mbox series

[V8,net-next,00/12] Peer to Peer One-Step time stamping

Message ID cover.1576956342.git.richardcochran@gmail.com
Headers show
Series Peer to Peer One-Step time stamping | expand

Message

Richard Cochran Dec. 21, 2019, 7:36 p.m. UTC
This series adds support for PTP (IEEE 1588) P2P one-step time
stamping along with a driver for a hardware device that supports this.

If the hardware supports p2p one-step, it subtracts the ingress time
stamp value from the Pdelay_Request correction field.  The user space
software stack then simply copies the correction field into the
Pdelay_Response, and on transmission the hardware adds the egress time
stamp into the correction field.

This new functionality extends CONFIG_NETWORK_PHY_TIMESTAMPING to
cover MII snooping devices, but it still depends on phylib, just as
that option does.  Expanding beyond phylib is not within the scope of
the this series.

User space support is available in the current linuxptp master branch.

- Patch 1 adds phy_device methods for existing time stamping fields.
- Patches 2-5 convert the stack and drivers to the new methods.
- Patch 6 moves code around the dp83640 driver.
- Patches 7-10 add support for MII time stamping in non-PHY devices.
- Patch 11 adds the new P2P 1-step option.
- Patch 12 adds a driver implementing the new option.

Thanks,
Richard

Changed in v8:
~~~~~~~~~~~~~~

- Avoided adding forward functional declarations in the dp83640 driver.
- Picked up Florian's new review tags and another one from Andrew.

Changed in v7:
~~~~~~~~~~~~~~

- Converted pr_debug|err to dev_ variants in new driver.
- Fixed device tree documentation per Rob's v6 review.
- Picked up Andrew's and Rob's review tags.
- Silenced sparse warnings in new driver.

Changed in v6:
~~~~~~~~~~~~~~

- Added methods for accessing the phy_device time stamping fields.
- Adjust the device tree documentation per Rob's v5 review.
- Fixed the build failures due to missing exports.

Changed in v5:
~~~~~~~~~~~~~~

- Fixed build failure in macvlan.
- Fixed latent bug with its gcc warning in the driver.

Changed in v4:
~~~~~~~~~~~~~~

- Correct error paths and PTR_ERR return values in the framework.
- Expanded KernelDoc comments WRT PHY locking.
- Pick up Andrew's review tag.

Changed in v3:
~~~~~~~~~~~~~~

- Simplify the device tree binding and document the time stamping
  phandle by itself.

Changed in v2:
~~~~~~~~~~~~~~

- Per the v1 review, changed the modeling of MII time stamping
  devices.  They are no longer a kind of mdio device.

Richard Cochran (12):
  net: phy: Introduce helper functions for time stamping support.
  net: macvlan: Use the PHY time stamping interface.
  net: vlan: Use the PHY time stamping interface.
  net: ethtool: Use the PHY time stamping interface.
  net: netcp_ethss: Use the PHY time stamping interface.
  net: phy: dp83640: Move the probe and remove methods around.
  net: Introduce a new MII time stamping interface.
  net: Add a layer for non-PHY MII time stamping drivers.
  dt-bindings: ptp: Introduce MII time stamping devices.
  net: mdio: of: Register discovered MII time stampers.
  net: Introduce peer to peer one step PTP time stamping.
  ptp: Add a driver for InES time stamping IP core.

 .../devicetree/bindings/ptp/ptp-ines.txt      |  35 +
 .../devicetree/bindings/ptp/timestamper.txt   |  42 +
 .../net/ethernet/broadcom/bnx2x/bnx2x_main.c  |   1 +
 drivers/net/ethernet/ti/netcp_ethss.c         |   8 +-
 drivers/net/macvlan.c                         |   4 +-
 drivers/net/phy/Makefile                      |   2 +
 drivers/net/phy/dp83640.c                     | 217 ++---
 drivers/net/phy/mii_timestamper.c             | 125 +++
 drivers/net/phy/phy.c                         |   4 +-
 drivers/net/phy/phy_device.c                  |   5 +
 drivers/of/of_mdio.c                          |  30 +-
 drivers/ptp/Kconfig                           |  10 +
 drivers/ptp/Makefile                          |   1 +
 drivers/ptp/ptp_ines.c                        | 852 ++++++++++++++++++
 include/linux/mii_timestamper.h               | 121 +++
 include/linux/phy.h                           |  85 +-
 include/uapi/linux/net_tstamp.h               |   8 +
 net/8021q/vlan_dev.c                          |   4 +-
 net/Kconfig                                   |   7 +-
 net/core/dev_ioctl.c                          |   1 +
 net/core/timestamping.c                       |  20 +-
 net/ethtool/ioctl.c                           |   4 +-
 22 files changed, 1432 insertions(+), 154 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/ptp/ptp-ines.txt
 create mode 100644 Documentation/devicetree/bindings/ptp/timestamper.txt
 create mode 100644 drivers/net/phy/mii_timestamper.c
 create mode 100644 drivers/ptp/ptp_ines.c
 create mode 100644 include/linux/mii_timestamper.h

Comments

David Miller Dec. 24, 2019, 11:47 p.m. UTC | #1
From: Richard Cochran <richardcochran@gmail.com>
Date: Sat, 21 Dec 2019 11:36:26 -0800

> This series adds support for PTP (IEEE 1588) P2P one-step time
> stamping along with a driver for a hardware device that supports this.
 ...

Series applied, thanks Richard.
David Miller Dec. 25, 2019, 12:15 a.m. UTC | #2
From: David Miller <davem@davemloft.net>
Date: Tue, 24 Dec 2019 15:47:13 -0800 (PST)

> From: Richard Cochran <richardcochran@gmail.com>
> Date: Sat, 21 Dec 2019 11:36:26 -0800
> 
>> This series adds support for PTP (IEEE 1588) P2P one-step time
>> stamping along with a driver for a hardware device that supports this.
>  ...
> 
> Series applied, thanks Richard.

Actually, I had to revert.  Richard, please fix this warning and resubmit:

drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c: In function ‘mlxsw_sp_ptp_get_message_types’:
drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c:915:2: warning: enumeration value ‘HWTSTAMP_TX_ONESTEP_P2P’ not handled in switch [-Wswitch]
  switch (tx_type) {
  ^~~~~~
Richard Cochran Dec. 26, 2019, 2:23 a.m. UTC | #3
On Tue, Dec 24, 2019 at 04:15:28PM -0800, David Miller wrote:
> Actually, I had to revert.  Richard, please fix this warning and resubmit:

Thanks for catching that.  I found yet another case (not flagged by
the compiler due to the driver not switching on the enum), and I put
the fixes into v9.

During my audit, I also found a couple of other unrelated problems
with drivers' HWTSTAMP ioctl implementations, but I'll put those up
separately.

Thanks,
Richard