netdev - Re: skb_shared_hwstamps ability to handle 64 bit seconds and nanoseconds

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: <CAL+tcoBTk8Bs1vEX=cAKfaN+qrVptkqJH74FE6=VD89nXh2QGw@mail.gmail.com>
Date: Sat, 28 Sep 2024 17:00:46 +0800
From: Jason Xing <kerneljasonxing@...il.com>
To: Greg Dowd <dowdgreg@...il.com>
Cc: netdev@...r.kernel.org, Willem de Bruijn <willemdebruijn.kernel@...il.com>
Subject: Re: skb_shared_hwstamps ability to handle 64 bit seconds and nanoseconds

On Sat, Sep 28, 2024 at 12:36 AM Greg Dowd <dowdgreg@...il.com> wrote:
>
> Hello all,
> I am not sure if this is the right list but I posted this in newbies
> and got referred here.  Anyone have any insight?
>
>
> I had a question regarding kernel timestamping.  I see definitions for
> SO_TIMESTAMPING_NEW in networking options to allow use of time
> structures containing 64bit timestamps.  However, I don't see any way
> for skbuff timestamps to pass around structures with 64 bits as the
> skb_shared_hwstamps use a typedef ktime_t which stacks the seconds and
> nanoseconds into a single 64bit value.
> I am not sure who maintains this section of the code.  Any ideas?

I don't know what the real question is, either. But I think I can help
you somehow:

1) try to understand Documentation/networking/timestamping.rst
2) try to run and test tools/testing/selftests/net/txtimestamp.c,
which is good for you to understand the SO_TIMESTAMPING feature.
3) I guess that you're looking for how the sk tsflags can be passed to
each last skb in one sendmsg()? If so, I recommend that you read the
code following this call trace:
tcp_sendmsg_locked()->tcp_tx_timestamp()->sock_tx_timestamp().
4) I also guess that you're looking for how the hardware timestamp is
passed to the userside during the report phase? If so, please see
sock_recv_timestamp() and skb_tstamp_tx().

For now, Willem is the maintainer of the SO_TIMESTAMPING* feature. Let
me CC him here.

Thanks,
Jason