Method and device using a redundant cache for preventing the loss of dirty data

• US 5,895,485 A
• Filed: 02/24/1997
• Issued: 04/20/1999
• Est. Priority Date: 02/24/1997
• Status: Expired due to Term

• Current Assignee: NortonLifeLock Inc.
• Original Assignee: ECCS, INC.
• Inventors: Loechel, Barbara, Giovannetti, Federico
• Primary Examiner(s): Swann, Tod R.
• Assistant Examiner(s): LANGJAHR, DAVID
• Application Number:
  US08/805,146
• Time in Patent Office: 785 Days
• Field of Search: 395/873, 711/136, 711/140, 711/143, 711/119
• US Class Current: 711/119
• CPC Class Codes: G06F 12/0804 with main memory updating G06F12/0806 takes precedence
• CPC Class Codes: G06F 12/0815 Cache consistency protocols
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache

Abstract

A method and device for writing blocks from a client to a storage device. First and second caches are provided between the client and the storage device. An I/O management process accepts a write request from the client. If the write request is a write-back request, the I/O management process stores the blocks in the first cache, and transferring the blocks from the first cache to the second cache. If, on the other hand, the write request is not a write-back request, a write-through request is performed in which blocks are stored in the first cache, stored in the storage device, but not stored in the second cache. When the blocks are flushed from the first cache to store the blocks on the storage device, the dirty map of the cache header corresponding to the blocks is cleared from the second cache.

Claims

1. In a system having a first cache and a second cache arranged between a client and a persistent memory device, a method for managing input/output requests which prevents data loss, the method comprising steps of:

a) requesting a cache line in the first cache;

b) if the input/output request is a write request to a volume with write-back cache enabled, flagging one or more blocks of dirty data in the first cache;

c) performing a write operation;

d) transmitting the one or more blocks of dirty data from the first cache to the second cache based on the flagging; and

e) releasing the cache line.

2. The method of claim 1 wherein the step of flagging one or more blocks of dirty data in the first cache is achieved by updating a dirty map in the first cache to indicate the one or more blocks of dirty data in the first cache.

3. The method of claim 2 further comprising steps of:

determining whether the requested cache line results in a cache miss; and

if the requested cache line results in a cache miss,

i) finding an empty cache line, and

ii) clearing the dirty map and a valid map of the empty cache line.

4. The method of claim 1 further comprising a step of:

setting, in response to an I/O request to a volume with write-back cache enabled, a cache mode to write-back in a header of the cache line.

5. The method of claim 1 further comprising steps of:

configuring each of the first and second caches to have a cache directory and cache data, the cache directory including a table of headers, each of the headers corresponding to a cache line.

6. A system for providing storage for a client, the system comprising:

a) a mass storage device;

b) a first controller, including a cache, for controlling the mass storage device in response to requests from the client; and

c) a second controller, including a cache, for controlling the mass storage device in response to requests from the client,

wherein one of the first and second controllers responding to an I/O request, sends dirty data to the cache of another of the first and second controllers.

7. The system of claim 6 wherein, upon a flush of the cache of the one of the first and second controllers, the dirty map of the cache of the other of the first and second controllers is cleared.

8. The system of claim 6 wherein, upon a flush of a cache line of the cache of the one of the first and second controllers, the corresponding dirty map of the cache line header of the cache of the other of the first and second controllers is cleared.

9. The system of claim 6 wherein at least one of the first and second controllers includes:

i) a client input/output interface unit for facilitating communication with the client;

ii) a target input/output interface unit for facilitating communication with the mass storage device;

iii) a cache;

iv) an I/O management unit for controlling the cache and the I/O request; and

v) a bus system for permitting the communication of data between the client input/output interface unit, the target input/output interface unit, the cache and the I/O management unit.

10. The system of claim 9 wherein the at least one of the client input/output interface unit and the target input/output interface unit supports the SCSI protocol.

11. The system of claim 9 wherein the cache has a data section which includes cache lines, each cache line storing one or more blocks of data, and

wherein the cache directory section includes a table of headers, each of the headers corresponding to a cache line.

12. The system of claim 11 wherein each of the headers includes a dirty map which identifies blocks of dirty data in the cache line corresponding to the header.

13. The system of claim 6 wherein at least one of the first and second controllers includes:

i) at least one client input/output interface unit;

ii) at least one target input/output interface unit;

iii) a processor;

iv) a program storage;

v) a data storage;

vi) a local bus shared by the processor and the program storage, the data storage, the cache, at least one of the client I/O units, and at least one of the target I/O units;

vii) an input/output bus shared by the at least one client input/output interface unit and the at least one target input/output interface unit; and

viii) a SCSI bus shared by the at least one client input/output interface unit, the at least one target input/output interface unit, the data storage, and the cache.

14. The system of claim 13 wherein the cache has a data section which includes cache lines, each cache line storing one or more blocks of data, and

wherein the cache directory section includes a table of headers, each of the headers corresponding to a cache line.

15. The system of claim 14 wherein each of the headers includes a dirty map which identifies blocks of dirty data in the cache line corresponding to the header.

16. A method for writing data from a client to a storage device, the method comprising steps of:

a) providing a first cache between the client and the storage device;

b) providing a second cache between the client and the storage device;

c) accepting a write request from the client; and

d) if the write request is a write-back request, storing the data in the first cache, and transferring the data from the first cache to the second cache.

17. The method of claim 16 further comprising steps of:

e) flushing the data from the first cache to store the data on the storage device; and

f) clearing a dirty map of a cache header corresponding to the blocks of the data from the second cache.

18. The method of claim 17 wherein if the write request is not a write-back request, a write-through request is performed in which blocks are stored in the first cache, stored in the storage device, but not stored in the second cache.