[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240624150254.053d0c00@kernel.org>
Date: Mon, 24 Jun 2024 15:02:54 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: David Wei <dw@...idwei.uk>
Cc: Michael Chan <michael.chan@...adcom.com>, Andy Gospodarek
<andrew.gospodarek@...adcom.com>, Adrian Alvarado
<adrian.alvarado@...adcom.com>, Somnath Kotur <somnath.kotur@...adcom.com>,
netdev@...r.kernel.org, Pavel Begunkov <asml.silence@...il.com>, David
Ahern <dsahern@...nel.org>, "David S. Miller" <davem@...emloft.net>, Eric
Dumazet <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next v3 2/2] bnxt_en: implement
netdev_queue_mgmt_ops
On Mon, 24 Jun 2024 11:20:59 -0700 David Wei wrote:
> > What's the warning you hit?
> > We should probably bring back page_pool_unlink_napi(),
> > if this is really needed.
>
> This one:
>
> https://elixir.bootlin.com/linux/v6.10-rc5/source/net/core/page_pool.c#L1030
>
> The cause is having two different bnxt_rx_ring_info referring to the
> same NAPI instance. One is the proper one in bp->rx_ring, the other is
> the temporarily allocated one for holding the "replacement" during the
> reset.
Makes sense, as I said please look thru the history - some form of
page_pool_unlink_napi() used to be exported for this use case, but
Olek(?) deleted it due to lack of in-tree users.
With that helper in place you can unlink the page pool while the NAPI
is stopped, without poking into internals at the driver level.
Powered by blists - more mailing lists