WO1991015825A1 - Procede et apprareil d'obtention d'une interface amelioree pour systeme informatique - Google Patents

Procede et apprareil d'obtention d'une interface amelioree pour systeme informatique Download PDF

Info

Publication number
WO1991015825A1
WO1991015825A1 PCT/US1991/001255 US9101255W WO9115825A1 WO 1991015825 A1 WO1991015825 A1 WO 1991015825A1 US 9101255 W US9101255 W US 9101255W WO 9115825 A1 WO9115825 A1 WO 9115825A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
sel
bus
continue
bytes
Prior art date
Application number
PCT/US1991/001255
Other languages
English (en)
Inventor
Kumar Gajjar
Kaushik S. Shah
Duc H. Trang
Original Assignee
Sf2 Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sf2 Corporation filed Critical Sf2 Corporation
Publication of WO1991015825A1 publication Critical patent/WO1991015825A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/38—Information transfer, e.g. on bus
    • G06F13/42—Bus transfer protocol, e.g. handshake; Synchronisation
    • G06F13/4204—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus
    • G06F13/4208—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being a system bus, e.g. VME bus, Futurebus, Multibus
    • G06F13/4213—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being a system bus, e.g. VME bus, Futurebus, Multibus with asynchronous protocol
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/38—Information transfer, e.g. on bus
    • G06F13/42—Bus transfer protocol, e.g. handshake; Synchronisation
    • G06F13/4204—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus
    • G06F13/4221—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being an input/output bus, e.g. ISA bus, EISA bus, PCI bus, SCSI bus
    • G06F13/4226—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being an input/output bus, e.g. ISA bus, EISA bus, PCI bus, SCSI bus with asynchronous protocol
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
    • G06F3/0671—In-line storage system
    • G06F3/0673—Single storage device

Definitions

  • This invention relates to an improved computer interface for interconnecting intelligent computer devices. More particularly, this invention relates to extensions and enhancements of the proposed American National Standard For Information Systems (ANSI) Small Computer System Interface-2 (SCSI-2) , Rev. 10, described in ANSI document number X3.131-198X which is hereby incorporated by reference herein.
  • ANSI National Standard For Information Systems
  • SCSI-2 Small Computer System Interface-2
  • the SCSI-2 standard defines an I/O bus for interconnecting computers and peripheral devices.
  • the bus can be operated over a wide range of data rates.
  • the SCSI-2 standard provides specifications for mechanical, electrical and functional characteristics of the bus, including definitions of the physical characteristics of the bus conductors, the electrical characteristics of the signals that the conductors carry, and the meanings of those signals (e.g., control or data) .
  • the standard further defines protocols for communicating between devices interconnected by the bus. Communication is allowed between only two devices at any given time (although up to eight devices may be interconnected by the same bus) .
  • the initiator originates an operation (i.e., requests an I/O process to be performed) and the target performs the operation.
  • Transfers on the bus are typically asynchronous (although a synchronous option is defined) and follow a "handshaking" protocol involving the exchange of a "Request” signal from a target and an "Acknowledge” signal from an initiator.
  • the standard defines a message protocol for managing transfers on the bus. For example, a target may send a "Disconnect" message to inform an initiator that a present connection is going to be broken, and that a later reconnect will be required to complete the current I/O process.
  • the standard defines a command and status structure. Commands are used by an initiator to request a target to perform particular I/O operations. At the completion of a command, or if for some reason a command cannot be completed by the target, the target sends a status byte to the initiator to inform the initiator of its condition.
  • the specifications of the SCSI-2 standard thus combine to define an interface having multiple protocol levels.
  • the defined interface provides computer systems with device independence within a class of devices. For example, a variety of mass storage devices (such as disk drives, tape drives, optical drives, and memory caches) , printers, microprocessors, and other devices can be added to a computer system without requiring modifications to system hardware or software.
  • mass storage devices such as disk drives, tape drives, optical drives, and memory caches
  • printers, microprocessors, and other devices can be added to a computer system without requiring modifications to system hardware or software.
  • special features and functions of individual devices can be handled through the use of device-dependent fields and codes in the command structure.
  • the SCSI-2 standard further defines that initiator and target devices are daisy-chained together using a common 50-conductor "A" cable and, optionally, a 68-conductor "B” cable.
  • a cable permits 8-bit wide data transfers.
  • B cable allows wider information transfers (data only) of 16 or 32 bits.
  • SCSI-2 interface provides a convenient and powerful means for system interconnection, it has limitations.
  • One important limitation is that messages, commands, and status information may only be transferred a single byte at a time on a designated group of eight data lines.
  • a message, command, or status communication is more than one byte, as some of them are in the SCSI-2 protocol, separate transfer operations must be executed for each byte of the message, command, or status communication. This limits the speed of the interface.
  • This limitation becomes particularly restrictive as computer devices become capable of responding to increasingly complicated (and longer) commands.
  • This limitation also tends to make the prior art SCSI-2 interface less desirable for use within I/O systems in which intelligence is distributed among several processors throughout various components of the system.
  • an interface (referred to hereafter as a SCSI-2E interface) comprising, in the preferred embodiment, a bus capable of simultaneously transferring at least 32 bits of data (plus at least 4 bits of parity data) , and a microsequencer-driven interface unit for connecting devices to the bus.
  • a SCSI-2E interface comprising, in the preferred embodiment, a bus capable of simultaneously transferring at least 32 bits of data (plus at least 4 bits of parity data) , and a microsequencer-driven interface unit for connecting devices to the bus.
  • a conventional 64-conductor cable has been found to be suitable for implementing the bus of this invention (referred to hereafter as a SCSI-2E cable) , although cables of other widths may also be used.
  • the SCSI-2E cable includes nine dedicated control lines. The remaining lines provide paths for terminator power and ground. The various data, parity, control, and other lines may be arranged within the cable as desired.
  • the SCSI-2E interface offers several advantages over the prior art. For example, a single SCSI-2E cable replaces two prior art SCSI-2 cables (i.e., cables A and B) . Also, the SCSI-2 A and B cables require separate "Request” and "Acknowledge” lines for each of the cables. These two sets of Request and Acknowledge lines are required to synchronize the transfer of data on cables A and B during multiple-byte parallel data transfers.
  • All four data bytes of a 32-bit wide parallel data transfer are synchronized with a single request signal and a single acknowledge signal.
  • a single request signal and a single acknowledge signal can be used to synchronize the simultaneous transfer of multiple bytes of message, command or status information.
  • the interface unit of the present invention can be implemented using commercially available integrated circuit components.
  • the interface unit preferably includes a circuit for controlling the timing of signals on the bus and for generating interrupts in response to changes in the signals on the bus (e.g., a commercially-available SCSI integrated circuit chip) , and a microsequencer circuit for controlling the operation of the timing control circuit.
  • the microsequencer operates under the control of a microprocessor in the device to which the interface unit is attached, and handles interrupts generated by the timing control circuit during message/command/status information transfers, thereby preventing that circuit from overburdening the microprocessor.
  • the SCSI-2E cable and interface unit of this invention preferably transfer all information, including data, commands, status information, and messages, in 32-bit words.
  • the prior art SCSI cables are capable of transferring commands, messages, and status information only on the data path of cable A, with a maximum size of eight bits.
  • the SCSI-2E interface of the present invention therefore significantly enhances command, message and status information transfers between devices connected via the SCSI-2E cable and interface units.
  • FIGS, la and lb are diagrams of the lines of a prior art SCSI-2 interface bus
  • FIG. 2 is a diagram of the lines of a SCSI- 2E interface bus constructed in accordance with the principles of this invention
  • FIG. 3 is a block diagram of a system in which the SCSI-2E interface bus of FIG. 2 is suitable for use in accordance with the principles of the present invention
  • FIG. 4 is a block diagram of a preferred embodiment of an interface unit constructed in accordance with the principles of this invention.
  • FIG. 5 is a table of exemplary messages, and their corresponding message code words and attributes, for use in implementing a message protocol to manage communication on the SCSI-2E interface bus of the present invention
  • FIGS. 6-9 are flowchart diagrams illustrating the steps of various exemplary operations for transferring information over the SCSI-2E interface bus of the present invention.
  • FIG. 10 is a state diagram of the SCSI-2E interface of the present invention.
  • FIGS, la and lb are diagrams of the prior art SCSI-2 cables.
  • Cable “A,” shown in Fig. la, is a 50-conductor cable, with nine control lines (including two lines for "request” and “acknowledge” functions) , eight data lines, and a single parity line. The remaining lines are used as terminator power and ground lines.
  • FIG. la the various lines of cable A are shown grouped according to their respective functions. Table 1 below illustrates the particular designations of the 50 conductors of cable A as specified by the SCSI standard. - 9 -
  • Cable “B,” shown in Fig. lb, is a 68- conductor cable.
  • Cable B consists of 24 data lines and three parity lines (one parity bit per eight data bits) , two lines for "request” and “acknowledge” functions, and terminator power and ground lines.
  • the various lines of cable B are grouped in FIG. lb according to their respective functions.
  • Table 2 below illustrates the particular designations of the 68 conductors of cable B as specified by the SCSI standard. TABLE 2
  • the A and B cables can be operated to transfer 32 bits of data in parallel by operating cable B as a slave to cable A. This is accomplished according to the SCSI-2 standard by synchronizing the ACKB and REQB signals on conductors 22 and 26 of cable B with their respective counterpart signals ACK and REQ on conductors 38 and 48 of cable A.
  • FIG. 2 is a diagram of a SCSI-2E bus in accordance with the principles of this invention.
  • the SCSI-2E cable preferably is a single, 64-conductor cable which includes nine control lines, 32 data lines, and four parity lines.
  • the data lines are segregated into four groups of eight, with one parity bit per eight data bits. The remaining lines may be used as terminator power and ground lines.
  • Table 3 below illustrates an exemplary assignment of particular bus signals to the 64 conductors of the cable of the present invention.
  • the SCSI-2E bus may, alternatively, include more than one cable, or may include more or less than 64 conductors (e.g., by varying the number of ground lines) .
  • the number of conductors may be selected to accommodate a standard cable size. Regardless of the type of cable or number of cables used in implementing the present invention, however, it will be appreciated that, upon consideration of the detailed description of the invention herein, the transfer of all data, message, command and status bytes across the interface can be synchronized with a single request signal from a first device connected to the SCSI-2E bus and a single acknowledge signal from a second device connected to the SCSI-2E bus.
  • the SCSI- 2E interface of the present invention can be used to interconnect a variety of devices.
  • Such devices may include mass storage devices (e.g., disk drives, tape drives, optical drives, and memory caches) , printers, and microprocessors.
  • the SCSI-2E interface of the present invention is used to interconnect a mass storage device controller to a plurality of mass storage devices. More particularly, the mass storage devices of the preferred embodiment each comprises one or more arrays of disk drives. Each array is operated under the control of an array controller that communicates with the aforementioned mass storage device controller via the SCSI-2E interface of the present invention.
  • FIG. 3 shows an example of a system 100 in which the SCSI-2E interface of the present invention can be used and which is well-suited to illustrate the present invention.
  • two SCSI-2E buses 102a and 102b are used to interconnect, respectively, two device controllers 104a and 104b with a plurality of mass storage devices 107.
  • Each of device controllers 104a and 104b has a plurality of data ports 106(0)-(3) for communicating with the mass storage devices 107.
  • Each storage device 107 includes two array controllers 108 and a pair of associated device sets 109 (e.g., disk drive arrays).
  • Each array controller 108 has two data ports 106a and 106b that are connected to permit the array controller to communicate independently (and, if desired, simultaneously) with each of device controllers 104a and 104b. More specifically, SCSI-2E bus 102a connects the data port 106(0) of device controller 104a to the data ports 106a of each array controller 108 in system 100. Likewise, SCSI-2E bus 102b connects the data port 106(0) of device controller 104b to the data ports 106b of each array controller 108 in system 100.
  • Device controllers 104a and 104b act as "initiators," and array controllers 108 as “targets,” as those terms have been previously defined.
  • Device controllers 104a and 104b are also referred to herein by the term “PAA” (Parallel Array Adapter)
  • array controllers 108 are also referred herein to by the term “PAC” (Parallel Array Controller) .
  • SCSI-2E buses 102a and 102b are identical. For purposes of convenience, reference will be made in the following discussion to SCSI-2E bus 102a only, it being understood that the following discussion applies to bus 102b as well.
  • Array controllers 108 are interconnected by SCSI-2E bus 102a in a daisy-chain manner. This can be accomplished by providing a bi-directional cable terminated at one end by data port 106(0) of device controller 104a and at the other end by one of data ports 106a, and having taps along its length for connecting branch cables to the other data ports 106a in system 100.
  • FIG. 4 shows a preferred embodiment of interface unit 110, constructed in accordance with the principles of this invention.
  • the particular embodiment of interface unit 110 shown in FIG. 4 and described hereafter is designed for use in array controller 108. It is to be appreciated, however, that the same circuit can be implemented for use in a variety of devices, including both initiator and target devices, with only minor modifications.
  • the interface of the present invention can be implemented using interface units other than the embodiment described herein.
  • the interface unit embodiment described herein is driven by a microsequencer circuit to reduce the amount of time that a processor must be involved in the operation of the interface unit.
  • a processor-time-saving design may be unnecessary, and an interface unit driven directly by a processor in the device may be preferred.
  • device controller 104a initiates a data transfer by "selecting" an array controller 108, and sends commands (e.g., read data) to the selected array controller via SCSI-2E bus 102a.
  • the interface unit 110 within the data port 106a of the selected array controller receives the commands, and then sends the commands to array controller 108. While array controller 108 performs the commands, interface unit 110 preferably "disconnects" from SCSI-2E bus 102a, to permit device controller 104a to perform other operations (i.e., with other interface units and with other storage devices) .
  • Interface unit 110 "reselects" the same device controller 104a when array controller 108 has executed the command and is ready to reconnect to controller 104a.
  • the manner in which the steps of "select,” “disconnect,” “reselect” and other operations relating to the management of communication on bus 102a are performed is determined by the protocol of the SCSI-2E interface. In many respects, this protocol may, if desired, follow the standard SCSI-2 protocol, modified only as necessary to account for the parallel transfer of multiple byte messages, commands and status information.
  • the SCSI-2 standard defines eight distinct bus phases: BUS FREE, ARBITRATION, SELECTION, RESELECTION, COMMAND, DATA, STATUS and MESSAGE. These phases are defined by certain bus control signals which are specified by the SCSI-2 standard and are identified below. Preferably, the interface of the present invention is implemented using these same bus control signals and bus phases.
  • the control information transferred during the COMMAND, STATUS and MESSAGE phases may follow the SCSI-2 standard, or a variation thereof.
  • FIG. 5 a subset of messages that might be used in implementing the interface of the present invention is shown in FIG. 5.
  • Some of these messages (e.g., '-.Command Complete”) are the same as in the SCSI-2 standard, except that they are transferred in words, and not single bytes.
  • a brief description of each message shown in FIG. 5 follows. "Command Complete (00)”: This message word is sent from the array controller to the device controller after the array controller performs an operation requested by the device controller and sends an end response packet (i.e., a status communication).
  • "Selected With Disconnect (01)” This message word is sent from the device controller to the array controller on initial selection.
  • Disconnect (03) This message word is sent from the device controller to the array controller when the device controller is not ready to continue or complete an operation and wants the array controller to disconnect.
  • Disconnect with Interrupt Pending (04) This message word is sent from the array controller to the device controller in response to the above Disconnect (03) Message Code.
  • Each message in the set of FIG. 5 is a multiple byte message.
  • Each Message Code is four bytes wide, although for each of the messages in the set of FIG. 5 the upper three bytes are undefined and may be any value.
  • the other columns of FIG. 5 indicate attributes that are associated with the
  • Message Codes and that form a part of each message. As can be seen, some messages have no attributes, and are thus only one word in length. Other messages have one, two or three words of attributes.
  • the messages are transferred across bus 102a one word at a time, and thus messages comprising the group generally designated 500 can be sent in a single parallel transfer, those in group 502 can be sent in two transfers, those in group 504 in three transfers and those in group 506 in four.
  • a sixteen byte message such as is shown in group 506 would require sixteen separate transfers on a conventional SCSI-2 bus (assuming all bytes, including the three bytes of the Message Code word having undefined values, are transferred) .
  • interface unit 110 includes a programmable microsequencer 112, SCSI chip 114, first-in first-out (FIFO) storage circuits 116a-116d, multiplexers 118, 119, 121, and various registers and drivers.
  • FIFO first-in first-out
  • interface unit 110 transfers device-dependent information (e.g., commands and data) between the SCSI-2E interface bus 102a and a bus (which is designated by reference numeral 120) connected to array controller 108.
  • Interface 110 also transmits and receives over the SCSI-2E interface bus 102a device-independent information (e.g., messages and control signals) for managing communication between the interface units 110 connected by the SCSI-2E bus.
  • device-independent information e.g., messages and control signals
  • the aforementioned control signals are handled by SCSI chip 114.
  • the messages on the other hand, like commands and data, are passed on to microsequencer 112 and array controller 108 for handling.
  • the data and parity lines of SCSI-2E bus 102a connect to interface unit 110 in parallel, via bi ⁇ directional SCSI chip 114 and bi-direct nal FIFO circuits 116b-ll6d.
  • the ground and power lines of bus 102a are not shown.
  • parity check circuits which may be connected and used in interface unit 110 in a conventional manner to validate transfers on the various data lines within, or connected to, the interface unit.
  • Each of FIFO circuits 116b-d receives eight data bits and one parity bit via the SCSI-2E bus. The remaining eight bits of data (D(00)-D(07) ) and the corresponding parity bit (D(P0)) pass through a FIFO circuit internal to SCSI chip 114, before being transmitted to FIFO circuit
  • Chips suitable for use as FIFO circuits 116a-d in the present invention are commercially available from Cypress Semiconductor, San Jose, California, model number CY7C409 64 x 9 FIFO. If desired, additional data and parity lines may be added by adding additional FIFO circuits, connected in the same manner as shown with respect to FIFO circuits 116b-116d. Cable width may be expanded and/or additional cables may be added if desired to overcome constraints caused by the cable width.
  • SCSI chip 114 also receives control signals from the nine control lines of bus 102a. These nine control lines, and the bus signals they carry, are the same as those of the "A" cable of the standard SCSI-2 interface. As set forth in Tables l and 3, these control signals are: “busy (BSY),” “select (SEL),” “control/data (C/D),” “input/output (I/O),” “message (MSG) ,” “request (REQ) ,” “acknowledge (ACK) ,” “attention (ATN) ,” and “reset (RST) .” For a further discussion of these bus signals, the reader is referred to the above-referenced SCSI-2 ANSI standard document.
  • SCSI chip 114 is the main communication and bus control hardware of interface unit 110. SCSI chip 114 controls the timing and handshaking (e.g., request and acknowledge signals) for transfers across bus 102a. Chips suitable for use as the SCSI chip 114 in the present invention are commercially available from NCR Microelectronics, Colorado Springs, Colorado, model Nos. NCR53C90A and 53C90B (either chip version may be used) .
  • SCSI chip 114 receives instructions from microsequencer 112 via a command register 114a in SCSI chip 114, and generates interrupt signals for the microsequencer when it has completed the assigned instructions (or when a change in the signals on bus 102a for which SCSI chip 114 has been instructed to wait takes place) .
  • the interrupt signals are generated by setting a bit in control register 130 associated with microsequencer 112, via interrupt line 114b.
  • Microsequencer 112 analyzes interrupts generated by SCSI chip 114 by reading interrupt status register 114c of the chip.
  • Interrupt status register 114c is an addressable register.
  • Microsequencer 112 accesses interrupt status register 114c, and other addressable registers of SCSI chip 114 such as command register 114a, by supplying the appropriate register address via select register 136 to the register address terminals 114e of chip 114 and selecting the chip via its chip select ("CS") terminal. The contents of the addressed register can then be read from (and written to) using data terminals 114d.
  • CS chip select
  • microsequencer 112 is initiated by a command from a general purpose microprocessor 150 associated with the array controller 108.
  • Microprocessor 150 interfaces with microsequencer 112 via a 9-bit wide bus (eight data bits plus one parity bit) .
  • This 9-bit wide bus comprises the least significant byte of a larger 36-bit wide local bus extension (LBE) 120 that connects microprocessor 150 with other components of array controller 108, and that connects the 36 bit wide data path of interface unit 110 to a data buffer 152 in array controller 108.
  • LBE local bus extension
  • the same 9-bit wide portion of bus 120 is shown at two separate locations (i.e., in the upper left hand corner and middle right of FIG. 4); both represent the least significant byte portion of bus 120.
  • microsequencer 112 In response to commands from microprocessor 150, microsequencer 112 prepares interface unit 110 for receiving or transmitting message/command/status information over the SCSI-2E bus 102a, and is also responsible for transferring information from SCSI chip 114 to FIFO circuit 116a. Microsequencer 112 handles all interrupts from SCSI chip 114 during message/command/status information transfers, thereby preventing the SCSI chip from overburdening microprocessor 150 of array controller 108. When data is being transferred across SCSI-2E bus 102a (i.e., when bus 102a is in the DATA phase) , interface unit 110 is controlled by the general purpose microprocessor 150, and microsequencer 112 is in an idle state.
  • a programmable microsequencer suitable for use in the present invention is commercially available from Advanced Micro Devices, Inc., Sunnyvale, California, model No. Am29CPL154.
  • An example of a program for this model microsequencer for use in interface unit 110 is provided as at the end of the description portion of this application.
  • a microprocessor could be used in place of microsequencer 112 without departing from the spirit of the invention.
  • Such a microprocessor, in addition to performing the functions of microsequencer 112 could perform other processes related to data handling and data transfers (e.g., error handling functions).
  • FIFO circuits 116b-116d are used to extend the number of data lines in the interface.
  • Each FIFO circuit adds data-transfer capability of, e.g., eight data bit signals plus a parity bit signal, to the interface, and includes the control logic needed to hold the data on a first-in, first-out basis.
  • FIFO circuits I16b-116d operate in synchronism with SCSI circuit 114 when transferring information to or from bus 102a.
  • a target device such as array controller 108
  • FIFO circuits 116b-116d are synchronized with SCSI chip 114 using the "Acknowledge (ACK)" signal from bus 102a, as shown in FIG. 4.
  • the "Request (REQ) " signal from bus 102a is used instead as the synchronizing signal.
  • the "Input/Output (I/O)" signal from bus 102a may be
  • FIFO circuits 116b- 116d ⁇ sed to provide a control signal to FIFO circuits 116b- 116d to determine whether the FIFO circuits are to read data from or write data to bus 102a in response to the synchronizing ACK/REQ signal.
  • FIFO circuits of different conventional types can be used to implement FIFOs 116b-116d, and thus the particular signals required to control the FIFOs (e.g., select, shift in, shift out, input ready, output ready, reset) are implementation specific. It is well within the skill, and the discretion, of one in the art of digital circuit design to provide any logic circuits that may be necessary, in addition to those shown in FIG. 4, to operate FIFO circuits 116b-116d in accordance with the principles of the present invention as described herein.
  • Control signals for managing the timing and flow of information through interface unit 110 are routed via various multiplexer, register and other logic circuits.
  • FIG. 4 illustrates the signal paths for exemplary ones of these control signals. Again, these control signals may vary depending on the particular circuit components used in implementing interface unit 110, and it is within the skill and discretion of the circuit designer to choose the particular manner in which the timing and flow of information through interface unit 110 is to be controlled.
  • bus 120 serves several purposes. It operates as the conduit for instructions to be passed from microprocessor 150 to microsequencer 112, and for status information to be sent in return from microsequencer 112 to microprocessor 150. It also operates as the conduit for microsequencer 112 to pass instructions to SCSI chip 114, and for SCSI chip 114 to - 24 -
  • microsequencer 112 provides status information to microsequencer 112 in return. It further operates as the conduit for commands and messages to be passed between bus 102a and microprocessor 150, and for data to be passed between bus 102a and the mass storage devices of device set 109. A typical I/O operation will require bus 120 to serve these various purposes at different times during the operation. Control of bus 120 is therefore shared by microprocessor 150, microsequencer 112 and SCSI chip 114, as described further below.
  • Transfers on bus 120 are accomplished using a handshaking protocol involving a request signal and an acknowledge signal between the particular components that are communicating on the bus. For example, when SCSI chip 114 and FIFO circuits 116b-116d receive data from bus 102a, and are ready to transfer the first word of data to buffer 152, SCSI chip 114 and FIFO circuits 116b-ll6d each place their respective byte (plus parity bit) on bus 120, and SCSI chip 114 generates a request signal on line 151 at the same time. SCSI chip 114 and FIFO circuits H6b-ll6d then wait for an acknowledge signal from buffer 152 before placing the next data word on bus 120. Line 151 supplies the request signal to buffer 152 via driver 153.
  • Buffer 152 responds to the request signal by reading the data on bus 120 and supplying an acknowledge signal on line 124 when the data on bus 120 has been read.
  • the acknowledge signal is provided to SCSI chip 114 via multiplexer 118, and to FIFO circuits 116b-116d via drivers 142.
  • the data transfer process repeats itself until all data words received by SCSI chip 114 and FIFO circuits 116b-116d from bus 102a have been transferred (SCSI chip 114 and FIFO circuits 116b-116d may continue to receive data from bus 102a while they are transferring previously received data on bus 120) , and - 25 -
  • SCSI chip 114 has a count register that is loaded by microprocessor 150 prior to the data transfer with the transfer size. The count is decremented by SCSI chip 114 with each data transfer on bus 102. When the transfer is complete SCSI chip 114 interrupts microprocessor 150 to notify the processor that the data transfer is done.
  • SCSI chip 114 When data is to be transferred from buffer 152 to bus 102a, SCSI chip 114 is again provided with a count of the data to be transferred and asserts a request signal on line 151, in response to which buffer 152 puts a word of data on bus 120 and supplies an acknowledge signal to SCSI chip 114 and FIFO circuits 116b-116d. This exchange repeats itself as previously described until the counter in SCSI chip 114 reaches zero and when the transfer is complete SCSI chip 114 interrupts microprocessor 150.
  • bus 120 is used to transfer messages and commands. Such transfers are controlled by microsequencer 112 under the direction of microprocessor 150. For example, when interface unit 110 receives a message from bus 102a, SCSI chip 114 interrupts microsequencer 112. Microsequencer 112 analyzes interrupt status register 114c to determine the nature of the interrupt, and transfers the first message byte received by SCSI chip 114 to FIFO 116a for microprocessor 150 to read. This message byte, which is the lowest byte of the first word of the message, is the Message Code, as shown in FIG. 5.
  • microsequencer 112 transfers the additional byte(s) from SCSI chip 114 to FIFO circuit 116a. If the message indicates that the interface unit is to receive a command, microsequencer 112 in turn instructs SCSI chip 114 to enter the COMMAND phase to receive the command.
  • Command bytes received by SCSI chip 114 (the lowest byte of each command word) are placed into FIFO circuit 116a by microsequencer 112 for microprocessor 150 to read.
  • the upper three bytes of each command word are received respectively by FIFO circuits 116b- 116d, are read directly from those FIFO circuits by microprocessor 150.
  • a command preferably includes a checksum byte which is checked by the microprocessor; if valid, the microprocessor instructs the microsequencer to disconnect interface unit from bus 102a while array controller 108 performs the command.
  • microprocessor 150 places the bytes of the message into FIFO circuits 116a-116d.
  • Microprocessor 150 then instructs microsequencer 112 to transfer the message, in response to which microsequencer 112 reads the first message byte in FIFO 116a and supplies the byte to the data terminals 114d of SCSI chip 114.
  • SCSI chip 114 transfers this byte to data lines D(00)-(07) of bus 102a and, at the same time, FIFO circuits 116b-116d place the message bytes stored therein onto data lines D(08)-D(31). Parity bits are also transferred to the bus.
  • Multiplexer 118 is used by micro- - 27 -
  • sequencer 112 to select data acknowledge signals (DAck) from data buffer 152 connected to bus 120 (via line 124) or from microsequencer 112 (via line 126) .
  • the output of multiplexer 118 is provided to a DAck input terminal of SCSI chip 114, and as previously discussed, indicates that data has been received by either buffer 152 or microsequencer 112.
  • a second multiplexer 119 multiplexes a "Direction" signal from microsequencer 112, another "Direction” signal from microprocessor 150, and a third "Direction” signal from buffer 152.
  • the output of multiplexer 119 is supplied to a "Direction" input terminal of SCSI chip 114.
  • the Direction signal determines the direction of information flow on bus 120.
  • a third multiplexer 121 multiplexes a first "FIFO Strobe" signal from microsequencer 112 and a second "FIFO Strobe” signal from microprocessor 150.
  • the output of multiplexer 121 is a write/read clock pulse to FIFO circuit 116a, which serves to control the flow of information between FIFO circuit 116a and bus 120.
  • Communication signals between microprocessor 150 and microsequencer 112 are channeled through two registers, "branch” register 128 and "control" register 130.
  • the microprocessor sets a branch address in branch register 128 to instruct microsequencer 112 to perform a function.
  • Microprocessor 150 then sets a "GO" bit in control register 130 via line 132 to cause the microsequencer to execute the function indicated by branch register 128.
  • microsequencer 112 responds to interrupts generated by SCSI chip 114.
  • Microsequencer 112 typically sets a status code in "status" register 134 after servicing an interrupt.
  • Status register 134 may be an eight bit register, and is used to inform microprocessor 150 of the status of - 28 -
  • Microsequencer 112 uses "select" registers 136 or 138 to communicate with various components of the hardware of interface unit 110. For example, microsequencer 112 selects SCSI chip 114 and addresses the registers of SCSI chip 114 via select register 136. Microsequencer 112 uses select register 138 to: reset FIFO circuits 116a-d (as discussed below) ; select control register 130, branch register 128, or "Data In" register 140; and to control multiplexer 118 and drivers 142 for enabling/disabling transfers between SCSI chip 114/FIFOs 116b-d and buffer 152.
  • Tables 4 and 5 show illustrative functions for certain bits of select registers 136 and 138. A more detailed assignment of the bits of registers 136 and 138 and other registers of interface unit 110 can be found in the definition portion of the microsequencer program at the end of this application.
  • Bit 0 SCSI register address (lsb) - 29 -
  • microsequencer 112 When microsequencer 112 is controlling interface unit 110 (e.g., during MESSAGE and COMMAND phases), each of the drivers 142, which enable FIFO circuits H6b-d to receive acknowledge signals from buffer 152, are disabled. When the microsequencer is not in control of interface unit 110 (e.g., during the DATA phase) , microsequencer 112 resets a bit (Bit 3) in select register 138, which enables drivers 142 via line 144. Data can then be transferred between FIFO circuits 116b-d (and SCSI chip 114) and buffer 152.
  • Microprocessor 150 also may read the internal registers of SCSI chip 114 via bus 120 when bit 3 of select register 138 is reset.
  • Data In register 140 and Data Out register 146 each include nine latched drivers (one for each of eight data bits and one parity bit) .
  • Registers 140 and 146 are connected between the "data input” and “data output” pins, respectively, of microsequencer 112, and the lines of data bus 120 which connect SCSI chip 114, and FIFO circuit 116a.
  • An "Enable Constant" signal 149 latches data from microsequencer 112 into register 146.
  • Register 146 is used to load the SCSI chip internal registers (the SCSI register address is selected by setting bits 3-0 of select register 136) .
  • register 140 registers data from SCSI chip 114 for microsequencer 112 to read.
  • Microsequencer 112 can also read data from FIFO 116a via register 140, and can - 30 -
  • microprocessor 150 provides diagnostic information to microprocessor 150 via register 146.
  • interface unit 110 is further illustrated by the following discussion describing the steps of typical "selection" and
  • interface unit 110 incorporated in a data port 106 of an array controller 108 of FIG. 3.
  • interface unit 110 provides particular illustrations of how interface unit 110 can be operated in accordance with the protocol illustrated by FIGS. 6-10.
  • interface unit 110 may be operated differently, and may implement a different protocol, and still operate in accordance with the principles of the present invention.
  • Microprocessor 150 of array controller 108 prepares interface unit 110 to receive commands from device controller 104 via SCSI-2E bus 102 as follows.
  • the microprocessor sends an instruction signal, via LBE 120, to branch register 128 to stage an instruction for microsequencer 112.
  • the microprocessor sets the "GO" bit of control register 130, which causes microsequencer 112 to begin executing the instruction staged in register 128.
  • Microsequencer 112 then tells SCSI chip 114 (via Data Out buffer 146 and bus 120) to wait for selection by device controller 104.
  • Microsequencer 112 does this by putting the register address of command register 114a of SCSI chip 114 in register 136, setting the SCSI chip select bit of register 136, and setting a bit in command register
  • Microsequencer may also initialize other registers in SCSI chip 114 at this time.
  • SCSI chip 114 (using the "select" control line of bus 102a) , SCSI chip 114 interrupts microsequencer 112, which analyzes the interrupt by reading interrupt status register 114c, and causes SCSI chip 114 to respond to the selection by requesting a message or command (depending on whether selection is with or without attention) . Assuming a message is requested, SCSI chip 114 receives the first byte of a message word ("disconnect") and stores the word in its internal FIFO circuit. At the same time, the other three bytes of the message word are automatically stored in FIFO circuits ll6b-d, respectively. SCSI chip 114 then interrupts microsequencer 112.
  • Microsequencer 112 moves the first byte of the message word from SCSI chip 114 to FIFO circuit 116a. Once all bytes of the message words are stored in FIFO circuits 116a-d, and if a command is to follow, microsequencer 112 then loads the command register 114a of SCSI chip 114 with a request command instruction, and SCSI chip 114 requests a command by asserting appropriate control signals on bus 102a. When a command has been received, microsequencer 112 loads a status code in status register 134 and interrupts the microprocessor.
  • the microprocessor then reads status register 134 to determine what type of information is in FIFO circuits 116a-d (e.g., selection, error data, etc.). The microprocessor then reads the message/command from FIFO circuits 116a-d, and checks its validity. Microprocessor 150 loads a "disconnect" message into FIFO circuits H6a-d. The microprocessor instructs - 32 -
  • microsequencer 112 to disconnect by loading a corresponding instruction in branch register 128 and setting the "GO" bit in control register 130.
  • Microsequencer 112 transfers the lowest byte of the "disconnect" message from FIFO circuit 116a to SCSI chip 114, and then instructs SCSI chip 114 to send the "disconnect” message to device controller 104.
  • SCSI chip 114 and FIFO circuits 116b-116d send the "disconnect” message to device controller 104a and then interrupts microsequencer 112 to tell the microsequencer that it has so informed device controller 104.
  • Microsequencer 112 then loads the command register 114a of SCSI chip 114 with the disconnect instruction, and SCSI chip 114 disconnects. SCSI chip 114 again interrupts microsequencer 112, which, in turn, interrupts the microprocessor. At this point, the microprocessor has the information sent by device controller 104a and can perform the necessary steps without tying up device controller 104 or the SCSI-2E bus (because it is no longer connected) . At the same time, it has instructed the microsequencer 112 to wait for another selection by device controller 104a. When the microprocessor has finished processing the command or information received from device controller 104, it will initiate "reselection" (discussed below) .
  • microprocessor 112 copies the first byte of data from FIFO circuit 116a to the - 33 -
  • this invention provides a single bus interface cable capable of simultaneously transferring at least 32 bits of data, and an interface unit for interfacing the cable with a device.
  • test_06 - t5 " Xfer Data " test_05 - t4 " Xfer Data " test_04 - t3 test_03 - t2 test_02 - tl test_01 - tO
  • bus_rst - t7 cmd_err - t6 discon — t5 bus_srv - t4 func__cmpl - t3 sel_attn - tl selected - tO
  • cmd_cnt_hi 00#h cmd_cnt_lo la#h " 104 Bytes " data_count 50 " 200 Bytes/Words discmsg 03#h flush_fifo 01#h init_id 07#h tgt_mode 05#h timeout 95#h
  • w4_sel sel_rl + s_chp_wrtc + s_cmdreg, continue; wrt + enasel, continue; sel_none + enasel, continue; sel rl + init rl, if (not fail) then call pl (w4_int) ; - 41 -
  • noattn noattn:sel_r2 + fifo_rst + sel_dt_in_dbl, continue; sel_r2 + sel_dt_in_dbl, if (not fail) then call pl (dmpidm) ; off, if (not fail) then call pl (dmpidm) ; off, if (not fail) then call pl (rcvcmd) ; off, if (bus_srv) then goto pl (err_00) ; sel_stat + t_sel_cmd + int29k, if (not fail) then goto pl (start);
  • err_00 sel_stat + t_sel_cmd_bsrv + int29k, if (not fail) then go to pl (start) ;
  • Attnl off, if (not fail) then call pl (rcvcmd); sel_stat + t_sel_msg_cmd + int29k, if (not fail) then goto pl (start) ;
  • r-econ off , cmp t (3f#h) to pl (33#h) ; off, if (eq) then goto pl (reconl) ; off , if (not eq) then goto pl (err_04) ;
  • err_04 sel_stat + t_sel_inv_msg + int29k, if (not fail) then goto pl (start) ;
  • resel2 sel_rl + e_fifo, continue; wrt + f_strb, continue; sel_rl + s_chp_wrtc + s_xferhi, continue; wrt + 00#h, continue; sel_none + 00#h, continue; sel_rl + init_rl, continue; sel_rl + s_chp_wrtc + s_xferlo, continue; wrt + 03#h, continue; sel_none + 03#h, if (not fail) then load pl (03#h) ; sel_rl + init_rl, if (not fail) then call pl (sdmsgl); off, if (not bus_srv) then goto pl (resel4) ; off, if (not fail) then call pl (rcvmgl) ; off, while (cre ⁇ > 0) loop to pl (rc
  • rcv_mg off, if (not fail) then call pl (rcvmg); sel_stat + t_msg_rcvd + int29k, if (not fail) then goto pl (start) ;
  • rcvmr off, if (not fail) then call pl (rcvmg); sel_stat + t_disc_msg + int29k, if (not fail) then goto pl
  • rcvmgl sel_r2 + fifo_rst_dbl, continue; sel_r2 + dis_bl, continue; sel_rl + s_chp_wrtc + s_cmdreg, continue; wrt + rcv_msg, continue; sel_none + rcv_msg, continue; sel rl + init rl, if (not fail) then call pl (w4int) ; off, if (not fail) then load pl (0) ;
  • xfr_rd sel_rl + s_chp + s_fifo + e_fifo, continue; off, continue; rd, continue; rd, continue; sel_r2 + init_r2 + rd_sel + f_strb_sel, continue; sel_r2 + sel_dt_in_dbl + rd_sel, continue; sel_rl + init_rl + rd_sel, if (not fail) then ret;
  • sdmsg sel_rl + e_fifo + wrt_sel, continue; sel_r2 + init_r2 + wrt_sel, continue; sel r2 + sel dt in dbl + wrt sel, continue;
  • xfrmsg sel_rl + init_rl, continue; sel_rl + s_chp_wrtc + s_xferhi, continue; wrt + 00#h, continue; sel_none + 00#h, continue; sel_rl + init_rl, continue;
  • sdmsgl sel_rl + s_chp_wrtc + s_cmdreg, continue; wrt + snddmsg, continue; sel_none + snddmsg, continue; sel_rl + init_rl, if (not fail) then call pl (dma_wt) ; off, if (not fail) then ret;
  • dma rd sel rl + e fifo, continue; sel_r2 + sel_cntl_dbl, continue; rd, continue; dmard: rd, if (s_int) then goto pl (chkdma) ; off, if (not fail) then call pl (onechp) ; off, if (not fail) then ret;
  • dack rd + s_strb, continue; rd + s_strb, continue; rd + s_strb + f_strb, continue; rd, if (not fail) then goto pl (dmard) ;
  • dackl wrt + s_strb, continue; wrt + s strb, continue; wrt, continue; wrt + f_strb, if (not fail) then goto pl (dmawrt) ;
  • test off, if (test_01) then goto pl (testl) ; off, if (test_02) then goto pl (test2); off, if (test_03) then goto pl (test3); off, if (test_04) then goto pl (test4) ; - 55 -
  • testl sel_r2 + sel_cntl_dbl, continue; sel_stat + t_end_of_tstl + int29k, if (not fail) then goto pl (start) ;
  • test2 sel_rl + e_fifo + wrt_sel, continue; sel_r2 + init_r2 + wrt_sel, continue; sel_r2 + sel_dt_in_dbl + wrt_sel, continue;
  • grt_ off, if (tl) then goto pl (grt_6) ; off, if (tO) then goto pl (out_5) ; sel_stat + 04#h + int29k, if (not fail) then goto pl
  • grt_6 off, if (tO) then goto pl (out_7); sel_stat + 06#h + int29k, if (not fail) then goto pl
  • grt_2 off, if (tO) then goto pl (out_3) ; sel_stat + 02#h + int29k, if (not fail) then goto pl
  • grt_8 off, if (t2) then goto pl (grt_c) ; off, if (tl) then goto pl (grt_a) ; off, if (tO) then goto pl (out_9); sel_stat + 08#h + int29k, if (not fail) then goto pl
  • grt_c off, if (tl) then goto pl (grt_e) ; off, if (tO) then goto pl (out_d) ; sel_stat + 0c#h + int29k, if (not fail) then goto pl (test2e); qut_d: sel_stat + Od#h + int29k, if (not fail) then goto pl
  • grt_a off, if (tO) then goto pl (out_b) ; sel_stat + Oa#h + int29k, if (not fail) then goto pl (test2e);
  • grt_e off, if (tO) then goto pl (out_f) sel stat + Oe#h + int29k, if (not fail) then goto pl (test2e) ;
  • test2e off, if (not fail) then goto pl (start);
  • test3 sel_rl + s_chp_wrtc + s_cmdreg, continue; wrt + flush_fifo, continue; sel j none + flush_fifo, continue; sel_none + flush_fifo, continue; wrt + nop, continue; sel_none + nop, continue; - 58 -
  • rd_nxt sel_rl + s_chp + s_fifo + e_fifo, while (creg ⁇ > 0) loop to pl (rd_ntl) ; sel_rl + init_rl, if (not fail) then goto pl (test3e) ;
  • rd_ntl off, continue; rd, continue; rd, continue; rd + f strb, if (not fail) then goto pl (rd_nxt) ;
  • test3e sel_stat + t_end_of_tst3 + int29k, if (not fail) then goto pl (start) ;
  • test4 sel_r2 + sel_cntl_dbl, continue; " Disable Buffer xfer " sel_rl + s_chp_wrtc + s_test, continue; wrt + tgt_mode, continue; sel_none + tgt_mode, continue; sel rl + init rl, continue;
  • d_req wrt, if (not s_dreq) then goto pl (d_req) ; wrt, while (creg ⁇ > 0) loop to pl (d_ack) ; off, if (not fail) then load pl (10#h); off, if (not fail) then goto pl (d_reql) ;
  • d_ack wrt + s_strb, continue; wrt + f strb, if (not fail) then goto pl (d_req) ;
  • test42 wrt, while (creg ⁇ > 0) loop to pl (d_reql) ; sel_rl + s_chp_wrtc + s_cmdreg, continue; wrt + rst_chp, continue; sel_none + rst_chp, continue; sel_none + rst_chp, continue; wrt + nop, continue; sel_none + nop, continue; sel_stat + t_end_of_tst4 + int29k, if (not fail) then goto pl (start) ;
  • test5 sel_r2 + sel_cntl + diag_bl, if (not fail) then load pl
  • dackd s_strb, if (bl_dack) then goto pl (dackd) ;
  • dacku off, if (not bl dack) then goto pl (dacku) ; off, while (creg ⁇ > 0) loop to pl (dackd); sel_stat + t_end_of_tst6 + int29k, if (not fail) then goto pl (start) ;

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Bus Control (AREA)

Abstract

Système d'interface amélioré basé en partie sur le standard SCSI.Un bus de données à un seul câble transfère simultanément plusieurs octets d'information entre deux dispositifs. Le système d'interface transfère des ordres, des messages, des informations d'états ou des données multi-octets en un seul transfert parallèle. Un micro-séquenceur permet le transfert de données par l'interface sans nécessiter d'attention particulière à partir d'un processeur dans un dispositif impliqué dans le transfert.
PCT/US1991/001255 1990-04-06 1991-02-27 Procede et apprareil d'obtention d'une interface amelioree pour systeme informatique WO1991015825A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US50578090A 1990-04-06 1990-04-06
US505,780 1990-04-06

Publications (1)

Publication Number Publication Date
WO1991015825A1 true WO1991015825A1 (fr) 1991-10-17

Family

ID=24011790

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1991/001255 WO1991015825A1 (fr) 1990-04-06 1991-02-27 Procede et apprareil d'obtention d'une interface amelioree pour systeme informatique

Country Status (3)

Country Link
EP (1) EP0524203A1 (fr)
AU (1) AU7478291A (fr)
WO (1) WO1991015825A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0537072A1 (fr) * 1991-10-09 1993-04-14 Lg Electronics Inc. Procédé et système pour réaliser l'interface entre un PC et des appareils CD-Rom

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Electronics & Wireless Worls, vol. 95, no. 1643, September 1989, (Sutton, Surrey, GB), G. Humphrey: "SCSI - Small computer system interface", pages 884-888 *
IBM Technical Disclosure Bulletin, vol. 25, no. 10, March 1983, IBM Corp., L. Andrews et al.: "Card-to-card communication", pages 5236-5239 *
Wescon, vol. 29, November 1985, (New York, US), D. Karr: "designing to full SCSI implementation for performance improvement", pages 1-5 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0537072A1 (fr) * 1991-10-09 1993-04-14 Lg Electronics Inc. Procédé et système pour réaliser l'interface entre un PC et des appareils CD-Rom

Also Published As

Publication number Publication date
AU7478291A (en) 1991-10-30
EP0524203A1 (fr) 1993-01-27

Similar Documents

Publication Publication Date Title
US5233692A (en) Enhanced interface permitting multiple-byte parallel transfers of control information and data on a small computer system interface (SCSI) communication bus and a mass storage system incorporating the enhanced interface
US4961140A (en) Apparatus and method for extending a parallel synchronous data and message bus
US5968143A (en) Information handling system for transfer of command blocks to a local processing side without local processor intervention
EP0185676B1 (fr) Processeur de donnees avec dimensionnement dynamique du bus
US5940866A (en) Information handling system having a local address queue for local storage of command blocks transferred from a host processing side
US4106092A (en) Interface system providing interfaces to central processing unit and modular processor-controllers for an input-output subsystem
US4860244A (en) Buffer system for input/output portion of digital data processing system
US4562533A (en) Data communications system to system adapter
US4381542A (en) System for interrupt arbitration
EP0341710A2 (fr) Séquence atomique pour transitions de phases
EP0432978A2 (fr) Appareil pour le conditionnement de l'arbitrage de priorité dans l'adressage tamponné direct de mémoire
US5925118A (en) Methods and architectures for overlapped read and write operations
US4827409A (en) High speed interconnect unit for digital data processing system
WO1991011767A1 (fr) Controleur d'acces direct a une memoire pour le transfert flexible en continu a haute vitesse de donnees d'une source a une destination
WO1983002021A1 (fr) Circuit d'interface pour controleur de sous-systeme
US20080144649A1 (en) Apparatus for multiplexing signals through I/O pins
EP0293860A2 (fr) Module de commande périphérique et interface pour adaptateur
EP0575042A1 (fr) Partage physique de bus logiquement continu
EP0121364A2 (fr) Système de lecture automatique pour circuit de commande périphérique
US4534013A (en) Automatic write system for peripheral-controller
WO1991015825A1 (fr) Procede et apprareil d'obtention d'une interface amelioree pour systeme informatique
EP0055741B1 (fr) Systeme d'entree/sortie et procede de communication pour des dispositifs peripheriques dans un systeme de traitement de donnees
KR100499350B1 (ko) 다중및단일채널을갖는다이렉트메모리억세스의데이터전송방법
EP0278263B1 (fr) Appareil de commande d'accès direct de mémoire à bus multiple
EP0055763B1 (fr) Processeur d'entree/sortie et procede de communication pour un systeme de traitement de donnees

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AT AU BB BG BR CA CH DE DK ES FI GB HU JP KP KR LK LU MC MG MW NL NO PL RO SD SE SU

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE BF BJ CF CG CH CM DE DK ES FR GA GB GR IT LU ML MR NL SE SN TD TG

WWE Wipo information: entry into national phase

Ref document number: 1991906225

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1991906225

Country of ref document: EP

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

NENP Non-entry into the national phase

Ref country code: CA

WWW Wipo information: withdrawn in national office

Ref document number: 1991906225

Country of ref document: EP