On 8/13/24 22:13, Mina Almasry wrote:

Implement a memory provider that allocates dmabuf devmem in the form of
net_iov.

The provider receives a reference to the struct netdev_dmabuf_binding
via the pool->mp_priv pointer. The driver needs to set this pointer for
the provider in the net_iov.

The provider obtains a reference on the netdev_dmabuf_binding which
guarantees the binding and the underlying mapping remains alive until
the provider is destroyed.

Usage of PP_FLAG_DMA_MAP is required for this memory provide such that
the page_pool can provide the driver with the dma-addrs of the devmem.

Support for PP_FLAG_DMA_SYNC_DEV is omitted for simplicity & p.order !=
0.

Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Kaiyuan Zhang <redacted>
Signed-off-by: Mina Almasry <redacted>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>

---

v19:
- Add PP_FLAG_ALLOW_UNREADABLE_NETMEM flag. It serves 2 purposes, (a)
   it guards drivers that don't support unreadable netmem (net_iov
   backed) from accidentally getting exposed to it, and (b) drivers that
   wish to create header pools can unset it for that pool to force
   readable netmem.
- Add page_pool_check_memory_provider, which verifies that the driver
   has created a page_pool with the expected configuration. This is used
   to report to the user if the mp configuration succeeded, and also
   verify that the driver is doing the right thing.
- Don't reset niov->dma_addr on allocation/free.

v17:
- Use ASSERT_RTNL (Jakub)

v16:
- Add DEBUG_NET_WARN_ON_ONCE(!rtnl_is_locked()), to catch cases if
   page_pool_init without rtnl_locking when the queue is provided. In
   this case, the queue configuration may be changed while we're initing
   the page_pool, which could be a race.

v13:
- Return on warning (Pavel).
- Fixed pool->recycle_stats not being freed on error (Pavel).
- Applied reviewed-by from Pavel.

v11:
- Rebase to not use the ops. (Christoph)

v8:
- Use skb_frag_size instead of frag->bv_len to fix patch-by-patch build
   error

v6:
- refactor new memory provider functions into net/core/devmem.c (Pavel)

v2:
- Disable devmem for p.order != 0

v1:
- static_branch check in page_is_page_pool_iov() (Willem & Paolo).
- PP_DEVMEM -> PP_IOV (David).
- Require PP_FLAG_DMA_MAP (Jakub).

...

Frankly, I pretty much don't like it.

1. We do it after reconfiguring the queue just to fail and reconfigure
it again.

2. It should be a part of the common path like netdev_rx_queue_restart(),
not specific to devmem TCP.

These two can be fixed by moving the check into
netdev_rx_queue_restart() just after ->ndo_queue_mem_alloc, assuming
that the callback where we init page pools.

3. That implicit check gives me bad feeling, instead of just getting
direct feedback from the driver, either it's a flag or an error
returned, we have to try to figure what exactly the driver did, with
a high chance this inference will fail us at some point.

And page_pool_check_memory_provider() is not that straightforward,
it doesn't walk through pools of a queue. Not looking too deep,
but it seems like the nested loop can be moved out with the same
effect, so it first looks for a pool in the device and the follows
with the bound_rxqs. And seems the bound_rxqs check would always turn
true, you set the binding into the map in
net_devmem_bind_dmabuf_to_queue() before the restart and it'll be there
after restart for page_pool_check_memory_provider(). Maybe I missed
something, but it's not super clear.

4. And the last thing Jakub mentioned is that we need to be prepared
to expose a flag to the userspace for whether a queue supports
netiov. Not really doable in a sane manner with such implicit
post configuration checks.

And that brings us back to the first approach I mentioned, where
we have a flag in the queue structure, drivers set it, and
netdev_rx_queue_restart() checks it before any callback. That's
where the thread with Jakub stopped, and it reads like at least
he's not against the idea.

...

-- 
Pavel Begunkov