drivers/net/xen-netfront.c

   1 /*
   2  * Virtual network driver for conversing with remote driver backends.
   3  *
   4  * Copyright (c) 2002-2005, K A Fraser
   5  * Copyright (c) 2005, XenSource Ltd
   6  *
   7  * This program is free software; you can redistribute it and/or
   8  * modify it under the terms of the GNU General Public License version 2
   9  * as published by the Free Software Foundation; or, when distributed
  10  * separately from the Linux kernel or incorporated into other
  11  * software packages, subject to the following license:
  12  *
  13  * Permission is hereby granted, free of charge, to any person obtaining a copy
  14  * of this source file (the "Software"), to deal in the Software without
  15  * restriction, including without limitation the rights to use, copy, modify,
  16  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
  17  * and to permit persons to whom the Software is furnished to do so, subject to
  18  * the following conditions:
  19  *
  20  * The above copyright notice and this permission notice shall be included in
  21  * all copies or substantial portions of the Software.
  22  *
  23  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  24  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  25  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  26  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  27  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  28  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  29  * IN THE SOFTWARE.
  30  */
  31
  32 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
  33
  34 #include <linux/module.h>
  35 #include <linux/kernel.h>
  36 #include <linux/netdevice.h>
  37 #include <linux/etherdevice.h>
  38 #include <linux/skbuff.h>
  39 #include <linux/ethtool.h>
  40 #include <linux/if_ether.h>
  41 #include <net/tcp.h>
  42 #include <linux/udp.h>
  43 #include <linux/moduleparam.h>
  44 #include <linux/mm.h>
  45 #include <linux/slab.h>
  46 #include <net/ip.h>
  47
  48 #include <xen/xen.h>
  49 #include <xen/xenbus.h>
  50 #include <xen/events.h>
  51 #include <xen/page.h>
  52 #include <xen/platform_pci.h>
  53 #include <xen/grant_table.h>
  54
  55 #include <xen/interface/io/netif.h>
  56 #include <xen/interface/memory.h>
  57 #include <xen/interface/grant_table.h>
  58
  59 /* Module parameters */
  60 #define MAX_QUEUES_DEFAULT 8
  61 static unsigned int xennet_max_queues;
  62 module_param_named(max_queues, xennet_max_queues, uint, 0644);
  63 MODULE_PARM_DESC(max_queues,
  64                  "Maximum number of queues per virtual interface");
  65
  66 static const struct ethtool_ops xennet_ethtool_ops;
  67
  68 struct netfront_cb {
  69         int pull_to;
  70 };
  71
  72 #define NETFRONT_SKB_CB(skb)    ((struct netfront_cb *)((skb)->cb))
  73
  74 #define RX_COPY_THRESHOLD 256
  75
  76 #define GRANT_INVALID_REF       0
  77
  78 #define NET_TX_RING_SIZE __CONST_RING_SIZE(xen_netif_tx, XEN_PAGE_SIZE)
  79 #define NET_RX_RING_SIZE __CONST_RING_SIZE(xen_netif_rx, XEN_PAGE_SIZE)
  80
  81 /* Minimum number of Rx slots (includes slot for GSO metadata). */
  82 #define NET_RX_SLOTS_MIN (XEN_NETIF_NR_SLOTS_MIN + 1)
  83
  84 /* Queue name is interface name with "-qNNN" appended */
  85 #define QUEUE_NAME_SIZE (IFNAMSIZ + 6)
  86
  87 /* IRQ name is queue name with "-tx" or "-rx" appended */
  88 #define IRQ_NAME_SIZE (QUEUE_NAME_SIZE + 3)
  89
  90 static DECLARE_WAIT_QUEUE_HEAD(module_load_q);
  91 static DECLARE_WAIT_QUEUE_HEAD(module_unload_q);
  92
  93 struct netfront_stats {
  94         u64                     packets;
  95         u64                     bytes;
  96         struct u64_stats_sync   syncp;
  97 };
  98
  99 struct netfront_info;
 100
 101 struct netfront_queue {
 102         unsigned int id; /* Queue ID, 0-based */
 103         char name[QUEUE_NAME_SIZE]; /* DEVNAME-qN */
 104         struct netfront_info *info;
 105
 106         struct napi_struct napi;
 107
 108         /* Split event channels support, tx_* == rx_* when using
 109          * single event channel.
 110          */
 111         unsigned int tx_evtchn, rx_evtchn;
 112         unsigned int tx_irq, rx_irq;
 113         /* Only used when split event channels support is enabled */
 114         char tx_irq_name[IRQ_NAME_SIZE]; /* DEVNAME-qN-tx */
 115         char rx_irq_name[IRQ_NAME_SIZE]; /* DEVNAME-qN-rx */
 116
 117         spinlock_t   tx_lock;
 118         struct xen_netif_tx_front_ring tx;
 119         int tx_ring_ref;
 120
 121         /*
 122          * {tx,rx}_skbs store outstanding skbuffs. Free tx_skb entries
 123          * are linked from tx_skb_freelist through skb_entry.link.
 124          *
 125          *  NB. Freelist index entries are always going to be less than
 126          *  PAGE_OFFSET, whereas pointers to skbs will always be equal or
 127          *  greater than PAGE_OFFSET: we use this property to distinguish
 128          *  them.
 129          */
 130         union skb_entry {
 131                 struct sk_buff *skb;
 132                 unsigned long link;
 133         } tx_skbs[NET_TX_RING_SIZE];
 134         grant_ref_t gref_tx_head;
 135         grant_ref_t grant_tx_ref[NET_TX_RING_SIZE];
 136         struct page *grant_tx_page[NET_TX_RING_SIZE];
 137         unsigned tx_skb_freelist;
 138
 139         spinlock_t   rx_lock ____cacheline_aligned_in_smp;
 140         struct xen_netif_rx_front_ring rx;
 141         int rx_ring_ref;
 142
 143         struct timer_list rx_refill_timer;
 144
 145         struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
 146         grant_ref_t gref_rx_head;
 147         grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];
 148 };
 149
 150 struct netfront_info {
 151         struct list_head list;
 152         struct net_device *netdev;
 153
 154         struct xenbus_device *xbdev;
 155
 156         /* Multi-queue support */
 157         struct netfront_queue *queues;
 158
 159         /* Statistics */
 160         struct netfront_stats __percpu *rx_stats;
 161         struct netfront_stats __percpu *tx_stats;
 162
 163         atomic_t rx_gso_checksum_fixup;
 164 };
 165
 166 struct netfront_rx_info {
 167         struct xen_netif_rx_response rx;
 168         struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
 169 };
 170
 171 static void skb_entry_set_link(union skb_entry *list, unsigned short id)
 172 {
 173         list->link = id;
 174 }
 175
 176 static int skb_entry_is_link(const union skb_entry *list)
 177 {
 178         BUILD_BUG_ON(sizeof(list->skb) != sizeof(list->link));
 179         return (unsigned long)list->skb < PAGE_OFFSET;
 180 }
 181
 182 /*
 183  * Access macros for acquiring freeing slots in tx_skbs[].
 184  */
 185
 186 static void add_id_to_freelist(unsigned *head, union skb_entry *list,
 187                                unsigned short id)
 188 {
 189         skb_entry_set_link(&list[id], *head);
 190         *head = id;
 191 }
 192
 193 static unsigned short get_id_from_freelist(unsigned *head,
 194                                            union skb_entry *list)
 195 {
 196         unsigned int id = *head;
 197         *head = list[id].link;
 198         return id;
 199 }
 200
 201 static int xennet_rxidx(RING_IDX idx)
 202 {
 203         return idx & (NET_RX_RING_SIZE - 1);
 204 }
 205
 206 static struct sk_buff *xennet_get_rx_skb(struct netfront_queue *queue,
 207                                          RING_IDX ri)
 208 {
 209         int i = xennet_rxidx(ri);
 210         struct sk_buff *skb = queue->rx_skbs[i];
 211         queue->rx_skbs[i] = NULL;
 212         return skb;
 213 }
 214
 215 static grant_ref_t xennet_get_rx_ref(struct netfront_queue *queue,
 216                                             RING_IDX ri)
 217 {
 218         int i = xennet_rxidx(ri);
 219         grant_ref_t ref = queue->grant_rx_ref[i];
 220         queue->grant_rx_ref[i] = GRANT_INVALID_REF;
 221         return ref;
 222 }
 223
 224 #ifdef CONFIG_SYSFS
 225 static const struct attribute_group xennet_dev_group;
 226 #endif
 227
 228 static bool xennet_can_sg(struct net_device *dev)
 229 {
 230         return dev->features & NETIF_F_SG;
 231 }
 232
 233
 234 static void rx_refill_timeout(struct timer_list *t)
 235 {
 236         struct netfront_queue *queue = from_timer(queue, t, rx_refill_timer);
 237         napi_schedule(&queue->napi);
 238 }
 239
 240 static int netfront_tx_slot_available(struct netfront_queue *queue)
 241 {
 242         return (queue->tx.req_prod_pvt - queue->tx.rsp_cons) <
 243                 (NET_TX_RING_SIZE - XEN_NETIF_NR_SLOTS_MIN - 1);
 244 }
 245
 246 static void xennet_maybe_wake_tx(struct netfront_queue *queue)
 247 {
 248         struct net_device *dev = queue->info->netdev;
 249         struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, queue->id);
 250
 251         if (unlikely(netif_tx_queue_stopped(dev_queue)) &&
 252             netfront_tx_slot_available(queue) &&
 253             likely(netif_running(dev)))
 254                 netif_tx_wake_queue(netdev_get_tx_queue(dev, queue->id));
 255 }
 256
 257
 258 static struct sk_buff *xennet_alloc_one_rx_buffer(struct netfront_queue *queue)
 259 {
 260         struct sk_buff *skb;
 261         struct page *page;
 262
 263         skb = __netdev_alloc_skb(queue->info->netdev,
 264                                  RX_COPY_THRESHOLD + NET_IP_ALIGN,
 265                                  GFP_ATOMIC | __GFP_NOWARN);
 266         if (unlikely(!skb))
 267                 return NULL;
 268
 269         page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
 270         if (!page) {
 271                 kfree_skb(skb);
 272                 return NULL;
 273         }
 274         skb_add_rx_frag(skb, 0, page, 0, 0, PAGE_SIZE);
 275
 276         /* Align ip header to a 16 bytes boundary */
 277         skb_reserve(skb, NET_IP_ALIGN);
 278         skb->dev = queue->info->netdev;
 279
 280         return skb;
 281 }
 282
 283
 284 static void xennet_alloc_rx_buffers(struct netfront_queue *queue)
 285 {
 286         RING_IDX req_prod = queue->rx.req_prod_pvt;
 287         int notify;
 288         int err = 0;
 289
 290         if (unlikely(!netif_carrier_ok(queue->info->netdev)))
 291                 return;
 292
 293         for (req_prod = queue->rx.req_prod_pvt;
 294              req_prod - queue->rx.rsp_cons < NET_RX_RING_SIZE;
 295              req_prod++) {
 296                 struct sk_buff *skb;
 297                 unsigned short id;
 298                 grant_ref_t ref;
 299                 struct page *page;
 300                 struct xen_netif_rx_request *req;
 301
 302                 skb = xennet_alloc_one_rx_buffer(queue);
 303                 if (!skb) {
 304                         err = -ENOMEM;
 305                         break;
 306                 }
 307
 308                 id = xennet_rxidx(req_prod);
 309
 310                 BUG_ON(queue->rx_skbs[id]);
 311                 queue->rx_skbs[id] = skb;
 312
 313                 ref = gnttab_claim_grant_reference(&queue->gref_rx_head);
 314                 WARN_ON_ONCE(IS_ERR_VALUE((unsigned long)(int)ref));
 315                 queue->grant_rx_ref[id] = ref;
 316
 317                 page = skb_frag_page(&skb_shinfo(skb)->frags[0]);
 318
 319                 req = RING_GET_REQUEST(&queue->rx, req_prod);
 320                 gnttab_page_grant_foreign_access_ref_one(ref,
 321                                                          queue->info->xbdev->otherend_id,
 322                                                          page,
 323                                                          0);
 324                 req->id = id;
 325                 req->gref = ref;
 326         }
 327
 328         queue->rx.req_prod_pvt = req_prod;
 329
 330         /* Try again later if there are not enough requests or skb allocation
 331          * failed.
 332          * Enough requests is quantified as the sum of newly created slots and
 333          * the unconsumed slots at the backend.
 334          */
 335         if (req_prod - queue->rx.rsp_cons < NET_RX_SLOTS_MIN ||
 336             unlikely(err)) {
 337                 mod_timer(&queue->rx_refill_timer, jiffies + (HZ/10));
 338                 return;
 339         }
 340
 341         wmb();          /* barrier so backend seens requests */
 342
 343         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->rx, notify);
 344         if (notify)
 345                 notify_remote_via_irq(queue->rx_irq);
 346 }
 347
 348 static int xennet_open(struct net_device *dev)
 349 {
 350         struct netfront_info *np = netdev_priv(dev);
 351         unsigned int num_queues = dev->real_num_tx_queues;
 352         unsigned int i = 0;
 353         struct netfront_queue *queue = NULL;
 354
 355         if (!np->queues)
 356                 return -ENODEV;
 357
 358         for (i = 0; i < num_queues; ++i) {
 359                 queue = &np->queues[i];
 360                 napi_enable(&queue->napi);
 361
 362                 spin_lock_bh(&queue->rx_lock);
 363                 if (netif_carrier_ok(dev)) {
 364                         xennet_alloc_rx_buffers(queue);
 365                         queue->rx.sring->rsp_event = queue->rx.rsp_cons + 1;
 366                         if (RING_HAS_UNCONSUMED_RESPONSES(&queue->rx))
 367                                 napi_schedule(&queue->napi);
 368                 }
 369                 spin_unlock_bh(&queue->rx_lock);
 370         }
 371
 372         netif_tx_start_all_queues(dev);
 373
 374         return 0;
 375 }
 376
 377 static void xennet_tx_buf_gc(struct netfront_queue *queue)
 378 {
 379         RING_IDX cons, prod;
 380         unsigned short id;
 381         struct sk_buff *skb;
 382         bool more_to_do;
 383
 384         BUG_ON(!netif_carrier_ok(queue->info->netdev));
 385
 386         do {
 387                 prod = queue->tx.sring->rsp_prod;
 388                 rmb(); /* Ensure we see responses up to 'rp'. */
 389
 390                 for (cons = queue->tx.rsp_cons; cons != prod; cons++) {
 391                         struct xen_netif_tx_response *txrsp;
 392
 393                         txrsp = RING_GET_RESPONSE(&queue->tx, cons);
 394                         if (txrsp->status == XEN_NETIF_RSP_NULL)
 395                                 continue;
 396
 397                         id  = txrsp->id;
 398                         skb = queue->tx_skbs[id].skb;
 399                         if (unlikely(gnttab_query_foreign_access(
 400                                 queue->grant_tx_ref[id]) != 0)) {
 401                                 pr_alert("%s: warning -- grant still in use by backend domain\n",
 402                                          __func__);
 403                                 BUG();
 404                         }
 405                         gnttab_end_foreign_access_ref(
 406                                 queue->grant_tx_ref[id], GNTMAP_readonly);
 407                         gnttab_release_grant_reference(
 408                                 &queue->gref_tx_head, queue->grant_tx_ref[id]);
 409                         queue->grant_tx_ref[id] = GRANT_INVALID_REF;
 410                         queue->grant_tx_page[id] = NULL;
 411                         add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, id);
 412                         dev_kfree_skb_irq(skb);
 413                 }
 414
 415                 queue->tx.rsp_cons = prod;
 416
 417                 RING_FINAL_CHECK_FOR_RESPONSES(&queue->tx, more_to_do);
 418         } while (more_to_do);
 419
 420         xennet_maybe_wake_tx(queue);
 421 }
 422
 423 struct xennet_gnttab_make_txreq {
 424         struct netfront_queue *queue;
 425         struct sk_buff *skb;
 426         struct page *page;
 427         struct xen_netif_tx_request *tx; /* Last request */
 428         unsigned int size;
 429 };
 430
 431 static void xennet_tx_setup_grant(unsigned long gfn, unsigned int offset,
 432                                   unsigned int len, void *data)
 433 {
 434         struct xennet_gnttab_make_txreq *info = data;
 435         unsigned int id;
 436         struct xen_netif_tx_request *tx;
 437         grant_ref_t ref;
 438         /* convenient aliases */
 439         struct page *page = info->page;
 440         struct netfront_queue *queue = info->queue;
 441         struct sk_buff *skb = info->skb;
 442
 443         id = get_id_from_freelist(&queue->tx_skb_freelist, queue->tx_skbs);
 444         tx = RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 445         ref = gnttab_claim_grant_reference(&queue->gref_tx_head);
 446         WARN_ON_ONCE(IS_ERR_VALUE((unsigned long)(int)ref));
 447
 448         gnttab_grant_foreign_access_ref(ref, queue->info->xbdev->otherend_id,
 449                                         gfn, GNTMAP_readonly);
 450
 451         queue->tx_skbs[id].skb = skb;
 452         queue->grant_tx_page[id] = page;
 453         queue->grant_tx_ref[id] = ref;
 454
 455         tx->id = id;
 456         tx->gref = ref;
 457         tx->offset = offset;
 458         tx->size = len;
 459         tx->flags = 0;
 460
 461         info->tx = tx;
 462         info->size += tx->size;
 463 }
 464
 465 static struct xen_netif_tx_request *xennet_make_first_txreq(
 466         struct netfront_queue *queue, struct sk_buff *skb,
 467         struct page *page, unsigned int offset, unsigned int len)
 468 {
 469         struct xennet_gnttab_make_txreq info = {
 470                 .queue = queue,
 471                 .skb = skb,
 472                 .page = page,
 473                 .size = 0,
 474         };
 475
 476         gnttab_for_one_grant(page, offset, len, xennet_tx_setup_grant, &info);
 477
 478         return info.tx;
 479 }
 480
 481 static void xennet_make_one_txreq(unsigned long gfn, unsigned int offset,
 482                                   unsigned int len, void *data)
 483 {
 484         struct xennet_gnttab_make_txreq *info = data;
 485
 486         info->tx->flags |= XEN_NETTXF_more_data;
 487         skb_get(info->skb);
 488         xennet_tx_setup_grant(gfn, offset, len, data);
 489 }
 490
 491 static struct xen_netif_tx_request *xennet_make_txreqs(
 492         struct netfront_queue *queue, struct xen_netif_tx_request *tx,
 493         struct sk_buff *skb, struct page *page,
 494         unsigned int offset, unsigned int len)
 495 {
 496         struct xennet_gnttab_make_txreq info = {
 497                 .queue = queue,
 498                 .skb = skb,
 499                 .tx = tx,
 500         };
 501
 502         /* Skip unused frames from start of page */
 503         page += offset >> PAGE_SHIFT;
 504         offset &= ~PAGE_MASK;
 505
 506         while (len) {
 507                 info.page = page;
 508                 info.size = 0;
 509
 510                 gnttab_foreach_grant_in_range(page, offset, len,
 511                                               xennet_make_one_txreq,
 512                                               &info);
 513
 514                 page++;
 515                 offset = 0;
 516                 len -= info.size;
 517         }
 518
 519         return info.tx;
 520 }
 521
 522 /*
 523  * Count how many ring slots are required to send this skb. Each frag
 524  * might be a compound page.
 525  */
 526 static int xennet_count_skb_slots(struct sk_buff *skb)
 527 {
 528         int i, frags = skb_shinfo(skb)->nr_frags;
 529         int slots;
 530
 531         slots = gnttab_count_grant(offset_in_page(skb->data),
 532                                    skb_headlen(skb));
 533
 534         for (i = 0; i < frags; i++) {
 535                 skb_frag_t *frag = skb_shinfo(skb)->frags + i;
 536                 unsigned long size = skb_frag_size(frag);
 537                 unsigned long offset = frag->page_offset;
 538
 539                 /* Skip unused frames from start of page */
 540                 offset &= ~PAGE_MASK;
 541
 542                 slots += gnttab_count_grant(offset, size);
 543         }
 544
 545         return slots;
 546 }
 547
 548 static u16 xennet_select_queue(struct net_device *dev, struct sk_buff *skb,
 549                                struct net_device *sb_dev,
 550                                select_queue_fallback_t fallback)
 551 {
 552         unsigned int num_queues = dev->real_num_tx_queues;
 553         u32 hash;
 554         u16 queue_idx;
 555
 556         /* First, check if there is only one queue */
 557         if (num_queues == 1) {
 558                 queue_idx = 0;
 559         } else {
 560                 hash = skb_get_hash(skb);
 561                 queue_idx = hash % num_queues;
 562         }
 563
 564         return queue_idx;
 565 }
 566
 567 #define MAX_XEN_SKB_FRAGS (65536 / XEN_PAGE_SIZE + 1)
 568
 569 static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 570 {
 571         struct netfront_info *np = netdev_priv(dev);
 572         struct netfront_stats *tx_stats = this_cpu_ptr(np->tx_stats);
 573         struct xen_netif_tx_request *tx, *first_tx;
 574         unsigned int i;
 575         int notify;
 576         int slots;
 577         struct page *page;
 578         unsigned int offset;
 579         unsigned int len;
 580         unsigned long flags;
 581         struct netfront_queue *queue = NULL;
 582         unsigned int num_queues = dev->real_num_tx_queues;
 583         u16 queue_index;
 584         struct sk_buff *nskb;
 585
 586         /* Drop the packet if no queues are set up */
 587         if (num_queues < 1)
 588                 goto drop;
 589         /* Determine which queue to transmit this SKB on */
 590         queue_index = skb_get_queue_mapping(skb);
 591         queue = &np->queues[queue_index];
 592
 593         /* If skb->len is too big for wire format, drop skb and alert
 594          * user about misconfiguration.
 595          */
 596         if (unlikely(skb->len > XEN_NETIF_MAX_TX_SIZE)) {
 597                 net_alert_ratelimited(
 598                         "xennet: skb->len = %u, too big for wire format\n",
 599                         skb->len);
 600                 goto drop;
 601         }
 602
 603         slots = xennet_count_skb_slots(skb);
 604         if (unlikely(slots > MAX_XEN_SKB_FRAGS + 1)) {
 605                 net_dbg_ratelimited("xennet: skb rides the rocket: %d slots, %d bytes\n",
 606                                     slots, skb->len);
 607                 if (skb_linearize(skb))
 608                         goto drop;
 609         }
 610
 611         page = virt_to_page(skb->data);
 612         offset = offset_in_page(skb->data);
 613
 614         /* The first req should be at least ETH_HLEN size or the packet will be
 615          * dropped by netback.
 616          */
 617         if (unlikely(PAGE_SIZE - offset < ETH_HLEN)) {
 618                 nskb = skb_copy(skb, GFP_ATOMIC);
 619                 if (!nskb)
 620                         goto drop;
 621                 dev_consume_skb_any(skb);
 622                 skb = nskb;
 623                 page = virt_to_page(skb->data);
 624                 offset = offset_in_page(skb->data);
 625         }
 626
 627         len = skb_headlen(skb);
 628
 629         spin_lock_irqsave(&queue->tx_lock, flags);
 630
 631         if (unlikely(!netif_carrier_ok(dev) ||
 632                      (slots > 1 && !xennet_can_sg(dev)) ||
 633                      netif_needs_gso(skb, netif_skb_features(skb)))) {
 634                 spin_unlock_irqrestore(&queue->tx_lock, flags);
 635                 goto drop;
 636         }
 637
 638         /* First request for the linear area. */
 639         first_tx = tx = xennet_make_first_txreq(queue, skb,
 640                                                 page, offset, len);
 641         offset += tx->size;
 642         if (offset == PAGE_SIZE) {
 643                 page++;
 644                 offset = 0;
 645         }
 646         len -= tx->size;
 647
 648         if (skb->ip_summed == CHECKSUM_PARTIAL)
 649                 /* local packet? */
 650                 tx->flags |= XEN_NETTXF_csum_blank | XEN_NETTXF_data_validated;
 651         else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
 652                 /* remote but checksummed. */
 653                 tx->flags |= XEN_NETTXF_data_validated;
 654
 655         /* Optional extra info after the first request. */
 656         if (skb_shinfo(skb)->gso_size) {
 657                 struct xen_netif_extra_info *gso;
 658
 659                 gso = (struct xen_netif_extra_info *)
 660                         RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 661
 662                 tx->flags |= XEN_NETTXF_extra_info;
 663
 664                 gso->u.gso.size = skb_shinfo(skb)->gso_size;
 665                 gso->u.gso.type = (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) ?
 666                         XEN_NETIF_GSO_TYPE_TCPV6 :
 667                         XEN_NETIF_GSO_TYPE_TCPV4;
 668                 gso->u.gso.pad = 0;
 669                 gso->u.gso.features = 0;
 670
 671                 gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
 672                 gso->flags = 0;
 673         }
 674
 675         /* Requests for the rest of the linear area. */
 676         tx = xennet_make_txreqs(queue, tx, skb, page, offset, len);
 677
 678         /* Requests for all the frags. */
 679         for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 680                 skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 681                 tx = xennet_make_txreqs(queue, tx, skb,
 682                                         skb_frag_page(frag), frag->page_offset,
 683                                         skb_frag_size(frag));
 684         }
 685
 686         /* First request has the packet length. */
 687         first_tx->size = skb->len;
 688
 689         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->tx, notify);
 690         if (notify)
 691                 notify_remote_via_irq(queue->tx_irq);
 692
 693         u64_stats_update_begin(&tx_stats->syncp);
 694         tx_stats->bytes += skb->len;
 695         tx_stats->packets++;
 696         u64_stats_update_end(&tx_stats->syncp);
 697
 698         /* Note: It is not safe to access skb after xennet_tx_buf_gc()! */
 699         xennet_tx_buf_gc(queue);
 700
 701         if (!netfront_tx_slot_available(queue))
 702                 netif_tx_stop_queue(netdev_get_tx_queue(dev, queue->id));
 703
 704         spin_unlock_irqrestore(&queue->tx_lock, flags);
 705
 706         return NETDEV_TX_OK;
 707
 708  drop:
 709         dev->stats.tx_dropped++;
 710         dev_kfree_skb_any(skb);
 711         return NETDEV_TX_OK;
 712 }
 713
 714 static int xennet_close(struct net_device *dev)
 715 {
 716         struct netfront_info *np = netdev_priv(dev);
 717         unsigned int num_queues = dev->real_num_tx_queues;
 718         unsigned int i;
 719         struct netfront_queue *queue;
 720         netif_tx_stop_all_queues(np->netdev);
 721         for (i = 0; i < num_queues; ++i) {
 722                 queue = &np->queues[i];
 723                 napi_disable(&queue->napi);
 724         }
 725         return 0;
 726 }
 727
 728 static void xennet_move_rx_slot(struct netfront_queue *queue, struct sk_buff *skb,
 729                                 grant_ref_t ref)
 730 {
 731         int new = xennet_rxidx(queue->rx.req_prod_pvt);
 732
 733         BUG_ON(queue->rx_skbs[new]);
 734         queue->rx_skbs[new] = skb;
 735         queue->grant_rx_ref[new] = ref;
 736         RING_GET_REQUEST(&queue->rx, queue->rx.req_prod_pvt)->id = new;
 737         RING_GET_REQUEST(&queue->rx, queue->rx.req_prod_pvt)->gref = ref;
 738         queue->rx.req_prod_pvt++;
 739 }
 740
 741 static int xennet_get_extras(struct netfront_queue *queue,
 742                              struct xen_netif_extra_info *extras,
 743                              RING_IDX rp)
 744
 745 {
 746         struct xen_netif_extra_info *extra;
 747         struct device *dev = &queue->info->netdev->dev;
 748         RING_IDX cons = queue->rx.rsp_cons;
 749         int err = 0;
 750
 751         do {
 752                 struct sk_buff *skb;
 753                 grant_ref_t ref;
 754
 755                 if (unlikely(cons + 1 == rp)) {
 756                         if (net_ratelimit())
 757                                 dev_warn(dev, "Missing extra info\n");
 758                         err = -EBADR;
 759                         break;
 760                 }
 761
 762                 extra = (struct xen_netif_extra_info *)
 763                         RING_GET_RESPONSE(&queue->rx, ++cons);
 764
 765                 if (unlikely(!extra->type ||
 766                              extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
 767                         if (net_ratelimit())
 768                                 dev_warn(dev, "Invalid extra type: %d\n",
 769                                         extra->type);
 770                         err = -EINVAL;
 771                 } else {
 772                         memcpy(&extras[extra->type - 1], extra,
 773                                sizeof(*extra));
 774                 }
 775
 776                 skb = xennet_get_rx_skb(queue, cons);
 777                 ref = xennet_get_rx_ref(queue, cons);
 778                 xennet_move_rx_slot(queue, skb, ref);
 779         } while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);
 780
 781         queue->rx.rsp_cons = cons;
 782         return err;
 783 }
 784
 785 static int xennet_get_responses(struct netfront_queue *queue,
 786                                 struct netfront_rx_info *rinfo, RING_IDX rp,
 787                                 struct sk_buff_head *list)
 788 {
 789         struct xen_netif_rx_response *rx = &rinfo->rx;
 790         struct xen_netif_extra_info *extras = rinfo->extras;
 791         struct device *dev = &queue->info->netdev->dev;
 792         RING_IDX cons = queue->rx.rsp_cons;
 793         struct sk_buff *skb = xennet_get_rx_skb(queue, cons);
 794         grant_ref_t ref = xennet_get_rx_ref(queue, cons);
 795         int max = XEN_NETIF_NR_SLOTS_MIN + (rx->status <= RX_COPY_THRESHOLD);
 796         int slots = 1;
 797         int err = 0;
 798         unsigned long ret;
 799
 800         if (rx->flags & XEN_NETRXF_extra_info) {
 801                 err = xennet_get_extras(queue, extras, rp);
 802                 cons = queue->rx.rsp_cons;
 803         }
 804
 805         for (;;) {
 806                 if (unlikely(rx->status < 0 ||
 807                              rx->offset + rx->status > XEN_PAGE_SIZE)) {
 808                         if (net_ratelimit())
 809                                 dev_warn(dev, "rx->offset: %u, size: %d\n",
 810                                          rx->offset, rx->status);
 811                         xennet_move_rx_slot(queue, skb, ref);
 812                         err = -EINVAL;
 813                         goto next;
 814                 }
 815
 816                 /*
 817                  * This definitely indicates a bug, either in this driver or in
 818                  * the backend driver. In future this should flag the bad
 819                  * situation to the system controller to reboot the backend.
 820                  */
 821                 if (ref == GRANT_INVALID_REF) {
 822                         if (net_ratelimit())
 823                                 dev_warn(dev, "Bad rx response id %d.\n",
 824                                          rx->id);
 825                         err = -EINVAL;
 826                         goto next;
 827                 }
 828
 829                 ret = gnttab_end_foreign_access_ref(ref, 0);
 830                 BUG_ON(!ret);
 831
 832                 gnttab_release_grant_reference(&queue->gref_rx_head, ref);
 833
 834                 __skb_queue_tail(list, skb);
 835
 836 next:
 837                 if (!(rx->flags & XEN_NETRXF_more_data))
 838                         break;
 839
 840                 if (cons + slots == rp) {
 841                         if (net_ratelimit())
 842                                 dev_warn(dev, "Need more slots\n");
 843                         err = -ENOENT;
 844                         break;
 845                 }
 846
 847                 rx = RING_GET_RESPONSE(&queue->rx, cons + slots);
 848                 skb = xennet_get_rx_skb(queue, cons + slots);
 849                 ref = xennet_get_rx_ref(queue, cons + slots);
 850                 slots++;
 851         }
 852
 853         if (unlikely(slots > max)) {
 854                 if (net_ratelimit())
 855                         dev_warn(dev, "Too many slots\n");
 856                 err = -E2BIG;
 857         }
 858
 859         if (unlikely(err))
 860                 queue->rx.rsp_cons = cons + slots;
 861
 862         return err;
 863 }
 864
 865 static int xennet_set_skb_gso(struct sk_buff *skb,
 866                               struct xen_netif_extra_info *gso)
 867 {
 868         if (!gso->u.gso.size) {
 869                 if (net_ratelimit())
 870                         pr_warn("GSO size must not be zero\n");
 871                 return -EINVAL;
 872         }
 873
 874         if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4 &&
 875             gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV6) {
 876                 if (net_ratelimit())
 877                         pr_warn("Bad GSO type %d\n", gso->u.gso.type);
 878                 return -EINVAL;
 879         }
 880
 881         skb_shinfo(skb)->gso_size = gso->u.gso.size;
 882         skb_shinfo(skb)->gso_type =
 883                 (gso->u.gso.type == XEN_NETIF_GSO_TYPE_TCPV4) ?
 884                 SKB_GSO_TCPV4 :
 885                 SKB_GSO_TCPV6;
 886
 887         /* Header must be checked, and gso_segs computed. */
 888         skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
 889         skb_shinfo(skb)->gso_segs = 0;
 890
 891         return 0;
 892 }
 893
 894 static RING_IDX xennet_fill_frags(struct netfront_queue *queue,
 895                                   struct sk_buff *skb,
 896                                   struct sk_buff_head *list)
 897 {
 898         RING_IDX cons = queue->rx.rsp_cons;
 899         struct sk_buff *nskb;
 900
 901         while ((nskb = __skb_dequeue(list))) {
 902                 struct xen_netif_rx_response *rx =
 903                         RING_GET_RESPONSE(&queue->rx, ++cons);
 904                 skb_frag_t *nfrag = &skb_shinfo(nskb)->frags[0];
 905
 906                 if (skb_shinfo(skb)->nr_frags == MAX_SKB_FRAGS) {
 907                         unsigned int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
 908
 909                         BUG_ON(pull_to <= skb_headlen(skb));
 910                         __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
 911                 }
 912                 BUG_ON(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS);
 913
 914                 skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
 915                                 skb_frag_page(nfrag),
 916                                 rx->offset, rx->status, PAGE_SIZE);
 917
 918                 skb_shinfo(nskb)->nr_frags = 0;
 919                 kfree_skb(nskb);
 920         }
 921
 922         return cons;
 923 }
 924
 925 static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
 926 {
 927         bool recalculate_partial_csum = false;
 928
 929         /*
 930          * A GSO SKB must be CHECKSUM_PARTIAL. However some buggy
 931          * peers can fail to set NETRXF_csum_blank when sending a GSO
 932          * frame. In this case force the SKB to CHECKSUM_PARTIAL and
 933          * recalculate the partial checksum.
 934          */
 935         if (skb->ip_summed != CHECKSUM_PARTIAL && skb_is_gso(skb)) {
 936                 struct netfront_info *np = netdev_priv(dev);
 937                 atomic_inc(&np->rx_gso_checksum_fixup);
 938                 skb->ip_summed = CHECKSUM_PARTIAL;
 939                 recalculate_partial_csum = true;
 940         }
 941
 942         /* A non-CHECKSUM_PARTIAL SKB does not require setup. */
 943         if (skb->ip_summed != CHECKSUM_PARTIAL)
 944                 return 0;
 945
 946         return skb_checksum_setup(skb, recalculate_partial_csum);
 947 }
 948
 949 static int handle_incoming_queue(struct netfront_queue *queue,
 950                                  struct sk_buff_head *rxq)
 951 {
 952         struct netfront_stats *rx_stats = this_cpu_ptr(queue->info->rx_stats);
 953         int packets_dropped = 0;
 954         struct sk_buff *skb;
 955
 956         while ((skb = __skb_dequeue(rxq)) != NULL) {
 957                 int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
 958
 959                 if (pull_to > skb_headlen(skb))
 960                         __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
 961
 962                 /* Ethernet work: Delayed to here as it peeks the header. */
 963                 skb->protocol = eth_type_trans(skb, queue->info->netdev);
 964                 skb_reset_network_header(skb);
 965
 966                 if (checksum_setup(queue->info->netdev, skb)) {
 967                         kfree_skb(skb);
 968                         packets_dropped++;
 969                         queue->info->netdev->stats.rx_errors++;
 970                         continue;
 971                 }
 972
 973                 u64_stats_update_begin(&rx_stats->syncp);
 974                 rx_stats->packets++;
 975                 rx_stats->bytes += skb->len;
 976                 u64_stats_update_end(&rx_stats->syncp);
 977
 978                 /* Pass it up. */
 979                 napi_gro_receive(&queue->napi, skb);
 980         }
 981
 982         return packets_dropped;
 983 }
 984
 985 static int xennet_poll(struct napi_struct *napi, int budget)
 986 {
 987         struct netfront_queue *queue = container_of(napi, struct netfront_queue, napi);
 988         struct net_device *dev = queue->info->netdev;
 989         struct sk_buff *skb;
 990         struct netfront_rx_info rinfo;
 991         struct xen_netif_rx_response *rx = &rinfo.rx;
 992         struct xen_netif_extra_info *extras = rinfo.extras;
 993         RING_IDX i, rp;
 994         int work_done;
 995         struct sk_buff_head rxq;
 996         struct sk_buff_head errq;
 997         struct sk_buff_head tmpq;
 998         int err;
 999
1000         spin_lock(&queue->rx_lock);
1001
1002         skb_queue_head_init(&rxq);
1003         skb_queue_head_init(&errq);
1004         skb_queue_head_init(&tmpq);
1005
1006         rp = queue->rx.sring->rsp_prod;
1007         rmb(); /* Ensure we see queued responses up to 'rp'. */
1008
1009         i = queue->rx.rsp_cons;
1010         work_done = 0;
1011         while ((i != rp) && (work_done < budget)) {
1012                 memcpy(rx, RING_GET_RESPONSE(&queue->rx, i), sizeof(*rx));
1013                 memset(extras, 0, sizeof(rinfo.extras));
1014
1015                 err = xennet_get_responses(queue, &rinfo, rp, &tmpq);
1016
1017                 if (unlikely(err)) {
1018 err:
1019                         while ((skb = __skb_dequeue(&tmpq)))
1020                                 __skb_queue_tail(&errq, skb);
1021                         dev->stats.rx_errors++;
1022                         i = queue->rx.rsp_cons;
1023                         continue;
1024                 }
1025
1026                 skb = __skb_dequeue(&tmpq);
1027
1028                 if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
1029                         struct xen_netif_extra_info *gso;
1030                         gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
1031
1032                         if (unlikely(xennet_set_skb_gso(skb, gso))) {
1033                                 __skb_queue_head(&tmpq, skb);
1034                                 queue->rx.rsp_cons += skb_queue_len(&tmpq);
1035                                 goto err;
1036                         }
1037                 }
1038
1039                 NETFRONT_SKB_CB(skb)->pull_to = rx->status;
1040                 if (NETFRONT_SKB_CB(skb)->pull_to > RX_COPY_THRESHOLD)
1041                         NETFRONT_SKB_CB(skb)->pull_to = RX_COPY_THRESHOLD;
1042
1043                 skb_shinfo(skb)->frags[0].page_offset = rx->offset;
1044                 skb_frag_size_set(&skb_shinfo(skb)->frags[0], rx->status);
1045                 skb->data_len = rx->status;
1046                 skb->len += rx->status;
1047
1048                 i = xennet_fill_frags(queue, skb, &tmpq);
1049
1050                 if (rx->flags & XEN_NETRXF_csum_blank)
1051                         skb->ip_summed = CHECKSUM_PARTIAL;
1052                 else if (rx->flags & XEN_NETRXF_data_validated)
1053                         skb->ip_summed = CHECKSUM_UNNECESSARY;
1054
1055                 __skb_queue_tail(&rxq, skb);
1056
1057                 queue->rx.rsp_cons = ++i;
1058                 work_done++;
1059         }
1060
1061         __skb_queue_purge(&errq);
1062
1063         work_done -= handle_incoming_queue(queue, &rxq);
1064
1065         xennet_alloc_rx_buffers(queue);
1066
1067         if (work_done < budget) {
1068                 int more_to_do = 0;
1069
1070                 napi_complete_done(napi, work_done);
1071
1072                 RING_FINAL_CHECK_FOR_RESPONSES(&queue->rx, more_to_do);
1073                 if (more_to_do)
1074                         napi_schedule(napi);
1075         }
1076
1077         spin_unlock(&queue->rx_lock);
1078
1079         return work_done;
1080 }
1081
1082 static int xennet_change_mtu(struct net_device *dev, int mtu)
1083 {
1084         int max = xennet_can_sg(dev) ? XEN_NETIF_MAX_TX_SIZE : ETH_DATA_LEN;
1085
1086         if (mtu > max)
1087                 return -EINVAL;
1088         dev->mtu = mtu;
1089         return 0;
1090 }
1091
1092 static void xennet_get_stats64(struct net_device *dev,
1093                                struct rtnl_link_stats64 *tot)
1094 {
1095         struct netfront_info *np = netdev_priv(dev);
1096         int cpu;
1097
1098         for_each_possible_cpu(cpu) {
1099                 struct netfront_stats *rx_stats = per_cpu_ptr(np->rx_stats, cpu);
1100                 struct netfront_stats *tx_stats = per_cpu_ptr(np->tx_stats, cpu);
1101                 u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
1102                 unsigned int start;
1103
1104                 do {
1105                         start = u64_stats_fetch_begin_irq(&tx_stats->syncp);
1106                         tx_packets = tx_stats->packets;
1107                         tx_bytes = tx_stats->bytes;
1108                 } while (u64_stats_fetch_retry_irq(&tx_stats->syncp, start));
1109
1110                 do {
1111                         start = u64_stats_fetch_begin_irq(&rx_stats->syncp);
1112                         rx_packets = rx_stats->packets;
1113                         rx_bytes = rx_stats->bytes;
1114                 } while (u64_stats_fetch_retry_irq(&rx_stats->syncp, start));
1115
1116                 tot->rx_packets += rx_packets;
1117                 tot->tx_packets += tx_packets;
1118                 tot->rx_bytes   += rx_bytes;
1119                 tot->tx_bytes   += tx_bytes;
1120         }
1121
1122         tot->rx_errors  = dev->stats.rx_errors;
1123         tot->tx_dropped = dev->stats.tx_dropped;
1124 }
1125
1126 static void xennet_release_tx_bufs(struct netfront_queue *queue)
1127 {
1128         struct sk_buff *skb;
1129         int i;
1130
1131         for (i = 0; i < NET_TX_RING_SIZE; i++) {
1132                 /* Skip over entries which are actually freelist references */
1133                 if (skb_entry_is_link(&queue->tx_skbs[i]))
1134                         continue;
1135
1136                 skb = queue->tx_skbs[i].skb;
1137                 get_page(queue->grant_tx_page[i]);
1138                 gnttab_end_foreign_access(queue->grant_tx_ref[i],
1139                                           GNTMAP_readonly,
1140                                           (unsigned long)page_address(queue->grant_tx_page[i]));
1141                 queue->grant_tx_page[i] = NULL;
1142                 queue->grant_tx_ref[i] = GRANT_INVALID_REF;
1143                 add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, i);
1144                 dev_kfree_skb_irq(skb);
1145         }
1146 }
1147
1148 static void xennet_release_rx_bufs(struct netfront_queue *queue)
1149 {
1150         int id, ref;
1151
1152         spin_lock_bh(&queue->rx_lock);
1153
1154         for (id = 0; id < NET_RX_RING_SIZE; id++) {
1155                 struct sk_buff *skb;
1156                 struct page *page;
1157
1158                 skb = queue->rx_skbs[id];
1159                 if (!skb)
1160                         continue;
1161
1162                 ref = queue->grant_rx_ref[id];
1163                 if (ref == GRANT_INVALID_REF)
1164                         continue;
1165
1166                 page = skb_frag_page(&skb_shinfo(skb)->frags[0]);
1167
1168                 /* gnttab_end_foreign_access() needs a page ref until
1169                  * foreign access is ended (which may be deferred).
1170                  */
1171                 get_page(page);
1172                 gnttab_end_foreign_access(ref, 0,
1173                                           (unsigned long)page_address(page));
1174                 queue->grant_rx_ref[id] = GRANT_INVALID_REF;
1175
1176                 kfree_skb(skb);
1177         }
1178
1179         spin_unlock_bh(&queue->rx_lock);
1180 }
1181
1182 static netdev_features_t xennet_fix_features(struct net_device *dev,
1183         netdev_features_t features)
1184 {
1185         struct netfront_info *np = netdev_priv(dev);
1186
1187         if (features & NETIF_F_SG &&
1188             !xenbus_read_unsigned(np->xbdev->otherend, "feature-sg", 0))
1189                 features &= ~NETIF_F_SG;
1190
1191         if (features & NETIF_F_IPV6_CSUM &&
1192             !xenbus_read_unsigned(np->xbdev->otherend,
1193                                   "feature-ipv6-csum-offload", 0))
1194                 features &= ~NETIF_F_IPV6_CSUM;
1195
1196         if (features & NETIF_F_TSO &&
1197             !xenbus_read_unsigned(np->xbdev->otherend, "feature-gso-tcpv4", 0))
1198                 features &= ~NETIF_F_TSO;
1199
1200         if (features & NETIF_F_TSO6 &&
1201             !xenbus_read_unsigned(np->xbdev->otherend, "feature-gso-tcpv6", 0))
1202                 features &= ~NETIF_F_TSO6;
1203
1204         return features;
1205 }
1206
1207 static int xennet_set_features(struct net_device *dev,
1208         netdev_features_t features)
1209 {
1210         if (!(features & NETIF_F_SG) && dev->mtu > ETH_DATA_LEN) {
1211                 netdev_info(dev, "Reducing MTU because no SG offload");
1212                 dev->mtu = ETH_DATA_LEN;
1213         }
1214
1215         return 0;
1216 }
1217
1218 static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
1219 {
1220         struct netfront_queue *queue = dev_id;
1221         unsigned long flags;
1222
1223         spin_lock_irqsave(&queue->tx_lock, flags);
1224         xennet_tx_buf_gc(queue);
1225         spin_unlock_irqrestore(&queue->tx_lock, flags);
1226
1227         return IRQ_HANDLED;
1228 }
1229
1230 static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
1231 {
1232         struct netfront_queue *queue = dev_id;
1233         struct net_device *dev = queue->info->netdev;
1234
1235         if (likely(netif_carrier_ok(dev) &&
1236                    RING_HAS_UNCONSUMED_RESPONSES(&queue->rx)))
1237                 napi_schedule(&queue->napi);
1238
1239         return IRQ_HANDLED;
1240 }
1241
1242 static irqreturn_t xennet_interrupt(int irq, void *dev_id)
1243 {
1244         xennet_tx_interrupt(irq, dev_id);
1245         xennet_rx_interrupt(irq, dev_id);
1246         return IRQ_HANDLED;
1247 }
1248
1249 #ifdef CONFIG_NET_POLL_CONTROLLER
1250 static void xennet_poll_controller(struct net_device *dev)
1251 {
1252         /* Poll each queue */
1253         struct netfront_info *info = netdev_priv(dev);
1254         unsigned int num_queues = dev->real_num_tx_queues;
1255         unsigned int i;
1256         for (i = 0; i < num_queues; ++i)
1257                 xennet_interrupt(0, &info->queues[i]);
1258 }
1259 #endif
1260
1261 static const struct net_device_ops xennet_netdev_ops = {
1262         .ndo_open            = xennet_open,
1263         .ndo_stop            = xennet_close,
1264         .ndo_start_xmit      = xennet_start_xmit,
1265         .ndo_change_mtu      = xennet_change_mtu,
1266         .ndo_get_stats64     = xennet_get_stats64,
1267         .ndo_set_mac_address = eth_mac_addr,
1268         .ndo_validate_addr   = eth_validate_addr,
1269         .ndo_fix_features    = xennet_fix_features,
1270         .ndo_set_features    = xennet_set_features,
1271         .ndo_select_queue    = xennet_select_queue,
1272 #ifdef CONFIG_NET_POLL_CONTROLLER
1273         .ndo_poll_controller = xennet_poll_controller,
1274 #endif
1275 };
1276
1277 static void xennet_free_netdev(struct net_device *netdev)
1278 {
1279         struct netfront_info *np = netdev_priv(netdev);
1280
1281         free_percpu(np->rx_stats);
1282         free_percpu(np->tx_stats);
1283         free_netdev(netdev);
1284 }
1285
1286 static struct net_device *xennet_create_dev(struct xenbus_device *dev)
1287 {
1288         int err;
1289         struct net_device *netdev;
1290         struct netfront_info *np;
1291
1292         netdev = alloc_etherdev_mq(sizeof(struct netfront_info), xennet_max_queues);
1293         if (!netdev)
1294                 return ERR_PTR(-ENOMEM);
1295
1296         np                   = netdev_priv(netdev);
1297         np->xbdev            = dev;
1298
1299         np->queues = NULL;
1300
1301         err = -ENOMEM;
1302         np->rx_stats = netdev_alloc_pcpu_stats(struct netfront_stats);
1303         if (np->rx_stats == NULL)
1304                 goto exit;
1305         np->tx_stats = netdev_alloc_pcpu_stats(struct netfront_stats);
1306         if (np->tx_stats == NULL)
1307                 goto exit;
1308
1309         netdev->netdev_ops      = &xennet_netdev_ops;
1310
1311         netdev->features        = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
1312                                   NETIF_F_GSO_ROBUST;
1313         netdev->hw_features     = NETIF_F_SG |
1314                                   NETIF_F_IPV6_CSUM |
1315                                   NETIF_F_TSO | NETIF_F_TSO6;
1316
1317         /*
1318          * Assume that all hw features are available for now. This set
1319          * will be adjusted by the call to netdev_update_features() in
1320          * xennet_connect() which is the earliest point where we can
1321          * negotiate with the backend regarding supported features.
1322          */
1323         netdev->features |= netdev->hw_features;
1324
1325         netdev->ethtool_ops = &xennet_ethtool_ops;
1326         netdev->min_mtu = ETH_MIN_MTU;
1327         netdev->max_mtu = XEN_NETIF_MAX_TX_SIZE;
1328         SET_NETDEV_DEV(netdev, &dev->dev);
1329
1330         np->netdev = netdev;
1331
1332         netif_carrier_off(netdev);
1333
1334         xenbus_switch_state(dev, XenbusStateInitialising);
1335         wait_event(module_load_q,
1336                            xenbus_read_driver_state(dev->otherend) !=
1337                            XenbusStateClosed &&
1338                            xenbus_read_driver_state(dev->otherend) !=
1339                            XenbusStateUnknown);
1340         return netdev;
1341
1342  exit:
1343         xennet_free_netdev(netdev);
1344         return ERR_PTR(err);
1345 }
1346
1347 /**
1348  * Entry point to this code when a new device is created.  Allocate the basic
1349  * structures and the ring buffers for communication with the backend, and
1350  * inform the backend of the appropriate details for those.
1351  */
1352 static int netfront_probe(struct xenbus_device *dev,
1353                           const struct xenbus_device_id *id)
1354 {
1355         int err;
1356         struct net_device *netdev;
1357         struct netfront_info *info;
1358
1359         netdev = xennet_create_dev(dev);
1360         if (IS_ERR(netdev)) {
1361                 err = PTR_ERR(netdev);
1362                 xenbus_dev_fatal(dev, err, "creating netdev");
1363                 return err;
1364         }
1365
1366         info = netdev_priv(netdev);
1367         dev_set_drvdata(&dev->dev, info);
1368 #ifdef CONFIG_SYSFS
1369         info->netdev->sysfs_groups[0] = &xennet_dev_group;
1370 #endif
1371
1372         return 0;
1373 }
1374
1375 static void xennet_end_access(int ref, void *page)
1376 {
1377         /* This frees the page as a side-effect */
1378         if (ref != GRANT_INVALID_REF)
1379                 gnttab_end_foreign_access(ref, 0, (unsigned long)page);
1380 }
1381
1382 static void xennet_disconnect_backend(struct netfront_info *info)
1383 {
1384         unsigned int i = 0;
1385         unsigned int num_queues = info->netdev->real_num_tx_queues;
1386
1387         netif_carrier_off(info->netdev);
1388
1389         for (i = 0; i < num_queues && info->queues; ++i) {
1390                 struct netfront_queue *queue = &info->queues[i];
1391
1392                 del_timer_sync(&queue->rx_refill_timer);
1393
1394                 if (queue->tx_irq && (queue->tx_irq == queue->rx_irq))
1395                         unbind_from_irqhandler(queue->tx_irq, queue);
1396                 if (queue->tx_irq && (queue->tx_irq != queue->rx_irq)) {
1397                         unbind_from_irqhandler(queue->tx_irq, queue);
1398                         unbind_from_irqhandler(queue->rx_irq, queue);
1399                 }
1400                 queue->tx_evtchn = queue->rx_evtchn = 0;
1401                 queue->tx_irq = queue->rx_irq = 0;
1402
1403                 if (netif_running(info->netdev))
1404                         napi_synchronize(&queue->napi);
1405
1406                 xennet_release_tx_bufs(queue);
1407                 xennet_release_rx_bufs(queue);
1408                 gnttab_free_grant_references(queue->gref_tx_head);
1409                 gnttab_free_grant_references(queue->gref_rx_head);
1410
1411                 /* End access and free the pages */
1412                 xennet_end_access(queue->tx_ring_ref, queue->tx.sring);
1413                 xennet_end_access(queue->rx_ring_ref, queue->rx.sring);
1414
1415                 queue->tx_ring_ref = GRANT_INVALID_REF;
1416                 queue->rx_ring_ref = GRANT_INVALID_REF;
1417                 queue->tx.sring = NULL;
1418                 queue->rx.sring = NULL;
1419         }
1420 }
1421
1422 /**
1423  * We are reconnecting to the backend, due to a suspend/resume, or a backend
1424  * driver restart.  We tear down our netif structure and recreate it, but
1425  * leave the device-layer structures intact so that this is transparent to the
1426  * rest of the kernel.
1427  */
1428 static int netfront_resume(struct xenbus_device *dev)
1429 {
1430         struct netfront_info *info = dev_get_drvdata(&dev->dev);
1431
1432         dev_dbg(&dev->dev, "%s\n", dev->nodename);
1433
1434         xennet_disconnect_backend(info);
1435         return 0;
1436 }
1437
1438 static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
1439 {
1440         char *s, *e, *macstr;
1441         int i;
1442
1443         macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
1444         if (IS_ERR(macstr))
1445                 return PTR_ERR(macstr);
1446
1447         for (i = 0; i < ETH_ALEN; i++) {
1448                 mac[i] = simple_strtoul(s, &e, 16);
1449                 if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
1450                         kfree(macstr);
1451                         return -ENOENT;
1452                 }
1453                 s = e+1;
1454         }
1455
1456         kfree(macstr);
1457         return 0;
1458 }
1459
1460 static int setup_netfront_single(struct netfront_queue *queue)
1461 {
1462         int err;
1463
1464         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->tx_evtchn);
1465         if (err < 0)
1466                 goto fail;
1467
1468         err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
1469                                         xennet_interrupt,
1470                                         0, queue->info->netdev->name, queue);
1471         if (err < 0)
1472                 goto bind_fail;
1473         queue->rx_evtchn = queue->tx_evtchn;
1474         queue->rx_irq = queue->tx_irq = err;
1475
1476         return 0;
1477
1478 bind_fail:
1479         xenbus_free_evtchn(queue->info->xbdev, queue->tx_evtchn);
1480         queue->tx_evtchn = 0;
1481 fail:
1482         return err;
1483 }
1484
1485 static int setup_netfront_split(struct netfront_queue *queue)
1486 {
1487         int err;
1488
1489         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->tx_evtchn);
1490         if (err < 0)
1491                 goto fail;
1492         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->rx_evtchn);
1493         if (err < 0)
1494                 goto alloc_rx_evtchn_fail;
1495
1496         snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name),
1497                  "%s-tx", queue->name);
1498         err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
1499                                         xennet_tx_interrupt,
1500                                         0, queue->tx_irq_name, queue);
1501         if (err < 0)
1502                 goto bind_tx_fail;
1503         queue->tx_irq = err;
1504
1505         snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name),
1506                  "%s-rx", queue->name);
1507         err = bind_evtchn_to_irqhandler(queue->rx_evtchn,
1508                                         xennet_rx_interrupt,
1509                                         0, queue->rx_irq_name, queue);
1510         if (err < 0)
1511                 goto bind_rx_fail;
1512         queue->rx_irq = err;
1513
1514         return 0;
1515
1516 bind_rx_fail:
1517         unbind_from_irqhandler(queue->tx_irq, queue);
1518         queue->tx_irq = 0;
1519 bind_tx_fail:
1520         xenbus_free_evtchn(queue->info->xbdev, queue->rx_evtchn);
1521         queue->rx_evtchn = 0;
1522 alloc_rx_evtchn_fail:
1523         xenbus_free_evtchn(queue->info->xbdev, queue->tx_evtchn);
1524         queue->tx_evtchn = 0;
1525 fail:
1526         return err;
1527 }
1528
1529 static int setup_netfront(struct xenbus_device *dev,
1530                         struct netfront_queue *queue, unsigned int feature_split_evtchn)
1531 {
1532         struct xen_netif_tx_sring *txs;
1533         struct xen_netif_rx_sring *rxs;
1534         grant_ref_t gref;
1535         int err;
1536
1537         queue->tx_ring_ref = GRANT_INVALID_REF;
1538         queue->rx_ring_ref = GRANT_INVALID_REF;
1539         queue->rx.sring = NULL;
1540         queue->tx.sring = NULL;
1541
1542         txs = (struct xen_netif_tx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1543         if (!txs) {
1544                 err = -ENOMEM;
1545                 xenbus_dev_fatal(dev, err, "allocating tx ring page");
1546                 goto fail;
1547         }
1548         SHARED_RING_INIT(txs);
1549         FRONT_RING_INIT(&queue->tx, txs, XEN_PAGE_SIZE);
1550
1551         err = xenbus_grant_ring(dev, txs, 1, &gref);
1552         if (err < 0)
1553                 goto grant_tx_ring_fail;
1554         queue->tx_ring_ref = gref;
1555
1556         rxs = (struct xen_netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1557         if (!rxs) {
1558                 err = -ENOMEM;
1559                 xenbus_dev_fatal(dev, err, "allocating rx ring page");
1560                 goto alloc_rx_ring_fail;
1561         }
1562         SHARED_RING_INIT(rxs);
1563         FRONT_RING_INIT(&queue->rx, rxs, XEN_PAGE_SIZE);
1564
1565         err = xenbus_grant_ring(dev, rxs, 1, &gref);
1566         if (err < 0)
1567                 goto grant_rx_ring_fail;
1568         queue->rx_ring_ref = gref;
1569
1570         if (feature_split_evtchn)
1571                 err = setup_netfront_split(queue);
1572         /* setup single event channel if
1573          *  a) feature-split-event-channels == 0
1574          *  b) feature-split-event-channels == 1 but failed to setup
1575          */
1576         if (!feature_split_evtchn || (feature_split_evtchn && err))
1577                 err = setup_netfront_single(queue);
1578
1579         if (err)
1580                 goto alloc_evtchn_fail;
1581
1582         return 0;
1583
1584         /* If we fail to setup netfront, it is safe to just revoke access to
1585          * granted pages because backend is not accessing it at this point.
1586          */
1587 alloc_evtchn_fail:
1588         gnttab_end_foreign_access_ref(queue->rx_ring_ref, 0);
1589 grant_rx_ring_fail:
1590         free_page((unsigned long)rxs);
1591 alloc_rx_ring_fail:
1592         gnttab_end_foreign_access_ref(queue->tx_ring_ref, 0);
1593 grant_tx_ring_fail:
1594         free_page((unsigned long)txs);
1595 fail:
1596         return err;
1597 }
1598
1599 /* Queue-specific initialisation
1600  * This used to be done in xennet_create_dev() but must now
1601  * be run per-queue.
1602  */
1603 static int xennet_init_queue(struct netfront_queue *queue)
1604 {
1605         unsigned short i;
1606         int err = 0;
1607         char *devid;
1608
1609         spin_lock_init(&queue->tx_lock);
1610         spin_lock_init(&queue->rx_lock);
1611
1612         timer_setup(&queue->rx_refill_timer, rx_refill_timeout, 0);
1613
1614         devid = strrchr(queue->info->xbdev->nodename, '/') + 1;
1615         snprintf(queue->name, sizeof(queue->name), "vif%s-q%u",
1616                  devid, queue->id);
1617
1618         /* Initialise tx_skbs as a free chain containing every entry. */
1619         queue->tx_skb_freelist = 0;
1620         for (i = 0; i < NET_TX_RING_SIZE; i++) {
1621                 skb_entry_set_link(&queue->tx_skbs[i], i+1);
1622                 queue->grant_tx_ref[i] = GRANT_INVALID_REF;
1623                 queue->grant_tx_page[i] = NULL;
1624         }
1625
1626         /* Clear out rx_skbs */
1627         for (i = 0; i < NET_RX_RING_SIZE; i++) {
1628                 queue->rx_skbs[i] = NULL;
1629                 queue->grant_rx_ref[i] = GRANT_INVALID_REF;
1630         }
1631
1632         /* A grant for every tx ring slot */
1633         if (gnttab_alloc_grant_references(NET_TX_RING_SIZE,
1634                                           &queue->gref_tx_head) < 0) {
1635                 pr_alert("can't alloc tx grant refs\n");
1636                 err = -ENOMEM;
1637                 goto exit;
1638         }
1639
1640         /* A grant for every rx ring slot */
1641         if (gnttab_alloc_grant_references(NET_RX_RING_SIZE,
1642                                           &queue->gref_rx_head) < 0) {
1643                 pr_alert("can't alloc rx grant refs\n");
1644                 err = -ENOMEM;
1645                 goto exit_free_tx;
1646         }
1647
1648         return 0;
1649
1650  exit_free_tx:
1651         gnttab_free_grant_references(queue->gref_tx_head);
1652  exit:
1653         return err;
1654 }
1655
1656 static int write_queue_xenstore_keys(struct netfront_queue *queue,
1657                            struct xenbus_transaction *xbt, int write_hierarchical)
1658 {
1659         /* Write the queue-specific keys into XenStore in the traditional
1660          * way for a single queue, or in a queue subkeys for multiple
1661          * queues.
1662          */
1663         struct xenbus_device *dev = queue->info->xbdev;
1664         int err;
1665         const char *message;
1666         char *path;
1667         size_t pathsize;
1668
1669         /* Choose the correct place to write the keys */
1670         if (write_hierarchical) {
1671                 pathsize = strlen(dev->nodename) + 10;
1672                 path = kzalloc(pathsize, GFP_KERNEL);
1673                 if (!path) {
1674                         err = -ENOMEM;
1675                         message = "out of memory while writing ring references";
1676                         goto error;
1677                 }
1678                 snprintf(path, pathsize, "%s/queue-%u",
1679                                 dev->nodename, queue->id);
1680         } else {
1681                 path = (char *)dev->nodename;
1682         }
1683
1684         /* Write ring references */
1685         err = xenbus_printf(*xbt, path, "tx-ring-ref", "%u",
1686                         queue->tx_ring_ref);
1687         if (err) {
1688                 message = "writing tx-ring-ref";
1689                 goto error;
1690         }
1691
1692         err = xenbus_printf(*xbt, path, "rx-ring-ref", "%u",
1693                         queue->rx_ring_ref);
1694         if (err) {
1695                 message = "writing rx-ring-ref";
1696                 goto error;
1697         }
1698
1699         /* Write event channels; taking into account both shared
1700          * and split event channel scenarios.
1701          */
1702         if (queue->tx_evtchn == queue->rx_evtchn) {
1703                 /* Shared event channel */
1704                 err = xenbus_printf(*xbt, path,
1705                                 "event-channel", "%u", queue->tx_evtchn);
1706                 if (err) {
1707                         message = "writing event-channel";
1708                         goto error;
1709                 }
1710         } else {
1711                 /* Split event channels */
1712                 err = xenbus_printf(*xbt, path,
1713                                 "event-channel-tx", "%u", queue->tx_evtchn);
1714                 if (err) {
1715                         message = "writing event-channel-tx";
1716                         goto error;
1717                 }
1718
1719                 err = xenbus_printf(*xbt, path,
1720                                 "event-channel-rx", "%u", queue->rx_evtchn);
1721                 if (err) {
1722                         message = "writing event-channel-rx";
1723                         goto error;
1724                 }
1725         }
1726
1727         if (write_hierarchical)
1728                 kfree(path);
1729         return 0;
1730
1731 error:
1732         if (write_hierarchical)
1733                 kfree(path);
1734         xenbus_dev_fatal(dev, err, "%s", message);
1735         return err;
1736 }
1737
1738 static void xennet_destroy_queues(struct netfront_info *info)
1739 {
1740         unsigned int i;
1741
1742         for (i = 0; i < info->netdev->real_num_tx_queues; i++) {
1743                 struct netfront_queue *queue = &info->queues[i];
1744
1745                 if (netif_running(info->netdev))
1746                         napi_disable(&queue->napi);
1747                 netif_napi_del(&queue->napi);
1748         }
1749
1750         kfree(info->queues);
1751         info->queues = NULL;
1752 }
1753
1754 static int xennet_create_queues(struct netfront_info *info,
1755                                 unsigned int *num_queues)
1756 {
1757         unsigned int i;
1758         int ret;
1759
1760         info->queues = kcalloc(*num_queues, sizeof(struct netfront_queue),
1761                                GFP_KERNEL);
1762         if (!info->queues)
1763                 return -ENOMEM;
1764
1765         for (i = 0; i < *num_queues; i++) {
1766                 struct netfront_queue *queue = &info->queues[i];
1767
1768                 queue->id = i;
1769                 queue->info = info;
1770
1771                 ret = xennet_init_queue(queue);
1772                 if (ret < 0) {
1773                         dev_warn(&info->xbdev->dev,
1774                                  "only created %d queues\n", i);
1775                         *num_queues = i;
1776                         break;
1777                 }
1778
1779                 netif_napi_add(queue->info->netdev, &queue->napi,
1780                                xennet_poll, 64);
1781                 if (netif_running(info->netdev))
1782                         napi_enable(&queue->napi);
1783         }
1784
1785         netif_set_real_num_tx_queues(info->netdev, *num_queues);
1786
1787         if (*num_queues == 0) {
1788                 dev_err(&info->xbdev->dev, "no queues\n");
1789                 return -EINVAL;
1790         }
1791         return 0;
1792 }
1793
1794 /* Common code used when first setting up, and when resuming. */
1795 static int talk_to_netback(struct xenbus_device *dev,
1796                            struct netfront_info *info)
1797 {
1798         const char *message;
1799         struct xenbus_transaction xbt;
1800         int err;
1801         unsigned int feature_split_evtchn;
1802         unsigned int i = 0;
1803         unsigned int max_queues = 0;
1804         struct netfront_queue *queue = NULL;
1805         unsigned int num_queues = 1;
1806
1807         info->netdev->irq = 0;
1808
1809         /* Check if backend supports multiple queues */
1810         max_queues = xenbus_read_unsigned(info->xbdev->otherend,
1811                                           "multi-queue-max-queues", 1);
1812         num_queues = min(max_queues, xennet_max_queues);
1813
1814         /* Check feature-split-event-channels */
1815         feature_split_evtchn = xenbus_read_unsigned(info->xbdev->otherend,
1816                                         "feature-split-event-channels", 0);
1817
1818         /* Read mac addr. */
1819         err = xen_net_read_mac(dev, info->netdev->dev_addr);
1820         if (err) {
1821                 xenbus_dev_fatal(dev, err, "parsing %s/mac", dev->nodename);
1822                 goto out_unlocked;
1823         }
1824
1825         rtnl_lock();
1826         if (info->queues)
1827                 xennet_destroy_queues(info);
1828
1829         err = xennet_create_queues(info, &num_queues);
1830         if (err < 0) {
1831                 xenbus_dev_fatal(dev, err, "creating queues");
1832                 kfree(info->queues);
1833                 info->queues = NULL;
1834                 goto out;
1835         }
1836         rtnl_unlock();
1837
1838         /* Create shared ring, alloc event channel -- for each queue */
1839         for (i = 0; i < num_queues; ++i) {
1840                 queue = &info->queues[i];
1841                 err = setup_netfront(dev, queue, feature_split_evtchn);
1842                 if (err)
1843                         goto destroy_ring;
1844         }
1845
1846 again:
1847         err = xenbus_transaction_start(&xbt);
1848         if (err) {
1849                 xenbus_dev_fatal(dev, err, "starting transaction");
1850                 goto destroy_ring;
1851         }
1852
1853         if (xenbus_exists(XBT_NIL,
1854                           info->xbdev->otherend, "multi-queue-max-queues")) {
1855                 /* Write the number of queues */
1856                 err = xenbus_printf(xbt, dev->nodename,
1857                                     "multi-queue-num-queues", "%u", num_queues);
1858                 if (err) {
1859                         message = "writing multi-queue-num-queues";
1860                         goto abort_transaction_no_dev_fatal;
1861                 }
1862         }
1863
1864         if (num_queues == 1) {
1865                 err = write_queue_xenstore_keys(&info->queues[0], &xbt, 0); /* flat */
1866                 if (err)
1867                         goto abort_transaction_no_dev_fatal;
1868         } else {
1869                 /* Write the keys for each queue */
1870                 for (i = 0; i < num_queues; ++i) {
1871                         queue = &info->queues[i];
1872                         err = write_queue_xenstore_keys(queue, &xbt, 1); /* hierarchical */
1873                         if (err)
1874                                 goto abort_transaction_no_dev_fatal;
1875                 }
1876         }
1877
1878         /* The remaining keys are not queue-specific */
1879         err = xenbus_printf(xbt, dev->nodename, "request-rx-copy", "%u",
1880                             1);
1881         if (err) {
1882                 message = "writing request-rx-copy";
1883                 goto abort_transaction;
1884         }
1885
1886         err = xenbus_printf(xbt, dev->nodename, "feature-rx-notify", "%d", 1);
1887         if (err) {
1888                 message = "writing feature-rx-notify";
1889                 goto abort_transaction;
1890         }
1891
1892         err = xenbus_printf(xbt, dev->nodename, "feature-sg", "%d", 1);
1893         if (err) {
1894                 message = "writing feature-sg";
1895                 goto abort_transaction;
1896         }
1897
1898         err = xenbus_printf(xbt, dev->nodename, "feature-gso-tcpv4", "%d", 1);
1899         if (err) {
1900                 message = "writing feature-gso-tcpv4";
1901                 goto abort_transaction;
1902         }
1903
1904         err = xenbus_write(xbt, dev->nodename, "feature-gso-tcpv6", "1");
1905         if (err) {
1906                 message = "writing feature-gso-tcpv6";
1907                 goto abort_transaction;
1908         }
1909
1910         err = xenbus_write(xbt, dev->nodename, "feature-ipv6-csum-offload",
1911                            "1");
1912         if (err) {
1913                 message = "writing feature-ipv6-csum-offload";
1914                 goto abort_transaction;
1915         }
1916
1917         err = xenbus_transaction_end(xbt, 0);
1918         if (err) {
1919                 if (err == -EAGAIN)
1920                         goto again;
1921                 xenbus_dev_fatal(dev, err, "completing transaction");
1922                 goto destroy_ring;
1923         }
1924
1925         return 0;
1926
1927  abort_transaction:
1928         xenbus_dev_fatal(dev, err, "%s", message);
1929 abort_transaction_no_dev_fatal:
1930         xenbus_transaction_end(xbt, 1);
1931  destroy_ring:
1932         xennet_disconnect_backend(info);
1933         rtnl_lock();
1934         xennet_destroy_queues(info);
1935  out:
1936         rtnl_unlock();
1937 out_unlocked:
1938         device_unregister(&dev->dev);
1939         return err;
1940 }
1941
1942 static int xennet_connect(struct net_device *dev)
1943 {
1944         struct netfront_info *np = netdev_priv(dev);
1945         unsigned int num_queues = 0;
1946         int err;
1947         unsigned int j = 0;
1948         struct netfront_queue *queue = NULL;
1949
1950         if (!xenbus_read_unsigned(np->xbdev->otherend, "feature-rx-copy", 0)) {
1951                 dev_info(&dev->dev,
1952                          "backend does not support copying receive path\n");
1953                 return -ENODEV;
1954         }
1955
1956         err = talk_to_netback(np->xbdev, np);
1957         if (err)
1958                 return err;
1959
1960         /* talk_to_netback() sets the correct number of queues */
1961         num_queues = dev->real_num_tx_queues;
1962
1963         if (dev->reg_state == NETREG_UNINITIALIZED) {
1964                 err = register_netdev(dev);
1965                 if (err) {
1966                         pr_warn("%s: register_netdev err=%d\n", __func__, err);
1967                         device_unregister(&np->xbdev->dev);
1968                         return err;
1969                 }
1970         }
1971
1972         rtnl_lock();
1973         netdev_update_features(dev);
1974         rtnl_unlock();
1975
1976         /*
1977          * All public and private state should now be sane.  Get
1978          * ready to start sending and receiving packets and give the driver
1979          * domain a kick because we've probably just requeued some
1980          * packets.
1981          */
1982         netif_carrier_on(np->netdev);
1983         for (j = 0; j < num_queues; ++j) {
1984                 queue = &np->queues[j];
1985
1986                 notify_remote_via_irq(queue->tx_irq);
1987                 if (queue->tx_irq != queue->rx_irq)
1988                         notify_remote_via_irq(queue->rx_irq);
1989
1990                 spin_lock_irq(&queue->tx_lock);
1991                 xennet_tx_buf_gc(queue);
1992                 spin_unlock_irq(&queue->tx_lock);
1993
1994                 spin_lock_bh(&queue->rx_lock);
1995                 xennet_alloc_rx_buffers(queue);
1996                 spin_unlock_bh(&queue->rx_lock);
1997         }
1998
1999         return 0;
2000 }
2001
2002 /**
2003  * Callback received when the backend's state changes.
2004  */
2005 static void netback_changed(struct xenbus_device *dev,
2006                             enum xenbus_state backend_state)
2007 {
2008         struct netfront_info *np = dev_get_drvdata(&dev->dev);
2009         struct net_device *netdev = np->netdev;
2010
2011         dev_dbg(&dev->dev, "%s\n", xenbus_strstate(backend_state));
2012
2013         switch (backend_state) {
2014         case XenbusStateInitialising:
2015         case XenbusStateInitialised:
2016         case XenbusStateReconfiguring:
2017         case XenbusStateReconfigured:
2018                 break;
2019
2020         case XenbusStateUnknown:
2021                 wake_up_all(&module_unload_q);
2022                 break;
2023
2024         case XenbusStateInitWait:
2025                 if (dev->state != XenbusStateInitialising)
2026                         break;
2027                 if (xennet_connect(netdev) != 0)
2028                         break;
2029                 xenbus_switch_state(dev, XenbusStateConnected);
2030                 break;
2031
2032         case XenbusStateConnected:
2033                 netdev_notify_peers(netdev);
2034                 break;
2035
2036         case XenbusStateClosed:
2037                 wake_up_all(&module_unload_q);
2038                 if (dev->state == XenbusStateClosed)
2039                         break;
2040                 /* Missed the backend's CLOSING state -- fallthrough */
2041         case XenbusStateClosing:
2042                 wake_up_all(&module_unload_q);
2043                 xenbus_frontend_closed(dev);
2044                 break;
2045         }
2046 }
2047
2048 static const struct xennet_stat {
2049         char name[ETH_GSTRING_LEN];
2050         u16 offset;
2051 } xennet_stats[] = {
2052         {
2053                 "rx_gso_checksum_fixup",
2054                 offsetof(struct netfront_info, rx_gso_checksum_fixup)
2055         },
2056 };
2057
2058 static int xennet_get_sset_count(struct net_device *dev, int string_set)
2059 {
2060         switch (string_set) {
2061         case ETH_SS_STATS:
2062                 return ARRAY_SIZE(xennet_stats);
2063         default:
2064                 return -EINVAL;
2065         }
2066 }
2067
2068 static void xennet_get_ethtool_stats(struct net_device *dev,
2069                                      struct ethtool_stats *stats, u64 * data)
2070 {
2071         void *np = netdev_priv(dev);
2072         int i;
2073
2074         for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
2075                 data[i] = atomic_read((atomic_t *)(np + xennet_stats[i].offset));
2076 }
2077
2078 static void xennet_get_strings(struct net_device *dev, u32 stringset, u8 * data)
2079 {
2080         int i;
2081
2082         switch (stringset) {
2083         case ETH_SS_STATS:
2084                 for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
2085                         memcpy(data + i * ETH_GSTRING_LEN,
2086                                xennet_stats[i].name, ETH_GSTRING_LEN);
2087                 break;
2088         }
2089 }
2090
2091 static const struct ethtool_ops xennet_ethtool_ops =
2092 {
2093         .get_link = ethtool_op_get_link,
2094
2095         .get_sset_count = xennet_get_sset_count,
2096         .get_ethtool_stats = xennet_get_ethtool_stats,
2097         .get_strings = xennet_get_strings,
2098 };
2099
2100 #ifdef CONFIG_SYSFS
2101 static ssize_t show_rxbuf(struct device *dev,
2102                           struct device_attribute *attr, char *buf)
2103 {
2104         return sprintf(buf, "%lu\n", NET_RX_RING_SIZE);
2105 }
2106
2107 static ssize_t store_rxbuf(struct device *dev,
2108                            struct device_attribute *attr,
2109                            const char *buf, size_t len)
2110 {
2111         char *endp;
2112         unsigned long target;
2113
2114         if (!capable(CAP_NET_ADMIN))
2115                 return -EPERM;
2116
2117         target = simple_strtoul(buf, &endp, 0);
2118         if (endp == buf)
2119                 return -EBADMSG;
2120
2121         /* rxbuf_min and rxbuf_max are no longer configurable. */
2122
2123         return len;
2124 }
2125
2126 static DEVICE_ATTR(rxbuf_min, 0644, show_rxbuf, store_rxbuf);
2127 static DEVICE_ATTR(rxbuf_max, 0644, show_rxbuf, store_rxbuf);
2128 static DEVICE_ATTR(rxbuf_cur, 0444, show_rxbuf, NULL);
2129
2130 static struct attribute *xennet_dev_attrs[] = {
2131         &dev_attr_rxbuf_min.attr,
2132         &dev_attr_rxbuf_max.attr,
2133         &dev_attr_rxbuf_cur.attr,
2134         NULL
2135 };
2136
2137 static const struct attribute_group xennet_dev_group = {
2138         .attrs = xennet_dev_attrs
2139 };
2140 #endif /* CONFIG_SYSFS */
2141
2142 static int xennet_remove(struct xenbus_device *dev)
2143 {
2144         struct netfront_info *info = dev_get_drvdata(&dev->dev);
2145
2146         dev_dbg(&dev->dev, "%s\n", dev->nodename);
2147
2148         if (xenbus_read_driver_state(dev->otherend) != XenbusStateClosed) {
2149                 xenbus_switch_state(dev, XenbusStateClosing);
2150                 wait_event(module_unload_q,
2151                            xenbus_read_driver_state(dev->otherend) ==
2152                            XenbusStateClosing ||
2153                            xenbus_read_driver_state(dev->otherend) ==
2154                            XenbusStateUnknown);
2155
2156                 xenbus_switch_state(dev, XenbusStateClosed);
2157                 wait_event(module_unload_q,
2158                            xenbus_read_driver_state(dev->otherend) ==
2159                            XenbusStateClosed ||
2160                            xenbus_read_driver_state(dev->otherend) ==
2161                            XenbusStateUnknown);
2162         }
2163
2164         xennet_disconnect_backend(info);
2165
2166         if (info->netdev->reg_state == NETREG_REGISTERED)
2167                 unregister_netdev(info->netdev);
2168
2169         if (info->queues) {
2170                 rtnl_lock();
2171                 xennet_destroy_queues(info);
2172                 rtnl_unlock();
2173         }
2174         xennet_free_netdev(info->netdev);
2175
2176         return 0;
2177 }
2178
2179 static const struct xenbus_device_id netfront_ids[] = {
2180         { "vif" },
2181         { "" }
2182 };
2183
2184 static struct xenbus_driver netfront_driver = {
2185         .ids = netfront_ids,
2186         .probe = netfront_probe,
2187         .remove = xennet_remove,
2188         .resume = netfront_resume,
2189         .otherend_changed = netback_changed,
2190 };
2191
2192 static int __init netif_init(void)
2193 {
2194         if (!xen_domain())
2195                 return -ENODEV;
2196
2197         if (!xen_has_pv_nic_devices())
2198                 return -ENODEV;
2199
2200         pr_info("Initialising Xen virtual ethernet driver\n");
2201
2202         /* Allow as many queues as there are CPUs inut max. 8 if user has not
2203          * specified a value.
2204          */
2205         if (xennet_max_queues == 0)
2206                 xennet_max_queues = min_t(unsigned int, MAX_QUEUES_DEFAULT,
2207                                           num_online_cpus());
2208
2209         return xenbus_register_frontend(&netfront_driver);
2210 }
2211 module_init(netif_init);
2212
2213
2214 static void __exit netif_exit(void)
2215 {
2216         xenbus_unregister_driver(&netfront_driver);
2217 }
2218 module_exit(netif_exit);
2219
2220 MODULE_DESCRIPTION("Xen virtual network device frontend");
2221 MODULE_LICENSE("GPL");
2222 MODULE_ALIAS("xen:vif");
2223 MODULE_ALIAS("xennet");