Re: [PATCH v5 08/13] xen/pvcalls: implement accept command

[Stefano Stabellini]

On Tue, 17 Oct 2017, Boris Ostrovsky wrote:
> On 10/06/2017 08:30 PM, Stefano Stabellini wrote:
> > Introduce a waitqueue to allow only one outstanding accept command at
> > any given time and to implement polling on the passive socket. Introduce
> > a flags field to keep track of in-flight accept and poll commands.
> > 
> > Send PVCALLS_ACCEPT to the backend. Allocate a new active socket. Make
> > sure that only one accept command is executed at any given time by
> > setting PVCALLS_FLAG_ACCEPT_INFLIGHT and waiting on the
> > inflight_accept_req waitqueue.
> > 
> > Convert the new struct sock_mapping pointer into an uint64_t and use it
> > as id for the new socket to pass to the backend.
> > 
> > Check if the accept call is non-blocking: in that case after sending the
> > ACCEPT command to the backend store the sock_mapping pointer of the new
> > struct and the inflight req_id then return -EAGAIN (which will respond
> > only when there is something to accept). Next time accept is called,
> > we'll check if the ACCEPT command has been answered, if so we'll pick up
> > where we left off, otherwise we return -EAGAIN again.
> > 
> > Note that, differently from the other commands, we can use
> > wait_event_interruptible (instead of wait_event) in the case of accept
> > as we are able to track the req_id of the ACCEPT response that we are
> > waiting.
> > 
> > Signed-off-by: Stefano Stabellini <stefano@xxxxxxxxxxx>
> > CC: boris.ostrovsky@xxxxxxxxxx
> > CC: jgross@xxxxxxxx
> > ---
> >  drivers/xen/pvcalls-front.c | 146 ++++++++++++++++++++++++++++++++++++++++++++
> >  drivers/xen/pvcalls-front.h |   3 +
> >  2 files changed, 149 insertions(+)
> > 
> > diff --git a/drivers/xen/pvcalls-front.c b/drivers/xen/pvcalls-front.c
> > index 5433fae..8958e74 100644
> > --- a/drivers/xen/pvcalls-front.c
> > +++ b/drivers/xen/pvcalls-front.c
> > @@ -77,6 +77,16 @@ struct sock_mapping {
> >  #define PVCALLS_STATUS_BIND          1
> >  #define PVCALLS_STATUS_LISTEN        2
> >  			uint8_t status;
> > +		/*
> > +		 * Internal state-machine flags.
> > +		 * Only one accept operation can be inflight for a socket.
> > +		 * Only one poll operation can be inflight for a given socket.
> > +		 */
> > +#define PVCALLS_FLAG_ACCEPT_INFLIGHT 0
> > +			uint8_t flags;
> > +			uint32_t inflight_req_id;
> > +			struct sock_mapping *accept_map;
> > +			wait_queue_head_t inflight_accept_req;
> >  		} passive;
> >  	};
> >  };
> > @@ -392,6 +402,8 @@ int pvcalls_front_bind(struct socket *sock, struct sockaddr *addr, int addr_len)
> >  	memcpy(req->u.bind.addr, addr, sizeof(*addr));
> >  	req->u.bind.len = addr_len;
> >  
> > +	init_waitqueue_head(&map->passive.inflight_accept_req);
> > +
> >  	map->active_socket = false;
> >  
> >  	bedata->ring.req_prod_pvt++;
> > @@ -470,6 +482,140 @@ int pvcalls_front_listen(struct socket *sock, int backlog)
> >  	return ret;
> >  }
> >  
> > +int pvcalls_front_accept(struct socket *sock, struct socket *newsock, int flags)
> > +{
> > +	struct pvcalls_bedata *bedata;
> > +	struct sock_mapping *map;
> > +	struct sock_mapping *map2 = NULL;
> > +	struct xen_pvcalls_request *req;
> > +	int notify, req_id, ret, evtchn, nonblock;
> > +
> > +	pvcalls_enter();
> > +	if (!pvcalls_front_dev) {
> > +		pvcalls_exit();
> > +		return -ENOTCONN;
> > +	}
> > +	bedata = dev_get_drvdata(&pvcalls_front_dev->dev);
> > +
> > +	map = (struct sock_mapping *) sock->sk->sk_send_head;
> > +	if (!map) {
> > +		pvcalls_exit();
> > +		return -ENOTSOCK;
> > +	}
> > +
> > +	if (map->passive.status != PVCALLS_STATUS_LISTEN) {
> > +		pvcalls_exit();
> > +		return -EINVAL;
> > +	}
> > +
> > +	nonblock = flags & SOCK_NONBLOCK;
> > +	/*
> > +	 * Backend only supports 1 inflight accept request, will return
> > +	 * errors for the others
> > +	 */
> > +	if (test_and_set_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +			     (void *)&map->passive.flags)) {
> > +		req_id = READ_ONCE(map->passive.inflight_req_id);
> > +		if (req_id != PVCALLS_INVALID_ID &&
> > +		    READ_ONCE(bedata->rsp[req_id].req_id) == req_id) {
> 
> 
> READ_ONCE (especially the second one)? I know I may sound fixated on
> this but I really don't understand how compiler may do anything wrong if
> straight reads were used.
> 
> For the first case, I guess, theoretically the compiler may decide to
> re-fetch map->passive.inflight_req_id. But even if it did, would that be
> a problem? Both of these READ_ONCE targets are updated below before
> PVCALLS_FLAG_ACCEPT_INFLIGHT is cleared so there should not be any
> change between re-fetching, I think. (The only exception is the noblock
> case, which does WRITE_ONCE that don't understand either)

READ_ONCE is reasonably cheap: do we really want to have this kind of
conversation every time we touch this code in the future? Personally, I
would have used READ/WRITE_ONCE everywhere for inflight_req_id and
req_id, because it makes the code easier to understand. 

We have already limited their usage, but at least we have followed a set
of guidelines. Doing further optimizations on this code seems
unnecessary and prone to confuse the reader.


> > +			map2 = map->passive.accept_map;
> > +			goto received;
> > +		}
> > +		if (nonblock) {
> > +			pvcalls_exit();
> > +			return -EAGAIN;
> > +		}
> > +		if (wait_event_interruptible(map->passive.inflight_accept_req,
> > +			!test_and_set_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +					  (void *)&map->passive.flags))) {
> > +			pvcalls_exit();
> > +			return -EINTR;
> > +		}
> > +	}
> > +
> > +	spin_lock(&bedata->socket_lock);
> > +	ret = get_request(bedata, &req_id);
> > +	if (ret < 0) {
> > +		clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +			  (void *)&map->passive.flags);
> > +		spin_unlock(&bedata->socket_lock);
> > +		pvcalls_exit();
> > +		return ret;
> > +	}
> > +	map2 = kzalloc(sizeof(*map2), GFP_KERNEL);
> > +	if (map2 == NULL) {
> > +		clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +			  (void *)&map->passive.flags);
> > +		spin_unlock(&bedata->socket_lock);
> > +		pvcalls_exit();
> > +		return -ENOMEM;
> > +	}
> > +	ret =  create_active(map2, &evtchn);
> > +	if (ret < 0) {
> > +		kfree(map2);
> > +		clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +			  (void *)&map->passive.flags);
> > +		spin_unlock(&bedata->socket_lock);
> > +		pvcalls_exit();
> > +		return -ENOMEM;
> 
> Why not ret?

yes, good idea.


> 
> > +	}
> > +	list_add_tail(&map2->list, &bedata->socket_mappings);
> > +
> > +	req = RING_GET_REQUEST(&bedata->ring, req_id);
> > +	req->req_id = req_id;
> > +	req->cmd = PVCALLS_ACCEPT;
> > +	req->u.accept.id = (uint64_t) map;
> > +	req->u.accept.ref = map2->active.ref;
> > +	req->u.accept.id_new = (uint64_t) map2;
> > +	req->u.accept.evtchn = evtchn;
> > +	map->passive.accept_map = map2;
> > +
> > +	bedata->ring.req_prod_pvt++;
> > +	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&bedata->ring, notify);
> > +	spin_unlock(&bedata->socket_lock);
> > +	if (notify)
> > +		notify_remote_via_irq(bedata->irq);
> > +	/* We could check if we have received a response before returning. */
> > +	if (nonblock) {
> > +		WRITE_ONCE(map->passive.inflight_req_id, req_id);
> > +		pvcalls_exit();
> > +		return -EAGAIN;
> > +	}
> > +
> > +	if (wait_event_interruptible(bedata->inflight_req,
> > +		READ_ONCE(bedata->rsp[req_id].req_id) == req_id)) {
> > +		pvcalls_exit();
> > +		return -EINTR;
> > +	}
> > +	/* read req_id, then the content */
> > +	smp_rmb();
> > +
> > +received:
> > +	map2->sock = newsock;
> > +	newsock->sk = kzalloc(sizeof(*newsock->sk), GFP_KERNEL);
> > +	if (!newsock->sk) {
> > +		bedata->rsp[req_id].req_id = PVCALLS_INVALID_ID;
> > +		map->passive.inflight_req_id = PVCALLS_INVALID_ID;
> > +		clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
> > +			  (void *)&map->passive.flags);
> > +		pvcalls_front_free_map(bedata, map2);
> > +		kfree(map2);
> > +		pvcalls_exit();
> > +		return -ENOMEM;
> > +	}
> > +	newsock->sk->sk_send_head = (void *)map2;
> > +
> > +	ret = bedata->rsp[req_id].ret;
> > +	bedata->rsp[req_id].req_id = PVCALLS_INVALID_ID;
> > +	map->passive.inflight_req_id = PVCALLS_INVALID_ID;
> > +
> > +	clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT, (void *)&map->passive.flags);
> > +	wake_up(&map->passive.inflight_accept_req);
> > +
> > +	pvcalls_exit();
> > +	return ret;
> > +}
> > +
> >  static const struct xenbus_device_id pvcalls_front_ids[] = {
> >  	{ "pvcalls" },
> >  	{ "" }
> > diff --git a/drivers/xen/pvcalls-front.h b/drivers/xen/pvcalls-front.h
> > index aa8fe10..ab4f1da 100644
> > --- a/drivers/xen/pvcalls-front.h
> > +++ b/drivers/xen/pvcalls-front.h
> > @@ -10,5 +10,8 @@ int pvcalls_front_bind(struct socket *sock,
> >  		       struct sockaddr *addr,
> >  		       int addr_len);
> >  int pvcalls_front_listen(struct socket *sock, int backlog);
> > +int pvcalls_front_accept(struct socket *sock,
> > +			 struct socket *newsock,
> > +			 int flags);
> >  
> >  #endif
> > 
>