netdev - Re: [PATCH net-next] tcp: force a PSH flag on TSO packets

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:   Thu, 19 Sep 2019 07:18:12 -0700
From:   Eric Dumazet <edumazet@...gle.com>
To:     Or Gerlitz <gerlitz.or@...il.com>
Cc:     Eric Dumazet <eric.dumazet@...il.com>,
        "David S . Miller" <davem@...emloft.net>,
        netdev <netdev@...r.kernel.org>,
        Soheil Hassas Yeganeh <soheil@...gle.com>,
        Neal Cardwell <ncardwell@...gle.com>,
        Yuchung Cheng <ycheng@...gle.com>,
        Daniel Borkmann <daniel@...earbox.net>,
        Tariq Toukan <tariqt@...lanox.com>
Subject: Re: [PATCH net-next] tcp: force a PSH flag on TSO packets

On Thu, Sep 19, 2019 at 7:01 AM Or Gerlitz <gerlitz.or@...il.com> wrote:
>
> On Thu, Sep 19, 2019 at 4:46 PM Eric Dumazet <eric.dumazet@...il.com> wrote:
> > On 9/19/19 5:17 AM, Or Gerlitz wrote:
> > > On Wed, Sep 11, 2019 at 12:54 AM Eric Dumazet <edumazet@...gle.com> wrote:
> > >> When tcp sends a TSO packet, adding a PSH flag on it
> > >> reduces the sojourn time of GRO packet in GRO receivers.
> > >>
> > >> This is particularly the case under pressure, since RX queues
> > >> receive packets for many concurrent flows.
> > >>
> > >> A sender can give a hint to GRO engines when it is
> > >> appropriate to flush a super-packet, especially when pacing
>
> > > Is this correct that we add here the push flag for the tcp header template
> > > from which all the tcp headers for SW GSO packets will be generated?
> > > Wouldn't that cause a too early flush on GRO engines at the receiver side?
>
> > If a TSO engine is buggy enough to add the PSH on all the segments, it needs
> > to be fixed urgently :)
>
> yeah, but I guess you were not able to test this over all the TSO HWs
> out there..
> so I guess if someone complains we will have to add a quirk to disable
> that, lets see..


The only known pain point for TSO is the ECN part (probably was missed
by first Microsoft specs)

This is why we have NETIF_F_TSO_ECN