Efficient maintenance of memory list

The present application is related to commonly-assigned and co-pending U.S. application Ser. No.10/990,882, entitled METHOD AND SYSTEM FOR PERFORMANCE ENHANCEMENT IN A POLLED DATA STORAGE ENVIRONMENT, filed on the filing date hereof, which application is incorporated herein by reference in its entirety.

The present invention relates generally to computer data systems and, in particular, to maintaining and updating entries in a memory list.

In a data processing system, or in a component such as a disk control unit, customer data is frequently moved between storage elements, such as external storage drives, internal processor memory, conventional volatile RAM, non-volatile memory, etc. A list is typically maintained of entries which are associated with the data. Entries may include memory addresses, pointers or other means of identifying the data which is being moved. Entry lists conventionally have been built and maintained through instructions executed by a controller processor. However, building and maintaining a list consumes valuable processor resources. Moreover, spurious writes to the list may be undetected or an entry in the list may be corrupted, possibly leading to data being lost or unavailable and placing the integrity of the list in question.

Consequently, a need remains for a means of maintaining an entry list which is efficient and relieves the controller processor from at least some of its former responsibilities. There remains a further need to maintain an entry list while enhancing the ability to detect errors.

The present invention provides a protocol for managing memory entry lists in a data management system, such as a disk control unit. Preferably, the protocol is implemented substantially in hardware, thus freeing other system resources to execute other processes. Additionally, improved error correction may be performed as the list is being built and updated.

The list is initially built by adding new entries to pre-allocated space, each space having an address. A first entry is added to space having a first address and a pointer is set to reference the first address. If error checking is enabled, an error check value is generated. The next entry is added to space having the next address, the pointer is incremented to point to the next address and the error check value is updated. The process continues until all entries have been added to the list.

The list is updated by first determining if the update is the removal of an interstitial entry or of the last (most recently added) entry. If the update is the removal of an interstitial entry, that entry is removed from the list and, if error checking is enabled, the error check value is updated. The last entry is then written to the space vacated by the removed interstitial entry and a predetermined value which is neutral relative to the error checking procedure (such as all zeros) is written to the space vacated by the former last entry. The pointer is decremented to now reference the address “below” the former last entry; the entry stored in this address is the new last entry.

If the update is the removal of the last entry, the last entry is removed from the list and, if error checking is enabled, the error check value is updated. A predetermined neutral value is then written to the space vacated by the last entry and the pointer is decremented to now reference the address “below” the former last entry; the entry stored in this address is the new last entry.

The protocol, including error checking if enabled, is efficient and may be implemented substantially in hardware.

FIG. 1 is a block diagram of a data management system of the present invention;

FIGS. 2A-2E graphically illustrate a process in which a memory entry list is built in accordance with the present invention;

FIGS. 3A-3D graphically illustrate a process in which an interstitial entry is removed in accordance with the present invention;

FIGS. 4A-4D graphically illustrate a process in which the last entry is removed in accordance with the present invention; and

FIGS. 5A and B are a flow diagram of the memory list assist logic in accordance with the present invention.

The present invention provides an improved entry list management protocol and employs hardware logic to reduce spurious writes to the list and to reduce the burden on the controller processor. Optionally, the present invention may also include improved error detection to enhance the integrity of the memory list. FIG. 1 is a block diagram of a data management system 100 of the present invention. Such a system may be any data management system in which a contiguous entry list is maintained, including fixed or removable magnetic disk storage systems, tape storage systems or other media storage systems. In one embodiment, the system 100 comprises a disk control unit. The system 100 includes a controller processor 102 which is attached to a data source 104 , such as a disk array, and to a memory device 106 , which may be used to store instructions executable by the processor 102 . The system 100 further includes entry assist logic 500 which interfaces with an entry list 110 and executes the list management protocol. Preferably, the system also includes error checking (EC) logic 108 for enhanced performance. In one embodiment, the error checking logic 108 may perform a longitudinal redundancy check (LRC); however, other error checking procedures, including those with correction capability, may be employed in the present invention.

The entry list 110 may reside in a memory (including the memory device 106 ) or may reside in a register. The entry list 110 initially comprises empty spaces which have been allocated to the list 110 , each of which has an address, such as 0-n, for a predetermined number of future entries.

As illustrated in FIGS. 2A-E , when the entry list 110 is being built each new entry is added to the empty space with the lowest address in the list 110 . This is the most recently added entry and is designated the “last entry”. Preferably, each new entry is processed through EC logic 108 ( FIG. 2A ) before being stored in the list. For example, the new entry data DZP (letters chosen at random) 200 A is optionally but preferably processed through the EC logic 108 which generates an error check value. The entry 200 A is then written to the lowest empty space, in this case the space having 00 as its address ( FIG. 2B ). A pointer 220 is incremented to reference the address in which the last entry is stored. The next new entry data GBM 200 B is processed through the EC logic 108 ( FIG. 2C ), which updates the error check value. The entry 200 B is then written to the next available space (address 01) and the pointer is incremented to reference the new last entry ( FIG. 2D ). The process continues until the entry list 110 is filled ( FIG. 2E ), the EC logic 108 has generated a fully updated error check value and the pointer references the most recently added (last) entry space (address 05 in the FIG.).

After the entry list has been built, entries may be updated or removed. FIGS. 3A-D illustrate the manner in which an interstitial entry, such as EAQ 200 C in address space 02 may be removed according to the present invention. First, a new entry 200 G is compared with the last entry 200 F; both entries are the same (HYH). The target address of the new entry 200 G is 02 and is compared with valid list addresses. If the target address is not a valid list address, an error is noted. Because the target address is determined to be a valid address, the old (interstitial) entry 200 C is then removed from the list 110 and processed through the EC logic 108 which updates the error check value ( FIG. 3A ). The old entry EAQ 200 C is then discarded and, to fill the now empty space at address 02, the new entry 200 G replaces the discarded entry ( FIG. 3B ). Because the last entry data 200 F is the same as the new entry data 200 G, the last entry 200 F is discarded and replaced by a value 200 H, such as 000, which is neutral with respect to the error checking logic 108 ( FIG. 300C ). Consequently, it is not necessary to again update the error check value again. Finally, the pointer 220 to the last entry is adjusted such that the entry 200 E (KVB) is referenced as the last entry. In this manner, the last entry space is freed and gaps in the list are closed.

FIGS. 4A-D illustrate the manner in which the last entry, such as HYH 200 F may be removed according to the present invention. First, the target address of the new entry is confirmed to be the address of the last entry 05 and the new entry data itself 000 is confirmed to be neutral with respect to the error checking logic 108 ( FIG. 4A ). Next, the last entry HYH 200 F is removed and processed through the error checking logic 108 ; the new entry 200 J replaces the former last entry ( FIG. 4B ). The former last entry HYH 200 F is then discarded ( FIG. 4C ) with only one entry being processed through the error checking logic 108 . Finally, the pointer is decremented such that the entry 200 E (KVB) in the space with the address 04 becomes the last entry ( FIG. 4D ).

FIGS. 5A and 5B are a flow diagram of the hardware-based entry list assist logic 500 used to execute the list management protocol of the present invention after the entry list has been built. The scope of the present invention is not limited to the steps being performed in the order described herein; it will be appreciated that the steps may be performed in other orders. After the logic is invoked ( 502 ), a determination is made whether ECHO assist logic has been enabled ( 504 ). ECHO assist ( 506 ) is described and claimed in the aforementioned commonly-assigned and co-pending U.S. application Ser. No. 10/990,882 and will not be described herein. If ECHO assist is not enabled, or upon completion of the ECHO assist if it is enabled, a determination is made whether the entry list assist of the present invention is enabled ( 508 ). If so, a determination is next made whether an entry (an interstitial entry or a last entry) is to be updated ( 510 ) or, if not, made whether there is space to add a new entry ( 512 ) to the list 110 . More specifically, the target address of the new entry is checked to determine if it is the same as the address of an existing entry or if it is the same as the address of the next available space in the list (that is, the target address is a valid list address). If neither decision is affirmative, then an error is likely indicated and an appropriate flag is set ( 514 ) after which the logic returns control to the original process ( 560 ).

If an existing entry is to be updated ( 510 ), the last entry in the list 110 is read ( 516 ) and replaced in the last space with a value, such as all 0's, which is neutral relative to the error checking logic 108 ( 518 ) and the address pointer (which actually designates the addresses of the entry spaces) is decremented ( 520 ) such that the next entry now becomes the last entry. A comparison is then made between the new entry and the old last entry ( 522 ). If the two do not match, an error is logged ( 524 ) and the process returns ( 560 ). If the two entries match and if the error checking logic is enabled ( 526 ), the entry being updated is read ( 528 ) and processed through the error checking logic to update the error check value ( 530 ). If the error checking assist is disabled ( 526 ), or after the error check value has been updated, the new entry is written (stored) in the existing space ( 532 ) and control returns ( 560 ).

If a new entry is to be added to the list and there is space available at the end of the list ( 512 ), the new entry is written to the empty space ( 534 ) and the address pointer is incremented ( 536 ), thereby making the new entry the Last Entry and adjusting the addresses of the other entries. If the error checking logic is enabled ( 538 ), the new entry is processed and the error check value updated ( 540 ). If the error checking logic is disabled, or upon updating the error check value if the error checking logic is enabled, control returns ( 560 ).

If the assist logic is disabled ( 508 ), a decision is made whether the error checking logic is enabled ( 542 ). If not, the entry is merely written to the list ( 544 ) and control returns ( 560 ). Otherwise, the entry is and processed through the error checking logic ( 546 ) before being written to the list ( 544 ). Control is then returned ( 560 ).

Consequently, because the read, compare, error detection and write operations are performed by the entry list assist logic 110 , the burden on other system resources is reduced. Additionally, the list protocol reduces the risk of unintended writes to the entry list, thereby improving its reliability.

It is important to note that while the present invention has been described in the context of a fully functioning data processing system, those of ordinary skill in the art will appreciated that the processes of the present invention are capable of being distributed in the form of a computer readable medium of instructions and a variety of forms and that the present invention applies regardless of the particular type of signal bearing media actually used to carry out the distribution. Examples of computer readable media include recordable-type storage media such as a floppy disk, a hard disk drive, a RAM, and CD-ROMs.

The description of the present invention has been presented for purposes of illustration and description, but is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiment was chosen and described in order to best explain the principles of the invention, the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated. Moreover, although described above with respect to an apparatus, the need in the art may also be met by a method of efficient maintenance of memory list, a computer program product containing instructions for efficient maintenance of memory list, or a method for deploying computing infrastructure comprising integrating computer readable code into a computing system for efficient maintenance of memory list.