netdev - Re: [PATCH net-next 0/7] tcp: implement rb-tree based retransmit queue

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] [day] [month] [year] [list]

Date:   Tue, 6 Feb 2018 09:05:11 -0800
From:   Eric Dumazet <edumazet@...gle.com>
To:     Tal Gilboa <talgi@...lanox.com>
Cc:     Eric Dumazet <eric.dumazet@...il.com>,
        David Laight <David.Laight@...lab.com>,
        David Miller <davem@...emloft.net>,
        "ncardwell@...gle.com" <ncardwell@...gle.com>,
        "ycheng@...gle.com" <ycheng@...gle.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Saeed Mahameed <saeedm@...lanox.com>,
        Tariq Toukan <tariqt@...lanox.com>,
        Amir Ancel <amira@...lanox.com>
Subject: Re: [PATCH net-next 0/7] tcp: implement rb-tree based retransmit queue

On Tue, Feb 6, 2018 at 8:27 AM, Tal Gilboa <talgi@...lanox.com> wrote:
> On 2/6/2018 5:52 PM, Eric Dumazet wrote:
>>
>> On Tue, 2018-02-06 at 15:22 +0000, David Laight wrote:
>>>
>>> From: Eric Dumazet
>>>>
>>>> Sent: 06 February 2018 14:20
>>>
>>>
>>> ...
>>>>
>>>> Please give exact details.
>>>> Sending 64, 128, 256 or 512 bytes at a time on TCP_STREAM makes little
>>>> sense.
>>>> We are not optimizing stack for pathological cases, sorry.
>>>
>>>
>>> There are plenty of workloads which are not bulk data and where multiple
>>> small buffers get sent at unknown intervals (which may be back to back).
>>> Such connections have to have Nagle disabled because the Nagle delays
>>> are 'horrid'.
>>> Clearly lost packets can cause delays, but they are rare on local
>>> networks.
>>
>>
>> Auto corking makes sure aggregation happens, even for when Nagle is in
>> the picture.
>
>
>>
>> netperf -- -m 256    will still cook 64KB TSO packets
>
>
> This is what we would have liked to see, but auto corking isn't forcing 64KB
> TSO packets. Under certain conditions, specifically when TX queue is empty,
> it would send the SKB to transmit even if it isn't full:

Yes.

Auto corking does not predict the future, nor arm a high resolution
timer when application
does a send(small_size)

This packet is sent immaediately, as instructed by application and TCP
normal behavior.

But second or third packet would detect the condition.

Unless a driver does skb_orphan() too early, breaking back pressure.

> static bool tcp_should_autocork(struct sock *sk, struct sk_buff *skb,
>                                 int size_goal)
> {
>         return skb->len < size_goal &&
>                sock_net(sk)->ipv4.sysctl_tcp_autocorking &&
>                skb != tcp_write_queue_head(sk) &&
>                refcount_read(&sk->sk_wmem_alloc) > skb->truesize;
> }
> When skb == tcp_write_queue_head(sk) corking is done. This is part of the
> optimization for mlx5 driver I've mentioned. If we can better utilize auto
> corking we shouldn't have an issue.

Or not issue expensive system calls for small payloads. stdio was
invented a while back :)

Meltdown/Spectre mitigation put high price to system calls nowadays.