Electrical computers and digital processing systems: memory – Address formation – Slip control – misaligning – boundary alignment
Reexamination Certificate
1999-10-21
2003-02-25
Kim, Matthew (Department: 2186)
Electrical computers and digital processing systems: memory
Address formation
Slip control, misaligning, boundary alignment
C710S120000
Reexamination Certificate
active
06526496
ABSTRACT:
TECHNICAL FIELD
The present invention relates in general to data processing systems, and in particular, to burst instruction transfers from I/O subsystems in data processing systems.
BACKGROUND INFORMATION
In a modem data processing systems, a stream of instructions is often transferred from main memory into a cache memory in a burst read operation. The burst read transfers a plurality of instruction “words” which fills a cache line. (An instruction “word” includes a predetermined number of bits which constitute the instruction, which number of bits may be depend on the processor implementation.) Burst instruction transfers from a subsystem designed and optimized for transfers on word boundaries and a system bus designed and optimized for transfers on double word boundaries may give rise to misaligned reads. This may typically occur through devices that interface to the system bus through an input/output (I/O) subsystem. Exemplary implementations in which an instruction stream may be accessed through an I/O subsystem include initial program load read-only storage (IPLROS) and initialization or diagnostic code resident on I/O adapters.
Typically, to prevent misaligned reads, the I/O subsystem must be selected to be compatible with the system bus. This complicates the development of data processing system, particularly, as processor technology advances rapidly with a concomitant increase in system bus widths. Thus, there is a need in the art for a mechanism to align instruction reads across otherwise incompatible subsystems on a system bus.
SUMMARY OF THE INVENTION
The aforementioned needs are addressed by the present invention. Accordingly, there is provided, in a first form, a burst transfer alignment method. The method includes invalidating a cache line, in which the cache line includes at least one entry containing a predetermined data value representing an invalid instruction, the predetermined data value being loaded in response to detecting an unaligned read. The cache line is refetched at the address of the invalid instruction whereby an aligned transfer of the refetched line is effected.
There is also provided, in a second form, a data processing system. The system contains circuitry operable for invalidating a cache line, wherein the cache line includes at least one entry containing a predetermined data value representing an invalid instruction. System circuitry loads the predetermined data value in the cache line in response to circuitry operable for detecting an unaligned read. Also included is circuitry operable for returning to a selected one of the at least one entry containing an invalid instruction. A refetch of the invalidated cache line generates an aligned transfer.
Additionally, there is provided, in a third form, a computer program product operable for storage in machine readable storage media, the program product operable for burst transfer alignment. The program product includes programming for invalidating a cache line, wherein said cache line includes at least one entry containing a predetermined data value representing an invalid instruction. The program product also has programming for returning to a preselected one of said at least one entry containing an invalid instruction. A refetch of the invalidated cache line generates an aligned transfer.
The foregoing has outlined rather broadly the features and technical advantages of the present invention in order that the detailed description of the invention that follows may be better understood. Additional features and advantages of the invention will be described hereinafter which form the subject of the claims of the invention.
REFERENCES:
patent: 5577200 (1996-11-01), Abramson et al.
patent: 5897654 (1999-04-01), Eisen et al.
patent: 5996048 (1999-11-01), Cherabuddi et al.
patent: 6438653 (2002-08-01), Akashi et al.
Mano, Computer System Architecture, pp. 312-317, 1982.*
Handy, The Cache Memory Book, pp. 48-76, 1998.
Hartman Steven Paul
Lee Van Hoa
Miller, II Milton Devon
International Business Machines - Corporation
McBurney Mark E.
Newberger Barry S.
Tran Denise
Winstead Sechrest & Minick P.C.
LandOfFree
Burst instruction alignment method apparatus and method... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Burst instruction alignment method apparatus and method..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Burst instruction alignment method apparatus and method... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3165883