Transfer of information within an ASIC using a slotted ring based protocol

Patent metadata

Field	Value
Publication number	US-9608839-B2
Application number	US-201414527385-A
Country	US
Kind code	B2
Filing date	Oct 29, 2014
Priority date	Oct 29, 2014
Publication date	Mar 28, 2017
Grant date	Mar 28, 2017

Abstract

A system, comprising: a first local controller (LC) having a first position in a ring network and comprising a first LC cycle counter; a second LC having a second position in the ring network and comprising a second LC cycle counter; and a central controller (CC) connected to the ring network and comprising: a data structure linking the first LC to the first position and linking the second LC to the second position; and a CC cycle counter.

First claim

What is claimed is:

1. A system, comprising:
a first local controller (LC) having a first position in a ring network and comprising a first LC cycle counter,
wherein the first LC is configured to send a first event message on the ring network during a first clock cycle in response to the first LC cycle counter matching the first position during the first clock cycle;
a second LC having a second position in the ring network and comprising a second LC cycle counter,
wherein the second LC is configured to send a second event message on the ring network during a second clock cycle in response to the second LC cycle counter matching the second position during the second clock cycle; and

a central controller (CC) connected to the ring network and comprising:
a data structure linking the first LC to the first position and linking the second LC to the second position; and
a CC cycle counter configured to store the first position during a third clock cycle and to store the second position during a fourth clock cycle,
wherein the CC is configured to send, on the ring network, a first command to the first LC during the third clock cycle and a second command to the second LC during the fourth clock cycle.

2. The system of claim 1 , wherein the CC is further configured to send, on the ring network, a first start value to trigger a training phase, a second start value to start the first LC cycle counter and the second LC cycle counter, and a third start value to set a maximum value of the first LC cycle counter and the second LC cycle counter.

3. The system of claim 1 , further comprising:
a first local unit connected to the first LC and comprising a circuit configured to generate the first event message;
a second local unit connected to the second LC and comprising a circuit configured to generate the second event message;
a central unit connected to the CC and comprising a circuit configured to:
generate the first command based on an identification value of the first LC and the first event message; and
generate the second command based on an identification value of the second LC and the second event message.

4. The system of claim 1 , wherein the CC further comprises a Single Error Correcting (SEC) unit to perform error checking and correction on the first event message received from the ring network.

5. The system of claim 1 , wherein the first LC further comprises a Single Error Correcting (SEC) unit to perform error checking and correction on the first command received from the ring network.

6. The system of claim 1 , wherein the ring network comprises a fatal event line, wherein the first LC and the second LC only send event messages regarding fatal events in response to the fatal event line being asserted.

7. The system of claim 1 , wherein the CC, the first LC, the second LC, and the ring network are located on a single chip.

8. The system of claim 2 , wherein the second LC further comprises:
a position register;
a node ID register storing an identification value of the second LC; and
a node type register storing a type of the second LC,
wherein the second LC is further configured to:
receive, from the ring network, the first start value and a set of identification values including an identification value of the first LC;
determine the second position of the second LC based on a cardinality of the set;
store the second position in the position register;
concatenate the identification value of the second LC to the set of identification values; and
send, on the ring network, the first start value and the set of identification values.

9. The system of claim 3 , wherein the first LC further comprises a first event tracker to store the first event message received from the first local unit, and wherein the second LC further comprises a second event tracker to store the second event message received from the second local unit.

10. The system of claim 3 , wherein the ring network is a bus of 16 lines, wherein the first local unit is one selected from a group consisting of a link unit, a protocol unit, and a memory controller unit, and wherein the second local unit is one selected from the group.

11. The system of claim 8 , wherein the data structure in the CC stores the set of identification values received from the ring network.

12. A method of operating a central controller (CC) connected to a first local controller (LC) and a second local controller (LC) by a ring network, comprising:
receiving, by the CC and from the ring network, a first event message during a first clock cycle and a second event message during a second clock cycle,
wherein the first event message is generated by the first LC and the second event message is generated by the second LC;
reading, by the CC, a first value from a CC cycle counter during the first clock cycle and a second value from the CC cycle counter during the second clock cycle;
retrieving, by the CC and from a data structure in the CC, an identification value of the first LC based on the first value and an identification value of the second LC based on the second value;
sending, in response to the CC cycle counter having the first value during a third clock cycle, a first command on the ring network for the first LC during the third clock cycle; and
sending, in response to the CC cycle counter having the second value during a fourth clock cycle, a second command on the ring network for the second LC during the fourth clock cycle.

13. The method of claim 12 , further comprising:
sending, by the CC on the ring network, a first start value to trigger a training phase;
starting, by the CC, the CC cycle counter in response to sending the first start value;
receiving, by the CC and from the ring network, the first start value;
receiving, by the CC and from the ring network, the identification value of the first LC after the first start value is received,
wherein the first LC sends the identification value of the first LC on the ring network in response to the first start value;
receiving, by the CC and from the ring network, the identification value of the second LC after the identification value of the first LC is received, wherein the second LC sends the identification value of the second LC in response to the first start value; and
setting, by the CC, a maximum value of the CC cycle counter in response to receiving the first start value, the identification value of the first LC, and the identification value of the second LC.

14. The method of claim 12 , wherein the first event message is generated by a circuit in a local unit connected to the first LC, wherein the first event message is a fatal error, wherein the first command is generated by a central unit connected to the CC, and wherein the first command is a cease operations instruction for the local unit.

15. The method of claim 13 , further comprising:
sending, by the CC and on the ring network, a second start value,
wherein the second start value triggers a first LC cycle counter in the first LC, and
wherein the second start value triggers a second LC cycle counter in the second CL;
receiving on the ring network, by the CC, the second start value; and
sending, by the CC and on the ring network, a third start value in response to receiving the second start value to trigger an active phase.

16. The method of claim 14 , wherein the ring network is a bus of 16 lines, wherein the local unit is one selected from a group consisting of a link unit, a protocol unit, and a memory controller unit.

17. A method of operating a plurality of local contro

Assignees

• Oracle Int Corp

Inventors

• Schulz Jurgen
• Chandrasekhar Vishak
• Chiu Yu-Cheng

Classifications

• H04L41/0803

  Configuration setting · CPC title

• H04L12/423 Primary

  with centralised control, e.g. polling · CPC title