2 * Virtual network driver for conversing with remote driver backends.
4 * Copyright (c) 2002-2005, K A Fraser
5 * Copyright (c) 2005, XenSource Ltd
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:
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:
20 * The above copyright notice and this permission notice shall be included in
21 * all copies or substantial portions of the Software.
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
32 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
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>
42 #include <linux/udp.h>
43 #include <linux/moduleparam.h>
45 #include <linux/slab.h>
48 #include <asm/xen/page.h>
50 #include <xen/xenbus.h>
51 #include <xen/events.h>
53 #include <xen/platform_pci.h>
54 #include <xen/grant_table.h>
56 #include <xen/interface/io/netif.h>
57 #include <xen/interface/memory.h>
58 #include <xen/interface/grant_table.h>
60 static const struct ethtool_ops xennet_ethtool_ops;
66 #define NETFRONT_SKB_CB(skb) ((struct netfront_cb *)((skb)->cb))
68 #define RX_COPY_THRESHOLD 256
70 #define GRANT_INVALID_REF 0
72 #define NET_TX_RING_SIZE __CONST_RING_SIZE(xen_netif_tx, PAGE_SIZE)
73 #define NET_RX_RING_SIZE __CONST_RING_SIZE(xen_netif_rx, PAGE_SIZE)
74 #define TX_MAX_TARGET min_t(int, NET_TX_RING_SIZE, 256)
76 struct netfront_stats {
81 struct u64_stats_sync syncp;
84 struct netfront_info {
85 struct list_head list;
86 struct net_device *netdev;
88 struct napi_struct napi;
90 /* Split event channels support, tx_* == rx_* when using
91 * single event channel.
93 unsigned int tx_evtchn, rx_evtchn;
94 unsigned int tx_irq, rx_irq;
95 /* Only used when split event channels support is enabled */
96 char tx_irq_name[IFNAMSIZ+4]; /* DEVNAME-tx */
97 char rx_irq_name[IFNAMSIZ+4]; /* DEVNAME-rx */
99 struct xenbus_device *xbdev;
102 struct xen_netif_tx_front_ring tx;
106 * {tx,rx}_skbs store outstanding skbuffs. Free tx_skb entries
107 * are linked from tx_skb_freelist through skb_entry.link.
109 * NB. Freelist index entries are always going to be less than
110 * PAGE_OFFSET, whereas pointers to skbs will always be equal or
111 * greater than PAGE_OFFSET: we use this property to distinguish
117 } tx_skbs[NET_TX_RING_SIZE];
118 grant_ref_t gref_tx_head;
119 grant_ref_t grant_tx_ref[NET_TX_RING_SIZE];
120 unsigned tx_skb_freelist;
122 spinlock_t rx_lock ____cacheline_aligned_in_smp;
123 struct xen_netif_rx_front_ring rx;
126 /* Receive-ring batched refills. */
127 #define RX_MIN_TARGET 8
128 #define RX_DFL_MIN_TARGET 64
129 #define RX_MAX_TARGET min_t(int, NET_RX_RING_SIZE, 256)
130 unsigned rx_min_target, rx_max_target, rx_target;
131 struct sk_buff_head rx_batch;
133 struct timer_list rx_refill_timer;
135 struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
136 grant_ref_t gref_rx_head;
137 grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];
139 unsigned long rx_pfn_array[NET_RX_RING_SIZE];
140 struct multicall_entry rx_mcl[NET_RX_RING_SIZE+1];
141 struct mmu_update rx_mmu[NET_RX_RING_SIZE];
144 struct netfront_stats __percpu *stats;
146 unsigned long rx_gso_checksum_fixup;
149 struct netfront_rx_info {
150 struct xen_netif_rx_response rx;
151 struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
154 static void skb_entry_set_link(union skb_entry *list, unsigned short id)
159 static int skb_entry_is_link(const union skb_entry *list)
161 BUILD_BUG_ON(sizeof(list->skb) != sizeof(list->link));
162 return (unsigned long)list->skb < PAGE_OFFSET;
166 * Access macros for acquiring freeing slots in tx_skbs[].
169 static void add_id_to_freelist(unsigned *head, union skb_entry *list,
172 skb_entry_set_link(&list[id], *head);
176 static unsigned short get_id_from_freelist(unsigned *head,
177 union skb_entry *list)
179 unsigned int id = *head;
180 *head = list[id].link;
184 static int xennet_rxidx(RING_IDX idx)
186 return idx & (NET_RX_RING_SIZE - 1);
189 static struct sk_buff *xennet_get_rx_skb(struct netfront_info *np,
192 int i = xennet_rxidx(ri);
193 struct sk_buff *skb = np->rx_skbs[i];
194 np->rx_skbs[i] = NULL;
198 static grant_ref_t xennet_get_rx_ref(struct netfront_info *np,
201 int i = xennet_rxidx(ri);
202 grant_ref_t ref = np->grant_rx_ref[i];
203 np->grant_rx_ref[i] = GRANT_INVALID_REF;
208 static int xennet_sysfs_addif(struct net_device *netdev);
209 static void xennet_sysfs_delif(struct net_device *netdev);
210 #else /* !CONFIG_SYSFS */
211 #define xennet_sysfs_addif(dev) (0)
212 #define xennet_sysfs_delif(dev) do { } while (0)
215 static bool xennet_can_sg(struct net_device *dev)
217 return dev->features & NETIF_F_SG;
221 static void rx_refill_timeout(unsigned long data)
223 struct net_device *dev = (struct net_device *)data;
224 struct netfront_info *np = netdev_priv(dev);
225 napi_schedule(&np->napi);
228 static int netfront_tx_slot_available(struct netfront_info *np)
230 return (np->tx.req_prod_pvt - np->tx.rsp_cons) <
231 (TX_MAX_TARGET - MAX_SKB_FRAGS - 2);
234 static void xennet_maybe_wake_tx(struct net_device *dev)
236 struct netfront_info *np = netdev_priv(dev);
238 if (unlikely(netif_queue_stopped(dev)) &&
239 netfront_tx_slot_available(np) &&
240 likely(netif_running(dev)))
241 netif_wake_queue(dev);
244 static void xennet_alloc_rx_buffers(struct net_device *dev)
247 struct netfront_info *np = netdev_priv(dev);
250 int i, batch_target, notify;
251 RING_IDX req_prod = np->rx.req_prod_pvt;
255 struct xen_netif_rx_request *req;
257 if (unlikely(!netif_carrier_ok(dev)))
261 * Allocate skbuffs greedily, even though we batch updates to the
262 * receive ring. This creates a less bursty demand on the memory
263 * allocator, so should reduce the chance of failed allocation requests
264 * both for ourself and for other kernel subsystems.
266 batch_target = np->rx_target - (req_prod - np->rx.rsp_cons);
267 for (i = skb_queue_len(&np->rx_batch); i < batch_target; i++) {
268 skb = __netdev_alloc_skb(dev, RX_COPY_THRESHOLD + NET_IP_ALIGN,
269 GFP_ATOMIC | __GFP_NOWARN);
273 /* Align ip header to a 16 bytes boundary */
274 skb_reserve(skb, NET_IP_ALIGN);
276 page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
280 /* Could not allocate any skbuffs. Try again later. */
281 mod_timer(&np->rx_refill_timer,
284 /* Any skbuffs queued for refill? Force them out. */
290 skb_add_rx_frag(skb, 0, page, 0, 0, PAGE_SIZE);
291 __skb_queue_tail(&np->rx_batch, skb);
294 /* Is the batch large enough to be worthwhile? */
295 if (i < (np->rx_target/2)) {
296 if (req_prod > np->rx.sring->req_prod)
301 /* Adjust our fill target if we risked running out of buffers. */
302 if (((req_prod - np->rx.sring->rsp_prod) < (np->rx_target / 4)) &&
303 ((np->rx_target *= 2) > np->rx_max_target))
304 np->rx_target = np->rx_max_target;
308 skb = __skb_dequeue(&np->rx_batch);
314 id = xennet_rxidx(req_prod + i);
316 BUG_ON(np->rx_skbs[id]);
317 np->rx_skbs[id] = skb;
319 ref = gnttab_claim_grant_reference(&np->gref_rx_head);
320 BUG_ON((signed short)ref < 0);
321 np->grant_rx_ref[id] = ref;
323 pfn = page_to_pfn(skb_frag_page(&skb_shinfo(skb)->frags[0]));
324 vaddr = page_address(skb_frag_page(&skb_shinfo(skb)->frags[0]));
326 req = RING_GET_REQUEST(&np->rx, req_prod + i);
327 gnttab_grant_foreign_access_ref(ref,
328 np->xbdev->otherend_id,
336 wmb(); /* barrier so backend seens requests */
338 /* Above is a suitable barrier to ensure backend will see requests. */
339 np->rx.req_prod_pvt = req_prod + i;
341 RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->rx, notify);
343 notify_remote_via_irq(np->rx_irq);
346 static int xennet_open(struct net_device *dev)
348 struct netfront_info *np = netdev_priv(dev);
350 napi_enable(&np->napi);
352 spin_lock_bh(&np->rx_lock);
353 if (netif_carrier_ok(dev)) {
354 xennet_alloc_rx_buffers(dev);
355 np->rx.sring->rsp_event = np->rx.rsp_cons + 1;
356 if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx))
357 napi_schedule(&np->napi);
359 spin_unlock_bh(&np->rx_lock);
361 netif_start_queue(dev);
366 static void xennet_tx_buf_gc(struct net_device *dev)
370 struct netfront_info *np = netdev_priv(dev);
373 BUG_ON(!netif_carrier_ok(dev));
376 prod = np->tx.sring->rsp_prod;
377 rmb(); /* Ensure we see responses up to 'rp'. */
379 for (cons = np->tx.rsp_cons; cons != prod; cons++) {
380 struct xen_netif_tx_response *txrsp;
382 txrsp = RING_GET_RESPONSE(&np->tx, cons);
383 if (txrsp->status == XEN_NETIF_RSP_NULL)
387 skb = np->tx_skbs[id].skb;
388 if (unlikely(gnttab_query_foreign_access(
389 np->grant_tx_ref[id]) != 0)) {
390 pr_alert("%s: warning -- grant still in use by backend domain\n",
394 gnttab_end_foreign_access_ref(
395 np->grant_tx_ref[id], GNTMAP_readonly);
396 gnttab_release_grant_reference(
397 &np->gref_tx_head, np->grant_tx_ref[id]);
398 np->grant_tx_ref[id] = GRANT_INVALID_REF;
399 add_id_to_freelist(&np->tx_skb_freelist, np->tx_skbs, id);
400 dev_kfree_skb_irq(skb);
403 np->tx.rsp_cons = prod;
406 * Set a new event, then check for race with update of tx_cons.
407 * Note that it is essential to schedule a callback, no matter
408 * how few buffers are pending. Even if there is space in the
409 * transmit ring, higher layers may be blocked because too much
410 * data is outstanding: in such cases notification from Xen is
411 * likely to be the only kick that we'll get.
413 np->tx.sring->rsp_event =
414 prod + ((np->tx.sring->req_prod - prod) >> 1) + 1;
415 mb(); /* update shared area */
416 } while ((cons == prod) && (prod != np->tx.sring->rsp_prod));
418 xennet_maybe_wake_tx(dev);
421 static void xennet_make_frags(struct sk_buff *skb, struct net_device *dev,
422 struct xen_netif_tx_request *tx)
424 struct netfront_info *np = netdev_priv(dev);
425 char *data = skb->data;
427 RING_IDX prod = np->tx.req_prod_pvt;
428 int frags = skb_shinfo(skb)->nr_frags;
429 unsigned int offset = offset_in_page(data);
430 unsigned int len = skb_headlen(skb);
435 /* While the header overlaps a page boundary (including being
436 larger than a page), split it it into page-sized chunks. */
437 while (len > PAGE_SIZE - offset) {
438 tx->size = PAGE_SIZE - offset;
439 tx->flags |= XEN_NETTXF_more_data;
444 id = get_id_from_freelist(&np->tx_skb_freelist, np->tx_skbs);
445 np->tx_skbs[id].skb = skb_get(skb);
446 tx = RING_GET_REQUEST(&np->tx, prod++);
448 ref = gnttab_claim_grant_reference(&np->gref_tx_head);
449 BUG_ON((signed short)ref < 0);
451 mfn = virt_to_mfn(data);
452 gnttab_grant_foreign_access_ref(ref, np->xbdev->otherend_id,
453 mfn, GNTMAP_readonly);
455 tx->gref = np->grant_tx_ref[id] = ref;
461 /* Grant backend access to each skb fragment page. */
462 for (i = 0; i < frags; i++) {
463 skb_frag_t *frag = skb_shinfo(skb)->frags + i;
464 struct page *page = skb_frag_page(frag);
466 len = skb_frag_size(frag);
467 offset = frag->page_offset;
469 /* Data must not cross a page boundary. */
470 BUG_ON(len + offset > PAGE_SIZE<<compound_order(page));
472 /* Skip unused frames from start of page */
473 page += offset >> PAGE_SHIFT;
474 offset &= ~PAGE_MASK;
479 BUG_ON(offset >= PAGE_SIZE);
481 bytes = PAGE_SIZE - offset;
485 tx->flags |= XEN_NETTXF_more_data;
487 id = get_id_from_freelist(&np->tx_skb_freelist,
489 np->tx_skbs[id].skb = skb_get(skb);
490 tx = RING_GET_REQUEST(&np->tx, prod++);
492 ref = gnttab_claim_grant_reference(&np->gref_tx_head);
493 BUG_ON((signed short)ref < 0);
495 mfn = pfn_to_mfn(page_to_pfn(page));
496 gnttab_grant_foreign_access_ref(ref,
497 np->xbdev->otherend_id,
498 mfn, GNTMAP_readonly);
500 tx->gref = np->grant_tx_ref[id] = ref;
509 if (offset == PAGE_SIZE && len) {
510 BUG_ON(!PageCompound(page));
517 np->tx.req_prod_pvt = prod;
521 * Count how many ring slots are required to send the frags of this
522 * skb. Each frag might be a compound page.
524 static int xennet_count_skb_frag_slots(struct sk_buff *skb)
526 int i, frags = skb_shinfo(skb)->nr_frags;
529 for (i = 0; i < frags; i++) {
530 skb_frag_t *frag = skb_shinfo(skb)->frags + i;
531 unsigned long size = skb_frag_size(frag);
532 unsigned long offset = frag->page_offset;
534 /* Skip unused frames from start of page */
535 offset &= ~PAGE_MASK;
537 pages += PFN_UP(offset + size);
543 static int xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
546 struct netfront_info *np = netdev_priv(dev);
547 struct netfront_stats *stats = this_cpu_ptr(np->stats);
548 struct xen_netif_tx_request *tx;
549 char *data = skb->data;
555 unsigned int offset = offset_in_page(data);
556 unsigned int len = skb_headlen(skb);
559 /* If skb->len is too big for wire format, drop skb and alert
560 * user about misconfiguration.
562 if (unlikely(skb->len > XEN_NETIF_MAX_TX_SIZE)) {
563 net_alert_ratelimited(
564 "xennet: skb->len = %u, too big for wire format\n",
569 slots = DIV_ROUND_UP(offset + len, PAGE_SIZE) +
570 xennet_count_skb_frag_slots(skb);
571 if (unlikely(slots > MAX_SKB_FRAGS + 1)) {
572 net_alert_ratelimited(
573 "xennet: skb rides the rocket: %d slots\n", slots);
577 spin_lock_irqsave(&np->tx_lock, flags);
579 if (unlikely(!netif_carrier_ok(dev) ||
580 (slots > 1 && !xennet_can_sg(dev)) ||
581 netif_needs_gso(skb, netif_skb_features(skb)))) {
582 spin_unlock_irqrestore(&np->tx_lock, flags);
586 i = np->tx.req_prod_pvt;
588 id = get_id_from_freelist(&np->tx_skb_freelist, np->tx_skbs);
589 np->tx_skbs[id].skb = skb;
591 tx = RING_GET_REQUEST(&np->tx, i);
594 ref = gnttab_claim_grant_reference(&np->gref_tx_head);
595 BUG_ON((signed short)ref < 0);
596 mfn = virt_to_mfn(data);
597 gnttab_grant_foreign_access_ref(
598 ref, np->xbdev->otherend_id, mfn, GNTMAP_readonly);
599 tx->gref = np->grant_tx_ref[id] = ref;
604 if (skb->ip_summed == CHECKSUM_PARTIAL)
606 tx->flags |= XEN_NETTXF_csum_blank | XEN_NETTXF_data_validated;
607 else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
608 /* remote but checksummed. */
609 tx->flags |= XEN_NETTXF_data_validated;
611 if (skb_shinfo(skb)->gso_size) {
612 struct xen_netif_extra_info *gso;
614 gso = (struct xen_netif_extra_info *)
615 RING_GET_REQUEST(&np->tx, ++i);
617 tx->flags |= XEN_NETTXF_extra_info;
619 gso->u.gso.size = skb_shinfo(skb)->gso_size;
620 gso->u.gso.type = (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) ?
621 XEN_NETIF_GSO_TYPE_TCPV6 :
622 XEN_NETIF_GSO_TYPE_TCPV4;
624 gso->u.gso.features = 0;
626 gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
630 np->tx.req_prod_pvt = i + 1;
632 xennet_make_frags(skb, dev, tx);
635 RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->tx, notify);
637 notify_remote_via_irq(np->tx_irq);
639 u64_stats_update_begin(&stats->syncp);
640 stats->tx_bytes += skb->len;
642 u64_stats_update_end(&stats->syncp);
644 /* Note: It is not safe to access skb after xennet_tx_buf_gc()! */
645 xennet_tx_buf_gc(dev);
647 if (!netfront_tx_slot_available(np))
648 netif_stop_queue(dev);
650 spin_unlock_irqrestore(&np->tx_lock, flags);
655 dev->stats.tx_dropped++;
660 static int xennet_close(struct net_device *dev)
662 struct netfront_info *np = netdev_priv(dev);
663 netif_stop_queue(np->netdev);
664 napi_disable(&np->napi);
668 static void xennet_move_rx_slot(struct netfront_info *np, struct sk_buff *skb,
671 int new = xennet_rxidx(np->rx.req_prod_pvt);
673 BUG_ON(np->rx_skbs[new]);
674 np->rx_skbs[new] = skb;
675 np->grant_rx_ref[new] = ref;
676 RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->id = new;
677 RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->gref = ref;
678 np->rx.req_prod_pvt++;
681 static int xennet_get_extras(struct netfront_info *np,
682 struct xen_netif_extra_info *extras,
686 struct xen_netif_extra_info *extra;
687 struct device *dev = &np->netdev->dev;
688 RING_IDX cons = np->rx.rsp_cons;
695 if (unlikely(cons + 1 == rp)) {
697 dev_warn(dev, "Missing extra info\n");
702 extra = (struct xen_netif_extra_info *)
703 RING_GET_RESPONSE(&np->rx, ++cons);
705 if (unlikely(!extra->type ||
706 extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
708 dev_warn(dev, "Invalid extra type: %d\n",
712 memcpy(&extras[extra->type - 1], extra,
716 skb = xennet_get_rx_skb(np, cons);
717 ref = xennet_get_rx_ref(np, cons);
718 xennet_move_rx_slot(np, skb, ref);
719 } while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);
721 np->rx.rsp_cons = cons;
725 static int xennet_get_responses(struct netfront_info *np,
726 struct netfront_rx_info *rinfo, RING_IDX rp,
727 struct sk_buff_head *list)
729 struct xen_netif_rx_response *rx = &rinfo->rx;
730 struct xen_netif_extra_info *extras = rinfo->extras;
731 struct device *dev = &np->netdev->dev;
732 RING_IDX cons = np->rx.rsp_cons;
733 struct sk_buff *skb = xennet_get_rx_skb(np, cons);
734 grant_ref_t ref = xennet_get_rx_ref(np, cons);
735 int max = MAX_SKB_FRAGS + (rx->status <= RX_COPY_THRESHOLD);
740 if (rx->flags & XEN_NETRXF_extra_info) {
741 err = xennet_get_extras(np, extras, rp);
742 cons = np->rx.rsp_cons;
746 if (unlikely(rx->status < 0 ||
747 rx->offset + rx->status > PAGE_SIZE)) {
749 dev_warn(dev, "rx->offset: %x, size: %u\n",
750 rx->offset, rx->status);
751 xennet_move_rx_slot(np, skb, ref);
757 * This definitely indicates a bug, either in this driver or in
758 * the backend driver. In future this should flag the bad
759 * situation to the system controller to reboot the backend.
761 if (ref == GRANT_INVALID_REF) {
763 dev_warn(dev, "Bad rx response id %d.\n",
769 ret = gnttab_end_foreign_access_ref(ref, 0);
772 gnttab_release_grant_reference(&np->gref_rx_head, ref);
774 __skb_queue_tail(list, skb);
777 if (!(rx->flags & XEN_NETRXF_more_data))
780 if (cons + slots == rp) {
782 dev_warn(dev, "Need more slots\n");
787 rx = RING_GET_RESPONSE(&np->rx, cons + slots);
788 skb = xennet_get_rx_skb(np, cons + slots);
789 ref = xennet_get_rx_ref(np, cons + slots);
793 if (unlikely(slots > max)) {
795 dev_warn(dev, "Too many slots\n");
800 np->rx.rsp_cons = cons + slots;
805 static int xennet_set_skb_gso(struct sk_buff *skb,
806 struct xen_netif_extra_info *gso)
808 if (!gso->u.gso.size) {
810 pr_warn("GSO size must not be zero\n");
814 if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4 &&
815 gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV6) {
817 pr_warn("Bad GSO type %d\n", gso->u.gso.type);
821 skb_shinfo(skb)->gso_size = gso->u.gso.size;
822 skb_shinfo(skb)->gso_type =
823 (gso->u.gso.type == XEN_NETIF_GSO_TYPE_TCPV4) ?
827 /* Header must be checked, and gso_segs computed. */
828 skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
829 skb_shinfo(skb)->gso_segs = 0;
834 static RING_IDX xennet_fill_frags(struct netfront_info *np,
836 struct sk_buff_head *list)
838 struct skb_shared_info *shinfo = skb_shinfo(skb);
839 RING_IDX cons = np->rx.rsp_cons;
840 struct sk_buff *nskb;
842 while ((nskb = __skb_dequeue(list))) {
843 struct xen_netif_rx_response *rx =
844 RING_GET_RESPONSE(&np->rx, ++cons);
845 skb_frag_t *nfrag = &skb_shinfo(nskb)->frags[0];
847 if (shinfo->nr_frags == MAX_SKB_FRAGS) {
848 unsigned int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
850 BUG_ON(pull_to <= skb_headlen(skb));
851 __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
853 BUG_ON(shinfo->nr_frags >= MAX_SKB_FRAGS);
855 skb_add_rx_frag(skb, shinfo->nr_frags, skb_frag_page(nfrag),
856 rx->offset, rx->status, PAGE_SIZE);
858 skb_shinfo(nskb)->nr_frags = 0;
865 static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
867 bool recalculate_partial_csum = false;
870 * A GSO SKB must be CHECKSUM_PARTIAL. However some buggy
871 * peers can fail to set NETRXF_csum_blank when sending a GSO
872 * frame. In this case force the SKB to CHECKSUM_PARTIAL and
873 * recalculate the partial checksum.
875 if (skb->ip_summed != CHECKSUM_PARTIAL && skb_is_gso(skb)) {
876 struct netfront_info *np = netdev_priv(dev);
877 np->rx_gso_checksum_fixup++;
878 skb->ip_summed = CHECKSUM_PARTIAL;
879 recalculate_partial_csum = true;
882 /* A non-CHECKSUM_PARTIAL SKB does not require setup. */
883 if (skb->ip_summed != CHECKSUM_PARTIAL)
886 return skb_checksum_setup(skb, recalculate_partial_csum);
889 static int handle_incoming_queue(struct net_device *dev,
890 struct sk_buff_head *rxq)
892 struct netfront_info *np = netdev_priv(dev);
893 struct netfront_stats *stats = this_cpu_ptr(np->stats);
894 int packets_dropped = 0;
897 while ((skb = __skb_dequeue(rxq)) != NULL) {
898 int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
900 if (pull_to > skb_headlen(skb))
901 __pskb_pull_tail(skb, pull_to - skb_headlen(skb));
903 /* Ethernet work: Delayed to here as it peeks the header. */
904 skb->protocol = eth_type_trans(skb, dev);
906 if (checksum_setup(dev, skb)) {
909 dev->stats.rx_errors++;
913 u64_stats_update_begin(&stats->syncp);
915 stats->rx_bytes += skb->len;
916 u64_stats_update_end(&stats->syncp);
919 napi_gro_receive(&np->napi, skb);
922 return packets_dropped;
925 static int xennet_poll(struct napi_struct *napi, int budget)
927 struct netfront_info *np = container_of(napi, struct netfront_info, napi);
928 struct net_device *dev = np->netdev;
930 struct netfront_rx_info rinfo;
931 struct xen_netif_rx_response *rx = &rinfo.rx;
932 struct xen_netif_extra_info *extras = rinfo.extras;
935 struct sk_buff_head rxq;
936 struct sk_buff_head errq;
937 struct sk_buff_head tmpq;
941 spin_lock(&np->rx_lock);
943 skb_queue_head_init(&rxq);
944 skb_queue_head_init(&errq);
945 skb_queue_head_init(&tmpq);
947 rp = np->rx.sring->rsp_prod;
948 rmb(); /* Ensure we see queued responses up to 'rp'. */
952 while ((i != rp) && (work_done < budget)) {
953 memcpy(rx, RING_GET_RESPONSE(&np->rx, i), sizeof(*rx));
954 memset(extras, 0, sizeof(rinfo.extras));
956 err = xennet_get_responses(np, &rinfo, rp, &tmpq);
960 while ((skb = __skb_dequeue(&tmpq)))
961 __skb_queue_tail(&errq, skb);
962 dev->stats.rx_errors++;
967 skb = __skb_dequeue(&tmpq);
969 if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
970 struct xen_netif_extra_info *gso;
971 gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
973 if (unlikely(xennet_set_skb_gso(skb, gso))) {
974 __skb_queue_head(&tmpq, skb);
975 np->rx.rsp_cons += skb_queue_len(&tmpq);
980 NETFRONT_SKB_CB(skb)->pull_to = rx->status;
981 if (NETFRONT_SKB_CB(skb)->pull_to > RX_COPY_THRESHOLD)
982 NETFRONT_SKB_CB(skb)->pull_to = RX_COPY_THRESHOLD;
984 skb_shinfo(skb)->frags[0].page_offset = rx->offset;
985 skb_frag_size_set(&skb_shinfo(skb)->frags[0], rx->status);
986 skb->data_len = rx->status;
987 skb->len += rx->status;
989 i = xennet_fill_frags(np, skb, &tmpq);
991 if (rx->flags & XEN_NETRXF_csum_blank)
992 skb->ip_summed = CHECKSUM_PARTIAL;
993 else if (rx->flags & XEN_NETRXF_data_validated)
994 skb->ip_summed = CHECKSUM_UNNECESSARY;
996 __skb_queue_tail(&rxq, skb);
998 np->rx.rsp_cons = ++i;
1002 __skb_queue_purge(&errq);
1004 work_done -= handle_incoming_queue(dev, &rxq);
1006 /* If we get a callback with very few responses, reduce fill target. */
1007 /* NB. Note exponential increase, linear decrease. */
1008 if (((np->rx.req_prod_pvt - np->rx.sring->rsp_prod) >
1009 ((3*np->rx_target) / 4)) &&
1010 (--np->rx_target < np->rx_min_target))
1011 np->rx_target = np->rx_min_target;
1013 xennet_alloc_rx_buffers(dev);
1015 if (work_done < budget) {
1018 napi_gro_flush(napi, false);
1020 local_irq_save(flags);
1022 RING_FINAL_CHECK_FOR_RESPONSES(&np->rx, more_to_do);
1024 __napi_complete(napi);
1026 local_irq_restore(flags);
1029 spin_unlock(&np->rx_lock);
1034 static int xennet_change_mtu(struct net_device *dev, int mtu)
1036 int max = xennet_can_sg(dev) ?
1037 XEN_NETIF_MAX_TX_SIZE - MAX_TCP_HEADER : ETH_DATA_LEN;
1045 static struct rtnl_link_stats64 *xennet_get_stats64(struct net_device *dev,
1046 struct rtnl_link_stats64 *tot)
1048 struct netfront_info *np = netdev_priv(dev);
1051 for_each_possible_cpu(cpu) {
1052 struct netfront_stats *stats = per_cpu_ptr(np->stats, cpu);
1053 u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
1057 start = u64_stats_fetch_begin_bh(&stats->syncp);
1059 rx_packets = stats->rx_packets;
1060 tx_packets = stats->tx_packets;
1061 rx_bytes = stats->rx_bytes;
1062 tx_bytes = stats->tx_bytes;
1063 } while (u64_stats_fetch_retry_bh(&stats->syncp, start));
1065 tot->rx_packets += rx_packets;
1066 tot->tx_packets += tx_packets;
1067 tot->rx_bytes += rx_bytes;
1068 tot->tx_bytes += tx_bytes;
1071 tot->rx_errors = dev->stats.rx_errors;
1072 tot->tx_dropped = dev->stats.tx_dropped;
1077 static void xennet_release_tx_bufs(struct netfront_info *np)
1079 struct sk_buff *skb;
1082 for (i = 0; i < NET_TX_RING_SIZE; i++) {
1083 /* Skip over entries which are actually freelist references */
1084 if (skb_entry_is_link(&np->tx_skbs[i]))
1087 skb = np->tx_skbs[i].skb;
1088 gnttab_end_foreign_access_ref(np->grant_tx_ref[i],
1090 gnttab_release_grant_reference(&np->gref_tx_head,
1091 np->grant_tx_ref[i]);
1092 np->grant_tx_ref[i] = GRANT_INVALID_REF;
1093 add_id_to_freelist(&np->tx_skb_freelist, np->tx_skbs, i);
1094 dev_kfree_skb_irq(skb);
1098 static void xennet_release_rx_bufs(struct netfront_info *np)
1100 struct mmu_update *mmu = np->rx_mmu;
1101 struct multicall_entry *mcl = np->rx_mcl;
1102 struct sk_buff_head free_list;
1103 struct sk_buff *skb;
1105 int xfer = 0, noxfer = 0, unused = 0;
1108 dev_warn(&np->netdev->dev, "%s: fix me for copying receiver.\n",
1112 skb_queue_head_init(&free_list);
1114 spin_lock_bh(&np->rx_lock);
1116 for (id = 0; id < NET_RX_RING_SIZE; id++) {
1117 ref = np->grant_rx_ref[id];
1118 if (ref == GRANT_INVALID_REF) {
1123 skb = np->rx_skbs[id];
1124 mfn = gnttab_end_foreign_transfer_ref(ref);
1125 gnttab_release_grant_reference(&np->gref_rx_head, ref);
1126 np->grant_rx_ref[id] = GRANT_INVALID_REF;
1129 skb_shinfo(skb)->nr_frags = 0;
1135 if (!xen_feature(XENFEAT_auto_translated_physmap)) {
1136 /* Remap the page. */
1137 const struct page *page =
1138 skb_frag_page(&skb_shinfo(skb)->frags[0]);
1139 unsigned long pfn = page_to_pfn(page);
1140 void *vaddr = page_address(page);
1142 MULTI_update_va_mapping(mcl, (unsigned long)vaddr,
1143 mfn_pte(mfn, PAGE_KERNEL),
1146 mmu->ptr = ((u64)mfn << PAGE_SHIFT)
1147 | MMU_MACHPHYS_UPDATE;
1151 set_phys_to_machine(pfn, mfn);
1153 __skb_queue_tail(&free_list, skb);
1157 dev_info(&np->netdev->dev, "%s: %d xfer, %d noxfer, %d unused\n",
1158 __func__, xfer, noxfer, unused);
1161 if (!xen_feature(XENFEAT_auto_translated_physmap)) {
1162 /* Do all the remapping work and M2P updates. */
1163 MULTI_mmu_update(mcl, np->rx_mmu, mmu - np->rx_mmu,
1166 HYPERVISOR_multicall(np->rx_mcl, mcl - np->rx_mcl);
1170 __skb_queue_purge(&free_list);
1172 spin_unlock_bh(&np->rx_lock);
1175 static void xennet_uninit(struct net_device *dev)
1177 struct netfront_info *np = netdev_priv(dev);
1178 xennet_release_tx_bufs(np);
1179 xennet_release_rx_bufs(np);
1180 gnttab_free_grant_references(np->gref_tx_head);
1181 gnttab_free_grant_references(np->gref_rx_head);
1184 static netdev_features_t xennet_fix_features(struct net_device *dev,
1185 netdev_features_t features)
1187 struct netfront_info *np = netdev_priv(dev);
1190 if (features & NETIF_F_SG) {
1191 if (xenbus_scanf(XBT_NIL, np->xbdev->otherend, "feature-sg",
1196 features &= ~NETIF_F_SG;
1199 if (features & NETIF_F_IPV6_CSUM) {
1200 if (xenbus_scanf(XBT_NIL, np->xbdev->otherend,
1201 "feature-ipv6-csum-offload", "%d", &val) < 0)
1205 features &= ~NETIF_F_IPV6_CSUM;
1208 if (features & NETIF_F_TSO) {
1209 if (xenbus_scanf(XBT_NIL, np->xbdev->otherend,
1210 "feature-gso-tcpv4", "%d", &val) < 0)
1214 features &= ~NETIF_F_TSO;
1217 if (features & NETIF_F_TSO6) {
1218 if (xenbus_scanf(XBT_NIL, np->xbdev->otherend,
1219 "feature-gso-tcpv6", "%d", &val) < 0)
1223 features &= ~NETIF_F_TSO6;
1229 static int xennet_set_features(struct net_device *dev,
1230 netdev_features_t features)
1232 if (!(features & NETIF_F_SG) && dev->mtu > ETH_DATA_LEN) {
1233 netdev_info(dev, "Reducing MTU because no SG offload");
1234 dev->mtu = ETH_DATA_LEN;
1240 static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
1242 struct netfront_info *np = dev_id;
1243 struct net_device *dev = np->netdev;
1244 unsigned long flags;
1246 spin_lock_irqsave(&np->tx_lock, flags);
1247 xennet_tx_buf_gc(dev);
1248 spin_unlock_irqrestore(&np->tx_lock, flags);
1253 static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
1255 struct netfront_info *np = dev_id;
1256 struct net_device *dev = np->netdev;
1258 if (likely(netif_carrier_ok(dev) &&
1259 RING_HAS_UNCONSUMED_RESPONSES(&np->rx)))
1260 napi_schedule(&np->napi);
1265 static irqreturn_t xennet_interrupt(int irq, void *dev_id)
1267 xennet_tx_interrupt(irq, dev_id);
1268 xennet_rx_interrupt(irq, dev_id);
1272 #ifdef CONFIG_NET_POLL_CONTROLLER
1273 static void xennet_poll_controller(struct net_device *dev)
1275 xennet_interrupt(0, dev);
1279 static const struct net_device_ops xennet_netdev_ops = {
1280 .ndo_open = xennet_open,
1281 .ndo_uninit = xennet_uninit,
1282 .ndo_stop = xennet_close,
1283 .ndo_start_xmit = xennet_start_xmit,
1284 .ndo_change_mtu = xennet_change_mtu,
1285 .ndo_get_stats64 = xennet_get_stats64,
1286 .ndo_set_mac_address = eth_mac_addr,
1287 .ndo_validate_addr = eth_validate_addr,
1288 .ndo_fix_features = xennet_fix_features,
1289 .ndo_set_features = xennet_set_features,
1290 #ifdef CONFIG_NET_POLL_CONTROLLER
1291 .ndo_poll_controller = xennet_poll_controller,
1295 static struct net_device *xennet_create_dev(struct xenbus_device *dev)
1298 struct net_device *netdev;
1299 struct netfront_info *np;
1301 netdev = alloc_etherdev(sizeof(struct netfront_info));
1303 return ERR_PTR(-ENOMEM);
1305 np = netdev_priv(netdev);
1308 spin_lock_init(&np->tx_lock);
1309 spin_lock_init(&np->rx_lock);
1311 skb_queue_head_init(&np->rx_batch);
1312 np->rx_target = RX_DFL_MIN_TARGET;
1313 np->rx_min_target = RX_DFL_MIN_TARGET;
1314 np->rx_max_target = RX_MAX_TARGET;
1316 init_timer(&np->rx_refill_timer);
1317 np->rx_refill_timer.data = (unsigned long)netdev;
1318 np->rx_refill_timer.function = rx_refill_timeout;
1321 np->stats = alloc_percpu(struct netfront_stats);
1322 if (np->stats == NULL)
1325 for_each_possible_cpu(i) {
1326 struct netfront_stats *xen_nf_stats;
1327 xen_nf_stats = per_cpu_ptr(np->stats, i);
1328 u64_stats_init(&xen_nf_stats->syncp);
1331 /* Initialise tx_skbs as a free chain containing every entry. */
1332 np->tx_skb_freelist = 0;
1333 for (i = 0; i < NET_TX_RING_SIZE; i++) {
1334 skb_entry_set_link(&np->tx_skbs[i], i+1);
1335 np->grant_tx_ref[i] = GRANT_INVALID_REF;
1338 /* Clear out rx_skbs */
1339 for (i = 0; i < NET_RX_RING_SIZE; i++) {
1340 np->rx_skbs[i] = NULL;
1341 np->grant_rx_ref[i] = GRANT_INVALID_REF;
1344 /* A grant for every tx ring slot */
1345 if (gnttab_alloc_grant_references(TX_MAX_TARGET,
1346 &np->gref_tx_head) < 0) {
1347 pr_alert("can't alloc tx grant refs\n");
1349 goto exit_free_stats;
1351 /* A grant for every rx ring slot */
1352 if (gnttab_alloc_grant_references(RX_MAX_TARGET,
1353 &np->gref_rx_head) < 0) {
1354 pr_alert("can't alloc rx grant refs\n");
1359 netdev->netdev_ops = &xennet_netdev_ops;
1361 netif_napi_add(netdev, &np->napi, xennet_poll, 64);
1362 netdev->features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
1364 netdev->hw_features = NETIF_F_SG |
1366 NETIF_F_TSO | NETIF_F_TSO6;
1369 * Assume that all hw features are available for now. This set
1370 * will be adjusted by the call to netdev_update_features() in
1371 * xennet_connect() which is the earliest point where we can
1372 * negotiate with the backend regarding supported features.
1374 netdev->features |= netdev->hw_features;
1376 SET_ETHTOOL_OPS(netdev, &xennet_ethtool_ops);
1377 SET_NETDEV_DEV(netdev, &dev->dev);
1379 netif_set_gso_max_size(netdev, XEN_NETIF_MAX_TX_SIZE - MAX_TCP_HEADER);
1381 np->netdev = netdev;
1383 netif_carrier_off(netdev);
1388 gnttab_free_grant_references(np->gref_tx_head);
1390 free_percpu(np->stats);
1392 free_netdev(netdev);
1393 return ERR_PTR(err);
1397 * Entry point to this code when a new device is created. Allocate the basic
1398 * structures and the ring buffers for communication with the backend, and
1399 * inform the backend of the appropriate details for those.
1401 static int netfront_probe(struct xenbus_device *dev,
1402 const struct xenbus_device_id *id)
1405 struct net_device *netdev;
1406 struct netfront_info *info;
1408 netdev = xennet_create_dev(dev);
1409 if (IS_ERR(netdev)) {
1410 err = PTR_ERR(netdev);
1411 xenbus_dev_fatal(dev, err, "creating netdev");
1415 info = netdev_priv(netdev);
1416 dev_set_drvdata(&dev->dev, info);
1418 err = register_netdev(info->netdev);
1420 pr_warn("%s: register_netdev err=%d\n", __func__, err);
1424 err = xennet_sysfs_addif(info->netdev);
1426 unregister_netdev(info->netdev);
1427 pr_warn("%s: add sysfs failed err=%d\n", __func__, err);
1434 free_netdev(netdev);
1435 dev_set_drvdata(&dev->dev, NULL);
1439 static void xennet_end_access(int ref, void *page)
1441 /* This frees the page as a side-effect */
1442 if (ref != GRANT_INVALID_REF)
1443 gnttab_end_foreign_access(ref, 0, (unsigned long)page);
1446 static void xennet_disconnect_backend(struct netfront_info *info)
1448 /* Stop old i/f to prevent errors whilst we rebuild the state. */
1449 spin_lock_bh(&info->rx_lock);
1450 spin_lock_irq(&info->tx_lock);
1451 netif_carrier_off(info->netdev);
1452 spin_unlock_irq(&info->tx_lock);
1453 spin_unlock_bh(&info->rx_lock);
1455 if (info->tx_irq && (info->tx_irq == info->rx_irq))
1456 unbind_from_irqhandler(info->tx_irq, info);
1457 if (info->tx_irq && (info->tx_irq != info->rx_irq)) {
1458 unbind_from_irqhandler(info->tx_irq, info);
1459 unbind_from_irqhandler(info->rx_irq, info);
1461 info->tx_evtchn = info->rx_evtchn = 0;
1462 info->tx_irq = info->rx_irq = 0;
1464 /* End access and free the pages */
1465 xennet_end_access(info->tx_ring_ref, info->tx.sring);
1466 xennet_end_access(info->rx_ring_ref, info->rx.sring);
1468 info->tx_ring_ref = GRANT_INVALID_REF;
1469 info->rx_ring_ref = GRANT_INVALID_REF;
1470 info->tx.sring = NULL;
1471 info->rx.sring = NULL;
1475 * We are reconnecting to the backend, due to a suspend/resume, or a backend
1476 * driver restart. We tear down our netif structure and recreate it, but
1477 * leave the device-layer structures intact so that this is transparent to the
1478 * rest of the kernel.
1480 static int netfront_resume(struct xenbus_device *dev)
1482 struct netfront_info *info = dev_get_drvdata(&dev->dev);
1484 dev_dbg(&dev->dev, "%s\n", dev->nodename);
1486 xennet_disconnect_backend(info);
1490 static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
1492 char *s, *e, *macstr;
1495 macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
1497 return PTR_ERR(macstr);
1499 for (i = 0; i < ETH_ALEN; i++) {
1500 mac[i] = simple_strtoul(s, &e, 16);
1501 if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
1512 static int setup_netfront_single(struct netfront_info *info)
1516 err = xenbus_alloc_evtchn(info->xbdev, &info->tx_evtchn);
1520 err = bind_evtchn_to_irqhandler(info->tx_evtchn,
1522 0, info->netdev->name, info);
1525 info->rx_evtchn = info->tx_evtchn;
1526 info->rx_irq = info->tx_irq = err;
1531 xenbus_free_evtchn(info->xbdev, info->tx_evtchn);
1532 info->tx_evtchn = 0;
1537 static int setup_netfront_split(struct netfront_info *info)
1541 err = xenbus_alloc_evtchn(info->xbdev, &info->tx_evtchn);
1544 err = xenbus_alloc_evtchn(info->xbdev, &info->rx_evtchn);
1546 goto alloc_rx_evtchn_fail;
1548 snprintf(info->tx_irq_name, sizeof(info->tx_irq_name),
1549 "%s-tx", info->netdev->name);
1550 err = bind_evtchn_to_irqhandler(info->tx_evtchn,
1551 xennet_tx_interrupt,
1552 0, info->tx_irq_name, info);
1557 snprintf(info->rx_irq_name, sizeof(info->rx_irq_name),
1558 "%s-rx", info->netdev->name);
1559 err = bind_evtchn_to_irqhandler(info->rx_evtchn,
1560 xennet_rx_interrupt,
1561 0, info->rx_irq_name, info);
1569 unbind_from_irqhandler(info->tx_irq, info);
1572 xenbus_free_evtchn(info->xbdev, info->rx_evtchn);
1573 info->rx_evtchn = 0;
1574 alloc_rx_evtchn_fail:
1575 xenbus_free_evtchn(info->xbdev, info->tx_evtchn);
1576 info->tx_evtchn = 0;
1581 static int setup_netfront(struct xenbus_device *dev, struct netfront_info *info)
1583 struct xen_netif_tx_sring *txs;
1584 struct xen_netif_rx_sring *rxs;
1586 struct net_device *netdev = info->netdev;
1587 unsigned int feature_split_evtchn;
1589 info->tx_ring_ref = GRANT_INVALID_REF;
1590 info->rx_ring_ref = GRANT_INVALID_REF;
1591 info->rx.sring = NULL;
1592 info->tx.sring = NULL;
1595 err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
1596 "feature-split-event-channels", "%u",
1597 &feature_split_evtchn);
1599 feature_split_evtchn = 0;
1601 err = xen_net_read_mac(dev, netdev->dev_addr);
1603 xenbus_dev_fatal(dev, err, "parsing %s/mac", dev->nodename);
1607 txs = (struct xen_netif_tx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1610 xenbus_dev_fatal(dev, err, "allocating tx ring page");
1613 SHARED_RING_INIT(txs);
1614 FRONT_RING_INIT(&info->tx, txs, PAGE_SIZE);
1616 err = xenbus_grant_ring(dev, virt_to_mfn(txs));
1618 goto grant_tx_ring_fail;
1620 info->tx_ring_ref = err;
1621 rxs = (struct xen_netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
1624 xenbus_dev_fatal(dev, err, "allocating rx ring page");
1625 goto alloc_rx_ring_fail;
1627 SHARED_RING_INIT(rxs);
1628 FRONT_RING_INIT(&info->rx, rxs, PAGE_SIZE);
1630 err = xenbus_grant_ring(dev, virt_to_mfn(rxs));
1632 goto grant_rx_ring_fail;
1633 info->rx_ring_ref = err;
1635 if (feature_split_evtchn)
1636 err = setup_netfront_split(info);
1637 /* setup single event channel if
1638 * a) feature-split-event-channels == 0
1639 * b) feature-split-event-channels == 1 but failed to setup
1641 if (!feature_split_evtchn || (feature_split_evtchn && err))
1642 err = setup_netfront_single(info);
1645 goto alloc_evtchn_fail;
1649 /* If we fail to setup netfront, it is safe to just revoke access to
1650 * granted pages because backend is not accessing it at this point.
1653 gnttab_end_foreign_access_ref(info->rx_ring_ref, 0);
1655 free_page((unsigned long)rxs);
1657 gnttab_end_foreign_access_ref(info->tx_ring_ref, 0);
1659 free_page((unsigned long)txs);
1664 /* Common code used when first setting up, and when resuming. */
1665 static int talk_to_netback(struct xenbus_device *dev,
1666 struct netfront_info *info)
1668 const char *message;
1669 struct xenbus_transaction xbt;
1672 /* Create shared ring, alloc event channel. */
1673 err = setup_netfront(dev, info);
1678 err = xenbus_transaction_start(&xbt);
1680 xenbus_dev_fatal(dev, err, "starting transaction");
1684 err = xenbus_printf(xbt, dev->nodename, "tx-ring-ref", "%u",
1687 message = "writing tx ring-ref";
1688 goto abort_transaction;
1690 err = xenbus_printf(xbt, dev->nodename, "rx-ring-ref", "%u",
1693 message = "writing rx ring-ref";
1694 goto abort_transaction;
1697 if (info->tx_evtchn == info->rx_evtchn) {
1698 err = xenbus_printf(xbt, dev->nodename,
1699 "event-channel", "%u", info->tx_evtchn);
1701 message = "writing event-channel";
1702 goto abort_transaction;
1705 err = xenbus_printf(xbt, dev->nodename,
1706 "event-channel-tx", "%u", info->tx_evtchn);
1708 message = "writing event-channel-tx";
1709 goto abort_transaction;
1711 err = xenbus_printf(xbt, dev->nodename,
1712 "event-channel-rx", "%u", info->rx_evtchn);
1714 message = "writing event-channel-rx";
1715 goto abort_transaction;
1719 err = xenbus_printf(xbt, dev->nodename, "request-rx-copy", "%u",
1722 message = "writing request-rx-copy";
1723 goto abort_transaction;
1726 err = xenbus_printf(xbt, dev->nodename, "feature-rx-notify", "%d", 1);
1728 message = "writing feature-rx-notify";
1729 goto abort_transaction;
1732 err = xenbus_printf(xbt, dev->nodename, "feature-sg", "%d", 1);
1734 message = "writing feature-sg";
1735 goto abort_transaction;
1738 err = xenbus_printf(xbt, dev->nodename, "feature-gso-tcpv4", "%d", 1);
1740 message = "writing feature-gso-tcpv4";
1741 goto abort_transaction;
1744 err = xenbus_write(xbt, dev->nodename, "feature-gso-tcpv6", "1");
1746 message = "writing feature-gso-tcpv6";
1747 goto abort_transaction;
1750 err = xenbus_write(xbt, dev->nodename, "feature-ipv6-csum-offload",
1753 message = "writing feature-ipv6-csum-offload";
1754 goto abort_transaction;
1757 err = xenbus_transaction_end(xbt, 0);
1761 xenbus_dev_fatal(dev, err, "completing transaction");
1768 xenbus_transaction_end(xbt, 1);
1769 xenbus_dev_fatal(dev, err, "%s", message);
1771 xennet_disconnect_backend(info);
1776 static int xennet_connect(struct net_device *dev)
1778 struct netfront_info *np = netdev_priv(dev);
1779 int i, requeue_idx, err;
1780 struct sk_buff *skb;
1782 struct xen_netif_rx_request *req;
1783 unsigned int feature_rx_copy;
1785 err = xenbus_scanf(XBT_NIL, np->xbdev->otherend,
1786 "feature-rx-copy", "%u", &feature_rx_copy);
1788 feature_rx_copy = 0;
1790 if (!feature_rx_copy) {
1792 "backend does not support copying receive path\n");
1796 err = talk_to_netback(np->xbdev, np);
1801 netdev_update_features(dev);
1804 spin_lock_bh(&np->rx_lock);
1805 spin_lock_irq(&np->tx_lock);
1807 /* Step 1: Discard all pending TX packet fragments. */
1808 xennet_release_tx_bufs(np);
1810 /* Step 2: Rebuild the RX buffer freelist and the RX ring itself. */
1811 for (requeue_idx = 0, i = 0; i < NET_RX_RING_SIZE; i++) {
1813 const struct page *page;
1814 if (!np->rx_skbs[i])
1817 skb = np->rx_skbs[requeue_idx] = xennet_get_rx_skb(np, i);
1818 ref = np->grant_rx_ref[requeue_idx] = xennet_get_rx_ref(np, i);
1819 req = RING_GET_REQUEST(&np->rx, requeue_idx);
1821 frag = &skb_shinfo(skb)->frags[0];
1822 page = skb_frag_page(frag);
1823 gnttab_grant_foreign_access_ref(
1824 ref, np->xbdev->otherend_id,
1825 pfn_to_mfn(page_to_pfn(page)),
1828 req->id = requeue_idx;
1833 np->rx.req_prod_pvt = requeue_idx;
1836 * Step 3: All public and private state should now be sane. Get
1837 * ready to start sending and receiving packets and give the driver
1838 * domain a kick because we've probably just requeued some
1841 netif_carrier_on(np->netdev);
1842 notify_remote_via_irq(np->tx_irq);
1843 if (np->tx_irq != np->rx_irq)
1844 notify_remote_via_irq(np->rx_irq);
1845 xennet_tx_buf_gc(dev);
1846 xennet_alloc_rx_buffers(dev);
1848 spin_unlock_irq(&np->tx_lock);
1849 spin_unlock_bh(&np->rx_lock);
1855 * Callback received when the backend's state changes.
1857 static void netback_changed(struct xenbus_device *dev,
1858 enum xenbus_state backend_state)
1860 struct netfront_info *np = dev_get_drvdata(&dev->dev);
1861 struct net_device *netdev = np->netdev;
1863 dev_dbg(&dev->dev, "%s\n", xenbus_strstate(backend_state));
1865 switch (backend_state) {
1866 case XenbusStateInitialising:
1867 case XenbusStateInitialised:
1868 case XenbusStateReconfiguring:
1869 case XenbusStateReconfigured:
1870 case XenbusStateUnknown:
1871 case XenbusStateClosed:
1874 case XenbusStateInitWait:
1875 if (dev->state != XenbusStateInitialising)
1877 if (xennet_connect(netdev) != 0)
1879 xenbus_switch_state(dev, XenbusStateConnected);
1882 case XenbusStateConnected:
1883 netdev_notify_peers(netdev);
1886 case XenbusStateClosing:
1887 xenbus_frontend_closed(dev);
1892 static const struct xennet_stat {
1893 char name[ETH_GSTRING_LEN];
1895 } xennet_stats[] = {
1897 "rx_gso_checksum_fixup",
1898 offsetof(struct netfront_info, rx_gso_checksum_fixup)
1902 static int xennet_get_sset_count(struct net_device *dev, int string_set)
1904 switch (string_set) {
1906 return ARRAY_SIZE(xennet_stats);
1912 static void xennet_get_ethtool_stats(struct net_device *dev,
1913 struct ethtool_stats *stats, u64 * data)
1915 void *np = netdev_priv(dev);
1918 for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
1919 data[i] = *(unsigned long *)(np + xennet_stats[i].offset);
1922 static void xennet_get_strings(struct net_device *dev, u32 stringset, u8 * data)
1926 switch (stringset) {
1928 for (i = 0; i < ARRAY_SIZE(xennet_stats); i++)
1929 memcpy(data + i * ETH_GSTRING_LEN,
1930 xennet_stats[i].name, ETH_GSTRING_LEN);
1935 static const struct ethtool_ops xennet_ethtool_ops =
1937 .get_link = ethtool_op_get_link,
1939 .get_sset_count = xennet_get_sset_count,
1940 .get_ethtool_stats = xennet_get_ethtool_stats,
1941 .get_strings = xennet_get_strings,
1945 static ssize_t show_rxbuf_min(struct device *dev,
1946 struct device_attribute *attr, char *buf)
1948 struct net_device *netdev = to_net_dev(dev);
1949 struct netfront_info *info = netdev_priv(netdev);
1951 return sprintf(buf, "%u\n", info->rx_min_target);
1954 static ssize_t store_rxbuf_min(struct device *dev,
1955 struct device_attribute *attr,
1956 const char *buf, size_t len)
1958 struct net_device *netdev = to_net_dev(dev);
1959 struct netfront_info *np = netdev_priv(netdev);
1961 unsigned long target;
1963 if (!capable(CAP_NET_ADMIN))
1966 target = simple_strtoul(buf, &endp, 0);
1970 if (target < RX_MIN_TARGET)
1971 target = RX_MIN_TARGET;
1972 if (target > RX_MAX_TARGET)
1973 target = RX_MAX_TARGET;
1975 spin_lock_bh(&np->rx_lock);
1976 if (target > np->rx_max_target)
1977 np->rx_max_target = target;
1978 np->rx_min_target = target;
1979 if (target > np->rx_target)
1980 np->rx_target = target;
1982 xennet_alloc_rx_buffers(netdev);
1984 spin_unlock_bh(&np->rx_lock);
1988 static ssize_t show_rxbuf_max(struct device *dev,
1989 struct device_attribute *attr, char *buf)
1991 struct net_device *netdev = to_net_dev(dev);
1992 struct netfront_info *info = netdev_priv(netdev);
1994 return sprintf(buf, "%u\n", info->rx_max_target);
1997 static ssize_t store_rxbuf_max(struct device *dev,
1998 struct device_attribute *attr,
1999 const char *buf, size_t len)
2001 struct net_device *netdev = to_net_dev(dev);
2002 struct netfront_info *np = netdev_priv(netdev);
2004 unsigned long target;
2006 if (!capable(CAP_NET_ADMIN))
2009 target = simple_strtoul(buf, &endp, 0);
2013 if (target < RX_MIN_TARGET)
2014 target = RX_MIN_TARGET;
2015 if (target > RX_MAX_TARGET)
2016 target = RX_MAX_TARGET;
2018 spin_lock_bh(&np->rx_lock);
2019 if (target < np->rx_min_target)
2020 np->rx_min_target = target;
2021 np->rx_max_target = target;
2022 if (target < np->rx_target)
2023 np->rx_target = target;
2025 xennet_alloc_rx_buffers(netdev);
2027 spin_unlock_bh(&np->rx_lock);
2031 static ssize_t show_rxbuf_cur(struct device *dev,
2032 struct device_attribute *attr, char *buf)
2034 struct net_device *netdev = to_net_dev(dev);
2035 struct netfront_info *info = netdev_priv(netdev);
2037 return sprintf(buf, "%u\n", info->rx_target);
2040 static struct device_attribute xennet_attrs[] = {
2041 __ATTR(rxbuf_min, S_IRUGO|S_IWUSR, show_rxbuf_min, store_rxbuf_min),
2042 __ATTR(rxbuf_max, S_IRUGO|S_IWUSR, show_rxbuf_max, store_rxbuf_max),
2043 __ATTR(rxbuf_cur, S_IRUGO, show_rxbuf_cur, NULL),
2046 static int xennet_sysfs_addif(struct net_device *netdev)
2051 for (i = 0; i < ARRAY_SIZE(xennet_attrs); i++) {
2052 err = device_create_file(&netdev->dev,
2061 device_remove_file(&netdev->dev, &xennet_attrs[i]);
2065 static void xennet_sysfs_delif(struct net_device *netdev)
2069 for (i = 0; i < ARRAY_SIZE(xennet_attrs); i++)
2070 device_remove_file(&netdev->dev, &xennet_attrs[i]);
2073 #endif /* CONFIG_SYSFS */
2075 static const struct xenbus_device_id netfront_ids[] = {
2081 static int xennet_remove(struct xenbus_device *dev)
2083 struct netfront_info *info = dev_get_drvdata(&dev->dev);
2085 dev_dbg(&dev->dev, "%s\n", dev->nodename);
2087 xennet_disconnect_backend(info);
2089 xennet_sysfs_delif(info->netdev);
2091 unregister_netdev(info->netdev);
2093 del_timer_sync(&info->rx_refill_timer);
2095 free_percpu(info->stats);
2097 free_netdev(info->netdev);
2102 static DEFINE_XENBUS_DRIVER(netfront, ,
2103 .probe = netfront_probe,
2104 .remove = xennet_remove,
2105 .resume = netfront_resume,
2106 .otherend_changed = netback_changed,
2109 static int __init netif_init(void)
2114 if (!xen_has_pv_nic_devices())
2117 pr_info("Initialising Xen virtual ethernet driver\n");
2119 return xenbus_register_frontend(&netfront_driver);
2121 module_init(netif_init);
2124 static void __exit netif_exit(void)
2126 xenbus_unregister_driver(&netfront_driver);
2128 module_exit(netif_exit);
2130 MODULE_DESCRIPTION("Xen virtual network device frontend");
2131 MODULE_LICENSE("GPL");
2132 MODULE_ALIAS("xen:vif");
2133 MODULE_ALIAS("xennet");