu\n",
new_port_speed);
switch (vnic_dev_port_speed(fnic->vdev)) {
case DCEM_PORTSPEED_10G:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_10GBIT;
fnic->lport->link_supported_speeds = FC_PORTSPEED_10GBIT;
break;
case DCEM_PORTSPEED_20G:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_20GBIT;
fnic->lport->link_supported_speeds = FC_PORTSPEED_20GBIT;
break;
case DCEM_PORTSPEED_25G:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_25GBIT;
fnic->lport->link_supported_speeds = FC_PORTSPEED_25GBIT;
break;
case DCEM_PORTSPEED_40G:
case DCEM_PORTSPEED_4x10G:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_40GBIT;
fnic->lport->link_supported_speeds = FC_PORTSPEED_40GBIT;
break;
case DCEM_PORTSPEED_100G:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_100GBIT;
fnic->lport->link_supported_speeds = FC_PORTSPEED_100GBIT;
break;
default:
fc_host_speed(fnic->lport->host) = FC_PORTSPEED_UNKNOWN;
fnic->lport->link_supported_speeds = FC_PORTSPEED_UNKNOWN;
break;
}
if (old_link_status == fnic->link_status) {
if (!fnic->link_status) {
/* DOWN -> DOWN */
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_fc_trace_set_data(fnic->lport->host->host_no,
FNIC_FC_LE, "Link Status: DOWN->DOWN",
strlen("Link Status: DOWN->DOWN"));
} else {
if (old_link_down_cnt != fnic->link_down_cnt) {
/* UP -> DOWN -> UP */
fnic->lport->host_stats.link_failure_count++;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_fc_trace_set_data(
fnic->lport->host->host_no,
FNIC_FC_LE,
"Link Status:UP_DOWN_UP",
strlen("Link_Status:UP_DOWN_UP")
);
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"link down\n");
fcoe_ctlr_link_down(&fnic->ctlr);
if (fnic->config.flags & VFCF_FIP_CAPABLE) {
/* start FCoE VLAN discovery */
fnic_fc_trace_set_data(
fnic->lport->host->host_no,
FNIC_FC_LE,
"Link Status: UP_DOWN_UP_VLAN",
strlen(
"Link Status: UP_DOWN_UP_VLAN")
);
fnic_fcoe_send_vlan_req(fnic);
return;
}
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"link up\n");
fcoe_ctlr_link_up(&fnic->ctlr);
} else {
/* UP -> UP */
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_fc_trace_set_data(
fnic->lport->host->host_no, FNIC_FC_LE,
"Link Status: UP_UP",
strlen("Link Status: UP_UP"));
}
}
} else if (fnic->link_status) {
/* DOWN -> UP */
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
if (fnic->config.flags & VFCF_FIP_CAPABLE) {
/* start FCoE VLAN discovery */
fnic_fc_trace_set_data(
fnic->lport->host->host_no,
FNIC_FC_LE, "Link Status: DOWN_UP_VLAN",
strlen("Link Status: DOWN_UP_VLAN"));
fnic_fcoe_send_vlan_req(fnic);
return;
}
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host, "link up\n");
fnic_fc_trace_set_data(fnic->lport->host->host_no, FNIC_FC_LE,
"Link Status: DOWN_UP", strlen("Link Status: DOWN_UP"));
fcoe_ctlr_link_up(&fnic->ctlr);
} else {
/* UP -> DOWN */
fnic->lport->host_stats.link_failure_count++;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host, "link down\n");
fnic_fc_trace_set_data(
fnic->lport->host->host_no, FNIC_FC_LE,
"Link Status: UP_DOWN",
strlen("Link Status: UP_DOWN"));
if (fnic->config.flags & VFCF_FIP_CAPABLE) {
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"deleting fip-timer during link-down\n");
del_timer_sync(&fnic->fip_timer);
}
fcoe_ctlr_link_down(&fnic->ctlr);
}
}
/*
* This function passes incoming fabric frames to libFC
*/
void fnic_handle_frame(struct work_struct *work)
{
struct fnic *fnic = container_of(work, struct fnic, frame_work);
struct fc_lport *lp = fnic->lport;
unsigned long flags;
struct sk_buff *skb;
struct fc_frame *fp;
while ((skb = skb_dequeue(&fnic->frame_queue))) {
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
dev_kfree_skb(skb);
return;
}
fp = (struct fc_frame *)skb;
/*
* If we're in a transitional state, just re-queue and return.
* The queue will be serviced when we get to a stable state.
*/
if (fnic->state != FNIC_IN_FC_MODE &&
fnic->state != FNIC_IN_ETH_MODE) {
skb_queue_head(&fnic->frame_queue, skb);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fc_exch_recv(lp, fp);
}
}
void fnic_fcoe_evlist_free(struct fnic *fnic)
{
struct fnic_event *fevt = NULL;
struct fnic_event *next = NULL;
unsigned long flags;
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (list_empty(&fnic->evlist)) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
list_for_each_entry_safe(fevt, next, &fnic->evlist, list) {
list_del(&fevt->list);
kfree(fevt);
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
void fnic_handle_event(struct work_struct *work)
{
struct fnic *fnic = container_of(work, struct fnic, event_work);
struct fnic_event *fevt = NULL;
struct fnic_event *next = NULL;
unsigned long flags;
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (list_empty(&fnic->evlist)) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
list_for_each_entry_safe(fevt, next, &fnic->evlist, list) {
if (fnic->stop_rx_link_events) {
list_del(&fevt->list);
kfree(fevt);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
/*
* If we're in a transitional state, just re-queue and return.
* The queue will be serviced when we get to a stable state.
*/
if (fnic->state != FNIC_IN_FC_MODE &&
fnic->state != FNIC_IN_ETH_MODE) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
list_del(&fevt->list);
switch (fevt->event) {
case FNIC_EVT_START_VLAN_DISC:
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_fcoe_send_vlan_req(fnic);
spin_lock_irqsave(&fnic->fnic_lock, flags);
break;
case FNIC_EVT_START_FCF_DISC:
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"Start FCF Discovery\n");
fnic_fcoe_start_fcf_disc(fnic);
break;
default:
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"Unknown event 0x%x\n", fevt->event);
break;
}
kfree(fevt);
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
/**
* is_fnic_fip_flogi_reject() - Check if the Received FIP FLOGI frame is rejected
* @fip: The FCoE controller that received the frame
* @skb: The received FIP frame
*
* Returns non-zero if the frame is rejected with unsupported cmd with
* insufficient resource els explanation.
*/
static inline int is_fnic_fip_flogi_reject(struct fcoe_ctlr *fip,
struct sk_buff *skb)
{
struct fc_lport *lport = fip->lp;
struct fip_header *fiph;
struct fc_frame_header *fh = NULL;
struct fip_desc *desc;
struct fip_encaps *els;
u16 op;
u8 els_op;
u8 sub;
size_t rlen;
size_t dlen = 0;
if (skb_linearize(skb))
return 0;
if (skb->len < sizeof(*fiph))
return 0;
fiph = (struct fip_header *)skb->data;
op = ntohs(fiph->fip_op);
sub = fiph->fip_subcode;
if (op != FIP_OP_LS)
return 0;
if (sub != FIP_SC_REP)
return 0;
rlen = ntohs(fiph->fip_dl_len) * 4;
if (rlen + sizeof(*fiph) > skb->len)
return 0;
desc = (struct fip_desc *)(fiph + 1);
dlen = desc->fip_dlen * FIP_BPW;
if (desc->fip_dtype == FIP_DT_FLOGI) {
if (dlen < sizeof(*els) + sizeof(*fh) + 1)
return 0;
els = (struct fip_encaps *)desc;
fh = (struct fc_frame_header *)(els + 1);
if (!fh)
return 0;
/*
* ELS command code, reason and explanation should be = Reject,
* unsupported command and insufficient resource
*/
els_op = *(u8 *)(fh + 1);
if (els_op == ELS_LS_RJT) {
shost_printk(KERN_INFO, lport->host,
"Flogi Request Rejected by Switch\n");
return 1;
}
shost_printk(KERN_INFO, lport->host,
"Flogi Request Accepted by Switch\n");
}
return 0;
}
static void fnic_fcoe_send_vlan_req(struct fnic *fnic)
{
struct fcoe_ctlr *fip = &fnic->ctlr;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
struct sk_buff *skb;
char *eth_fr;
struct fip_vlan *vlan;
u64 vlan_tov;
fnic_fcoe_reset_vlans(fnic);
fnic->set_vlan(fnic, 0);
if (printk_ratelimit())
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"Sending VLAN request...\n");
skb = dev_alloc_skb(sizeof(struct fip_vlan));
if (!skb)
return;
eth_fr = (char *)skb->data;
vlan = (struct fip_vlan *)eth_fr;
memset(vlan, 0, sizeof(*vlan));
memcpy(vlan->eth.h_source, fip->ctl_src_addr, ETH_ALEN);
memcpy(vlan->eth.h_dest, fcoe_all_fcfs, ETH_ALEN);
vlan->eth.h_proto = htons(ETH_P_FIP);
vlan->fip.fip_ver = FIP_VER_ENCAPS(FIP_VER);
vlan->fip.fip_op = htons(FIP_OP_VLAN);
vlan->fip.fip_subcode = FIP_SC_VL_REQ;
vlan->fip.fip_dl_len = htons(sizeof(vlan->desc) / FIP_BPW);
vlan->desc.mac.fd_desc.fip_dtype = FIP_DT_MAC;
vlan->desc.mac.fd_desc.fip_dlen = sizeof(vlan->desc.mac) / FIP_BPW;
memcpy(&vlan->desc.mac.fd_mac, fip->ctl_src_addr, ETH_ALEN);
vlan->desc.wwnn.fd_desc.fip_dtype = FIP_DT_NAME;
vlan->desc.wwnn.fd_desc.fip_dlen = sizeof(vlan->desc.wwnn) / FIP_BPW;
put_unaligned_be64(fip->lp->wwnn, &vlan->desc.wwnn.fd_wwn);
atomic64_inc(&fnic_stats->vlan_stats.vlan_disc_reqs);
skb_put(skb, sizeof(*vlan));
skb->protocol = htons(ETH_P_FIP);
skb_reset_mac_header(skb);
skb_reset_network_header(skb);
fip->send(fip, skb);
/* set a timer so that we can retry if there no response */
vlan_tov = jiffies + msecs_to_jiffies(FCOE_CTLR_FIPVLAN_TOV);
mod_timer(&fnic->fip_timer, round_jiffies(vlan_tov));
}
static void fnic_fcoe_process_vlan_resp(struct fnic *fnic, struct sk_buff *skb)
{
struct fcoe_ctlr *fip = &fnic->ctlr;
struct fip_header *fiph;
struct fip_desc *desc;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
u16 vid;
size_t rlen;
size_t dlen;
struct fcoe_vlan *vlan;
u64 sol_time;
unsigned long flags;
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"Received VLAN response...\n");
fiph = (struct fip_header *) skb->data;
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"Received VLAN response... OP 0x%x SUB_OP 0x%x\n",
ntohs(fiph->fip_op), fiph->fip_subcode);
rlen = ntohs(fiph->fip_dl_len) * 4;
fnic_fcoe_reset_vlans(fnic);
spin_lock_irqsave(&fnic->vlans_lock, flags);
desc = (struct fip_desc *)(fiph + 1);
while (rlen > 0) {
dlen = desc->fip_dlen * FIP_BPW;
switch (desc->fip_dtype) {
case FIP_DT_VLAN:
vid = ntohs(((struct fip_vlan_desc *)desc)->fd_vlan);
shost_printk(KERN_INFO, fnic->lport->host,
"process_vlan_resp: FIP VLAN %d\n", vid);
vlan = kzalloc(sizeof(*vlan), GFP_ATOMIC);
if (!vlan) {
/* retry from timer */
spin_unlock_irqrestore(&fnic->vlans_lock,
flags);
goto out;
}
vlan->vid = vid & 0x0fff;
vlan->state = FIP_VLAN_AVAIL;
list_add_tail(&vlan->list, &fnic->vlans);
break;
}
desc = (struct fip_desc *)((char *)desc + dlen);
rlen -= dlen;
}
/* any VLAN descriptors present ? */
if (list_empty(&fnic->vlans)) {
/* retry from timer */
atomic64_inc(&fnic_stats->vlan_stats.resp_withno_vlanID);
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"No VLAN descriptors in FIP VLAN response\n");
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
goto out;
}
vlan = list_first_entry(&fnic->vlans, struct fcoe_vlan, list);
fnic->set_vlan(fnic, vlan->vid);
vlan->state = FIP_VLAN_SENT; /* sent now */
vlan->sol_count++;
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
/* start the solicitation */
fcoe_ctlr_link_up(fip);
sol_time = jiffies + msecs_to_jiffies(FCOE_CTLR_START_DELAY);
mod_timer(&fnic->fip_timer, round_jiffies(sol_time));
out:
return;
}
static void fnic_fcoe_start_fcf_disc(struct fnic *fnic)
{
unsigned long flags;
struct fcoe_vlan *vlan;
u64 sol_time;
spin_lock_irqsave(&fnic->vlans_lock, flags);
vlan = list_first_entry(&fnic->vlans, struct fcoe_vlan, list);
fnic->set_vlan(fnic, vlan->vid);
vlan->state = FIP_VLAN_SENT; /* sent now */
vlan->sol_count = 1;
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
/* start the solicitation */
fcoe_ctlr_link_up(&fnic->ctlr);
sol_time = jiffies + msecs_to_jiffies(FCOE_CTLR_START_DELAY);
mod_timer(&fnic->fip_timer, round_jiffies(sol_time));
}
static int fnic_fcoe_vlan_check(struct fnic *fnic, u16 flag)
{
unsigned long flags;
struct fcoe_vlan *fvlan;
spin_lock_irqsave(&fnic->vlans_lock, flags);
if (list_empty(&fnic->vlans)) {
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
return -EINVAL;
}
fvlan = list_first_entry(&fnic->vlans, struct fcoe_vlan, list);
if (fvlan->state == FIP_VLAN_USED) {
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
return 0;
}
if (fvlan->state == FIP_VLAN_SENT) {
fvlan->state = FIP_VLAN_USED;
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
return 0;
}
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
return -EINVAL;
}
static void fnic_event_enq(struct fnic *fnic, enum fnic_evt ev)
{
struct fnic_event *fevt;
unsigned long flags;
fevt = kmalloc(sizeof(*fevt), GFP_ATOMIC);
if (!fevt)
return;
fevt->fnic = fnic;
fevt->event = ev;
spin_lock_irqsave(&fnic->fnic_lock, flags);
list_add_tail(&fevt->list, &fnic->evlist);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
schedule_work(&fnic->event_work);
}
static int fnic_fcoe_handle_fip_frame(struct fnic *fnic, struct sk_buff *skb)
{
struct fip_header *fiph;
int ret = 1;
u16 op;
u8 sub;
if (!skb || !(skb->data))
return -1;
if (skb_linearize(skb))
goto drop;
fiph = (struct fip_header *)skb->data;
op = ntohs(fiph->fip_op);
sub = fiph->fip_subcode;
if (FIP_VER_DECAPS(fiph->fip_ver) != FIP_VER)
goto drop;
if (ntohs(fiph->fip_dl_len) * FIP_BPW + sizeof(*fiph) > skb->len)
goto drop;
if (op == FIP_OP_DISC && sub == FIP_SC_ADV) {
if (fnic_fcoe_vlan_check(fnic, ntohs(fiph->fip_flags)))
goto drop;
/* pass it on to fcoe */
ret = 1;
} else if (op == FIP_OP_VLAN && sub == FIP_SC_VL_NOTE) {
/* set the vlan as used */
fnic_fcoe_process_vlan_resp(fnic, skb);
ret = 0;
} else if (op == FIP_OP_CTRL && sub == FIP_SC_CLR_VLINK) {
/* received CVL request, restart vlan disc */
fnic_event_enq(fnic, FNIC_EVT_START_VLAN_DISC);
/* pass it on to fcoe */
ret = 1;
}
drop:
return ret;
}
void fnic_handle_fip_frame(struct work_struct *work)
{
struct fnic *fnic = container_of(work, struct fnic, fip_frame_work);
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
unsigned long flags;
struct sk_buff *skb;
struct ethhdr *eh;
while ((skb = skb_dequeue(&fnic->fip_frame_queue))) {
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
dev_kfree_skb(skb);
return;
}
/*
* If we're in a transitional state, just re-queue and return.
* The queue will be serviced when we get to a stable state.
*/
if (fnic->state != FNIC_IN_FC_MODE &&
fnic->state != FNIC_IN_ETH_MODE) {
skb_queue_head(&fnic->fip_frame_queue, skb);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
eh = (struct ethhdr *)skb->data;
if (eh->h_proto == htons(ETH_P_FIP)) {
skb_pull(skb, sizeof(*eh));
if (fnic_fcoe_handle_fip_frame(fnic, skb) <= 0) {
dev_kfree_skb(skb);
continue;
}
/*
* If there's FLOGI rejects - clear all
* fcf's & restart from scratch
*/
if (is_fnic_fip_flogi_reject(&fnic->ctlr, skb)) {
atomic64_inc(
&fnic_stats->vlan_stats.flogi_rejects);
shost_printk(KERN_INFO, fnic->lport->host,
"Trigger a Link down - VLAN Disc\n");
fcoe_ctlr_link_down(&fnic->ctlr);
/* start FCoE VLAN discovery */
fnic_fcoe_send_vlan_req(fnic);
dev_kfree_skb(skb);
continue;
}
fcoe_ctlr_recv(&fnic->ctlr, skb);
continue;
}
}
}
/**
* fnic_import_rq_eth_pkt() - handle received FCoE or FIP frame.
* @fnic: fnic instance.
* @skb: Ethernet Frame.
*/
static inline int fnic_import_rq_eth_pkt(struct fnic *fnic, struct sk_buff *skb)
{
struct fc_frame *fp;
struct ethhdr *eh;
struct fcoe_hdr *fcoe_hdr;
struct fcoe_crc_eof *ft;
/*
* Undo VLAN encapsulation if present.
*/
eh = (struct ethhdr *)skb->data;
if (eh->h_proto == htons(ETH_P_8021Q)) {
memmove((u8 *)eh + VLAN_HLEN, eh, ETH_ALEN * 2);
eh = skb_pull(skb, VLAN_HLEN);
skb_reset_mac_header(skb);
}
if (eh->h_proto == htons(ETH_P_FIP)) {
if (!(fnic->config.flags & VFCF_FIP_CAPABLE)) {
printk(KERN_ERR "Dropped FIP frame, as firmware "
"uses non-FIP mode, Enable FIP "
"using UCSM\n");
goto drop;
}
if ((fnic_fc_trace_set_data(fnic->lport->host->host_no,
FNIC_FC_RECV|0x80, (char *)skb->data, skb->len)) != 0) {
printk(KERN_ERR "fnic ctlr frame trace error!!!");
}
skb_queue_tail(&fnic->fip_frame_queue, skb);
queue_work(fnic_fip_queue, &fnic->fip_frame_work);
return 1; /* let caller know packet was used */
}
if (eh->h_proto != htons(ETH_P_FCOE))
goto drop;
skb_set_network_header(skb, sizeof(*eh));
skb_pull(skb, sizeof(*eh));
fcoe_hdr = (struct fcoe_hdr *)skb->data;
if (FC_FCOE_DECAPS_VER(fcoe_hdr) != FC_FCOE_VER)
goto drop;
fp = (struct fc_frame *)skb;
fc_frame_init(fp);
fr_sof(fp) = fcoe_hdr->fcoe_sof;
skb_pull(skb, sizeof(struct fcoe_hdr));
skb_reset_transport_header(skb);
ft = (struct fcoe_crc_eof *)(skb->data + skb->len - sizeof(*ft));
fr_eof(fp) = ft->fcoe_eof;
skb_trim(skb, skb->len - sizeof(*ft));
return 0;
drop:
dev_kfree_skb_irq(skb);
return -1;
}
/**
* fnic_update_mac_locked() - set data MAC address and filters.
* @fnic: fnic instance.
* @new: newly-assigned FCoE MAC address.
*
* Called with the fnic lock held.
*/
void fnic_update_mac_locked(struct fnic *fnic, u8 *new)
{
u8 *ctl = fnic->ctlr.ctl_src_addr;
u8 *data = fnic->data_src_addr;
if (is_zero_ether_addr(new))
new = ctl;
if (ether_addr_equal(data, new))
return;
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host, "update_mac %pM\n", new);
if (!is_zero_ether_addr(data) && !ether_addr_equal(data, ctl))
vnic_dev_del_addr(fnic->vdev, data);
memcpy(data, new, ETH_ALEN);
if (!ether_addr_equal(new, ctl))
vnic_dev_add_addr(fnic->vdev, new);
}
/**
* fnic_update_mac() - set data MAC address and filters.
* @lport: local port.
* @new: newly-assigned FCoE MAC address.
*/
void fnic_update_mac(struct fc_lport *lport, u8 *new)
{
struct fnic *fnic = lport_priv(lport);
spin_lock_irq(&fnic->fnic_lock);
fnic_update_mac_locked(fnic, new);
spin_unlock_irq(&fnic->fnic_lock);
}
/**
* fnic_set_port_id() - set the port_ID after successful FLOGI.
* @lport: local port.
* @port_id: assigned FC_ID.
* @fp: received frame containing the FLOGI accept or NULL.
*
* This is called from libfc when a new FC_ID has been assigned.
* This causes us to reset the firmware to FC_MODE and setup the new MAC
* address and FC_ID.
*
* It is also called with FC_ID 0 when we're logged off.
*
* If the FC_ID is due to point-to-point, fp may be NULL.
*/
void fnic_set_port_id(struct fc_lport *lport, u32 port_id, struct fc_frame *fp)
{
struct fnic *fnic = lport_priv(lport);
u8 *mac;
int ret;
FNIC_FCS_DBG(KERN_DEBUG, lport->host, "set port_id %x fp %p\n",
port_id, fp);
/*
* If we're clearing the FC_ID, change to use the ctl_src_addr.
* Set ethernet mode to send FLOGI.
*/
if (!port_id) {
fnic_update_mac(lport, fnic->ctlr.ctl_src_addr);
fnic_set_eth_mode(fnic);
return;
}
if (fp) {
mac = fr_cb(fp)->granted_mac;
if (is_zero_ether_addr(mac)) {
/* non-FIP - FLOGI already accepted - ignore return */
fcoe_ctlr_recv_flogi(&fnic->ctlr, lport, fp);
}
fnic_update_mac(lport, mac);
}
/* Change state to reflect transition to FC mode */
spin_lock_irq(&fnic->fnic_lock);
if (fnic->state == FNIC_IN_ETH_MODE || fnic->state == FNIC_IN_FC_MODE)
fnic->state = FNIC_IN_ETH_TRANS_FC_MODE;
else {
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"Unexpected fnic state %s while"
" processing flogi resp\n",
fnic_state_to_str(fnic->state));
spin_unlock_irq(&fnic->fnic_lock);
return;
}
spin_unlock_irq(&fnic->fnic_lock);
/*
* Send FLOGI registration to firmware to set up FC mode.
* The new address will be set up when registration completes.
*/
ret = fnic_flogi_reg_handler(fnic, port_id);
if (ret < 0) {
spin_lock_irq(&fnic->fnic_lock);
if (fnic->state == FNIC_IN_ETH_TRANS_FC_MODE)
fnic->state = FNIC_IN_ETH_MODE;
spin_unlock_irq(&fnic->fnic_lock);
}
}
static void fnic_rq_cmpl_frame_recv(struct vnic_rq *rq, struct cq_desc
*cq_desc, struct vnic_rq_buf *buf,
int skipped __attribute__((unused)),
void *opaque)
{
struct fnic *fnic = vnic_dev_priv(rq->vdev);
struct sk_buff *skb;
struct fc_frame *fp;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
u8 type, color, eop, sop, ingress_port, vlan_stripped;
u8 fcoe = 0, fcoe_sof, fcoe_eof;
u8 fcoe_fc_crc_ok = 1, fcoe_enc_error = 0;
u8 tcp_udp_csum_ok, udp, tcp, ipv4_csum_ok;
u8 ipv6, ipv4, ipv4_fragment, rss_type, csum_not_calc;
u8 fcs_ok = 1, packet_error = 0;
u16 q_number, completed_index, bytes_written = 0, vlan, checksum;
u32 rss_hash;
u16 exchange_id, tmpl;
u8 sof = 0;
u8 eof = 0;
u32 fcp_bytes_written = 0;
unsigned long flags;
dma_unmap_single(&fnic->pdev->dev, buf->dma_addr, buf->len,
DMA_FROM_DEVICE);
skb = buf->os_buf;
fp = (struct fc_frame *)skb;
buf->os_buf = NULL;
cq_desc_dec(cq_desc, &type, &color, &q_number, &completed_index);
if (type == CQ_DESC_TYPE_RQ_FCP) {
cq_fcp_rq_desc_dec((struct cq_fcp_rq_desc *)cq_desc,
&type, &color, &q_number, &completed_index,
&eop, &sop, &fcoe_fc_crc_ok, &exchange_id,
&tmpl, &fcp_bytes_written, &sof, &eof,
&ingress_port, &packet_error,
&fcoe_enc_error, &fcs_ok, &vlan_stripped,
&vlan);
skb_trim(skb, fcp_bytes_written);
fr_sof(fp) = sof;
fr_eof(fp) = eof;
} else if (type == CQ_DESC_TYPE_RQ_ENET) {
cq_enet_rq_desc_dec((struct cq_enet_rq_desc *)cq_desc,
&type, &color, &q_number, &completed_index,
&ingress_port, &fcoe, &eop, &sop,
&rss_type, &csum_not_calc, &rss_hash,
&bytes_written, &packet_error,
&vlan_stripped, &vlan, &checksum,
&fcoe_sof, &fcoe_fc_crc_ok,
&fcoe_enc_error, &fcoe_eof,
&tcp_udp_csum_ok, &udp, &tcp,
&ipv4_csum_ok, &ipv6, &ipv4,
&ipv4_fragment, &fcs_ok);
skb_trim(skb, bytes_written);
if (!fcs_ok) {
atomic64_inc(&fnic_stats->misc_stats.frame_errors);
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"fcs error. dropping packet.\n");
goto drop;
}
if (fnic_import_rq_eth_pkt(fnic, skb))
return;
} else {
/* wrong CQ type*/
shost_printk(KERN_ERR, fnic->lport->host,
"fnic rq_cmpl wrong cq type x%x\n", type);
goto drop;
}
if (!fcs_ok || packet_error || !fcoe_fc_crc_ok || fcoe_enc_error) {
atomic64_inc(&fnic_stats->misc_stats.frame_errors);
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"fnic rq_cmpl fcoe x%x fcsok x%x"
" pkterr x%x fcoe_fc_crc_ok x%x, fcoe_enc_err"
" x%x\n",
fcoe, fcs_ok, packet_error,
fcoe_fc_crc_ok, fcoe_enc_error);
goto drop;
}
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
goto drop;
}
fr_dev(fp) = fnic->lport;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
if ((fnic_fc_trace_set_data(fnic->lport->host->host_no, FNIC_FC_RECV,
(char *)skb->data, skb->len)) != 0) {
printk(KERN_ERR "fnic ctlr frame trace error!!!");
}
skb_queue_tail(&fnic->frame_queue, skb);
queue_work(fnic_event_queue, &fnic->frame_work);
return;
drop:
dev_kfree_skb_irq(skb);
}
static int fnic_rq_cmpl_handler_cont(struct vnic_dev *vdev,
struct cq_desc *cq_desc, u8 type,
u16 q_number, u16 completed_index,
void *opaque)
{
struct fnic *fnic = vnic_dev_priv(vdev);
vnic_rq_service(&fnic->rq[q_number], cq_desc, completed_index,
VNIC_RQ_RETURN_DESC, fnic_rq_cmpl_frame_recv,
NULL);
return 0;
}
int fnic_rq_cmpl_handler(struct fnic *fnic, int rq_work_to_do)
{
unsigned int tot_rq_work_done = 0, cur_work_done;
unsigned int i;
int err;
for (i = 0; i < fnic->rq_count; i++) {
cur_work_done = vnic_cq_service(&fnic->cq[i], rq_work_to_do,
fnic_rq_cmpl_handler_cont,
NULL);
if (cur_work_done) {
err = vnic_rq_fill(&fnic->rq[i], fnic_alloc_rq_frame);
if (err)
shost_printk(KERN_ERR, fnic->lport->host,
"fnic_alloc_rq_frame can't alloc"
" frame\n");
}
tot_rq_work_done += cur_work_done;
}
return tot_rq_work_done;
}
/*
* This function is called once at init time to allocate and fill RQ
* buffers. Subsequently, it is called in the interrupt context after RQ
* buffer processing to replenish the buffers in the RQ
*/
int fnic_alloc_rq_frame(struct vnic_rq *rq)
{
struct fnic *fnic = vnic_dev_priv(rq->vdev);
struct sk_buff *skb;
u16 len;
dma_addr_t pa;
int r;
len = FC_FRAME_HEADROOM + FC_MAX_FRAME + FC_FRAME_TAILROOM;
skb = dev_alloc_skb(len);
if (!skb) {
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"Unable to allocate RQ sk_buff\n");
return -ENOMEM;
}
skb_reset_mac_header(skb);
skb_reset_transport_header(skb);
skb_reset_network_header(skb);
skb_put(skb, len);
pa = dma_map_single(&fnic->pdev->dev, skb->data, len, DMA_FROM_DEVICE);
if (dma_mapping_error(&fnic->pdev->dev, pa)) {
r = -ENOMEM;
printk(KERN_ERR "PCI mapping failed with error %d\n", r);
goto free_skb;
}
fnic_queue_rq_desc(rq, skb, pa, len);
return 0;
free_skb:
kfree_skb(skb);
return r;
}
void fnic_free_rq_buf(struct vnic_rq *rq, struct vnic_rq_buf *buf)
{
struct fc_frame *fp = buf->os_buf;
struct fnic *fnic = vnic_dev_priv(rq->vdev);
dma_unmap_single(&fnic->pdev->dev, buf->dma_addr, buf->len,
DMA_FROM_DEVICE);
dev_kfree_skb(fp_skb(fp));
buf->os_buf = NULL;
}
/**
* fnic_eth_send() - Send Ethernet frame.
* @fip: fcoe_ctlr instance.
* @skb: Ethernet Frame, FIP, without VLAN encapsulation.
*/
void fnic_eth_send(struct fcoe_ctlr *fip, struct sk_buff *skb)
{
struct fnic *fnic = fnic_from_ctlr(fip);
struct vnic_wq *wq = &fnic->wq[0];
dma_addr_t pa;
struct ethhdr *eth_hdr;
struct vlan_ethhdr *vlan_hdr;
unsigned long flags;
if (!fnic->vlan_hw_insert) {
eth_hdr = (struct ethhdr *)skb_mac_header(skb);
vlan_hdr = skb_push(skb, sizeof(*vlan_hdr) - sizeof(*eth_hdr));
memcpy(vlan_hdr, eth_hdr, 2 * ETH_ALEN);
vlan_hdr->h_vlan_proto = htons(ETH_P_8021Q);
vlan_hdr->h_vlan_encapsulated_proto = eth_hdr->h_proto;
vlan_hdr->h_vlan_TCI = htons(fnic->vlan_id);
if ((fnic_fc_trace_set_data(fnic->lport->host->host_no,
FNIC_FC_SEND|0x80, (char *)eth_hdr, skb->len)) != 0) {
printk(KERN_ERR "fnic ctlr frame trace error!!!");
}
} else {
if ((fnic_fc_trace_set_data(fnic->lport->host->host_no,
FNIC_FC_SEND|0x80, (char *)skb->data, skb->len)) != 0) {
printk(KERN_ERR "fnic ctlr frame trace error!!!");
}
}
pa = dma_map_single(&fnic->pdev->dev, skb->data, skb->len,
DMA_TO_DEVICE);
if (dma_mapping_error(&fnic->pdev->dev, pa)) {
printk(KERN_ERR "DMA mapping failed\n");
goto free_skb;
}
spin_lock_irqsave(&fnic->wq_lock[0], flags);
if (!vnic_wq_desc_avail(wq))
goto irq_restore;
fnic_queue_wq_eth_desc(wq, skb, pa, skb->len,
0 /* hw inserts cos value */,
fnic->vlan_id, 1);
spin_unlock_irqrestore(&fnic->wq_lock[0], flags);
return;
irq_restore:
spin_unlock_irqrestore(&fnic->wq_lock[0], flags);
dma_unmap_single(&fnic->pdev->dev, pa, skb->len, DMA_TO_DEVICE);
free_skb:
kfree_skb(skb);
}
/*
* Send FC frame.
*/
static int fnic_send_frame(struct fnic *fnic, struct fc_frame *fp)
{
struct vnic_wq *wq = &fnic->wq[0];
struct sk_buff *skb;
dma_addr_t pa;
struct ethhdr *eth_hdr;
struct vlan_ethhdr *vlan_hdr;
struct fcoe_hdr *fcoe_hdr;
struct fc_frame_header *fh;
u32 tot_len, eth_hdr_len;
int ret = 0;
unsigned long flags;
fh = fc_frame_header_get(fp);
skb = fp_skb(fp);
if (unlikely(fh->fh_r_ctl == FC_RCTL_ELS_REQ) &&
fcoe_ctlr_els_send(&fnic->ctlr, fnic->lport, skb))
return 0;
if (!fnic->vlan_hw_insert) {
eth_hdr_len = sizeof(*vlan_hdr) + sizeof(*fcoe_hdr);
vlan_hdr = skb_push(skb, eth_hdr_len);
eth_hdr = (struct ethhdr *)vlan_hdr;
vlan_hdr->h_vlan_proto = htons(ETH_P_8021Q);
vlan_hdr->h_vlan_encapsulated_proto = htons(ETH_P_FCOE);
vlan_hdr->h_vlan_TCI = htons(fnic->vlan_id);
fcoe_hdr = (struct fcoe_hdr *)(vlan_hdr + 1);
} else {
eth_hdr_len = sizeof(*eth_hdr) + sizeof(*fcoe_hdr);
eth_hdr = skb_push(skb, eth_hdr_len);
eth_hdr->h_proto = htons(ETH_P_FCOE);
fcoe_hdr = (struct fcoe_hdr *)(eth_hdr + 1);
}
if (fnic->ctlr.map_dest)
fc_fcoe_set_mac(eth_hdr->h_dest, fh->fh_d_id);
else
memcpy(eth_hdr->h_dest, fnic->ctlr.dest_addr, ETH_ALEN);
memcpy(eth_hdr->h_source, fnic->data_src_addr, ETH_ALEN);
tot_len = skb->len;
BUG_ON(tot_len % 4);
memset(fcoe_hdr, 0, sizeof(*fcoe_hdr));
fcoe_hdr->fcoe_sof = fr_sof(fp);
if (FC_FCOE_VER)
FC_FCOE_ENCAPS_VER(fcoe_hdr, FC_FCOE_VER);
pa = dma_map_single(&fnic->pdev->dev, eth_hdr, tot_len, DMA_TO_DEVICE);
if (dma_mapping_error(&fnic->pdev->dev, pa)) {
ret = -ENOMEM;
printk(KERN_ERR "DMA map failed with error %d\n", ret);
goto free_skb_on_err;
}
if ((fnic_fc_trace_set_data(fnic->lport->host->host_no, FNIC_FC_SEND,
(char *)eth_hdr, tot_len)) != 0) {
printk(KERN_ERR "fnic ctlr frame trace error!!!");
}
spin_lock_irqsave(&fnic->wq_lock[0], flags);
if (!vnic_wq_desc_avail(wq)) {
dma_unmap_single(&fnic->pdev->dev, pa, tot_len, DMA_TO_DEVICE);
ret = -1;
goto irq_restore;
}
fnic_queue_wq_desc(wq, skb, pa, tot_len, fr_eof(fp),
0 /* hw inserts cos value */,
fnic->vlan_id, 1, 1, 1);
irq_restore:
spin_unlock_irqrestore(&fnic->wq_lock[0], flags);
free_skb_on_err:
if (ret)
dev_kfree_skb_any(fp_skb(fp));
return ret;
}
/*
* fnic_send
* Routine to send a raw frame
*/
int fnic_send(struct fc_lport *lp, struct fc_frame *fp)
{
struct fnic *fnic = lport_priv(lp);
unsigned long flags;
if (fnic->in_remove) {
dev_kfree_skb(fp_skb(fp));
return -1;
}
/*
* Queue frame if in a transitional state.
* This occurs while registering the Port_ID / MAC address after FLOGI.
*/
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->state != FNIC_IN_FC_MODE && fnic->state != FNIC_IN_ETH_MODE) {
skb_queue_tail(&fnic->tx_queue, fp_skb(fp));
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return 0;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return fnic_send_frame(fnic, fp);
}
/**
* fnic_flush_tx() - send queued frames.
* @fnic: fnic device
*
* Send frames that were waiting to go out in FC or Ethernet mode.
* Whenever changing modes we purge queued frames, so these frames should
* be queued for the stable mode that we're in, either FC or Ethernet.
*
* Called without fnic_lock held.
*/
void fnic_flush_tx(struct fnic *fnic)
{
struct sk_buff *skb;
struct fc_frame *fp;
while ((skb = skb_dequeue(&fnic->tx_queue))) {
fp = (struct fc_frame *)skb;
fnic_send_frame(fnic, fp);
}
}
/**
* fnic_set_eth_mode() - put fnic into ethernet mode.
* @fnic: fnic device
*
* Called without fnic lock held.
*/
static void fnic_set_eth_mode(struct fnic *fnic)
{
unsigned long flags;
enum fnic_state old_state;
int ret;
spin_lock_irqsave(&fnic->fnic_lock, flags);
again:
old_state = fnic->state;
switch (old_state) {
case FNIC_IN_FC_MODE:
case FNIC_IN_ETH_TRANS_FC_MODE:
default:
fnic->state = FNIC_IN_FC_TRANS_ETH_MODE;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
ret = fnic_fw_reset_handler(fnic);
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->state != FNIC_IN_FC_TRANS_ETH_MODE)
goto again;
if (ret)
fnic->state = old_state;
break;
case FNIC_IN_FC_TRANS_ETH_MODE:
case FNIC_IN_ETH_MODE:
break;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
static void fnic_wq_complete_frame_send(struct vnic_wq *wq,
struct cq_desc *cq_desc,
struct vnic_wq_buf *buf, void *opaque)
{
struct sk_buff *skb = buf->os_buf;
struct fc_frame *fp = (struct fc_frame *)skb;
struct fnic *fnic = vnic_dev_priv(wq->vdev);
dma_unmap_single(&fnic->pdev->dev, buf->dma_addr, buf->len,
DMA_TO_DEVICE);
dev_kfree_skb_irq(fp_skb(fp));
buf->os_buf = NULL;
}
static int fnic_wq_cmpl_handler_cont(struct vnic_dev *vdev,
struct cq_desc *cq_desc, u8 type,
u16 q_number, u16 completed_index,
void *opaque)
{
struct fnic *fnic = vnic_dev_priv(vdev);
unsigned long flags;
spin_lock_irqsave(&fnic->wq_lock[q_number], flags);
vnic_wq_service(&fnic->wq[q_number], cq_desc, completed_index,
fnic_wq_complete_frame_send, NULL);
spin_unlock_irqrestore(&fnic->wq_lock[q_number], flags);
return 0;
}
int fnic_wq_cmpl_handler(struct fnic *fnic, int work_to_do)
{
unsigned int wq_work_done = 0;
unsigned int i;
for (i = 0; i < fnic->raw_wq_count; i++) {
wq_work_done += vnic_cq_service(&fnic->cq[fnic->rq_count+i],
work_to_do,
fnic_wq_cmpl_handler_cont,
NULL);
}
return wq_work_done;
}
void fnic_free_wq_buf(struct vnic_wq *wq, struct vnic_wq_buf *buf)
{
struct fc_frame *fp = buf->os_buf;
struct fnic *fnic = vnic_dev_priv(wq->vdev);
dma_unmap_single(&fnic->pdev->dev, buf->dma_addr, buf->len,
DMA_TO_DEVICE);
dev_kfree_skb(fp_skb(fp));
buf->os_buf = NULL;
}
void fnic_fcoe_reset_vlans(struct fnic *fnic)
{
unsigned long flags;
struct fcoe_vlan *vlan;
struct fcoe_vlan *next;
/*
* indicate a link down to fcoe so that all fcf's are free'd
* might not be required since we did this before sending vlan
* discovery request
*/
spin_lock_irqsave(&fnic->vlans_lock, flags);
if (!list_empty(&fnic->vlans)) {
list_for_each_entry_safe(vlan, next, &fnic->vlans, list) {
list_del(&vlan->list);
kfree(vlan);
}
}
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
}
void fnic_handle_fip_timer(struct fnic *fnic)
{
unsigned long flags;
struct fcoe_vlan *vlan;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
u64 sol_time;
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
if (fnic->ctlr.mode == FIP_MODE_NON_FIP)
return;
spin_lock_irqsave(&fnic->vlans_lock, flags);
if (list_empty(&fnic->vlans)) {
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
/* no vlans available, try again */
if (unlikely(fnic_log_level & FNIC_FCS_LOGGING))
if (printk_ratelimit())
shost_printk(KERN_DEBUG, fnic->lport->host,
"Start VLAN Discovery\n");
fnic_event_enq(fnic, FNIC_EVT_START_VLAN_DISC);
return;
}
vlan = list_first_entry(&fnic->vlans, struct fcoe_vlan, list);
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"fip_timer: vlan %d state %d sol_count %d\n",
vlan->vid, vlan->state, vlan->sol_count);
switch (vlan->state) {
case FIP_VLAN_USED:
FNIC_FCS_DBG(KERN_DEBUG, fnic->lport->host,
"FIP VLAN is selected for FC transaction\n");
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
break;
case FIP_VLAN_FAILED:
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
/* if all vlans are in failed state, restart vlan disc */
if (unlikely(fnic_log_level & FNIC_FCS_LOGGING))
if (printk_ratelimit())
shost_printk(KERN_DEBUG, fnic->lport->host,
"Start VLAN Discovery\n");
fnic_event_enq(fnic, FNIC_EVT_START_VLAN_DISC);
break;
case FIP_VLAN_SENT:
if (vlan->sol_count >= FCOE_CTLR_MAX_SOL) {
/*
* no response on this vlan, remove from the list.
* Try the next vlan
*/
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"Dequeue this VLAN ID %d from list\n",
vlan->vid);
list_del(&vlan->list);
kfree(vlan);
vlan = NULL;
if (list_empty(&fnic->vlans)) {
/* we exhausted all vlans, restart vlan disc */
spin_unlock_irqrestore(&fnic->vlans_lock,
flags);
FNIC_FCS_DBG(KERN_INFO, fnic->lport->host,
"fip_timer: vlan list empty, "
"trigger vlan disc\n");
fnic_event_enq(fnic, FNIC_EVT_START_VLAN_DISC);
return;
}
/* check the next vlan */
vlan = list_first_entry(&fnic->vlans, struct fcoe_vlan,
list);
fnic->set_vlan(fnic, vlan->vid);
vlan->state = FIP_VLAN_SENT; /* sent now */
}
spin_unlock_irqrestore(&fnic->vlans_lock, flags);
atomic64_inc(&fnic_stats->vlan_stats.sol_expiry_count);
vlan->sol_count++;
sol_time = jiffies + msecs_to_jiffies
(FCOE_CTLR_START_DELAY);
mod_timer(&fnic->fip_timer, round_jiffies(sol_time));
break;
}
}