Köry Maincent wrote:
> From: Richard Cochran <richardcochran@xxxxxxxxx>
>
> Time stamping on network packets may happen either in the MAC or in
> the PHY, but not both. In preparation for making the choice
> selectable, expose both the current and available layers via sysfs.
>
> In accordance with the kernel implementation as it stands, the current
> layer will always read as "phy" when a PHY time stamping device is
> present. Future patches will allow changing the current layer
> administratively.
>
> Signed-off-by: Richard Cochran <richardcochran@xxxxxxxxx>
> Signed-off-by: Kory Maincent <kory.maincent@xxxxxxxxxxx>
> ---
>
> Notes:
> Changes in v2:
> - Move the introduction of selected_timestamping_layer variable in next
> patch.
>
> .../ABI/testing/sysfs-class-net-timestamping | 17 ++++++
> net/core/net-sysfs.c | 60 +++++++++++++++++++
> 2 files changed, 77 insertions(+)
> create mode 100644 Documentation/ABI/testing/sysfs-class-net-timestamping
>
> diff --git a/Documentation/ABI/testing/sysfs-class-net-timestamping b/Documentation/ABI/testing/sysfs-class-net-timestamping
> new file mode 100644
> index 000000000000..529c3a6eb607
> --- /dev/null
> +++ b/Documentation/ABI/testing/sysfs-class-net-timestamping
> @@ -0,0 +1,17 @@
> +What: /sys/class/net/<iface>/available_timestamping_providers
> +Date: January 2022
> +Contact: Richard Cochran <richardcochran@xxxxxxxxx>
> +Description:
> + Enumerates the available providers for SO_TIMESTAMPING.
> + The possible values are:
> + - "mac" The MAC provides time stamping.
> + - "phy" The PHY or MII device provides time stamping.
> +
> +What: /sys/class/net/<iface>/current_timestamping_provider
> +Date: January 2022
> +Contact: Richard Cochran <richardcochran@xxxxxxxxx>
> +Description:
> + Show the current SO_TIMESTAMPING provider.
> + The possible values are:
> + - "mac" The MAC provides time stamping.
> + - "phy" The PHY or MII device provides time stamping.
> diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
> index 8409d41405df..26095634fb31 100644
> --- a/net/core/net-sysfs.c
> +++ b/net/core/net-sysfs.c
> @@ -620,6 +620,64 @@ static ssize_t threaded_store(struct device *dev,
> }
> static DEVICE_ATTR_RW(threaded);
>
> +static ssize_t available_timestamping_providers_show(struct device *dev,
> + struct device_attribute *attr,
> + char *buf)
> +{
> + const struct ethtool_ops *ops;
> + struct net_device *netdev;
> + struct phy_device *phydev;
> + int ret = 0;
> +
> + netdev = to_net_dev(dev);
> + phydev = netdev->phydev;
> + ops = netdev->ethtool_ops;
> +
> + if (!rtnl_trylock())
> + return restart_syscall();
> +
> + ret += sprintf(buf, "%s\n", "mac");
> + buf += 4;
> +
Should advertising mac be subject to having ops->get_ts_info?
> + if (phy_has_tsinfo(phydev)) {
> + ret += sprintf(buf, "%s\n", "phy");
> + buf += 4;
> + }
> +
> + rtnl_unlock();
> +
> + return ret;
> +}
> +static DEVICE_ATTR_RO(available_timestamping_providers);
