On 2/25/25 11:54 AM, Mina Almasry wrote:

Will do, although I think maybe I'll add that comment above the
refcount_t definition in net_devmem_dmabuf_binding, if that's OK with
you. That was meant to explain how the refcounting on the binding
works.

That is fine, I just think the lifecycle of the dmabuf and what the
unbinding is really doing should be clearly documented to avoid surprises.

Maybe worthy of note, this is not that new, in fact something similar
is happening in the RX path. In the RX path each skb waiting in the
receive queue to be recvmsg()'d will hold a reference on the
underlying page_pool, which in-turn holds a reference on the
underlying dmabuf. It's just that a similar thing is happening in the
TX path.