From 7fbb9d8415d4a51cf542e87cf3a717a9f7e6aedc Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Tue, 24 Feb 2015 11:17:59 +0000 Subject: xen-netback: release pending index before pushing Tx responses If the pending indexes are released /after/ pushing the Tx response then a stale pending index may be used if a new Tx request is immediately pushed by the frontend. The may cause various WARNINGs or BUGs if the stale pending index is actually still in use. Fix this by releasing the pending index before pushing the Tx response. The full barrier for the pending ring update is not required since the the Tx response push already has a suitable write barrier. Signed-off-by: David Vrabel Reviewed-by: Wei Liu Signed-off-by: David S. Miller --- drivers/net/xen-netback/netback.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-) (limited to 'drivers/net/xen-netback/netback.c') diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index f7a31d2cb3f1..c4d68d768408 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -655,9 +655,15 @@ static void xenvif_tx_err(struct xenvif_queue *queue, unsigned long flags; do { + int notify; + spin_lock_irqsave(&queue->response_lock, flags); make_tx_response(queue, txp, XEN_NETIF_RSP_ERROR); + RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); spin_unlock_irqrestore(&queue->response_lock, flags); + if (notify) + notify_remote_via_irq(queue->tx_irq); + if (cons == end) break; txp = RING_GET_REQUEST(&queue->tx, cons++); @@ -1649,17 +1655,28 @@ static void xenvif_idx_release(struct xenvif_queue *queue, u16 pending_idx, { struct pending_tx_info *pending_tx_info; pending_ring_idx_t index; + int notify; unsigned long flags; pending_tx_info = &queue->pending_tx_info[pending_idx]; + spin_lock_irqsave(&queue->response_lock, flags); + make_tx_response(queue, &pending_tx_info->req, status); - index = pending_index(queue->pending_prod); + + /* Release the pending index before pusing the Tx response so + * its available before a new Tx request is pushed by the + * frontend. + */ + index = pending_index(queue->pending_prod++); queue->pending_ring[index] = pending_idx; - /* TX shouldn't use the index before we give it back here */ - mb(); - queue->pending_prod++; + + RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); + spin_unlock_irqrestore(&queue->response_lock, flags); + + if (notify) + notify_remote_via_irq(queue->tx_irq); } @@ -1669,7 +1686,6 @@ static void make_tx_response(struct xenvif_queue *queue, { RING_IDX i = queue->tx.rsp_prod_pvt; struct xen_netif_tx_response *resp; - int notify; resp = RING_GET_RESPONSE(&queue->tx, i); resp->id = txp->id; @@ -1679,9 +1695,6 @@ static void make_tx_response(struct xenvif_queue *queue, RING_GET_RESPONSE(&queue->tx, ++i)->status = XEN_NETIF_RSP_NULL; queue->tx.rsp_prod_pvt = ++i; - RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); - if (notify) - notify_remote_via_irq(queue->tx_irq); } static struct xen_netif_rx_response *make_rx_response(struct xenvif_queue *queue, -- cgit v1.2.3 From 49d9991a18f9aae7b14abbd9c1cc87555330a769 Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Wed, 4 Mar 2015 11:14:47 +0000 Subject: xen-netback: unref frags when handling a from-guest skb with a frag list Every time a VIF is destroyed up to 256 pages may be leaked if packets with more than MAX_SKB_FRAGS frags were transmitted from the guest. Even worse, if another user of ballooned pages allocated one of these ballooned pages it would not handle the unexpectedly >1 page count (e.g., gntdev would deadlock when unmapping a grant because the page count would never reach 1). When handling a from-guest skb with a frag list, unref the frags before releasing them so they are freed correctly when the VIF is destroyed. Signed-off-by: David Vrabel Signed-off-by: David S. Miller --- drivers/net/xen-netback/netback.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'drivers/net/xen-netback/netback.c') diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index c4d68d768408..f1d84fb1eba8 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -1349,7 +1349,7 @@ static int xenvif_handle_frag_list(struct xenvif_queue *queue, struct sk_buff *s { unsigned int offset = skb_headlen(skb); skb_frag_t frags[MAX_SKB_FRAGS]; - int i; + int i, f; struct ubuf_info *uarg; struct sk_buff *nskb = skb_shinfo(skb)->frag_list; @@ -1389,6 +1389,11 @@ static int xenvif_handle_frag_list(struct xenvif_queue *queue, struct sk_buff *s frags[i].page_offset = 0; skb_frag_size_set(&frags[i], len); } + + /* Release all the original (foreign) frags. */ + for (f = 0; f < skb_shinfo(skb)->nr_frags; f++) + skb_frag_unref(skb, f); + /* swap out with old one */ memcpy(skb_shinfo(skb)->frags, frags, -- cgit v1.2.3 From b0c21badf174eb00160f842398f3918d7b365853 Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Wed, 4 Mar 2015 11:14:48 +0000 Subject: xen-netback: refactor xenvif_handle_frag_list() When handling a from-guest frag list, xenvif_handle_frag_list() replaces the frags before calling the destructor to clean up the original (foreign) frags. Whilst this is safe (the destructor doesn't actually use the frags), it looks odd. Reorder the function to be less confusing. Signed-off-by: David Vrabel Signed-off-by: David S. Miller --- drivers/net/xen-netback/netback.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) (limited to 'drivers/net/xen-netback/netback.c') diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index f1d84fb1eba8..cab9f5257f57 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -1390,27 +1390,24 @@ static int xenvif_handle_frag_list(struct xenvif_queue *queue, struct sk_buff *s skb_frag_size_set(&frags[i], len); } + /* Copied all the bits from the frag list -- free it. */ + skb_frag_list_init(skb); + xenvif_skb_zerocopy_prepare(queue, nskb); + kfree_skb(nskb); + /* Release all the original (foreign) frags. */ for (f = 0; f < skb_shinfo(skb)->nr_frags; f++) skb_frag_unref(skb, f); - - /* swap out with old one */ - memcpy(skb_shinfo(skb)->frags, - frags, - i * sizeof(skb_frag_t)); - skb_shinfo(skb)->nr_frags = i; - skb->truesize += i * PAGE_SIZE; - - /* remove traces of mapped pages and frag_list */ - skb_frag_list_init(skb); uarg = skb_shinfo(skb)->destructor_arg; /* increase inflight counter to offset decrement in callback */ atomic_inc(&queue->inflight_packets); uarg->callback(uarg, true); skb_shinfo(skb)->destructor_arg = NULL; - xenvif_skb_zerocopy_prepare(queue, nskb); - kfree_skb(nskb); + /* Fill the skb with the new (local) frags. */ + memcpy(skb_shinfo(skb)->frags, frags, i * sizeof(skb_frag_t)); + skb_shinfo(skb)->nr_frags = i; + skb->truesize += i * PAGE_SIZE; return 0; } -- cgit v1.2.3 From c8a4d29988edb0db9ee80669f2e5e21bd9f7e0d0 Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Wed, 11 Mar 2015 15:27:59 +0000 Subject: xen-netback: notify immediately after pushing Tx response. This fixes a performance regression introduced by 7fbb9d8415d4a51cf542e87cf3a717a9f7e6aedc (xen-netback: release pending index before pushing Tx responses) Moving the notify outside of the spin locks means it can be delayed a long time (if the dealloc thread is descheduled or there is an interrupt or softirq). Signed-off-by: David Vrabel Reviewed-by: Zoltan Kiss Acked-by: Wei Liu Signed-off-by: David S. Miller --- drivers/net/xen-netback/netback.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) (limited to 'drivers/net/xen-netback/netback.c') diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index cab9f5257f57..997cf0901ac2 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -96,6 +96,7 @@ static void xenvif_idx_release(struct xenvif_queue *queue, u16 pending_idx, static void make_tx_response(struct xenvif_queue *queue, struct xen_netif_tx_request *txp, s8 st); +static void push_tx_responses(struct xenvif_queue *queue); static inline int tx_work_todo(struct xenvif_queue *queue); @@ -655,15 +656,10 @@ static void xenvif_tx_err(struct xenvif_queue *queue, unsigned long flags; do { - int notify; - spin_lock_irqsave(&queue->response_lock, flags); make_tx_response(queue, txp, XEN_NETIF_RSP_ERROR); - RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); + push_tx_responses(queue); spin_unlock_irqrestore(&queue->response_lock, flags); - if (notify) - notify_remote_via_irq(queue->tx_irq); - if (cons == end) break; txp = RING_GET_REQUEST(&queue->tx, cons++); @@ -1657,7 +1653,6 @@ static void xenvif_idx_release(struct xenvif_queue *queue, u16 pending_idx, { struct pending_tx_info *pending_tx_info; pending_ring_idx_t index; - int notify; unsigned long flags; pending_tx_info = &queue->pending_tx_info[pending_idx]; @@ -1673,12 +1668,9 @@ static void xenvif_idx_release(struct xenvif_queue *queue, u16 pending_idx, index = pending_index(queue->pending_prod++); queue->pending_ring[index] = pending_idx; - RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); + push_tx_responses(queue); spin_unlock_irqrestore(&queue->response_lock, flags); - - if (notify) - notify_remote_via_irq(queue->tx_irq); } @@ -1699,6 +1691,15 @@ static void make_tx_response(struct xenvif_queue *queue, queue->tx.rsp_prod_pvt = ++i; } +static void push_tx_responses(struct xenvif_queue *queue) +{ + int notify; + + RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->tx, notify); + if (notify) + notify_remote_via_irq(queue->tx_irq); +} + static struct xen_netif_rx_response *make_rx_response(struct xenvif_queue *queue, u16 id, s8 st, -- cgit v1.2.3