mbox series

[RFC,v2,00/13] virtio/vsock: introduce SOCK_SEQPACKET support.

Message ID 20210115053553.1454517-1-arseny.krasnov@kaspersky.com (mailing list archive)
Headers show
Series virtio/vsock: introduce SOCK_SEQPACKET support. | expand

Message

Arseny Krasnov Jan. 15, 2021, 5:35 a.m. UTC
This patchset impelements support of SOCK_SEQPACKET for virtio
transport.
	As SOCK_SEQPACKET guarantees to save record boundaries, so to
do it, new packet operation was added: it marks start of record (with
record length in header), such packet doesn't carry any data.  To send
record, packet with start marker is sent first, then all data is sent
as usual 'RW' packets. On receiver's side, length of record is known
from packet with start record marker. Now as  packets of one socket
are not reordered neither on vsock nor on vhost transport layers, such
marker allows to restore original record on receiver's side. If user's
buffer is smaller that record length, when all out of size data is
dropped.
	Maximum length of datagram is not limited as in stream socket,
because same credit logic is used. Difference with stream socket is
that user is not woken up until whole record is received or error
occurred. Implementation also supports 'MSG_EOR' and 'MSG_TRUNC' flags.
	Tests also implemented.

 Arseny Krasnov (13):
  af_vsock: implement 'vsock_wait_data()'.
  af_vsock: separate rx loops for STREAM/SEQPACKET.
  af_vsock: implement rx loops entry point
  af_vsock: replace previous stream rx loop.
  af_vsock: implement send logic for SOCK_SEQPACKET
  af_vsock: general support of SOCK_SEQPACKET type.
  af_vsock: update comments for stream sockets.
  virtio/vsock: dequeue callback for SOCK_SEQPACKET.
  virtio/vsock: implement fetch of record length
  virtio/vsock: update receive logic
  virtio/vsock: rest of SOCK_SEQPACKET support
  vhost/vsock: support for SOCK_SEQPACKET socket.
  vsock_test: add SOCK_SEQPACKET tests.

 drivers/vhost/vsock.c                   |   7 +-
 include/linux/virtio_vsock.h            |  12 +
 include/net/af_vsock.h                  |   6 +
 include/uapi/linux/virtio_vsock.h       |   9 +
 net/vmw_vsock/af_vsock.c                | 483 ++++++++++++++++------
 net/vmw_vsock/virtio_transport.c        |   4 +
 net/vmw_vsock/virtio_transport_common.c | 294 +++++++++++--
 tools/testing/vsock/util.c              |  32 +-
 tools/testing/vsock/util.h              |   3 +
 tools/testing/vsock/vsock_test.c        | 126 ++++++
 10 files changed, 824 insertions(+), 152 deletions(-)

 v1 -> v2:
 - patches reordered: af_vsock.c changes now before virtio vsock
 - patches reorganized: more small patches, where +/- are not mixed
 - tests for SOCK_SEQPACKET added
 - all commit messages updated
 - af_vsock.c: 'vsock_pre_recv_check()' inlined to
   'vsock_connectible_recvmsg()'
 - af_vsock.c: 'vsock_assign_transport()' returns ENODEV if transport
   was not found
 - virtio_transport_common.c: transport callback for seqpacket dequeue
 - virtio_transport_common.c: simplified
   'virtio_transport_recv_connected()'
 - virtio_transport_common.c: send reset on socket and packet type
			      mismatch.

Signed-off-by: Arseny Krasnov <arseny.krasnov@kaspersky.com>

Comments

Stas Sergeev Jan. 15, 2021, 9:59 a.m. UTC | #1
15.01.2021 08:35, Arseny Krasnov пишет:
> 	This patchset impelements support of SOCK_SEQPACKET for virtio
> transport.
> 	As SOCK_SEQPACKET guarantees to save record boundaries, so to
> do it, new packet operation was added: it marks start of record (with
> record length in header), such packet doesn't carry any data.  To send
> record, packet with start marker is sent first, then all data is sent
> as usual 'RW' packets. On receiver's side, length of record is known
> from packet with start record marker. Now as  packets of one socket
> are not reordered neither on vsock nor on vhost transport layers, such
> marker allows to restore original record on receiver's side. If user's
> buffer is smaller that

than


>   record length, when

then


>   v1 -> v2:
>   - patches reordered: af_vsock.c changes now before virtio vsock
>   - patches reorganized: more small patches, where +/- are not mixed

If you did this because I asked, then this
is not what I asked. :)
You can't just add some static func in a
separate patch, as it will just produce the
compilation warning of an unused function.
I only asked to separate the refactoring from
the new code. I.e. if you move some code
block to a separate function, you shouldn't
split that into 2 patches, one that adds a
code block and another one that removes it.
It should be in one patch, so that it is clear
what was moved, and no new warnings are
introduced.
What I asked to separate, is the old code
moves with the new code additions. Such
things can definitely go in a separate patches.

NB: just trying to help, as I already played
with your code a bit. I am neither a
maintainer nor a contributor here, but
it would be cool to have the vsock SEQPACKET
support.
Stefano Garzarella Jan. 18, 2021, 3:16 p.m. UTC | #2
On Fri, Jan 15, 2021 at 12:59:30PM +0300, stsp wrote:
>15.01.2021 08:35, Arseny Krasnov пишет:
>>      This patchset impelements support of SOCK_SEQPACKET for virtio
>>transport.
>>      As SOCK_SEQPACKET guarantees to save record boundaries, so to
>>do it, new packet operation was added: it marks start of record (with
>>record length in header), such packet doesn't carry any data.  To send
>>record, packet with start marker is sent first, then all data is sent
>>as usual 'RW' packets. On receiver's side, length of record is known
>>from packet with start record marker. Now as  packets of one socket
>>are not reordered neither on vsock nor on vhost transport layers, such
>>marker allows to restore original record on receiver's side. If user's
>>buffer is smaller that
>
>than
>
>
>>  record length, when
>
>then
>
>
>>  v1 -> v2:
>>  - patches reordered: af_vsock.c changes now before virtio vsock
>>  - patches reorganized: more small patches, where +/- are not mixed
>
>If you did this because I asked, then this
>is not what I asked. :)
>You can't just add some static func in a
>separate patch, as it will just produce the
>compilation warning of an unused function.
>I only asked to separate the refactoring from
>the new code. I.e. if you move some code
>block to a separate function, you shouldn't
>split that into 2 patches, one that adds a
>code block and another one that removes it.
>It should be in one patch, so that it is clear
>what was moved, and no new warnings are
>introduced.
>What I asked to separate, is the old code
>moves with the new code additions. Such
>things can definitely go in a separate patches.

Arseny, thanks for the v2.
I appreciated that you moved the af_vsock changes before the transport
and also the test, but I agree with stsp about split patches.

As stsp suggested, you can have some "preparation" patches that touch
the already existing code (e.g. rename vsock_stream_sendmsg in
vsock_connectible_sendmsg() and call it inside the new
vsock_stream_sendmsg, etc.), then a patch that adds seqpacket stuff in
af_vsock.

Also for virtio/vhost transports, you can have some patches that add
support in virtio_transport_common, then a patch that enable it in
virtio_transport and a patch for vhost_vsock, as you rightly did in
patch 12.

So, I'd suggest moving out the code that touches virtio_transport.c
from patch 11.

These changes should simplify the review.

In addition, you can also remove the . from the commit titles.


I left other comments in the single patches.

Thanks,
Stefano