On Wed, 2011-02-23 at 19:19 +0100, Marco d'Itri wrote: > On Apr 28, Marco d'Itri <md@linux.it> wrote: > > > On some of my systems, some time after boot (from about 1 to 8 days) > > the kernel stops emitting packets (?). Using tcpdump on eth0 on the > > guest I could see incoming ARP requests from the host but no answers. > > No errors have been logged by the kernel and removing and loading again > > virtio_net fixed the issue. > We later determined that this is reproducible by opening a few hundreds > of parallel TCP connections to the guest. This is very unlikely to be fixed in lenny now. Please try the kernel package from squeeze with the addition of the attached patches. (These patches fix different bugs which can also cause virtio_net to stop passing traffic. They are not applicable to Linux 2.6.26.) Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse.
From: Bruce Rogers <brogers@novell.com>
Date: Thu, 10 Feb 2011 11:03:31 -0800
Subject: [PATCH] virtio_net: Add schedule check to napi_enable call
commit 3e9d08ec0a68f6faf718d5a7e050fe5ca0ba004f upstream.
Under harsh testing conditions, including low memory, the guest would
stop receiving packets. With this patch applied we no longer see any
problems in the driver while performing these tests for extended periods
of time.
Make sure napi is scheduled subsequent to each napi_enable.
Signed-off-by: Bruce Rogers <brogers@novell.com>
Signed-off-by: Olaf Kirch <okir@suse.de>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[bwh: Adjust for 2.6.32]
---
drivers/net/virtio_net.c | 27 ++++++++++++++++-----------
1 files changed, 16 insertions(+), 11 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index b6d4028..c1689ba 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -446,6 +446,20 @@ static void skb_recv_done(struct virtqueue *rvq)
}
}
+static void virtnet_napi_enable(struct virtnet_info *vi)
+{
+ napi_enable(&vi->napi);
+
+ /* If all buffers were filled by other side before we napi_enabled, we
+ * won't get another interrupt, so process any outstanding packets
+ * now. virtnet_poll wants re-enable the queue, so we disable here.
+ * We synchronize against interrupts via NAPI_STATE_SCHED */
+ if (napi_schedule_prep(&vi->napi)) {
+ vi->rvq->vq_ops->disable_cb(vi->rvq);
+ __napi_schedule(&vi->napi);
+ }
+}
+
static void refill_work(struct work_struct *work)
{
struct virtnet_info *vi;
@@ -454,7 +468,7 @@ static void refill_work(struct work_struct *work)
vi = container_of(work, struct virtnet_info, refill.work);
napi_disable(&vi->napi);
still_empty = !try_fill_recv(vi, GFP_KERNEL);
- napi_enable(&vi->napi);
+ virtnet_napi_enable(vi);
/* In theory, this can happen: if we don't get any buffers in
* we will *never* try to fill again. */
@@ -638,16 +652,7 @@ static int virtnet_open(struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
- napi_enable(&vi->napi);
-
- /* If all buffers were filled by other side before we napi_enabled, we
- * won't get another interrupt, so process any outstanding packets
- * now. virtnet_poll wants re-enable the queue, so we disable here.
- * We synchronize against interrupts via NAPI_STATE_SCHED */
- if (napi_schedule_prep(&vi->napi)) {
- vi->rvq->vq_ops->disable_cb(vi->rvq);
- __napi_schedule(&vi->napi);
- }
+ virtnet_napi_enable(vi);
return 0;
}
--
1.7.4.1
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Fri, 2 Jul 2010 16:34:01 +0000
Subject: [PATCH] virtio_net: fix oom handling on tx
commit 58eba97d0774c69b1cf3e5a8ac74419409d1abbf upstream.
virtio net will never try to overflow the TX ring, so the only reason
add_buf may fail is out of memory. Thus, we can not stop the
device until some request completes - there's no guarantee anything
at all is outstanding.
Make the error message clearer as well: error here does not
indicate queue full.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (...and avoid TX_BUSY)
Cc: stable@kernel.org # .34.x (s/virtqueue_/vi->svq->vq_ops->/)
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/virtio_net.c | 21 +++++++++++++--------
1 files changed, 13 insertions(+), 8 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index ee75711..bb6b67f 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -562,7 +562,6 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
struct virtnet_info *vi = netdev_priv(dev);
int capacity;
-again:
/* Free up any pending old buffers before queueing new ones. */
free_old_xmit_skbs(vi);
@@ -571,14 +570,20 @@ again:
/* This can happen with OOM and indirect buffers. */
if (unlikely(capacity < 0)) {
- netif_stop_queue(dev);
- dev_warn(&dev->dev, "Unexpected full queue\n");
- if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
- vi->svq->vq_ops->disable_cb(vi->svq);
- netif_start_queue(dev);
- goto again;
+ if (net_ratelimit()) {
+ if (likely(capacity == -ENOMEM)) {
+ dev_warn(&dev->dev,
+ "TX queue failure: out of memory\n");
+ } else {
+ dev->stats.tx_fifo_errors++;
+ dev_warn(&dev->dev,
+ "Unexpected TX queue failure: %d\n",
+ capacity);
+ }
}
- return NETDEV_TX_BUSY;
+ dev->stats.tx_dropped++;
+ kfree_skb(skb);
+ return NETDEV_TX_OK;
}
vi->svq->vq_ops->kick(vi->svq);
--
1.7.4.1
Attachment:
signature.asc
Description: This is a digitally signed message part