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