netdev - [PATCH v1 0/2] tcp bugs for window shrinking, OoO ACKs

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-next>] [day] [month] [year] [list]

Date:   Thu, 20 Oct 2022 11:22:40 -0700
From:   Kamaljit Singh <kamaljit.singh1@....com>
To:     edumazet@...gle.com, davem@...emloft.net, yoshfuji@...ux-ipv6.org,
        dsahern@...nel.org, kuba@...nel.org, pabeni@...hat.com
Cc:     netdev@...r.kernel.org, Niklas.Cassel@....com,
        Damien.LeMoal@....com, kamaljit.singh1@....com
Subject: [PATCH v1 0/2] tcp bugs for window shrinking, OoO ACKs

We have found significant performance impacts for 4K Writes to an NVMe/TCP
target under heavy congestion conditions due to improper SEQ-NUM number 
provided by the host TCP layer. In this case the target shrunk the window,
which seemed to cause the host to send a stale SEQ-NUM for a subsequent
ACK.

It was also observed that on occasions that a host sends ACKs out-of-order
it caused data retransmissions thus affecting performance for NVMe/TCP
Writes.

A network trace is available and can be provided for further reference.

Kamaljit Singh (2):
  tcp: Fix for stale host ACK when tgt window shrunk
  tcp: Ignore OOO handling for TCP ACKs

 net/ipv4/tcp_output.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

-- 
2.25.1