Data Communications Preprocessor

The microprogrammable data communications preprocessor of the present invention cooperates in a data communications system employing a microprogrammable central processing unit. In the preferred embodiment, the central processor is of the type disclosed in copending application Ser. No. 825,569 for "Polymorphic Programmable Units Employing Sub-Instruction Sets" filed May 19, 1969 in the name of U. Faber et al. Further, the preferred embodiment utilizes a central processing unit having a port select unit (PSU) of the type disclosed in copending application Ser. No. 392,116 for "Dynamic System Firmware" filed Aug. 27, 1973 in the name of K. Kistler et al.

The data communications preprocessor of the present invention employs a microprogrammable serial byte microprocessor of the type disclosed in U.S. Pat. No. 3,878,514 issued to U. Faber on Apr. 15, 1975 for an "LSI Programmable Processor" filed Nov. 20, 1972. More specifically, the preferred embodiment of the present invention uses an improved version of the Faber processor of the type disclosed in copending application Ser. No. 455,060, now U.S. Pat. 3,972,024, for "An Improved Programmable Microprocessor" filed Mar. 27, 1974 in the name of F. Schroeder et al.

The above-cited cross references provide much useful technical and background information which is not unnecessarily repeated in the present specification. Therefore, a review of these cross references will aid in an understanding of the present invention.

The present invention relates to a microprogrammable preprocessor for interfacing between a multiplicity of data communications lines and a central microprogrammable processor by tending to the detailed line procedures and requiring central processor intervention on a message basis only.

A data processing communications system generally includes a central processing unit (CPU) and a plurality of data communications channels including direct connect lines, MODEMS, and the like. As the system becomes more complex, inefficiencies increase as the CPU data processing functions are interrupted more frequently for line procedures and data communications details. The addition to the system of a microprocessor to handle line procedures and details may relieve the CPU of these burdens but may otherwise limit the data communications aspects of the system to the necessarily limited capabilities and throughout speed of the microprocessor.

Therefore it is an object of the present invention to provide a high speed, flexible, data communications preprocessor.

It is a further object of the present invention to provide a microprogrammable data communications preprocessor for use with a microprogrammable CPU in a data communications system.

A data communications preprocessor constructed in accordance with the present invention includes a line adapter associated with each data communications line for interface purposes, a scratch pad memory for storing the data line parameters of each data line and a microprogrammable serial byte microprocessor for exercising detailed control over a plurality of data communications lines while requiring central processor intervention on a per message basis only. Operational speed is enhanced through a direct memory access channel which permits data transfer directly from the preprocessor to the main memory of the central processor, and through the inclusion of automatic operation logic which effectively by-passes the serial byte microprocessor for an automatic transfer of two bytes of data.

The system configuration and operational details given above have been presented in simplified form. Other objects, features, and advantages of the invention will be apparent from the following description of the preferred embodiment of the invention and the accompanying drawings thereof.

FIG. 1 is a block diagram of a data communications system using the data communications preprocessor of the present invention;

FIG. 2 is a block diagram of the data communications preprocessor of the present invention;

FIG. 3 is a diagram of the scratch pad memory used in the preprocessor of FIG. 2;

FIG. 4 is a more detailed diagram of the preprocessor of FIG. 2;

FIG. 5 is an illustration of the word format of transmit status, receive status, line frequency select, and adapter descriptor control words;

FIG. 6 is a representation of the data paths and logic functions of the preprocessor of FIG. 4;

FIG. 7 is a representation of an adapter I/O register used in the preprocessor of FIG. 4;

FIg. 8 is a representation of the status words and bit isolation unit used in the preprocessor of FIG. 4;.

FIG. 9 is an illustration of the scratch pad memory addressing and control instructions used for the scratch pad memory of FIG. 3;

FIG. 10 is an illustration of the adapter descriptor register which serves to define the functional operation of the adapter; and

FIG. 11 is an illustration of the data set description register which serves to define the functional operation of the data set.

Following is a listing of abbreviations frequently used in the detailed description of the preferred embodiment of the invention and related cross referenced.

With reference to FIG. 1, the data communication preprocessor, hereinafter called the DCPP 11, of the present invention provides the data communications interface between the central processing unit CPU 13 and multiple data communications lines 15. The data communications interface is with either standard data sets 17 or direct connect lines, and include full duplex, half duplex, synchronous, and asynchronous applications. Each data communication line 15 serviced requires an individual line adapter 19. Each line adaptor 19 contains the transmit and receive logic necessary for transferring data to and from the data communication line 15, and the logic required to control the communication interface, both for normal operating control and status interrogation.

The DCPP 11 incorporates its own microprocessing unit, hereinafter called the MPU 21, which permits the DCPP 11 to operate in a semi-autonomous mode relative to the CPU 13. The MPY performs many of the functions normally required of a control processor, thus relieving the CPU 13 of much of the burden associated with detailed control of the DCPP 11. The MPU 21 is loadable with commands from the CPU 13 via a DCPP 11 microprogram memory, hereinafter called the MPM 23. The MPU 21 utilizes these commands to generate the internal firmware set used for DCPP 11 function control. Thus, the functional characteristics of the DCPP 11 are determined by the program loaded from the CPU 13. A direct access channel 25 provides for information transfer directly between the direct memory access logic, hereinafter called the DMA 26, in the DCPP 11, and the main memory 28 of the CPU 13. This direct transfer further reduces the need for the CPU 13 intervention.

The DCPP 11 interface 27 with the CPU 13 is via a port select unit, hereinafter referred to as the PSU 29, contained within the CPU 13. The PSU 29 controls and synchronizes the interface 27 between the CPU 13 and DCPP 11. The DCPP 11 acts upon control words from the CPU 13, performs the specified operation, and upon completion of the operation, generates and returns a status word containing operation status and/or error status information.

The primary features of the DCPP 11 are itemized as follows:

1. Multiple data communication lines are serviced concurrently requiring central processor unit CPU 13 attention only a per message basis.

2. Line procedures are handled without CPU 13 intervention. Line procedures are defined by the DCPP 11 microprogram thereby providing application flexibility.

3. Direct memory access enables the DCPP 11 to handle message transfers without intervening the CPU 13 attention; the CPU 13 is notified upon message transfer completion. Buffer area is assigned in the main memory 28 for each line. Buffer size and location is specified by the CPU 13.

4. Upon exceeding a specified number of retries for a message, or upon line failure, the DCPP 11 status word is up-dated accordingly and the CPU 13 is interrupted. Failure on one line will not effect transactions on another line.

5. The data communications line adaptor 19 is designed to interface with the standard data sets 17 or direct connect line including full or hald duplex, synchronous or asynchronous applications. Standard baud rates from 75 to 19.2K baud are programmably selectable. The data communication line adaptor 19 can be preset for most line requirements by programmatic control.

FIG. 2 is a functional diagram of the DCPP 11. The diagram shows the following functional blocks:

The CPU 13 interface logic 31, the loader 33, the MPM 23, the MPU 21, the function control logic 35, the microprocessor support logic 37, the direct memory access logic 26, and the universal line adaptor 19. Each of these functions is briefly described below.

The CPU interface logic 31 allows the CPU 13 to load or modify the contents of the memory MPM 23. This provides the CPU 13 with complete control over the program contents of the MPM 23 memory. The CPU interface logic 31 permits the CPU 13 to direct DCPP 11 operations such as transmit data, receive message, disconnect line, etc. The CPU interface logic 31 also allows the DCPP 11 via status interrupt to report status to the CPU 13.

The loader logic 33 is used only when the CPU 13 loads or modifies the DCPP 11 microprogram memory MPM 23. The CPU 13 selects load mode via a control word register. In load mode, random or sequential loading of the microprogram memory MPM 23 is possible.

The MPM 23 stores program information for the microprocessor MPU 21. The storage capacity is expandable in the preferred embodiment from 2k to 6k words, in 2K word increments. Each word consists of 13 bits (12 data bits plus one parity bit).

The microprocessor MPU 21 controls the operations of the DCPP 11. It initiates and executes line message transactions and special functions as directed by the CPU 13. When a plurality of lines are active, the microprocessor MPU 21 will sequentially interrogate each line adaptor 19 and provide service on an as-requested basis. In the preferred embodiment the microprocessor MPU 21 contains an eight bit wide arithmetic unit and performs instructions at a one megaHertz rate.

The function control logic 35 allows the microprocessor MPU 21 to address and transfer information to and from other functional units including up to four line adaptors 19.

In the microprocessor support logic 37, a scratch pad memory 39 provides temporary data storage for the microprocessor MPU 21. The scratch pad memory 39 is used to store information for each data communication line 15 such as buffer location in current address, line procedure details, job description, current status, etc. Thus, the scratch pad memory 39 is used by the microprocessor MPU 21 to keep track of what the DCPP 11 is doing. In the preferred embodiment the scratch pad memory 39 is expandable as the number of data communication lines 15 is increased. A storage capacity of up to 128 bytes per data communication line 15 is provided. The microprocessor support logic 37 also includes the automatic operation logic 43.

The direct memory access (DMA) logic 26 provides the DCPP 11 with a direct access path to the main memory 28. The main memory 28 is used for message input and output buffering for each data communication line 15. In addition, the main memory 28 is used to store control and data descriptors for each data communication line 15. The main memory 28 which is accessed both by the CPU 13 and the DCPP 11 is used as the transfer medium for the exchange of data and control of result descriptors.

The data communication line adaptors 19 provides the interface between the DCPP 11 and the data sets 17 or direct connect lines. An individual line adaptor 19 is required for each data communication line. It contains transmit and receive logic 45 capable of operating at selectable baud rates, synchronous or asynchronous and full or half duplex. The microprocessor MPU 21 controls the functional operation of the data set 17 and line adaptor 19 by means of control descriptors. The microprocessor 21 interrogates the line adaptor 19 status and the data set 17 status by reading the respective status register in the included data set control and status logic 47 and the adaptor control line status logic 49.

The DCPP 11 with its own microprocessing unit 21 is a programmable multiline control device dependent port (DDP). In combination with a CPU 13 an overall data communications system consists of two processors, each able to operate independently by containing an interpreter and program within its own memory. The first processor, the CPU 13, has control of the usual peripherals whereas the second processor, the DCPP 11, controls up to four data communication lines 15. Since the processor DCPP 11 is microprogrammable a program language, NDL (network definition language) is provided in the preferred embodiment to handle the data communication aspects. In order that data received by the DCPP 11 can be, for example, printed, there has to be a means of communications between the DCPP 11 and the CPU 13. When a message is received by the DCPP 11 the NDL program will cause it to be stored in an area of the DCPP 11 to which the CPU 13 has access. This may be termed a receive buffer. Likewise, the CPU 13 may desire to read a keyboard message and sent it to a terminal. Therefore, as well as the receive buffer there must also be a transmit buffer. In order that data programs being run may be related in some way, one processor or the other must have ultimate control. The CPU 13 is given this ultimate control. The program it is running is the main program which will, for example, contain data it wishes to send to a data communication line 15 from some peripheral and decide which terminal to send it to. Thus the source language has instructions such as "transmit message to . " These instructions are decoded by the data communications controller portion of the source language and formatted to suit NDL. This is stored in an area of memory considered the operation buffer. Now it can be clearly seen that the NDL program and the DCPP 11 can investigate an instruction in the operation buffer. Dependent on this instruction it will branch through a specific routine. The NDL instructions within the subroutine will be broken down into microstrings by the NDL interpreter and performed by the hardware.

The NDL program is specific to a given network of terminals. It will provide all the necessary controls for that network, determining such things as: is the requested terminal valid, is it already busy, how often should it be serviced, what is its line frequency, what line is it on, etc. Thus the main source language program needs only to say "send this message to terminal " and NDL along with DCPP 11 does the rest.

As part of the pairing-up procedure a warm start is performed wherein the NDL is read from a disc and stored into the DCPP 11 memory. This is done by the usual peripheral control method of control words and data words. Once loaded the DCPP 11 is inhibited from operation until suh time as a program is initiated which requires data communication.

When a data communication program is called from disc into the main memory 28, the first part of the load operation loads any soft controllers, followed by a special instruction set of source level data communication instructions.

In the preferred embodiment, a set of NDL tables are produced containing information about software to be used (size-location number of entries - size of transmission, etc.), the lines in use (speed-number of stations - delays, etc.), the stations in use (description of terminal - type of parity-frequency-buffer size, etc.), and many other parameters used for operation. Each table also references the location of the next in chain, but this reference is calculated at the NDL program load time.

Once loaded, the NDL program will start automatically. It releases the DCPP 11 for operation and selects from the NDL tables the various words frequently used and stores information relating thereto in the internal scratch pad memory 39 for fast access. The information stored contains the reference addresses to the NDL tables, the start address of a microstring to be returned to, the address of the CPU 13 memory of the next NDL instruction, etc. Once the initial process is completed the NDL program goes into idle where it waits for further instructions from the main program.

A very important feature of the DCPP 11 is the scratch pad memory 39. With reference to FIG. 3, the scratch pad memory 39 contains in the preferred embodiment 512 one-byte word locations. Since there can be up to four data communication lines 15, and since the memory is divided into four blocks of 128 bytes with one block allocated per line, part of the scratch pad memory 39 addressing is devoted to selecting the direct block of 128 bytes. Each block of 128 bytes is subdivided by page addressing logic 50 into four pages of 32 bytes each. Within a selected page, byte addressing logic 52 selects a specific byte. As will be detailed hereinafter, bits of four bytes in each page may be isolated and selected by further address logic 54. In the preferred embodiment, two pages of scratch pad memory 39 (64 characters or bytes) are required to handle a half duplex communications line and four pages to handle a full duplex communications line.

In operation, the DCPP 11 fetches an NDL instruction from the main memory 28 and decodes it into OP code. Performance of the OP code may require data to be transferred to or from main memory 28. This is done by stealing. A status word is used to indicate to the CPU 13 a microprogram memory 23 or a scratch pad memory 39 error in the DCPP 11. If this condition occurs a status interrupt is sent to the CPU 13 and the MPU 21 is inhibited from running.

In the initiation of information transfer between the CPU 13 and the DCPP 11, the port select unit 29 decodes the three bit groups from the CPU 13 to completely define the operation that is to take place. The command field (nanobits 51 through 54) establishes whether the operation is to be a device read or a device write. The four least significant bits contain the specific device write, and the most significnt bits in conjunction with the command type distinguishes between control, data, and status words.

The interface between a microprogrammable processor and a microprogrammable microprocessor is not characterized by a complex array of hardwired logic interconnections but rather by the control, data, and status words that the two processors use to communicate with and to control each other. Four basic word formats are used to tranfer control, data and status information between the main processor CPU 13 and the preprocessor DCPP 11. Each word format comprises 16 bits with the first bit designated as the most significant bit (MSB) and the sixteenth bit as the least significant bit (LSB). The four basic word formats are a control word (DW.multidot.INST), a write data word (DW.multidot.INST), a read data word (DR.multidot.INST/), and a status word (DR.multidot.INST). The significant bits of the four basic word formats are detailed below.

DW.multidot.INST allows the CPU 13 to perform the following:

Bit 1: Enter or exit load mode. Load mode enables the CPU (via DW.multidot.INST/) to load or verify the contents of the MPM 23 or to have the MPU 21 operate using the MIR bus 51 as an instruction source.

Bit 2: Clear MPU 21 MPCR. Used only in conjunction with enter or exit load mode. Clear MPU 21 MPCR provides the capability to start loading or execution at either MPM 23 location zero or the current address.

Bit 3: Clear DCPP 11. Sets or resets the clear DCPP 11 flip-flop. The clear DCPP 11 if set will hold the DCPP 11 in a cleared state. Once set it will remain set until a command is received to reset it.

Bit 4: MTR Override. When set, inhibits SPM 39 or MPU 21 parity errors from shutting off run.

Bit 15: Execute mode. Used by MTR to disable writing into MAR 53 and MDR 55 from the MIR but 51 with a data write (DW.multidot.INST/). Also, disables MTR request to steal.

Bit 16: MTR Mode.

Used by MTR for the following:

1. Disables MAR select.

2. Enables parity error while in load mode.

3. Complements MPM 23 parity checkergenerator.

4. Gates ASYNCLKO to DINT.

5. complements SPM parity generator.

A DR.multidot.BEX.multidot.INST/ will transfer the contents (16 bits) of the MDR 55 into the B-register of the CPU 13.

A DW.multidot.INST/ will transfer 16 bits of MIR data into MAR 53 and MDR 55. If in load mode, the following functions are also enabled:

1. If MIR01 and 02 are both zero than MIR06-16 (12 bits will be loaded into the MPM 23 at the address defined by the MPU 21 MPCR and then the MPU 21 MPCR will be incremented.

2. IF MIR01 (EXECUTE INST) is set then MPU 21 will execute the instruction (12 bits) contained in MIR05-16, thus the CPU 13 can single-step instructions to MPU 21 and/or DCPP 11.

3. if MIR01/ and MIR02 (MPU 21 to MDR 55) is set then a copy of the instruction in MPM 23 location per MPU 21 MPCR is transferred into the MDR 55 and MPU 21 MPCR is subsequently incremented. This feature is provided to allow the CPU 13 to read and check the contents of the MPM 23 after loading.

4. If MIR03/ (MTR STEAL) is set, the CPU 13 will execute an MRT request to steal. This allows CPU 13 to execute a cycle steal independent of MPU 21. The MTR data word (DW.multidot.INST/) is defined as follows:

this mode enables the CPU 13 to generate a steal cycle from a data write. The address is determined by the previous data loaded into MAR 53.

5. the MTR parity bit is used by the MIR to bring the parity bit to the MDR 55 and MAR 53 registers for testing.

The following listing describes functions of various combinations of control and data words.

In addition to device address, the status word contains 4 bits placed on the external bus 57 (EXT1 -- 16):

Ext01: execption bit: This bit is set if either EXT02 or EXT03 is set. This bit is reset when a status read DR.multidot.INST) is executed. It is not reset with an enable status (ENST/).

Ext02: mpm23 parity error: This bit is set by a MPU 21 parity error and is reset by a status read or clear DCPP 11. When MPM 23 parity error is set, MPU 21 clock will be stopped.

Ext03: spm 39 parity error is set by an SPM 39 parity error and is reset by a clear DCPP 11. When SPM 39 parity is set, the MPU 21 clock will be stopped.

Ext04: run indicate status of DCPP 11. If set, indicates that DCPP 11 is executing instructions. If reset, indicates that DCPP 11 is stopped. This bit is used by MIR only.

Status interrupt: this level is set by an MPM 23 or SPM 39 parity error or by firmware control of MPU 21. It is reset upon a status read or clear DCPP 11. If status interrupt is set and the exception bit is not set a "Soft Interrupt" is indicated (normal completion of most operations).

The DCPP 11 includes the following logic (see FIG. 4):

A. function control register (FCR) 35 -- controls data path and functional operations within the DCPP 11.

B. adapter I/O register 61 -- used for exchange of data between MPU 21 and the data comm line adapters 19.

C. memory address register (MAR) 53 -- used to address memory for direct memory access (DMA) operations.

D. memory data register (MDR) 55 -- used for data storage for DMA read or write operations.

E. bit isolation unit (BIU) 63 -- used to select a particular bit of status or SPM information to be applied to the external interrupt input of MPU 21.

F. scratch pad memory (SPM) 39 -- used by MPU 21 for the storage of control information.

G. automatic operation logic -- used for the automatic transfer of 16 bits of information between SPM 39 and working registers. In some cases an index value may be added during the transfer.

H. adapter 19 -- multiple data comm line adapters may be contained in a single DCPP 11. The adapter 19 contains registers and logic as required to interface with data sets 17.

This register is loaded by an MPU 21 DEVO instruction. It controls the data path and logic functions of the DCPP 11 as shown in FIG. 5 and FIG. 6.

A very important feature of the present invention is the capability to automatically transfer two bytes of SPM data not using MPU 21. The starting byte address of the SPM 39 data is defined by a preceding MPU 21 DEVI instruction and must be an even-numbered address. The automatic transfer may include the addition of an index to the SPM 39 data. The index of one byte is contained in the adapter I/O register 61 and is added if the "index" bit is set. The results of the automatic function will be transferred to the register specified by the "destination" bits; also, if the "SPM updata" bit is set, the results will be stored into SPM 39 in place of the original data. When MAR 53 is the selected destination, a DMA operation to main memory 28 is initiated by requesting "steal" upon completion of the automatic transfer. Whether the memory operation is a read or a write is defined by the "DMA write" bit. The clock to MPU 21 will be stopped for the duration of the automatic operation and if a "DMA read" is initiated, the MPU 21 clock will not be restarted until the completion of the memory read operation. It is possible for an automatic operation to result in an overflow. In this case the "auto overflow" flip-flop is set for subsequent testing by MPU 21. The "overflow" flip-flop will not reset until the next "AUTO OP" is initiated. The function code calling forth an automatic operation is 100 in bits MD5-MD8 of a DEVO instruction. Functional descriptions of the automatic operation codes are given below.

The following controls are used for adapter 19 receive and transmit data exchange and sync detection, break, and second stop bit control.

Another important feature of the present invention is the RTS dealy bit, which when sent, allows the transmitter logic to send one more character. This character is inhibited from being transmitted but is used as a delay for turning off the request to send (RTS) line.

In the asynchronous mode, the RTS delay time is one bit time plus the number of first "0" in the RTS delay character that was loaded into the transmitter logic; for example:

In the synchronous mode, it is just the number of "0". Using the above example, the RTS delay would be 5 bit times.

Notes:

(1) At the same time that the RTS delay ends, and in a half duplex mode, the data to the receiver logic is enabled. This allows the adapter 19 to flip from the XMIT mode to the RCV mode when the RTS delay time expires.

(2) A jumper is available on each adapter 19 for direct connect modes of operation. This allows the adapter 19 to flip to the RCV mode without waiting for the RTS delay time to expire.

The following word formats are used for testing adapter 19 receive (RCV) status, testing adapter transmit (XMIT) status, loading the adapter control descriptor, or selecting line frequency, in the respective order.

This register, see FIG. 4, is used to exchange data between MPU 21 and the selected data comm line adapter 19. It is loaded by MPU 21 by an OUTO or DEV3 instruction, see FIG. 5. Once loaded, the contents of the adapter I/O register 61 can be transferred to any one of eight destinations by loading the appropriate command into the FCR (Function Control Register) 35 by an MPU 21 DEVO instruction. Receive data is read from the adapter 19 as follows: The FCR 35 is loaded (via a DEVO MPU 21 instruction) to select RCV data, then a one-instruction delay is required to allow access time. An MPU 21 BEXO instruction will now transfer the contents of the RCV logic data buffer to the MPU 21 B-register via the adapter I/O register 61.

Formats of the various data and control words handled by the adapter I/O register 61 are described in detail below with reference to FIG. 7.

Used to select baud rate for each line in asynchronous operations. During synchronous operations the respective line should be set to off. A separate frequency select register is provided for each line. The following listing defines the register content required for each frequency selection.

The BIU 63, see FIG. 4, provides the capability to select one of up to eight bits in a status word and apply it to the EXT input of the MPU 21.

Status word selection and bit selection is accomplished by an MPU 21 DEVO instruction loading the appropriate command into the function control register 35.

If SPM bit selection is desired, this is accomplished by an MPU 21 DEV1 instruction loading the appropriate command into the SPM address and control register 56. In this case the MPU 21 condition test operation (EXT TEST) must immediately follow the DEV1 instruction. The selected SPM bit will be applied to the MPU 21 EXT input for only one instruction period after execution of the DEV1 because EXT will revert back to the previously selected status bit.

The scratch pad memory 39 is designed in the preferred embodiment to be expandable in blocks of 512 .times. 9.

The SPM 39 may be accessed in a character or bit mode operation. Bit mode can access one of 32 bits per page. Character mode can access all of the 32 bytes in a page including the 4 bytes used by bit mode.

SPM 39 organization provides for a maximum of four lines with up to four pages per line. Page addressing is accomplished by the FCR 35 via "line select" and "page select" commands. Line select will define a group of four pages and page select will define one of those four pages. The SPM address and control register is loaded via a DEV1 instruction (see FIG. 6 and FIG. 9) from MPU 21.

CHAR or BIT/: Determines whether access will be character or bit mode.

Character mode will either:

1. Read the specified location into the SPM 39 I/O register which then may be read into MPU 21 via a BEX1 instruction.

2. Write the contents of the SPM I/O register into the specified location in the SPM 39. The SPM I/O register is previously loaded from MPU 21 may an OUT1 or OUT3 instruction.

3. If update is set, the specified address of SMP 39 is in the upper page. If update is reset, the specified address of SPM is defined by SPM page select.

Bit mode will either:

1. Read the specified location into the SPM I/O register and apply the selected bit 0 thru 7 to the EXT input of MPU 21.

2. set or reset the specified bit location in SPM 39.

1. To access RCV data, wait one instruction, after the DEVO command before doing a BEXO to allow access time on the receive logic.

2. SPM bit access (BIU) correlates to character mode locations 28, 29, 30 and 31.

3. An SPM bit read must be tested on the following instruction by MPU 21.

4. MPM 23 parity or SPM 39 parity stop DCPP 11 can be reset with a clear only.

5. After a DMA operation, MDR 55 must be selected before addressing information.

The data communications line adapter 19, see FIG. 4, includes the necessary logic to interface the DCPP 11 to a data communications network either by a data set 17 or by direct connect means. Thus the line adapter 19 provides the output interface of the DCPP 11. One line adapter 19 is required for each output line. In the preferred embodiment, the output line adaptor 19 is programmable to adapt to a variety of data set 17 requirements. In alternate embodiments, it is realized of course, that the line adapter 19 could be hardwired to provide the necessary interface between a specific set of data set 17 requirements and the DCPP 11.

The Line Adapter 19 includes the following functional blocks:

1. The Adapter Descriptor Register 69 which serves to define the functional operation of the Line Adapter 19. The Adapter Descriptor Register 69 is loaded from the Adapter I/O Register 61 and includes the following information bits (see FIG. 10) Receiver Parity Bit 75, Even Bit 77, Asynchronous Bit 79, Character (Y) Bit 81, Size (X) Bit 83, Transmitter Parity Bit 85, Enable Transmitter Logic Bit 87, and Enable Receiver Logic Bit 89.

2. Data Set Descriptor Register 91 which defines the functional operation of the specific data set 17 associated with a specific line adapter 19. The data set descriptor register 91 includes the following information bits (see FIG. 11): Second Stop Bit 93, Standby Rate Bit 95, Rate Bit 97, New Synchronization Bit 99, D Mode Bit 101, DTR Bit 103, ORIG BIT 105 and Request-To-Send Bit 107. The data set descriptor register 91 is loaded from the Adapter I/O Register 61.

3. The transmitter logic 109 which is a synchronous/asynchronous data communications adapter that accepts parallel binary data in the form of characters and serially transmits the data to a modem. Internally generated parity bit and appended control bits for asynchronous mode are transmitted at the same time if applicable.

4. The receiver logic 111 which is a synchronous/asynchronous data communications adapter for receiving serial digital data from a modem or other source. The receiver logic 111 organizes the received data into fixed word lengths corresponding to characters, and transmits these characters to a buffer register from which the character may be accessed in parallel format.

Each line adapter 19 also includes the following status registers:

1. The Receiver Status Logic 113 which provides the capability to select status information relative to receive mode and apply it to the bit isolation unit 63 for interrogation by the MPU 21.

2. The Transmitter Status Logic 115 which provides the capability to select status information relative to the transmit mode and apply it to the bit isolation unit 63 for interrogation by the MPU 21.

3. The Manager Status Logic 117 which provides data to the bit isolation unit 63 for interrogation by the MPU 21. Specifically, three bits of information are provided. The DWI Bit 119 which indicates that a device write instruction has been received from the CPU 13 which wants to communicate with the MPU 21. The DWI Bit 119 is reset by a reset DWI command 121 generated from the function control register 35. An AOV bit 123 indicates that an adder overflow occurred during the last automatic operation. The AOV bit 123 remains set until a subsequent automatic operation is initiated. A TI bit 125 is set when a strobe timer instruction is executed and a one millisecond time has expired. All bits from the Manager Status Logic 117 are presented to the bit isolation unit 63 for interrogation by the MPU 21.

Thus, the function of the line adapter 19 is to interface between the data sets 17 and the DCPP 11. In the preferred embodiment, as above described, the line adapter 19 is, in essence, microprogrammable and therefore universally applicable to a wide variety of output interface requirements.

Although an operative data communications preprocessor has been described it will be appreciated by those skilled in the art that modifications and additional features may be added without departing from the scope of the invention. As an example, provisions may be added for automatic dialing (AUTOCAL) and for cyclical redundancy checking (CRC). Furthermore, while the data communication preprocessing invention has been particularly shown and described with reference to a preferred embodiment thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein.