Device-to-device link training

Patent metadata

Field	Value
Publication number	US-2020259936-A1
Application number	US-202016859689-A
Country	US
Kind code	A1
Filing date	Apr 27, 2020
Priority date	Apr 27, 2020
Publication date	Aug 13, 2020
Grant date	—

Abstract

Examples described herein relate to a network interface comprising physical medium dependent (PMD) circuitry, the PMD circuitry to during link training of at least one lane consistent with IEEE 802.3, exit to TIME_OUT state during TRAIN_LOCAL state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state. In some examples, during link training for at least one lane consistent with IEEE 802.3, the PMD circuitry is to exit to TIME_OUT state during TRAIN_REMOTE state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state. In some examples, link training consistent with IEEE 802.3 comprises performance of the PMD control function in Section 162.8.11 of IEEE 802.3ck.

First claim

What is claimed is:

1 . A method comprising:
at a first device, during link training of at least one lane consistent with IEEE 802.3 with a potential link partner, exiting to TIME_OUT state from TRAIN_LOCAL state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state.

2 . The method of claim 1 , comprising:
at the first device, during link training for at least one lane consistent with IEEE 802.3, exiting to TIME_OUT state from TRAIN_REMOTE state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state.

3 . The method of claim 1 , wherein the link training consistent with IEEE 802.3 comprises performance of a PMD control function in Section 162.8.11 of IEEE 802.3ck.

4 . The method of claim 1 , wherein the local_tf_lock state comprises indication of successful decoding of training frames received at the first device from the potential link partner.

5 . The method of claim 1 , wherein the remote_tf_lock state comprises indication of successful decoding of training frames from the first device by the potential link partner.

6 . The method of claim 1 , comprising restarting link training with a potential link partner after the TIME_OUT state.

7 . The method of claim 1 , wherein the first device comprises one or more of a switch, network interface, or server and the potential link partner comprises one or more of a switch, network interface, or server.

8 . The method of claim 1 , wherein a connection between the first device and the potential link partner comprises one or more of: a backplane or copper cable.

9 . The method of claim 1 , comprising:
at the first device, applying auto-negotiation with the potential link partner before the link training.

10 . The method of claim 1 , comprising:
receiving communication of loss of remote_tf_lock state via an out-of-band communication or from a system manager.

11 . An apparatus comprising:
physical medium dependent (PMD) circuitry, the PMD circuitry to:
during link training of at least one lane consistent with IEEE 802.3 with a second device, exit to TIME_OUT state from TRAIN_LOCAL state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state.

12 . The apparatus of claim 11 , comprising a physical layer interface (PHY) that comprises the PMD circuitry.

13 . The apparatus of claim 11 , comprising a network interface that comprises the PMD circuitry.

14 . The apparatus of claim 11 , wherein the PMD circuitry is to:
during link training for at least one lane consistent with IEEE 802.3 with the second device, exit to TIME_OUT state from TRAIN_REMOTE state based on consideration of expiration of a wait timer, loss of local_tf_lock state, and loss of remote_tf_lock state.

15 . The apparatus of claim 11 , wherein the link training consistent with IEEE 802.3 comprises performance of a PMD control function in Section 162.8.11 of IEEE 802.3ck.

16 . The apparatus of claim 11 , wherein the local_tf_lock state comprises indication of successful decoding of training frames received at the PMD circuitry from the second device.

17 . The apparatus of claim 11 , wherein the remote_tf_lock state comprises indication of successful decoding of training frames by the second device.

18 . The apparatus of claim 11 , wherein the PMD circuitry is to:
restart link training with the second device after the TIME_OUT state.

19 . The apparatus of claim 11 , wherein the PMD circuitry is to:
apply auto-negotiation with the second device before the link training.

20 . The apparatus of claim 11 , wherein the PMD circuitry is to:
receive communication of loss of remote_tf_lock state via an out-of-band communication or from a system manager.

21 . The apparatus of claim 11 , comprising a server, data center, switch, or rack.

Assignees

• Intel Corp

Inventors

• Lusted Kent C

Classifications

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• H04L69/324

  in the data link layer [OSI layer 2], e.g. HDLC · CPC title

• H04L49/351

  for local area network [LAN], e.g. Ethernet switches · CPC title

• H04L49/3054

  Auto-negotiation, e.g. access control between switch gigabit interface connector [GBIC] and link · CPC title