netdev - RE: [Linuxptp-devel] strangeness

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: <02874ECE860811409154E81DA85FBB58923253F1@ORSMSX121.amr.corp.intel.com>
Date:   Tue, 12 Mar 2019 21:40:59 +0000
From:   "Keller, Jacob E" <jacob.e.keller@...el.com>
To:     Paul Thomas <pthomas8589@...il.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Harini Katakam <harinik@...inx.com>
CC:     "linuxptp-devel@...ts.sourceforge.net" 
        <linuxptp-devel@...ts.sourceforge.net>
Subject: RE: [Linuxptp-devel] strangeness



> -----Original Message-----
> From: Paul Thomas [mailto:pthomas8589@...il.com]
> Sent: Tuesday, March 12, 2019 1:00 PM
> To: Keller, Jacob E <jacob.e.keller@...el.com>; netdev@...r.kernel.org; Harini
> Katakam <harinik@...inx.com>
> Cc: linuxptp-devel@...ts.sourceforge.net
> Subject: Re: [Linuxptp-devel] strangeness
> 
> On Tue, Mar 12, 2019 at 3:22 PM Keller, Jacob E
> <jacob.e.keller@...el.com> wrote:
> >
> >
> > Hmm. I haven't been able to reproduce this using other device drivers, so I wonder
> if the specific device driver is doing something weird and indicating that it can Tx
> timestamp every packet, in its hard_xmit routine... But it should only be doing that for
> packets which are requested to timestamp from the socket.
> 
> Yes! I think this could be it. I just submitted this patch so that we
> can talk about it:
> https://lore.kernel.org/netdev/20190312195053.21741-1-pthomas8589@gmail.com/
> 
> This fixes the issue I'm having with nc and ssh.
> 
> -Paul
> 

The patch looks good. I think it's correct, since only sockets which expect to get Tx timestamps should be waiting and checking for them. Other sockets won't be checking, but also won't get betting timestamps.

I'm not sure if there is some improvement we could do in the future for if a Tx timestamp is sent for a packet that didn't request it.

Maybe others know more about that flow?

Thanks,
Jake