Documentation/networking/timestamping.rst

1 .. SPDX-License-Identifier: GPL-2.0
32 IP_MULTICAST_LOOP + SO_TIMESTAMP[NS]
43 -------------------------------------------------------------
59 -------------------------------------------------------------------
62 Its struct timespec allows for higher resolution (ns) timestamps than the
72 ----------------------------------------------------------------------
102   Request rx timestamps generated by the network adapter.
105   Request rx timestamps when data enters the kernel. These timestamps
122   transmit latency is, if long, often dominated by queuing delay. The
131   grained measurement of queuing delay. This flag can be enabled
138   over-report measurement, because the timestamp is generated when all
180   is derived from a per-socket u32 counter (that wraps). For datagram
207   a timestamp with counter N-1. SOF_TIMESTAMPING_OPT_ID_TCP
257   received the packet and its length at layer 2. A valid (non-zero)
285 Using this feature, applications can sample timestamps per sendmsg()
292   cmsg->cmsg_level	       = SOL_SOCKET;
293   cmsg->cmsg_type	       = SO_TIMESTAMPING;
294   cmsg->cmsg_len	       = CMSG_LEN(sizeof(__u32));
312 -------------------------
322 correlating a timestamp with data is non-trivial. A range of bytes
351 relevant sequence number in skb_shinfo(skb)->tskey. Because an skbuff
381 ----------------------------
404 feature. At least one field is non-zero at any time. Most timestamps
411 as linuxptp. For the PTP clock API, see Documentation/driver-api/ptp.rst.
448 is the first if ts[2] is non-zero, the second otherwise, in which
476 Reading from the error queue is always a non-blocking operation. To
512 the requested fine-grained filtering for incoming packets is not
539     /* possible values for hwtstamp_config->tx_type */
557     /* possible values for hwtstamp_config->rx_filter */
577 --------------------------------------------------------
597 - In hard_start_xmit(), check if (skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)
598   is set no-zero. If yes, then the driver is expected to do hardware time
600 - If this is possible for the skb and requested, then declare
602   SKBTX_IN_PROGRESS in skb_shinfo(skb)->tx_flags , e.g. with::
604       skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
610 - Driver should call skb_tx_timestamp() as close to passing sk_buff to hardware
613 - As soon as the driver has sent the packet and/or obtained a
625 ----------------------------------------------------------
642 interface (redirecting to the host port on TX, and intercepting frames on RX).
645 suffer, since the switch's variable queuing delay introduces a path delay
666 - ``.port_txtstamp()``: a hook called prior to the transmission of
668   This is required for two-step timestamping, since the hardware
671   packet so that it can re-enqueue the packet back into the socket's
674   in skb->cb and enqueue a tx skb queue. Typically, a switch will have a
677   key-value pairs of PTP sequence ID/message type/domain number and the
685   One-step TX timestamping do not require packet cloning, since there is
686   no follow-up message required by the PTP protocol (because the
689   to be re-enqueued into its socket's error queue.
691 - ``.port_rxtstamp()``: On RX, the BPF classifier is run by DSA to
696   timestamps might either be available in-band (through metadata in the
697   DSA header, or attached in other ways to the packet), or out-of-band
698   (through another RX timestamping FIFO). Deferral on RX is typically
713 mii_timestamper`` and add a pointer to it in ``phydev->mii_ts``. The presence
721 - Checking, in ``.ndo_eth_ioctl``, whether ``phy_has_hwtstamp(netdev->phydev)``
725 - On RX, special intervention may or may not be needed, depending on the
728   ``skb_defer_rx_timestamp(skb)`` is necessary or not - and if it is, don't
730   enabled, and ``skb->dev->phydev->mii_ts`` exists, its ``.rxtstamp()`` hook
732   deferral for RX timestamping is necessary.  Again like DSA, it becomes the
741 - On TX, again, special intervention might or might not be needed.  The
742   function that calls the ``mii_ts->txtstamp()`` hook is named
763 Stacked PHCs, especially DSA (but not only) - since that doesn't require any
765 all possible code paths - is that they uncover bugs which were impossible to
769       skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
780    the ``.ndo_eth_ioctl`` ("``priv->hwtstamp_tx_enabled == true``") and the
781    current skb requires a TX timestamp ("``skb_shinfo(skb)->tx_flags &
783    "``skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS``" flag. Note: as
791    "``skb_shinfo(skb)->tx_flags & SKBTX_IN_PROGRESS``" was set. With a stacked
798 "``skb_shinfo(skb)->tx_flags & SKBTX_IN_PROGRESS``", but also for
799 "``priv->hwtstamp_tx_enabled == true``". Because the rest of the system ensures