netdev - Re: [PATCH net-next v1 0/3] bnxt_en: implement netdev_queue_mgmt

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <b2dadafd-48c3-4598-bee5-a088ae5a4bc7@davidwei.uk>
Date: Mon, 10 Jun 2024 20:41:55 -0700
From: David Wei <dw@...idwei.uk>
To: David Ahern <dsahern@...nel.org>, 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
Cc: Pavel Begunkov <asml.silence@...il.com>, Jakub Kicinski
 <kuba@...nel.org>, "David S. Miller" <davem@...emloft.net>,
 Eric Dumazet <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next v1 0/3] bnxt_en: implement netdev_queue_mgmt_ops

On 2024-06-10 19:40, David Ahern wrote:
> On 6/10/24 8:33 PM, David Wei wrote:
>> Implement netdev_queue_mgmt_ops for bnxt added in [1]. This will be used
>> in the io_uring ZC Rx patchset to configure queues with a custom page
>> pool w/ a special memory provider for zero copy support.
>>
> 
> 
> I do not see it explicitly called out, so asking here: does this change
> enable / require header split if using memory from this "special memory
> provider"?

This patchset is orthogonal to header split and page pool memory
providers. It implements netdev_queue_mgmt_ops which enables dynamically
resetting a single Rx queue without needing a full device close/open,
and pre-allocates queue descriptor memory _before_ stopping the queue to
minimise downtime. It shows that, with FW support and carefully written
code, it is possible for drivers to move away from doing full resets on
changes.

My first intent is to use it for reconfiguring an Rx queue with a
different page pool, but that isn't the only purpose. For example, it
opens up the possibility of dynamically creating queues that share a
single napi instance.