net/9p/virtio: Fix hard lockup in req_done

When client has multiple threads that issue io requests
all the time, and the server has a very good performance,
it may cause cpu is running in the irq context for a long
time because it can check virtqueue has buf in the *while*
loop.

So we should keep chan->lock in the whole loop.

[ Dominique: reworded subject line ]
Link: http://lkml.kernel.org/r/5B503AEC.5080404@huawei.com
Signed-off-by: Yiwen Jiang <jiangyiwen@huawei.com>
To: Andrew Morton <akpm@linux-foundation.org>
To: Eric Van Hensbergen <ericvh@gmail.com>
To: Ron Minnich <rminnich@sandia.gov>
To: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
This commit is contained in:
jiangyiwen 2018-07-19 15:17:00 +08:00 committed by Dominique Martinet
parent c7ebbae7cf
commit 31934da810

View File

@ -144,24 +144,25 @@ static void req_done(struct virtqueue *vq)
struct virtio_chan *chan = vq->vdev->priv;
unsigned int len;
struct p9_req_t *req;
bool need_wakeup = false;
unsigned long flags;
p9_debug(P9_DEBUG_TRANS, ": request done\n");
while (1) {
spin_lock_irqsave(&chan->lock, flags);
req = virtqueue_get_buf(chan->vq, &len);
if (req == NULL) {
spin_unlock_irqrestore(&chan->lock, flags);
break;
spin_lock_irqsave(&chan->lock, flags);
while ((req = virtqueue_get_buf(chan->vq, &len)) != NULL) {
if (!chan->ring_bufs_avail) {
chan->ring_bufs_avail = 1;
need_wakeup = true;
}
chan->ring_bufs_avail = 1;
spin_unlock_irqrestore(&chan->lock, flags);
/* Wakeup if anyone waiting for VirtIO ring space. */
wake_up(chan->vc_wq);
if (len)
p9_client_cb(chan->client, req, REQ_STATUS_RCVD);
}
spin_unlock_irqrestore(&chan->lock, flags);
/* Wakeup if anyone waiting for VirtIO ring space. */
if (need_wakeup)
wake_up(chan->vc_wq);
}
/**