netdev - Re: [PATCH net-next] bonding: add software timestamping support

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZCUJgmGacqI5Aw+L@Laptop-X1>
Date:   Thu, 30 Mar 2023 12:01:06 +0800
From:   Hangbin Liu <liuhangbin@...il.com>
To:     Jay Vosburgh <jay.vosburgh@...onical.com>
Cc:     netdev@...r.kernel.org, "David S . Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Jonathan Toppins <jtoppins@...hat.com>,
        Paolo Abeni <pabeni@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Miroslav Lichvar <mlichvar@...hat.com>,
        Richard Cochran <richardcochran@...il.com>
Subject: Re: [PATCH net-next] bonding: add software timestamping support

On Tue, Mar 28, 2023 at 08:36:58PM -0700, Jay Vosburgh wrote:
> Hangbin Liu <liuhangbin@...il.com> wrote:
> 
> >At present, bonding attempts to obtain the timestamp (ts) information of
> >the active slave. However, this feature is only available for mode 1, 5,
> >and 6. For other modes, bonding doesn't even provide support for software
> >timestamping. To address this issue, let's call ethtool_op_get_ts_info
> >when there is no primary active slave. This will enable the use of software
> >timestamping for the bonding interface.
> 
> 	If I'm reading the patch below correctly, the actual functional
> change here is to additionally set SOF_TIMESTAMPING_TX_SOFTWARE in
> so_timestamping for the active-backup, balance-tlb and balance-alb modes

No. In the description. I said for other modes, bonding doesn't even provide
support for software timestamping. So this patch is to address this issue.
i.e. add sw timestaming for all bonding modes.

For mode 1,5,6. We will try find the active slave and get it's ts info
directly. If there is no ops->get_ts_info, just use sw timestamping.

For other modes, use sw timestamping directly.

This is because some users want to use PTP over bond with other modes. e.g. LACP.
They are satisfied with just sw timestamping as it's difficult to support hw
timestamping for LACP bonding.

Before this patch, bond mode with 0, 2, 3, 4 only has software-receive.

# ethtool -T bond0
Time stamping parameters for bond0:
Capabilities:
        software-receive
        software-system-clock
PTP Hardware Clock: none
Hardware Transmit Timestamp Modes: none
Hardware Receive Filter Modes: none

# ptp4l -m -S -i bond0
ptp4l[66296.154]: interface 'bond0' does not support requested timestamping mode
failed to create a clock

After this patch:

# ethtool -T bond0
Time stamping parameters for bond0:
Capabilities:
        software-transmit
        software-receive
        software-system-clock
PTP Hardware Clock: none
Hardware Transmit Timestamp Modes: none
Hardware Receive Filter Modes: none

# ptp4l -m -S -i bond0
ptp4l[66952.474]: port 1: INITIALIZING to LISTENING on INIT_COMPLETE
ptp4l[66952.474]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
ptp4l[66952.474]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
ptp4l[66981.681]: port 1: LISTENING to MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
ptp4l[66981.681]: selected local clock 007c50.fffe.70cdb6 as best master
ptp4l[66981.682]: port 1: assuming the grand master role

Thanks
Hangbin