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 #include <linux/bpf.h>
  48 #include <net/page_pool.h>
  49 #include <linux/bpf_trace.h>
  50
  51 #include <xen/xen.h>
  52 #include <xen/xenbus.h>
  53 #include <xen/events.h>
  54 #include <xen/page.h>
  55 #include <xen/platform_pci.h>
  56 #include <xen/grant_table.h>
  57
  58 #include <xen/interface/io/netif.h>
  59 #include <xen/interface/memory.h>
  60 #include <xen/interface/grant_table.h>
  61
  62 /* Module parameters */
  63 #define MAX_QUEUES_DEFAULT 8
  64 static unsigned int xennet_max_queues;
  65 module_param_named(max_queues, xennet_max_queues, uint, 0644);
  66 MODULE_PARM_DESC(max_queues,
  67                  "Maximum number of queues per virtual interface");
  68
  69 #define XENNET_TIMEOUT  (5 * HZ)
  70
  71 static const struct ethtool_ops xennet_ethtool_ops;
  72
  73 struct netfront_cb {
  74         int pull_to;
  75 };
  76
  77 #define NETFRONT_SKB_CB(skb)    ((struct netfront_cb *)((skb)->cb))
  78
  79 #define RX_COPY_THRESHOLD 256
  80
  81 #define GRANT_INVALID_REF       0
  82
  83 #define NET_TX_RING_SIZE __CONST_RING_SIZE(xen_netif_tx, XEN_PAGE_SIZE)
  84 #define NET_RX_RING_SIZE __CONST_RING_SIZE(xen_netif_rx, XEN_PAGE_SIZE)
  85
  86 /* Minimum number of Rx slots (includes slot for GSO metadata). */
  87 #define NET_RX_SLOTS_MIN (XEN_NETIF_NR_SLOTS_MIN + 1)
  88
  89 /* Queue name is interface name with "-qNNN" appended */
  90 #define QUEUE_NAME_SIZE (IFNAMSIZ + 6)
  91
  92 /* IRQ name is queue name with "-tx" or "-rx" appended */
  93 #define IRQ_NAME_SIZE (QUEUE_NAME_SIZE + 3)
  94
  95 static DECLARE_WAIT_QUEUE_HEAD(module_wq);
  96
  97 struct netfront_stats {
  98         u64                     packets;
  99         u64                     bytes;
 100         struct u64_stats_sync   syncp;
 101 };
 102
 103 struct netfront_info;
 104
 105 struct netfront_queue {
 106         unsigned int id; /* Queue ID, 0-based */
 107         char name[QUEUE_NAME_SIZE]; /* DEVNAME-qN */
 108         struct netfront_info *info;
 109
 110         struct bpf_prog __rcu *xdp_prog;
 111
 112         struct napi_struct napi;
 113
 114         /* Split event channels support, tx_* == rx_* when using
 115          * single event channel.
 116          */
 117         unsigned int tx_evtchn, rx_evtchn;
 118         unsigned int tx_irq, rx_irq;
 119         /* Only used when split event channels support is enabled */
 120         char tx_irq_name[IRQ_NAME_SIZE]; /* DEVNAME-qN-tx */
 121         char rx_irq_name[IRQ_NAME_SIZE]; /* DEVNAME-qN-rx */
 122
 123         spinlock_t   tx_lock;
 124         struct xen_netif_tx_front_ring tx;
 125         int tx_ring_ref;
 126
 127         /*
 128          * {tx,rx}_skbs store outstanding skbuffs. Free tx_skb entries
 129          * are linked from tx_skb_freelist through skb_entry.link.
 130          *
 131          *  NB. Freelist index entries are always going to be less than
 132          *  PAGE_OFFSET, whereas pointers to skbs will always be equal or
 133          *  greater than PAGE_OFFSET: we use this property to distinguish
 134          *  them.
 135          */
 136         union skb_entry {
 137                 struct sk_buff *skb;
 138                 unsigned long link;
 139         } tx_skbs[NET_TX_RING_SIZE];
 140         grant_ref_t gref_tx_head;
 141         grant_ref_t grant_tx_ref[NET_TX_RING_SIZE];
 142         struct page *grant_tx_page[NET_TX_RING_SIZE];
 143         unsigned tx_skb_freelist;
 144
 145         spinlock_t   rx_lock ____cacheline_aligned_in_smp;
 146         struct xen_netif_rx_front_ring rx;
 147         int rx_ring_ref;
 148
 149         struct timer_list rx_refill_timer;
 150
 151         struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
 152         grant_ref_t gref_rx_head;
 153         grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];
 154
 155         struct page_pool *page_pool;
 156         struct xdp_rxq_info xdp_rxq;
 157 };
 158
 159 struct netfront_info {
 160         struct list_head list;
 161         struct net_device *netdev;
 162
 163         struct xenbus_device *xbdev;
 164
 165         /* Multi-queue support */
 166         struct netfront_queue *queues;
 167
 168         /* Statistics */
 169         struct netfront_stats __percpu *rx_stats;
 170         struct netfront_stats __percpu *tx_stats;
 171
 172         /* XDP state */
 173         bool netback_has_xdp_headroom;
 174         bool netfront_xdp_enabled;
 175
 176         atomic_t rx_gso_checksum_fixup;
 177 };
 178
 179 struct netfront_rx_info {
 180         struct xen_netif_rx_response rx;
 181         struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
 182 };
 183
 184 static void skb_entry_set_link(union skb_entry *list, unsigned short id)
 185 {
 186         list->link = id;
 187 }
 188
 189 static int skb_entry_is_link(const union skb_entry *list)
 190 {
 191         BUILD_BUG_ON(sizeof(list->skb) != sizeof(list->link));
 192         return (unsigned long)list->skb < PAGE_OFFSET;
 193 }
 194
 195 /*
 196  * Access macros for acquiring freeing slots in tx_skbs[].
 197  */
 198
 199 static void add_id_to_freelist(unsigned *head, union skb_entry *list,
 200                                unsigned short id)
 201 {
 202         skb_entry_set_link(&list[id], *head);
 203         *head = id;
 204 }
 205
 206 static unsigned short get_id_from_freelist(unsigned *head,
 207                                            union skb_entry *list)
 208 {
 209         unsigned int id = *head;
 210         *head = list[id].link;
 211         return id;
 212 }
 213
 214 static int xennet_rxidx(RING_IDX idx)
 215 {
 216         return idx & (NET_RX_RING_SIZE - 1);
 217 }
 218
 219 static struct sk_buff *xennet_get_rx_skb(struct netfront_queue *queue,
 220                                          RING_IDX ri)
 221 {
 222         int i = xennet_rxidx(ri);
 223         struct sk_buff *skb = queue->rx_skbs[i];
 224         queue->rx_skbs[i] = NULL;
 225         return skb;
 226 }
 227
 228 static grant_ref_t xennet_get_rx_ref(struct netfront_queue *queue,
 229                                             RING_IDX ri)
 230 {
 231         int i = xennet_rxidx(ri);
 232         grant_ref_t ref = queue->grant_rx_ref[i];
 233         queue->grant_rx_ref[i] = GRANT_INVALID_REF;
 234         return ref;
 235 }
 236
 237 #ifdef CONFIG_SYSFS
 238 static const struct attribute_group xennet_dev_group;
 239 #endif
 240
 241 static bool xennet_can_sg(struct net_device *dev)
 242 {
 243         return dev->features & NETIF_F_SG;
 244 }
 245
 246
 247 static void rx_refill_timeout(struct timer_list *t)
 248 {
 249         struct netfront_queue *queue = from_timer(queue, t, rx_refill_timer);
 250         napi_schedule(&queue->napi);
 251 }
 252
 253 static int netfront_tx_slot_available(struct netfront_queue *queue)
 254 {
 255         return (queue->tx.req_prod_pvt - queue->tx.rsp_cons) <
 256                 (NET_TX_RING_SIZE - XEN_NETIF_NR_SLOTS_MIN - 1);
 257 }
 258
 259 static void xennet_maybe_wake_tx(struct netfront_queue *queue)
 260 {
 261         struct net_device *dev = queue->info->netdev;
 262         struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, queue->id);
 263
 264         if (unlikely(netif_tx_queue_stopped(dev_queue)) &&
 265             netfront_tx_slot_available(queue) &&
 266             likely(netif_running(dev)))
 267                 netif_tx_wake_queue(netdev_get_tx_queue(dev, queue->id));
 268 }
 269
 270
 271 static struct sk_buff *xennet_alloc_one_rx_buffer(struct netfront_queue *queue)
 272 {
 273         struct sk_buff *skb;
 274         struct page *page;
 275
 276         skb = __netdev_alloc_skb(queue->info->netdev,
 277                                  RX_COPY_THRESHOLD + NET_IP_ALIGN,
 278                                  GFP_ATOMIC | __GFP_NOWARN);
 279         if (unlikely(!skb))
 280                 return NULL;
 281
 282         page = page_pool_dev_alloc_pages(queue->page_pool);
 283         if (unlikely(!page)) {
 284                 kfree_skb(skb);
 285                 return NULL;
 286         }
 287         skb_add_rx_frag(skb, 0, page, 0, 0, PAGE_SIZE);
 288
 289         /* Align ip header to a 16 bytes boundary */
 290         skb_reserve(skb, NET_IP_ALIGN);
 291         skb->dev = queue->info->netdev;
 292
 293         return skb;
 294 }
 295
 296
 297 static void xennet_alloc_rx_buffers(struct netfront_queue *queue)
 298 {
 299         RING_IDX req_prod = queue->rx.req_prod_pvt;
 300         int notify;
 301         int err = 0;
 302
 303         if (unlikely(!netif_carrier_ok(queue->info->netdev)))
 304                 return;
 305
 306         for (req_prod = queue->rx.req_prod_pvt;
 307              req_prod - queue->rx.rsp_cons < NET_RX_RING_SIZE;
 308              req_prod++) {
 309                 struct sk_buff *skb;
 310                 unsigned short id;
 311                 grant_ref_t ref;
 312                 struct page *page;
 313                 struct xen_netif_rx_request *req;
 314
 315                 skb = xennet_alloc_one_rx_buffer(queue);
 316                 if (!skb) {
 317                         err = -ENOMEM;
 318                         break;
 319                 }
 320
 321                 id = xennet_rxidx(req_prod);
 322
 323                 BUG_ON(queue->rx_skbs[id]);
 324                 queue->rx_skbs[id] = skb;
 325
 326                 ref = gnttab_claim_grant_reference(&queue->gref_rx_head);
 327                 WARN_ON_ONCE(IS_ERR_VALUE((unsigned long)(int)ref));
 328                 queue->grant_rx_ref[id] = ref;
 329
 330                 page = skb_frag_page(&skb_shinfo(skb)->frags[0]);
 331
 332                 req = RING_GET_REQUEST(&queue->rx, req_prod);
 333                 gnttab_page_grant_foreign_access_ref_one(ref,
 334                                                          queue->info->xbdev->otherend_id,
 335                                                          page,
 336                                                          0);
 337                 req->id = id;
 338                 req->gref = ref;
 339         }
 340
 341         queue->rx.req_prod_pvt = req_prod;
 342
 343         /* Try again later if there are not enough requests or skb allocation
 344          * failed.
 345          * Enough requests is quantified as the sum of newly created slots and
 346          * the unconsumed slots at the backend.
 347          */
 348         if (req_prod - queue->rx.rsp_cons < NET_RX_SLOTS_MIN ||
 349             unlikely(err)) {
 350                 mod_timer(&queue->rx_refill_timer, jiffies + (HZ/10));
 351                 return;
 352         }
 353
 354         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->rx, notify);
 355         if (notify)
 356                 notify_remote_via_irq(queue->rx_irq);
 357 }
 358
 359 static int xennet_open(struct net_device *dev)
 360 {
 361         struct netfront_info *np = netdev_priv(dev);
 362         unsigned int num_queues = dev->real_num_tx_queues;
 363         unsigned int i = 0;
 364         struct netfront_queue *queue = NULL;
 365
 366         if (!np->queues)
 367                 return -ENODEV;
 368
 369         for (i = 0; i < num_queues; ++i) {
 370                 queue = &np->queues[i];
 371                 napi_enable(&queue->napi);
 372
 373                 spin_lock_bh(&queue->rx_lock);
 374                 if (netif_carrier_ok(dev)) {
 375                         xennet_alloc_rx_buffers(queue);
 376                         queue->rx.sring->rsp_event = queue->rx.rsp_cons + 1;
 377                         if (RING_HAS_UNCONSUMED_RESPONSES(&queue->rx))
 378                                 napi_schedule(&queue->napi);
 379                 }
 380                 spin_unlock_bh(&queue->rx_lock);
 381         }
 382
 383         netif_tx_start_all_queues(dev);
 384
 385         return 0;
 386 }
 387
 388 static void xennet_tx_buf_gc(struct netfront_queue *queue)
 389 {
 390         RING_IDX cons, prod;
 391         unsigned short id;
 392         struct sk_buff *skb;
 393         bool more_to_do;
 394
 395         BUG_ON(!netif_carrier_ok(queue->info->netdev));
 396
 397         do {
 398                 prod = queue->tx.sring->rsp_prod;
 399                 rmb(); /* Ensure we see responses up to 'rp'. */
 400
 401                 for (cons = queue->tx.rsp_cons; cons != prod; cons++) {
 402                         struct xen_netif_tx_response *txrsp;
 403
 404                         txrsp = RING_GET_RESPONSE(&queue->tx, cons);
 405                         if (txrsp->status == XEN_NETIF_RSP_NULL)
 406                                 continue;
 407
 408                         id  = txrsp->id;
 409                         skb = queue->tx_skbs[id].skb;
 410                         if (unlikely(gnttab_query_foreign_access(
 411                                 queue->grant_tx_ref[id]) != 0)) {
 412                                 pr_alert("%s: warning -- grant still in use by backend domain\n",
 413                                          __func__);
 414                                 BUG();
 415                         }
 416                         gnttab_end_foreign_access_ref(
 417                                 queue->grant_tx_ref[id], GNTMAP_readonly);
 418                         gnttab_release_grant_reference(
 419                                 &queue->gref_tx_head, queue->grant_tx_ref[id]);
 420                         queue->grant_tx_ref[id] = GRANT_INVALID_REF;
 421                         queue->grant_tx_page[id] = NULL;
 422                         add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, id);
 423                         dev_kfree_skb_irq(skb);
 424                 }
 425
 426                 queue->tx.rsp_cons = prod;
 427
 428                 RING_FINAL_CHECK_FOR_RESPONSES(&queue->tx, more_to_do);
 429         } while (more_to_do);
 430
 431         xennet_maybe_wake_tx(queue);
 432 }
 433
 434 struct xennet_gnttab_make_txreq {
 435         struct netfront_queue *queue;
 436         struct sk_buff *skb;
 437         struct page *page;
 438         struct xen_netif_tx_request *tx; /* Last request */
 439         unsigned int size;
 440 };
 441
 442 static void xennet_tx_setup_grant(unsigned long gfn, unsigned int offset,
 443                                   unsigned int len, void *data)
 444 {
 445         struct xennet_gnttab_make_txreq *info = data;
 446         unsigned int id;
 447         struct xen_netif_tx_request *tx;
 448         grant_ref_t ref;
 449         /* convenient aliases */
 450         struct page *page = info->page;
 451         struct netfront_queue *queue = info->queue;
 452         struct sk_buff *skb = info->skb;
 453
 454         id = get_id_from_freelist(&queue->tx_skb_freelist, queue->tx_skbs);
 455         tx = RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 456         ref = gnttab_claim_grant_reference(&queue->gref_tx_head);
 457         WARN_ON_ONCE(IS_ERR_VALUE((unsigned long)(int)ref));
 458
 459         gnttab_grant_foreign_access_ref(ref, queue->info->xbdev->otherend_id,
 460                                         gfn, GNTMAP_readonly);
 461
 462         queue->tx_skbs[id].skb = skb;
 463         queue->grant_tx_page[id] = page;
 464         queue->grant_tx_ref[id] = ref;
 465
 466         tx->id = id;
 467         tx->gref = ref;
 468         tx->offset = offset;
 469         tx->size = len;
 470         tx->flags = 0;
 471
 472         info->tx = tx;
 473         info->size += tx->size;
 474 }
 475
 476 static struct xen_netif_tx_request *xennet_make_first_txreq(
 477         struct netfront_queue *queue, struct sk_buff *skb,
 478         struct page *page, unsigned int offset, unsigned int len)
 479 {
 480         struct xennet_gnttab_make_txreq info = {
 481                 .queue = queue,
 482                 .skb = skb,
 483                 .page = page,
 484                 .size = 0,
 485         };
 486
 487         gnttab_for_one_grant(page, offset, len, xennet_tx_setup_grant, &info);
 488
 489         return info.tx;
 490 }
 491
 492 static void xennet_make_one_txreq(unsigned long gfn, unsigned int offset,
 493                                   unsigned int len, void *data)
 494 {
 495         struct xennet_gnttab_make_txreq *info = data;
 496
 497         info->tx->flags |= XEN_NETTXF_more_data;
 498         skb_get(info->skb);
 499         xennet_tx_setup_grant(gfn, offset, len, data);
 500 }
 501
 502 static struct xen_netif_tx_request *xennet_make_txreqs(
 503         struct netfront_queue *queue, struct xen_netif_tx_request *tx,
 504         struct sk_buff *skb, struct page *page,
 505         unsigned int offset, unsigned int len)
 506 {
 507         struct xennet_gnttab_make_txreq info = {
 508                 .queue = queue,
 509                 .skb = skb,
 510                 .tx = tx,
 511         };
 512
 513         /* Skip unused frames from start of page */
 514         page += offset >> PAGE_SHIFT;
 515         offset &= ~PAGE_MASK;
 516
 517         while (len) {
 518                 info.page = page;
 519                 info.size = 0;
 520
 521                 gnttab_foreach_grant_in_range(page, offset, len,
 522                                               xennet_make_one_txreq,
 523                                               &info);
 524
 525                 page++;
 526                 offset = 0;
 527                 len -= info.size;
 528         }
 529
 530         return info.tx;
 531 }
 532
 533 /*
 534  * Count how many ring slots are required to send this skb. Each frag
 535  * might be a compound page.
 536  */
 537 static int xennet_count_skb_slots(struct sk_buff *skb)
 538 {
 539         int i, frags = skb_shinfo(skb)->nr_frags;
 540         int slots;
 541
 542         slots = gnttab_count_grant(offset_in_page(skb->data),
 543                                    skb_headlen(skb));
 544
 545         for (i = 0; i < frags; i++) {
 546                 skb_frag_t *frag = skb_shinfo(skb)->frags + i;
 547                 unsigned long size = skb_frag_size(frag);
 548                 unsigned long offset = skb_frag_off(frag);
 549
 550                 /* Skip unused frames from start of page */
 551                 offset &= ~PAGE_MASK;
 552
 553                 slots += gnttab_count_grant(offset, size);
 554         }
 555
 556         return slots;
 557 }
 558
 559 static u16 xennet_select_queue(struct net_device *dev, struct sk_buff *skb,
 560                                struct net_device *sb_dev)
 561 {
 562         unsigned int num_queues = dev->real_num_tx_queues;
 563         u32 hash;
 564         u16 queue_idx;
 565
 566         /* First, check if there is only one queue */
 567         if (num_queues == 1) {
 568                 queue_idx = 0;
 569         } else {
 570                 hash = skb_get_hash(skb);
 571                 queue_idx = hash % num_queues;
 572         }
 573
 574         return queue_idx;
 575 }
 576
 577 static int xennet_xdp_xmit_one(struct net_device *dev,
 578                                struct netfront_queue *queue,
 579                                struct xdp_frame *xdpf)
 580 {
 581         struct netfront_info *np = netdev_priv(dev);
 582         struct netfront_stats *tx_stats = this_cpu_ptr(np->tx_stats);
 583         int notify;
 584
 585         xennet_make_first_txreq(queue, NULL,
 586                                 virt_to_page(xdpf->data),
 587                                 offset_in_page(xdpf->data),
 588                                 xdpf->len);
 589
 590         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->tx, notify);
 591         if (notify)
 592                 notify_remote_via_irq(queue->tx_irq);
 593
 594         u64_stats_update_begin(&tx_stats->syncp);
 595         tx_stats->bytes += xdpf->len;
 596         tx_stats->packets++;
 597         u64_stats_update_end(&tx_stats->syncp);
 598
 599         xennet_tx_buf_gc(queue);
 600
 601         return 0;
 602 }
 603
 604 static int xennet_xdp_xmit(struct net_device *dev, int n,
 605                            struct xdp_frame **frames, u32 flags)
 606 {
 607         unsigned int num_queues = dev->real_num_tx_queues;
 608         struct netfront_info *np = netdev_priv(dev);
 609         struct netfront_queue *queue = NULL;
 610         unsigned long irq_flags;
 611         int drops = 0;
 612         int i, err;
 613
 614         if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK))
 615                 return -EINVAL;
 616
 617         queue = &np->queues[smp_processor_id() % num_queues];
 618
 619         spin_lock_irqsave(&queue->tx_lock, irq_flags);
 620         for (i = 0; i < n; i++) {
 621                 struct xdp_frame *xdpf = frames[i];
 622
 623                 if (!xdpf)
 624                         continue;
 625                 err = xennet_xdp_xmit_one(dev, queue, xdpf);
 626                 if (err) {
 627                         xdp_return_frame_rx_napi(xdpf);
 628                         drops++;
 629                 }
 630         }
 631         spin_unlock_irqrestore(&queue->tx_lock, irq_flags);
 632
 633         return n - drops;
 634 }
 635
 636
 637 #define MAX_XEN_SKB_FRAGS (65536 / XEN_PAGE_SIZE + 1)
 638
 639 static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 640 {
 641         struct netfront_info *np = netdev_priv(dev);
 642         struct netfront_stats *tx_stats = this_cpu_ptr(np->tx_stats);
 643         struct xen_netif_tx_request *tx, *first_tx;
 644         unsigned int i;
 645         int notify;
 646         int slots;
 647         struct page *page;
 648         unsigned int offset;
 649         unsigned int len;
 650         unsigned long flags;
 651         struct netfront_queue *queue = NULL;
 652         unsigned int num_queues = dev->real_num_tx_queues;
 653         u16 queue_index;
 654         struct sk_buff *nskb;
 655
 656         /* Drop the packet if no queues are set up */
 657         if (num_queues < 1)
 658                 goto drop;
 659         /* Determine which queue to transmit this SKB on */
 660         queue_index = skb_get_queue_mapping(skb);
 661         queue = &np->queues[queue_index];
 662
 663         /* If skb->len is too big for wire format, drop skb and alert
 664          * user about misconfiguration.
 665          */
 666         if (unlikely(skb->len > XEN_NETIF_MAX_TX_SIZE)) {
 667                 net_alert_ratelimited(
 668                         "xennet: skb->len = %u, too big for wire format\n",
 669                         skb->len);
 670                 goto drop;
 671         }
 672
 673         slots = xennet_count_skb_slots(skb);
 674         if (unlikely(slots > MAX_XEN_SKB_FRAGS + 1)) {
 675                 net_dbg_ratelimited("xennet: skb rides the rocket: %d slots, %d bytes\n",
 676                                     slots, skb->len);
 677                 if (skb_linearize(skb))
 678                         goto drop;
 679         }
 680
 681         page = virt_to_page(skb->data);
 682         offset = offset_in_page(skb->data);
 683
 684         /* The first req should be at least ETH_HLEN size or the packet will be
 685          * dropped by netback.
 686          */
 687         if (unlikely(PAGE_SIZE - offset < ETH_HLEN)) {
 688                 nskb = skb_copy(skb, GFP_ATOMIC);
 689                 if (!nskb)
 690                         goto drop;
 691                 dev_consume_skb_any(skb);
 692                 skb = nskb;
 693                 page = virt_to_page(skb->data);
 694                 offset = offset_in_page(skb->data);
 695         }
 696
 697         len = skb_headlen(skb);
 698
 699         spin_lock_irqsave(&queue->tx_lock, flags);
 700
 701         if (unlikely(!netif_carrier_ok(dev) ||
 702                      (slots > 1 && !xennet_can_sg(dev)) ||
 703                      netif_needs_gso(skb, netif_skb_features(skb)))) {
 704                 spin_unlock_irqrestore(&queue->tx_lock, flags);
 705                 goto drop;
 706         }
 707
 708         /* First request for the linear area. */
 709         first_tx = tx = xennet_make_first_txreq(queue, skb,
 710                                                 page, offset, len);
 711         offset += tx->size;
 712         if (offset == PAGE_SIZE) {
 713                 page++;
 714                 offset = 0;
 715         }
 716         len -= tx->size;
 717
 718         if (skb->ip_summed == CHECKSUM_PARTIAL)
 719                 /* local packet? */
 720                 tx->flags |= XEN_NETTXF_csum_blank | XEN_NETTXF_data_validated;
 721         else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
 722                 /* remote but checksummed. */
 723                 tx->flags |= XEN_NETTXF_data_validated;
 724
 725         /* Optional extra info after the first request. */
 726         if (skb_shinfo(skb)->gso_size) {
 727                 struct xen_netif_extra_info *gso;
 728
 729                 gso = (struct xen_netif_extra_info *)
 730                         RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 731
 732                 tx->flags |= XEN_NETTXF_extra_info;
 733
 734                 gso->u.gso.size = skb_shinfo(skb)->gso_size;
 735                 gso->u.gso.type = (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) ?
 736                         XEN_NETIF_GSO_TYPE_TCPV6 :
 737                         XEN_NETIF_GSO_TYPE_TCPV4;
 738                 gso->u.gso.pad = 0;
 739                 gso->u.gso.features = 0;
 740
 741                 gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
 742                 gso->flags = 0;
 743         }
 744
 745         /* Requests for the rest of the linear area. */
 746         tx = xennet_make_txreqs(queue, tx, skb, page, offset, len);
 747
 748         /* Requests for all the frags. */
 749         for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 750                 skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 751                 tx = xennet_make_txreqs(queue, tx, skb, skb_frag_page(frag),
 752                                         skb_frag_off(frag),
 753                                         skb_frag_size(frag));
 754         }
 755
 756         /* First request has the packet length. */
 757         first_tx->size = skb->len;
 758
 759         /* timestamp packet in software */
 760         skb_tx_timestamp(skb);
 761
 762         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->tx, notify);
 763         if (notify)
 764                 notify_remote_via_irq(queue->tx_irq);
 765
 766         u64_stats_update_begin(&tx_stats->syncp);
 767         tx_stats->bytes += skb->len;
 768         tx_stats->packets++;
 769         u64_stats_update_end(&tx_stats->syncp);
 770
 771         /* Note: It is not safe to access skb after xennet_tx_buf_gc()! */
 772         xennet_tx_buf_gc(queue);
 773
 774         if (!netfront_tx_slot_available(queue))
 775                 netif_tx_stop_queue(netdev_get_tx_queue(dev, queue->id));
 776
 777         spin_unlock_irqrestore(&queue->tx_lock, flags);
 778
 779         return NETDEV_TX_OK;
 780
 781  drop:
 782         dev->stats.tx_dropped++;
 783         dev_kfree_skb_any(skb);
 784         return NETDEV_TX_OK;
 785 }
 786
 787 static int xennet_close(struct net_device *dev)
 788 {
 789         struct netfront_info *np = netdev_priv(dev);
 790         unsigned int num_queues = dev->real_num_tx_queues;
 791         unsigned int i;
 792         struct netfront_queue *queue;
 793         netif_tx_stop_all_queues(np->netdev);
 794         for (i = 0; i < num_queues; ++i) {
 795                 queue = &np->queues[i];
 796                 napi_disable(&queue->napi);
 797         }
 798         return 0;
 799 }
 800
 801 static void xennet_move_rx_slot(struct netfront_queue *queue, struct sk_buff *skb,
 802                                 grant_ref_t ref)
 803 {
 804         int new = xennet_rxidx(queue->rx.req_prod_pvt);
 805
 806         BUG_ON(queue->rx_skbs[new]);
 807         queue->rx_skbs[new] = skb;
 808         queue->grant_rx_ref[new] = ref;
 809         RING_GET_REQUEST(&queue->rx, queue->rx.req_prod_pvt)->id = new;
 810         RING_GET_REQUEST(&queue->rx, queue->rx.req_prod_pvt)->gref = ref;
 811         queue->rx.req_prod_pvt++;
 812 }
 813
 814 static int xennet_get_extras(struct netfront_queue *queue,
 815                              struct xen_netif_extra_info *extras,
 816                              RING_IDX rp)
 817
 818 {
 819         struct xen_netif_extra_info *extra;
 820         struct device *dev = &queue->info->netdev->dev;
 821         RING_IDX cons = queue->rx.rsp_cons;
 822         int err = 0;
 823
 824         do {
 825                 struct sk_buff *skb;
 826                 grant_ref_t ref;
 827
 828                 if (unlikely(cons + 1 == rp)) {
 829                         if (net_ratelimit())
 830                                 dev_warn(dev, "Missing extra info\n");
 831                         err = -EBADR;
 832                         break;
 833                 }
 834
 835                 extra = (struct xen_netif_extra_info *)
 836                         RING_GET_RESPONSE(&queue->rx, ++cons);
 837
 838                 if (unlikely(!extra->type ||
 839                              extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
 840                         if (net_ratelimit())
 841                                 dev_warn(dev, "Invalid extra type: %d\n",
 842                                         extra->type);
 843                         err = -EINVAL;
 844                 } else {
 845                         memcpy(&extras[extra->type - 1], extra,
 846                                sizeof(*extra));
 847                 }
 848
 849                 skb = xennet_get_rx_skb(queue, cons);
 850                 ref = xennet_get_rx_ref(queue, cons);
 851                 xennet_move_rx_slot(queue, skb, ref);
 852         } while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);
 853
 854         queue->rx.rsp_cons = cons;
 855         return err;
 856 }
 857
 858 static u32 xennet_run_xdp(struct netfront_queue *queue, struct page *pdata,
 859                    struct xen_netif_rx_response *rx, struct bpf_prog *prog,
 860                    struct xdp_buff *xdp, bool *need_xdp_flush)
 861 {
 862         struct xdp_frame *xdpf;
 863         u32 len = rx->status;
 864         u32 act;
 865         int err;
 866
 867         xdp->data_hard_start = page_address(pdata);
 868         xdp->data = xdp->data_hard_start + XDP_PACKET_HEADROOM;
 869         xdp_set_data_meta_invalid(xdp);
 870         xdp->data_end = xdp->data + len;
 871         xdp->rxq = &queue->xdp_rxq;
 872         xdp->frame_sz = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM;
 873
 874         act = bpf_prog_run_xdp(prog, xdp);
 875         switch (act) {
 876         case XDP_TX:
 877                 get_page(pdata);
 878                 xdpf = xdp_convert_buff_to_frame(xdp);
 879                 err = xennet_xdp_xmit(queue->info->netdev, 1, &xdpf, 0);
 880                 if (unlikely(err < 0))
 881                         trace_xdp_exception(queue->info->netdev, prog, act);
 882                 break;
 883         case XDP_REDIRECT:
 884                 get_page(pdata);
 885                 err = xdp_do_redirect(queue->info->netdev, xdp, prog);
 886                 *need_xdp_flush = true;
 887                 if (unlikely(err))
 888                         trace_xdp_exception(queue->info->netdev, prog, act);
 889                 break;
 890         case XDP_PASS:
 891         case XDP_DROP:
 892                 break;
 893
 894         case XDP_ABORTED:
 895                 trace_xdp_exception(queue->info->netdev, prog, act);
 896                 break;
 897
 898         default:
 899                 bpf_warn_invalid_xdp_action(act);
 900         }
 901
 902         return act;
 903 }
 904
 905 static int xennet_get_responses(struct netfront_queue *queue,
 906                                 struct netfront_rx_info *rinfo, RING_IDX rp,
 907                                 struct sk_buff_head *list,
 908                                 bool *need_xdp_flush)
 909 {
 910         struct xen_netif_rx_response *rx = &rinfo->rx;
 911         int max = XEN_NETIF_NR_SLOTS_MIN + (rx->status <= RX_COPY_THRESHOLD);
 912         RING_IDX cons = queue->rx.rsp_cons;
 913         struct sk_buff *skb = xennet_get_rx_skb(queue, cons);
 914         struct xen_netif_extra_info *extras = rinfo->extras;
 915         grant_ref_t ref = xennet_get_rx_ref(queue, cons);
 916         struct device *dev = &queue->info->netdev->dev;
 917         struct bpf_prog *xdp_prog;
 918         struct xdp_buff xdp;
 919         unsigned long ret;
 920         int slots = 1;
 921         int err = 0;
 922         u32 verdict;
 923
 924         if (rx->flags & XEN_NETRXF_extra_info) {
 925                 err = xennet_get_extras(queue, extras, rp);
 926                 if (!err) {
 927                         if (extras[XEN_NETIF_EXTRA_TYPE_XDP - 1].type) {
 928                                 struct xen_netif_extra_info *xdp;
 929
 930                                 xdp = &extras[XEN_NETIF_EXTRA_TYPE_XDP - 1];
 931                                 rx->offset = xdp->u.xdp.headroom;
 932                         }
 933                 }
 934                 cons = queue->rx.rsp_cons;
 935         }
 936
 937         for (;;) {
 938                 if (unlikely(rx->status < 0 ||
 939                              rx->offset + rx->status > XEN_PAGE_SIZE)) {
 940                         if (net_ratelimit())
 941                                 dev_warn(dev, "rx->offset: %u, size: %d\n",
 942                                          rx->offset, rx->status);
 943                         xennet_move_rx_slot(queue, skb, ref);
 944                         err = -EINVAL;
 945                         goto next;
 946                 }
 947
 948                 /*
 949                  * This definitely indicates a bug, either in this driver or in
 950                  * the backend driver. In future this should flag the bad
 951                  * situation to the system controller to reboot the backend.
 952                  */
 953                 if (ref == GRANT_INVALID_REF) {
 954                         if (net_ratelimit())
 955                                 dev_warn(dev, "Bad rx response id %d.\n",
 956                                          rx->id);
 957                         err = -EINVAL;
 958                         goto next;
 959                 }
 960
 961                 ret = gnttab_end_foreign_access_ref(ref, 0);
 962                 BUG_ON(!ret);
 963
 964                 gnttab_release_grant_reference(&queue->gref_rx_head, ref);
 965
 966                 rcu_read_lock();
 967                 xdp_prog = rcu_dereference(queue->xdp_prog);
 968                 if (xdp_prog) {
 969                         if (!(rx->flags & XEN_NETRXF_more_data)) {
 970                                 /* currently only a single page contains data */
 971                                 verdict = xennet_run_xdp(queue,
 972                                                          skb_frag_page(&skb_shinfo(skb)->frags[0]),
 973                                                          rx, xdp_prog, &xdp, need_xdp_flush);
 974                                 if (verdict != XDP_PASS)
 975                                         err = -EINVAL;
 976                         } else {
 977                                 /* drop the frame */
 978                                 err = -EINVAL;
 979                         }
 980                 }
 981                 rcu_read_unlock();
 982 next:
 983                 __skb_queue_tail(list, skb);
 984                 if (!(rx->flags & XEN_NETRXF_more_data))
 985                         break;
 986
 987                 if (cons + slots == rp) {
 988                         if (net_ratelimit())
 989                                 dev_warn(dev, "Need more slots\n");
 990                         err = -ENOENT;
 991                         break;
 992                 }
 993
 994                 rx = RING_GET_RESPONSE(&queue->rx, cons + slots);
 995                 skb = xennet_get_rx_skb(queue, cons + slots);
 996                 ref = xennet_get_rx_ref(queue, cons + slots);
 997                 slots++;
 998         }
 999
1000         if (unlikely(slots > max)) {
1001                 if (net_ratelimit())
1002                         dev_warn(dev, "Too many slots\n");
1003                 err = -E2BIG;
1004         }
1005
1006         if (unlikely(err))
1007                 queue->rx.rsp_cons = cons + slots;
1008
1009         return err;
1010 }
1011
1012 static int xennet_set_skb_gso(struct sk_buff *skb,
1013                               struct xen_netif_extra_info *gso)
1014 {
1015         if (!gso->u.gso.size) {
1016                 if (net_ratelimit())
1017                         pr_warn("GSO size must not be zero\n");
1018                 return -EINVAL;
1019         }
1020
1021         if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4 &&
1022             gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV6) {
1023                 if (net_ratelimit())
1024                         pr_warn("Bad GSO type %d\n", gso->u.gso.type);
1025                 return -EINVAL;
1026         }
1027
1028         skb_shinfo(skb)->gso_size = gso->u.gso.size;
1029         skb_shinfo(skb)->gso_type =
1030                 (gso->u.gso.type == XEN_NETIF_GSO_TYPE_TCPV4) ?
1031                 SKB_GSO_TCPV4 :
1032                 SKB_GSO_TCPV6;
1033
1034         /* Header must be checked, and gso_segs computed. */
1035         skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
1036         skb_shinfo(skb)->gso_segs = 0;
1037
1038         return 0;
1039 }
1040
1041 static int xennet_fill_frags(struct netfront_queue *queue,
1042                              struct sk_buff *skb,
1043                              struct sk_buff_head *list)
1044 {
1045         RING_IDX cons = queue->rx.rsp_cons;
1046         struct sk_buff *nskb;
1047
1048         while ((nskb = __skb_dequeue(list))) {
1049                 struct xen_netif_rx_response *rx =
1050                         RING_GET_RESPONSE(&queue->rx, ++cons);
1051                 skb_frag_t *nfrag = &skb_shinfo(nskb)->frags[0];
1052
1053                 if (skb_shinfo(skb)->nr_frags == MAX_SKB_FRAGS) {
1054                         unsigned int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
1055
1056                         BUG_ON(pull_to < skb_headlen(skb));
1057                         __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
1058                 }
1059                 if (unlikely(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS)) {
1060                         queue->rx.rsp_cons = ++cons + skb_queue_len(list);
1061                         kfree_skb(nskb);
1062                         return -ENOENT;
1063                 }
1064
1065                 skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
1066                                 skb_frag_page(nfrag),
1067                                 rx->offset, rx->status, PAGE_SIZE);
1068
1069                 skb_shinfo(nskb)->nr_frags = 0;
1070                 kfree_skb(nskb);
1071         }
1072
1073         queue->rx.rsp_cons = cons;
1074
1075         return 0;
1076 }
1077
1078 static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
1079 {
1080         bool recalculate_partial_csum = false;
1081
1082         /*
1083          * A GSO SKB must be CHECKSUM_PARTIAL. However some buggy
1084          * peers can fail to set NETRXF_csum_blank when sending a GSO
1085          * frame. In this case force the SKB to CHECKSUM_PARTIAL and
1086          * recalculate the partial checksum.
1087          */
1088         if (skb->ip_summed != CHECKSUM_PARTIAL && skb_is_gso(skb)) {
1089                 struct netfront_info *np = netdev_priv(dev);
1090                 atomic_inc(&np->rx_gso_checksum_fixup);
1091                 skb->ip_summed = CHECKSUM_PARTIAL;
1092                 recalculate_partial_csum = true;
1093         }
1094
1095         /* A non-CHECKSUM_PARTIAL SKB does not require setup. */
1096         if (skb->ip_summed != CHECKSUM_PARTIAL)
1097                 return 0;
1098
1099         return skb_checksum_setup(skb, recalculate_partial_csum);
1100 }
1101
1102 static int handle_incoming_queue(struct netfront_queue *queue,
1103                                  struct sk_buff_head *rxq)
1104 {
1105         struct netfront_stats *rx_stats = this_cpu_ptr(queue->info->rx_stats);
1106         int packets_dropped = 0;
1107         struct sk_buff *skb;
1108
1109         while ((skb = __skb_dequeue(rxq)) != NULL) {
1110                 int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
1111
1112                 if (pull_to > skb_headlen(skb))
1113                         __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
1114
1115                 /* Ethernet work: Delayed to here as it peeks the header. */
1116                 skb->protocol = eth_type_trans(skb, queue->info->netdev);
1117                 skb_reset_network_header(skb);
1118
1119                 if (checksum_setup(queue->info->netdev, skb)) {
1120                         kfree_skb(skb);
1121                         packets_dropped++;
1122                         queue->info->netdev->stats.rx_errors++;
1123                         continue;
1124                 }
1125
1126                 u64_stats_update_begin(&rx_stats->syncp);
1127                 rx_stats->packets++;
1128                 rx_stats->bytes += skb->len;
1129                 u64_stats_update_end(&rx_stats->syncp);
1130
1131                 /* Pass it up. */
1132                 napi_gro_receive(&queue->napi, skb);
1133         }
1134
1135         return packets_dropped;
1136 }
1137
1138 static int xennet_poll(struct napi_struct *napi, int budget)
1139 {
1140         struct netfront_queue *queue = container_of(napi, struct netfront_queue, napi);
1141         struct net_device *dev = queue->info->netdev;
1142         struct sk_buff *skb;
1143         struct netfront_rx_info rinfo;
1144         struct xen_netif_rx_response *rx = &rinfo.rx;
1145         struct xen_netif_extra_info *extras = rinfo.extras;
1146         RING_IDX i, rp;
1147         int work_done;
1148         struct sk_buff_head rxq;
1149         struct sk_buff_head errq;
1150         struct sk_buff_head tmpq;
1151         int err;
1152         bool need_xdp_flush = false;
1153
1154         spin_lock(&queue->rx_lock);
1155
1156         skb_queue_head_init(&rxq);
1157         skb_queue_head_init(&errq);
1158         skb_queue_head_init(&tmpq);
1159
1160         rp = queue->rx.sring->rsp_prod;
1161         rmb(); /* Ensure we see queued responses up to 'rp'. */
1162
1163         i = queue->rx.rsp_cons;
1164         work_done = 0;
1165         while ((i != rp) && (work_done < budget)) {
1166                 memcpy(rx, RING_GET_RESPONSE(&queue->rx, i), sizeof(*rx));
1167                 memset(extras, 0, sizeof(rinfo.extras));
1168
1169                 err = xennet_get_responses(queue, &rinfo, rp, &tmpq,
1170                                            &need_xdp_flush);
1171
1172                 if (unlikely(err)) {
1173 err:
1174                         while ((skb = __skb_dequeue(&tmpq)))
1175                                 __skb_queue_tail(&errq, skb);
1176                         dev->stats.rx_errors++;
1177                         i = queue->rx.rsp_cons;
1178                         continue;
1179                 }
1180
1181                 skb = __skb_dequeue(&tmpq);
1182
1183                 if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
1184                         struct xen_netif_extra_info *gso;
1185                         gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
1186
1187                         if (unlikely(xennet_set_skb_gso(skb, gso))) {
1188                                 __skb_queue_head(&tmpq, skb);
1189                                 queue->rx.rsp_cons += skb_queue_len(&tmpq);
1190                                 goto err;
1191                         }
1192                 }
1193
1194                 NETFRONT_SKB_CB(skb)->pull_to = rx->status;
1195                 if (NETFRONT_SKB_CB(skb)->pull_to > RX_COPY_THRESHOLD)
1196                         NETFRONT_SKB_CB(skb)->pull_to = RX_COPY_THRESHOLD;
1197
1198                 skb_frag_off_set(&skb_shinfo(skb)->frags[0], rx->offset);
1199                 skb_frag_size_set(&skb_shinfo(skb)->frags[0], rx->status);
1200                 skb->data_len = rx->status;
1201                 skb->len += rx->status;
1202
1203                 if (unlikely(xennet_fill_frags(queue, skb, &tmpq)))
1204                         goto err;
1205
1206                 if (rx->flags & XEN_NETRXF_csum_blank)
1207                         skb->ip_summed = CHECKSUM_PARTIAL;
1208                 else if (rx->flags & XEN_NETRXF_data_validated)
1209                         skb->ip_summed = CHECKSUM_UNNECESSARY;
1210
1211                 __skb_queue_tail(&rxq, skb);
1212
1213                 i = ++queue->rx.rsp_cons;
1214                 work_done++;
1215         }
1216         if (need_xdp_flush)
1217                 xdp_do_flush();
1218
1219         __skb_queue_purge(&errq);
1220
1221         work_done -= handle_incoming_queue(queue, &rxq);
1222
1223         xennet_alloc_rx_buffers(queue);
1224
1225         if (work_done < budget) {
1226                 int more_to_do = 0;
1227
1228                 napi_complete_done(napi, work_done);
1229
1230                 RING_FINAL_CHECK_FOR_RESPONSES(&queue->rx, more_to_do);
1231                 if (more_to_do)
1232                         napi_schedule(napi);
1233         }
1234
1235         spin_unlock(&queue->rx_lock);
1236
1237         return work_done;
1238 }
1239
1240 static int xennet_change_mtu(struct net_device *dev, int mtu)
1241 {
1242         int max = xennet_can_sg(dev) ? XEN_NETIF_MAX_TX_SIZE : ETH_DATA_LEN;
1243
1244         if (mtu > max)
1245                 return -EINVAL;
1246         dev->mtu = mtu;
1247         return 0;
1248 }
1249
1250 static void xennet_get_stats64(struct net_device *dev,
1251                                struct rtnl_link_stats64 *tot)
1252 {
1253         struct netfront_info *np = netdev_priv(dev);
1254         int cpu;
1255
1256         for_each_possible_cpu(cpu) {
1257                 struct netfront_stats *rx_stats = per_cpu_ptr(np->rx_stats, cpu);
1258                 struct netfront_stats *tx_stats = per_cpu_ptr(np->tx_stats, cpu);
1259                 u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
1260                 unsigned int start;
1261
1262                 do {
1263                         start = u64_stats_fetch_begin_irq(&tx_stats->syncp);
1264                         tx_packets = tx_stats->packets;
1265                         tx_bytes = tx_stats->bytes;
1266                 } while (u64_stats_fetch_retry_irq(&tx_stats->syncp, start));
1267
1268                 do {
1269                         start = u64_stats_fetch_begin_irq(&rx_stats->syncp);
1270                         rx_packets = rx_stats->packets;
1271                         rx_bytes = rx_stats->bytes;
1272                 } while (u64_stats_fetch_retry_irq(&rx_stats->syncp, start));
1273
1274                 tot->rx_packets += rx_packets;
1275                 tot->tx_packets += tx_packets;
1276                 tot->rx_bytes   += rx_bytes;
1277                 tot->tx_bytes   += tx_bytes;
1278         }
1279
1280         tot->rx_errors  = dev->stats.rx_errors;
1281         tot->tx_dropped = dev->stats.tx_dropped;
1282 }
1283
1284 static void xennet_release_tx_bufs(struct netfront_queue *queue)
1285 {
1286         struct sk_buff *skb;
1287         int i;
1288
1289         for (i = 0; i < NET_TX_RING_SIZE; i++) {
1290                 /* Skip over entries which are actually freelist references */
1291                 if (skb_entry_is_link(&queue->tx_skbs[i]))
1292                         continue;
1293
1294                 skb = queue->tx_skbs[i].skb;
1295                 get_page(queue->grant_tx_page[i]);
1296                 gnttab_end_foreign_access(queue->grant_tx_ref[i],
1297                                           GNTMAP_readonly,
1298                                           (unsigned long)page_address(queue->grant_tx_page[i]));
1299                 queue->grant_tx_page[i] = NULL;
1300                 queue->grant_tx_ref[i] = GRANT_INVALID_REF;
1301                 add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, i);
1302                 dev_kfree_skb_irq(skb);
1303         }
1304 }
1305
1306 static void xennet_release_rx_bufs(struct netfront_queue *queue)
1307 {
1308         int id, ref;
1309
1310         spin_lock_bh(&queue->rx_lock);
1311
1312         for (id = 0; id < NET_RX_RING_SIZE; id++) {
1313                 struct sk_buff *skb;
1314                 struct page *page;
1315
1316                 skb = queue->rx_skbs[id];
1317                 if (!skb)
1318                         continue;
1319
1320                 ref = queue->grant_rx_ref[id];
1321                 if (ref == GRANT_INVALID_REF)
1322                         continue;
1323
1324                 page = skb_frag_page(&skb_shinfo(skb)->frags[0]);
1325
1326                 /* gnttab_end_foreign_access() needs a page ref until
1327                  * foreign access is ended (which may be deferred).
1328                  */
1329                 get_page(page);
1330                 gnttab_end_foreign_access(ref, 0,
1331                                           (unsigned long)page_address(page));
1332                 queue->grant_rx_ref[id] = GRANT_INVALID_REF;
1333
1334                 kfree_skb(skb);
1335         }
1336
1337         spin_unlock_bh(&queue->rx_lock);
1338 }
1339
1340 static netdev_features_t xennet_fix_features(struct net_device *dev,
1341         netdev_features_t features)
1342 {
1343         struct netfront_info *np = netdev_priv(dev);
1344
1345         if (features & NETIF_F_SG &&
1346             !xenbus_read_unsigned(np->xbdev->otherend, "feature-sg", 0))
1347                 features &= ~NETIF_F_SG;
1348
1349         if (features & NETIF_F_IPV6_CSUM &&
1350             !xenbus_read_unsigned(np->xbdev->otherend,
1351                                   "feature-ipv6-csum-offload", 0))
1352                 features &= ~NETIF_F_IPV6_CSUM;
1353
1354         if (features & NETIF_F_TSO &&
1355             !xenbus_read_unsigned(np->xbdev->otherend, "feature-gso-tcpv4", 0))
1356                 features &= ~NETIF_F_TSO;
1357
1358         if (features & NETIF_F_TSO6 &&
1359             !xenbus_read_unsigned(np->xbdev->otherend, "feature-gso-tcpv6", 0))
1360                 features &= ~NETIF_F_TSO6;
1361
1362         return features;
1363 }
1364
1365 static int xennet_set_features(struct net_device *dev,
1366         netdev_features_t features)
1367 {
1368         if (!(features & NETIF_F_SG) && dev->mtu > ETH_DATA_LEN) {
1369                 netdev_info(dev, "Reducing MTU because no SG offload");
1370                 dev->mtu = ETH_DATA_LEN;
1371         }
1372
1373         return 0;
1374 }
1375
1376 static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
1377 {
1378         struct netfront_queue *queue = dev_id;
1379         unsigned long flags;
1380
1381         spin_lock_irqsave(&queue->tx_lock, flags);
1382         xennet_tx_buf_gc(queue);
1383         spin_unlock_irqrestore(&queue->tx_lock, flags);
1384
1385         return IRQ_HANDLED;
1386 }
1387
1388 static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
1389 {
1390         struct netfront_queue *queue = dev_id;
1391         struct net_device *dev = queue->info->netdev;
1392
1393         if (likely(netif_carrier_ok(dev) &&
1394                    RING_HAS_UNCONSUMED_RESPONSES(&queue->rx)))
1395                 napi_schedule(&queue->napi);
1396
1397         return IRQ_HANDLED;
1398 }
1399
1400 static irqreturn_t xennet_interrupt(int irq, void *dev_id)
1401 {
1402         xennet_tx_interrupt(irq, dev_id);
1403         xennet_rx_interrupt(irq, dev_id);
1404         return IRQ_HANDLED;
1405 }
1406
1407 #ifdef CONFIG_NET_POLL_CONTROLLER
1408 static void xennet_poll_controller(struct net_device *dev)
1409 {
1410         /* Poll each queue */
1411         struct netfront_info *info = netdev_priv(dev);
1412         unsigned int num_queues = dev->real_num_tx_queues;
1413         unsigned int i;
1414         for (i = 0; i < num_queues; ++i)
1415                 xennet_interrupt(0, &info->queues[i]);
1416 }
1417 #endif
1418
1419 #define NETBACK_XDP_HEADROOM_DISABLE    0
1420 #define NETBACK_XDP_HEADROOM_ENABLE     1
1421
1422 static int talk_to_netback_xdp(struct netfront_info *np, int xdp)
1423 {
1424         int err;
1425         unsigned short headroom;
1426
1427         headroom = xdp ? XDP_PACKET_HEADROOM : 0;
1428         err = xenbus_printf(XBT_NIL, np->xbdev->nodename,
1429                             "xdp-headroom", "%hu",
1430                             headroom);
1431         if (err)
1432                 pr_warn("Error writing xdp-headroom\n");
1433
1434         return err;
1435 }
1436
1437 static int xennet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
1438                           struct netlink_ext_ack *extack)
1439 {
1440         unsigned long max_mtu = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM;
1441         struct netfront_info *np = netdev_priv(dev);
1442         struct bpf_prog *old_prog;
1443         unsigned int i, err;
1444
1445         if (dev->mtu > max_mtu) {
1446                 netdev_warn(dev, "XDP requires MTU less than %lu\n", max_mtu);
1447                 return -EINVAL;
1448         }
1449
1450         if (!np->netback_has_xdp_headroom)
1451                 return 0;
1452
1453         xenbus_switch_state(np->xbdev, XenbusStateReconfiguring);
1454
1455         err = talk_to_netback_xdp(np, prog ? NETBACK_XDP_HEADROOM_ENABLE :
1456                                   NETBACK_XDP_HEADROOM_DISABLE);
1457         if (err)
1458                 return err;
1459
1460         /* avoid the race with XDP headroom adjustment */
1461         wait_event(module_wq,
1462                    xenbus_read_driver_state(np->xbdev->otherend) ==
1463                    XenbusStateReconfigured);
1464         np->netfront_xdp_enabled = true;
1465
1466         old_prog = rtnl_dereference(np->queues[0].xdp_prog);
1467
1468         if (prog)
1469                 bpf_prog_add(prog, dev->real_num_tx_queues);
1470
1471         for (i = 0; i < dev->real_num_tx_queues; ++i)
1472                 rcu_assign_pointer(np->queues[i].xdp_prog, prog);
1473
1474         if (old_prog)
1475                 for (i = 0; i < dev->real_num_tx_queues; ++i)
1476                         bpf_prog_put(old_prog);
1477
1478         xenbus_switch_state(np->xbdev, XenbusStateConnected);
1479
1480         return 0;
1481 }
1482
1483 static int xennet_xdp(struct net_device *dev, struct netdev_bpf *xdp)
1484 {
1485         switch (xdp->command) {
1486         case XDP_SETUP_PROG:
1487                 return xennet_xdp_set(dev, xdp->prog, xdp->extack);
1488         default:
1489                 return -EINVAL;
1490         }
1491 }
1492
1493 static const struct net_device_ops xennet_netdev_ops = {
1494         .ndo_open            = xennet_open,
1495         .ndo_stop            = xennet_close,
1496         .ndo_start_xmit      = xennet_start_xmit,
1497         .ndo_change_mtu      = xennet_change_mtu,
1498         .ndo_get_stats64     = xennet_get_stats64,
1499         .ndo_set_mac_address = eth_mac_addr,
1500         .ndo_validate_addr   = eth_validate_addr,
1501         .ndo_fix_features    = xennet_fix_features,
1502         .ndo_set_features    = xennet_set_features,
1503         .ndo_select_queue    = xennet_select_queue,
1504         .ndo_bpf            = xennet_xdp,
1505         .ndo_xdp_xmit       = xennet_xdp_xmit,
1506 #ifdef CONFIG_NET_POLL_CONTROLLER
1507         .ndo_poll_controller = xennet_poll_controller,
1508 #endif
1509 };
1510
1511 static void xennet_free_netdev(struct net_device *netdev)
1512 {
1513         struct netfront_info *np = netdev_priv(netdev);
1514
1515         free_percpu(np->rx_stats);
1516         free_percpu(np->tx_stats);
1517         free_netdev(netdev);
1518 }
1519
1520 static struct net_device *xennet_create_dev(struct xenbus_device *dev)
1521 {
1522         int err;
1523         struct net_device *netdev;
1524         struct netfront_info *np;
1525
1526         netdev = alloc_etherdev_mq(sizeof(struct netfront_info), xennet_max_queues);
1527         if (!netdev)
1528                 return ERR_PTR(-ENOMEM);
1529
1530         np                   = netdev_priv(netdev);
1531         np->xbdev            = dev;
1532
1533         np->queues = NULL;
1534
1535         err = -ENOMEM;
1536         np->rx_stats = netdev_alloc_pcpu_stats(struct netfront_stats);
1537         if (np->rx_stats == NULL)
1538                 goto exit;
1539         np->tx_stats = netdev_alloc_pcpu_stats(struct netfront_stats);
1540         if (np->tx_stats == NULL)
1541                 goto exit;
1542
1543         netdev->netdev_ops      = &xennet_netdev_ops;
1544
1545         netdev->features        = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
1546                                   NETIF_F_GSO_ROBUST;
1547         netdev->hw_features     = NETIF_F_SG |
1548                                   NETIF_F_IPV6_CSUM |
1549                                   NETIF_F_TSO | NETIF_F_TSO6;
1550
1551         /*
1552          * Assume that all hw features are available for now. This set
1553          * will be adjusted by the call to netdev_update_features() in
1554          * xennet_connect() which is the earliest point where we can
1555          * negotiate with the backend regarding supported features.
1556          */
1557         netdev->features |= netdev->hw_features;
1558
1559         netdev->ethtool_ops = &xennet_ethtool_ops;
1560         netdev->min_mtu = ETH_MIN_MTU;
1561         netdev->max_mtu = XEN_NETIF_MAX_TX_SIZE;
1562         SET_NETDEV_DEV(netdev, &dev->dev);
1563
1564         np->netdev = netdev;
1565         np->netfront_xdp_enabled = false;
1566
1567         netif_carrier_off(netdev);
1568
1569         do {
1570                 xenbus_switch_state(dev, XenbusStateInitialising);
1571                 err = wait_event_timeout(module_wq,
1572                                  xenbus_read_driver_state(dev->otherend) !=
1573                                  XenbusStateClosed &&
1574                                  xenbus_read_driver_state(dev->otherend) !=
1575                                  XenbusStateUnknown, XENNET_TIMEOUT);
1576         } while (!err);
1577
1578         return netdev;
1579
1580  exit:
1581         xennet_free_netdev(netdev);
1582         return ERR_PTR(err);
1583 }
1584
1585 /**
1586  * Entry point to this code when a new device is created.  Allocate the basic
1587  * structures and the ring buffers for communication with the backend, and
1588  * inform the backend of the appropriate details for those.
1589  */
1590 static int netfront_probe(struct xenbus_device *dev,
1591                           const struct xenbus_device_id *id)
1592 {
1593         int err;
1594         struct net_device *netdev;
1595         struct netfront_info *info;
1596
1597         netdev = xennet_create_dev(dev);
1598         if (IS_ERR(netdev)) {
1599                 err = PTR_ERR(netdev);
1600                 xenbus_dev_fatal(dev, err, "creating netdev");
1601                 return err;
1602         }
1603
1604         info = netdev_priv(netdev);
1605         dev_set_drvdata(&dev->dev, info);
1606 #ifdef CONFIG_SYSFS
1607         info->netdev->sysfs_groups[0] = &xennet_dev_group;
1608 #endif
1609
1610         return 0;
1611 }
1612
1613 static void xennet_end_access(int ref, void *page)
1614 {
1615         /* This frees the page as a side-effect */
1616         if (ref != GRANT_INVALID_REF)
1617                 gnttab_end_foreign_access(ref, 0, (unsigned long)page);
1618 }
1619
1620 static void xennet_disconnect_backend(struct netfront_info *info)
1621 {
1622         unsigned int i = 0;
1623         unsigned int num_queues = info->netdev->real_num_tx_queues;
1624
1625         netif_carrier_off(info->netdev);
1626
1627         for (i = 0; i < num_queues && info->queues; ++i) {
1628                 struct netfront_queue *queue = &info->queues[i];
1629
1630                 del_timer_sync(&queue->rx_refill_timer);
1631
1632                 if (queue->tx_irq && (queue->tx_irq == queue->rx_irq))
1633                         unbind_from_irqhandler(queue->tx_irq, queue);
1634                 if (queue->tx_irq && (queue->tx_irq != queue->rx_irq)) {
1635                         unbind_from_irqhandler(queue->tx_irq, queue);
1636                         unbind_from_irqhandler(queue->rx_irq, queue);
1637                 }
1638                 queue->tx_evtchn = queue->rx_evtchn = 0;
1639                 queue->tx_irq = queue->rx_irq = 0;
1640
1641                 if (netif_running(info->netdev))
1642                         napi_synchronize(&queue->napi);
1643
1644                 xennet_release_tx_bufs(queue);
1645                 xennet_release_rx_bufs(queue);
1646                 gnttab_free_grant_references(queue->gref_tx_head);
1647                 gnttab_free_grant_references(queue->gref_rx_head);
1648
1649                 /* End access and free the pages */
1650                 xennet_end_access(queue->tx_ring_ref, queue->tx.sring);
1651                 xennet_end_access(queue->rx_ring_ref, queue->rx.sring);
1652
1653                 queue->tx_ring_ref = GRANT_INVALID_REF;
1654                 queue->rx_ring_ref = GRANT_INVALID_REF;
1655                 queue->tx.sring = NULL;
1656                 queue->rx.sring = NULL;
1657
1658                 page_pool_destroy(queue->page_pool);
1659         }
1660 }
1661
1662 /**
1663  * We are reconnecting to the backend, due to a suspend/resume, or a backend
1664  * driver restart.  We tear down our netif structure and recreate it, but
1665  * leave the device-layer structures intact so that this is transparent to the
1666  * rest of the kernel.
1667  */
1668 static int netfront_resume(struct xenbus_device *dev)
1669 {
1670         struct netfront_info *info = dev_get_drvdata(&dev->dev);
1671
1672         dev_dbg(&dev->dev, "%s\n", dev->nodename);
1673
1674         xennet_disconnect_backend(info);
1675         return 0;
1676 }
1677
1678 static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
1679 {
1680         char *s, *e, *macstr;
1681         int i;
1682
1683         macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
1684         if (IS_ERR(macstr))
1685                 return PTR_ERR(macstr);
1686
1687         for (i = 0; i < ETH_ALEN; i++) {
1688                 mac[i] = simple_strtoul(s, &e, 16);
1689                 if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
1690                         kfree(macstr);
1691                         return -ENOENT;
1692                 }
1693                 s = e+1;
1694         }
1695
1696         kfree(macstr);
1697         return 0;
1698 }
1699
1700 static int setup_netfront_single(struct netfront_queue *queue)
1701 {
1702         int err;
1703
1704         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->tx_evtchn);
1705         if (err < 0)
1706                 goto fail;
1707
1708         err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
1709                                         xennet_interrupt,
1710                                         0, queue->info->netdev->name, queue);
1711         if (err < 0)
1712                 goto bind_fail;
1713         queue->rx_evtchn = queue->tx_evtchn;
1714         queue->rx_irq = queue->tx_irq = err;
1715
1716         return 0;
1717
1718 bind_fail:
1719         xenbus_free_evtchn(queue->info->xbdev, queue->tx_evtchn);
1720         queue->tx_evtchn = 0;
1721 fail:
1722         return err;
1723 }
1724
1725 static int setup_netfront_split(struct netfront_queue *queue)
1726 {
1727         int err;
1728
1729         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->tx_evtchn);
1730         if (err < 0)
1731                 goto fail;
1732         err = xenbus_alloc_evtchn(queue->info->xbdev, &queue->rx_evtchn);
1733         if (err < 0)
1734                 goto alloc_rx_evtchn_fail;
1735
1736         snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name),
1737                  "%s-tx", queue->name);
1738         err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
1739                                         xennet_tx_interrupt,
1740                                         0, queue->tx_irq_name, queue);
1741         if (err < 0)
1742                 goto bind_tx_fail;
1743         queue->tx_irq = err;
1744
1745         snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name),
1746                  "%s-rx", queue->name);
1747         err = bind_evtchn_to_irqhandler(queue->rx_evtchn,
1748                                         xennet_rx_interrupt,
1749                                         0, queue->rx_irq_name, queue);
1750         if (err < 0)
1751                 goto bind_rx_fail;
1752         queue->rx_irq = err;
1753
1754         return 0;
1755
1756 bind_rx_fail:
1757         unbind_from_irqhandler(queue->tx_irq, queue);
1758         queue->tx_irq = 0;
1759 bind_tx_fail:
1760         xenbus_free_evtchn(queue->info->xbdev, queue->rx_evtchn);
1761         queue->rx_evtchn = 0;
1762 alloc_rx_evtchn_fail:
1763         xenbus_free_evtchn(queue->info->xbdev, queue->tx_evtchn);
1764         queue->tx_evtchn = 0;
1765 fail:
1766         return err;
1767 }
1768
1769 static int setup_netfront(struct xenbus_device *dev,
1770                         struct netfront_queue *queue, unsigned int feature_split_evtchn)
1771 {
1772         struct xen_netif_tx_sring *txs;
1773         struct xen_netif_rx_sring *rxs;
1774         grant_ref_t gref;
1775         int err;
1776
1777         queue->tx_ring_ref = GRANT_INVALID_REF;
1778         queue->rx_ring_ref = GRANT_INVALID_REF;
1779         queue->rx.sring = NULL;
1780         queue->tx.sring = NULL;
1781
1782         txs = (struct xen_netif_tx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1783         if (!txs) {
1784                 err = -ENOMEM;
1785                 xenbus_dev_fatal(dev, err, "allocating tx ring page");
1786                 goto fail;
1787         }
1788         SHARED_RING_INIT(txs);
1789         FRONT_RING_INIT(&queue->tx, txs, XEN_PAGE_SIZE);
1790
1791         err = xenbus_grant_ring(dev, txs, 1, &gref);
1792         if (err < 0)
1793                 goto grant_tx_ring_fail;
1794         queue->tx_ring_ref = gref;
1795
1796         rxs = (struct xen_netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1797         if (!rxs) {
1798                 err = -ENOMEM;
1799                 xenbus_dev_fatal(dev, err, "allocating rx ring page");
1800                 goto alloc_rx_ring_fail;
1801         }
1802         SHARED_RING_INIT(rxs);
1803         FRONT_RING_INIT(&queue->rx, rxs, XEN_PAGE_SIZE);
1804
1805         err = xenbus_grant_ring(dev, rxs, 1, &gref);
1806         if (err < 0)
1807                 goto grant_rx_ring_fail;
1808         queue->rx_ring_ref = gref;
1809
1810         if (feature_split_evtchn)
1811                 err = setup_netfront_split(queue);
1812         /* setup single event channel if
1813          *  a) feature-split-event-channels == 0
1814          *  b) feature-split-event-channels == 1 but failed to setup
1815          */
1816         if (!feature_split_evtchn || (feature_split_evtchn && err))
1817                 err = setup_netfront_single(queue);
1818
1819         if (err)
1820                 goto alloc_evtchn_fail;
1821
1822         return 0;
1823
1824         /* If we fail to setup netfront, it is safe to just revoke access to
1825          * granted pages because backend is not accessing it at this point.
1826          */
1827 alloc_evtchn_fail:
1828         gnttab_end_foreign_access_ref(queue->rx_ring_ref, 0);
1829 grant_rx_ring_fail:
1830         free_page((unsigned long)rxs);
1831 alloc_rx_ring_fail:
1832         gnttab_end_foreign_access_ref(queue->tx_ring_ref, 0);
1833 grant_tx_ring_fail:
1834         free_page((unsigned long)txs);
1835 fail:
1836         return err;
1837 }
1838
1839 /* Queue-specific initialisation
1840  * This used to be done in xennet_create_dev() but must now
1841  * be run per-queue.
1842  */
1843 static int xennet_init_queue(struct netfront_queue *queue)
1844 {
1845         unsigned short i;
1846         int err = 0;
1847         char *devid;
1848
1849         spin_lock_init(&queue->tx_lock);
1850         spin_lock_init(&queue->rx_lock);
1851
1852         timer_setup(&queue->rx_refill_timer, rx_refill_timeout, 0);
1853
1854         devid = strrchr(queue->info->xbdev->nodename, '/') + 1;
1855         snprintf(queue->name, sizeof(queue->name), "vif%s-q%u",
1856                  devid, queue->id);
1857
1858         /* Initialise tx_skbs as a free chain containing every entry. */
1859         queue->tx_skb_freelist = 0;
1860         for (i = 0; i < NET_TX_RING_SIZE; i++) {
1861                 skb_entry_set_link(&queue->tx_skbs[i], i+1);
1862                 queue->grant_tx_ref[i] = GRANT_INVALID_REF;
1863                 queue->grant_tx_page[i] = NULL;
1864         }
1865
1866         /* Clear out rx_skbs */
1867         for (i = 0; i < NET_RX_RING_SIZE; i++) {
1868                 queue->rx_skbs[i] = NULL;
1869                 queue->grant_rx_ref[i] = GRANT_INVALID_REF;
1870         }
1871
1872         /* A grant for every tx ring slot */
1873         if (gnttab_alloc_grant_references(NET_TX_RING_SIZE,
1874                                           &queue->gref_tx_head) < 0) {
1875                 pr_alert("can't alloc tx grant refs\n");
1876                 err = -ENOMEM;
1877                 goto exit;
1878         }
1879
1880         /* A grant for every rx ring slot */
1881         if (gnttab_alloc_grant_references(NET_RX_RING_SIZE,
1882                                           &queue->gref_rx_head) < 0) {
1883                 pr_alert("can't alloc rx grant refs\n");
1884                 err = -ENOMEM;
1885                 goto exit_free_tx;
1886         }
1887
1888         return 0;
1889
1890  exit_free_tx:
1891         gnttab_free_grant_references(queue->gref_tx_head);
1892  exit:
1893         return err;
1894 }
1895
1896 static int write_queue_xenstore_keys(struct netfront_queue *queue,
1897                            struct xenbus_transaction *xbt, int write_hierarchical)
1898 {
1899         /* Write the queue-specific keys into XenStore in the traditional
1900          * way for a single queue, or in a queue subkeys for multiple
1901          * queues.
1902          */
1903         struct xenbus_device *dev = queue->info->xbdev;
1904         int err;
1905         const char *message;
1906         char *path;
1907         size_t pathsize;
1908
1909         /* Choose the correct place to write the keys */
1910         if (write_hierarchical) {
1911                 pathsize = strlen(dev->nodename) + 10;
1912                 path = kzalloc(pathsize, GFP_KERNEL);
1913                 if (!path) {
1914                         err = -ENOMEM;
1915                         message = "out of memory while writing ring references";
1916                         goto error;
1917                 }
1918                 snprintf(path, pathsize, "%s/queue-%u",
1919                                 dev->nodename, queue->id);
1920         } else {
1921                 path = (char *)dev->nodename;
1922         }
1923
1924         /* Write ring references */
1925         err = xenbus_printf(*xbt, path, "tx-ring-ref", "%u",
1926                         queue->tx_ring_ref);
1927         if (err) {
1928                 message = "writing tx-ring-ref";
1929                 goto error;
1930         }
1931
1932         err = xenbus_printf(*xbt, path, "rx-ring-ref", "%u",
1933                         queue->rx_ring_ref);
1934         if (err) {
1935                 message = "writing rx-ring-ref";
1936                 goto error;
1937         }
1938
1939         /* Write event channels; taking into account both shared
1940          * and split event channel scenarios.
1941          */
1942         if (queue->tx_evtchn == queue->rx_evtchn) {
1943                 /* Shared event channel */
1944                 err = xenbus_printf(*xbt, path,
1945                                 "event-channel", "%u", queue->tx_evtchn);
1946                 if (err) {
1947                         message = "writing event-channel";
1948                         goto error;
1949                 }
1950         } else {
1951                 /* Split event channels */
1952                 err = xenbus_printf(*xbt, path,
1953                                 "event-channel-tx", "%u", queue->tx_evtchn);
1954                 if (err) {
1955                         message = "writing event-channel-tx";
1956                         goto error;
1957                 }
1958
1959                 err = xenbus_printf(*xbt, path,
1960                                 "event-channel-rx", "%u", queue->rx_evtchn);
1961                 if (err) {
1962                         message = "writing event-channel-rx";
1963                         goto error;
1964                 }
1965         }
1966
1967         if (write_hierarchical)
1968                 kfree(path);
1969         return 0;
1970
1971 error:
1972         if (write_hierarchical)
1973                 kfree(path);
1974         xenbus_dev_fatal(dev, err, "%s", message);
1975         return err;
1976 }
1977
1978 static void xennet_destroy_queues(struct netfront_info *info)
1979 {
1980         unsigned int i;
1981
1982         for (i = 0; i < info->netdev->real_num_tx_queues; i++) {
1983                 struct netfront_queue *queue = &info->queues[i];
1984
1985                 if (netif_running(info->netdev))
1986                         napi_disable(&queue->napi);
1987                 netif_napi_del(&queue->napi);
1988         }
1989
1990         kfree(info->queues);
1991         info->queues = NULL;
1992 }
1993
1994
1995
1996 static int xennet_create_page_pool(struct netfront_queue *queue)
1997 {
1998         int err;
1999         struct page_pool_params pp_params = {
2000                 .order = 0,
2001                 .flags = 0,
2002                 .pool_size = NET_RX_RING_SIZE,
2003                 .nid = NUMA_NO_NODE,
2004                 .dev = &queue->info->netdev->dev,
2005                 .offset = XDP_PACKET_HEADROOM,
2006                 .max_len = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM,
2007         };
2008
2009         queue->page_pool = page_pool_create(&pp_params);
2010         if (IS_ERR(queue->page_pool)) {
2011                 err = PTR_ERR(queue->page_pool);
2012                 queue->page_pool = NULL;
2013                 return err;
2014         }
2015
2016         err = xdp_rxq_info_reg(&queue->xdp_rxq, queue->info->netdev,
2017                                queue->id);
2018         if (err) {
2019                 netdev_err(queue->info->netdev, "xdp_rxq_info_reg failed\n");
2020                 goto err_free_pp;
2021         }
2022
2023         err = xdp_rxq_info_reg_mem_model(&queue->xdp_rxq,
2024                                          MEM_TYPE_PAGE_POOL, queue->page_pool);
2025         if (err) {
2026                 netdev_err(queue->info->netdev, "xdp_rxq_info_reg_mem_model failed\n");
2027                 goto err_unregister_rxq;
2028         }
2029         return 0;
2030
2031 err_unregister_rxq:
2032         xdp_rxq_info_unreg(&queue->xdp_rxq);
2033 err_free_pp:
2034         page_pool_destroy(queue->page_pool);
2035         queue->page_pool = NULL;
2036         return err;
2037 }
2038
2039 static int xennet_create_queues(struct netfront_info *info,
2040                                 unsigned int *num_queues)
2041 {
2042         unsigned int i;
2043         int ret;
2044
2045         info->queues = kcalloc(*num_queues, sizeof(struct netfront_queue),
2046                                GFP_KERNEL);
2047         if (!info->queues)
2048                 return -ENOMEM;
2049
2050         for (i = 0; i < *num_queues; i++) {
2051                 struct netfront_queue *queue = &info->queues[i];
2052
2053                 queue->id = i;
2054                 queue->info = info;
2055
2056                 ret = xennet_init_queue(queue);
2057                 if (ret < 0) {
2058                         dev_warn(&info->xbdev->dev,
2059                                  "only created %d queues\n", i);
2060                         *num_queues = i;
2061                         break;
2062                 }
2063
2064                 /* use page pool recycling instead of buddy allocator */
2065                 ret = xennet_create_page_pool(queue);
2066                 if (ret < 0) {
2067                         dev_err(&info->xbdev->dev, "can't allocate page pool\n");
2068                         *num_queues = i;
2069                         return ret;
2070                 }
2071
2072                 netif_napi_add(queue->info->netdev, &queue->napi,
2073                                xennet_poll, 64);
2074                 if (netif_running(info->netdev))
2075                         napi_enable(&queue->napi);
2076         }
2077
2078         netif_set_real_num_tx_queues(info->netdev, *num_queues);
2079
2080         if (*num_queues == 0) {
2081                 dev_err(&info->xbdev->dev, "no queues\n");
2082                 return -EINVAL;
2083         }
2084         return 0;
2085 }
2086
2087 /* Common code used when first setting up, and when resuming. */
2088 static int talk_to_netback(struct xenbus_device *dev,
2089                            struct netfront_info *info)
2090 {
2091         const char *message;
2092         struct xenbus_transaction xbt;
2093         int err;
2094         unsigned int feature_split_evtchn;
2095         unsigned int i = 0;
2096         unsigned int max_queues = 0;
2097         struct netfront_queue *queue = NULL;
2098         unsigned int num_queues = 1;
2099
2100         info->netdev->irq = 0;
2101
2102         /* Check if backend supports multiple queues */
2103         max_queues = xenbus_read_unsigned(info->xbdev->otherend,
2104                                           "multi-queue-max-queues", 1);
2105         num_queues = min(max_queues, xennet_max_queues);
2106
2107         /* Check feature-split-event-channels */
2108         feature_split_evtchn = xenbus_read_unsigned(info->xbdev->otherend,
2109                                         "feature-split-event-channels", 0);
2110
2111         /* Read mac addr. */
2112         err = xen_net_read_mac(dev, info->netdev->dev_addr);
2113         if (err) {
2114                 xenbus_dev_fatal(dev, err, "parsing %s/mac", dev->nodename);
2115                 goto out_unlocked;
2116         }
2117
2118         info->netback_has_xdp_headroom = xenbus_read_unsigned(info->xbdev->otherend,
2119                                                               "feature-xdp-headroom", 0);
2120         if (info->netback_has_xdp_headroom) {
2121                 /* set the current xen-netfront xdp state */
2122                 err = talk_to_netback_xdp(info, info->netfront_xdp_enabled ?
2123                                           NETBACK_XDP_HEADROOM_ENABLE :
2124                                           NETBACK_XDP_HEADROOM_DISABLE);
2125                 if (err)
2126                         goto out_unlocked;
2127         }
2128
2129         rtnl_lock();
2130         if (info->queues)
2131                 xennet_destroy_queues(info);
2132
2133         err = xennet_create_queues(info, &num_queues);
2134         if (err < 0) {
2135                 xenbus_dev_fatal(dev, err, "creating queues");
2136                 kfree(info->queues);
2137                 info->queues = NULL;
2138                 goto out;
2139         }
2140         rtnl_unlock();
2141
2142         /* Create shared ring, alloc event channel -- for each queue */
2143         for (i = 0; i < num_queues; ++i) {
2144                 queue = &info->queues[i];
2145                 err = setup_netfront(dev, queue, feature_split_evtchn);
2146                 if (err)
2147                         goto destroy_ring;
2148         }
2149
2150 again:
2151         err = xenbus_transaction_start(&xbt);
2152         if (err) {
2153                 xenbus_dev_fatal(dev, err, "starting transaction");
2154                 goto destroy_ring;
2155         }
2156
2157         if (xenbus_exists(XBT_NIL,
2158                           info->xbdev->otherend, "multi-queue-max-queues")) {
2159                 /* Write the number of queues */
2160                 err = xenbus_printf(xbt, dev->nodename,
2161                                     "multi-queue-num-queues", "%u", num_queues);
2162                 if (err) {
2163                         message = "writing multi-queue-num-queues";
2164                         goto abort_transaction_no_dev_fatal;
2165                 }
2166         }
2167
2168         if (num_queues == 1) {
2169                 err = write_queue_xenstore_keys(&info->queues[0], &xbt, 0); /* flat */
2170                 if (err)
2171                         goto abort_transaction_no_dev_fatal;
2172         } else {
2173                 /* Write the keys for each queue */
2174                 for (i = 0; i < num_queues; ++i) {
2175                         queue = &info->queues[i];
2176                         err = write_queue_xenstore_keys(queue, &xbt, 1); /* hierarchical */
2177                         if (err)
2178                                 goto abort_transaction_no_dev_fatal;
2179                 }
2180         }
2181
2182         /* The remaining keys are not queue-specific */
2183         err = xenbus_printf(xbt, dev->nodename, "request-rx-copy", "%u",
2184                             1);
2185         if (err) {
2186                 message = "writing request-rx-copy";
2187                 goto abort_transaction;
2188         }
2189
2190         err = xenbus_printf(xbt, dev->nodename, "feature-rx-notify", "%d", 1);
2191         if (err) {
2192                 message = "writing feature-rx-notify";
2193                 goto abort_transaction;
2194         }
2195
2196         err = xenbus_printf(xbt, dev->nodename, "feature-sg", "%d", 1);
2197         if (err) {
2198                 message = "writing feature-sg";
2199                 goto abort_transaction;
2200         }
2201
2202         err = xenbus_printf(xbt, dev->nodename, "feature-gso-tcpv4", "%d", 1);
2203         if (err) {
2204                 message = "writing feature-gso-tcpv4";
2205                 goto abort_transaction;
2206         }
2207
2208         err = xenbus_write(xbt, dev->nodename, "feature-gso-tcpv6", "1");
2209         if (err) {
2210                 message = "writing feature-gso-tcpv6";
2211                 goto abort_transaction;
2212         }
2213
2214         err = xenbus_write(xbt, dev->nodename, "feature-ipv6-csum-offload",
2215                            "1");
2216         if (err) {
2217                 message = "writing feature-ipv6-csum-offload";
2218                 goto abort_transaction;
2219         }
2220
2221         err = xenbus_transaction_end(xbt, 0);
2222         if (err) {
2223                 if (err == -EAGAIN)
2224                         goto again;
2225                 xenbus_dev_fatal(dev, err, "completing transaction");
2226                 goto destroy_ring;
2227         }
2228
2229         return 0;
2230
2231  abort_transaction:
2232         xenbus_dev_fatal(dev, err, "%s", message);
2233 abort_transaction_no_dev_fatal:
2234         xenbus_transaction_end(xbt, 1);
2235  destroy_ring:
2236         xennet_disconnect_backend(info);
2237         rtnl_lock();
2238         xennet_destroy_queues(info);
2239  out:
2240         rtnl_unlock();
2241 out_unlocked:
2242         device_unregister(&dev->dev);
2243         return err;
2244 }
2245
2246 static int xennet_connect(struct net_device *dev)
2247 {
2248         struct netfront_info *np = netdev_priv(dev);
2249         unsigned int num_queues = 0;
2250         int err;
2251         unsigned int j = 0;
2252         struct netfront_queue *queue = NULL;
2253
2254         if (!xenbus_read_unsigned(np->xbdev->otherend, "feature-rx-copy", 0)) {
2255                 dev_info(&dev->dev,
2256                          "backend does not support copying receive path\n");
2257                 return -ENODEV;
2258         }
2259
2260         err = talk_to_netback(np->xbdev, np);
2261         if (err)
2262                 return err;
2263         if (np->netback_has_xdp_headroom)
2264                 pr_info("backend supports XDP headroom\n");
2265
2266         /* talk_to_netback() sets the correct number of queues */
2267         num_queues = dev->real_num_tx_queues;
2268
2269         if (dev->reg_state == NETREG_UNINITIALIZED) {
2270                 err = register_netdev(dev);
2271                 if (err) {
2272                         pr_warn("%s: register_netdev err=%d\n", __func__, err);
2273                         device_unregister(&np->xbdev->dev);
2274                         return err;
2275                 }
2276         }
2277
2278         rtnl_lock();
2279         netdev_update_features(dev);
2280         rtnl_unlock();
2281
2282         /*
2283          * All public and private state should now be sane.  Get
2284          * ready to start sending and receiving packets and give the driver
2285          * domain a kick because we've probably just requeued some
2286          * packets.
2287          */
2288         netif_carrier_on(np->netdev);
2289         for (j = 0; j < num_queues; ++j) {
2290                 queue = &np->queues[j];
2291
2292                 notify_remote_via_irq(queue->tx_irq);
2293                 if (queue->tx_irq != queue->rx_irq)
2294                         notify_remote_via_irq(queue->rx_irq);
2295
2296                 spin_lock_irq(&queue->tx_lock);
2297                 xennet_tx_buf_gc(queue);
2298                 spin_unlock_irq(&queue->tx_lock);
2299
2300                 spin_lock_bh(&queue->rx_lock);
2301                 xennet_alloc_rx_buffers(queue);
2302                 spin_unlock_bh(&queue->rx_lock);
2303         }
2304
2305         return 0;
2306 }
2307
2308 /**
2309  * Callback received when the backend's state changes.
2310  */
2311 static void netback_changed(struct xenbus_device *dev,
2312                             enum xenbus_state backend_state)
2313 {
2314         struct netfront_info *np = dev_get_drvdata(&dev->dev);
2315         struct net_device *netdev = np->netdev;
2316
2317         dev_dbg(&dev->dev, "%s\n", xenbus_strstate(backend_state));
2318
2319         wake_up_all(&module_wq);
2320
2321         switch (backend_state) {
2322         case XenbusStateInitialising:
2323         case XenbusStateInitialised:
2324         case XenbusStateReconfiguring:
2325         case XenbusStateReconfigured:
2326         case XenbusStateUnknown:
2327                 break;
2328
2329         case XenbusStateInitWait:
2330                 if (dev->state != XenbusStateInitialising)
2331                         break;
2332                 if (xennet_connect(netdev) != 0)
2333                         break;
2334                 xenbus_switch_state(dev, XenbusStateConnected);
2335                 break;
2336
2337         case XenbusStateConnected:
2338                 netdev_notify_peers(netdev);
2339                 break;
2340
2341         case XenbusStateClosed:
2342                 if (dev->state == XenbusStateClosed)
2343                         break;
2344                 fallthrough;    /* Missed the backend's CLOSING state */
2345         case XenbusStateClosing:
2346                 xenbus_frontend_closed(dev);
2347                 break;
2348         }
2349 }
2350
2351 static const struct xennet_stat {
2352         char name[ETH_GSTRING_LEN];
2353         u16 offset;
2354 } xennet_stats[] = {
2355         {
2356                 "rx_gso_checksum_fixup",
2357                 offsetof(struct netfront_info, rx_gso_checksum_fixup)
2358         },
2359 };
2360
2361 static int xennet_get_sset_count(struct net_device *dev, int string_set)
2362 {
2363         switch (string_set) {
2364         case ETH_SS_STATS:
2365                 return ARRAY_SIZE(xennet_stats);
2366         default:
2367                 return -EINVAL;
2368         }
2369 }
2370
2371 static void xennet_get_ethtool_stats(struct net_device *dev,
2372                                      struct ethtool_stats *stats, u64 * data)
2373 {
2374         void *np = netdev_priv(dev);
2375         int i;
2376
2377         for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
2378                 data[i] = atomic_read((atomic_t *)(np + xennet_stats[i].offset));
2379 }
2380
2381 static void xennet_get_strings(struct net_device *dev, u32 stringset, u8 * data)
2382 {
2383         int i;
2384
2385         switch (stringset) {
2386         case ETH_SS_STATS:
2387                 for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
2388                         memcpy(data + i * ETH_GSTRING_LEN,
2389                                xennet_stats[i].name, ETH_GSTRING_LEN);
2390                 break;
2391         }
2392 }
2393
2394 static const struct ethtool_ops xennet_ethtool_ops =
2395 {
2396         .get_link = ethtool_op_get_link,
2397
2398         .get_sset_count = xennet_get_sset_count,
2399         .get_ethtool_stats = xennet_get_ethtool_stats,
2400         .get_strings = xennet_get_strings,
2401         .get_ts_info = ethtool_op_get_ts_info,
2402 };
2403
2404 #ifdef CONFIG_SYSFS
2405 static ssize_t show_rxbuf(struct device *dev,
2406                           struct device_attribute *attr, char *buf)
2407 {
2408         return sprintf(buf, "%lu\n", NET_RX_RING_SIZE);
2409 }
2410
2411 static ssize_t store_rxbuf(struct device *dev,
2412                            struct device_attribute *attr,
2413                            const char *buf, size_t len)
2414 {
2415         char *endp;
2416         unsigned long target;
2417
2418         if (!capable(CAP_NET_ADMIN))
2419                 return -EPERM;
2420
2421         target = simple_strtoul(buf, &endp, 0);
2422         if (endp == buf)
2423                 return -EBADMSG;
2424
2425         /* rxbuf_min and rxbuf_max are no longer configurable. */
2426
2427         return len;
2428 }
2429
2430 static DEVICE_ATTR(rxbuf_min, 0644, show_rxbuf, store_rxbuf);
2431 static DEVICE_ATTR(rxbuf_max, 0644, show_rxbuf, store_rxbuf);
2432 static DEVICE_ATTR(rxbuf_cur, 0444, show_rxbuf, NULL);
2433
2434 static struct attribute *xennet_dev_attrs[] = {
2435         &dev_attr_rxbuf_min.attr,
2436         &dev_attr_rxbuf_max.attr,
2437         &dev_attr_rxbuf_cur.attr,
2438         NULL
2439 };
2440
2441 static const struct attribute_group xennet_dev_group = {
2442         .attrs = xennet_dev_attrs
2443 };
2444 #endif /* CONFIG_SYSFS */
2445
2446 static void xennet_bus_close(struct xenbus_device *dev)
2447 {
2448         int ret;
2449
2450         if (xenbus_read_driver_state(dev->otherend) == XenbusStateClosed)
2451                 return;
2452         do {
2453                 xenbus_switch_state(dev, XenbusStateClosing);
2454                 ret = wait_event_timeout(module_wq,
2455                                    xenbus_read_driver_state(dev->otherend) ==
2456                                    XenbusStateClosing ||
2457                                    xenbus_read_driver_state(dev->otherend) ==
2458                                    XenbusStateClosed ||
2459                                    xenbus_read_driver_state(dev->otherend) ==
2460                                    XenbusStateUnknown,
2461                                    XENNET_TIMEOUT);
2462         } while (!ret);
2463
2464         if (xenbus_read_driver_state(dev->otherend) == XenbusStateClosed)
2465                 return;
2466
2467         do {
2468                 xenbus_switch_state(dev, XenbusStateClosed);
2469                 ret = wait_event_timeout(module_wq,
2470                                    xenbus_read_driver_state(dev->otherend) ==
2471                                    XenbusStateClosed ||
2472                                    xenbus_read_driver_state(dev->otherend) ==
2473                                    XenbusStateUnknown,
2474                                    XENNET_TIMEOUT);
2475         } while (!ret);
2476 }
2477
2478 static int xennet_remove(struct xenbus_device *dev)
2479 {
2480         struct netfront_info *info = dev_get_drvdata(&dev->dev);
2481
2482         xennet_bus_close(dev);
2483         xennet_disconnect_backend(info);
2484
2485         if (info->netdev->reg_state == NETREG_REGISTERED)
2486                 unregister_netdev(info->netdev);
2487
2488         if (info->queues) {
2489                 rtnl_lock();
2490                 xennet_destroy_queues(info);
2491                 rtnl_unlock();
2492         }
2493         xennet_free_netdev(info->netdev);
2494
2495         return 0;
2496 }
2497
2498 static const struct xenbus_device_id netfront_ids[] = {
2499         { "vif" },
2500         { "" }
2501 };
2502
2503 static struct xenbus_driver netfront_driver = {
2504         .ids = netfront_ids,
2505         .probe = netfront_probe,
2506         .remove = xennet_remove,
2507         .resume = netfront_resume,
2508         .otherend_changed = netback_changed,
2509 };
2510
2511 static int __init netif_init(void)
2512 {
2513         if (!xen_domain())
2514                 return -ENODEV;
2515
2516         if (!xen_has_pv_nic_devices())
2517                 return -ENODEV;
2518
2519         pr_info("Initialising Xen virtual ethernet driver\n");
2520
2521         /* Allow as many queues as there are CPUs inut max. 8 if user has not
2522          * specified a value.
2523          */
2524         if (xennet_max_queues == 0)
2525                 xennet_max_queues = min_t(unsigned int, MAX_QUEUES_DEFAULT,
2526                                           num_online_cpus());
2527
2528         return xenbus_register_frontend(&netfront_driver);
2529 }
2530 module_init(netif_init);
2531
2532
2533 static void __exit netif_exit(void)
2534 {
2535         xenbus_unregister_driver(&netfront_driver);
2536 }
2537 module_exit(netif_exit);
2538
2539 MODULE_DESCRIPTION("Xen virtual network device frontend");
2540 MODULE_LICENSE("GPL");
2541 MODULE_ALIAS("xen:vif");
2542 MODULE_ALIAS("xennet");