Re: [RFC 00/12] io_uring zerocopy send

From: Pavel Begunkov <asml.silence@gmail.com>
To: Martin KaFai Lau <kafai@fb.com>, David Ahern <dsahern@gmail.com>
Cc: io-uring@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org, Jakub Kicinski <kuba@kernel.org>,
	Jonathan Lemon <jonathan.lemon@gmail.com>,
	"David S . Miller" <davem@davemloft.net>,
	Willem de Bruijn <willemb@google.com>,
	Eric Dumazet <edumazet@google.com>,
	Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
	David Ahern <dsahern@kernel.org>, Jens Axboe <axboe@kernel.dk>
Subject: Re: [RFC 00/12] io_uring zerocopy send
Date: Wed, 1 Dec 2021 23:18:36 +0000	[thread overview]
Message-ID: <3e2caa38-27a1-986f-7187-b4e333105371@gmail.com> (raw)
In-Reply-To: <20211201230757.fuhnhgtbx5o22wgs@kafai-mbp.dhcp.thefacebook.com>

On 12/1/21 23:07, Martin KaFai Lau wrote:
> On Wed, Dec 01, 2021 at 03:35:49PM -0700, David Ahern wrote:
>> On 12/1/21 2:51 PM, Martin KaFai Lau wrote:
>>>
>>> To tx out dummy, I did:
>>> #> ip a add 10.0.0.1/24 dev dummy0
>>                ^^^^^^^^
>>>
>>> #> ip -4 r
>>> 10.0.0.0/24 dev dummy0 proto kernel scope link src 10.0.0.1
>>>
>>> #> ./send-zc -4 -D 10.0.0.(2) -t 10 udp
>>                       ^^^^^^^^^^
>>
>> Pavel's commands have: 'send-zc -4 -D <dummy_ip_addr> -t 10 udp'
>>
>> I read dummy_ip_addr as the address assigned to dummy0; that's an
>> important detail. You are sending to an address on that network, not the
>> address assigned to the device, in which case packets are created and
>> then dropped by the dummy driver - nothing actually makes it to the server.
> Right, I assumed "dropped by dummy driver" is the usual intention
> for using dummy, so just in case if it was the intention for
> testing tx only.  You are right and it seems the intention
> of this command is to have server receiving the packets.

I see, it seems we found the misunderstanding:

For dummy device, I indeed was testing only tx part without
a server, in my understanding it better approximates an actual
fast NIC. I don't think dummy is even capable to pass the data,
so was thinking that it's given that there is no receive side,
and so no "msg_zerocopy -r".

For localhost testing (with the hack), there was a server
verifying data.

>>
>>> ip -s link show dev dummy0
>>> 2: dummy0: <BROADCAST,NOARP,UP,LOWER_UP> mtu 65535 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
>>>     link/ether 82:0f:e0:dc:f7:e6 brd ff:ff:ff:ff:ff:ff
>>>     RX:    bytes packets errors dropped  missed   mcast
>>>                0       0      0       0       0       0
>>>     TX:    bytes packets errors dropped carrier collsns
>>>     140800890299 2150397      0       0       0       0
>>>
>>

-- 
Pavel Begunkov