CRC error detection system and method

[0002] This invention relates generally to data storage systems, and more particularly to data storage systems having redundancy arrangements to protect against total system failure in the event of a failure in a component or subassembly of the storage system.

[0003] As is known in the art, large host computers and servers (collectively referred to herein as “host computer/servers”) require large capacity data storage systems. These large computer/servers generally includes data processors, which perform many operations on data introduced to the host computer/server through peripherals including the data storage system. The results of these operations are output to peripherals, including the storage system.

[0004] One type of data storage system is a magnetic disk storage system. Here a bank of disk drives and the host computer/server are coupled together through an interface. The interface includes “front end” or host computer/server controllers (or directors) and “back-end” or disk controllers (or directors). The interface operates the controllers (or directors) in such a way that they are transparent to the host computer/server. That is, data is stored in, and retrieved from, the bank of disk drives in such a way that the host computer/server merely thinks it is operating with its own local disk drive. One such system is described in U.S. Pat. No. 5,206,939, entitled “System and Method for Disk Mapping and Data Retrieval”, inventors Moshe Yanai, Natan Vishlitzky, Bruno Alterescu and Daniel Castel, issued Apr. 27, 1993, and assigned to the same assignee as the present invention.

[0005] As described in such U.S. Patent, the interface may also include, in addition to the host computer/server controllers (or directors) and disk controllers (or directors), addressable cache memories. The cache memory is a semiconductor memory and is provided to rapidly store data from the host computer/server before storage in the disk drives, and, on the other hand, store data from the disk drives prior to being sent to the host computer/server. The cache memory being a semiconductor memory, as distinguished from a magnetic memory as in the case of the disk drives, is much faster than the disk drives in reading and writing data.

[0006] The host computer/server controllers, disk controllers and cache memory are interconnected through a backplane printed circuit board. More particularly, disk controllers are mounted on disk controller printed circuit boards. The host computer/server controllers are mounted on host computer/server controller printed circuit boards. And, cache memories are mounted on cache memory printed circuit boards. The disk directors, host computer/server directors, and cache memory printed circuit boards plug into the backplane printed circuit board. In order to provide data integrity in case of a failure in a director, the backplane printed circuit board has a pair of buses. One set the disk directors is connected to one bus and another set of the disk directors is connected to the other bus. Likewise, one set the host computer/server directors is connected to one bus and another set of the host computer/server directors is directors connected to the other bus. The cache memories are connected to both buses. Each one of the buses provides data, address and control information.

[0007] The arrangement is shown schematically in FIG. 1. Thus, the use of two buses B1, B2 provides a degree of redundancy to protect against a total system failure in the event that the controllers or disk drives connected to one bus, fail. Further, the use of two buses increases the data transfer bandwidth of the system compared to a system having a single bus. Thus, in operation, when the host computer/server 12 wishes to store data, the host computer 12 issues a write request to one of the front-end directors 14 (i.e., host computer/server directors) to perform a write command. One of the front-end directors 14 replies to the request and asks the host computer 12 for the data. After the request has passed to the requesting one of the front-end directors 14, the director 14 determines the size of the data and reserves space in the cache memory 18 to store the request. The front-end director 14 then produces control signals on one of the address memory busses B1, B2 connected to such front-end director 14 to enable the transfer to the cache memory 18. The host computer/server 12 then transfers the data to the front-end director 14. The front-end director 14 then advises the host computer/server 12 that the transfer is complete. The front-end director 14 looks up in a Table, not shown, stored in the cache memory 18 to determine which one of the back-end directors 20 (i.e., disk directors) is to handle this request. The Table maps the host computer/server 12 addresses into an address in the bank 14 of disk drives. The front-end director 14 then puts a notification in a “mail box” (not shown and stored in the cache memory 18) for the back-end director 20, which is to handle the request, the amount of the data and the disk address for the data. Other back-end directors 20 poll the cache memory 18 when they are idle to check their “mail boxes”. If the polled “mail box” indicates a transfer is to be made, the back-end director 20 processes the request, addresses the disk drive in the bank 22, reads the data from the cache memory 18 and writes it into the addresses of a disk drive in the bank 22.

[0008] When data is to be read from a disk drive in bank 22 to the host computer/server 12 the system operates in a reciprocal manner. More particularly, during a read operation, a read request is instituted by the host computer/server 12 for data at specified memory locations (i.e., a requested data block). One of the front-end directors 14 receives the read request and examines the cache memory 18 to determine whether the requested data block is stored in the cache memory 18. If the requested data block is in the cache memory 18, the requested data block is read from the cache memory 18 and is sent to the host computer/server 12. If the front-end director 14 determines that the requested data block is not in the cache memory 18 (i.e., a so-called “cache miss”) and the director 14 writes a note in the cache memory 18 (i.e., the “mail box”) that it needs to receive the requested data block. The back-end directors 20 poll the cache memory 18 to determine whether there is an action to be taken (i.e., a read operation of the requested block of data). The one of the back- end directors 20 which poll the cache memory 18 mail box and detects a read operation reads the requested data block and initiates storage of such requested data block stored in the cache memory 18. When the storage is completely written into the cache memory 18, a read complete indication is placed in the “mail box” in the cache memory 18. It is to be noted that the front-end directors 14 are polling the cache memory 18 for read complete indications. When one of the polling front-end directors 14 detects a read complete indication, such front- end director 14 completes the transfer of the requested data which is now stored in the cache memory 18 to the host computer/server 12.

[0009] The use of mailboxes and polling requires time to transfer data between the host computer/server 12 and the bank 22 of disk drives thus reducing the operating bandwidth of the interface.

[0010] In accordance with the present invention, a method is provided for protecting erroneous data from being stored in a memory, such DATA comprising a series of data words terminating in a Cyclic Redundancy Check (CRC). The method includes: checking the CRC of the data words while delaying the DATA from passing to an output; corrupting the delayed data words if such checking determines a CRC error, such corruption of one of the data words being performed prior to the data words pass to said output; detecting whether such data word at the output is corrupt; and inhibiting storage of such data words in the memory if such one of the data words at the output is detected as being corrupt.

[0011] In one embodiment, the corrupting comprising corrupting a parity byte of such data words.

[0012] In accordance with another feature of the invention, a system is provided for protecting erroneous data from being stored in a memory, such DATA comprising a series of data words terminating in a Cyclic Redundancy Check (CRC). The system includes: a source of DATA, such DATA comprising a series of bytes each byte having a parity bit, such series of bytes terminating in a Cyclic Redundancy Check (CRC) portion associated with the series of bytes of the DATA; a source of a the CRC portion; a CRC checker fed by the series of bytes of the DATA and the source of the CRC portion, for determining a CRC from the series of bytes and for comparing such determined CRC with the CRC fed by the CRC source; a delay fed by the series of bytes and the parity bits thereof, a selector having a first input thereof fed by the parity bits and a second input thereof fed by the complement of such parity bits, such selector coupling the first input thereof to an output of such selector when the determined CRC is the same as the CRC fed by the CRC source and for coupling the second input thereof to the output when the determined CRC is different from the CRC fed by the CRC source, the output of the selector providing an appended parity bit for the data bytes after such data bytes pass through the delay; and, a detector and control logic for detecting whether such data word at the output is corrupt and for inhibiting storage of such data words in the memory if such one of the data words at the output is detected as being corrupt.

[0013] In accordance with another feature of the invention, a system is provided for protecting erroneous data from being stored in a memory, such DATA comprising a series of data words terminating in a Cyclic Redundancy Check (CRC). The system includes: a source of DATA, such DATA comprising a series of data words, each data word having a parity bit, each data word in the series being associated with a clock pulse, such series of data words terminating in a Cyclic Redundancy Check (CRC) portion associated with the series of bytes of the DATA, such CRC portion comprising a predetermined number of CRC words, each one of such CRC words being associated with one of the clock pulses; a source of a the CRC portion; a CRC checker fed by the series of data words and the source of the CRC portion, for determining a CRC from the series of data words and for comparing such determined CRC with the CRC fed by the CRC source; a delay fed by the series of DATA, such delay delaying the DATA by at least the number of CRC words; a selector having a first input thereof fed by the parity bits and a second input thereof fed by the complement of such parity bits, such selector coupling the first input thereof to an output of such selector when the determined CRC is the same as the CRC fed by the CRC source and for coupling the second input thereof to the output when the determined CRC is different from the CRC fed by the CRC source, the output of the selector providing an appended parity bit for the data words after such DATA has passed through the delay; and a detector and control logic for detecting whether such data word at the output is corrupt and for inhibiting storage of such data words in the memory if such one of the data words at the output is detected as being corrupt.

[0014] In one embodiment, a second selector is provided. The second selector has a first input fed the DATA and a second input fed by the output of the first-mentioned selector, such second selector coupling either the first input thereof or the second input thereof to an output of the second selector selectively in accordance with a control signal fed to such second selector.

[0015] The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.

[0016] These and other features of the invention will become more readily apparent from the following detailed description when read together with the accompanying drawings, in which:

[0017] FIG. 1 is a block diagram of a data storage system according to the PRIOR ART;

[0018] FIG. 2 is a block diagram of a data storage system according to the invention;

[0019] FIG. 2A shows the fields of a descriptor used in the system interface of the data storage system of FIG. 2;

[0020] FIG. 2B shows the filed used in a MAC packet used in the system interface of the data storage system of FIG. 2;

[0021] FIG. 3 is a sketch of an electrical cabinet storing a system interface used in the data storage system of FIG. 2;

[0022] FIG. 4 is a diagramatical, isometric sketch showing printed circuit boards providing the system interface of the data storage system of FIG. 2;

[0023] FIG. 5 is a block diagram of the system interface used in the data storage system of FIG. 2;

[0024] FIG. 6 shows the relationship between FIGS. 6A and 6B which when taken together is a block diagram showing the connections between front-end and back-end directors to one of a pair of message network boards used in the system interface of the data storage system of FIG. 2;

[0025] FIG. 7 is a block diagram of an exemplary one of the director boards used in the system interface of he data storage system of FIG. 2;

[0026] FIG. 8 is a block diagram of the system interface used in the data storage system of FIG. 2;

[0027] FIG. 8A is a diagram of an exemplary global cache memory board used in the system interface of FIG. 8;

[0028] FIG. 8B is a diagram showing a pair of director boards coupled between a pair of host processors and global cache memory boards used in the system interface of FIG. 8;

[0029] FIGS. 9A, 9B and 9C are a more detailed block diagram of the exemplary cache memory board of FIG. 8A;

[0030] FIG. 10 is a block diagram of a crossbar switch used in the memory board of FIGS. 9A, 9B and 9C;

[0031] FIGS. 11A, 11B, 11C and 11D are a block diagram of an upper port interface section used in the crossbar switch of FIG. 10;

[0032] FIGS. 12A, 12B, 12C and 12D are a block diagram of a lower port interface section used in the crossbar switch of FIG. 10;

[0033] FIGS. 13A, 13B, 13C, 13D and 13E are a block diagram of a pair of logic sections used in the memory board of FIGS. 9A, 9B and 9C;

[0034] FIGS. 14A, 14B, 14C and 14D are a block diagram of a pair of port controllers used in the pair of logic sections of FIGS. 13A, 13B, 13C, 13D and 13E;

[0035] FIGS. 15A, 15B, 15C, 15D and 15E are a block diagram of a pair of arbitration logics used in the pair of logic sections of FIGS. 13A, 13B, 13C, 13D and 13E and of a watchdog section used for such pair of logic sections;

[0036] FIG. 16 is a diagram showing words that make up exemplary information cycle used in the memory board of FIGS. 9A, 9B and 9C;

[0037] FIG. 17 is a Truth Table for a majority gate used in the memory board of FIGS. 9A, 9B and 9C;

[0038] FIG. 18 is a block diagram shown interconnections between one of the arbitration units used in one of the pair of port controllers of FIGS. 13A, 13B, 13C, 13D and 13E and a filter used in the arbitration unit of the other one of such pair of controllers of FIGS. 13A, 13B, 13C, 13D and 13E;

[0039] FIG. 19 is a timing diagram of signals in arbitration units of FIG. 18 used of one of the pair of port controllers of FIGS. 14A, 14B, 14C and 14D and a filter used in the arbitration unit used in the other one of such pair of controllers of FIGS. 14A, 14B, 14C and 14D; and

[0040] FIGS. 20A, 20B and 20C are a more detailed block diagram of arbitrations used in the arbitration logics of FIGS. 15A, 15B, 15C, 15D and 15E.

[0041] Referring now to FIG. 2, a data storage system 100 is shown for transferring data between a host computer/server 120 and a bank of disk drives 140 through a system interface 160. The system interface 160 includes: a plurality of, here 32 front-end directors 180₁-180₃₂coupled to the host computer/server 120 via ports-123₃₂; a plurality of back-end directors 200₁-200₃₂coupled to the bank of disk drives 140 via ports 123₃₃-123₆₄; a data transfer section 240, having a global cache memory 220, coupled to the plurality of front-end directors 180₁-180₁₆and the back-end directors 200₁-200₁₆; and a messaging network 260, operative independently of the data transfer section 240, coupled to the plurality of front-end directors 180₁-180₃₂and the plurality of back-end directors 200₁-200₃₂, as shown. The front-end and back-end directors 180₁-180₃₂, 200₁-200₃₂are functionally similar and include a microprocessor (μP) 299 (i.e., a central processing unit (CPU) and RAM), a message engine/CPU controller 314 and a data pipe 316 to be described in detail in connection with FIGS. 5, 6 and 7. Suffice it to say here, however, that the front-end and back-end directors 180₁-180₃₂, 200₁-200₃₂control data transfer between the host computer/server 120 and the bank of disk drives 140 in response to messages passing between the directors 180₁-180₃₂, 200₁-200₃₂through the messaging network 260. The messages facilitate the data transfer between host computer/server 120 and the bank of disk drives 140 with such data passing through the global cache memory 220 via the data transfer section 240. More particularly, in the case of the front-end directors 180₁-180₃₂, the data passes between the host computer to the global cache memory 220 through the data pipe 316 in the front-end directors 180₁-180₃₂and the messages pass through the message engine/CPU controller 314 in such front-end directors 180₁-180₃₂. In the case of the back-end directors 200₁-200₃₂the data passes between the back-end directors 200₁-200₃₂and the bank of disk drives 140 and the global cache memory 220 through the data pipe 316 in the back-end directors 200₁-200₃₂and again the messages pass through the message engine/CPU controller 314 in such back-end director 200₁-200₃₂.

[0042] With such an arrangement, the cache memory 220 in the data transfer section 240 is not burdened with the task of transferring the director messaging. Rather the messaging network 260 operates independent of the data transfer section 240 thereby increasing the operating bandwidth of the system interface 160.

[0043] In operation, and considering first a read request by the host computer/server 120 (i.e., the host computer/server 120 requests data from the bank of disk drives 140), the request is passed from one of a plurality of, here 32, host computer processors 121₁-121₃₂in the host computer 120 to one or more of the pair of the front-end directors 180₁-180₃₂connected to such host computer processor 121₁-121₃₂. (It is noted that in the host computer 120, each one of the host computer processors 121₁-121₃₂is coupled to here a pair (but not limited to a pair) of the front-end directors 180₁-180₃₂, to provide redundancy in the event of a failure in one of the front end-directors 1811-18132 coupled thereto. Likewise, the bank of disk drives 140 has a plurality of, here 32, disk drives 141₁-141₃₂, each disk drive 141₁-141₃₂being coupled to here a pair (but not limited to a pair) of the back-end directors 200₁-200₃₂, to provide redundancy in the event of a failure in one of the back-end directors 200₁-200₃₂coupled thereto). Each front-end director 180₁-180₃₂includes a microprocessor (μP) 299 (i.e., a central processing unit (CPU) and RAM) and will be described in detail in connection with FIGS. 5 and 7. Suffice it to say here, however, that the microprocessor 299 makes a request for the data from the global cache memory 220. The global cache memory 220 has a resident cache management table, not shown. Every director 180₁-180₃₂, 200₁-200₃₂has access to the resident cache management table and every time a front-end director 180₁-180₃₂requests a data transfer, the front-end director 180₁-180₃₂must query the global cache memory 220 to determine whether the requested data is in the global cache memory 220. If the requested data is in the global cache memory 220 (i.e., a read “hit”), the front-end director 180₁-180₃₂, more particularly the microprocessor 299 therein, mediates a DMA (Direct Memory Access) operation for the global cache memory 220 and the requested data is transferred to the requesting host computer processor 121₁-121₃₂.

[0044] If, on the other hand, the front-end director 180₁-180₃₂receiving the data request determines that the requested data is not in the global cache memory 220 (i.e., a “miss”) as a result of a query of the cache management table in the global cache memory 220, such front-end director 180₁-180₃₂concludes that the requested data is in the bank of disk drives 140. Thus the front-end director 180₁-180₃₂that received the request for the data must make a request for the data from one of the back-end directors 200₁-200₃₂in order for such back-end director 200₁-200₃₂to request the data from the bank of disk drives 140. The mapping of which back-end directors 200₁-200₃₂control which disk drives 141₁-141₃₂in the bank of disk drives 140 is determined during a power-up initialization phase. The map is stored in the global cache memory 220. Thus, when the front-end director 180₁-180₃₂makes a request for data from the global cache memory 220 and determines that the requested data is not in the global cache memory 220 (i.e., a “miss”), the front-end director 180₁-180₃₂is also advised by the map in the global cache memory 220 of the back-end director 200₁-200₃₂responsible for the requested data in the bank of disk drives 140. The requesting front-end director 180₁-180₃₂then must make a request for the data in the bank of disk drives 140 from the map designated back-end director 200₁-200₃₂. This request between the front-end director 180₁-180₃₂and the appropriate one of the back-end directors 200₁-200₃₂(as determined by the map stored in the global cache memory 200) is by a message which passes from the front-end director 180₁-180₃₂through the message network 260 to the appropriate back-end director 200₁-200₃₂. It is noted then that the message does not pass through the global cache memory 220 (i.e., does not pass through the data transfer section 240) but rather passes through the separate, independent message network 260. Thus, communication between the directors 180₁-180₃₂, 200₁-200₃₂is through the message network 260 and not through the global cache memory 220. Consequently, valuable bandwidth for the global cache memory 220 is not used for messaging among the directors 180₁-180₃₂, 200₁-200₃₂.

[0045] Thus, on a global cache memory 220 “read miss”, the front-end director 180₁-180₃₂sends a message to the appropriate one of the back-end directors 200₁-200₃₂through the message network 260 to instruct such back-end director 200₁-200₃₂to transfer the requested data from the bank of disk drives 140 to the global cache memory 220. When accomplished, the back-end director 200₁-200₃₂advises the requesting front-end director 180₁-180₃₂that the transfer is accomplished by a message, which passes from the back-end director 200₁-200₃₂to the front-end director 180₁-180₃₂through the message network 260. In response to the acknowledgement signal, the front-end director 180₁-180₃₂is thereby advised that such front-end director 180₁-180₃₂can transfer the data from the global cache memory 220 to the requesting host computer processor 121₁-121₃₂as described above when there is a cache “read hit”.

[0046] It should be noted that there might be one or more back-end directors 200₁-200₃₂responsible for the requested data. Thus, if only one back-end director 200₁-200₃₂is responsible for the requested data, the requesting front-end director 180₁-180₃₂sends a uni-cast message via the message network 260 to only that specific one of the back-end directors 200₁-200₃₂. On the other hand, if more than one of the back-end directors 200₁-200₃₂is responsible for the requested data, a multi-cast message (here implemented as a series of uni-cast messages) is sent by the requesting one of the front-end directors 180₁-180₃₂to all of the back-end directors 200₁-200₃₂having responsibility for the requested data. In any event, with both a uni-cast or multi-cast message, such message is passed through the message network 260 and not through the data transfer section 240 (i.e., not through the global cache memory 220).

[0047] Likewise, it should be noted that while one of the host computer processors 121₁-121₃₂might request data, the acknowledgement signal may be sent to the requesting host computer processor 121₁or one or more other host computer processors 121₁-121₃₂via a multi-cast (i.e., sequence of uni-cast) messages through the message network 260 to complete the data read operation.

[0048] Considering a write operation, the host computer 120 wishes to write data into storage (i.e., into the bank of disk drives 140). One of the front-end directors 180₁-180₃₂receives the data from the host computer 120 and writes it into the global cache memory 220. The front-end director 180₁-180₃₂then requests the transfer of such data after some period of time when the back-end director 200₁-200₃₂determines that the data can be removed from such cache memory 220 and stored in the bank of disk drives 140. Before the transfer to the bank of disk drives 140, the data in the cache memory 220 is tagged with a bit as “fresh data” (i.e., data which has not been transferred to the bank of disk drives 140, that is data which is “write pending”). Thus, if there are multiple write requests for the same memory location in the global cache memory 220 (e.g., a particular bank account) before being transferred to the bank of disk drives 140, the data is overwritten in the cache memory 220 with the most recent data. Each time data is transferred to the global cache memory 220, the front-end director 180₁-180₃₂controlling the transfer also informs the host computer 120 that the transfer is complete to thereby free-up the host computer 120 for other data transfers.

[0049] When it is time to transfer the data in the global cache memory 220 to the bank of disk drives 140, as determined by the back-end director 200₁-200₃₂, the back-end director 200₁-200₃₂transfers the data from the global cache memory 220 to the bank of disk drives 140 and resets the tag associated with data in the global cache memory 220 (i.e., un-tags the data) to indicate that the data in the global cache memory 220 has been transferred to the bank of disk drives 140. It is noted that the un-tagged data in the global cache memory 220 remains there until overwritten with new data.

[0050] Referring now to FIGS. 3 and 4, the system interface 160 is shown to include an electrical cabinet 300 having stored therein: a plurality of, here eight front-end director boards 190₁-190₈, each one having here four of the front-end directors 180₁-180₃₂; a plurality of, here eight back-end director boards 210₁-210₈, each one having here four of the back-end directors 200₁-200₃₂; and a plurality of, here eight, memory boards 220′ which together make up the global cache memory 220. These boards plug into the front side of a backplane 302. (It is noted that the backplane 302 is a mid-plane printed circuit board). Plugged into the backside of the backplane 302 are message network boards 304₁, 304₂. The backside of the backplane 302 has plugged into it adapter boards, not shown in FIGS. 2-4, which couple the boards plugged into the back-side of the backplane 302 with the computer 120 and the bank of disk drives 140 as shown in FIG. 2. That is, referring again briefly to FIG. 2, an I/O adapter, not shown, is coupled between each one of the front-end directors 180₁-180₃₂and the host computer 120 and an I/O adapter, not shown, is coupled between each one of the back-end directors 200₁-200₃₂and the bank of disk drives 140.

[0051] Referring now to FIG. 5, the system interface 160 is shown to include the director boards 190₁-190₈, 210₁-210₈and the global cache memory 220, plugged into the backplane 302 and the disk drives 141₁-141₃₂in the bank of disk drives along with the host computer 120 also plugged into the backplane 302 via I/O adapter boards, not shown. The message network 260 (FIG. 2) includes the message network boards 304₁and 304₂. Each one of the message network boards 304₁and 304₂is identical in construction. A pair of message network boards 304₁and 304₂is used for redundancy and for message load balancing. Thus, each message network board 304₁, 304₂, includes a controller 306, (i.e., an initialization and diagnostic processor comprising a CPU, system controller interface and memory, as shown in FIG. 6 for one of the message network boards ³⁰⁴₁, 304₂, here board 304₁) and a crossbar switch section 308 (e.g., a switching fabric made up of here four switches 308₁-308₄).

[0052] Referring again to FIG. 5, each one of the director boards 190₁-210₈includes, as noted above four of the directors 180₁-180₃₂, 200₁-200₃₂(FIG. 2). It is noted that the director boards 190₁-190₈having four front-end directors per board, 180₁-180₃₂are referred to as front-end directors and the director boards 210₁-210₈having four back-end directors per board, 200₁-200₃₂are referred to as back-end directors. Each one of the directors 180₁-180₃₂, 200₁-200₃₂includes a CPU 310, a RAM 312 (which make up the microprocessor 299 referred to above), the message engine/CPU controller 314, and the data pipe 316.

[0053] Each one of the director boards 190₁-210₈includes a crossbar switch 318. The crossbar switch 318 has four input/output ports 319, each one being coupled to the data pipe 316 of a corresponding one of the four directors 180₁-180₃₂, 200₁-200₃₂on the director board. 190₁-210₈. The crossbar switch 318 has eight output/input ports collectively identified in FIG. 5 by numerical designation 321 (which plug into the backplane 302. The crossbar switch 318 on the front-end director boards 191₁-191₈is used for coupling the data pipe 316 of a selected one of the four front-end directors 180₁-180₃₂on the front-end director board 190₁-190₈to the global cache memory 220 via the backplane 302 and I/O adapter, not shown. The crossbar switch 318 on the back-end director boards 210₁-210₈is used for coupling the data pipe 316 of a selected one of the four back-end directors 200₁-200₃₂on the back-end director board 210₁-210₈to the global cache memory 220 via the backplane 302 and I/O adapter, not shown. Thus, referring to FIG. 2, the data pipe 316 in the front-end directors 180₁-180₃₂couples data between the host computer 120 and the global cache memory 220 while the data pipe 316 in the back-end directors 200₁-200₃₂couples data between the bank of disk drives 140 and the global cache memory 220. It is noted that there are separate point-to-point data paths P₁-P₆₄(FIG. 2) between each one of the directors 180₁-180₃₂, 200₁-200₃₂and the global cache memory 220. It is also noted that the backplane 302 is a passive backplane because it is made up of only etched conductors on one or more layers of a printed circuit board. That is, the backplane 302 does not have any active components.

[0054] Referring again to FIG. 5, each one of the director boards 190₁-210₈includes a crossbar switch 320. Each crossbar switch 320 has four input/output ports 323, each one of the four input/output ports 323 being coupled to the message engine/CPU controller 314 of a corresponding one of the four directors 180₁-180₃₂, 200₁-200₃₂on the director board 190₁-210₈. Each crossbar switch 320 has a pair of output/input ports 325₁, 325₂, which plug into the backplane 302. Each port 325₁-325₂is coupled to a corresponding one of the message network boards 304₁, 304₂, respectively, through the backplane 302. The crossbar switch 320 on the front-end director boards 190₁-190₈is used to couple the messages between the message engine/CPU controller 314 of a selected one of the four front-end directors 180₁-180₃₂on the front-end director boards 190₁-190₈and the message network 260, FIG. 2. Likewise, the back-end director boards 210₁-210₈are used to couple the messages produced by a selected one of the four back-end directors 200₁-200₃₂on the back-end director board 210₁-210₈between the message engine/CPU controller 314 of a selected one of such four back-end directors and the message network 260 (FIG. 2). Thus, referring also to FIG. 2, instead of having a separate dedicated message path between each one of the directors 180₁-180₃₂, 200₁-200₃₂and the message network 260 (which would require M individual connections to the backplane 302 for each of the directors, where M is an integer), here only M/4 individual connections are required). Thus, the total number of connections between the directors 180₁-180₃₂, 200₁-200₃₂and the backplane 302 is reduced to ¼th. Thus, it should be noted from FIGS. 2 and 5 that the message network 260 (FIG. 2) includes the crossbar switch 320 and the message network boards 304₁, 304₂.

[0055] Each message is a 64-byte descriptor, shown in FIG. 2A, which is created by the CPU 310 (FIG. 5) under software control and is stored in a send queue in RAM 312. When the message is to be read from the send queue in RAM 312 and transmitted through the message network 260 (FIG. 2) to one or more other directors via a DMA operation to be described, it is packetized in the packetizer portion of packetizer/de-packetizer 428 (FIG. 7) into a MAC type packet, shown in FIG. 2B, here using the NGIO protocol specification. There are three types of packets: a message packet section; an acknowledgement packet; and a message network fabric management packet, the latter being used to establish the message network routing during initialization (i.e., during power-up). Each one of the MAC packets has: an 8-byte header which includes source (i.e., transmitting director) and destination (i.e., receiving director) address; a payload; and terminates with a 4-byte Cyclic Redundancy Check (CRC), as shown in FIG. 2B. The acknowledgement packet (i.e., signal) has a 4-byte acknowledgment payload section. The message packet has a 32-byte payload section. The Fabric Management Packet (FMP) has a 256-byte payload section. The MAC packet is sent to the crossbar switch 320. The destination portion of the packet is used to indicate the destination for the message and is decoded by the switch 320 to determine which port the message is to be routed. The decoding process uses a decoder table 327 in the switch 318, such table being initialized by controller during power-up by the initialization and diagnostic processor (controller) 306 (FIG. 5). The table 327 (FIG. 7) provides the relationship between the destination address portion of the MAC packet, which identifies the routing for the message and the one of the four directors 180₁-180₃₂, 200₁-200₃₂on the director board 190₁-190₈, 210₁-210₈or to one of the message network boards 304₁, 304₂to which the message is to be directed.

[0056] More particularly, and referring to FIG. 5, a pair of output/input ports 325₁, 325₂is provided for each one of the crossbar switches 320, each one being coupled to a corresponding one of the pair of message network boards 304₁, 304₂. Thus, each one of the message network boards 304₁, 304₂has sixteen input/output ports 322₁-322₁₆, each one being coupled to a corresponding one of the output/input ports 325₁, 325₂, respectively, of a corresponding one of the director boards 190₁-190₈, 210₁-210₈through the backplane 302, as shown. Thus, considering exemplary message network board 304₁, FIG. 6, each switch 308₁-308₄also includes three coupling ports 324₁-324₃. The coupling ports 324₁-324₃are used to interconnect the switches 322₁-322₄, as shown in FIG. 6. Thus, considering message network board 304₁, input/output ports 322₁-322₈are coupled to output/input ports 325, of front-end director boards 190_i-190₈and input/output ports 322₉-322₁₆are coupled to output/input ports 325₁of back-end director boards 210₁-210₈, as shown. Likewise, considering message network board 304₂, input/output ports 322₁-322₈thereof are coupled, via the backplane 302, to output/input ports 325₂of front-end director boards 190₁-190₈and input/output ports 322₉-322₁₆are coupled, via the backplane 302, to output/input ports 325₂of back-end director boards 210₁-210₈.

[0057] As noted above, each one of the message network boards 304₁, 304₂includes a processor 306 (FIG. 5) and a crossbar switch section 308 having four switches 308₁-308₄, as shown in FIGS. 5 and 6. The switches 308₁-308₄are interconnected as shown so that messages can pass between any pair of the input/output ports 322₁-322₁₆. Thus, it follow that a message from any one of the front-end directors 180₁-180₃₂can be coupled to another one of the front-end directors 180₁-180₃₂and/or to any one of the back-end directors 200₁-200₃₂. Likewise, a message from any one of the back-end directors 180₁-180₃₂can be coupled to another one of the back-end directors 180₁-180₃₂and/or to any one of the front-end directors 200₁-200₃₂.

[0058] As noted above, each MAC packet (FIG. 2B) includes in an address destination portion and a data payload portion. The MAC header is used to indicate the destination for the MAC packet and such MAC header is decoded by the switch to determine which port the MAC packet is to be routed. The decoding process uses a table in the switch 308₁-308₄, such table being initialized by processor 306 during power-up. The table provides the relationship between the MAC header, which identifies the destination for the MAC packet and the route to be taken through the message network. Thus, after initialization, the switches 320 and the switches 308₁-308₄in switch section 308 provides packet routing which enables each one of the directors 180₁-180₃₂, 200₁-200₃₂to transmit a message between itself and any other one of the directors, regardless of whether such other director is on the same director board 190₁-190₈, 210₁-210₈or on a different director board. Further, the MAC packet has an additional bit B in the header thereof, as shown in FIG. 2B, which enables the message to pass through message network board 304₁or through message network board 304₂. During normal operation, this additional bit B is toggled between a logic 1 and a logic 0 so that one message asses through one of the redundant message network boards 304₁, 304₂and the next message to pass through the other one of the message network boards 304₁, 304₂to balance he load requirement on the system. However, in the event of a failure in one of the message network boards 304₁, 304₂, the non-failed one of the boards 304₁, 304₂is used exclusively until the failed message network board is replaced.

[0059] Referring now to FIG. 7, an exemplary one of the director boards 190₁-190₈, 210₁-210₈, here director board 190₁is shown to include directors 180₁, 180₃, 180₅and 180₇. An exemplary one of the directors 180₁-180₄, here director 180₁is shown in detail to include the data pipe 316, the message engine/CPU controller 314, the RAM 312, and the CPU 310 all coupled to the CPU interface bus 317, as shown. The exemplary director 180₁also includes: a local cache memory 319 (which is coupled to the CPU 310); the crossbar switch 318; and, the crossbar switch 320, described briefly above in connection with FIGS. 5 and 6. The data pipe 316 includes a protocol translator 400, a quad port RAM 402 and a quad port RAM controller 404 arranged as shown. Briefly, the protocol translator 400 converts between the protocol of the host computer 120, in the case of a front-end director 180₁-180₃₂, (and between the protocol used by the disk drives in bank 140 in the case of a back-end director 200₁-200₃₂) and the protocol between the directors 180₁-180₃, 200₁-200₃₂and the global memory 220 (FIG. 2). More particularly, the protocol used the host computer 120 may, for example, be fibre channel, SCSI, ESCON or FICON, for example, as determined by the manufacture of the host computer 120 while the protocol used internal to the system interface 160 (FIG. 2) may be selected by the manufacturer of the interface 160. The quad port RAM 402 is a FIFO controlled by controller 404 because the rate data coming into the RAM 402 may be different from the rate data leaving the RAM 402. The RAM 402 has four ports, each adapted to handle an 18 bit digital word. Here, the protocol translator 400 produces 36 bit digital words for the system interface 160 (FIG. 2) protocol, one 18 bit portion of the word is coupled to one of a pair of the ports of the quad port RAM 402 and the other 18 bit portion of the word is coupled to the other one of the pair of the ports of the quad port RAM 402. The quad port RAM has a pair of ports 402A, 402B, each one of to ports 402A, 402B being adapted to handle an 18 bit digital word. Each one of the ports 402A, 402B is independently controllable and has independent, but arbitrated, access to the memory array within the RAM 402. Data is transferred between the ports 402A, 402B and the cache memory 220 (FIG. 2) through the crossbar switch 318, as shown.

[0060] The crossbar switch 318 includes a pair of switches 406A, 406B. Each one of the switches 406A, 406B includes four input/output director-side ports D₁-D₄(collectively referred to above in connection with FIG. 5 as port 319) and four input/output memory-side ports M₁-M₄, M₅-M₈, respectively, as indicated. The input/output memory-side ports M₁-M₄, M₅-M₈were collectively referred to above in connection with FIG. 5 as port 317). The director-side ports D₁-D₄of switch 406A are connected to the 402A ports of the quad port RAMs 402 in each one the directors 180₁, 180₃, 180₅and 180₇, as indicated. Likewise, director-side ports of switch 406B are connected to the 402B ports of the quad port RAMs 402 in each one the directors 180₁, 180₃, 180₅, and 180₇, as indicated. The ports D₁-D₄are selectively coupled to the ports M₁-M₄in accordance with control words provided to the switch 406A by the controllers in directors 180₁, 180₃, 180₅, 180₇on busses R_A1-R_A4, respectively, and the ports D₁-D₄are coupled to ports M₅-M₈in accordance with the control words provided to switch 406B by the controllers in directors 180₁, 180₃, 180₅, 180₇on busses R_B1-R_B4, as indicated. The signals on buses R_A1-R_A4are request signals. Thus, port 402A of any one of the directors 180₁, 180₃, 180₅, 180₇may be coupled to any one of the ports M₁-M₄of switch 406A, selectively in accordance with the request signals on buses R_A1-R_A4. Likewise, port 402B of any one of the directors 180₁-180₄may be coupled to any one of the ports M₅-M₈of switch 406B, selectively in accordance with the request signals on buses R_B1-R_B4. The coupling between the director boards 190₁-190₈, 210₁-210₈and the global cache memory 220 is shown in FIG. 8.

[0061] More particularly, and referring also to FIG. 2, as noted above, each one of the host computer processors 121₁-121₃₂in the host computer 120 is coupled to a pair of the front-end directors 180₁-180₃₂, to provide redundancy in the event of a failure in one of the front end-directors 181₁-181₃₂coupled thereto. Likewise, the bank of disk drives 140 has a plurality of, here 32, disk drives 141₁-141₃₂, each disk drive 141₁-141₃₂being coupled to a pair of the back-end directors 200₁-200₃₂, to provide redundancy in the event of a failure in one of the back-end directors 200₁-200₃₂coupled thereto). Thus, considering exemplary host computer processor 121₁, such processor 121₁is coupled to a pair of front-end directors 180₁, 180₂. Thus, if director 180₁fails, the host computer processor 121₁can still access the system interface 160, albeit by the other front-end director 180₂. Thus, directors 180, and 180₂are considered redundancy pairs of directors. Likewise, other redundancy pairs of front-end directors are: front-end directors 180₃, 180₄; 180₅, 180₆; 180₇, 180₈; 180₉, 180₁₀; 180₁₁, 180₁₂; 180₁₃, 180₁₄; 180₁₅, 180₁₆; 180₁₇, 180₁₈; 180₁₉, 180₂₀; 180₂₁, 180₂₂; 180₂₃, 180₂₄; 180₂₅, 180₂₆; 180₂₇, 180₂₈; 180₂₉, 180₃₀; and 180₃₁, 180₃₂(only directors 180₃₁and 180₃₂being shown in FIG. 2).

[0062] Likewise, disk drive 141₁is coupled to a pair of back-end directors 200₁, 200₂. Thus, if director 200₁fails, the disk drive 141₁can still access the system interface 160, albeit by the other back-end director 180₂. Thus, directors 200₁and 200₂are considered redundancy pairs of directors. Likewise, other redundancy pairs of back-end directors are: back-end directors 200₃, 200₄; 200₅, 200₆; 200₇, 200₈; 200₉, 200₁₀; 200₁₁, 200₁₂; 200₁₃, 200₁₄; 200₁₅, 200₁₆; 200₁₇, 200₁₈; 200₁₉, 200₂₀; 200₂₁, 200₂₂; 200₂₃, 200₂₄; 200₂₅, 200₂₆; 200₂₇, 200₂₈; 200₂₉, 200₃₀; and 200₃₁, 200₃₂(only directors 200₃₁and 200₃₂being shown in FIG. 2). Further, referring also to FIG. 8, the global cache memory 220 includes a plurality of, here eight, cache memory boards 220₁-220₈, as shown. Still further, referring to FIG. 8A, an exemplary one of the cache memory boards, here board 220₁is shown in detail and will be described in detail in connection with FIGS. 23-29. Here, each cache memory board includes four memory array regions, an exemplary one thereof being shown and described in connection with FIG. 6 of U. S. Pat. No. 5,943,287 entitled “Fault Tolerant Memory System”, John K. Walton, inventor, issued Aug. 24, 1999 and assigned to the same assignee as the present invention, the entire subject matter therein being incorporated herein by reference. Further detail of the exemplary one of the cache memory boards.

[0063] As shown in FIG. 8A, the board 220₁includes a plurality of, here four RAM memory arrays, each one of the arrays has a pair of redundant ports, i.e., an A port and a B port. The board itself has sixteen ports; a set of eight A ports M_A1-M_A8and a set of eight B ports M_B1-M_B8. Four of the eight A port, here A ports M_A1-M_A4are coupled to the M₁port of each of the front-end director boards 190₁, 190₃, 190₅, and 190₇, respectively, as indicated in FIG. 8. Four of the eight B port, here B ports M_B1-M_B4are coupled to the M₁port of each of the front-end director boards 190₂, 190₄, 190₆, and 190₈, respectively, as indicated in FIG. 8. The other four of the eight A port, here A ports M_A5-M_A8are coupled to the M₁port of each of the back-end director boards 210₁, 210₃, 210₅, and 210₇, respectively, as indicated in FIG. 8. The other four of the eight B port, here B ports MB_B5-M₄₈are coupled to the M₁port of each of the back-end director boards 210₂, 210₄, 210₆, and 210₈, respectively, as indicated in FIG. 8

[0064] Considering the exemplary four A ports M_A1-M_A4, each one of the four A ports M_A1-M_A4can be coupled to the A port of any one of the memory arrays through the logic network 221_1A, to be described in more detail in connection with FIGS. 25, 126 and 27. Thus, considering port M_A1, such port can be coupled to the A port of the four memory arrays. Likewise, considering the four A ports M_A5-M_A8, each one of the four A ports M_A5-M_A8can be coupled to the A port of any one of the memory arrays through the logic network 221_1BLikewise, considering the four B ports M_B1-M_B4, each one of the four B ports M_B3-M_B4can be coupled to the B port of any one of the memory arrays through logic network 221_1B. Likewise, considering the four B ports M_B5-M_B8, each one of the four B ports M_B5-M_B8can be coupled to the B port of any one of the memory arrays through the logic network 221_2B. Thus, considering port M_B1, such port can be coupled to the B port of the four memory arrays. Thus, there are two paths data and control from either a front-end director 180₁-180₃₂or a back-end director 200₁-200₃₂can reach each one of the four memory arrays on the memory board. Thus, there are eight sets of redundant ports on a memory board, i.e., ports M_A1, M_B1; M_A2, M_B2; M_A3, M_B3; M_A4, M_B4; M_A5, M_B5; M_A6, M_B6; M_A7, M_B7; and M_A8, M_B8. Further, as noted above each one of the directors has a pair of redundant ports, i.e. a 402A port and a 402 B port (FIG. 7). Thus, for each pair of redundant directors, the A port (i.e., port 402A) of one of the directors in the pair is connected to one of the pair of redundant memory ports and the B port (i.e., 402B) of the other one of the directors in such pair is connected to the other one of the pair of redundant memory ports.

[0065] More particularly, referring to FIG. 8B, an exemplary pair of redundant directors is shown, here, for example, front-end director 180₁and front end-director 180₂. It is first noted that the directors 180₁, 180₂in each redundant pair of directors must be on different director boards, here boards 190₁, 190₂, respectively. Thus, here front-end director boards 190₁-190₈have thereon: front-end directors 180₁, 180₃, 180₅and 180₇; front-end directors 180₂, 180₄, 180₆and 180₈; front end directors 180₉, 180₁₁, 180₁₃and 180₁₅; front end directors 180₁₀, 180₁₂, 180₁₄and 180₁₆; front-end directors 180₁₇, 180₁₉, 180₂₁and 180₂₃; front-end directors 180₁₈, 180₂₀, 180₂₂and 180₂₄; front-end directors 180₂₅, 180₂₇, 180₂₉and 180₃₁; front-end directors 180₁₈, 180₂₀, 180₂₂and 180₂₄. Thus, here back-end director boards 210₁-210₈have thereon: back-end directors 200₁, 200₃, 200₅and 200₇; back-end directors 200₂, 200₄, 200₆and 200₈; back-end directors 200₉, 200₁₁, 200₁₃and 200₁₅; back-end directors 200₁₀, 200₁₂, 200₁₄and 200₁₆; back-end directors 200₁₇, 200₁₉, 200₂₁and 200₂₃; back-end directors 200₁₈, 200₂₀, 200₂₂and 200₂₄; back-end directors 200₂₅, 200₂₇, 200₂₉and 200₃₁; back-end directors 200₁₈, 200₂₀, 200₂₂and 200₂₄.

[0066] Thus, here front-end director 1801, shown in FIG. 8A, is on front-end director board 190₁and its redundant front-end director 180₂, shown in FIG. 8B, is on anther front-end director board, here for example, front-end director board 190₂. As described above, the port 402A of the quad port RAM 402 (i.e., the A port referred to above) is connected to switch 406A of crossbar switch 318 and the port 402B of the quad port RAM 402 (i.e., the B port referred to above) is connected to switch 406B of crossbar switch 318. Likewise, for redundant director 180₂, However, the ports M₁-M₄of switch 406A of director 180₁are connected to the M_A1ports of global cache memory boards 220₁-200₄, as shown, while for its redundancy director 180₂, the ports M₁-M₄of switch 406A are connected to the redundant M_B1ports of global cache memory boards 220₁-200₄, as shown.

[0067] Further details are provided in co-pending patent application Ser. No 09/561,531 filed Apr. 28, 2000 and Ser. No. 09/561,161 assigned to the same assignee as the present patent application, the entire subject matter thereof being incorporated herein by reference.

[0068] Referring again to FIG. 8, the system includes a plurality of, here eight, memory boards. As described above in connection with FIG. 8A, each one of the memory boards includes four memory array regions R₁-R₄. Referring now to FIGS. 9A, 9B and 90C, an exemplary one of the cache memory boards in the cache memory 220 (FIG. 8), here cache memory board 220₁, is shown in more detail to include, here, the four logic networks 221_1B, 221_2B, 221_1A, and 221_2Aand, here eight interface, or memory region control, sections, here logic sections 5010₁-5010₈, arranged as shown.

[0069] Each one of the four logic networks 221_1B, 221_2B, 221_1A, and 221_2Aincludes four sets of serial-to-parallel converters (S/P), each one of the sets having four of the S/P converters. The sets of S/P converters are coupled between ports M_B1-M_B4, M_B5-M_B8, M_A1-M_A4, and M_A5-M_A5, respectively, and a corresponding one of four crossbar switches 5004₁-5004₄. The S/Ps convert between a serial stream of information (i.e., data, address, and control, Cyclic Redundancy Checks (CRCs), signaling semaphores, etc.) at ports M_B1-M_B8, M_A1-M_A8, and a parallel stream of the information which passes through the crossbar switches 5004₁-5004₄. Thus, here the crossbar switches 5004₁-5004₄process parallel information. Information is transferred between directors and the crossbar switches as transfers, or information cycles. An exemplary information transfer for information passing for storage in the memory array region is shown in FIG. 16. Each information cycle is shown to include a plurality of sixteen bit words, each word being associated with a clock pulse. Thus, first word 0 is shown to include protocol signaling (e.g., semaphore) and a terminating “start-frame” indication. The next word 1 includes memory control information. The next three words, 2-4, include memory address (ADDR) information. The next word, 5, is a “tag” which indicated the memory board, memory array region, and other information to be described. The next two words, 6 and 7, provide Cyclic Redundancy Checks (CRC) information regarding the address (ADDR_CRC). The DATA to be written into the memory then follows. The number of words of DATA is variable and here is between 4 words and 256 words. The information cycle terminates with two words, X and Y which include DATA CRC information.

[0070] As will be described in more detail below, the cache memory board 220, is a multi-ported design which allows equal access to one of several, here four, regions of memory (i.e., here memory array regions R₁-R₄) from any of here sixteen ports M_B1-M_B8, M_A1-M_A8. The sixteen ports M_B1-M_B8, M_A1-M_A8are grouped into four sets S₁-S4. Each one of the sets S₁-S₄is associated with, i.e., coupled to, a corresponding one of the four crossbar switches 5004₁-5004₄, respectively, as indicated. Each one of the crossbar switches 5004₁-5004₄interconnects its upper four ports 5006₁-5006₄to a corresponding one of the four memory regions R₁-R₄in a point-to-point fashion. Thus, between the four crossbar switches 5004₁-5004₄and the four memory regions R₁-R₄there are sixteen potential unique interconnects.

[0071] The communication between any port M_B1-M_B8, M_A1-M_A8and its corresponding crossbar switch 5004₁-5004₄is protected by Cyclic Redundancy Check (CRC) defined by CCITT-V.41. The communication between a crossbar switch 5004₁-5004₄and the memory array region R₁-R₄is protected by byte parity (p). There is a pipelined architecture from the port M_B1-M_B8, M_A1-M_A8. Such architecture includes a pipeline having the crossbar switches 5004₁-5004₄, the logic sections 5010₁-5010₈and, the memory array regions R₁-R₄.

[0072] Each one of the memory regions R₁-R₄is here comprised of SDRAM memory chips, as noted above. Each one of these regions R₁-R₄is coupled to the four crossbar switches 5004₁-5004₄through a pair of memory region controller, herein referred to as logic sections, here logic sections 5010₁, 5010₂; . . . 5010₇, 5010₈, respectively. Each logic section 5010₁-5010₈is dual ported, (i.e., Port_A, (A) and Port_B, (B)) with each port being coupled to one of the crossbar switches. The two logic sections 5010₁, 5010₂; . . . 5010₇, 5010₈(i.e., region controllers) associated with one of the memory regions R₁-R₄, respectively, share control of the SDRAM in such memory region. More particularly, and as will be described in more detail below, each pair of logic section, such as for example pair 5010₁and 5010₂, share a common DATA port of memory array region R₁. However, each one of the logic sections 5010₁and 5010₂is coupled to a different control port P_Aand P_B, respectively, of memory array region R₁, as indicated.

[0073] More particularly, each one of the crossbar switches 5004₁-5004₄has, here, four lower ports 5008₁-5008₄and four upper ports 5006₁-5006₄. Each one of the four upper ports 5006₁-5006₄, is, as noted above, coupled to a corresponding one of the four sets S₁-S₄, respectively, of four of the S/P converters. As noted above, the cache memory board 220₁also includes eight logic sections coupled 5010₁-5010₈(to be described in detail in connection with FIGS. 13A, 13B, 13C, 13D and 13E) as well as the four memory array regions R₁-R₄. An exemplary one of the memory array regions R₁-R₄is described in connection with FIG. 6 of U. S. Pat. No. 5,943,287. As described in such U. S. Patent, each one of the memory array regions includes a pair of redundant control ports P_A, P_Band a data/chip select port (here designated as DATA). As described in such U. S. Patent, data may be written into, or read from, one of the memory array regions by control signals fed to either port P_Aor to port P_B. In either case, the data fed to, or read from, the memory array region is on the common DATA port.

[0074] An exemplary one of the logic sections 5010₁-5010₈will be discussed below in detail in connection with FIGS. 13A-15E and an exemplary one of the crossbar switches 5004₁-5004₄in the logic networks 221_1B-221_2Awill be discussed below in detail in connection with FIGS. 10-12D. Suffice it to say here, however, each one of the memory array regions R₁-R₄is coupled to a pair of the logic sections 5010₁, 5010₂; 5010₃, 5010₄; 5010₅, 5010₆; 5010₇, 5010₈, respectively, as shown. More particularly, each one of the logic sections 5010₁, 5010₂; 5010₃, 5010₄5010₅, 5010₆; 5010₇, 5010₈includes: a pair of upper ports, Port_A (A), Port_B (B); a control port, C; and a data port, D, as indicated. The control port C of one each one of the logic sections 5010₁, 5010₃, 5010₅, 5010₇, is coupled to port P_Aof a corresponding one of the four memory array regions R₁-R₄. In like manner, the control port C of one of each one of the logic sections 5010₂, 5010₄, 5010₆, 5010₈is coupled to port P_Bof a corresponding one of the four memory array regions R₁-R₄, respectively as shown. Thus, each one of the memory array regions R₁-R₄is coupled to a redundant pair of the logic sections 5010₁, 5010₂; 5010₃, 5010₄; 5010₅, 5010₆; 5010₇, 5010₈, respectively. The data ports D of logic section pairs 5010₁, 5010₂; 5010₃, 5010₄; 5010₅, 5010₆; 5010₇, 5010₈, respectively, are coupled together and to the DATA port of a corresponding one of the memory regions, R₁-R₄, respectively, as indicated.

[0075] It should be noted that each one of the crossbar switches 5004₁-5004₄is adapted to couple the upper ports 5006₁-5006₄thereof to the lower ports 5008₁-5008₄thereof selectively in accordance with a portion (i.e., a “tag” portion) of the information fed to the crossbar switch. In response to such “tag” portion, a transfer of information between a selected one of the memory array regions R₁-R₄and a selected the of the directors coupled to the crossbar switch is enabled. The memory control portion (e.g., read, write, row address select, column address select, etc.) of the information passes between either port A or port B of a logic sections 5010₁, 5010₃, 5010₅, 5010₇, and port P_Aof the memory array region R₁-R₄coupled to such logic section and the data (DATA) portion of the information passes to the DATA port of such coupled memory array region R₁-R₄, respectively. Likewise, the control portion of the information passes between port A or port B of a logic sections 5010₂, 5010₄, 5010₆, 5010₈, and port P_Bof the memory array region R₁-R₄coupled to such logic section and the data portion of the information passes to the DATA port of such coupled memory array region R₁-R₄, respectively.

[0076] Thus, each one of the logic sections 5010₁-5010₈includes a pair of redundant upper ports, A and B. The lower ports 5008₁-5008₄of crossbar switch 5004₁are coupled to the A port of logic sections 5010₁, 5010₃, 5010₅, and 5010₇, respectively, while the lower ports 5008₁-5008₄of crossbar switch 5004₂are coupled to the B port of logic sections 5010₁, 5010₃, 5010₅, and 5010₇, respectively. The lower ports 5008₁-5008₄of crossbar switch 5004₃are coupled to the A port of logic sections 5010₁, 5010₃, 5010₅, and 5010₇, respectively, while he lower ports 5008₁-5008₄of crossbar switch 5004₄are coupled to the B port of logic sections 5010₂, 5010₄, 5010₆, and 5010₈, respectively.

[0077] As noted above in connection with FIG. 2, each one of the host computer,processors 121₁-121₃₂is coupled to here a pair (but not limited to a pair) of the front-end directors 180₁-180₃₂, to provide redundancy in the event of a failure in one of the front end-directors 181₁-181₃₂coupled thereto. Likewise, the bank of disk drives 140 has a plurality of, here 32, disk drives 141₁-141₃₂, each disk drive 141₁-141₃₂is coupled to here a pair (but not limited to a pair) of the back-end directors 200₁-200₃₂, to provide redundancy in the event of a failure in one of the back-end directors 200₁-200₃₂coupled thereto. Thus, the system has redundant front-end processor pairs 121₁, 121₂through 121₃₁, 121₃₂and redundant back-end processor pairs 141₁, 141₂through 141₃₁, 141₃₂. Considering the exemplary logic network 220₁shown in FIGS. 9A, 9B and 9C, as noted above in connection with FIG. 8B, redundant front-end processor pairs 121₁and 121₂, are able to be coupled to ports M_A1and M_B1of a cache memory board. Thus, the ports M_A1and M_B1may be considered as redundant memory board ports. In like manner, the following may be considered as redundant memory ports because the are able to be coupled to a pair of redundant processors: M_A2and M_B2; M_A3and M_B3; M_A4and M_B4; M_A5and M_B5; M_A6and M_B6; M_A7and M_B7; and, M_A8and M_B8. It is noted that ports M_A1and M_B1; M_A2and M_B2; M_A3and M_B3; M_A4and M_B4are coupled to the front-end processors through front-end directors and ports M_A5and M_B5; M_A6and M_B6; M_A7and M_B7; M_A8and M_B8are coupled to the disk drives through back-end directors.

[0078] Referring again to FIGS. 9A, 9B and 9C, from the above it should be noted then that logic networks 221_IBand 221_1Amay be considered as a pair of redundant logic networks (i.e., pair 1) because they are able to be coupled to redundant pairs of processors, here front-end processors. Likewise, logic networks 221_2Band 221_2Amay be considered as a pair of redundant logic networks (i.e., pair 2) because they are able to be coupled to redundant pairs of disk drives. Further, logic network 221_1Bof pair 1 is coupled to upper port A of logic sections 5010₁, 5010₃, 5010₅, and 5010₇while logic network 221_1Aof pair 1 is coupled to port A of the logic sections 5010₂, 5010₄, 5010₆, and 5010₈. Logic network 221_2Bof pair 2 is coupled to port B of logic sections 5010₁, 5010₃, 5010₅, and 5010₇while logic network 221_2Aof pair 2 is coupled to port B of the logic sections 5010₂, 5010₄, 5010₆, and 5010₈.

[0079] Thus, from the above it is noted that ports M_B1-MB₄, which are coupled to one of a pair of redundant processors, are adapted to be coupled to one of the ports in a pair of redundant control ports, here port P_Aof the four memory array regions R₁-R₄while ports M_A1-M_A4, of the other one of the pair of redundant processors are adapted to be coupled to the other one of the ports of the redundant control ports, here port P_Bof the four memory array regions R₁-R₄. Likewise, ports M_B5-M_B8, which are coupled to one of a pair of redundant processors, are adapted to be coupled to one of the ports in a pair of redundant control ports, here port P_Aof the four memory array regions R₁-R₄while ports M_A5-M_A8, of the other one of the pair of redundant processors are adapted to be coupled to the other one of the ports of the redundant control ports, here port P_Bof the four memory array regions R₁-R₄.

[0080] Thus, the memory board 220₁(FIGS. 9A, 9B and 9C) is arranged with a pair of independent fault domains: One fault domain, Fault Domain A, is associated with logic networks 221_1Band 221_2B, logic sections 5010₁, 5010₃5010₅, 5010₇, and ports P_Aof the memory array regions R₁-R₄and, the other fault domain, Fault Domain B, is associated with logic networks 221_1Aand 221_2A, logic sections 5010₂, 5010₄, 5010₆, 5010₈and port P_Bof the memory array regions R₁-R₄. The logic in each one of the fault domains is operated by a corresponding one of a pair of independent clocks, Clock 1 and Clock 2 (FIGS. 9A, 9B and 9C). More generally, a fault domain is defined as a collection of devices which share one or more common points of failure. Here, Fault Domain A includes: logic networks 221_1B, 221_2B(i.e., the S/Ps and crossbar switches 5004₁-5004₂therein) and logic sections 5010₁, 5010₃, 5010₅, 5010₇, such devices being indicated by lines which slope from lower left to upper right (i.e., ///). The other fault domain, Fault Domain B, includes: logic networks 221_1A, 221_AB(i.e., the S/Ps and crossbar switches 5004₃-5004₄therein) and logic sections 5010₂, 5010₄, 5010₆, 5010₈, such devices being indicated by lines which slope from upper left to lower right (i. e., \\\\). It is noted from FIGS. 9A, 9B and 9C that port P_Aof each one of the memory array regions R₁-R₄is coupled to Fault Domain A while port P_Bis coupled to fault domain B. Thus, each one of the fault domains includes the devices used to couple one of a pair of redundant processors to one of a pair of redundant control ports P_A, P_Bof the memory array regions R₁-R₄and the other fault domain includes the devices used to couple the other one of the pair of redundant processors to the other one of a pair of redundant control ports P_A, P_Bof the memory array regions R₁-R₄. As noted above each fault domain operates with a clock (i.e., clock 1, clock 2) separate from and independent of the clock used to operate the other fault domain.

[0081] Referring now to FIG. 10, an exemplary one of the crossbar switches 5004₁-5004₄, here crossbar switch 5004₁is shown in detail to include four upper port interface sections A-D and lower port interface sections W-Z. The details of an exemplary one of the upper port interface sections A-D, here upper port interface section A, will be described in more detail in connection with FIGS. 11A, 11B, 11C and 11D and the details of an exemplary one of the lower port interface sections W-Z, here lower port interface section W, will be described in more detail in connection with FIGS. 12A, 12B, 12C and 12D. The function of the exemplary crossbar switch 5004₁is to mediate the information cycle at the request of an initiating one of the directors coupled to one of the upper 5006₁-5006₄and one logic section 5010₁-5010₈indicated by the “tag” portion of the information (FIG. 16).

[0082] More particularly, the crossbar switches request, negotiate, and then effect a transfer between the upper thereof 5006₁-5006₄and the lower ports 5008₁-5008₄thereof in a manner to be described below. Suffice it to say here, however, that the upper interface section A-D handle the protocol between the director requesting a information cycle and the memory board 220₁(FIG. 8). It also provides a control and data interface to the serial-to-parallel (S-P) converters (e.g., serializer-deserializer). These interface sections A-D are also responsible for generating parity across the address, control, DATA, and CRC received from the director. There are here two parity bits, one per cycle as described in co-pending patent application entitled “Fault Tolerant Parity Generation” filed May 20, 1999, Ser. No. 99/315,437, and assigned to the same assignee as the present invention, the entire subject matter being incorporated herein by reference. As described in such patent application, the parity is generated such that one byte has odd parity and the other byte has even parity. The sense of these parity bits alternate on successive clocks.

[0083] The lower port interface sections W-Z provides address, control, DATA and routing to one of the four of the logic sections 5010₁-5010₈(FIGS. 9A, 9B and 9C) in a manner to be described. Each one of the lower interface sections W-Z is adapted to couple a corresponding one of the four memory array regions R₁-R₄(FIGS. 9A, 9B and 9C), respectively, via logic sections 5010₁-5010₈. Each one of the four lower interface sections W-Z independently acts as an arbiter between the four upper interface sections A-D and the logic section 5010₁-5010₈coupled thereto. This allows for simultaneous transfers (i.e., information cycles) to multiple memory array regions R₁-R₄from multiple upper interface sections A-D. The upper interface section A-D are single threaded, i.e., one information cycle must be complete before another information cycle is allowed to the same memory array regions R₁-R₄.

[0084] The lower interfaces W-Z deliver control, address and the “tag” field (to be described in more detail below) to the logic section 5010₁-5010₈. The parity across these fields are generated in the upper interface sections A-D and then pass unmodified such that the memory array region can check for alternating parity sense. For write transfers, the lower interface sections W-Z also deliver the write data to the memory array region, checking for correct CRC across the data. If any error is detected, and if the control field indicates a “Wait-and-Validate” process to be described, the parity of the last double byte of data is corrupted (e.g., a fault is induced in the parity (p) thereof) such that the logic section 5010₁-5010₈coupled thereto detects the corrupted parity and inhibits execution of the information cycle. Otherwise, the alternating parity of the data is unmodified. For read transfers, the lower interface sections W-Z accept the data from the memory array regions R₁-R₄via the logic sections 5010₁-5010₈, check the alternating parity, and generates CRC to be returned to the director.

[0085] More particularly, assume for example that information at upper port 5006₄(FIGS. 9A, 9B and 9C) of crossbar switch 5004₄is to be transferred to memory array region R₁. Referring to FIG. 10 a negotiation, i.e., arbitration, must be made by lower port interface W as a result of a request made by the upper port interface section D of crossbar switch 5004₄to section interface W thereof. When interface section W is available to satisfy such request, (i.e., not satisfying request from other one of the upper port interface sections A-C) interface W issues a grant to upper interface section D.

[0086] Thus, each one of the upper port sections A-D sends requests signals (REQs) to the lower port sections W-Z when such upper port sections A-D wants access to (i.e., wants to be coupled to) such lower port sections. Conversely, each one of the upper port sections A-D receives grant signals (GR) from the lower port sections W-Z when such lower port sections W-Z grants access to (i.e., wants to be coupled to) such upper port sections A-D. The request (REQ) and grant (GR) signals, produced by and received from the upper port sections A-D and lower port sections W-Z are as follows:

[0087] where:

[0088] For upper port section A:

[0089] RWA is a request signal sent by upper port section A to lower port section W;

[0090] RXA is a request signal sent by upper port section A to lower port section X;

[0091] RYA is a request signal sent by upper port section A to lower port section Y;

[0092] RZA is a request signal sent by upper port section A to lower port section Z;

[0093] GWA is a grant signal from lower port section W to upper port section A;

[0094] GXA is a grant signal from lower port section X to upper port section A;

[0095] GYA is a grant signal from lower port section Y to upper port section A;

[0096] GZA is a grant signal from lower port section Z to upper port section A;

[0097] For upper port B:

[0098] RWB is a request signal sent by upper port section B to lower port section W;

[0099] RXB is a request signal sent by upper port section B to lower port section X;

[0100] RYB is a request signal sent by upper port section B to upper port section Y;

[0101] RZB is a request signal sent by upper port section B to lower port section Z;

[0102] GWB is a grant signal from lower port section W to upper port section B;

[0103] GXB is a grant signal from lower port section X to upper port section B;

[0104] GYB is a grant signal from lower port section Y to upper port section B;

[0105] GZB is a grant signal from lower port section Z to upper port section B;

[0106] and so forth for the remaining upper and lower port sections C-D and W-Z.

[0107] Each one of the upper port sections A-D has four ports A₁-A₄, through D₁-D₄, respectively, as shown. Each one of the lower port sections W-Z has four ports W₁-W₄, through Z₁-Z₄, respectively, as shown. Ports A₁-A₄are connected to ports W₁-Z₁, respectively, as shown. In like manner, Ports B₁-B₄are connected to ports W₂-Z₂, respectively, as shown, ports C₁-C₄are connected to ports W₃-Z₃, as shown, and Ports D₁-D₄are connected to ports W₄-Z₄, as shown. Lower ports 5008₁-5008₄are connected to lower port sections W-Z, respectively, as shown.

[0108] As noted above, an exemplary one of the upper port interface sections A-D and an exemplary one of the lower port interface sections W-Z will be described in more detail in connection with FIGS. 11A-11D and 12A-12D, respectively. Suffice it to say here, however, that information fed to port 5006₁is coupled to ports 5008₁-5008₄selectively in accordance with a “tag” portion such information. In a reciprocal manner, information fed to port 5008₁is coupled to ports 5006₁-5006₄selectively in accordance with the “tag” portion in such information. Further, ports 5006₂-5006₄operate in like manner to port 5006₁, so that information at such ports 5006₂-5006₄may be coupled to ports 5008₁-5008₄. Still further, ports 5008₂-5008₄operate in like manner to port 5008₁, so that information at such ports 5008₂-5008₄may be coupled to ports 5006₁-5006₄. It should also be noted that information may appear simultaneously at ports 5008₁-5008₄with the information at one of such ports being coupled simultaneously to one of the ports 5006₁-5006₄while information at another one of the ports 5008₁-5008₄is coupled to a different one of the ports 5006₁-5006₄. It is also noted that, in a reciprocal manner, information may appear simultaneously at ports 5006₁-5006₄with the information at one of such ports being coupled simultaneously to one of the ports 5008₁-5008₄and with information at another one of the ports 5006₁-5006₄being coupled to a different one of the ports 5008₁-5008₄.

[0109] Referring now to FIGS. 11A, 11B, 11C and 11D, an exemplary one of the upper port interface sections A-D, here upper port interface section A is shown in more detail. It is first noted that the information at port 5006₁includes: the “tag” portion referred to above; an address CRC ADDR_CRC portion, an address ADDR portion, a memory control portion (i.e., read/write, transfer length, “Wait and Validate”, etc.); a data portion, (DATA); and a DATA Cyclic Redundancy Check (CRC) portion (DATA_CRC).

[0110] The “tag” portion includes: a two bit word indicating the one of the four memory array regions R₁-R₄where the data is to be stored/read; a three bit word indicating the one of the eight memory boards having the desired array region R₁-R₄; a four bit word indicating the one of the 16 director boards 190₁-190₈, 210₁-210₈(FIG. 8) having the director which initiated the transfer; a two bit word indicating which one of the four directors on such one of the director boards is making the requested data transfer; and a five bit random number designating, (i.e., uniquely identifying) the particular information cycle.

[0111] The information described above passing from the director to the crossbar switch (i.e., the “tag”, the ADDR_CRC, the ADDR, the memory control, the DATA, and the DATA_CRC) for the entire information cycle (FIG. 17) are successively stored in a register 5100, in response to clock pulses Clock 1, in the order described above in connection with FIG. 17. The information stored in the register 5100 is passed to a parity generator (PG) 5102 for appending to such information a byte parity (p). After passing through the parity generator (PG) 5102, the different portions of the information are stored in registers 5104₁-5104₆, as follows: Register 5104₁stores the DATA_CRC portion (with the generated parity); register 5104₂, here a FIFO, stores the data portion, DATA, (with the generated parity); register 5104₃stores the memory control portion (with the generated parity); register 5104₄stores the address ADDR portion (with the generated parity), register 5104₅stores the address ADDR_CRC portion (with the generated parity); and register 5104₆stores the “tag” portion (with the generated parity) in the order shown in FIG. 17. Each clock pulse (Clock 1 or Clock 2) results in one of the words described above in connection with FIG. 17. Here, each word has two bytes and is stored in register 5100. The word stored in register 5100 is then shifted out of register 5100 with the next clock pulse, as new information becomes stored in such register 5100.

[0112] The portions stored in the registers 5104₁-5104₄and 5104₆(not register 5104₅which stores ADDR_CRC) are fed to selectors 5106₁-5106₄, and 5106₆, respectively, as indicated. An exemplary one of the selectors 5106₁-5106₄, and 5106₆,, here selector 5106₆is shown to include four registers 5108₁-5108₄. The four registers 5108₁-5108₄are connected to the same input port I of the selector 5106₆to thereby store four copies of the information portion, here the “tag” portion, fed to such input port I in this example. The output of each of the four registers 5108₁-5108₄is fed to a corresponding one of four gated buffers 5110₁-5110₄, respectively, as indicated. With such an arrangement, one of the stored four copies is coupled to a selected one of the output ports A₁-A₄selectively (and hence to ports W₁-Z₁, respectively) in accordance with enable memory control signals on lines EAW-EAZ as a result of decoding the two-bit portion of “tag” indicating the selected one of the four memory array regions R₁-R₄. More particularly, each one of the lines EAW-EAZ is coupled to a corresponding one of the enable inputs of the four gated buffers 5110₁-5110₄, respectively, as indicated.

[0113] More particularly, as noted above, the “tag” includes 2 bits which indicates the one of the four memory array regions R₁-R₄which is to receive the information at port 5006₁(i.e., the “tag”, the ADDR_CRC, the ADDR, the memory control, the DATA, and the DATA_CRC). The “tag” is fed to a memory control logic/ADDR_CRC checker 5112. In response to this two bit portion of the “tag”, the memory control logic/ADDR CRC checker 5112 activates one of the four lines EAW-EAZ to thereby enable a selected one of the four copies stored in the four registers 5108₁-5108₄to pass to one of the ports A₁-A₄. It is noted that the lines EAW-EAZ are also fed to selectors 5106₁-5106₅in a similar manner with the result that the information at port 5006₁(i.e., the “tag”, the ADDR_CRC, the ADDR, the memory control, the DATA, and the DATA_CRC) portions Data CRC, Data, memory control, ADDR, and ADDR_CRC is fed to the same selected one of the ports A₁-A₄and thus to the one of the four memory array regions R₁-R₄described by the two-bit portion of the “tag”.

[0114] It is noted that the upper port section A also includes a memory board checker 5114. Each of the here eight memory board 220₁-220₈(FIG. 8) plugs into the backplane 302 as discussed above in connection with FIG. 3. As noted above, here the backplane 302 is adapted to a plurality of, here up to eight memory boards. Thus, here the backplane 302 has eight memory board slots. Pins P₁-P₃(FIGS. 9A, 9B and 9C) are provided for each backplane 320 memory board slot and produce logic voltage levels indicating the slot position in the backplane. Thus, here the slot position may be indicated with the logic signals on the three pins P₁-P₃to produce a three bit logic signal representative of the backplane slot position. Referring again to FIGS. 9A, 9B and 9C, the exemplary memory board 220₁is shown plugged into a slot in the backplane 302. As noted above, the slot has pins P₁-P₃which provides the slot position three bit logic signal indicative of the slot or “memory board” number in the backplane. The logic signals produced by the pins P₁-P₃are fed to the memory board checker 5114 (FIGS. 11A, 11B, 11C and 11D). Also fed to the memory board checker 5114 are the 3-bits of the “tag” which indicates the one of the memory array boards which is to receive the data (i.e., a 3-bit “memory board code”). If the three bit memory board indication provided by “tag” is the same as the backplane slot or “memory board number” indication provided by the pins P₁-P₃, the director routed the information cycle to the proper one of the eight memory boards and such “accept” indication is provided to the decode logic/ADDR CRC checker 5112 via line A/R. On the other hand, if the three bit memory board indication provided by “tag” is different from the backplane slot indication provided by the pins P₁-P₃, the information cycle was not received by the correct one of the memory boards and such “reject” indication is provided to the decode logic/ADDR CRC checker 5112 via line A/R. When a reject indication is provided to the decode logic/ADDR CRC checker 5112, the intended transfer in prevented and the indication is provided by the decode logic/ADDR CRC checker 5112 to the initiating director via the A/R line. Thus, if the “memory board number” provided by pins P₁-P₃does not match the “memory board code” contained in the “tag” the transfer request from the director is rejected and such error indication is sent back to the director. In this manner, a routing error in the director is detected immediately and is not propagated along.

[0115] On the other hand, if the “memory board number” and the “memory board code” do match, the crossbar switch will forward the requested transfer to one of the four memory regions (i.e., the “memory region number”, R₁-R₄) designated by the “tag”.

[0116] The decode logic and ADDR_CRC checker 5112 also produces load signals L₁-L₆to the registers 5104₁-5104₆, respectively, in response to the “start-frame” signal in word 0 described above in connection with FIG. 16

[0117] Also fed to the decode logic/ADDR_CRC checker 5112 is the ADDR_CRC portion stored in registers 5104₃5104₆(i.e., control, ADDR, ADDR_CRC, and “tag”). The decode logic/ADDR_CRC 5112 performs a check of the CRC of the control, ADDR, ADDR_CRC, and “tag” and if such checker 5112 detects an error such error is reported back to the transfer initiating director via line ADDR_CRC_CHECK, as indicated. Detection of such an ADDR_CRC_CHECK error also results in termination of the transfer.

[0118] When data is read from a selected one of the memory array region R₁-R₄as indicated by the “tag” stored in register 5104₆, the decode logic/ADDR_CRC checker 5112 activates the proper one of the lines EAW-WAZ to coupled the proper one of the ports A₁-A₄coupled to such selected one of the memory array regions R₁-R₄to a register 5120. Thus, read data passes via selector 5118 to the register 5120 and is then sent to the transfer-requesting director via pot 5006₁.

[0119] It is noted that the decode logic and ADDR CRC checker 5112 in upper port interface logic A also produces request signals RWA, RXA, RYA, and RZA and sends such request signal to lower port sections W-Z, respectively. Such requests are fed to an arbitration logic 5114 (FIGS. 12A, 12B, 12C and 12D) included within each of the lower port sections W, X, Y and Z, respectively. Thus, because the other upper port sections B-D operate in like manner to upper port section A, the arbitration 5114 in lower port interface section W may receive requests RWB, RWC, and RWD from such other upper port sections B-D, respectively. In accordance with a predetermined arbitration rule, such as, for example, first-come, first-served, the arbitration logic 5114 of lower port interface section W grants for access to lower port 5008₁of lower port section W to one of the requesting upper port sections A-D via a grant signal on one of the lines GWA, GWB, GWC and GWD, respectively.

[0120] Thus, referring again to FIGS. 11A, 11B, 11C and 11D, the decode logic/CRC ADR checker 5112 issues a request on line RWA when port 5008₁(FIG. 10) desires, based on the two bit information in the “tag”, memory array region R₁(FIGS. 9A, 9B and 9C). In like manner, if memory array regions R₂-R₄are indicted by the “tag”, requests are made by the upper port section on lines RXA, RYA, RZA, respectively. The other upper port sections B-D operate in like manner. The grants (GR) produced by the lower port sections W, X, Y and Z are fed to the upper port sections A-D as indicated above. Thus, considering exemplary upper port section A (FIGS. 11A, 11B, 11C and 11D), the grant signals from lower port sections W-Z are fed to the decode logic/CRC checker 5112 therein on lines GWA, GXA, GYA and GZA, respectively. When a grant on one of these four lines GWA, GXA, GYA and GZA is received by the decode logic/CRC checker 5112, such checker 5112 enables the gating signal to be produced on the one of the enable lines EAW, EAX, EAY, EAZ indicated by the “tag” portion. For example, if the “tag” indicates that memory array region R₃(which is adapted for coupling to port 5008₃of lower port section Y) the checker 5112 issues a request on line RYA. When after the arbitration logic 5114 in section Y determines that lower port logic A is to be granted access to port 5008₃, such lower port section Y issues a grant signal on line GYA. In response to such grant, the checker 5112 issues an enable signal on line EAY to thereby enable information to pass to port A₃(FIGS. 11A, 11B, 11C and 11D).

[0121] In a reciprocal manner, when data is to be transferred from a memory array region to the requesting director, the information sent by the requesting director is processed as described above. Now, however, the checker 5112 sends a control signal to one of the lines EAW-EAZ to selector section 5118 to enable data on one of the ports A₁-A₄coupled to the addressed memory array regions R₁-R₄to pass to register 5120 and then to upper port 5006₁.

[0122] Referring now to FIGS. 12A, 12B, 12C and 12D, exemplary lower port section W is shown to include arbitration logic 5114 described above, and the selector 5120 fed by signals on ports W₁-W₄. (Referring again to FIG. 10, ports W₁-W₄are coupled to ports A₁, B₁, C₁and D₁, respectively, of upper port interface sections A-D, respectively.) Thus, when the arbitration logic 5114 grants access to one of the upper port sections A-D, the decoder 5122 decodes the grant information produced by the arbitration logic and produces a two bit control signal for the selector 5120. In response to the two bit control signal produced by the decoder 5122, the selector couples one of the ports W₁-W₄(and hence one of the upper port sections A-D, respectively), to the output of the selector 5120 and hence to lower port 5008₁in a manner to be described.

[0123] As noted above, the communication between any port M_B1-M_B8, M_A1-M_A8and its corresponding crossbar switches 5004₁-5004₄is protected by Cyclic Redundancy Check (CRC) defined by CCITT-V.41. The communication between a crossbar switch 5004₁-5004₄and its corresponding memory array region R₁-R₄is protected by byte parity (p). There is a pipelined architecture from the port M_B1-M_B8, M_A1-M_A8, and through the crossbar switch, and through the logic sections 5010₁-5010₈.

[0124] The nature of CRC calculation is such that an error in the data is not detected until the entire transfer is completed and the checksum of the CRC is known. In the case of a write of data into the memory, by the time the CRC is checked, most of the data is already through the pipeline and written into memory.

[0125] Here, the memory control field has a specific bit “Wait and Validate” in the control word 1 in FIG. 16 which is at the director's control. If the bit is set, the logic sections 5010₁-5010₈buffers the entire information cycle, pending the CRC calculation, performed at the lower port interface sections W-Z. If the CRC check indicates no CRC error, then the data is written into the memory array region. If the CRC check does indicate an error, then the memory array region is informed of the error, here by the lower interface section W-Z corrupting the data into a fault. Such fault is detected in the logic section 5010₁-5010₈and such information is prevented from being stored in the memory region R₁-R₄, in a manner to be described. Suffice it to say here, however, that this “Wait and Validate” technique enables the director to flag certain data transfers as critical, and if an error occurs, prevents corruption of the data stored in the memory array. That is, the data having a CRC error is detected and prevented from being stored in the memory array region. For those transfers not indicated as critical by the director, the “Wait and Validate” bit is not set thereby maximum performance of the memory is obtained.

[0126] More particularly, the DATA, memory control, ADDR, and “tag” portions (with their byte parity (p) generated by parity generator 5102 (FIGS. 11A, 11B, 11C and 11D)) of the information coupled to the output of selector 5120 is stored in the register 5124. As noted above in connection with FIG. 16, the DATA_CRC portion (i.e., the words X and Y) occurs after the last DATA word.

[0127] Thus, as the words in the DATA clock through register 5124 they pass into the DATA_CRC checker 5132 where the CRC of the DATA is determined (i.e., the DATA_CRC checker 5132 determine X and Y words of the DATA fed to such checker 5132). The actual X and Y words (i.e., DATA_CRC stored in register 5128, both content (n) and parity (p)) are stored successively in register 5128 and are then passed to checker 5132 where they are checked against the X and Y words determined by the checker 5132. As noted above, the DATA has appended to it its parity (p). Thus, the “information” whether in register 5124 or register 5128 has a content portion indicated by “n” and its parity indicated by “p”. Thus, the DATA_CRC register 5128 includes the DATA_CRC previously stored in register 5104₁(FIGS. 11A, 11B, 11C and 11D) (i.e., the content portion designated by “n”) and its parity (designated by “p”). The DATA, memory control, ADDR, and “tag” portions, (with their parity (p) (i.e., content “n” plus its appended parity “p”) stored in register 5124 may be coupled through a selector 5149 through one of two paths: One path is a direct path when the “Wait and Validate” command is not issued by the director; and, a second path which includes a delay network 5130, here a three clock pulse delay network 5130.

[0128] More particularly, it is noted that the DATA, control, ADDR, “tag”, both content (n) and parity (p) are also fed to a DATA_CRC checker 5132. Also fed to the DATA_CRC checker 5132 is the output of DATA_CRC register 5128. The CRC checker 5132 checks whether the DATA_CRC (content “n” plus its parity “p”) is the same as the CRC of the DATA, such DATA having been previously stored in register 5104₂(FIGS. 11A, 11B, 11C and 11D), i.e., the content “n” plus its parity “p” of the DATA previously stored in register 51042 (FIGS. 11A, 11B, 11C and 11D). If they are the same, (i.e., no DATA_CRC_ERROR), a logic 0 is produced by the CRC checker 5132. If, on the other hand, they are not the same, (i.e., a DATA_CRC_ERROR), the CRC checker 5132 produces a logic 1. The output of the Data_CRC checker 5132 thereby indicates whether there is an error in the CRC of the DATA. Note that a DATA_CRC_ERROR is not known until three clock cycles after the last sixteen-bit portion of the DATA (i.e., the word of the DATA, FIG. 16) is calculated due to the nature of the CRC algorithm. Such indication is fed to a selector 5152 via an OR gate 5141. If there is a DATA_CRC_ERROR, the “information” at the output of the delay network 5130 (i.e., the last word of the DATA (FIG. 16)) with its parity (p)) is corrupted. Here, the content (n) of such “information” (i.e., the “information” at the output of the delay network 5130 (i.e., the last word of the DATA (FIG. 16))) is fed to a second input 12 of the selector 5140. The parity (p) of such “information” (i.e., the last word of the DATA (FIG. 16)) is fed non-inverted to one input of selector 5152 and inverted, via inverter 5150, to a second input of the selector 5152. If there is a DATA_CRC_ERROR detected by data CRC checker 5132, the inverted parity is passed through the selector 5152 and appended to the content portion (n) of the “information” (i.e., the last word of the DATA (FIG. 16)) provided at the output of the delay network 5130 and both “n” and appended “p” are fed to the second input I₂of selector 5140 thereby corrupting such “information”. It should be noted that the remaining portions of the information cycle (i.e., the memory control, address (ADDR), “tag”, and all but the last word of the DATA (FIG. 16)) pass through the delay network 5130 without having their parity (p) corrupted.

[0129] If there is a no “Wait and Validate” transfer, logic decoder 5122 selects the first input I₁as the output of the selector 5140. If there is a “Wait and Validate” transfer, the logic decoder 5122 selects the second input I₂as the output of the selector 5140. It is noted, however, that that because the last word of DATA (FIG. 16) is delayed three clock pulses (from Clock 1) by registers 5142, 5144, and 5146 (such registers 5142, 5144 and 5146 being fed by such Clock 1), the DATA_CRC check is performed before the last word of the DATA appears at the output of register 5146. Thus, the last word of the DATA is corrupted in byte parity before being passed to the logic section 5010₁-5010₈. That is, because of the delay network 5130, the DATA_CRC is evaluated before the last word of the DATA has passed to port 5008₁. This corruption in parity (p), as a result of a detected DATA_CRC error, is detected by a parity checker 6106 (FIGS. 14A, 14B, 14C and 14D) in the following logic section 5010₁-5010₈in a manner to be described. Suffice it to say here, however, that detection of the parity error (produced by the detected CRC error) prevents such corrupted information from storage in the SDRAMs.

[0130] On the other hand, if there is no DATA_CRC_ERROR (and no error in the parity of the DATA_CRC detected by the parity checker 6106 (FIGS. 14A, 14B, 14C and 14D) in a manner to be described) the non-inverted parity (p) is appended to the “information” (i.e., DATA, memory control, ADDR, and “tag”) provided at the output of the delay network 5130 and such information is fed to the proper memory address region R₁-R₄as indicated by “tag”.

[0131] More particularly, it is noted that the selector 5140 is also fed the “information” (i.e., DATA, memory control, ADDR, and “tag”) without such “information” passing through the delay 5130. The director issuing the transfer may not require that the transfer have the DATA_CRC check result preclude the writing of information into the memory (i.e., no “Wait and Validate”), in which case the “information” is passed directly through the selector 5140. On the other hand, if such DATA_CRC check is to be effected, the delay network 5130 output, with a possible corruption as described above, is passed through the selector 5140. The director provides the indication as part of the control field in the described “Wait and Validate” bit. Such bit is decoded by the logic decoder 5122. In response to such director indication, a “Wait and Validate” control signal is sent by the logic decoder 5122 to the selector 5140.

[0132] As noted above, the communication between any port and its corresponding crossbar switch is protected by Cyclic Redundancy Check (CRC) defined by CCITT-V.41. The communication between a crossbar switch and a memory array region R₁-R₄is protected by byte parity (p). This implies that the crossbar switch must translate between CRC protection and parity protection.

[0133] As a further check of the validity of the DATA CRC, the generated parity p of the CRC of such DATA is checked. However, because the CRC is generated by the director, and the CRC parity is also generated by upper interface section A-D, a CRC generation fault would yield an undetectable CRC parity fault.

[0134] It has been discovered that the parity (p) of the DATA_CRC must be the same as the parity of the DATA parity (p). Thus, one merely has to check whether the parity of the DATA_CRC is the same as the parity of the DATA parity (p). Therefore, such detection DATA_CRC parity checking method is accomplished without using the DATA_CRC itself.

[0135] More particularly, since the DATA over which the DATA_CRC is being calculated is already parity protected, one can use the DATA parity (p) to calculate the DATA_CRC parity: i.e., the DATA_CRC parity is equal to the parity of all the DATA parity bits. Still more particularly, if there are N bytes of DATA:

[0136] [D(0), D(1), . . . D(N−1)]

[0137] and each byte is protected by a parity bit p, then the DATA_CRC parity is the parity of

[0138] [p(0), p(1), . . . p(N−1)].

[0139] Thus, if there is a fault in the generation of the DATA_CRC, it is immediately detected and isolated from the director.

[0140] Thus, the exemplary lower port interface section W (FIGS. 12A, 12B, 12C and 12D) includes a parity generator made up of an exclusive OR gate 5134 and register 5136 arranged as shown fed by the parity (p) of the DATA portion stored in register 5124. The generated parity p is fed to a comparator 5138 along with the parity (p) of the DATA_CRC (i.e., DATA_CRC_PARITY), as indicated. If the two are the same at the end of the DATA portion of the information cycle (FIG. 16), a logic 0 is produced by the comparator 5138 and such logic 0 passes to the selector 5152 to enable the non-inverted parity to pass through such selector 5152. If there is an error in the parity bit of the CRC, a logic 1 is produced by the comparator 5138 and the inverted parity is passed through the selector 5152. The logic 1 output of comparator 5138 passes through OR gate 5141 to couple the inverted parity (p) through selector 5152 to append to the content port (n) of DATA control, ADDR, and “tag” at port I₂of selector 5140. Thus, if there is either a DATA_CRC_ERROR or if DATA_CRC_PARITY is different from parity of the DATA_PARITY at the end of the DATA portion of the information cycle as indicated by a signal produced on line COMP_ENABLE by the logic decoder 5122, a logic 1 is produced at the output of OR gate 5141 thereby coupling the inverted parity through selector 5152. Otherwise, the non-inverted parity passes through selector 5152. That is, the COMP_EN is produced at the end of the DATA in the information cycle (FIG. 16).

[0141] It is noted that information read from the memory region passes to a register 5170 and a CRC generator 5172. The generated CRC is appended to the information clocked out of the register 5170. Four copies of the information with appended CRC are stored in registers 5174₁-5174₄, respectively. In response to the “tag” portion fed to logic decoder 5122, a selected one of the registers 5174₁-5174₄is coupled to one of the port W₁-W₄by selector 5180 and gates 5182₁-5182₄in a manner similar to that described in connection with FIGS. 11A, 11B, 11C and 11D.

[0142] Referring now to FIGS. 13A, 13B, 13C, 13D and 13E, a pair of the logic sections 5010₁-5010₈(memory array region controllers), here logic sections 5010₁and 5010₂are shown. As noted above in connection with FIGS. 9A, 9B and 9C, both logic sections 5010₁and 5010₂are coupled to the same memory array region, here memory array region R₁. As was also noted above in connection with FIGS. 9A, 9B and 9C, the logic section 5010₁is in one fault domain, here fault domain A, and logic section 5010₂is in a different fault domain, here fault domain B. Thus, logic section 5010₁operates in response to clock pulses from Clock 1 and logic section 5010₂operates in response to clock pulses from Clock 2.

[0143] As noted above, each logic section 5010₁-5010₈(FIGS. 9A, 9B and 9C) includes a pair of upper ports, A and B, a control port C and a data port D. Referring to FIGS. 13A, 13B, 13C, 13D and 13E, an exemplary logic section 5010₁is shown in detail to include a upper port A controller 6002A coupled to upper port A, a upper port B controller 6002B coupled to upper port B, and a memory refresh section 6002R.

[0144] Both port A and port B controllers 5010₁, 5010₂have access to the data stored in the same memory array region R₁. Further, while each can provide different, independent control and address information, (i.e., memory control, ADDR, and “tag” (hereinafter sometimes referred to as ADDR/CONTROL)), both share the same DATA port. As noted above, the details of the memory array region 1 are described in detail in connection with FIG. 6 of U.S. Pat. No. 5,943,287. Thus, arbitration is required for access to the common memory array region R₁when both the port A and port B controllers 5010₁and 5010₂desire access to the memory array region R₁. Further, the SDRAMs in the memory array region R₁require periodic refresh signals from the memory refresh section 6002R. Thus, access or request for, the memory array region R₁may come from: the upper port A controller 6002A (i.e., REQUEST A); the upper port B controller 6002B (i.e., REQUEST B); and from the memory refresh section 6002R (i.e., REFRESH REQUEST). These request are fed to an arbitration logic 6004 included within the logic section 5010₁-5010₈. The arbitration sections 6004₁, 6004₂in the redundant paired logic sections, here logic sections 5010₁, 5010₂, respectively, arbitrate in accordance with an arbitration algorithm to be described and thereby to issue a grant for access to the memory array region R₁to either: the upper port A controller 6002A (i.e., GRANT A); the upper port B controller 6002B (i.e., GRANT B); or the memory refresh section 6002R (i.e., REFRESH GRANT).

[0145] Here, the arbitration algorithm is an asymmetric round robin sharing of the common memory array region R₁. The arbitration logic 6004₁, 6004₂and the algorithm executed therein will be described in more detail in connection with FIGS. 15A, 15B, 15C, 15D and 15E. Suffice it to say here however that the arbitration grants access to the common memory array region based on the following conditions:

[0146] Condition I—If both the logic sections 5010₁and 5010₂are operating properly (i.e., produce Memory Output Enable (MOE) and Memory Refresh Enable (MRE) signals, to be described, properly), the port A controller 6002A memory refresh controller 6002R is used exclusively for memory refresh during the round-robin arbitration). Thus, there is asymmetric round robin arbitration because the memory refresh section 6002R of logic section 5010₂is not used when operating in this normal Condition I. The states of the arbitration sequences are as follows:

[0147] State 1—The upper port A controller 6002A of logic section 5010₁is granted access to the memory array region R₁;

[0148] State 2—The memory refresh section 6002R of logic section 5010₁is granted access to the memory array region R₁;

[0149] State 3—The upper port B controller 6002B of logic section 5010₁is granted access to the memory array region R₁;

[0150] State 4—The memory refresh section 6002R of logic section 5010₁is granted access to the memory array region R₁;

[0151] State 4—A check is made as to whether the of logic section 5010₂requests access to the memory array region R₁. If such a request exist:

[0152] (a) The upper port A controller 6002A of logic section 5010₂is granted access to the memory array region R₁if such access is requested;

[0153] (b) The upper port B controller 6002B of logic section 5010₂is granted access to the memory array region R₁if such access is requested;

[0154] State 5—The process returns to State 1.

[0155] (It should be noted that the process uses the memory refresh section 6002R of logic section 5010₁but does not use the memory refresh section 6002R of logic section 5010₂. Thus the round robin is asymmetric.)

[0156] Condition II—If the logic section 5010₂is disabled (i.e., does not produce MOE and MRE signals properly), the logic section 5010₂is not part of the round-robin arbitration and memory refresh is provided, as in Condition I, exclusively by the logic section 5010₁memory refresh controller 6002R. The logic section 5010₁no longer receives request signals FROM the logic section 5010₂. Also the logic section 5010₁is granted access to the memory array region R₁all the time. Thus, the states of the arbitration sequence are in Condition II as follows:

[0157] State 1—The upper port A controller 6002A of logic section 5010₁is granted access to the memory array region R₁;

[0158] State 2—The memory refresh section 6002R of logic section 5010₁is granted access to the memory array region R₁;

[0159] State 3—The upper port B controller 6002B of logic section 5010₁is granted access to the memory array region R₁;

[0160] State 4—The memory refresh section 6002R of logic section 5010₁, is granted access to the memory array region R₁;

[0161] State 5—The process returns to State 1.

[0162] Condition III—The logic section 5010₁is disabled (i.e., does not produce MOE and MRE signals properly) and thus the logic section 5010₁is not part of the round-robin arbitration. Memory refresh is provided exclusively by the memory refresh section 6002R (not shown) in the logic section 5010₂. The logic section 5010₂is granted access to the memory array region R₁all the time. Thus the states of the arbitration sequence in Condition III are as follows:

[0163] State 1—The upper port A controller 6002A of logic section 5010₂is granted access to the memory array region R₁;

[0164] State 2—The memory refresh section 6002R of logic section 5010₂is granted access to the memory array region R₁;

[0165] State 3—The upper port B controller 6002B of logic section 5010₂is granted access to the memory array region R₁;

[0166] State 4—The memory refresh section 6002R of logic section 5010₂is granted access to the memory array region R₁;

[0167] State 5—The process returns to State 1.

[0168] Condition IV—Reset (the arbitration is reset into Condition I from either Condition II or from condition III).

[0169] Referring again to FIGS. 13A, 13B, 13C, 13D and 13E, the arbitration logic 6004₁, 6004₂in each one of the logic sections 5010₁, 5010₂produces: a memory output enable (MOE) signal; a memory refresh enable (MRE) signal (to be described in more detail in connection with FIGS. 15A-15E and 19); and, a memory grant (MG) signal, (to be described in more detail in connection with FIGS. 15A-15E and 19). Thus, logic section 5010₁produces a memory output enable signal MOEA (to be described in more detail in connection with FIGS. 15A-15E and 19), a memory refresh enable signal MREA (to be described in more detail in connection with FIGS. 15A-15E and 19) and a memory grant signal MGA (to be described in more detail in connection with FIGS. 15A-15E and 19). Likewise, logic section 5010₂produces a memory output enable signal MOEB (to be described in more detail in connection with FIGS. 15A-15E and 19), a memory refresh enable signal MREB (to be described in more detail in connection with FIGS. 15A-15E and 19) and a memory grant signal MGB (to be described in more detail in connection with FIGS. 15A-15E and 19). Suffice it to say here, however, that the MOEA signal is a triplicate signal MOE_I-1, MOE_I-2, MOE_I-3and the MGA signal is also a triplicate signal MGE_IA, MGE_IIA, and MGE_IIIA.

[0170] The MOEA and MREA signals from the logic section 5010₁and the MOEB and MREB signals from the logic section 5010₂are fed to a watch dog (WD) section 6006, to be described in more detail in connection with FIGS. 15A, 15B, 15C, 15D and 15E. Suffice it to say here, however, that, as noted above, the arbitration algorithm is a function of the operating/non-operating condition of the logic sections 5010₁, 5010₂. This operating/non-operating condition is determined by the watchdog section 6006 and more particularly by examining the MOEA, MREA, MOEB, MREB signals produced by the logic sections 5010₁and 5010₂, 6002B, respectively. The MOEA, MREA, MOEB, MREB signals are asserted when there is a grant. Such signals MOEA, MREA, MOEB, MREB are fed to the watchdog section 6006. As will be described, the watchdog section 6006 examines the time history of these signals to determine if the logic section 5010₁or 5010₂asserting them is operating properly. Based on the results of such examination, the watchdog selects the Condition I, Condition II, or Condition III, described above.

[0171] More particularly, consider, for example, a case where the MOEA signal is asserted for too long a predetermined time interval. It should be recalled that the logic section 5010₁producing such MOEA signal is granted access to the memory in State 1 of the normal arbitration condition (i.e., Condition I, above). The watchdog section 6006 thus detects a fault in logic section 5010₁. When such a fault is detected, the watchdog section 6006 issues a Condition III signal on in triplicate on lines MSAB to the arbitration sections 6004₁, 6004₂in both the logic sections 5010₁, 5010₂, respectively, indicating that the arbitration algorithm will operate in accordance with the States set forth above for Condition III. Further, the watchdog 6006 issues a data output enable signal in triplicate on lines DOEA (i.e., DOEA₀, DOEA₁, and DOEA₂). This triplicate signal DOEA (i.e., DOEA₀, DOEA₁, and DOEA₂) is fed to a majority gate (MG) 6007 (FIGS. 13A, 13B, 13C, 13D and 13E), in accordance with the majority of the triplicate data fed to it, provides an enable/disable signal for gate 6009. If the majority indicates a fault, the gate 6009 inhibits DATA from passing between the logic section 5010₁and the data port D thereof.

[0172] Consider the case where the arbitration is in Condition I. Consider also that in such condition I, the MREA signal is not produced after a predetermined time interval which ensures proper refreshing on the SDRAMs in the memory array region R₁. The watchdog section 6006 will again detect a fault in the logic section 5010₁port A controller 6002A. When such a fault is detected, the watchdog section 6006 issues a Condition III signal on in triplicate on lines MSAB (i.e., MSAB₀, MSAB₁, MSAB₂) to the arbitration sections 6004₁, 6004₂in both the logic sections 5010₁, 5010₂, respectively. Further, the watchdog 6006 issues a data output enable signal in triplicate on lines DOEA (i.e., DOEA₀, DOEA₁, and DOEA₂) (FIGS. 13A, 13B, 13C, 13D and 13E) to inhibit DATA from passing between the logic section 5010₁and the data port D thereof.

[0173] Consider, for example, a case where the arbitration is in Condition I and the MOEB signal from the logic section 5010₂is asserted for too long a predetermined time interval. The watchdog section 6006 thus detects a fault in the logic section 5010₂. When such a fault is detected, the watchdog section 6006 issues a Condition II signal on line MSAB to the arbitration sections 6004₁, 6004₂in both the logic sections 5010₁, 5010₂. Further, the watchdog 6006 issues a data output enable signal in triplicate on lines DOEB to inhibit DATA from passing between the logic section 5010₂and the data port D thereof.

[0174] It should be noted that the algorithm allows a transition between Condition II and Condition IV (i.e., reset) or from Condition III and Condition IV.

[0175] Thus, the arbitration logics 6004₁and 6004₂are adapted to issue the following signals:

[0176] GRANT A (GA)—grant port A controller 6002B access to the memory array region R₁;

[0177] GRANT B (GB)—grant port B controller 6002B access to the memory array region R₁

[0178] REFRESH GRANT (GR)—grant the memory refresh section 6002R of logic section 5010₁access to the memory array region R₁in Condition I and II or grant the memory refresh section 6002R of logic section 5010₂access to the memory array region R₁in Condition III.

[0179] It should be noted that the details of GA and the other signal GB and GR are shown in more detail in connection with FIG. 19.

[0180] Thus, referring to FIGS. 13A, 13B, 13C, 13D and 13E, the memory array region R₁may be coupled to either Port_A (A) or Port_B (B) of the logic sections 5010₁, 5010₂or to the memory refresh section 6002R therein selectively in accordance with a Port_A_SELECT, Port_B_SELECT, Port_R_SELECT signal fed to a pair of selectors 6010_c, 6010_D, shown in more detail for exemplary logic section 5010₁. Access by the upper port A controller 6002A (i.e., Port_A), by the upper port B controller 6002B, or the memory refresh section 6002R to the memory array region R₁is in accordance with the algorithm described above

[0181] An exemplary one of the upper port A and port B logic controllers 6002A and 6002B, here controller 6002A, will be described in more detail in connection with FIGS. 14A, 14B, 14C and 14D. Suffice it to say here, however, that it is noted that the output of selector ⁶⁰¹⁰_Cis coupled to the control port C of the exemplary logic section 5101₁and the output of selector 6010_Dis coupled to the data port D of the exemplary logic section 5101₁through the gate 6009. Each one of the selectors 6010_Cand 6010_Dhas three inputs A, B. and R, as shown. The A, B and R inputs of selector 6010_Care coupled to: the ADR/CONTROL produced at the output of upper port A controller 6002A; the ADR/CONTROL produced at the output of upper port B controller 6002B; and, the portion REFRESH_C of the refresh signal produced by the memory refresh section 6002R, respectively as indicated. The A, B and R inputs of selector 6010D are coupled to: the WRITE DATA produced at the output of upper port A controller 6002A; the WRITE DATA produced at the output of upper port B controller 6002B; and, the portion REFRESH_D of the refresh signal produced by the memory refresh section 6002R, respectively as indicated. The Port_A_SELECT, Port_B_SELECT are produced by the upper port A controller 6002A, upper port B controller 6002B in a manner to be described. The Port_R_SELECT signal is produced by the memory refresh section 6002R in a manner to be described to enable proper operation of the above described arbitration algorithm and to proper a refresh signal to the SDRAMs in the memory array region R_at the proper time. Suffice it to say here, however, that when port A controller 6002A produces the Port_A_SELECT signal, the ADR/CONTROL at the output of port A controller 6002A passes to the output of the selector 6010C and the DATA_WRITE at the output of the port A controller 6002A passes to the output of the selector 6010D. Likewise, when port B controller 6002B produces the Port_B_SELECT signal, the ADR/CONTROL at the output of port B controller 6002B passes to the output of the selector 6010C and the DATA_WRITE at the output of the port B controller 6002B passes to the output of the selector 6010D. In like manner, when refresh memory section 6002R produces the Port_R_SELECT_C signal, the REFRESH_C at the output of refresh memory section 8002R passes to the output of the selector 6010C and in response to the Port_R_SELECT signal, the REFRESH_D at the output of the refresh memory section 8002R passes to the output of the selector 6010D.

[0182] It is noted that data read from the memory array R₁(i.e., READ_DATA) is fed from the data port D to both the upper Port A controller 6002A and the upper Port B controller 6002B.

[0183] Referring now to FIGS. 14A, 14B, 14C and 14D, the exemplary port A controller 6002A is shown in more detail to include a Port A primary control section 6100P and a Port A secondary control section 6100S. The two sections 6100P and 6100S are both coupled to port A and both implement the identical control logic. Thus, each one of the two sections 6100P and 6100S should produce the same results unless there is an error, here a hardware fault, in one of the two sections 6100P and 6100S. Such a fault is detected by a fault detector 6102 in a manner to be described.

[0184] Thus, referring to the details of one of the two sections 6100P and 6100S, here section 6100P, it is first noted that the information at Port_A is fed to a parity checker 6101. It is noted that is there is an error in parity induced by the CRC check described in FIGS. 12A, 12B, 12C and 12D in connection with selector 5152, such detected parity error is reported to a control and DATA path logic 6112. In response to a detected parity error, control and DATA path logic 6112 prevents memory control signals (e.g., suppress the Column Address Select signal to the SDRAMs) from being produced on the CONTROL_P line. Thus, absent control signal, DATA will not be stored in the memory region.

[0185] The information at Port_A is also fed to a control register 6104 for storing the memory control portion of the information at port A, an ADDR register 6106 for storing the address portion (ADDR) of the information at port A, a write data register 6108 (here a FIFO) for storing the DATA portion of the information at port A, such being the data which is to be written into the memory array region R The control portion stored in register 6104 is fed also to the control and data path logic 6112. Such logic 6112 produces: a memory array region request Request_Port_A_Primary (RAP) signal when the control portion in register 6104 indicates that there is data to be stored in the memory array region R₁; a Port A Primary Select (Port_A_P_SELECT) signal when the grant has been issued thereto via a Grant_A_P signal (GAP) produced by the arbitration logic 6004₁; and passes the control portion (CONTROL_P) stored in register 6104 to the output of the upper port A controller 6002A, as indicated. It should be noted that the port A secondary control section 6100S being fed the same information as the primary controller 6100P should produce the same signals: here indicated as a memory array region request Request_Port_A_SECONDARY (RAS) signal when the control portion in register 6104 indicates that there is data to be stored in the memory array region R₁; a Port A Secondary Select (Port_A_S_SELECT) signal when the grant has been issued thereto via a Grant_A_S signal (GAS) produced by the arbitration logic 6004₁.

[0186] The address portion stored in the ADDR register 6106 (ADDR_P) is combined with the address portion ADDR_P stored in register 6106. Both CONTOL_P and ADDR_P are fed to a parity generator 6109 to produce ADDR/CONTROL_P (which has both a content portion (n) and parity (p). The content portion (n) of ADDR/CONTROL_P is fed to a parity generator 6120 to generate byte parity (p′) from the content portion (n) of ADDR/CONTROL_P. The generated parity (p′) is inverted by inverter 6122 and the inverted parity is fed to a first input I₁of the selector 6124. The content portion (n) of ADDR/CONRTOL_P is combined with a parity (p) produced at the output of selector 6124 in a manner to be described. The parity (p) of ADDR/CONTROL_P is fed to a second input I₂of the selector 6124 and such parity (p) is also fed to an exclusive OR gate 6130. Also fed to the exclusive OR gate 6130 is the parity (p) of the equivalent ADDR/CONTROL_S signal produced by the Port A secondary control section 6100S. As noted above, since both sections 600P and 600S are fed the same information and implement the same logic functions, ADDR/CONTROL_P should be the same as ADDR/CONTROL_S unless there is a hardware fault in one of the sections 6100P, 6100S. If there is a fault (i.e., if ADDR/CONTROL_S and ADDR/CONTROL_P are different), the exclusive OR gate 6130 will produce a logic 1 and in the absence of a fault, (i.e., ADDR/CONTROL_S is the same as ADDR/CONTROL_P), the exclusive OR gate 6130 will produce a logic 0.

[0187] In like manner, the content (n) of ADDR/CONTROL_P is fed to an exclusive OR gate 6128. Also fed to the exclusive OR gate 6128 is the content (n) of the equivalent ADDR/CONTROL_S signal produced by the Port A secondary control section 6100S. As noted above, since both sections 600P and 600S are fed the same information and implement the same logic functions, ADDR/CONTROL_P should be the same as ADDR/CONTROL_S unless there is a hardware fault in one of the sections 6100P, 6100S. If there is a fault (i.e., if ADDR/CONTROL_S and ADDR/CONTROL_P are different), the exclusive OR gate 6128 will produce a logic 1 and in the absence of a fault, (i.e., ADDR/CONTROL_S is the same as ADDR/CONTROL_P), the exclusive OR gate 6128 will produce a logic 0.

[0188] The outputs of exclusive OR gates 6128 and 6130 are fed to an OR gate 6126. Thus, if there is an error in either the content (n) or the parity (p), the OR gate produces a logic 1; otherwise it produces a logic 0. The output of OR gate 6126 is fed to a fault detector 6102 which detects such a fault and reports such detected fault to the director. The output of OR gate 6126 is also fed as a control signal to selector 6124. If the OR gate produces a logic 1 (i.e., there is a fault), the selector couples the inverted parity of input I₁to the output of selector 6124. This inverted parity is appended to the content (n) of ADDR/CONTROL_P to thereby corrupt such information. This corrupted information is detected by the memory array region and converted into a “no-operation” command as described in the above-referenced U.S. Pat. No. 5,943,287. On the other hand, if the OR gate 6126 produces a logic 0 (i.e., no fault), the non-inverted parity at input I₂of selector 6124 passes through selector 6124 and is appended to the content portion (n) of ADDR/CONTROL/P.

[0189] A similar check is made with the DATA to be written into the memory array region. Thus, the DATA in register 6108 of primary controller 6100P (WRITE_DATA_P) is fed to an exclusive OR gate 6116 along with the write DATA in the secondary controller 6100S (WRITE_DATA_S). (It is noted the data in the write register 6108 of the primary controller 6100P (DATA_WRITE_P) is fed to output DATA_WRITE bus while the write data in the secondary controller 6100S (DATA_WRITER_S) is fed only to the exclusive OR gate 6118.) Thus, the exclusive OR gate 6116 produces a logic 0 if WRITE_DATA_P and WRITE_DATA_S are the same and produces a logic 1 if they are different. The fault detector 6102 detects such logic 1 and reports the detected fault to the transfer requesting director.

[0190] In like manner, a check is made of the DATA read (READ_DATA) from the memory array region R₁which becomes stored in Read data register 6119, here a FIFO. The READ_DATA is fed to a read data register (here a FIFO) for transmission to the director via Port_A. Such READ_DATA in register 6119 indicated as READ_DATA_P is fed to an exclusive OR gate 6118. In like manner, secondary controller 6100S should produce the same signals on output READ_DATA_S. READ_DATA_P and READ_DATA_S are fed to an exclusive OR gate 6118. Thus, the exclusive OR gate 6118 produces a logic 0 if READ_DATA_P and READ_DATA_S are the same and produces a logic 1 if they are different. The fault detector 6102 detects such logic 1 and reports the detected fault to the transfer requesting director.

[0191] It is noted that the RAP and PAS signals are both sent to the arbitration logic 6004₁(FIGS. 13A, 13B, 13C, 13D and 13E) as composite signal REQUEST A. The arbitration section 6004₁considers a valid request only if both signals RAP and RAS are the same. In like manner, the arbitration logic 6004₁issues separate grant signals GAP and GAS which are shown in FIGS. 13A, 13B, 13C, 13D and 13E as a composite signal GRANT_A. Likewise, PORT_A_P_SELECT and PORT_A_S_SELECT signals are both sent to the arbitration logic 6004₁(FIGS. 13A, 13B, 13C, 13D and 13E) as composite signal PORT₁₃A_SELECT. The arbitration section 6004₁considers a valid request only if both signals PORT_A_P_SELECT and PORTA_S_SELECT are the same.

[0192] As noted above, the upper port B controller 6002B provides signals: RBP, GBP, PORT_B_P_SELECT, ADDR/CONTROL, DATA_WRITE RBS, GBS, PORT B_SELECT, and READ_DATA, which are equivalent to RAP, GAP, PORT A_SELECT, ADR/CONTROL, DATA_WRITE, RAS, GAS, PORT A_SELECT, and READ_DATA, respectively, which are provided by the upper port A controller 6002A

[0193] Referring now to FIGS. 15A, 15B, 15C, 15D and 15E, the arbitration logics 6004₁, 6004₂of the logic sections 5010₁, 5010₂, respectively, are shown along with the watchdog section 6006. It is first noted that the arbitration logic 6004₁, 6004₂are identical in construction.

[0194] Arbitration logic 6004₁is fed by:

[0195] REQUEST A (i.e., RAP, RAS) from upper port A controller 6002A of logic section 5010₁(FIGS. 13A, 13B, 13C, 13D and 13E);

[0196] REQUEST B (RBP, RBS) from upper port B controller 6002B of logic section 5010₁(FIGS. 13A, 13B, 13C, 13D and 13E);

[0197] REQUEST R from upper memory refresh section 6002R of logic section 5010₁(FIGS. 13A, 13B, 13C, 13D and 13E) (It is to be noted that the REQUEST R is made up of two signals, each being produced by identical primary and secondary identical memory refresh units, not shown, in memory refresh section 6002R both of which have to produce the same refresh signal in order for the arbitration logic 6004₁to respond to the refresh request). Arbitration logic 6004₂is fed by:

[0198] REQUEST A from upper port A controller 6002A of logic section 5010₂(FIGS. 13A, 13B, 13C, 13D and 13E);

[0199] REQUEST B from upper port B controller 6002B of logic section 5010₂(FIGS. 13A, 13B, 13C, 13D and 13E);

[0200] REQUEST R from upper memory refresh section 6002R of logic section 5010₂.

[0201] As shown in FIGS. 15A, 15B, 15C, 15D and 15E, each one of the three request signals REQUEST A, REQUEST B, and REQUEST R, produced in logic section 5010₁is fed, in triplicate, to three identical arbitration units, (i.e., arbitration unit I, arbitration unit II, and arbitration unit III) in the arbitration logic 6004₁of such logic section 5010₁, as indicated. (See also FIG. 19). Likewise, each one of the three request signals REQUEST A, REQUEST B, and REQUEST R, produced in logic section 5010₂is fed, in triplicate, to three identical arbitration units, (i.e., arbitration unit I, arbitration unit II, and arbitration unit III, in the arbitration logic 6004₂of such logic section 5010₂as indicated.

[0202] In response to such request signals, REQUEST A, REQUEST B, and REQUEST R, each arbitration unit I, II, and III determines from the three requests; i.e., REQUEST A, REQUEST B, and REQUEST R, fed to it and in accordance with the algorithm described above, whether upper port A controller 6002A, upper port B controller 6002B, or the memory refresh 6002R is to be given access to the memory array region R₁. As noted above, the operating Condition (i.e., Condition I, Condition II, or Condition III) is a function of whether the logic section 5010₁, is operating properly and whether the logic section 5010₂is operating properly. The watchdog section 2006 determines whether such logic sections 5010₁, 5010₂are operating properly. More particularly, when the arbitration units I, II, and III make their decision, they also produce a memory output enable (MOE) signals MOEI, MOEII and MOEIII, respectively, (when either logic section 5010₁or 5010₂is to be granted access to the memory array region R₁) and a memory refresh signal MREs (i.e., MREI, MREII and MREIII, respectively, when memory refresh section 6002R is to be granted access to the memory array region R₁). Thus, MOE signals MOEI₁, MOEII₁, and MOEIII₁, are produced by arbitration units I, II, and III, respectively, in arbitration logic 6004₁. Also, MRE signals MREI₁, MREII₁, and MREIII₁are produced by arbitration units I, II, and III, respectively, in arbitration logic 6004₁. In like manner, MOE signals MOEI₂, MOEII₂, and MOEIII₂are produced by arbitration units I, II, and III, respectively, in arbitration logic 6004₂. Also, MRE signals MREI₂, MREII₂, and MREIII₂are produced by arbitration units I, II, and III, respectively, in arbitration logic 6004₂. (See also FIG. 19).

[0203] These signals are fed to each of three identical watchdogs, WD_I, WD_II, WD_IIIas follows:

[0204] The MOE and MRE signals produced by the arbitration unit I in arbitration logics 6004₁and 6004₂(i.e., MOEI₁, MOEI₂, MREI₁and MREI₂) are fed to watchdog WD_III.

[0205] The MOE and MRE signals produced by the arbitration unit II in arbitration logics 6004₁and 6004₂(i.e., MOEII₁, MOEII₂, MREII₁and MREII₂) are fed to watchdog WD_II; and

[0206] The MOE and MRE signals produced by the arbitration unit III in arbitration logics 6004₁and 6004₂(i.e., MOEIII₁, MOEIII₂, MREIII₁and MREIII₂) are fed to watchdog WD_III.

[0207] Each one of the watchdogs I, II, III is implemented and arranged identical to perform the same logic functions; however, they preferably implemented with components manufactured independently of each other. Further, each one of the watchdogs I, II, and III operates in response to its own independent clock, i.e., Clock I, Clock II, and Clock III, respectively. Thus, each watchdog makes an independent determination as to whether these signals are in proper time and rate and thus, determine, in accordance with the “Condition algorithm” described above, the proper one of the Conditions (i.e., Condition I, Condition II, or Condition III) for the system. An indication of the Condition is provided by each of the watchdogs WD_I, WD_IIand WD_IIIas a two-bit word MSAB_I, MSAB_II, and MSAB_III, respectively. The two-bit word is produces as follows:

[0208] 00=Condition I

[0209] 01=Condition II

[0210] 10=condition III

[0211] 11=Reset (i.e., Condition IV)

[0212] These three words MSAB_I, MSAB_II, and MSAB_IIIare fed to both arbitration logics 6004₁and 6004₂, as indicated. It should be remembered that each one of the arbitration logics 6004₁and 6004₂(and hence the arbitration logics 6004₁and 6004₂therein), operate with a separate independent clock, Clock 1, and Clock 2, respectively. In order to synchronize the three words MSAB_I, MSAB_II, and MSAB_IIIare fed to logic section 5010₁and fed to logic section 5010₂. Each one of the arbitration logics 6004₁, 6004₂has a synchronization filter 6200₁, 6200₂to be described. Suffice it to say here, however, that the filter 6200₁produces corresponding signals MSAB_I1, MSAB_II1, and MSAB_III1, respectively, and filter 6200₂produce corresponding signals MSAB_I2, MSAB_II2, and MSAB_III2, respectively, as indicated

[0213] The signals MSAB_I1, MSAB_II1, and MSAB_III1, are fed to the arbitration units I, II, and III, respectively, in arbitration logic 6004₁. In like manner, the signals MSAB₁₂, MSAB_II2, and MSAB_III2, are fed to the arbitration units I, II, and III, respectively, in arbitration logic 6004₂. In response to such signals, each one of the arbitration units I, II, and III, makes an independent determination of whether logic section 5010₁(FIG. 13A, 13B, 13C, 13D and 13E) or logic section 5010₂will be granted access to the memory array region R₁. A grant by logic section 5010₁to logic section 5010₂is indicated by a Memory Grant (MG) signal. Thus, arbitration units I, II, and III of logic section 5010₁produce Memory Grant signals MGI₁, MGII₁, and MGIII₁, respectively. Such signals are fed to a synchronization filter 6202₂in arbitration logic 6004₂. The synchronization filter 6202₂operates as is constructed in the same manner as synchronization filters 6200₁and 6200₂. In like manner arbitration units I, II, and III of logic section 5010₂produce Memory Grant signals MGI₂, MGII₂, and MGIII₂, respectively. Such signals are fed to a synchronization filter 6202₁in arbitration logic 6004₁. The synchronization filter 6202₁operates as is constructed in the same manner as synchronization filter 6202₂.

[0214] Thus, considering exemplary synchronization filter 6202₂, such filter is fed by the three Memory Grant (MG) signals MGI₂, MGII₂, and MGIII₂. as indicated. The three signals are stored in registers 6204I, 6204II and 6204III, respectively, in response to a clock pulse produced by the Clock 2. Each of the three registers 6204I, 6204II and 6204 III, send the information stored therein to each of three majority gates MGI, MGII, and MGIII, respectively, as indicated. The majority gates produce an output which is the majority of the three inputs fed thereto. The outputs of the three majority gates MGI, MGII and MGIII are the arbitration units I, II and III, respectively, in the arbitration logic 6004₂, as indicated.

[0215] More particularly, referring to FIG. 16, portions of arbitration logics 6004₁and 6004₂are shown. The data to be fed to the output of arbitration logic 6004₁is clocked into register 7000₁of arbitration I, register 7000₂of arbitration II, and register 7000₃of arbitration III simultaneously in response to the same clock pulse produced by Clock 1. Thus, each of the registers 7000₁, 7000₂, 7000₃should store the same data at the clock pulse produced by Clock 1, as indicated in FIG. 18. The data is then fed to registers 7002₁, 7002₂, 7002₃of filter 6202₂of arbitration logic 6004₂. The data at the registers 7002₁, 7002₂, 7002₃are stored therein in response to the same clock produced by Clock 2. Because of the data in registers 7000₁, 7000₂, 7000₃arrive at registers 7002₁, 7002₂, 7002₃with different delays as indicated in FIG. 18, while the data in 7000₁, 7000₂7000₃is the same, here the data stored in registers 7002₁, 7002₂, 7002₃may be different as shown in FIG. 18. The data stored in register 7002₁is fed to majority gates (MGs) 7004₁, 7004₂and 7004₃. The data stored in register 7002₂is also fed to majority gates (MGs) 7004₁, 7004₂and 7004₃. Likewise, the data stored in register 7002₃is fed to majority gates (MGs) 7004₁, 7004₂and 7004₃. Each one of the majority gates MGs produces an output representative of the majority of the logic signals fed thereto as indicated in FIG. 17.

[0216] Referring now to FIGS. 20A, 20B and 20C, the three arbitrations I, II, and III of exemplary arbitration logic 6004₁are the signals fed thereto and produced thereby are shown in more detail. It is first noted that the primary signal REQUEST₁₃A_P, (RAP), and the secondary request signal REQUEST_A_S (RAS) are each fed in triplicate; one copy to each of the arbitrations I, II, and III, as indicated. The one of the triplicate RAP and RAS fed to arbitration I are fed to an AND gate 8000₁, a second one of the triplicate RAP and RAS fed to arbitration II are fed to an AND gate 8000₂, and the third one of the triplicate RAP and RAS fed to arbitration III are fed to an AND gate 8000₃, as indicated. Likewise, the signals REQUEST_B_P, (RBP), and REQUEST_B_S (RBS) are each fed in triplicate; one copy to each of the arbitrations I, II, and III, as indicated. The one of the triplicate RBP and RBS fed to arbitration I are fed to an AND gate 8002₁, a second one of the triplicate RBP and RBS fed to arbitration II are fed to an AND gate 8002₂, and the third one of the triplicate RBP and RBS fed to arbitration III are fed to an AND gate 8002₃, as indicated. As mentioned briefly above, there are two memory refresh units in the memory refresh section 6002R (FIGS. 13A, 13B, 13C, 13D and 13E). One, a primary unit (not shown), issues a request RRP and the other, a secondary unit (not shown), issues a request RRS. Above, in connection with FIGS. 13A, 13B, 13C, 13D and 13E, these two requests were considered as a composite request (REFRESH_REQUEST) to simplify the discussion presented above. Here, in connection with FIG. 19, the individual signals RRP, RRS are shown in more detail. Thus, the signals RRP, RRS are each fed in triplicate; one copy to each of the arbitrations I, II, and III, as indicated. The one of the triplicate RRP and RRS is fed to arbitration I are fed to an AND gate 8004₁, a second one of the triplicate RRP and RRS fed to arbitration II are fed to an AND gate 8004₂, and the third one of the triplicate RRP and RS fed to arbitration III are fed to an AND gate 8004₃, as indicated.

[0217] Thus, in the case of each pair, in order for the request to be issued to the arbitration I, II, or III. the AND gate associated therewith must see the same request from both the primary signal and the secondary signal fed to it.

[0218] Each arbitration I, II and II issues pairs of grants, i.e., a primary grant to the primary unit and a secondary grant to the secondary unit. Thus, each of the arbitrations I, II and III issues: the primary and secondary grants (GAP and GAS, respectively) to the Port A primary control section 6100P (FIGS. 14A, 14B, 14C and 14D) and Port A secondary control section 6100S of Port A controller 6002A; the primary and secondary grants (GBP and GBS, respectively) to the Port B primary control section and Port A secondary control section of Port B controller 6002B; and the primary and secondary grants (GRP and GRS, respectively) to the memory refresh primary unit memory refresh secondary unit of the memory refresh section 6002R (FIGS. 13A, 13B, 13C, 13D and 13E).

[0219] The arbitrations I, II, and III produce Memory Output Enable signals MOE_I-1, MOE_II-1, and MOE_III-1, respectively, as indicated, for the watchdogs WD_I, WD_IIand WD_II, respectively, as shown in FIGS. 15A, 15B, 15C, 15D and 15E. The arbitrations I, II, and III produce Memory Refresh Enable signals MRE_I-1, MRE_{II-1, and MREIII-1}, respectively, as indicated, for the watchdogs WD₁, WD_IIand WD_III, respectively, as shown in FIGS. 15A, 153B, 15C, 15D and 15E. The arbitrations I, II, and III produce Memory Grant signals MG_I, MG_I, and MG_III, respectively, as indicated, for the registers 6204₁, 6204_IIand 6204_III, respectively, of filter 6202₂of logic section 5010₂, as shown in FIGS. 15A, 15B, 15C, 15D and 15E.

[0220] Thus, it should be noted that while each one of the registers 7002₁, 7002₂, 7002₃, of filters 6002₁, 6002₂(FIGS. 19), are fed the same data from registers 7000₁, 7000₂, and 7000₃, respectively, because of the time skew shown in FIG. 18, such registers 7002₁. 7002₂, 7002₃, may not store the data which is in registers 7000₁, 7000₂, and 7000₃, respectively. However, the majority gates MG 7004₁-7004₃will produce the same data according to FIG. 17. Therefore, the three arbitrations I, II, and III of arbitration logic 6004₂will receive the same data (i.e., the data produced by the majority gates MG 7004₁-7004₃) thereby providing coherency (i.e., synchronization) to the arbitrations I, II, and III even though the arbitrations are operating independently of each other.

[0221] Other embodiments are within the spirit and scope of the appended claims.