Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1697571
| From | Boris Ostrovsky <boris.ostrovsky@oracle.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [PATCH v2 09/13] xen/pvcalls: implement recvmsg |
| Date | 2017-07-26 23:30 +0200 |
| Message-ID | <u7C2u-5CF-17@gated-at.bofh.it> (permalink) |
| References | <u7fyV-8aW-3@gated-at.bofh.it> <u7fyV-8aW-5@gated-at.bofh.it> <u7fyX-8aW-59@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On 07/25/2017 05:22 PM, Stefano Stabellini wrote:
> Implement recvmsg by copying data from the "in" ring. If not enough data
> is available and the recvmsg call is blocking, then wait on the
> inflight_conn_req waitqueue. Take the active socket in_mutex so that
> only one function can access the ring at any given time.
>
> If not enough data is available on the ring, rather than returning
> immediately or sleep-waiting, spin for up to 5000 cycles. This small
> optimization turns out to improve performance and latency significantly.
>
> Signed-off-by: Stefano Stabellini <stefano@aporeto.com>
> CC: boris.ostrovsky@oracle.com
> CC: jgross@suse.com
> ---
> drivers/xen/pvcalls-front.c | 106 ++++++++++++++++++++++++++++++++++++++++++++
> drivers/xen/pvcalls-front.h | 4 ++
> 2 files changed, 110 insertions(+)
>
> diff --git a/drivers/xen/pvcalls-front.c b/drivers/xen/pvcalls-front.c
> index d8ed280..b4ca569 100644
> --- a/drivers/xen/pvcalls-front.c
> +++ b/drivers/xen/pvcalls-front.c
> @@ -96,6 +96,20 @@ static int pvcalls_front_write_todo(struct sock_mapping *map)
> return size - pvcalls_queued(prod, cons, size);
> }
>
> +static bool pvcalls_front_read_todo(struct sock_mapping *map)
> +{
> + struct pvcalls_data_intf *intf = map->active.ring;
> + RING_IDX cons, prod;
> + int32_t error;
> +
> + cons = intf->in_cons;
> + prod = intf->in_prod;
> + error = intf->in_error;
> + return (error != 0 ||
> + pvcalls_queued(prod, cons,
> + XEN_FLEX_RING_SIZE(intf->ring_order))) != 0;
> +}
> +
> static irqreturn_t pvcalls_front_event_handler(int irq, void *dev_id)
> {
> struct xenbus_device *dev = dev_id;
> @@ -418,6 +432,98 @@ int pvcalls_front_sendmsg(struct socket *sock, struct msghdr *msg,
> return tot_sent;
> }
>
> +static int __read_ring(struct pvcalls_data_intf *intf,
> + struct pvcalls_data *data,
> + struct iov_iter *msg_iter,
> + size_t len, int flags)
> +{
> + RING_IDX cons, prod, size, masked_prod, masked_cons;
> + RING_IDX array_size = XEN_FLEX_RING_SIZE(intf->ring_order);
> + int32_t error;
> +
> + cons = intf->in_cons;
> + prod = intf->in_prod;
> + error = intf->in_error;
> + /* get pointers before reading from the ring */
> + virt_rmb();
> + if (error < 0)
> + return error;
> +
> + size = pvcalls_queued(prod, cons, array_size);
> + masked_prod = pvcalls_mask(prod, array_size);
> + masked_cons = pvcalls_mask(cons, array_size);
> +
> + if (size == 0)
> + return 0;
> +
> + if (len > size)
> + len = size;
> +
> + if (masked_prod > masked_cons) {
> + copy_to_iter(data->in + masked_cons, len, msg_iter);
> + } else {
> + if (len > (array_size - masked_cons)) {
> + copy_to_iter(data->in + masked_cons,
> + array_size - masked_cons, msg_iter);
> + copy_to_iter(data->in,
> + len - (array_size - masked_cons),
> + msg_iter);
> + } else {
> + copy_to_iter(data->in + masked_cons, len, msg_iter);
> + }
> + }
> + /* read data from the ring before increasing the index */
> + virt_mb();
> + if (!(flags & MSG_PEEK))
> + intf->in_cons += len;
> +
> + return len;
> +}
> +
> +int pvcalls_front_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
> + int flags)
> +{
> + struct pvcalls_bedata *bedata;
> + int ret = -EAGAIN;
> + struct sock_mapping *map;
> + int count = 0;
> +
> + if (!pvcalls_front_dev)
> + return -ENOTCONN;
> + bedata = dev_get_drvdata(&pvcalls_front_dev->dev);
> +
> + map = (struct sock_mapping *) READ_ONCE(sock->sk->sk_send_head);
> + if (!map)
> + return -ENOTSOCK;
> +
> + if (flags & (MSG_CMSG_CLOEXEC|MSG_ERRQUEUE|MSG_OOB|MSG_TRUNC))
> + return -EOPNOTSUPP;
> +
> + mutex_lock(&map->active.in_mutex);
> + if (len > XEN_FLEX_RING_SIZE(map->active.ring->ring_order))
> + len = XEN_FLEX_RING_SIZE(map->active.ring->ring_order);
> +
> + while (!(flags & MSG_DONTWAIT) && !pvcalls_front_read_todo(map)) {
> + if (count < PVCALLS_FRONT_MAX_SPIN)
> + count++;
> + else
> + wait_event_interruptible(map->active.inflight_conn_req,
> + pvcalls_front_read_todo(map));
> + }
Should we be using PVCALLS_FRONT_MAX_SPIN here? In sendmsg it is
counting non-sleeping iterations but here we are sleeping so
PVCALLS_FRONT_MAX_SPIN (5000) may take a while.
In fact, what shouldn't this waiting be a function of MSG_DONTWAIT
and/or socket's O_NONBLOCK?
-boris
> + ret = __read_ring(map->active.ring, &map->active.data,
> + &msg->msg_iter, len, flags);
> +
> + if (ret > 0)
> + notify_remote_via_irq(map->active.irq);
> + if (ret == 0)
> + ret = -EAGAIN;
> + if (ret == -ENOTCONN)
> + ret = 0;
> +
> + mutex_unlock(&map->active.in_mutex);
> + return ret;
> +}
> +
> int pvcalls_front_bind(struct socket *sock, struct sockaddr *addr, int addr_len)
> {
> struct pvcalls_bedata *bedata;
> diff --git a/drivers/xen/pvcalls-front.h b/drivers/xen/pvcalls-front.h
> index d937c24..de24041 100644
> --- a/drivers/xen/pvcalls-front.h
> +++ b/drivers/xen/pvcalls-front.h
> @@ -16,5 +16,9 @@ int pvcalls_front_accept(struct socket *sock,
> int pvcalls_front_sendmsg(struct socket *sock,
> struct msghdr *msg,
> size_t len);
> +int pvcalls_front_recvmsg(struct socket *sock,
> + struct msghdr *msg,
> + size_t len,
> + int flags);
>
> #endif
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH v2 00/13] introduce the Xen PV Calls frontend Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
[PATCH v2 13/13] xen: introduce a Kconfig option to enable the pvcalls frontend Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
[PATCH v2 10/13] xen/pvcalls: implement poll command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 10/13] xen/pvcalls: implement poll command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-27 01:30 +0200
Re: [PATCH v2 10/13] xen/pvcalls: implement poll command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 02:30 +0200
[PATCH v2 06/13] xen/pvcalls: implement listen command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
[PATCH v2 07/13] xen/pvcalls: implement accept command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 07/13] xen/pvcalls: implement accept command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 20:00 +0200
Re: [PATCH v2 07/13] xen/pvcalls: implement accept command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 01:30 +0200
[PATCH v2 01/13] xen/pvcalls: introduce the pvcalls xenbus frontend Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
[PATCH v2 12/13] xen/pvcalls: implement frontend disconnect Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
[PATCH v2 05/13] xen/pvcalls: implement bind command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 05/13] xen/pvcalls: implement bind command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 17:00 +0200
Re: [PATCH v2 05/13] xen/pvcalls: implement bind command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 02:00 +0200
Re: [PATCH v2 05/13] xen/pvcalls: implement bind command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-27 16:50 +0200
[PATCH v2 04/13] xen/pvcalls: implement connect command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 04/13] xen/pvcalls: implement connect command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 17:00 +0200
Re: [PATCH v2 04/13] xen/pvcalls: implement connect command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 01:30 +0200
[PATCH v2 09/13] xen/pvcalls: implement recvmsg Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 09/13] xen/pvcalls: implement recvmsg Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 23:30 +0200
Re: [Xen-devel] [PATCH v2 09/13] xen/pvcalls: implement recvmsg Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 23:40 +0200
Re: [Xen-devel] [PATCH v2 09/13] xen/pvcalls: implement recvmsg Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 02:10 +0200
Re: [Xen-devel] [PATCH v2 09/13] xen/pvcalls: implement recvmsg Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-27 17:00 +0200
[PATCH v2 11/13] xen/pvcalls: implement release command Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 11/13] xen/pvcalls: implement release command Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-27 20:40 +0200
[PATCH v2 02/13] xen/pvcalls: connect to the backend Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 02/13] xen/pvcalls: connect to the backend Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 15:40 +0200
Re: [PATCH v2 02/13] xen/pvcalls: connect to the backend Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 02:30 +0200
Re: [PATCH v2 02/13] xen/pvcalls: connect to the backend Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-27 17:10 +0200
[PATCH v2 03/13] xen/pvcalls: implement socket command and handle events Stefano Stabellini <sstabellini@kernel.org> - 2017-07-25 23:30 +0200
Re: [PATCH v2 03/13] xen/pvcalls: implement socket command and handle events Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-07-26 16:30 +0200
Re: [PATCH v2 03/13] xen/pvcalls: implement socket command and handle events Stefano Stabellini <sstabellini@kernel.org> - 2017-07-27 01:20 +0200
csiph-web