On 1/9/25 14:42, Stefano Garzarella wrote:

On Thu, Jan 09, 2025 at 02:34:28PM +0100, Michal Luczaj wrote:

quoted

...
That said, when I apply this patch, but drop the `sk->sk_state !=
TCP_LISTEN &&`: no more splats.

We can't drop `sk->sk_state != TCP_LISTEN &&` because listener socket 
doesn't have any transport (vsk->transport == NULL), so every connection 
request will receive an error, so maybe this is the reason of no splats.

Bah, sorry, I didn't run the test suit.

I'm cooking some more patches to fix Hyunwoo's scenario handling better 
the close work when the virtio destructor is called.

I'll run your reproduces to test it, thanks for that!

Stefano