netdev - Re: TCP connection/socket gets stuck - Customer requests are dropped with SocketTimeoutException

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]

Message-ID: <CANn89iJLjyzn_=E7fHXJGM1Geg_3WnVTsqVfxGAu-QAyBByHAQ@mail.gmail.com>
Date: Wed, 10 Sep 2025 07:04:11 -0700
From: Eric Dumazet <edumazet@...gle.com>
To: Neal Cardwell <ncardwell@...gle.com>
Cc: Ramakant Badolia <Ramakant.Badolia@...tom.com>, "kuniyu@...gle.com" <kuniyu@...gle.com>, 
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>, Ozan Sengul <Ozan.Sengul@...tom.com>, 
	Raja Sekhar Pula Venkata <RAJASEKHAR.PULAVENKATA@...tom.com>, 
	Jean-Christophe Duberga <Jean-Christophe.Duberga@...tom.com>
Subject: Re: TCP connection/socket gets stuck - Customer requests are dropped
 with SocketTimeoutException

On Wed, Sep 10, 2025 at 6:58 AM Neal Cardwell <ncardwell@...gle.com> wrote:
>
> On Wed, Sep 10, 2025 at 6:16 AM Eric Dumazet <edumazet@...gle.com> wrote:
> >
> > On Wed, Sep 10, 2025 at 1:49 AM Ramakant Badolia
> > <Ramakant.Badolia@...tom.com> wrote:
> > >
> > > Hi Linux TCP Maintainers,
> > >
> > > I am writing to get insight on this bug report - https://bugzilla.kernel.org/show_bug.cgi?id=219221
> > > Unfortunately, we at TomTom have also been stuck with this issue for the last two months and our customer requests are getting dropped intermittently several times a day.
> > >
> > > Currently we are using Linux version 5.14.0-570.37.1.el9_6.x86_64 which is causing this issue.
> > >
> > > As reported in https://bugzilla.kernel.org/show_bug.cgi?id=219221, we don't have possibility to rollback to previous working version.
> > >
> > > I want to check if you acknowledged this bug and what solution was provided? Which version should we switch to in order to have this fixed?
> > >
> >
> > No idea. This might be a question for Redhat support ?
> >
> > I do not think you shared a pcap with us ?
>
> Looks like the bug report at
> https://bugzilla.kernel.org/show_bug.cgi?id=219221 posted a working
> and non-working ("not working TCP connection PCAP file"
> non-working_tcp_packets.pcap) pcap file, though there was only a text
> update for the working case.

This was a one year old report/pcap, I was assuming Ramakant Badolia
had maybe a different issue.

What I saw in it was a suspicious/tiny initial RWIN of  585 bytes in a
SYNACK, of less than one MSS...

 00:00:02.896476 IP 10.51.51.211.57738 > 10.51.51.75.31421: Flags [S],
seq 3801883815, win 64240, options [mss 1460,nop,wscale
8,nop,nop,sackOK], length 0
 00:00:00.000065 IP 10.51.51.75.31421 > 10.51.51.211.57738: Flags
[S.], seq 1230473431, ack 3801883816, win 585, options [mss
1460,nop,nop,sackOK,nop,wscale 0], length 0

Thus my recommendation against tiny and non functional SO_RCVBUF.