2023-09-15 02:27:15

by Stefano Garzarella

[permalink] [raw]
Subject: Re: [PATCH net-next v8 0/4] vsock/virtio/vhost: MSG_ZEROCOPY preparations

Hi Arseniy,

On Mon, Sep 11, 2023 at 11:22:30PM +0300, Arseniy Krasnov wrote:
>Hello,
>
>this patchset is first of three parts of another big patchset for
>MSG_ZEROCOPY flag support:
>https://lore.kernel.org/netdev/[email protected]/
>
>During review of this series, Stefano Garzarella <[email protected]>
>suggested to split it for three parts to simplify review and merging:
>
>1) virtio and vhost updates (for fragged skbs) <--- this patchset
>2) AF_VSOCK updates (allows to enable MSG_ZEROCOPY mode and read
> tx completions) and update for Documentation/.
>3) Updates for tests and utils.
>
>This series enables handling of fragged skbs in virtio and vhost parts.
>Newly logic won't be triggered, because SO_ZEROCOPY options is still
>impossible to enable at this moment (next bunch of patches from big
>set above will enable it).
>
>I've included changelog to some patches anyway, because there were some
>comments during review of last big patchset from the link above.

Thanks, I left some comments on patch 4, the others LGTM.
Sorry to not having spotted them before, but moving
virtio_transport_alloc_skb() around the file, made the patch a little
confusing and difficult to review.

In addition, I started having failures of test 14 (server: host,
client: guest), so I looked better to see if there was anything wrong,
but it fails me even without this series applied.

It happens to me intermittently (~30%), does it happen to you?
Can you take a look at it?

host$ ./vsock_test --mode=server --control-port=12345 --peer-cid=4
...
14 - SOCK_STREAM virtio skb merge...expected recv(2) returns 8 bytes, got 3

guest$ ./vsock_test --mode=client --control-host=192.168.133.2 --control-port=12345 --peer-cid=2

Thanks,
Stefano