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