Method and system for gathering and buffering sequential data for a transaction comprising multiple data access requests

• US 6,425,023 B1
• Filed: 03/24/1999
• Issued: 07/23/2002
• Est. Priority Date: 03/24/1999
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Leabo, Dell Patrick, Wade, Forrest Lee, Jones, Carl Evan, Batchelor, Gary William, Medlin, Robert Earl
• Primary Examiner(s): Gaffin, Jeffrey
• Assistant Examiner(s): Perveen, Rehana
• Application Number:
  US09/275,603
• Time in Patent Office: 1,217 Days
• Field of Search: 710/36, 710/33, 710/52, 710/39, 709/1, 709/2
• US Class Current: 710/39
• CPC Class Codes: G06F 13/4027 using bus bridges G06F13/4022 takes precedence

Abstract

Disclosed is a bridge system for processing read and write transactions over a bus in which in a preferred embodiment continuous read data obtained from a target device in a number of separate read operations over a secondary bus may be gathered by the bridge and assembled into a larger block of data before forwarding the data over the primary bus to the requesting agent. As a consequence, the transmission of optimal, address boundary-aligned bursts of read data over the primary bus may be increased and conversely, the transmission of fractionated, nonaligned read data over the primary bus may be reduced. Because each agent is assigned particular buffers, read data may be gathered concurrently in the assigned bridge buffers without assertion of a read request by one agent causing the flushing of the data being gathered for a different agent. In a similar manner, in delayed write operations, the write data may also be gathered from several write operations and assembled into an address boundary-aligned block of write data before the bridge circuit forwards the write data to the target device.

Claims

1. A method of processing a transaction involving the transfer of data between an agent and a target device, comprising:

processing the transaction as a plurality of data access requests;

gathering data for at least one data access request involved in the transaction at a plurality of sequential addresses and storing said data into a buffer until the addresses of said gathered data reach a boundary address; and

transferring said gathered data between the agent and target device to complete the transaction.

2. The method of claim 1, wherein the transaction is a read operation and the data access requests are read requests, wherein the step of gathering the data comprises gathering read data in a plurality of read requests from a plurality of sequential addresses at said target device, and wherein the step of transferring said gathered data comprises sending said gathered read data to said agent.

3. The method of claim 2, further comprising issuing a plurality of read requests to read data from said target device in response to the read operation, wherein the step of gathering the read data occurs in response to said plurality of read requests and further comprises storing said gathered plurality of blocks of data in a combined block of data until an address of said data stored in said buffer reaches a boundary, wherein the step of sending the gathered read data to said agent comprises sending the plurality of gathered blocks of data to said agent in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered blocks of data is aligned with an address boundary, and further comprising repeating said issuing, gathering and sending until all of the data requested by the agent in the read operation has been read, gathered and sent to the requesting agent.

4. The method of claim 1, wherein the transaction is a write operation and the data access requests are write requests, wherein the step of gathering the data comprises gathering write data having a plurality of sequential addresses from a plurality of write requests by said requesting agent and writing said gathered write data into a buffer until the addresses of said gathered write data reach a boundary address and wherein the step of transferring said gathered data comprises sending said gathered write data to said target device.

5. The method of claim 4, wherein the step of gathering said write data further comprises gathering said write data in response to a plurality of write requests until an address of said write data stored in said buffer reaches a boundary, wherein the step of sending the gathered write data to said target device comprises sending gathered write data in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered write data is aligned with a boundary address, further comprising repeating said gathering and sending for said requesting agent until all of the data requested to be written by said requesting agent has been written, gathered and sent to the target device.

6. The method of claim 1 wherein adjacent boundary addresses define a block length of said data transferred between the agent and target device as a predetermined length.

7. The method of claim 6 wherein said predetermined length is at least 32 bytes long.

8. The method of claim 6 wherein said predetermined length is 2^N bytes long where N is a positive integer greater than 4.

9. The method of claim 1 further comprising repeating said gathering and transferring until all of the data involved in the transaction has been gathered and transferred between the agent and the target device.

10. The method of claim 9 wherein said transferring of gathered data transfers data in a combined block in which at least one of the beginning and ending addresses of the transferred combined block of gathered data is aligned with a boundary address.

41. The method of claim 1, further comprising:

buffering a last address of data gathered in response to a previous data access request stored into a buffer;

receiving a subsequent data access request to a target address;

determining whether said target address is adjacent to said last address; and

if said target address is adjacent to said last address, then gathering data for the subsequent data access request at a plurality of sequential addresses and storing said data into said buffer until the addresses of said gathered data reaches said boundary address.

42. The method of claim 1, further comprising:

continuing to gather data subject to the read request if all of the data for the read request was not fetched before determining that the address of the gathered data reached the boundary data and transferring the buffered data.

43. The method of claim 1, wherein sequential data gathered from multiple read requests is stored in the buffer before the boundary address is reached and the data is transferred between the agent and the target device.

11. A method of processing transactions involving the transfer of data between a first agent and a second agent and a target device, comprising:

gathering data involved in a first transaction at a plurality of sequential addresses and storing said data into a first buffer until the addresses of said gathered data reach a boundary address, wherein the first buffer is assigned to store data exclusively for transactions involving the first agent;

gathering data involved in a second transaction at a plurality of sequential addresses and storing said data into a second buffer wherein said second buffer is assigned to store data exclusively for said second agent;

transferring at least a portion of said gathered data in said first buffer when said gathered data in said first buffer reaches a predetermined limit; and

transferring at least a portion of said gathered data in said second buffer when said gathered data in said second buffer reaches a predetermined limit.

12. The method of claim 11 wherein said predetermined limit is reached when the addresses of gathered data reach a predetermined boundary address.

13. A bridge system for processing transactions involving the transfer of data between an agent and a target device over a bus, comprising:

a buffer for storing data involved in transactions having a plurality of sequential addresses;

a logic circuit for processing the transaction as a plurality of data access requests;

gathering data for at least one data access request involved in the transaction at a plurality of sequential addresses and storing said data into a buffer until the addresses of said gathered data reach a boundary address; and

transferring said gathered data between the agent and target device to complete the transaction.

14. The bridge system of claim 13, wherein the transaction is a read operation and the data access requests are read requests, wherein the logic circuit gathers read data in a plurality of read requests from a plurality of sequential addresses at said target device and transfers said gathered read data to said agent.

15. The bridge system of claim 14, wherein the logic circuit further issues a plurality of read requests to read data from said target device in response to the read operation, gathers the read data in response to said plurality of read requests and stores said gathered plurality of blocks of data in a combined block of data until an address of said data stored in said buffer reaches a boundary, and sends the gathered read data to said agent comprises sending the plurality of gathered blocks of data to said agent in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered blocks of data is aligned with an address boundary, and repeats said issuing, gathering and sending until all of the data requested by the agent in the read operation has been read, gathered and sent to the requesting agent.

16. The bridge system of claim 13, wherein the transaction is a write operation and the data access requests are write requests, wherein the logic gathers write data having a plurality of sequential addresses from a plurality of write requests by said requesting agent and writes said gathered write data into a buffer until the addresses of said gathered write data reaches a boundary address, wherein the logic circuit transfers said gathered data comprises sending said gathered write data to said target device.

17. The bridge system of claim 16, wherein the logic circuit gathers said write data in response to a plurality of write requests until an address of said write data stored in said buffer reaches a boundary, sends gathered write data in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered write data is aligned with a boundary address, and repeats said gathering and sending for said requesting agent until all of the data requested to be written by said requesting agent has been written, gathered and sent to the target device.

18. The bridge system of claim 13 wherein adjacent boundary addresses define a block length of said data transferred between the agent and target device as a predetermined length.

19. The bridge system of claim 18 wherein said predetermined length is at least 32 bytes long.

20. The bridge system of claim 18 wherein said predetermined length is 2^N bytes long where N is a positive integer greater than 4.

21. The bridge system of claim 13 wherein the logic circuit further repeats said gathering and transferring until all of the data involved in the transaction has been gathered and transferred between the agent and the target device.

22. The bridge system of claim 21, wherein the logic circuit transfers gathered data in a combined block in which at least one of the beginning and ending addresses of the transferred combined block of gathered data is aligned with a boundary address.

44. The bridge system of claim 13, wherein the logic circuit further performs:

buffering a last address of data gathered in response to a previous data access request stored into a buffer;

receiving a subsequent data access request to a target address;

determining whether said target address is adjacent to said last address; and

if said target address is adjacent to said last address, then gathering data for the subsequent data access request at a plurality of sequential addresses and storing said data into said buffer until the addresses of said gathered data reaches said boundary address.

45. The bridge system of claim 13, wherein the logic circuit further performs:

continuing to gather data subject to the read request if all of the data for the read request was not fetched before determining that the address of the gathered data reached the boundary address and transferring the buffered data.

46. The bridge system of claim 13, wherein sequential data gathered from multiple read requests is stored in the buffer before the address boundary is reached and the data is transferred between the agent and the target device.

23. A bridge system for processing transactions involving the transfer of data between a first agent and a second agent and a target device over a bus, comprising:

a first buffer for storing data involved in transactions having a plurality of sequential addresses, wherein the first buffer is assigned to store data exclusively for transactions involving the first agent;

a second buffer for storing data involved in transactions having a plurality of sequential addresses, wherein said second buffer is assigned to store data exclusively for the second agent;

a logic circuit for gathering data involved in a transaction at a plurality of sequential addresses and storing said data into the first buffer until the addresses of said gathered data reach a boundary address;

transferring said gathered data between the first agent and target device to complete the transaction;

wherein the logic circuit further gathers data involved in a second transaction at a plurality of sequential addresses and stores said data into the second buffer, and wherein the logic circuit transfers said gathered data by transferring at least a portion of said gathered data in said first buffer when said gathered data in said first buffer reaches a predetermined limit and transferring at least a portion of said gathered data in said second buffer when said gathered data in said second buffer reaches a predetermined limit.

24. The bridge system of claim 23 wherein said predetermined limit is reached when the addresses of gathered data reach a predetermined boundary address.

25. An article of manufacture for use in programming a bridge system to process a transaction involving the transfer of data between an agent and a target device, the article of manufacture comprising computer useable media accessible to the bridge system, wherein the computer usable media includes at least one computer program that is capable of causing the bridge system to perform:

processing the transaction as a plurality of data access requests;

gathering data for at least one data access request involved in the transaction at a plurality of sequential addresses and storing said data into a buffer until the addresses of said gathered data reach a boundary address; and

transferring said gathered data between the agent and target device to complete the transaction.

26. The article of manufacture of claim 25, wherein the transaction is a read operation and the data access requests are read requests, wherein the step of gathering the data comprises gathering read data in a plurality of read requests from a plurality of sequential addresses at said target device, and wherein the step of transferring said gathered data comprises sending said gathered read data to said agent.

27. The article of manufacture of claim 26, further comprising issuing a plurality of read requests to read data from said target device in response to the read operation, wherein the step of gathering the read data occurs in response to said plurality of read requests and further comprises storing said gathered plurality of blocks of data in a combined block of data until an address of said data stored in said buffer reaches a boundary, wherein the step of sending the gathered read data to said agent comprises sending the plurality of gathered blocks of data to said agent in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered blocks of data is aligned with an address boundary, and further comprising repeating said issuing, gathering and sending until all of the data requested by the agent in the read operation has been read, gathered and sent to the requesting agent.

28. The article of manufacture of claim 25, wherein the transaction is a write operation and the data access requests are write requests, wherein the step of gathering the data comprises gathering write data having a plurality of sequential addresses from a plurality of write requests by said requesting agent and writing said gathered write data into a buffer until the addresses of said gathered write data reach a boundary address and wherein the step of transferring said gathered data comprises sending said gathered write data to said target device.

29. The article of manufacture of claim 28, wherein the step of gathering said write data further comprises gathering said write data in response to a plurality of write requests until an address of said write data stored in said buffer reaches a boundary, wherein the step of sending the gathered write data to said target device comprises sending gathered write data in a combined block in which at least one of the beginning and ending addresses of the sent combined block of gathered write data is aligned with a boundary address, further comprising repeating said gathering and sending for said requesting agent until all of the data requested to be written by said requesting agent has been written, gathered and sent to the target device.

30. The article of manufacture of claim 25, wherein adjacent boundary addresses define a block length of said data transferred between the agent and target device as a predetermined length.

31. The article of manufacture of claim 30 wherein said predetermined length is at least 32 bytes long.

32. The article of manufacture of claim 31 wherein said predetermined length is 2^N bytes long where N is a positive integer greater than 4.

33. The article of manufacture of claim 25, further comprising repeating said gathering and transferring until all of the data involved in the transaction has been gathered and transferred between the agent and the target device.

34. The article of manufacture of claim 33 wherein said transferring of gathered data transfers data in a combined block in which at least one of the beginning and ending addresses of the transferred combined block of gathered data is aligned with a boundary address.

47. The article of manufacture method of claim 25, further comprising:

buffering a last address of data gathered in response to a previous data access request stored into a buffer;

receiving a subsequent data access request to a target address;

determining whether said target address is adjacent to said last address; and

if said target address is adjacent to said last address, then gathering data for the subsequent data access request at a plurality of sequential addresses and storing said data into said buffer until the addresses of said gathered data reaches said boundary address.

48. The article of manufacture method of claim 25, further comprising:

continuing to gather data subject to the read request if all of the data for the read request was not fetched before determining that the address of the gathered data reached the boundary address and transferring the buffered data.

49. The article of manufacture method of claim 25, wherein sequential data gathered from multiple read requests is stored in the buffer before the address boundary is reached and the data is transferred between the agent and the target device.

35. An article of manufacture for use in programming a bridge system to process transactions involving the transfer of data between a first agent and a second agent and a target device, the article of manufacture comprising computer useable media accessible to the bridge system, wherein the computer usable media includes at least one computer program that is capable of causing the bridge system to perform:

gathering data involved in a first transaction at a plurality of sequential addresses and storing said data into a first buffer until the addresses of said gathered data reach a boundary address, wherein the first buffer is assigned to store data exclusively for transactions involving the first agent;

gathering data involved in a second transaction at a plurality of sequential addresses and storing said data into a second buffer wherein said second buffer is assigned to store data exclusively for a second agent;

transferring at least a portion of said gathered data in said first buffer when said gathered data in said first buffer reaches a predetermined limit; and

transferring at least a portion of said gathered data in said second buffer when said gathered data in said second buffer reaches a predetermined limit.

36. The article of manufacture of claim 35 wherein said predetermined limit is reached when the addresses of gathered data reach a predetermined boundary address.

37. A data processing system for processing a transaction involving the transfer of data;

comprising;

an adaptor for receiving the transaction;

a processor including processor memory;

a bridge system for processing transactions involving the transfer of data between the adaptor and the processor memory over a bus, comprising;

(i) a buffer for storing data involved in transactions having a plurality of sequential addresses; and

(ii) a logic circuit for processing the transaction as a plurality of data access requests;

gathering data for at least one data access request involved in the transaction at a plurality of sequential addresses and storing said data into a buffer until the addresses of said gathered data reach a boundary address; and

transferring said gathered data between the processor memory and the adaptor to complete the transaction.

38. The data processing system of claim 37, wherein the transaction is a read operation comprising the adaptor requesting data from the processor memory, wherein the logic circuit gathers read data in a plurality of read operations from a plurality of sequential addresses at said processor memory and transfers said gathered read data to said adaptor.

39. The data processing system of claim 37, wherein the transaction is a write operation comprising the adaptor writing data to the processor memory, wherein the logic gathers write data having a plurality of sequential addresses from a plurality of write operations by said requesting adaptor and writes said gathered write data into a buffer until the addresses of said gathered write data reaches a boundary address, wherein the logic circuit transfers said gathered data by sending said gathered write data to said processor memory.

40. A data processing system for processing transactions involving the transfer of data, comprising:

a first adaptor;

a second adaptor;

a processor including processor memory;

a bridge system for processing transactions involving the transfer of data between the first and second adaptors and the processor memory over a bus, comprising;

(i) a first buffer for storing data involved in transactions having a plurality of sequential addresses, wherein the first buffer is assigned to store data exclusively for transactions involving the first adaptor;

(ii) a second buffer for storing data involved in transactions having a plurality of sequential addresses, wherein said second buffer is assigned to store data exclusively for the second adaptor;

(iii) a logic circuit for gathering data involved in a first transaction at a plurality of sequential addresses and storing said data into the first buffer until the addresses of said gathered data reach a boundary address; and

transferring said gathered data between the processor memory and the first adaptor to complete the first transaction wherein the logic circuit further gathers data involved in a second transaction at a plurality of sequential addresses and stores said data into the second buffer, and wherein the logic circuit transfers said gathered data by transferring at least a portion of said gathered data in said first buffer when said gathered data in said first buffer reaches a predetermined limit and transferring at least a portion of said gathered data in said second buffer when said gathered data in said second buffer reaches a predetermined limit.