On 27.07.2022 15:17, Stefano Garzarella wrote:

On Mon, Jul 25, 2022 at 07:56:59AM +0000, Arseniy Krasnov wrote:

quoted

Both bits indicate, that next data read call won't be blocked, but when
sk_rcvlowat is not 1, these bits will be set by poll anyway, thus when
user tries to dequeue data,it will wait until sk_rcvlowat bytes of data
will be available.

The patch LGTM, but I suggest you to rewrite the title and commit of the message to better explain what this patch does (pass sock_rcvlowat to notify_poll_in as target) and then explain why as you already did (to set POLLIN/POLLRDNORM only when target is reached).

Ok, i see. Ack

Thanks,
Stefano

quoted

Signed-off-by: Arseniy Krasnov <redacted>
---
net/vmw_vsock/af_vsock.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/vmw_vsock/af_vsock.c b/net/vmw_vsock/af_vsock.c
index f04abf662ec6..63a13fa2686a 100644
--- a/net/vmw_vsock/af_vsock.c
+++ b/net/vmw_vsock/af_vsock.c

@@ -1066,8 +1066,9 @@ static __poll_t vsock_poll(struct file *file, struct socket *sock,

        if (transport && transport->stream_is_active(vsk) &&
            !(sk->sk_shutdown & RCV_SHUTDOWN)) {
            bool data_ready_now = false;
+            int target = sock_rcvlowat(sk, 0, INT_MAX);
            int ret = transport->notify_poll_in(
-                    vsk, 1, &data_ready_now);
+                    vsk, target, &data_ready_now);
            if (ret < 0) {
                mask |= EPOLLERR;
            } else {
-- 
2.25.1