| Message ID | 20210508163558.3432246-1-arseny.krasnov@kaspersky.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | virtio/vsock: introduce SOCK_SEQPACKET support | expand |
On Sat, May 08, 2021 at 07:35:54PM +0300, Arseny Krasnov wrote: >This adds rest of logic for SEQPACKET: >1) Send SHUTDOWN on socket close for SEQPACKET type. >2) Set SEQPACKET packet type during send. >3) 'seqpacket_allow' flag to virtio transport. Please update this commit message, point 3 is not included anymore in this patch, right? >4) Set 'VIRTIO_VSOCK_SEQ_EOR' bit in flags for last > packet of message. > >Signed-off-by: Arseny Krasnov <arseny.krasnov@kaspersky.com> >--- > v8 -> v9: > 1) Use cpu_to_le32() to set VIRTIO_VSOCK_SEQ_EOR. > > include/linux/virtio_vsock.h | 4 ++++ > net/vmw_vsock/virtio_transport_common.c | 17 +++++++++++++++-- > 2 files changed, 19 insertions(+), 2 deletions(-) > >diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h >index 02acf6e9ae04..7360ab7ea0af 100644 >--- a/include/linux/virtio_vsock.h >+++ b/include/linux/virtio_vsock.h >@@ -80,6 +80,10 @@ virtio_transport_dgram_dequeue(struct vsock_sock *vsk, > struct msghdr *msg, > size_t len, int flags); > >+int >+virtio_transport_seqpacket_enqueue(struct vsock_sock *vsk, >+ struct msghdr *msg, >+ size_t len); > ssize_t > virtio_transport_seqpacket_dequeue(struct vsock_sock *vsk, > struct msghdr *msg, >diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c >index 7fea0a2192f7..b6608b4ac7c2 100644 >--- a/net/vmw_vsock/virtio_transport_common.c >+++ b/net/vmw_vsock/virtio_transport_common.c >@@ -74,6 +74,10 @@ virtio_transport_alloc_pkt(struct virtio_vsock_pkt_info *info, > err = memcpy_from_msg(pkt->buf, info->msg, len); > if (err) > goto out; >+ >+ if (info->msg->msg_iter.count == 0) Also here is better `msg_data_left(info->msg)` >+ pkt->hdr.flags = cpu_to_le32(info->flags | >+ VIRTIO_VSOCK_SEQ_EOR); Re-thinking an alternative could be to set EOR here... info->flags |= VIRTIO_VSOCK_SEQ_EOR; > } ... and move pkt->hdr.flags assignment after this block: pkt->hdr.flags = cpu_to_le32(info->flags); But I don't have a strong opinion on that. > > trace_virtio_transport_alloc_pkt(src_cid, src_port,
sdf On Thu, May 13, 2021 at 2:27 PM Stefano Garzarella <sgarzare@redhat.com> wrote: > > On Sat, May 08, 2021 at 07:35:54PM +0300, Arseny Krasnov wrote: > >This adds rest of logic for SEQPACKET: > >1) Send SHUTDOWN on socket close for SEQPACKET type. > >2) Set SEQPACKET packet type during send. > >3) 'seqpacket_allow' flag to virtio transport. > > Please update this commit message, point 3 is not included anymore in > this patch, right? > > >4) Set 'VIRTIO_VSOCK_SEQ_EOR' bit in flags for last > > packet of message. > > > >Signed-off-by: Arseny Krasnov <arseny.krasnov@kaspersky.com> > >--- > > v8 -> v9: > > 1) Use cpu_to_le32() to set VIRTIO_VSOCK_SEQ_EOR. > > > > include/linux/virtio_vsock.h | 4 ++++ > > net/vmw_vsock/virtio_transport_common.c | 17 +++++++++++++++-- > > 2 files changed, 19 insertions(+), 2 deletions(-) > > > >diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h > >index 02acf6e9ae04..7360ab7ea0af 100644 > >--- a/include/linux/virtio_vsock.h > >+++ b/include/linux/virtio_vsock.h > >@@ -80,6 +80,10 @@ virtio_transport_dgram_dequeue(struct vsock_sock *vsk, > > struct msghdr *msg, > > size_t len, int flags); > > > >+int > >+virtio_transport_seqpacket_enqueue(struct vsock_sock *vsk, > >+ struct msghdr *msg, > >+ size_t len); > > ssize_t > > virtio_transport_seqpacket_dequeue(struct vsock_sock *vsk, > > struct msghdr *msg, > >diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c > >index 7fea0a2192f7..b6608b4ac7c2 100644 > >--- a/net/vmw_vsock/virtio_transport_common.c > >+++ b/net/vmw_vsock/virtio_transport_common.c > >@@ -74,6 +74,10 @@ virtio_transport_alloc_pkt(struct virtio_vsock_pkt_info *info, > > err = memcpy_from_msg(pkt->buf, info->msg, len); > > if (err) > > goto out; > >+ > >+ if (info->msg->msg_iter.count == 0) > > Also here is better `msg_data_left(info->msg)` > > >+ pkt->hdr.flags = cpu_to_le32(info->flags | > >+ VIRTIO_VSOCK_SEQ_EOR); > > Re-thinking an alternative could be to set EOR here... > > info->flags |= VIRTIO_VSOCK_SEQ_EOR; Or just `pkt->hdr.flags |= cpu_to_le32(VIRTIO_VSOCK_SEQ_EOR)`, as you did in vhost-vsock :-)
diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h index 02acf6e9ae04..7360ab7ea0af 100644 --- a/include/linux/virtio_vsock.h +++ b/include/linux/virtio_vsock.h @@ -80,6 +80,10 @@ virtio_transport_dgram_dequeue(struct vsock_sock *vsk, struct msghdr *msg, size_t len, int flags); +int +virtio_transport_seqpacket_enqueue(struct vsock_sock *vsk, + struct msghdr *msg, + size_t len); ssize_t virtio_transport_seqpacket_dequeue(struct vsock_sock *vsk, struct msghdr *msg, diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c index 7fea0a2192f7..b6608b4ac7c2 100644 --- a/net/vmw_vsock/virtio_transport_common.c +++ b/net/vmw_vsock/virtio_transport_common.c @@ -74,6 +74,10 @@ virtio_transport_alloc_pkt(struct virtio_vsock_pkt_info *info, err = memcpy_from_msg(pkt->buf, info->msg, len); if (err) goto out; + + if (info->msg->msg_iter.count == 0) + pkt->hdr.flags = cpu_to_le32(info->flags | + VIRTIO_VSOCK_SEQ_EOR); } trace_virtio_transport_alloc_pkt(src_cid, src_port, @@ -187,7 +191,7 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk, struct virtio_vsock_pkt *pkt; u32 pkt_len = info->pkt_len; - info->type = VIRTIO_VSOCK_TYPE_STREAM; + info->type = virtio_transport_get_type(sk_vsock(vsk)); t_ops = virtio_transport_get_ops(vsk); if (unlikely(!t_ops)) @@ -477,6 +481,15 @@ virtio_transport_seqpacket_dequeue(struct vsock_sock *vsk, } EXPORT_SYMBOL_GPL(virtio_transport_seqpacket_dequeue); +int +virtio_transport_seqpacket_enqueue(struct vsock_sock *vsk, + struct msghdr *msg, + size_t len) +{ + return virtio_transport_stream_enqueue(vsk, msg, len); +} +EXPORT_SYMBOL_GPL(virtio_transport_seqpacket_enqueue); + int virtio_transport_dgram_dequeue(struct vsock_sock *vsk, struct msghdr *msg, @@ -911,7 +924,7 @@ void virtio_transport_release(struct vsock_sock *vsk) struct sock *sk = &vsk->sk; bool remove_sock = true; - if (sk->sk_type == SOCK_STREAM) + if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) remove_sock = virtio_transport_close(vsk); if (remove_sock) {
This adds rest of logic for SEQPACKET: 1) Send SHUTDOWN on socket close for SEQPACKET type. 2) Set SEQPACKET packet type during send. 3) 'seqpacket_allow' flag to virtio transport. 4) Set 'VIRTIO_VSOCK_SEQ_EOR' bit in flags for last packet of message. Signed-off-by: Arseny Krasnov <arseny.krasnov@kaspersky.com> --- v8 -> v9: 1) Use cpu_to_le32() to set VIRTIO_VSOCK_SEQ_EOR. include/linux/virtio_vsock.h | 4 ++++ net/vmw_vsock/virtio_transport_common.c | 17 +++++++++++++++-- 2 files changed, 19 insertions(+), 2 deletions(-)