Real-time operating system and virtual digital signal processor for the control of a digital signal processor

• US 5,283,900 A
• Filed: 10/02/1989
• Issued: 02/01/1994
• Est. Priority Date: 10/02/1989
• Status: Expired due to Fees

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Spectron Microsystems, Inc.
• Inventors: Russo, David A., Wong, David Y., Frankel, Robert E., Lindsay, David M.
• Primary Examiner(s): Lee, Thomas C.
• Assistant Examiner(s): Loomis, John C.
• Application Number:
  US07/415,882
• Time in Patent Office: 1,583 Days
• Field of Search: 364/721, 364/DIG. 1, 364/DIG. 2, 395/200, 395/275, 395/650, 395/700, 395/705, 395/800, 395/500, 340/717
• US Class Current: 718/100
• CPC Class Codes: G06F 17/10 Complex mathematical operations function generation by table look-up G06F1/03; evaluation of elementary functions by calculation G06F7/544
• CPC Class Codes: G06F 17/16 Matrix or vector computation , e.g. matrix-matrix or matrix-vector multiplication, matrix factorization matrix transposition G06F7/78

Abstract

A data processor, such as a digital signal processor, that has augmented memory, I/O and math units for real-time performance of complex functions, is placed under the control of a group of abstract object-oriented modules arranged with an underlying operational nucleus that includes a real-time kernel. The modules are hierarchically configured, with the lowest being an array object type that references memory allocations. A stream object type, based on the arrays, defines channels between application software and data devices. A vector object type, also based on the arrays, establishes structure within allocated blocks and also enables vector math functions to be undertaken by the vector module. Matrix and filter object types utilize the arrays and vectors in sequences controlled by the corresponding matrix and vector modules. The system provides a model of DSP functionality that is hardware independent, and an interface between high level language calls and highly efficient routines executed in assembly language. With this arrangement a large library of math functions is held available for use in real-time operations of complex nature.

Claims

1. An apparatus for managing the operation f a digital signal processor (DSP), comprising:

(a) a digital signal processor (DSP); and

(b) virtual DSP machine means, performed by the DSP, for insulating computer programs executed by the DSP from the DSP'"'"'s specific architecture, wherein functions of the DSP are accessed by the computer programs via a high-level programming interface, wherein the virtual DSP machine means comprises a plurality of modules arranged in a hierarchy, comprising;

(1) an array module for creating arrays of data and allocating memory blocks therefor through the memory management means;

(2) a stream module for processing continuous data streams being input and output by the I/O devices coupled to the DSP, comprising means for binding the I/O devices to the computer programs, means for transferring the data between the arrays and the I/O devices, and means for invoking device-dependent control functions for the I/O devices;

(3) a vector module for creating vectors of data that provide a one dimensional view of the arrays, and for performing vector math operations on the vectors;

(4) a matrix module for creating matrices of data that provide a two dimensional view of the arrays, and for performing matrix math functions on the matrices; and

(5) a filter module for creating finite impulse response filters, infinite impulse response filters, and all-pole filters, and for applying the filters to the vectors.

2. A digital signal processing apparatus, comprising:

(a) a digital signal processor (DSP) having a plurality of memory segments comprising memories of different sizes and speeds coupled thereto;

(b) array module means, performed by the DSP, for creating arrays of data and allocating memory blocks therefor from the memory segments;

(c) stream module means, performed by the DSP, for processing continuous data streams being input and output by I/O devices coupled to the DSP, comprising means for binding the I/O devices to computer programs executed by the DSP, means for transferring the data between the arrays and the I/O devices, and means for invoking device-dependent control functions for the I/O devices;

(d) vector module means, performed by the DSP, for creating vectors of data that provide a one dimensional view of the arrays, and for performing vector math operations on the vectors;

(e) matrix module means, performed by the DSP, for creating matrices of data that provide a two dimensional view of the arrays, and for performing matrix math functions on the matrices; and

(f) filter module means, performed by the DSP, for creating finite impulses response filters, infinite impulse response filters, and all-pole filters, and for applying the filters to the vectors.

3. A digital signal processing apparatus, comprising:

(a) a digital signal processor (DSP);

(b) real-time kernel means, performed by the DSP, for multi-tasking a plurality of concurrently operating computer programs executed by the DSP;

(c) memory management means, coupled to the real-time kernel means and performed by the DSP, for managing memory requests from the computer programs executing on the DSP;

(d) interrupt handling means, coupled to the real-time kernel means and performed by the DSP, for processing interrupts from I/O devices coupled to the DSP with minimum overhead;

(e) device driver support means, coupled to the real-time kernel means, invoked by the computer programs and the interrupt handling means, and performed by the DSP, for handling data input and output by the I/O devices as signaled by the interrupt handler means; and

(f) virtual DSP machine means, coupled to the real-time kernel means and performed by the DSP, for insulating the computer programs executing on the DSP from the DSP'"'"'s specific architecture, wherein functions of the DSP are accessed by the computer programs via a high-level programming interface, and wherein the virtual DSP machine means comprises a plurality of modules arranged in a hierarchy comprising;

(1) an array module for creating arrays of data and allocating memory blocks therefor through the memory management means;

(2) a stream module for processing continuous data streams being input and output by the I/O devices coupled to the DSP, comprising means for binding the I/O devices to the computer programs, means for transferring the data between the arrays and the I/O devices, and means for invoking devicedependent control functions for the I/O devices;

(3) a vector module for creating vectors of data that provide a one dimensional view of the arrays, and for performing vector math operations on the vectors;

(4) a matrix module for creating matrices of data that provide a two dimensional view of the arrays, and for performing matrix math functions on the matrices;

(5) a filter module for creating finite impulse response filters, infinite impulse response filters, and all-pole filters, and for applying the filters to the vectors; and

(6) wherein the modules are selectively executed by the DSP.

4. The invention as set forth in claim 3 above, wherein the virtual DSP machine means further comprise an exception handler module for managing system and application exceptions, for classifying the exceptions as major and minor, and for binding handlers to the exceptions.

5. The invention as set forth in claim 3 above, wherein the stream module further comprises software file device means for inserting and extracting the data in the data streams.

6. The invention as set forth in claim 3 above, wherein the virtual DSP machine means further comprises macro means for retrieving and manipulating arrays, vectors, matrices and filters.

7. The invention as set forth in claim 3 above, wherein the arrays each reference a block of memory in at least one of the memory segments.

8. The invention as set forth in claim 3 above, wherein the virtual DSP machine means further comprises a global definition module for assigning symbolic names to the memory blocks, for assigning symbolic names to the I/O devices coupled to the DSP, and for defining configuration parameters associated with the apparatus.

9. The invention as set forth in claim 3 above, wherein the interrupt handling means further comprises means for managing software interrupts triggered from hardware interrupt service routines.

10. The invention as set forth in claim 3 above, further comprising means, performed by the DSP, for providing software signal generators comprising random noise, sinusoidal wave and constant value generators.

11. The invention as set forth in claim 3 above, wherein the device driver support means further comprises:

(1means for using a first buffer in a first I/O operation, wherein the first buffer is identified by a first pointer;

(2) means for waiting for an I/O request from a computer program executed by the DSP, wherein the request includes a second pointed to a second buffer; and

(3) means for replacing the second pointer in the I/O request with the first pointer, so that the first buffer is returned to the computer program while the second buffer is sued in a second I/O operation.

12. The invention as set forth in claim 11 above, further comprising a plurality of buffers maintained in a queue for a single I/O stream.

13. The invention as set forth in claim 11 above, wherein there are a plurality of I/O devices, a plurality of first and second I/O operations directed concurrently at various ones of the plurality of I/O devices, and a plurality of first and second buffers for use in each of the plurality of first and second I/O operations.

14. The invention as set forth in claim 11 above, wherein the first and second buffers used in each of the plurality of first and second I/O operations have different sizes.

15. The invention as set forth in claim 3 above, wherein the memory management means comprises means for allocating and de-allocating variable-sized blocks of memory from a plurality of memory segments comprising memories of different sizes and speeds coupled to the DSP.

16. The invention as set forth in claim 4 above, wherein each request to the memory management means comprises at least two attributes associated therewith, wherein the attributes comprises a physical size of the block and the memory segment where the block is located.

17. The invention as set forth in claim 16 above, wherein the attributes further comprise an alignment field indicating that the block is aligned on a particular boundary.

18. The invention as set forth in claim 15 above, wherein the memory management means comprises means for mapping the blocks to physical locations in the memory segments.

19. The invention as set forth in claim 3 above, further comprising:

(1) a host computer operatively coupled to the DSP, wherein the host computer executes at least one computer program concurrently with the DSP and the operating system; and

(2) means, executed by the host computer, for transferring data between the host computer and the DSP.

20. The invention as set forth in claim 19 above, wherein the host computer operates under control of a separate operating system and operates concurrently with the computer programs executed by the DSP.

21. The invention as set forth in claim 19 above, wherein the means for transferring comprises means, executed by the DSP, for accessing a file system on the host computer.

22. The invention as set forth in claim 19 above, wherein the DSP includes a file device, and the host computer includes a computer program for servicing requests from the file device, wherein the means for transferring further comprises means for communicating with the file device using a stream I/O module.

23. The invention as set forth in claim 19 above, wherein the means for communicating comprises a remote procedure call.

24. The invention as set forth in claim 19 above, further comprising external memory coupled to the DSP.

25. The invention as set forth in claim 24 above, further comprising means for sharing the external memory coupled to the DSP with the host computer.