Patent US5253353A successfully refreshed from Google Patents.
US5253353A
Expired / Lapsed
US
⏱️ 0.0 years remaining
System and method for efficiently supporting access to I/O devices through large direct-mapped data caches
Assignee: Hewlett Packard Development Co LP
•
Filing: 1990-01-02
•
Granted: 1993-10-12
•
Expires: 2010-01-02
Abstract
Abstract A data processing system (10) includes a CPU (12) connected to a direct-mapped cache (14) by address bus (16) and data bus (18). The cache (14) includes a first-level cache (20) connected to a second-level cache (22) by address bus (24) and data bus (26). The second-level cache (22) of the cache (14) is connected to address bus (28) and data bus (30) by address bus (32) and data bus (34). The address and data busses (28) and (30) are connected to memory (36) and I/O device (41) by address bus (40), data bus (42), address bus (44) and data bus (46), respectively. In the system (10), I/O interface (38) decodes physical memory addresses and responds to addresses in specific ranges using first and second addresses alternately, which are chosen to collide in the data cache (14). I/O software alternates between the two addresses instead of alternating between a device register address and a reserved-region address as in prior art systems.
Inventors
- 👤 Jeffrey C. Mogul
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-08 08:38
CPC Classifications
G; G06; G06F; G06F12/00; G06F12/02; G06F12/08; G06F12/0802; G06F12/0893; G06F12/0897
Claims Analysis
Total Claims: 21
Claim 1
Independent Claim
144 words
A data processing system, which comprises: a central processing unit; a main memory; at least one input/output device; a direct-mapped cache connected between said central processing unit and said main memory and between said central processing unit and said at least one input/output device; wherein said at least one input/output device having at least one register being addressed by a first address and a second address; both the first and second addresses being chosen to collide in said direct-mapped cache; said direct-mapped cache, said main memory and said at least one input/output device being addressed by means of addresses having a common format; and said central processing unit operates under control of an input/output program to address said at least one input/output device by alternately addressing the first and second addresses, with the first and second addresses chosen to collide in said direct-mapped cache.
Claim 2
Dependent Claim
144 words
A data processing system, which comprises: a central processing unit; a main memory; at least one input/output device; a direct-mapped cache connected between said central processing unit and said main memory and between said central processing unit and said at least one input/output device; wherein said at least one input/output device having at least one register being addressed by a first address and a second address; both the first and second addresses being chosen to collide in said direct-mapped cache; said direct-mapped cache, said main memory and said at least one input/output device being addressed by means of addresses having a common format; and said central processing unit operates under control of an input/output program to address said at least one input/output device by alternately addressing the first and second addresses, with the first and second addresses chosen to collide in said direct-mapped cache.
Claim 3
Dependent Claim
20 words
The data processing system of claim 1 in which said direct-mapped cache comprises a first-level cache and a second-level cache.
Claim 4
Dependent Claim
35 words
The data processing system of claim 2 in which an access time between said central processing unit and said first-level cache is less than an access time between said second-level cache and said main memory.
Claim 5
Dependent Claim
19 words
The data processing system of claim 2 additionally comprising a floating point unit connected to said central processing unit.
Claim 6
Dependent Claim
30 words
The data processing system of claim 4 in which said central processing unit, said floating point unit and said first-level cache are implemented together on a single integrated circuit chip.
Claim 7
Dependent Claim
57 words
The data processing system of claim 1 in which said central processing unit is connected to said direct-mapped cache by a first address bus and a first data bus and said direct-mapped cache is connected to said main memory and to said at least one input/output device by a second address bus and a second data bus.
Claim 8
Dependent Claim
30 words
The data processing system of claim 1 in which said system operates with memory addresses having a format including a tag field, a line field and a byte offset field.
Claim 9
Dependent Claim
67 words
The data processing system of claim 7 in which memory addresses for said at least one input/output device have a bit position ignored by said at least one input/output device and said central processing unit operates with a bitmask to use the bit position ignored by said at least one input/output device to address said at least one input/output device alternately with the first and second addresses.
Claim 10
Dependent Claim
25 words
The data processing system of claim 8 in which said cache and a register of said at least one input/output register have equal line sizes.
Claim 11
Dependent Claim
119 words
The data processing system of claim 1 in which said at least one input/output device is connected to said direct-mapped cache by an address bus and a data bus, said system additionally comprising an address recognizer connected to said address bus, said address recognizer including a comparator connected to receive addresses on said address bus as a first input and a value which will address the at least one register as a second input for comparison with the first input, said comparator being configured to ignore a bit in the addresses, an output of said comparator being connected as a control input to the at least one register, the at least one register being connected to said data bus.
Claim 12
Dependent Claim
30 words
The data processing system of claim 1 in which the at least one register of said at least one input/output device is a buffer memory comprising a plurality of registers.
Claim 13
Dependent Claim
141 words
A data processing method, comprising the steps of: providing a data processing system including a central processing unit, a main memory, at least one input/output device, a direct-mapped cache connected between the central processing unit and the main memory and between the central processing unit and the at least one input/output device, with the at least one input/output device having at least one register being addressable by first and second addresses chosen to collide in the direct-mapped cache; addressing the direct-mapped cache, the main memory and the at least one input/output device with addresses having a common format; and addressing the at least one register of the at least one input/output device with the first address and the second address by alternately addressing the first and second addresses, with the first and second addresses chosen to collide in the direct-mapped cache.
Claim 14
Dependent Claim
141 words
A data processing method, comprising the steps of: providing a data processing system including a central processing unit, a main memory, at least one input/output device, a direct-mapped cache connected between the central processing unit and the main memory and between the central processing unit and the at least one input/output device, with the at least one input/output device having at least one register being addressable by first and second addresses chosen to collide in the direct-mapped cache; addressing the direct-mapped cache, the main memory and the at least one input/output device with addresses having a common format; and addressing the at least one register of the at least one input/output device with the first address and the second address by alternately addressing the first and second addresses, with the first and second addresses chosen to collide in the direct-mapped cache.
Claim 15
Dependent Claim
22 words
The data processing method of claim 12 in which the direct-mapped cache is provided with a first-level cache and a second-level cache.
Claim 16
Dependent Claim
35 words
The data processing method of claim 13 in which an access time between the central processing unit and the first-level cache is less than an access time between the second-level cache and the main memory.
Claim 17
Dependent Claim
47 words
The data processing method of claim 12 in which the cache, the main memory and the at least one register of the at least one input/output device are addressed with memory addresses having a format including a tag field, a line field and a byte offset field.
Claim 18
Dependent Claim
96 words
The data processing method of claim 15 in which memory addresses for the at least one register of the at least one input/output device have a bit position ignored by the at least one register of the at least one input/output device, the method further comprising the step of examining the memory addresses with a bitmask to use the bit position ignored by the at least one register of the at least one input/output device to address the at least one register of the at least one input/output device alternately with the first and second addresses.
Claim 19
Dependent Claim
30 words
The data processing method of claim 16 in which the cache and the at least one register of the at least one input/output device are provided with equal line sizes.
Claim 20
Dependent Claim
74 words
The data processing method of claim 12 in which the at least one register of the at least one input/output device is addressed with the first address and the second address by comparing addresses with a value representing the address of the at least one register, ignoring one bit in the first address and the second address while comparing addresses, and enabling the at least one register when the value and the addresses match.
Claim 21
Dependent Claim
23 words
The data processing method of claim 12 in which the at least one register is a buffer memory comprising a plurality of registers.
References Cited (Backward Citations)
US4802085A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1996-10-11 | 1997-04-12 | 1997-10-12 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2000-10-11 | 2001-04-12 | 2001-10-12 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2004-10-11 | 2005-04-12 | 2005-10-12 | $7700.00 | Window Closed / Lapsed |