 |
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
| Datasheet File OCR Text: |
| User's Manual RX850 Pro Real-Time Operating System Basics Target Devices V850 SeriesTM Target Real-Time OS RX850 Pro Ver. 3.15 Document No. U13773EJ3V0UM00 (3rd edition) Date Published April 2003 N CP(K) 1999 Printed in Japan [MEMO] 2 User's Manual U13773EJ3V0UM V850 Series, V851, V852, V853, V854, V850/SA1, V850/SB1, V850/SB2, V850/SC1, V850/SC2, V850/SC3, V850/SV1, V850/SF1, V850E/MS1, V850E/MS2, V850E/MA1, V850E/MA2, V850E/IA1, V850E/IA2, V850ES/SA2, V850ES/SA3, V850ES/KF1, V850ES/KG1, and V850ES/KJ1 are trademarks of NEC Electronics Corporation. MS-DOS, Windows, and Windows NT are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. UNIX is a registered trademark in the United States and other countries licensed exclusively through X/Open Company, Ltd. PC/AT is a trademark of International Business Machines Corporation. Green Hills Software and MULTI are trademarks of United States Green Hills Software, Inc. SPARCstation is a trademark of SPARC International, Inc. Solaris is a trademark of Sun Microsystems, Inc. TRON is an abbreviation for The Real-time Operating system Nucleus. ITRON is an abbreviation for Industrial TRON. ITRON is an abbreviation for "Micro Industrial TRON". User's Manual U13773EJ3V0UM 3 * The information in this document is current as of November, 2002. The information is subject to change without notice. For actual design-in, refer to the latest publications of NEC Electronics data sheets or data books, etc., for the most up-to-date specifications of NEC Electronics products. Not all products and/or types are available in every country. Please check with an NEC Electronics sales representative for availability and additional information. * No part of this document may be copied or reproduced in any form or by any means without the prior written consent of NEC Electronics. NEC Electronics assumes no responsibility for any errors that may appear in this document. * NEC Electronics does not assume any liability for infringement of patents, copyrights or other intellectual property rights of third parties by or arising from the use of NEC Electronics products listed in this document or any other liability arising from the use of such products. No license, express, implied or otherwise, is granted under any patents, copyrights or other intellectual property rights of NEC Electronics or others. * Descriptions of circuits, software and other related information in this document are provided for illustrative purposes in semiconductor product operation and application examples. The incorporation of these circuits, software and information in the design of a customer's equipment shall be done under the full responsibility of the customer. NEC Electronics assumes no responsibility for any losses incurred by customers or third parties arising from the use of these circuits, software and information. * While NEC Electronics endeavors to enhance the quality, reliability and safety of NEC Electronics products, customers agree and acknowledge that the possibility of defects thereof cannot be eliminated entirely. To minimize risks of damage to property or injury (including death) to persons arising from defects in NEC Electronics products, customers must incorporate sufficient safety measures in their design, such as redundancy, fire-containment and anti-failure features. * NEC Electronics products are classified into the following three quality grades: "Standard", "Special" and "Specific". The "Specific" quality grade applies only to NEC Electronics products developed based on a customerdesignated "quality assurance program" for a specific application. The recommended applications of an NEC Electronics product depend on its quality grade, as indicated below. Customers must check the quality grade of each NEC Electronics product before using it in a particular application. "Standard": Computers, office equipment, communications equipment, test and measurement equipment, audio and visual equipment, home electronic appliances, machine tools, personal electronic equipment and industrial robots. "Special": Transportation equipment (automobiles, trains, ships, etc.), traffic control systems, anti-disaster systems, anti-crime systems, safety equipment and medical equipment (not specifically designed for life support). "Specific": Aircraft, aerospace equipment, submersible repeaters, nuclear reactor control systems, life support systems and medical equipment for life support, etc. The quality grade of NEC Electronics products is "Standard" unless otherwise expressly specified in NEC Electronics data sheets or data books, etc. If customers wish to use NEC Electronics products in applications not intended by NEC Electronics, they must contact an NEC Electronics sales representative in advance to determine NEC Electronics' willingness to support a given application. (Note) (1) "NEC Electronics" as used in this statement means NEC Electronics Corporation and also includes its majority-owned subsidiaries. (2) "NEC Electronics products" means any product developed or manufactured by or for NEC Electronics (as defined above). M8E 02. 11-1 4 User's Manual U13773EJ3V0UM Regional Information Some information contained in this document may vary from country to country. Before using any NEC Electronics product in your application, pIease contact the NEC Electronics office in your country to obtain a list of authorized representatives and distributors. They will verify: * * * * * Device availability Ordering information Product release schedule Availability of related technical literature Development environment specifications (for example, specifications for third-party tools and components, host computers, power plugs, AC supply voltages, and so forth) Network requirements * In addition, trademarks, registered trademarks, export restrictions, and other legal issues may also vary from country to country. NEC Electronics America, Inc. (U.S.) Santa Clara, California Tel: 408-588-6000 800-366-9782 Fax: 408-588-6130 800-729-9288 * Filiale Italiana Milano, Italy Tel: 02-66 75 41 Fax: 02-66 75 42 99 * Branch The Netherlands Eindhoven, The Netherlands Tel: 040-244 58 45 Fax: 040-244 45 80 * Tyskland Filial Taeby, Sweden Tel: 08-63 80 820 Fax: 08-63 80 388 * United Kingdom Branch Milton Keynes, UK Tel: 01908-691-133 Fax: 01908-670-290 NEC Electronics Hong Kong Ltd. Hong Kong Tel: 2886-9318 Fax: 2886-9022/9044 NEC Electronics Hong Kong Ltd. Seoul Branch Seoul, Korea Tel: 02-528-0303 Fax: 02-528-4411 NEC Electronics (Europe) GmbH Duesseldorf, Germany Tel: 0211-65 03 01 Fax: 0211-65 03 327 * Sucursal en Espana Madrid, Spain Tel: 091-504 27 87 Fax: 091-504 28 60 * Succursale Francaise Velizy-Villacoublay, France Tel: 01-30-67 58 00 Fax: 01-30-67 58 99 NEC Electronics Shanghai, Ltd. Shanghai, P.R. China Tel: 021-6841-1138 Fax: 021-6841-1137 NEC Electronics Taiwan Ltd. Taipei, Taiwan Tel: 02-2719-2377 Fax: 02-2719-5951 NEC Electronics Singapore Pte. Ltd. Novena Square, Singapore Tel: 6253-8311 Fax: 6250-3583 J02.11 User's Manual U13773EJ3V0UM 5 Major Revisions in This Edition Page Throughout p.20 p.21 p.22 p.68 p.80 p.193 pp.199, 200 Modification of V850 Family to V850 Series Modification of description in 1.7 Execution Environment Modification of description in 1.8 Development Environment Addition of /librxpm.a to nucleus library in 1.9 System Construction Procedure Modification of description of Caution 2 in 6.3.4 Returning a memory block Modification of description in 7.6.5 Interrupts in cyclically activated handler Modification of description of Caution for rel_blk in 11.8.5 Memory pool management system calls Modification of description of Structure of system clock SYSTIME for set_tim and get_tim in 11.8.6 Time management system calls Addition of APPENDIX B Q & A Description p.238 The mark shows major revised points. 6 User's Manual U13773EJ3V0UM INTRODUCTION Target Readers This manual is intended for users engaged in the design or development of application systems of the V850 Series. Purpose This manual is intended to give users an understanding of the functions of the RX850 Pro described in the Organization below. Organization This manual is roughly organized into the following sections. * * * * * * * * * * * Overview Nucleus Task management function Synchronous communication functions Interrupt management function Memory pool management function Time management function Scheduler System initialization Interface library System calls How to Use This Manual It is assumed that the reader of this manual has general knowledge in the fields of electrical engineering, logic circuits, microcontrollers, C language, and assembly language. To learn about the hardware functions or command functions of the V850 Series: Refer to the user's manual for the relevant product. Conventions Note: Caution: Remark: Numerical representation: Footnote for item marked with Note in the text Information requiring particular attention Supplementary information Binary ... XXXX or B'XXXX Decimal ... XXXX Hexadecimal ... 0xXXXX or H'XXXX Prefix indicating power of 2 (address space, memory capacity): K (kilo) 2 = 1024 M (mega) 2 = 1024 20 2 10 User's Manual U13773EJ3V0UM 7 Related Documents The following documents may be referenced together with this manual. The related documents indicated in this publication may include preliminary versions. However, preliminary versions are not marked as such. Documents related to development tools (user's manuals) Document Name IE-703002-MC (In-circuit emulator for V851TM, V852TM, V853TM, V854TM, V850/SA1TM, V850/SB1TM, V850/SB2TM, V850/SV1TM, V850/SF1TM, V850/SC1TM, V850/SC2TM, and V850/SC3TM) IE-703003-MC-EM1 (In-circuit emulator option board for V853) IE-703008-MC-EM1 (In-circuit emulator option board for V854) IE-703017-MC-EM1 (In-circuit emulator option board for V850/SA1) IE-703037-MC-EM1 (In-circuit emulator option board for V850/SB1 and V850/SB2) IE-703040-MC-EM1 (In-circuit emulator option board for V850/SV1) IE-703079-MC-EM1 (In-circuit emulator option board for V850/SF1) IE-703089-MC-EM1 (In-circuit emulator option board for V850/SC1, V850/SC2, and V850/SC3) IE-703102-MC (In-circuit emulator for V850E/MS1 and V850E/MS2 ) IE-703102-MC-EM1 (In-circuit emulator option board for V850E/MS1 and V850E/MS2), IE-703102-MC-EM1-A (In-circuit emulator option board for V850E/MS1) IE-V850E-MC (In-circuit emulator for V850E/IA1TM and V850E/IA2TM), IE-V850E-MC-A (In-circuit emulator for V850E1 (NB85E core), V850E/MA1TM, and V850E/MA2TM) IE-V850E-MC-EM1-A (In-circuit emulator option board for V850E1 (NB85E core)) IE-V850E-MC-EM1-B, IE-V850E-MC-MM2 (In-circuit emulator option board for V850E1 (NB85E core)) IE-703107-MC-EM1 (In-circuit emulator option board for V850E/MA1 and V850E/MA2) IE-703116-MC-EM1 (In-circuit emulator option board for V850E/IA1) IE-703114-MC-EM1 (In-circuit emulator option board for V850E/IA2) CA850 Ver. 2.50 C compiler package Operation C Language PM plus Assembly Language ID850 Ver. 2.40 Integrated debugger SM850 Ver. 2.40 System simulator SM850 Ver. 2.00 or later System simulator RX850 Ver. 3.13 or later Real-time OS Operation Windows Based Operation Windows Based External Part User Open Interface Specifications Basics Installation Technical RX850 Pro Ver. 3.15 Real-time OS Basics Installation Technical RD850 Ver. 3.01 Task debugger RD850 Pro Ver. 3.01 Task debugger AZ850 Ver. 3.10 System performance analyzer PG-FP4 Flash memory programmer TM TM TM Document No. U11595E U11596E U12420E U12898E U14151E U14337E U15447E U15776E U13875E U13876E U14487E To be prepared U14482E U14481E U14700E To be prepared U16053E U16054E To be prepared U16042E U15181E U15182E U14873E U13430E U13410E U13431E This manual U13774E U13772E U13737E U13916E U14410E U15260E 8 User's Manual U13773EJ3V0UM CONTENTS CHAPTER 1 OVERVIEW..........................................................................................................................16 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Overview........................................................................................................................................................16 Real-Time OS ................................................................................................................................................16 Multitask OS ..................................................................................................................................................17 Features.........................................................................................................................................................17 Configuration ................................................................................................................................................19 Applications ..................................................................................................................................................20 Execution Environment................................................................................................................................20 Development Environment ..........................................................................................................................21 1.8.1 1.8.2 1.9 Hardware environment........................................................................................................................21 Software environment .........................................................................................................................21 System Construction Procedure .................................................................................................................22 CHAPTER 2 NUCLEUS............................................................................................................................27 2.1 2.2 Overview........................................................................................................................................................27 Functions.......................................................................................................................................................28 CHAPTER 3 TASK MANAGEMENT FUNCTION ..................................................................................30 3.1 3.2 3.3 3.4 3.5 3.6 3.7 Overview........................................................................................................................................................30 Task States....................................................................................................................................................30 Creating Tasks ..............................................................................................................................................33 Activating Tasks ...........................................................................................................................................33 Terminating Tasks ........................................................................................................................................33 Deleting Tasks ..............................................................................................................................................34 Internal Processing of Task.........................................................................................................................34 3.7.1 3.7.2 Acquiring task information ..................................................................................................................35 Acquiring ID number ...........................................................................................................................36 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS ........................................................37 4.1 4.2 Overview........................................................................................................................................................37 Semaphores ..................................................................................................................................................37 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.2.7 4.3 4.3.1 4.3.2 4.3.3 Generating semaphores .....................................................................................................................38 Deleting semaphores ..........................................................................................................................38 Returning resources ...........................................................................................................................38 Acquiring resources ............................................................................................................................39 Acquiring semaphore information .......................................................................................................40 Acquiring ID number ...........................................................................................................................40 Exclusive control using semaphores ..................................................................................................40 Generating event flags........................................................................................................................43 Deleting event flags ............................................................................................................................44 Setting a bit pattern.............................................................................................................................44 User's Manual U13773EJ3V0UM Event Flags....................................................................................................................................................43 9 4.3.4 4.3.5 4.3.6 4.3.7 4.3.8 4.4 4.4.1 4.4.2 4.4.3 4.4.4 4.4.5 4.4.6 4.4.7 4.4.8 Clearing a bit pattern...........................................................................................................................44 Checking a bit pattern .........................................................................................................................44 Acquiring event flag information..........................................................................................................45 Acquiring ID number ...........................................................................................................................46 Wait function using event flags ...........................................................................................................46 Generating mailboxes .........................................................................................................................48 Deleting mailboxes..............................................................................................................................49 Transmitting a message......................................................................................................................49 Receiving a message..........................................................................................................................50 Messages............................................................................................................................................51 Acquiring mailbox information .............................................................................................................51 Acquiring ID number ...........................................................................................................................52 Inter task communication using mailboxes .........................................................................................52 Mailboxes.......................................................................................................................................................48 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION ....................................................................... 54 5.1 5.2 5.3 Overview ........................................................................................................................................................54 Interrupt Handler ...........................................................................................................................................54 Directly Activated Interrupt Handler............................................................................................................55 5.3.1 5.3.2 5.4 5.4.1 5.4.2 5.5 5.6 5.7 5.8 5.9 Registering directly activated interrupt handler...................................................................................55 Processing in directly activated interrupt handler ...............................................................................55 Registering indirectly activated interrupt handler ................................................................................58 Processing in indirectly activated interrupt handler.............................................................................59 Indirectly Activated Interrupt Handler.........................................................................................................58 Disabling/Resuming Maskable Interrupt Acknowledgement ....................................................................61 Changing/Acquiring Interrupt Control Register .........................................................................................62 Non-Maskable Interrupts ..............................................................................................................................63 Clock Interrupts ............................................................................................................................................63 Multiple Interrupts.........................................................................................................................................63 CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION ............................................................... 64 6.1 6.2 6.3 Overview ........................................................................................................................................................64 Management Objects ....................................................................................................................................65 Memory Pool and Memory Blocks...............................................................................................................66 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.3.7 Generating a memory pool..................................................................................................................66 Deleting a memory pool ......................................................................................................................67 Acquiring a memory block...................................................................................................................67 Returning a memory block ..................................................................................................................68 Acquiring memory pool information.....................................................................................................69 Acquiring ID number ...........................................................................................................................70 Dynamic management of memory block by memory pool ..................................................................70 CHAPTER 7 TIME MANAGEMENT FUNCTION ................................................................................... 72 7.1 7.2 Overview ........................................................................................................................................................72 System Clock ................................................................................................................................................72 10 User's Manual U13773EJ3V0UM 7.2.1 7.3 7.4 7.5 7.6 Setting and reading the system clock .................................................................................................72 Timer Operations ..........................................................................................................................................72 Delayed Task Wake-Up ................................................................................................................................73 Timeout..........................................................................................................................................................74 Cyclically Activated Handler........................................................................................................................76 7.6.1 7.6.2 7.6.3 7.6.4 7.6.5 7.6.6 Registering a cyclically activated handler ...........................................................................................76 Activity state of cyclically activated handler........................................................................................77 Internal processing performed by cyclically activated handler............................................................78 Acquiring cyclically activated handler information ..............................................................................79 Interrupts in cyclically activated handler .............................................................................................80 Activation order of cyclically activated handler ...................................................................................80 CHAPTER 8 SCHEDULER.......................................................................................................................81 8.1 8.2 8.3 Overview........................................................................................................................................................81 Drive Method .................................................................................................................................................81 Scheduling Method.......................................................................................................................................82 8.3.1 8.3.2 8.4 8.5 8.6 8.7 Priority method....................................................................................................................................82 FCFS method......................................................................................................................................82 Implementing a Round-Robin Method ........................................................................................................83 Scheduling Lock Function...........................................................................................................................86 Scheduling While Handler Is Operating......................................................................................................88 Idle Handler ...................................................................................................................................................88 8.7.1 Idle handler .........................................................................................................................................88 CHAPTER 9 SYSTEM INITIALIZATION .................................................................................................89 9.1 9.2 9.3 9.4 9.5 Overview........................................................................................................................................................89 Boot Processing ...........................................................................................................................................90 Hardware Initialization Section....................................................................................................................90 Nucleus Initialization Section ......................................................................................................................91 Software Initialization Section .....................................................................................................................91 CHAPTER 10 INTERFACE LIBRARY.....................................................................................................92 10.1 Overview........................................................................................................................................................92 10.2 Processing in the Interface Library.............................................................................................................93 10.3 Types of Interface Libraries.........................................................................................................................93 10.4 Supported Interface Libraries......................................................................................................................93 CHAPTER 11 SYSTEM CALLS ..............................................................................................................94 11.1 Overview........................................................................................................................................................94 11.2 Calling System Calls ....................................................................................................................................96 11.3 System Call Function Codes .......................................................................................................................96 11.4 Data Types of Parameters............................................................................................................................97 11.5 Parameter Value Range................................................................................................................................98 11.6 System Call Return Values ..........................................................................................................................99 11.7 System Call Extension .................................................................................................................................99 User's Manual U13773EJ3V0UM 11 11.8 Explanation of System Calls ......................................................................................................................100 11.8.1 Task management system calls........................................................................................................102 11.8.2 Task-associated synchronization system calls .................................................................................121 11.8.3 Synchronous communication system calls .......................................................................................129 11.8.4 Interrupt management system calls ..................................................................................................169 11.8.5 Memory pool management system calls...........................................................................................183 11.8.6 Time management system calls........................................................................................................198 11.8.7 System management system calls....................................................................................................207 APPENDIX A PROGRAMMING METHODS......................................................................................... 214 A.1 A.2 A.3 A.4 Overview ......................................................................................................................................................214 Keywords.....................................................................................................................................................215 Reserved Words..........................................................................................................................................215 Tasks............................................................................................................................................................216 A.4.1 CA850-supported version .................................................................................................................216 A.4.2 CCV850-supported version...............................................................................................................218 A.5 Directly Activated Interrupt Handler..........................................................................................................220 A.5.1 CA850-supported version .................................................................................................................220 A.5.2 CCV850-supported version...............................................................................................................223 A.6 Indirectly Activated Interrupt Handler.......................................................................................................226 A.6.1 CA850-supported version .................................................................................................................226 A.6.2 CCV850-supported version...............................................................................................................228 A.7 Cyclically Activated Handler ......................................................................................................................230 A.7.1 CA850-supported version .................................................................................................................230 A.7.2 CCV850-supported version...............................................................................................................232 A.8 Extended SVC Handler ...............................................................................................................................234 A.8.1 CA850-supported version .................................................................................................................234 A.8.2 CCV850-supported version...............................................................................................................236 APPENDIX B Q & A ............................................................................................................................... 238 APPENDIX C INDEX .............................................................................................................................. 284 APPENDIX D REVISION HISTORY...................................................................................................... 290 12 User's Manual U13773EJ3V0UM LIST OF FIGURES (1/2) Figure No. 1-1 1-2 2-1 3-1 4-1 4-2 4-3 4-4 4-5 4-6 4-7 4-8 4-9 4-10 5-1 5-2 5-3 5-4 5-5 6-1 6-2 6-3 6-4 7-1 7-2 7-3 8-1 8-2 8-3 8-4 8-5 8-6 8-7 8-8 9-1 Title Page System Construction Procedure (When CA850 Is Used) ............................................................................23 System Construction Procedure (When CCV850 Is Used) ..........................................................................25 Nucleus Configuration..................................................................................................................................27 Task State Transition ...................................................................................................................................32 State of Semaphore Counter .......................................................................................................................41 State of Wait Queue (When wai_sem Is Issued) .........................................................................................41 State of Wait Queue (When sig_sem Is Issued) ..........................................................................................41 Exclusive Control Using Semaphores..........................................................................................................42 State of Wait Queue (When wai_flg Is Issued) ............................................................................................46 State of Wait Queue (When set_flg Is Issued) .............................................................................................47 Wait and Control by Event Flags..................................................................................................................47 State of Task Wait Queue (When rcv_msg Is Issued) .................................................................................52 State of Task Wait Queue (When snd_msg Is Issued) ................................................................................53 Inter-Task Communication Using Mailboxes................................................................................................53 Flow of Processing Performed by Directly Activated Interrupt Handler .......................................................55 Operation Flow of Indirectly Activated Interrupt Handler..............................................................................58 Control Flow if Interrupt Mask Processing Is Not Performed (Normal) ........................................................61 Control Flow if loc_cpu System Call Is Issued .............................................................................................62 Processing Flow for Handling Multiple Interrupts.........................................................................................63 Typical Arrangement of Management Objects.............................................................................................65 State of Wait Queue (When get_blk Is Issued) ............................................................................................70 State of Wait Queue (When rel_blk Is Issued) .............................................................................................71 Dynamic Use of Memory by Memory Pool ...................................................................................................71 Flow of Processing After Issuance of dly_tsk ..............................................................................................73 Flow of Processing After Issuance of act_cyc (TCY_ON)............................................................................77 Flow of Processing After Issuance of act_cyc (TCY_ON|TCY_INI) .............................................................78 Ready Queue State (1) ................................................................................................................................83 Ready Queue State (2) ................................................................................................................................84 Ready Queue State (3) ................................................................................................................................84 Flow of Processing by Using Round-Robin Method.....................................................................................85 Flow of Control if Scheduling Processing Is Not Delayed (Normal) .............................................................86 Flow of Control if dis_dsp System Call Is Issued .........................................................................................87 Flow of Control if loc_cpu System Call Is Issued .........................................................................................87 Flow of Control if wup_tsk System Call Is Issued ........................................................................................88 Flow of System Initialization.........................................................................................................................89 User's Manual U13773EJ3V0UM 13 LIST OF FIGURES (2/2) Figure No. 10-1 11-1 A-1 A-2 A-3 A-4 A-5 A-6 A-7 A-8 A-9 A-10 A-11 A-12 A-13 A-14 A-15 A-16 A-17 A-18 Title Page Positioning of Interface Library.................................................................................................................... 92 System Call Description Format................................................................................................................ 100 Task Description Format When Using CA850 (C Language).................................................................... 216 Task Description Format When Using CA850 (Assembly Language)....................................................... 217 Task Description Format When Using CCV850 (C Language) ................................................................. 218 Task Description Format When Using CCV850 (Assembly Language) .................................................... 219 Description Format of Directly Activated Interrupt Handler When Using CA850 (Assembly Language) ....... 220 Description Format of Directly Activated Interrupt Handler When Using CCV850 (Assembly Language) ............................................................................................................................... 223 Description Format of Indirectly Activated Interrupt Handler When Using CA850 (C Language) ............. 226 Description Format of Indirectly Activated Interrupt Handler When Using CA850 (Assembly Language) ............................................................................................................................... 227 Description Format of Indirectly Activated Interrupt Handler When Using CCV850 (C Language)........... 228 Description Format of Indirectly Activated Interrupt Handler When Using CCV850 (Assembly Language) ............................................................................................................................... 229 Description Format of Cyclically Activated Handler When Using CA850 (C Language) ........................... 230 Description Format of Cyclically Activated Handler When Using CA850 (Assembly Language) .............. 231 Description Format of Cyclically Activated Handler When Using CCV850 (C Language)......................... 232 Description Format of Cyclically Activated Handler When Using CCV850 (Assembly Language)............ 233 Description Format of Extended SVC Handler When Using CA850 (C Language) .................................. 234 Description Format of Extended SVC Handler When Using CA850 (Assembly Language) ..................... 235 Description Format of Extended SVC Handler When Using CCV850 (C Language)................................ 236 Description Format of Extended SVC Handler When Using CCV850 (Assembly Language)................... 237 14 User's Manual U13773EJ3V0UM LIST OF TABLES Table No. 6-1 11-1 11-2 11-3 11-4 11-5 11-6 11-7 11-8 11-9 11-10 11-11 Title Page Memory Information Allocation Combination ...............................................................................................64 System Call Function Codes........................................................................................................................96 Data Types of Parameters ...........................................................................................................................97 Ranges of Parameter Values .......................................................................................................................98 System Call Return Values ..........................................................................................................................99 Task Management System Calls ...............................................................................................................102 Task-Associated Synchronization System Calls ........................................................................................121 Synchronous Communication System Calls ..............................................................................................129 Interrupt Management System Calls..........................................................................................................169 Memory Pool Management System Calls ..................................................................................................183 Time Management System Calls ...............................................................................................................198 System Management System Calls ...........................................................................................................207 User's Manual U13773EJ3V0UM 15 CHAPTER 1 OVERVIEW Rapid advances in semiconductor technologies have led to the explosive spread of microprocessors, to the extent that they are now to be found in more fields than many would have imagined only a few years ago. In line with this spread, the number of processing programs that must be created for ever-newer high-performance, multi-function microprocessors is also increasing. This rule of growth makes it difficult to create processing programs specific to given hardware. For this reason, there is a need for operating systems (OSs) that can fully exploit the capabilities of the latest generation of microprocessors. Operating systems are broadly classified into two types: program-development OSs and control OSs. Programdevelopment OSs are to be found in those environments in which standard OSs (e.g., MS-DOSTM, Windows, and UNIX TM OS) predominate because the hardware configuration to be used for development can be limited to some extent (e.g., personal computers). Conversely, control OSs are incorporated into control units. That is, these OSs are found in those environments where standard OSs cannot easily be applied because the hardware configuration varies from system to system and because efficient operation matching the application is required. In order to adequately exploit the functions of the V850 Series of high-performance microcontrollers, developed in consideration of the current market conditions, as well as support the creation of more systematic software, NEC Electronics has now released the RX850 Pro. The RX850 Pro is a control OS for real-time, multitask processing that it has been developed to increase the application range of high-performance, multi-function microcontrollers and further improve their flexibility. 1.1 Overview The RX850 Pro is an embedded real-time, multitask control OS that provides a highly efficient real-time, multitasking environment to increase the application range of processor control units. The RX850 Pro is a high-speed, compact OS capable of being stored in and run from the ROM of a target system. 1.2 Real-Time OS Control equipment demands systems that can rapidly respond to events occurring both internal and external to the equipment. Conventional systems have utilized simple interrupt handling as a means of satisfying this demand. As control equipment has become more powerful, however, it has proved difficult for systems to satisfy these requirements by means of simple interrupt handling alone. In other words, the task of managing the order in which internal and external events are processed has become increasingly difficult as systems have increased in complexity and programs have become larger. Real-time operating systems have been designed to overcome this problem. The main purpose of a real-time OS is to respond to internal and external events rapidly and execute programs in the optimum order. 16 User's Manual U13773EJ3V0UM CHAPTER 1 OVERVIEW 1.3 Multitask OS A "task" is the minimum unit in which a program can be executed by an OS. "Multitasking" is the name given to the mode of operation in which a single processor processes multiple tasks concurrently. Actually, the processor can handle no more than one program (instruction) at a time. processed simultaneously. A multitask OS enables the parallel processing of tasks by switching the tasks to be executed as determined by the system. One important purpose of a multitask OS is to improve the throughput of the overall system through the parallel processing of multiple tasks. But, by switching the processor's attention to individual tasks on a regular basis (at a certain timing) it appears that the tasks are being 1.4 Features The RX850 Pro has the following features. (1) Conformity with ITRON3.0 specification As a representative embedded type control OS architecture, the RX850 Pro performs design which is compatible with ITRON3.0 specifications, and includes all the functions up to level E. The ITRON3.0 specification applies to an embedded, real-time control OS. (2) High flexibility The RX850 Pro supports system-specific system calls (7 types), as well as those defined in the ITRON3.0 specification (67 types). The RX850 Pro thus offers superior application system flexibility. The RX850 Pro can be used to create a real-time, multitask OS that is compact and that can satisfy the user's needs because the functions (system calls) to be used by the application system can be selected at system construction. (3) Realization of real-time processing and multitasking The RX850 Pro supports the following functions to realize complete real-time processing and multitasking. * Task management function * Task-associated synchronization function * Synchronous communication function * Interrupt management function * Memory pool management function * Time management function * System management function * Scheduling function (4) Scheduling lock function The RX850 Pro supports a function to disable/resume dispatch processing (task scheduling processing). This function enables users to disable/resume dispatch processing from the processing program level. User's Manual U13773EJ3V0UM 17 CHAPTER 1 OVERVIEW (5) ROMization The RX850 Pro is a real-time, multitask OS that has been designed on the assumption that it will be incorporated into the target system, and has been made as compact as possible to enable it to be loaded into a system's ROM. (6) Use of original instructions The RX850 Pro realizes high-speed processing through the high-speed instruction execution of the V850 Series microcontrollers and by using original instructions. (7) Use of internal ROM/RAM The RX850 Pro realizes high-speed instruction execution and high-speed data access by using the V850 Series internal ROM and RAM. (8) Application utility support The RX850 Pro supports the following two utilities to aid in application system construction. * Configurator CF850 Pro * High-level language interface library (9) C compiler The RX850 Pro supports the following C compilers for the V850 Series. * CA850 (NEC Electronics) * CCV850 (Green Hills Software , Inc.) TM 18 User's Manual U13773EJ3V0UM CHAPTER 1 OVERVIEW 1.5 Configuration The RX850 Pro consists of four subsystems: the nucleus, system initialization, interface library, and system configurator. These subsystems are outlined below. (1) Nucleus The nucleus forms the heart of the RX850 Pro, a system that supports real-time, multitask control. The nucleus provides the following functions. * Creation/initialization of management objects * Processing of system calls issued by the processing program (task/non-task) * Selection of the processing program (task/non-task) to be executed next, according to an event that occurs internal or external to the target system Management object creation/initialization and system call processing are executed by management modules. Processing program selection is performed by a scheduler. (2) System initialization System initialization includes the hardware initialization and software initialization necessary for the RX850 Pro to run. When the system is started, therefore, system initialization is executed first. Among the system initialization processes, sample source files are supplied for the portion that is dependent on the hardware configuration of the execution environment (boot processing and hardware initialization block) and the portion that makes the software environment conformable (software initialization section). These sample source files improve transplantability to various target systems and facilitate customization. (3) Interface library When a processing program (task/non-task) is written in C language, the external function format is used to issue a system call or call an extended SVC handler. The issue format that can be understood by the nucleus (nucleus issue format), however, differs from the external function format. Therefore, the interface library is supported to translate a system call, issued in external function format or an extended SVC handler called in that format, into the nucleus issue format. The interface library thus acts as an agent between processing programs and the nucleus. Furthermore, an interface library compatible with the CA850 C compiler for the NEC Electronics V850 Series and the C cross V800 compiler CCV850 manufactured by Green Hills Software, Inc. are available with the RX850 Pro. (4) Configurator CF850 Pro To organize a system using the RX850 Pro, information files holding various data to be supplied to the RX850 Pro (system information table, branch table, and system information header file) are necessary. Because these files are basically an enumeration of data in specified formats, they can be described by using editors. In this case, however, description and legibility are poor. Therefore, the RX850 Pro supplies a utility that converts files created in an original description format that has excellent description and legibility (configuration files) to information files. This utility, the "configurator CF850 Pro", reads a configuration file created in an original format as an input file and outputs information files such as the system information table, branch table, and system information header file. User's Manual U13773EJ3V0UM 19 CHAPTER 1 OVERVIEW 1.6 Applications The RX850 Pro is suitable for the following devices. * Systems using motor controllers PPCs, printers, FAXes * Systems requiring low power consumption Cellular phones, personal handyphones (PHS), digital still cameras 1.7 Execution Environment The RX850 Pro was developed as an OS for embedded control. It therefore runs on a target system equipped with the following hardware. (1) Operating CPU * V850 core * V850E1 core * V850E2 core * V850ES core * V850ES/Kx1 Series * V850ES/Kx1+ Series (2) Peripheral controller The RX850 Pro extracts the hardware-dependent blocks of the nucleus and supplies them as sample source files to support various execution environments. The support is available by rewriting the sample source files for each target system, therefore, no special peripheral controller is required. (3) Data access to external RAM The SPOL0 area that is allocated to the RX850 Pro's management area requires support for 8-bit (1-byte) data access. In other words, in cases where 8-bit access to the external RAM area is not supported, SPOL0 area must not be allocated to that area. If it is allocated there, the data being operated on will be lost and normal operation cannot be guaranteed. As for access to the stack or memory pool, since this does not require 8-bit access, allocation poses no problems as far as the SPOL1 area is concerned. However, compile options or other means are needed to suppress 8-bit access in the code that accesses data in any memory blocks or other blocks that are created in the SPOL1 area. This problem can be resolved allocating the SPOL0 area to the internal RAM area in a V850 Series device. 20 User's Manual U13773EJ3V0UM CHAPTER 1 OVERVIEW 1.8 Development Environment The hardware and software environments required for system development are shown below. 1.8.1 Hardware environment (1) Host machine * PC that runs Windows * SPARCstation TM (2) In-circuit emulator Select an in-circuit emulator that is compatible with the CPU to be used. For details, see the relevant pamphlet or other reference document. (3) I/O board for in-circuit emulator Select an I/O board that is compatible with the CPU to be used. For details, see the relevant pamphlet or other reference document. (4) PC interface board Select a PC interface board that is compatible with the in-circuit emulator and host machine to be used. 1.8.2 Software environment (1) OS (host machine in parentheses) * Windows 98, Me, 2000, Windows NT 4.0 (PC-9800 series, IBM PC/AT and compatibles) TM * Solaris TM 2.x (SPARCstation) (2) Cross tools * CA850 (product of NEC Electronics) * CCV850 (product of GHS) (3) Debuggers * ID850 (product of NEC Electronics) * SM850 (product of NEC Electronics) * MULTI , MULTI2000 (product of GHS) TM * PARTNER (product of KMC) (4) Task debugger * RD850 Pro (product of NEC Electronics) Remark Included with RX850 Pro package (5) System performance analyzer * AZ850 (product of NEC Electronics) User's Manual U13773EJ3V0UM 21 CHAPTER 1 OVERVIEW 1.9 System Construction Procedure System construction involves incorporating created load modules into a target system, using the file group copied from the RX850 Pro distribution media (CGMT or floppy disk) to the user development environment (host machine). The RX850 Pro system construction procedure is outlined below. For details, refer to the RX850 Pro Installation User's Manual (U13774E). (1) Creating a configuration file (2) Creating an information definition file * System information table (SIT) * System call table (SCT) * System information header file Note that these information tables are created using a configurator. (3) Creating system initialization processing * Boot processing * Hardware initialization section * Software initialization section (4) Creating processing programs * Task * Interrupt handler * Cyclically activated handler * Extended SVC handler * Interface library for an extended SVC handler These programs are created using C language or assembly language. (5) Creating an initialization data save area (only when CA850 is used) (6) Creating a link directive file (section map file) (7) Creating a load module (8) Incorporating the load module into the system Remark When using the CCV850 C cross V800 compiler manufactured by Green Hills Software, Inc., the initialization data save area does not need to be created. For details of the creation of the initialization data save area, refer to the CA850 C Compiler Package Operation User's Manual (U16053E). An example of the system construction procedure when the NEC Electronics V850 Series C compiler CA850 is used is shown in Figure 1-1 and an example of the system construction procedure when the C cross V800 compiler CCV850 manufactured by Green Hills Software, Inc. is used is shown in Figure 1-2. 22 User's Manual U13773EJ3V0UM CHAPTER 1 OVERVIEW Figure 1-1. System Construction Procedure (When CA850 Is Used) Configuration file sys.cf Configurator Information file sys.s sct.s sys.h System initialization boot.s init.c entry.s varfunc.c Processing program task.c inthdr.c cychdr.c svchdr.c svcif.s Initialization data save area rompcrt.c C compiler/assembler Information file sys.o sct.o System initialization boot.o init.o entry.o varfunc.o Processing program task.o inthdr.o cychdr.o svchdr.o svcif.o Initialization data save area rompcrt.o Link directive file sample.lnk Library file rxcore.o/rxtmcore.o librxp.a/librxpm.a libchp.a/libncp.a Runtime Library Link editor Load module (not including ROMization information) sample.out ROMization processor Load module (including ROMization information) sample.rom Hex converter Load module (HEX format) sample.hex User's Manual U13773EJ3V0UM 23 CHAPTER 1 OVERVIEW The files shown in Figure 1-1 are outlined below. These files are provided as samples. * Configuration file sys.cf: * Information files sys.s: sct.s: sys.h: System information table Branch table System information header file Configuration file * System initialization boot.s: init.c: entry.s: varfunc.c: Boot processing Hardware initialization section (interrupt controller initialization) Hardware initialization section (interrupt/exception entry) Software initialization section * Processing program task.c: inthdr.c: cychdr.c: svchdr.c: svcif.s: Task Interrupt handler Cyclically activated handler Extended SVC handler Interface library for an extended SVC handler * Initialization data save area rompcrt.s: * Link directive file sample.lnk: * Nucleus object rxcore.o/rxtmcore.o: Nucleus common section librxp.a/librxpm.a: libchp.a/libncp.a: * Load modules sample.out: sample.rom: sample.hex: Not including ROMization information Including ROMization information HEX format Nucleus library Interface library for system calls Link directive file Initialization data save area 24 User's Manual U13773EJ3V0UM CHAPTER 1 OVERVIEW Figure 1-2. System Construction Procedure (When CCV850 Is Used) Configuration file sys.cf Configurator Information file sys.850 sct.850 sys.h System initialization boot.850 init.c entry.850 varfunc.c Processing program task.c inthdr.c cychdr.c svchdr.c svcif.850 C compiler/assembler Information file sys.o sct.o System initialization boot.o init.o entry.o varfunc.o Processing program task.o inthdr.o cychdr.o svchdr.o svcif.o Link directive file sample.lnk Library file rxcore.o/rxtmcore.o librxp.a/librxpm.a libchp.a/libncp.a Runtime Library Link editor Load module sample.out Hex converter Load module (HEX format) sample.hex User's Manual U13773EJ3V0UM 25 CHAPTER 1 OVERVIEW The files shown in Figure 1-2 are outlined below. These files are provided as samples. * Configuration file sys.cf: * Information files sys.850: sct.850: sys.h: System information table Branch table System information header file Configuration file * System initialization boot.850: init.c: entry.850: varfunc.c: Boot processing Hardware initialization section (interrupt controller initialization) Hardware initialization section (interrupt/exception entry) Software initialization section * Processing program task.c: inthdr.c: cychdr.c: svchdr.c: svcif.850: * Link directive file sample.lnk: * Nucleus object rxcore.o/rxtmcore.o: Nucleus common section librxp.a/librxpm.a: libchp.a/libncp.a: * Load modules sample.out: sample.hex: Including/not including ROMization information HEX format Nucleus library Interface library for system calls Link directive file Task Interrupt handler Cyclically activated handler Extended SVC handler Interface library for an extended SVC handler 26 User's Manual U13773EJ3V0UM CHAPTER 2 NUCLEUS This chapter describes the nucleus, which is the core of the RX850 Pro. 2.1 Overview The nucleus forms the heart of the RX850 Pro, a system that supports real-time, multitask control. The nucleus provides the following functions. * Creation/initialization of management objects * Processing of system calls issued by processing program (task/non-task) * Selection of the processing program (task/non-task) to be executed next, according to an event that occurs internal or external to the target system Management object creation/initialization and system call processing are executed by management modules. Program selection is performed by a scheduler. The configuration of the RX850 Pro nucleus is shown below. Figure 2-1. Nucleus Configuration Other Task management Task-associated handler Task-associated synchronization Synchronous communication management Scheduler System management Interrupt management Time management Memory pool management User's Manual U13773EJ3V0UM 27 CHAPTER 2 NUCLEUS 2.2 Functions The nucleus consists of various kinds of management modules and a scheduler. This section outlines the functions of the management modules and scheduler. See CHAPTERS 3 TASK MANAGEMENT FUNCTION through 8 SCHEDULER for details of the individual functions. (1) Task management function This module manipulates and manages the states of a task, the minimum unit in which processing is performed by the RX850 Pro. For example, the module can create, start, run, stop, terminate, and delete a task. (2) Synchronous communication function This module enables three functions related to synchronous communication between tasks: exclusive control, wait, and communication. Exclusive control function: Wait function: Communication function: Semaphore Event flag Mailbox (3) Interrupt management function This module executes the processing related to maskable interrupts, such as the registration of an indirectly activated interrupt mask, return from a directly activated interrupt handler, and change or acquisition of the interrupt-enable level. (4) Memory pool management function This module manages the memory area specified at configuration, dividing it into the following two areas. * RX850 Pro area Management objects Memory pool * Processing program (task/non-task) area Text area Data area Stack area The RX850 Pro also applies dynamic memory pool management. For example, the RX850 Pro provides a function for acquiring and returning a memory area to be used as a work area as required. By exploiting this ability to dynamically manage memory, the user can utilize a limited memory area with maximum efficiency. 28 User's Manual U13773EJ3V0UM CHAPTER 2 NUCLEUS (5) Time management function This module supports a timer operation function (such as delayed wake-up of a task or activation of a cyclically activated handler) that is based on clock interrupts generated by hardware (such as a clock controller). (6) Scheduler By monitoring the dynamically changing states of tasks, this module manages and determines the order in which tasks are executed and optimally assigns tasks a processing time. The RX850 Pro determines the task execution order according to assigned priority levels and by applying the FCFS method. When started, the scheduler determines the priority levels assigned to the tasks, selects an optimum task from those ready to be executed (run or ready state), and optimally assigns tasks a processing time. Remark In the RX850 Pro, the smaller the value of the priority assigned to the task, the higher the priority. User's Manual U13773EJ3V0UM 29 CHAPTER 3 TASK MANAGEMENT FUNCTION This chapter describes the task management function performed by the RX850 Pro. 3.1 Overview Tasks are execution entities of arbitrary sizes, making them difficult to manage directly. The RX850 Pro manages task states and tasks themselves by using management objects that correspond to tasks on a one-to-one basis. Remark A task uses the execution environment information provided by the program counter, work registers, and the like when it executes processing. This information is called the task context. When the task execution is switched, the current task context is saved and the task context for the next task is loaded. 3.2 Task States The task changes its state according to how resources required to execute the processing are acquired, whether an event occurs, and so on. The RX850 Pro classifies task states into the following seven types. (1) Non-existent state A task in this state has not been created or has been deleted. A task in the non-existent state is not managed by the RX850 Pro even if its execution entity is located in memory. (2) Dormant state A task in this state has just been created or has already completed its processing. A task in the dormant state is not scheduled by the RX850 Pro. This state differs from the wait state in the following points: * All resources are released. * The task context is initialized when the processing is resumed. * A state manipulation system call (ter_tsk, chg_pri, etc.) causes an error. (3) Ready state A task in this state is ready to perform its processing. This task has been waiting for a processing time to be assigned because another task having a higher (or the same) priority level is being executed. A task in the ready state is scheduled by the RX850 Pro. (4) Run state A task in this state has been assigned a processing time and is currently performing its processing. Within the entire system, only a single task can be in the run state at any one time. 30 User's Manual U13773EJ3V0UM CHAPTER 3 TASK MANAGEMENT FUNCTION (5) Wait state A task in this state has been stopped because the requirements for performing its processing are not satisfied. The processing of this task is resumed from the point at which it was stopped, so the values that were being used immediately before the stop are restored to the task context required to resume the processing. The RX850 Pro further divides tasks in the wait state into the following six groups, according to the conditions which caused the transition to the wait state. Wake-up wait state: A task enters this state if the counter for the task (registering the number of times the wake-up request has been issued) indicates 0x0 upon the issuance of an slp_tsk or tslp_tsk system call. A task enters this state if it cannot acquire a resource from the relevant semaphore upon the issuance of a wai_sem or twai_sem system call. A task enters this state if a relevant event flag does not satisfy a predetermined condition upon the issuance of a wai_flg or twai_flg system call. A task enters this state if it cannot receive a message from the relevant mailbox upon the issuance of a rcv_msg or trcv_msg system call. A task enters this state if it cannot acquire a memory block from the relevant memory pool upon the issuance of a get_blk or tget_blk system call. A task enters this state upon the issuance of a dly_tsk system call. Resource wait state: Event flag wait state: Message wait state: Memory block wait state: Timeout wait state: (6) Suspend state A task in this state has been forcibly stopped by another task. The processing of this task is resumed from the point at which it was stopped, so the values that were being used immediately before the stop are restored to the task context required for resuming the processing. Remark RX850 Pro supports nesting of more than one level of the suspend state (up to 127 times). (7) Wait-suspend state This state is a combination of the wait and suspend states. A task in this state has entered the suspend state upon exiting from the wait state, or has entered the wait state upon exiting from the suspend state. Task state transitions are shown in Figure 3-1. User's Manual U13773EJ3V0UM 31 CHAPTER 3 TASK MANAGEMENT FUNCTION Figure 3-1. Task State Transition Dispatch Ready state Preempt Run state Wait released Wait condition Wait state Stop Resume Wait-suspend state Wait released Stop Suspend state Resume Forced termination Dormant state Activation Deletion Forced termination Termination Creation Non-existent state Termination and deletion 32 User's Manual U13773EJ3V0UM CHAPTER 3 TASK MANAGEMENT FUNCTION 3.3 Creating Tasks Two types of interfaces are provided in the RX850 Pro to create tasks: A task is created statically at system initialization (in the nucleus initialization section), or dynamically according to a system call issued from a processing program. Task in the RX850 Pro consists of three steps: A task management area (management object) is allocated in system memory. Then, the allocated task management area is initialized. Finally, the task state is changed from the non-existent state to the dormant state. (1) Static registration of a task To register a task statically, specify that task during configuration. The RX850 Pro creates a task according to the information defined in the information files (system information table and system information header file) at system initialization, and makes the task manageable. (2) Dynamic registration of a task To register a task dynamically, issue the cre_tsk system call from a processing program (task). The RX850 Pro generates a task according to the information specified with parameters upon the issuance of the cre_tsk system call, and makes the task manageable. 3.4 Activating Tasks In task activation in the RX850 Pro, a task is switched from the dormant state to the ready state, and scheduled. A task is activated by issuing the sta_tsk system call, specifying the task by the parameters. * sta_tsk system call A task specified by the parameters is switched from the dormant state to the ready state. 3.5 Terminating Tasks In task termination in the RX850 Pro, a task is switched from the ready state, run state, wait state, suspend state, or wait-suspend state to the dormant state and excluded from the schedule by the RX850 Pro. In the RX850 Pro, a task can be terminated in either of the following two ways. Normal termination: A task terminates upon completing all processing and when it need not be subsequently scheduled. Forced termination: When a number of troubles occur during processing and processing must be terminated immediately, this enables termination from another task. The task terminates upon the issuance of the following system calls. * ext_tsk system call The task that issued the ext_tsk system call is switched from the run state to the dormant state. * exd_tsk system call The task that issued the exd_tsk system call is switched from the run state to the non-existent state. * ter_tsk system call The task specified by the parameters is forcibly switched to the dormant state. User's Manual U13773EJ3V0UM 33 CHAPTER 3 TASK MANAGEMENT FUNCTION 3.6 Deleting Tasks In task deletion in the RX850 Pro, a task is switched from the run or dormant state to the non-existent state, and excluded from management by the RX850 Pro. A task is deleted upon the issuance of the following system calls. * exd_tsk system call The task that issued the exd_tsk system call is switched from the run state to the non-existent state. * del_tsk system call The task specified by the parameters is switched from the dormant state to the non-existent state. 3.7 Internal Processing of Task The RX850 Pro utilizes a unique means of scheduling to switch tasks. Therefore, when describing a task's processing, observe the following points. (1) Saving/restoring registers When switching tasks, the RX850 Pro saves and restores the contents of work registers in line with the function call conventions of the C compiler (CA850 or CCV850). This eliminates the need for coding processing to save the contents at the beginning of a task and to restore the contents at the end. If a task coded in assembly language uses a register for a register variable, however, the processing for saving the contents of that register must be coded at the beginning of the task, and the processing for restoring the contents at the end. (2) Stack switching When switching tasks, the RX850 Pro switches to the special task stack of the selected task. The processing for switching the stack need not be coded at the beginning and end of the task. (3) Limitations imposed on system calls Some of the RX850 Pro system calls cannot be issued within a task. The following system calls can be issued within a task: * Task management system calls cre_tsk ter_tsk rel_wai del_tsk dis_dsp get_tid sta_tsk ena_dsp ref_tsk ext_tsk chg_pri vget_tid exd_tsk rot_rdq * Task-associated synchronization system calls sus_tsk wup_tsk rsm_tsk can_wup frsm_tsk slp_tsk tslp_tsk 34 User's Manual U13773EJ3V0UM CHAPTER 3 TASK MANAGEMENT FUNCTION * Synchronous communication system calls cre_sem twai_sem set_flg ref_flg rcv_msg del_sem ref_sem clr_flg vget_fid prcv_msg sig_sem vget_sid wai_flg cre_mbx trcv_msg wai_sem cre_flg pol_flg del_mbx ref_mbx preq_sem del_flg twai_flg snd_msg vget_mid * Interrupt management system calls def_int chg_icr ena_int ref_icr dis_int loc_cpu unl_cpu * Memory pool management system calls cre_mpl rel_blk del_mpl ref_mpl get_blk vget_pid pget_blk tget_blk * Time management system calls set_tim ref_cyc * System management system calls get_ver ref_sys def_svc viss_svc get_tim dly_tsk def_cyc act_cyc 3.7.1 Acquiring task information Task information is acquired upon the issuance of the ref_tsk system call. * ref_tsk system call Task information (such as extended information or the current priority) for the task specified by the parameters is acquired. The contents of the task information are as follows: * Extended information * Current priority * Task state * Type of wait state * ID number of object to be processed for wait (semaphore, event flag, etc.) * Number of wake-up requests * Number of suspend requests * Key ID number User's Manual U13773EJ3V0UM 35 CHAPTER 3 TASK MANAGEMENT FUNCTION 3.7.2 Acquiring ID number The ID number of a task can be acquired by issuing the vget_tid system call. * vget_tid system call Acquires the ID number of the task specified by the parameter. To manipulate a task with a system call, the ID number of the task is necessary. Whether the ID number is determined univocally by the user or automatically assigned can be specified when a task is created. If automatic assignment of the ID number is specified, however, the user cannot learn the ID number of a task. To do so, a "key ID number" is necessary. The key ID number is univocally specified when a task is created. By issuing the vget_tid system call with this key ID number as a parameter, the ID number of the task having that key ID number can be acquired. 36 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS This chapter describes the synchronous communication functions performed by the RX850 Pro. 4.1 Overview In an environment where multiple tasks are executed concurrently (multitasking), a result produced by a preceding task may determine the next task to be executed or affect the processing performed by the subsequent task. In other words, some task execution conditions vary with the processing performed by another task, or the processing performed by some tasks is related. Therefore, liaison functions between tasks are required, so that task execution will be suspended to await the result output by another task or until necessary conditions have been established to enable the processing to be continued. In the RX850 Pro, these functions are called "synchronization functions." The synchronization functions include an exclusive control function and a wait function. The RX850 Pro provides semaphores that act as the exclusive control function and event flags that act as the wait function. For multitasking, an inter task communication function is also required to enable one task to receive the processing result from another. In the RX850 Pro, this function is called a "communication function". The RX850 Pro provides mailboxes that act as the communication function. 4.2 Semaphores Multitasking requires a function to prevent the resource contention that would occur when concurrently operating multiple tasks attempt to use a limited number of resources such as an A/D converter, coprocessors, files, and programs. To implement this contention preventive function, the RX850 Pro provides non-negative counter-type semaphores. The following system calls are used to dynamically manipulate a semaphore: cre_sem: del_sem: sig_sem: wai_sem: preq_sem: twai_sem: ref_sem: vget_sid: Remark Generates a semaphore Deletes a semaphore Returns a resource Acquires a resource Acquires a resource (by polling) Acquires a resource (with timeout setting) Acquires semaphore information Acquires semaphore ID number In other words, In RX850 Pro, those elements required to execute tasks are called resources. as well as software components such as files and programs. resources comprehensively refer to hardware components such as the A/D converter and coprocessor, User's Manual U13773EJ3V0UM 37 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.2.1 Generating semaphores The RX850 Pro provides two interfaces for generating semaphores. semaphore by issuing a system call from within a processing program. To generate a semaphore in the RX850 Pro, an area in system memory is allocated for managing that semaphore (as an object of management by the RX850 Pro), then initialized. (1) Static registration of a semaphore To statically register a semaphore, specify it during configuration. The RX850 Pro generates that semaphore according to the semaphore information defined in the information file (including system information tables and system information header files) during system initialization. The semaphore is subsequently managed by the RX850 Pro. (2) Dynamic registration of a semaphore To dynamically register a semaphore, issue the cre_sem system call from within a processing program (task). The RX850 Pro generates that semaphore according to the information specified with parameters when the cre_sem system call is issued. The semaphore is subsequently managed by the RX850 Pro. 4.2.2 Deleting semaphores A semaphore is deleted by issuing the del_sem system call. * del_sem system call The del_sem system call deletes the semaphore specified by the parameter. That semaphore is then no longer managed by the RX850 Pro. If a task is queued into the wait queue of the semaphore specified by this system call parameter, that task is removed from the wait queue, after which it leaves the wait state (the resource wait state) and enters the ready state. E_DLT is returned to the task released from the wait state as the value returned in response to the system call (wai_sem or twai_sem) that triggered the transition of the task to the wait state. 4.2.3 Returning resources A resource is returned by issuing the sig_sem system call. * sig_sem system call By issuing the sig_sem system call, the task returns a resource to the semaphore specified by the parameter (the semaphore counter is incremented by 0x1). If a task or tasks are queued into the wait queue of the semaphore specified by these system call parameter, the relevant resource is passed to the first task in the wait queue without being returned to the semaphore (thus, the semaphore counter is not incremented). That task is then removed from the wait queue, after which it either leaves the wait state (the resource wait state) and enters the ready state, or leaves the wait-suspend state and enters the suspend state. One enables the static generation of a The other dynamically generates a semaphore during system initialization (in the nucleus initialization section). 38 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.2.4 Acquiring resources A resource is acquired by issuing the wai_sem, preq_sem, or twai_sem system call. * wai_sem system call By issuing the wai_sem system call, the task acquires a resource from the semaphore specified by the parameter (the semaphore counter is decremented by 0x1.) After issuing this system call, if the task cannot acquire the resource from the specified semaphore (no idle resource exists), the task itself is queued into the wait queue of this semaphore. Thus, the task leaves the run state and enters the wait state (the resource wait state). The resource wait state is canceled in the following cases, and the task returns to the ready state. * When the sig_sem system call is issued. * When the del_sem system call is issued and the specified semaphore is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the wait queue of the specified semaphore, it is executed in the order (FIFO order or priority order) specified when that semaphore was generated (during configuration or when the cre_sem system call was issued). * preq_sem system call By issuing the preq_sem system call, the task acquires a resource from the semaphore specified by the parameter (the semaphore counter is decremented by 0x1.) After this system call is issued, if the task cannot acquire the resource from the specified semaphore (no idle resource exists), E_TMOUT is returned as the return value. * twai_sem system call By issuing the twai_sem system call, the task acquires a resource from the semaphore specified by the parameter (the semaphore counter is decremented by 0x1.) After issuing this system call, if the task cannot acquire the resource from the specified semaphore (no idle resource exists), the task itself is queued into the wait queue of this semaphore. Thus, the task leaves the run state and enters the wait state (the resource wait state). The resource wait state is canceled in the following cases, and the task returns to the ready state. * When the given wait time specified by a parameter has elapsed. * When the sig_sem system call is issued. * When the del_sem system call is issued and the specified semaphore is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the wait queue of the specified semaphore, it is executed in the order (FIFO order or priority order) specified when that semaphore was generated (during configuration or when the cre_sem system call was issued). User's Manual U13773EJ3V0UM 39 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.2.5 Acquiring semaphore information Semaphore information is acquired by issuing the ref_sem system call. * ref_sem system call By issuing the ref_sem system call, the task acquires the semaphore information (extended information, queued tasks, etc. ) for the semaphore specified by the parameter. The semaphore information consists of the following: * Extended information * Whether tasks are queued * The number of currently available resources * The maximum number of resources specified when the semaphore was generated * Key ID number 4.2.6 Acquiring ID number The ID number of a semaphore can be acquired by issuing the vget_sid system call. * vget_sid system call Acquires the ID number of a semaphore specified by the parameter. To manipulate a semaphore with a system call, the ID number of the semaphore is necessary. Whether the ID number is determined univocally by the user or automatically assigned can be specified when a task is created. If automatic assignment of the ID number is specified, however, the user cannot learn the ID number of a semaphore. To do so, a "key ID number" is necessary. The key ID number is univocally specified when a semaphore is created. By issuing the vget_sid system call with this key ID number as a parameter, the ID number of the semaphore having that key ID number can be acquired. 4.2.7 Exclusive control using semaphores The following is an example of using semaphores to manipulate the tasks under exclusive control. Conditions * Task priority Task A > Task B * State of tasks Task A: Run state Task B: Ready state * Semaphore attributes Number of resources initially assigned to the semaphore: Maximum number of resources that can be assigned to the semaphore: Task queuing order: 0x1 0x5 FIFO 40 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS (1) Task A issues the wai_sem system call. The number of resources assigned to this semaphore and managed by the RX850 Pro is 0x1. Thus, the RX850 Pro decrements the semaphore counter by 0x1. At this time, task A does not enter the wait state (the resource wait state). Instead, it remains in the run state. The relevant semaphore counter changes as shown in Figure 4-1. Figure 4-1. State of Semaphore Counter Before issuing the system call Number of resources: 0x1 Task A: wai_sem After issuing the system call Number of resources: 0x0 (2) Task A issues the wai_sem system call. The number of resources assigned to this semaphore and managed by the RX850 Pro is 0x0. Thus, the RX850 Pro changes the state of task A from run to the wait state (resource wait state) and places the task at the end of the wait queue for this semaphore. The wait queue of this semaphore changes as shown in Figure 4-2. Figure 4-2. State of Wait Queue (When wai_sem Is Issued) Before issuing the system call Wait queue Task A: wai_sem After issuing the system call Wait queue Task A (3) As task A enters the resource wait state, the state of task B changes from ready to run. (4) Task B issues the sig_sem system call. At this time, the state of task A that has been placed in the wait queue of this semaphore changes from the resource wait state to ready state. The wait queue of this semaphore changes as shown in Figure 4-3. Figure 4-3. State of Wait Queue (When sig_sem Is Issued) Before issuing the system call Wait queue Task B: sig_sem Task A After issuing the system call Wait queue User's Manual U13773EJ3V0UM 41 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS (5) The state of task A having the higher priority changes from ready to run. At the same time, task B leaves the run state and enters the ready state. Figure 4-4 shows the transition of exclusive control in steps (1) to (5). Figure 4-4. Exclusive Control Using Semaphores Task A Priority: High Task B Priority: Low wai_sem wai_sem sig_sem 42 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.3 Event Flags In multitasking, an intertask wait function, in which other tasks wait to resume execution of processing until the results of processing by a given task are output, is necessary. In such a case, it is good to have a function for other tasks to judge whether or not the "processing results output" event has occurred or not, and in the RX850 Pro, an event flag is provided in order to realize this kind of function. An event flag is a set of data consisting of 1-bit flags that indicate whether a particular event has occurred. 32-bit event flags are used in the RX850 Pro. 32 bits are handled as a set of information with each bit or a combination of bits having a specific meaning. The following system calls regarding event flags are used to dynamically manipulate an event flag. cre_flg: del_flg: set_flg: clr_flg: wai_flg: pol_flg: twai_flg: ref_flg: Generates an event flag. Deletes an event flag. Sets a bit pattern. Clears a bit pattern. Checks a bit pattern. Checks a bit pattern (by polling). Checks a bit pattern (with timeout setting). Acquires event flag information. vget_flg: Acquires event flag ID number. 4.3.1 Generating event flags The RX850 Pro provides two interfaces for generating event flags. One is for statically generating an event flag during system initialization (in the nucleus initialization section). The other is for dynamically generating an event flag by issuing a system call from within a processing program. To generate an event flag in the RX850 Pro, an area in system memory is allocated for managing that event flag (as an object of management by the RX850 Pro), then initialized. (1) Static registration of an event flag To statically register an event flag, specify it during configuration. The RX850 Pro generates that event flag according to the event flag information defined in the information file (including system information tables and system information header files) during system initialization. Subsequently, the event flag is managed by the RX850 Pro. (2) Dynamic registration of an event flag To dynamically register an event flag, issue the cre_flg system call from within a processing program (task). The RX850 Pro generates that event flag according to the information specified by a parameter when the cre_flg system call is issued. Subsequently, the event flag is managed by the RX850 Pro. User's Manual U13773EJ3V0UM 43 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.3.2 Deleting event flags An event flag is deleted by issuing a del_flg system call. * del_flg system call The del_flg system call deletes the event flag specified by the parameter. That event flag is then no longer managed by the RX850 Pro. If a task is queued into the wait queue of the event flag specified by this system call parameter, that task is removed from the wait queue, after which it leaves the wait state (the event flag wait state) and enters the ready state. E_DLT is returned to the task released from the wait state as the return value for the system call (wai_flg or twai_flg) that triggered the transition of the task to the wait state. 4.3.3 Setting a bit pattern The event flag bit pattern is set by issuing the set_flg system call. * set_flg system call The set_flg system call sets a bit pattern for the event flag specified by the parameter. When this system call is issued, if the given condition for a task queued into the wait queue of the specified event flag is satisfied, that task is removed from the wait queue. The task then either leaves the wait state (the event flag wait state) and enters the ready state, or leaves the wait-suspend state and enters the suspend state. 4.3.4 Clearing a bit pattern The event flag bit pattern is cleared by issuing the clr_flg system call. * clr_flg system call The clr_flg system call clears the bit pattern of the event flag specified by the parameter. Note that when this system call is issued, if the bit pattern of the specified event flag has already been cleared to zero, it is not regarded as an error. 4.3.5 Checking a bit pattern The event flag bit pattern is checked by issuing the wai_flg, pol_flg, or twai_flg system call. * wai_flg system call The wai_flg system call checks whether the bit pattern is set to satisfy the wait condition required for the event flag specified by the parameter. If the bit pattern does not satisfy the wait condition required this task is queued at the end of the wait queue of this event flag. Thus, the task leaves the run state and enters the wait state (the event flag wait state). The event flag wait state is canceled in the following cases, and the task returns to the ready state. * When the set_flg system call is issued and the required wait condition is set. * When the del_mbx system call is issued and this event flag is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. 44 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS * pol_flg system call The pol_flg system call checks whether the bit pattern is set to satisfy the wait condition required for the event flag specified by the parameter. If the bit pattern does not satisfy the wait condition required for the event flag specified by this system call parameter, E_TMOUT is returned as the return value. * twai_flg system call The twai_flg system call checks whether the bit pattern is set to satisfy the wait condition required for the event flag specified by the parameter. If the bit pattern does not satisfy the wait condition required for the event flag specified by this system call parameter, the task that issues this system call is queued at the end of the wait queue for this event flag. Thus, the task leaves the run state and enters the wait state (the event flag wait state). The event flag wait state is canceled in the following cases, and the task returns to the ready state. * Once the given wait time specified by the parameter has elapsed. * When the set_flg system call is issued and the required wait condition is set. * When the del_mbx system call is issued and this event flag is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Also, the event flag wait conditions and processing when the conditions are established can be specified as follows in the RX850 Pro. (1) Wait conditions * AND wait The wait state continues until all bits to be set to 1 in the required bit pattern have been set in the relevant event flag. * OR wait The wait state continues until any bit to be set to 1 in the required bit pattern has been set in the relevant event flag. (2) When the condition is satisfied * Clearing a bit pattern When the wait condition specified for the event flag is satisfied, the bit pattern for the event flag is cleared. 4.3.6 Acquiring event flag information Event flag information is acquired by issuing the ref_flg system call. * ref_flg system call By issuing the ref_flg system call, the task acquires the event flag information (extended information, queued tasks, etc.) for the event flag specified by the parameter. Details of event flag information are as follows: * Extended information * Whether tasks are queued * Current bit pattern * Key ID number User's Manual U13773EJ3V0UM 45 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.3.7 Acquiring ID number The ID number of an event flag can be acquired by issuing the vget_fid system call. * vget_fid system call Acquires the ID number of the event flag specified by the parameter. To manipulate an event flag with a system call, the ID number of the event flag is necessary. Whether the ID number is determined univocally by the user or automatically assigned can be specified when an event flag is created. If automatic assignment of the ID number is specified, however, the user cannot learn the ID number of an event flag. To do so, a "key ID number" is necessary. The key ID number is univocally specified when an event flag is created. By issuing the vget_fid system call with this key ID number as a parameter, the ID number of the event flag having that key ID number can be acquired. 4.3.8 Wait function using event flags The following is an example of manipulating the tasks under wait and control using event flags. Conditions * Task priority Task A > Task B * State of tasks Task A: Task B: Run state Ready state 0x0 One task * Event flag attributes Initial bit pattern: The number of tasks that can be placed in the wait queue: (1) Task A issues the wai_flg system call. The required bit pattern is 0x1 and the wait condition is TWF_ANDW|TWF_CLR. The current bit pattern of the relevant event flag managed by the RX850 Pro is 0x0. Thus, the RX850 Pro changes the state of task A from run to wait (the event flag wait state). Task A is then queued at the end of the wait queue for this event flag. The wait queue of this event flag changes as shown in Figure 4-5. Figure 4-5. State of Wait Queue (When wai_flg Is Issued) Before issuing the system call Wait queue Task A: wai_flg After issuing the system call Wait queue Task A 46 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS (2) As task A enters the event flag wait state, the state of task B changes from ready to run. (3) Task B issues the set_flg system call. The bit pattern is set to 0x1. This bit pattern satisfies the wait condition for task A that has been queued into the wait queue of the relevant event flag. Thus, task A leaves the event flag wait state and enters the ready state. Since TWF_CLR was specified when task A issued the wai_flag system call, the bit pattern of this event flag is cleared. The wait queue for this event flag changes as shown in Figure 4-6. Figure 4-6. State of Wait Queue (When set_flg Is Issued) Before issuing the system call Wait queue Task B: set_flg Task A After issuing the system call Wait queue (4) The state of task A having the higher priority changes from ready to run. At the same time, task B leaves the run state and enters the ready state. Figure 4-7 shows the transition of wait and control by event flags in steps (1) to (4). Figure 4-7. Wait and Control by Event Flags Task A Priority: High Task B Priority: Low wai_flg set_flg User's Manual U13773EJ3V0UM 47 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.4 Mailboxes Multitasking requires an inter task communication function, so that the tasks can be informed of the results output by other tasks. To implement this function, the RX850 Pro provides mailboxes. The mailboxes used in the RX850 Pro have two different queues, one dedicated to tasks and the other dedicated to messages. They can be used for both an inter task message communication function and an inter task wait function. The following mailbox-related system calls are used to dynamically operate a mailbox. cre_mbx: del_mbx: snd_msg: rcv_msg: prcv_msg: trcv_msg: ref_mbx: vget_mid: Generates a mailbox. Deletes a mailbox. Sends a message. Receives a message. Receives a message (by polling). Receives a message (with timeout setting). Acquires mailbox information Acquires mailbox ID number 4.4.1 Generating mailboxes The RX850 Pro provides two interfaces for generating mailboxes. One is for statically generating a mailbox during system initialization (in the nucleus initialization section). The other is for dynamically generating a mailbox by issuing a system call from within a processing program. To generate a mailbox in the RX850 Pro, an area in system memory is allocated for managing that mailbox (as an RX850 Pro management object), then initialized. (1) Static registration of a mailbox To statically register a mailbox, specify it during configuration. The RX850 Pro generates the mailbox according to the mailbox information defined in the information file (including system information tables and system information header files) during system initialization. Subsequently, the mailbox is managed by the RX850 Pro. (2) Dynamic registration of a mailbox To dynamically register a mailbox, issue the cre_mbx system call from within a processing program (task). The RX850 Pro generates the mailbox according to the information specified by the parameter when the cre_mbx system call is issued. Subsequently, the mailbox is managed by the RX850 Pro. 48 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.4.2 Deleting mailboxes A mailbox is deleted by issuing the del_mbx system call. * del_mbx system call The del_mbx system call deletes the mailbox specified by the parameter. That mailbox is then no longer managed by the RX850 Pro. If a task is queued into the task wait queue of the mailbox specified by this system call parameter, that task is removed from the task wait queue, after which it will leave the wait state (the message wait state) and enter the ready state. E_DLT is returned to the task released from the wait state as the return value for the system call (rcv_msg or trcv_msg) that triggered the transition of the task to the wait state. If a message is queued to the message wait queue of the specified mailbox when this system call is issued, the message is released from the wait queue and is returned to the memory pool from which the message area is acquired. Consequently, if an area other than the memory block acquired from a memory pool is used as a message area, the operation of deleting a mailbox is not guaranteed. Be sure to use a memory block acquired from a memory pool as the message area for the mailbox that may be deleted by this system call. 4.4.3 Transmitting a message A message is transmitted from the task by issuing the snd_msg system call. * snd_msg system call Upon the issuance of the snd_msg system call, the task transmits a message to the mailbox specified by the parameter. If a task or tasks are queued into the task wait queue of the mailbox specified by this system call parameter, the message is delivered to the first task in the task wait queue without being queued into the mailbox. The first task is then removed from the wait queue, after which it either leaves the wait state (the message wait state) and enters the ready state, or leaves the wait-suspend state and enters the suspend state. If no tasks are queued in the task wait queue of the object mailbox, the message is placed in the message wait queue of the object mailbox. Remark When a message queues into the message wait queue of the specified mailbox, it is executed in the order (FIFO order or priority order) specified when the mailbox was generated (during configuration or when the cre_mbx system call was issued). User's Manual U13773EJ3V0UM 49 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.4.4 Receiving a message A message is received by the task upon the issuance of the rcv_msg, prcv_msg, or trcv_msg system call. * rcv_msg system call Upon the issuance of the rcv_msg system call, the task receives a message from the mailbox specified by the parameter. If the task cannot receive a message from the mailbox specified by this system call parameter (no message exists in the message wait queue of that mailbox), the task that issued this system call is queued at the end of the task wait queue for this mailbox. Thus, the task leaves the run state and enters the wait state (the message wait state). The message wait state is canceled in the following cases and the task returns to the ready state. * When the snd_msg system call is issued. * When the del_mbx system call is issued and this mailbox is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the task wait queue of the specified mailbox, it is executed in the order (FIFO order or priority order) specified when that mailbox was generated (during configuration or when a cre_mbx system call was issued). * prcv_msg system call Upon the issuance of the prcv_msg system call, the task receives a message from the mailbox specified by the parameter. If the task cannot receive a message from the mailbox specified by this system call parameter (no message exists in the message wait queue for that mailbox), E_TMOUT is returned as the return value. * trcv_msg system call Upon the issue of the trcv_msg system call, the task receives a message from the mailbox specified by the parameter. If the task cannot receive a message from the mailbox specified by this system call parameter (no message exists in the message wait queue for that mailbox), the task that issued this system call is queued at the end of the task wait queue for this mailbox. Thus, the task leaves the run state and enters the wait state (the message wait state). The message wait state is canceled in the following cases and the task returns to the ready state. * When the given time specified by the parameter has elapsed. * When the snd_msg system call is issued. * When the del_mbx system call is issued and this mailbox is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the task wait queue of the specified mailbox, it is executed in the order (FIFO order or priority order) specified when that mailbox was generated (during configuration or when the cre_mbx system call was issued). 50 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.4.5 Messages In the RX850 Pro, all items of information exchanged between tasks, via mailboxes, are called "messages." Messages can be transmitted to an arbitrary task via a mailbox. In inter task communication in the RX850 Pro, however, only the start address of a message is delivered to a receiving task, enabling the task to access the message. The contents of the message are not copied to any other area. (1) Allocating message areas NEC Electronics recommends that the memory pool managed by the RX850 Pro be allocated for messages. To make a memory pool area available for a message, the task should issue the get_blk, pget_blk, or tget_blk system call. The first four bytes of each message are used as the block for linkage to the message wait queue when queued. Therefore, save messages after the first four bytes of the message area. (2) Composition of messages RX850 Pro does not prescribe the length and composition of messages to be transmitted to mailboxes. The message length, except for the first four bytes, and its composition are defined by the tasks that communicate with each other via mailboxes. Caution The RX850 Pro prescribes that the first four bytes of each message are used as the block for linkage to the message wait queue when queued. For this reason, when a message is transmitted to the relevant mailbox, the first four bytes of the message must be set to 0x0 before the snd_msg system call is issued. If the first four bytes of the message are set to a value other than 0x0 when the snd_msg system call is issued, the RX850 Pro determines that this message has already been queued into the message wait queue. Thus, the RX850 Pro does not send the message to the mailbox and returns E_OBJ as the return value. (3) Priority of messages The RX850 Pro can specify the priority according to which a message is to be queued. To specify the priority of a message, two bytes are necessary in addition to the four bytes of the link area that is used to queue the message to the message wait queue. Therefore, store the message in an area six bytes after the beginning of the message area. The message priority is specified by a positive integer of 1 to 0x7fff. The lower the value, the higher the priority. 4.4.6 Acquiring mailbox information Mailbox information is acquired by issuing the ref_mbx system call. * ref_mbx system call Upon the issuance of a ref_mbx system call, the task acquires the mailbox information (extended information, queued tasks, etc.) for the mailbox specified by the parameter. The mailbox information consists of the following: * Extended information * Whether tasks are queued * Whether messages are queued * Key ID number User's Manual U13773EJ3V0UM 51 CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS 4.4.7 Acquiring ID number The ID number of a mailbox can be acquired by issuing the vget_mid system call. * vget_mid system call Acquires the ID number of a mailbox specified by the parameter. To manipulate a mailbox with a system call, the ID number of the mailbox is necessary. Whether the ID number is determined univocally by the user or automatically assigned can be specified when a mailbox is created. If automatic assignment of the ID number is specified, however, the user cannot learn the ID number of a mailbox. To do so, a "key ID number" is necessary. The key ID number is univocally specified when a mailbox is created. By issuing the vget_mid system call with this key ID number as a parameter, the ID number of the mailbox having that key ID number can be acquired. 4.4.8 Inter task communication using mailboxes The following is an example of manipulating the tasks in inter task communication using mailboxes. Conditions * Task priority Task A > Task B * State of tasks Task A: Run state Task B: Ready state * Mailbox attributes Task queuing order: Message queuing order: FIFO FIFO (1) Task A issues a rcv_msg system call. No message is queued in the message wait queue of the relevant mailbox managed by the RX850 Pro. Thus, the RX850 Pro changes the state of task A from run to wait (the message wait state). The task is queued at the end of the task wait queue for this mailbox. The task wait queue for this mailbox changes as shown in Figure 4-8. Figure 4-8. State of Task Wait Queue (When rcv_msg Is Issued) Before issuing the system call Wait queue Task A: rcv_msg After issuing the system call Wait queue Task A (2) As task A enters the message wait state, the state of task B changes from ready to run. (3) Task B issues the get_blk system call. By means of this system call, a memory pool area is allocated for a message (as a memory block). 52 User's Manual U13773EJ3V0UM CHAPTER 4 SYNCHRONOUS COMMUNICATION FUNCTIONS (4) Task B writes a message into this memory block. (5) Task B issues the snd_msg system call. This changes the state of task A that has been placed in the task wait for the relevant mailbox from the message wait state to ready state. The task wait queue for this mailbox changes as shown in Figure 4-9. Figure 4-9. State of Task Wait Queue (When snd_msg Is Issued) Before issuing the system call Wait queue Task B: snd_msg Task A After issuing the system call Wait queue (6) The state of task A having the higher priority changes from ready to run. At the same time, task B leaves the run state and enters the ready state. (7) Task A issues the rel_blk system call. This releases the memory block allocated for the message in the memory pool. The flow of communications between tasks as explained in (1) to (7) is shown in Figure 4-10. Figure 4-10. Inter-Task Communication Using Mailboxes Task A Priority: High Task B Priority: Low rcv_msg get_blk A message is generated snd_msg rel_blk User's Manual U13773EJ3V0UM 53 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION This chapter describes the interrupt management function provided by the RX850 Pro. 5.1 Overview The RX850 Pro interrupt management function enables the following: * Registration of an interrupt handler * Activation of an interrupt handler * Return from an interrupt handler * Change or acquisition of the interrupt enable level 5.2 Interrupt Handler An interrupt handler is a routine dedicated to interrupt processing. Upon the occurrence of an interrupt, the interrupt handler is initiated immediately and handled independently of all other tasks. Therefore, if a task having the highest priority in the system is being executed upon the occurrence of an interrupt, its processing is suspended and control is passed to the interrupt handler. The RX850 Pro supports the following two interrupt handler interfaces considering the response from the occurrence of interrupts to the activation of interrupt handler. * Directly activated interrupt handler A routine dedicated to interrupt processing activated without the RX850 Pro. * Indirectly activated interrupt handler A routine dedicated to interrupt processing activated upon the completion of the interrupt preprocessing by the RX850 Pro (such as saving the contents of the registers or switching the stack). If a system call is issued while the interrupt handler is performing processing, scheduling is performed in a way specific to the RX850 Pro. That is, if a system call (chg_pri, sig_sem, etc.) that requires task scheduling is issued during processing by the interrupt handler, the RX850 Pro merely queues the tasks into the wait queue. system call and return instruction). The actual processing of task scheduling is batched and deferred until a return from the interrupt handler has been made (by issuing the ret_int 54 User's Manual U13773EJ3V0UM CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION 5.3 Directly Activated Interrupt Handler A directly activated interrupt handler is a routine dedicated to interrupt processing without using the RX850 Pro upon the occurrence of an interrupt. Accordingly, a high-speed response close to the hardware limitation is expected. The flow of the interrupt handler's operation is shown in Figure 5-1. Figure 5-1. Flow of Processing Performed by Directly Activated Interrupt Handler Task Occurrence of interrupt RX850 Pro Directly activated interrupt handler ret_int Scheduling processing 5.3.1 Registering directly activated interrupt handler The directly activated interrupt handler is registered by allocating the handler to the handler address to which the processor transfers control if an interrupt occurs, or by setting a branch instruction that branches execution to the directly activated interrupt handler. For details, refer to A.5 Directly Activated Interrupt Handler. 5.3.2 Processing in directly activated interrupt handler When describing the processing to be performed by the directly activated interrupt handler, note the following: (1) Saving/restoring the registers The contents of the work registers when control is transferred to the directly activated interrupt handler are the same as when an interrupt occurred. To use the work registers in the directly activated interrupt handler, therefore, the contents of the work registers must be saved at the start of the handler and restored at the end. To mitigate the workload of the user when describing saving and restoring the work registers in an assembly language, the RX850 Pro provides a macro for the directly activated interrupt handler. This macro helps the user to describe the saving and restoring of the work registers easily. (2) Stack switching The stack when control is transferred to the directly activated interrupt handler is the same as the stack when an interrupt occurred. To use the stack for the interrupt handler, therefore, it is necessary to switch the task between that for the handler and that for ordinary purposes at the beginning and end of the indirectly activated interrupt handler. If a macro for the directly activated interrupt handler is used, the stack is switched over between that for the interrupt handler and that for ordinary purposes in the macro. User's Manual U13773EJ3V0UM 55 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION (3) Limitations imposed on system calls The following lists the system calls that can be issued during the processing of a directly activated interrupt handler: * Task management system calls sta_tsk ref_tsk chg_pri vget_tid rot_rdq rel_wai get_tid * Task-associated synchronization system calls sus_tsk rsm_tsk frsm_tsk wup_tsk can_wup * Synchronous communication system calls sig_sem clr_flg prcv_msg preq_sem pol_flg ref_mbx ref_sem ret_flg vget_mid vget_sid vget_fid set_flg snd_msg * Interrupt management system calls def_int chg_icr ret_int ref_icr ret_wup ena_int dis_int * Memory pool management system calls pget_blk rel_blk ref_mpl vget_pid * Time management system call set_tim get_tim def_cyc act_cyc ref_cyc * System management system calls get_ver ref_sys def_svc viss_svc 56 User's Manual U13773EJ3V0UM CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION (4) Return processing from the directly activated interrupt handler Return processing from the directly activated interrupt handler is performed by issuing the ret_int or ret_wup system call upon the completion of interrupt handler operation. * ret_int system call Performs return from the directly activated interrupt handler. * ret_wup system call Issues a wake-up request to the task specified by the parameters, then returns from the directly activated interrupt handler. When a system call (chg_pri, sig_sem, etc.) that requires task scheduling is issued during processing by the directly activated interrupt handler, the RX850 Pro merely queues the tasks into the wait queue. The actual processing of task scheduling is batched and deferred until return from the directly activated interrupt handler has been made (by issuing the ret_int or ret_wup system call). Cautions 1. The ret_int and ret_wup system calls do not notify the external interrupt controllers that operation of the interrupt handler has terminated (the EOI command is not issued). Therefore, if a return is made from the directly activated interrupt handler that was initiated by an external interrupt request, notification of the termination of interrupt handler operation must be posted to the external interrupt controller before these system calls are issued. 2. The RX850 Pro provides a macro that can be used when a directly activated interrupt handler is described. It is recommended that this macro is also used when execution returns from the handler. In this macro, the necessary registers are restored and the ret_int and ret_wup system calls are issued. To restore the directly activated interrupt handler when this macro is used, therefore, it is not necessary to issue these system calls. For details, see APPENDIX A PROGRAMMING METHODS. 3. The values of the GP (global pointer) and TP (text pointer) used by the directly activated interrupt handler become undefined. Therefore, the values of the GP and TP should be set again at the start of the directly activated interrupt handler (after macro description). Since the restoring processing is executed by the RX850 Pro, it is not necessary to be described. For details of the description method, refer to A.5 Directly Activated Interrupt Handler. User's Manual U13773EJ3V0UM 57 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION 5.4 Indirectly Activated Interrupt Handler The indirectly activated interrupt handler is an interrupt processing routine that is activated after the interrupt preprocessing of the RX850 Pro (such as saving the registers and switching the stack) has been performed if an interrupt occurs. This interrupt handler is inferior to the directly activated interrupt handler in terms of response speed. However, because interrupt preprocessing of the RX850 Pro is performed, the processing in the handler is simplified. Figure 5-2 shows the flow of the operation of the indirectly activated interrupt handler. Figure 5-2. Operation Flow of Indirectly Activated Interrupt Handler Task Occurrence of interrupt RX850 Pro Indirectly activated interrupt handler Interrupt preprocessing return (TSK_NULL) Interrupt post processing Scheduling processing 5.4.1 Registering indirectly activated interrupt handler The RX850 Pro has two types of interfaces for registering an indirectly activated interrupt handler: "statically register the handler by system initialization (nucleus initialization block)" and "dynamically register the handler by issuing a system call from the processing program". Registration of an indirectly activated interrupt handler with the RX850 Pro means allocating an area that manages the indirectly activated interrupt handler (management object) from the system memory and initializing this area. (1) Static registration To statically register an indirectly activated interrupt handler, specify so when configuration is executed. The RX850 Pro registers and manages the indirectly activated interrupt handler based on the information defined in the information files (system information table and system information header file) when system initialization is performed. (2) Dynamic registration To dynamically register an indirectly activated interrupt handler, issue the def_int system call from the processing program (task or non-task). The RX850 Pro registers and manages the indirectly activated interrupt handler based on the information specified by the parameter when the def_int system call is issued. 58 User's Manual U13773EJ3V0UM CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION 5.4.2 Processing in indirectly activated interrupt handler Keep in mind the following points when describing the processing of an indirectly activated interrupt handler. (1) Saving and restoring registers The RX850 Pro saves and restores the contents of the work registers in compliance with the function calling convention of the C compiler (CA850 or CCV850) when it transfers control to an indirectly activated interrupt handler or when execution returns from the handler. It is therefore not necessary to save the work registers at the beginning of the indirectly activated interrupt handler and to restore the registers at the end. (2) Switching stack The RX850 Pro switches the stack when it transfers control to an indirectly activated interrupt handler and when execution returns from the handler. It is therefore not necessary to switch the task between that for the handler and that for ordinary purposes at the beginning and end of the indirectly activated interrupt handler. If the stack for handler is not defined when configuration is performed, however, the stack is not switched, and the stack for ordinary purposes is used. (3) Issuing system calls Here is a list of the system calls that can be issued in the indirectly activated interrupt handler. * Task management system calls sta_tsk rer_tsk chg_pri vget_tid rot_rdq rel_wai get_tid * Task-associated synchronization system calls sus_tsk rsm_tsk frsm_tsk wup_tsk can_wup * Synchronous communication system calls sig_sem clr_flg prcv_msg preq_sem pol_flg ref_mbx ref_sem ref_flg vget_mid vget_sid vget_fid set_flg snd_msg * Interrupt management system calls def_int ena_int dis_int ref_icr chg_icr * Memory pool management system calls pget_blk rel_blk ref_mpl vget_pid * Time management system calls set_tim get_tim def_cyc act_cyc ref_cyc * System management system calls get_ver ref_sys def_svc viss_svc User's Manual U13773EJ3V0UM 59 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION (4) Return processing from indirectly activated interrupt handler To exit an indirectly activated interrupt handler, issue the return instruction at the end of the handler. * return (TSK_NULL) instruction Performs return from the indirectly activated interrupt handler. * return (ID tskid) instruction Issues a wake-up request to the task specified by the parameters then returns from the indirectly activated interrupt handler. The RX850 Pro only manipulates the queues if a system call requiring scheduling of a task (such as chg_pri and sig_sem) is issued in an indirectly activated interrupt handler. The actual scheduling is postponed until execution returns from the indirectly activated interrupt handler, and is then performed all at once. Caution The return instruction does not notify an external interrupt controller of the end of processing (by issuing the EOI command). To exit from an indirectly activated interrupt handler that has been activated by an external interrupt request, therefore, notify the external interrupt controller of the end of the processing before issuing these system calls. 60 User's Manual U13773EJ3V0UM CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION 5.5 Disabling/Resuming Maskable Interrupt Acknowledgement RX850 Pro provides a function for disabling or resuming the acknowledgement of maskable interrupts, so that whether maskable interrupts are acknowledged can be specified from a user processing program. This function is used by issuing the following system calls from within a task or interrupt handler. * loc_cpu system call The loc_cpu system call disables the acknowledgement of maskable interrupts, as well as the performing of dispatch processing (task scheduling). Once this system call has been issued, control is not passed to any other task or interrupt handler until the unl_cpu system call is issued. * unl_cpu system call The issue of the unl_cpu system call enables the acknowledgement of maskable interrupts, and resuming dispatch processing (task scheduling). This system call enables the acknowledgement of maskable interrupts which is disabled by the loc_cpu system call and resumes dispatch processing. Figure 5-3 shows the flow of control if an interrupt is not masked (normal) and Figure 5-4 shows the flow of control if the loc_cpu system call is issued. Figure 5-3. Control Flow if Interrupt Mask Processing Is Not Performed (Normal) Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk Occurrence of interrupt ret_int wup_tsk slp_tsk User's Manual U13773EJ3V0UM 61 CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION Figure 5-4. Control Flow if loc_cpu System Call Is Issued Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk loc_cpu Occurrence of interrupt wup_tsk unl_cpu ret_int slp_tsk 5.6 Changing/Acquiring Interrupt Control Register The interrupt control register is changed or acquired by issuing the chg_icr or ref_icr system call. * chg_icr system call This system call changes the interrupt control register specified by the parameter. * ref_icr system call This system call is available for acquiring the interrupt control register specified by the parameter. Caution When the RX850 Pro is operated on the V850E core, if the interrupt control register-related system calls chg_icr and ref_icr are issued, the desired interrupt control register may not be operated. In the RX850 Pro, the interrupt control register address is calculated from the interrupt source number. However, in the V850E core, the correct register address cannot be obtained since the alignment of the interrupt source numbers and interrupt control registers differs from other V850 Series products. chg_icr and ref_icr is restricted. Therefore, use of the system calls For manipulating the interrupt control register via an application, directly manipulate the register without using these system calls. 62 User's Manual U13773EJ3V0UM CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION 5.7 Non-Maskable Interrupts A non-maskable interrupt is not subject to management based on interrupt priority and has priority over all other interrupts. It can be acknowledged even if the processor is placed in the interrupt disabled state (setting the ID flag of PSW). Therefore, even while processing is being executed by the RX850 Pro or an interrupt handler, a non-maskable interrupt can be acknowledged. If a system call is issued during the processing of an interrupt handler that supports non-maskable interrupts, its operation cannot be assured in the RX850 Pro. 5.8 Clock Interrupts In the RX850 Pro, time management is performed using clock interrupts, which can be generated periodically by hardware (clock controller, etc.). If a clock interrupt is issued, RX850 Pro system clock processing is called and the processing related to time, such as the timeout wait of a task or the activation of the cyclically activated handler, is performed. For details about the time management, see CHAPTER 7 TIME MANAGEMENT FUNCTION. 5.9 Multiple Interrupts The occurrence of another interrupt while processing is being performed by an interrupt handler is called "multiple interrupts". The RX850 Pro also responds to multiple interrupts. All interrupt handlers, however, start their operation in the interrupt-disabled state (setting the ID flag of PSW). To enable the acknowledgement of multiple interrupts, the canceling of the interrupt disabled state should be described in the interrupt handler. Figure 5-5 shows the flow of the processing for handling multiple interrupts. Figure 5-5. Processing Flow for Handling Multiple Interrupts Occurrence of interrupt Task RX850 Pro Directly activated interrupt handler Directly activated interrupt handler Interrupt-disabled state Interrupt-disabled state canceled Occurrence of interrupt Interrupt-enabled state Interrupt-disabled state ret_int Interrupt-enabled state ret_int Scheduling processing User's Manual U13773EJ3V0UM 63 CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION This chapter describes the memory pool management function of the RX850 Pro. 6.1 Overview The information table to manage systems, memory areas where the management blocks for implementing functions are allocated, and memory areas to use the memory pool are required for the RX850 Pro. The items above are allocated in the following four types of memory areas. * System Memory Pool 0 * System Memory Pool 1 * User Memory Pool 0 * User Memory Pool 1 (Keyword: SPOL0) (Keyword: SPOL1) (Keyword: UPOL0) (Keyword: UPOL1) The resource management table, task stack, interrupt handler stack, and memory for memory pool are allocated in the above memory areas. The combination of allocatable areas is as follows. Table 6-1. Memory Information Allocation Combination Resource Management Table SPOL0 Task Stack SPOL0 or SPOL1 Interrupt Stack SPOL0 or SPOL1 Memory Pool UPOL0 or UPOL1 The start address and size of each memory area are set using the configuration file. SPOL0 must be created. SPOL1 needs to be created when the task stack and interrupt stack are to be allocated in other than SPOL0. UPOL0 and UPOL1 need to be created if the memory pool management function is to be used. In addition, UPOL1 can be created if UPOL0 has already been created. 64 User's Manual U13773EJ3V0UM CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION 6.2 Management Objects These The objects required for implementing the functions provided by the RX850 Pro are listed below. management objects are generated and initialized during system initialization, according to the information specified at configuration. These management objects are allocated to SPOL0 (SPOL1 also is available for task stacks and interrupt handler stacks). * System base table * Task management block * Semaphore management block * Event flag management block * Mailbox management block * Memory pool management block * Memory block management block * Cyclically activated handler management block * Extended SVC handler management block * Memory pool * Task stack * Interrupt handler stack * Interrupt handler management block Figure 6-1 shows a typical arrangement of the management objects. Figure 6-1. Typical Arrangement of Management Objects High address Task stack Interrupt handler management block Extended SVC handler management block Cyclically activated handler management block Memory pool Mailbox management block Event flag management block Semaphore management block Task management block Interrupt handler stack System base table Low address User's Manual U13773EJ3V0UM 65 CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION 6.3 Memory Pool and Memory Blocks The RX850 Pro executes a dynamic memory pool management function through which memory areas are acquired and released during application. Using this function, the memory area is acquired if required for working, and if it becomes unnecessary, the memory area is released. This function enables efficient use of limited memory area. The memory area that can be used as a memory pool is UPOL0 or UPOL1. Specify which area of the UPOL0 or UPOL1 is to be used when the memory pool is defined at configuration, or when the memory pool is created by issuing a system call (cre_mpl). The RX850 Pro provides a variable-length memory pool, but not a fixed-length memory pool. The memory pool consists of memory blocks and is allocated in units of memory blocks. Dynamic generation of a memory pool and access to the memory pool are performed using the following memory pool-related system calls: cre_mpl: del_mpl: get_blk: pget_blk: tget_blk: rel_blk: ref_mpl: vget_pid: Generates a memory pool. Deletes the memory pool. Acquires a memory block. Acquires a memory block (by polling). Acquires a memory block (with timeout setting). Release a memory block. Acquires memory pool information. Acquires ID information of the memory pool. 6.3.1 Generating a memory pool The RX850 Pro provides two interfaces for generating (registering) a memory pool. One enables static generation during system initialization (in the nucleus initialization section). The other enables dynamic generation by issuing a system call from within a processing program. To generate a memory pool in the RX850 Pro, certain areas in system memory are allocated to enable management of the memory pool (as an object of RX850 Pro management) and for the memory pool main body, then initialized. (1) Static registration of a memory pool To register a memory pool statically, specify it during configuration. The RX850 Pro generates the memory pool, based on the information defined in the information file (including system information tables and system information header files) during system initialization. The memory pool is subsequently managed by the RX850 Pro. (2) Dynamic registration of a memory pool To dynamically register a memory pool, issue the cre_mpl system call from within a processing program (task). The RX850 Pro generates the memory pool, according to the information specified by the parameters when the cre_mpl system call is issued. The memory pool is subsequently managed by the RX850 Pro. Remark When a memory pool is created, the RX850 Pro uses the first 8 bytes of the memory pool as a memory pool management area, in addition to the specified size of memory. Therefore, the size of the created memory pool is "specified size + 8 bytes". 66 User's Manual U13773EJ3V0UM CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION 6.3.2 Deleting a memory pool A memory pool is deleted upon the issuance of the del_mpl system call. * del_mpl system call The del_mpl system call deletes the memory pool specified by the parameter. Subsequently, that memory pool is no longer subject to management by the RX850 Pro. If a task is queued into the wait queue of the memory pool specified by this system call parameter, that task is removed from the wait queue, leaves the wait state (the memory block wait state) and enters the ready state. E_DLT is returned to the task released from the wait state as the return value for the system call (get_blk or tget_blk) that triggered the transition of the task to the wait state. If this system call is issued, the RX850 Pro excludes the memory block managed by the specified memory pool from management. If the task has already acquired a memory block from the memory pool before this system call is issued, the operation of the memory block is not guaranteed, and care must be exercised in deleting a memory pool. 6.3.3 Acquiring a memory block A memory block is acquired by issuing a get_blk, pget_blk, or tget_blk system call. Caution In the RX850 Pro, memory clear is not performed when a memory block is acquired. Therefore, the acquired memory block's contents are undefined. When a memory block is acquired, the RX850 Pro uses eight bytes of the memory pool as a memory management area, in addition to the requested size of memory. The RX850 Pro also aligns the requested size by four bytes. Check the remaining memory block size. The size of the acquired memory block can be calculated by this expression: Size of memory block (blk_siz) = align 4 (user requested size) + 8 * get_blk system call Upon the issuance of the get_blk system call, the processing program (task) acquires a memory block from the memory pool specified by the parameter. After the issue of this system call, if the task cannot acquire the block from the specified memory pool (because no free block of the required size exists), the task itself is queued in the wait queue of this memory pool. Thus, the task leaves the run state and enters the wait state (the memory block wait state). The memory block wait state is canceled in the following cases and the task returns to the ready state. * When the rel_blk system call is issued and a memory block of the required size is returned. * When the del_mpl system call is issued and the specified memory pool is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the wait queue of the specified memory pool, it is executed in the order (FIFO order or priority order) specified when that memory pool was generated (during configuration or when a cre_mpl system call was issued). * pget_blk system call Upon the issuance of the pget_blk system call, the processing program (task) acquires a memory block from the memory pool specified by a parameter. For this system call, if the task cannot acquire the block from the memory pool specified by this system call parameter (because no free block of the required size exists), E_TMOUT is returned as the return value. User's Manual U13773EJ3V0UM 67 CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION * tget_blk system call By issuing a tget_blk system call, the processing program (task) acquires a memory block from the memory pool specified by a parameter. After the issue of this system call, if the task cannot acquire the block from the specified memory pool (because no free block of the required size exists), the task itself is queued into the wait queue of this memory pool. Thus, the task leaves the run state and enters the wait state (the memory block wait state). The memory block wait state is canceled in the following cases and the task returns to the ready state. * When the wait time specified by the parameter has elapsed. * When the rel_blk system call is issued and a memory block of the required size is returned. * When the del_mpl system call is issued and the specified memory pool is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. Remark When a task queues in the wait queue of the specified memory pool, it is executed in the order (FIFO order or priority order) specified when that memory pool was generated (during configuration or when the cre_mpl system call was issued). 6.3.4 Returning a memory block A memory block is returned upon the issuance of the rel_blk system call. * rel_blk system call Upon the issuance of the rel_blk system call, a processing program (task) returns a memory block to the memory pool specified by the parameter. For this system, if the memory block returned by this system call is of the size required by the first task in the wait queue of the specified memory pool, this block is passed to that task. Thus, the first task is removed from the wait queue, leaves the wait state (the memory block wait state), and enters the ready state, or leaves the wait-suspend state and enters the suspend state. Cautions 1. The contents of a returned memory block are not cleared by the RX850 Pro. Thus, the contents of a memory block may be undefined when that memory block is returned. 2. The RX850 Pro includes two different specifications for the rel_blk system call. (1) When a memory block is returned by a rel_blk system call, if the first four bytes of the memory block are not filled with zeros, the return value E_OBJ is used for termination instead of returning the memory block. (2) When the rel_blk system call is issued, the memory block is returned even if the first four bytes of the memory block are not filled with zeros (return value = E_OK). The first specification applies when the memory block is used as a mailbox's message area, and this is the specification that has been used for the rel_blk system call as it has been implemented thus far in the RX850 Pro. 68 User's Manual U13773EJ3V0UM CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION When the memory block is used as a mailbox's message area, the first four bytes serve as the link area for the message's wait queue. In other words, if messages are queued in the mailbox, when the rel_blk system call is issued and the memory block must be returned, in which case it is the message area linked to the queue that is returned. To prevent this, the specification requires the first four bytes that comprise the link area to be filled with zeros, otherwise it will be recognized as the memory block used as the message area and the return value E_OBJ will be used for termination instead of returning the memory block. Under this specification, the first four bytes must be cleared to zero in order to use rel_blk to return the memory block. These specifications of rel_blk are stored in separate libraries so that one or the other rel_blk specification can be used. Link to the library of the rel_blk specification to be used. (1) Library containing rel_blk that requires zero-clearing of first four bytes of memory block librxp.a (2) Library containing rel_blk that does not require zero-clearing of first four bytes of memory block librxpm.a 3. Treat a memory pool that returns a memory block the same as a memory pool specified when issuing the get_blk, pget_blk, or tget_blk system call. 6.3.5 Acquiring memory pool information Memory pool information is acquired by issuing the ref_mpl system call. * ref_mpl system call Upon the issuance of the ref_mpl system call, the processing program (task) acquires the memory pool information (extended information, queued tasks, etc.) for the memory pool specified by the parameter. The memory pool information consists of the following: * Extended information * Whether tasks are queued * Total amount of free space * The maximum memory block size to be acquired * Key ID number User's Manual U13773EJ3V0UM 69 CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION 6.3.6 Acquiring ID number The ID number of a memory pool can be acquired by issuing the vget_pid system call. * vget_pid system call Acquires the ID number of a memory pool specified by the parameter. To manipulate a memory pool with a system call, the ID number of the memory pool is necessary. Whether the ID number is determined univocally by the user or automatically assigned can be specified when a memory pool is created. If automatic assignment of the ID number is specified, however, the user cannot learn the ID number of a memory pool. To do so, a "key ID number" is necessary. The key ID number is univocally specified when a memory pool is created. By issuing the vget_pid system call with this key ID number as the parameter, the ID number of the memory pool having that key ID number can be acquired. 6.3.7 Dynamic management of memory block by memory pool Here is an example of an operation to dynamically use the memory for tasks by using a memory pool. Conditions * Task priority Task A > Task B * State of tasks Task A: Run state Task B: Ready state * Memory pool attributes Vacant memory block size: Task queuing order: 0x20 FIFO (1) Task A issues the get_blk system call. The requested memory block size is "0x30". At present, the vacant memory block size of the memory pool under management of the RX850 Pro is "0x20". Therefore, the RX850 Pro changes the state of task A from run to wait (waiting for a memory block), and queues the task to the end of the wait queue of tasks waiting for a memory pool. At this time, this wait queue is in the state as shown in Figure 6-2. Figure 6-2. State of Wait Queue (When get_blk Is Issued) Before issuing the system call Wait queue Task A: get_blk After issuing the system call Wait queue Task A 70 User's Manual U13773EJ3V0UM CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION (2) As the state of task A changes from run to wait, the state of task B changes from ready to run. (3) Task B issues the rel_blk system call. The returned memory block size is "0x16". As a result, the requested memory block size of task A queued waiting for a memory pool is satisfied and task A changes its state from wait to ready. At this time the wait queue of tasks waiting for a memory pool is as shown in Figure 6-3. Figure 6-3. State of Wait Queue (When rel_blk Is Issued) Before issuing the system call Wait queue Task B: rel_blk Task A After issuing the system call Wait queue (4) The task A with the higher priority changes its state from ready to run. Task B changes its state from run to ready. Figure 6-4 shows the flow of dynamic use of memory by the memory pool explained in (1) through (4) above. Figure 6-4. Dynamic Use of Memory by Memory Pool Task A Priority: High Task B Priority: Low get_blk rel_blk User's Manual U13773EJ3V0UM 71 CHAPTER 7 TIME MANAGEMENT FUNCTION This chapter describes the time management function of the RX850 Pro. 7.1 Overview Time management in the RX850 Pro is performed using clock interrupts which can be generated periodically by hardware (clock controller, etc.). If a clock interrupt is issued, the RX850 Pro system clock processing is called and system clock update as well as processing related to time, such as delayed task wake-up, timeout, and starting of the cyclically activated handler, is executed. 7.2 System Clock The system clock is a software timer that provides the time (in units of milliseconds, with a width of 48 bits) used for time management by the RX850 Pro. The system clock is set to 0x0 at system initialization and updated in units of the basic clock cycle (specified at configuration) each time system clock processing is performed. Caution The system clock managed by the RX850 Pro shall have a fixed width of 48 bits. The RX850 Pro ignores any overflow (that exceeding 48 bits) for the clock value. 7.2.1 Setting and reading the system clock The system clock setting is executed by issuing the set_tim system call, and reading by issuing the get_tim system call. * set_tim system call The set_tim system call sets the time specified by the parameter to the system clock. * get_tim system call The get_tim system call stores the current time of the system clock into the packet specified by the parameter. 7.3 Timer Operations Real-time processing requires functions synchronized with time (timer operation functions) such as stopping the processing of a certain task for a specific time and executing the processing of a handler for specific time. The RX850 Pro therefore provides the functions of delayed wake-up of a task, timeout, and starting of a cyclically activated handler, as timer operation functions. 72 User's Manual U13773EJ3V0UM CHAPTER 7 TIME MANAGEMENT FUNCTION 7.4 Delayed Task Wake-Up Delayed task wake-up changes the state of a task from run to wait (the timeout wait state) and leaves the task in this state for a given period. Once this period elapses, the task is released from the wait state and returns to the ready state. Delayed task wake-up is performed by issuing the dly_tsk system call. * dly_tsk system call Upon the issue of the dly_tsk system call, the state of the task from which this system call was issued changes from run to wait (the timeout wait state). The timeout wait state is canceled in the following cases and the task returns to the ready state. * Upon the elapse of the delay specified by a parameter. * Upon the issue of a rel_wai system call and the forcible cancelation of the wait state. Figure 7-1 shows the flow of the processing after the issue of the dly_tsk system call. Figure 7-1. Flow of Processing After Issuance of dly_tsk Task A Priority: High Task B Priority: Low dly_tsk (Delay time) Delay time User's Manual U13773EJ3V0UM 73 CHAPTER 7 TIME MANAGEMENT FUNCTION 7.5 Timeout If the conditions required for a certain action are not satisfied when that action is requested by a task, the timeout function changes the state of the task from run to wait (wake-up wait state, resource wait state, etc.) and leaves the task in the wait state for a given period. Once that period elapses, the timeout function releases the task from the wait state. The task then returns to the ready state. The timeout function is enabled by issuing the tslp_tsk, twai_sem, twai_flg, trcv_msg, or tget_blk system call. * tslp_tsk system call Upon the issuance of the tslp_tsk system call, one request for wake-up, issued for the task from which this system call is issued, is canceled (the wake-up request counter is decremented by 0x1). If the wake-up request counter of the task from which this system call is issued currently indicates 0x0, the wake-up request is not canceled (decrement of the wake-up request counter) and the task enters the wait state (the wake-up wait state) from the run state. The wake-up wait state is canceled in the following cases, and the task returns to the ready state. * When the given wait time specified by a parameter has elapsed. * When the wup_tsk system call is issued. * When the ret_wup system call is issued. * When the rel_wai system call is issued and the wait state is forcibly canceled. * twai_sem system call Upon the issuance of the twai_sem system call, the task acquires a resource from the semaphore specified by a parameter (the semaphore counter is decremented by 0x1). After the issuance of this system call, if the task cannot acquire a resource from the semaphore specified by the parameter (no free resource exists), the task itself is queued in the wait queue of this semaphore. Thus, the task leaves the run state and enters the wait state (the resource wait state). The resource wait state is canceled in the following cases, and the task returns to the ready state. * When the given wait time specified by a parameter has elapsed. * When the sig_sem system call is issued. * When the del_sem system call is issued and the specified semaphore is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. * twai_flg system call The twai_flg system call checks whether the bit pattern is set so as to satisfy the wait condition required for the event flag specified by the parameter. If the bit pattern does not satisfy the wait condition required for the event flag specified by this system call parameter, the task from which this system call is issued is queued at the end of the wait queue of this event flag. Thus, the task leaves the run state and enters the wait state (the event flag wait state). The event flag wait state is canceled in the following cases, and the task returns to the ready state. * When the given wait time specified by a parameter has elapsed. * When the set_flg system call is issued and the required wait condition is satisfied. * When the del_flg system call is issued and the specified event flag is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. 74 User's Manual U13773EJ3V0UM CHAPTER 7 TIME MANAGEMENT FUNCTION * trcv_msg system call Upon the issuance of the trcv_msg system call, the task receives a message from the mailbox specified by the parameter. After the issuance of this system call, if the task cannot receive a message from the specified mailbox (no messages exist in the message wait queue of that mailbox), the task itself is queued at the end of the task wait queue of this mailbox. Thus, the task leaves the run state and enters the wait state (the message wait state). The message wait state is canceled in the following cases, and the task returns to the ready state. * When the given time specified by a parameter has elapsed. * When the snd_msg system call is issued. * When the del_mbx system call is issued and this mailbox is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. * tget_blk system call Upon the issuance of the tget_blk system call, the task acquires a memory block from the memory pool specified by the parameter. After the issuance of this system call, if the task cannot acquire the block from the specified memory pool (because no free block of the required size exists), the task itself is queued in the wait queue of this memory pool. Thus, the task leaves the run state and enters the wait state (the memory block wait state). The memory block wait state is canceled in the following cases, and the task returns to the ready state. * When the given wait time specified by a parameter has elapsed. * When the rel_blk system call is issued and a memory block of the required size is returned. * When the del_mpl system call is issued and the specified memory pool is deleted. * When the rel_wai system call is issued and the wait state is forcibly canceled. User's Manual U13773EJ3V0UM 75 CHAPTER 7 TIME MANAGEMENT FUNCTION 7.6 Cyclically Activated Handler The cyclically activated handler is an exclusive period processing routine which starts immediately when a predetermined start time arrives, and is a processing program which has optimally small overhead within the periodic processing program described by the user until execution is started. The cyclically activated handler is treated as independent of the task. For this reason, even if a task with the highest priority order is being executed in the system, that processing is interrupted and the system switches to the cyclically activated handler's control. The following system calls and instructions relevant to a cyclically activated handler are used in the dynamic operation of a cyclically activated handler. def_cyc: act_cyc: ref_cyc: return: Registers a cyclically activated handler. Controls the activity state of the cyclically activated handler. Acquires cyclically activated handler information. Performs return from the cyclically activated handler. 7.6.1 Registering a cyclically activated handler The RX850 Pro provides two interfaces for registering a cyclically activated handler. by issuing a system call from within a processing program. To register a cyclically activated handler with the RX850 Pro, an area in system memory is allocated for managing the cyclically activated handler (to be managed by the RX850 Pro), then initialized. (1) Static registration of a cyclically activated handler To statically register a cyclically activated handler, specify it during configuration. The RX850 Pro performs the processing for registering the cyclically activated handler, based on the information defined in the information file (including system information tables and system information header files) during system initialization. The cyclically activated handler is subsequently managed by the RX850 Pro. (2) Dynamic registration of a cyclically activated handler To dynamically register a cyclically activated handler, issue the def_cyc system call from within a processing program (task or non-task). The RX850 Pro performs the processing for registering the cyclically activated handler, according to the information specified by the parameter when the def_cyc system call is issued. The cyclically activated handler is subsequently managed by the RX850 Pro. One enables static registration during system initialization (in the nucleus initialization section). The other enables dynamic registration 76 User's Manual U13773EJ3V0UM CHAPTER 7 TIME MANAGEMENT FUNCTION 7.6.2 Activity state of cyclically activated handler The activity state of a cyclically activated handler is used as a criterion for determining whether the RX850 Pro activated the cyclically activated handler. The activity state is set when the cyclically activated handler is registered (during configuration or when the def_cyc system call is issued). However, the RX850 Pro allows the user to change the activity state of the cyclically activated handler from a user processing program. * act_cyc system call Upon the issuance of the act_cyc system call, the activity state of the cyclically activated handler is switched ON/OFF, as specified by the parameter. TCY_OFF: TCY_ON: TCY_INI: Switches the activity state of the cyclically activated handler to OFF. Switches the activity state of the cyclically activated handler to ON. Initializes the cycle counter of the cyclically activated handler. While the RX850 Pro is running, the cycle counter continues to count even when the activity state of the cyclically activated handler is OFF. In some cases, when an act_cyc system call is issued to switch the activity state of the cyclically activated handler from OFF to ON, the first restart request could be issued sooner than the activation time interval specified when it was registered (during configuration or upon the issuance of the def_cyc system call). To prevent this, the user must specify TCY_INI to initialize the cycle counter as well as TCY_ON to restart the cyclically activated handler when issuing the act_cyc system call. The first restart request will then be issued in sync with the time interval, specified when it was registered. Figures 7-2 and 7-3 show the flow of processing after the issuance of the act_cyc system call from a processing program to switch the activity state of the cyclically activated handler from OFF to ON. In those figures, T indicates the activation time interval specified upon the registration of the cyclically activated handler. In addition, the relationship between t and T in Figure 7-2 is assumed to be t T. Figure 7-2. Flow of Processing After Issuance of act_cyc (TCY_ON) Task Cyclically activated handler act_cyc (TCY_ON) t return T return T act_cyc (TCY_OFF) User's Manual U13773EJ3V0UM 77 CHAPTER 7 TIME MANAGEMENT FUNCTION Figure 7-3. Flow of Processing After Issuance of act_cyc (TCY_ON|TCY_INI) Task Cyclically activated handler act_cyc (TCY_ON|TCY_INI) T return T T return act_cyc (TCY_OFF) 7.6.3 Internal processing performed by cyclically activated handler After the occurrence of a timer interrupt, the RX850 Pro performs preprocessing for interruption before control is passed to the cyclically activated handler. When control is returned from the cyclically activated handler, the RX850 Pro performs interrupt postprocessing. When describing the processing to be performed by the activated interrupt handler, note the following: (1) Saving/restoring the registers Based on the function call protocol for the C compiler (CA850 or CCV850), the RX850 Pro saves the work registers when control is passed to the cyclically activated handler, and restores them upon the return of control from the handler. Therefore, the cyclically activated handler does not have to save the work registers when it starts, nor restore them upon the completion of its processing. Save/restoration of the registers should not be coded in the description of the cyclically activated handler. (2) Stack switching The RX850 Pro performs stack switching when control is passed to the cyclically activated handler and upon a return from the handler. Therefore, the cyclically activated handler does not have to switch to the interrupt handler stack when it starts, nor switch to the original stack upon the completion of its processing. However, if the interrupt handler stack is not defined during configuration, stack switching is not performed and system continues to use that stack being used upon the occurrence of an interrupt. 78 User's Manual U13773EJ3V0UM CHAPTER 7 TIME MANAGEMENT FUNCTION (3) Limitations imposed on system calls The following lists the system calls that can be issued during the processing performed by a cyclically activated handler: * Task management system calls sta_tsk ref_tsk chg_pri vget_tid rot_rdq rel_wai get_tid * Task-associated synchronization system calls sus_tsk rsm_tsk frsm_tsk wup_tsk can_wup * Synchronous communication system calls sig_sem clr_flg prcv_msg preq_sem pol_flg ref_mbx ref_sem ref_flg vget_mid vget_sid vget_fid set_flg snd_msg * Interrupt management system calls def_int ena_int dis_int chg_icr ref_icr * Memory pool management system calls pget_blk rel_blk ref_mpl vget_pid * Time management system calls set_tim get_tim def_cyc act_cyc ref_cyc * System management system calls get_ver ref_sys def_svc viss_svc (4) Return processing from the cyclically activated handler Return processing from the cyclically activated handler is performed by issuing a return instruction upon the completion of the processing performed by cyclically activated handler. When a system call (chg_pri, sig_sem, etc.) that requires task scheduling is issued during the processing of a cyclically activated handler, RX850 Pro merely queues that task into the wait queue. issuing a return instruction). 7.6.4 Acquiring cyclically activated handler information Information related to a cyclically activated handler is acquired by issuing the ref_cyc system call. * ref_cyc system call By issuing the ref_cyc system call, the task acquires information (including extended information, remaining time, etc.) related to the cyclically activated handler specified by a parameter. The cyclically activated handler information consists of the following: * Extended information * Time remaining until the next start of the cyclically activated handler * Current activity state User's Manual U13773EJ3V0UM The actual task scheduling is batched and deferred until return from the cyclically activated handler has been completed (by 79 CHAPTER 7 TIME MANAGEMENT FUNCTION 7.6.5 Interrupts in cyclically activated handler The cyclically activated handler starts processing in the interrupt enabled state. To disable interrupts during use of the cyclically activated handler, disable interrupts at the start of the handler. Since the RX850 Pro provides two types of nucleus common parts (rxcore.o and rxtmcore.o), the interrupts that can be acknowledged within the cyclically activated handler differ depending on the nucleus common part used. * When rxcore.o is used Although the cyclically activated handler is called from the timer handler, all levels of interrupts can be acknowledged because the interrupt processing is performed during timer handler execution. * When rxtmcore.o is used Although the cyclically activated handler is called from the timer handler, only the interrupts with a higher priority than timer interrupts can be acknowledged because the interrupt processing is not performed during timer handler execution. In addition, since timer interrupts are held pending even when interrupts are enabled, to execute a time-consuming processing within the cyclically activated handler, caution is required because displacement may occur between the time that has actually elapsed and the time managed by the RX850 Pro. Because the cyclically activated handler is developed as an indirectly activated interrupt handler, it operates on the handler stack at execution. 7.6.6 Activation order of cyclically activated handler When multiple cyclically activated handlers for which the activation interval time simultaneously elapsed exist, activated in order from the handler with shorter activation interval time specified. In addition, if the activation interval time elapsed during the another cyclically activated handler execution, the cyclically activated handler is not activated immediately but activated after the end of the cyclically activated handler currently executed. 80 User's Manual U13773EJ3V0UM CHAPTER 8 SCHEDULER This chapter explains the task scheduling performed by the RX850 Pro. 8.1 Overview By monitoring the dynamically changing task states, the RX850 Pro scheduler manages and determines the sequence in which tasks are executed, and assigns a processing time to a specific task. 8.2 Drive Method The RX850 Pro scheduler uses an event-driven technique, in which the scheduler operates in response to the occurrence of some event. The "occurrence of some event" means the issue of a system call that may cause a task state change, the issue of a return instruction that causes a return from a handler, or the occurrence of a clock interrupt. When these phenomena occur, task scheduling processing is executed with the scheduler driving. The following system calls can be used to drive the scheduler. * Task management system calls sta_tsk rot_rdq ext_tsk rel_wai exd_tsk ena_dsp chg_pri * Task-associated synchronization system calls rsm_tsk frsm_tsk slp_tsk tslp_tsk wup_tsk * Synchronous communication system calls del_sem set_flg rcv_msg sig_sem wai_flg trcv_msg wai_sem twai_flg twai_sem del_mbx del_flg snd_msg * Interrupt management system calls ret_int ret_wup vret_clk unl_cpu * Memory pool management system calls del_mpl get_blk tget_blk rel_blk * Time management system call dly_tsk User's Manual U13773EJ3V0UM 81 CHAPTER 8 SCHEDULER 8.3 Scheduling Method The RX850 Pro uses the priority and FCFS (First-Come, First-Served) scheduling method. When driven, the scheduler checks the priority of each task that can be executed (in the run or ready state), selects the optimum task, and assigns a processing time to the selected task. 8.3.1 Priority method Each task is assigned a priority that determines the sequence in which it will be executed. The scheduler checks the priority of each task that can be executed (in the run or ready state), selects the task having the highest priority, and assigns a processing time to the selected task. Remark In the RX850 Pro, a task to which a smaller value is assigned as the priority level has a higher priority. 8.3.2 FCFS method The RX850 Pro can assign the same priority to more than one task. Because the priority method is used for task scheduling, there is the possibility of more than one task having the highest priority being selected. Among those tasks having the highest priority, the scheduler selects the first to become executable (the task that has been in the ready state for the longest time) and assigns a processing time to the selected task. 82 User's Manual U13773EJ3V0UM CHAPTER 8 SCHEDULER 8.4 Implementing a Round-Robin Method In scheduling based on the priority and FCFS methods, even if a task has the same priority as that currently running, it cannot be executed unless the task to which a processing time has been assigned first enters another state or relinquishes control of the processor. The RX850 Pro provides system calls such as rot_rdq to implement a scheduling method (round-robin method) that can overcome the problems incurred by the priority and FCFS methods. The round-robin method can be implemented as follows: Conditions * Task priority Task A = Task B = Task C * State of tasks Task A: Run state Task B: Ready state Task C: Ready state * Cyclically activated handler attributes Activity state: Activation interval: Processing: ON T (unit: Basic clock cycle) Rotation of the ready queues (issue of the rot_rdq system call) (1) Task A is currently running. The other tasks (B and C) have the same priority as task A, but they cannot be executed unless task A enters another state or relinquishes control of the processor. The ready queue becomes as shown in Figure 8-1. Figure 8-1. Ready Queue State (1) Activation wait state High Ready queue Handler Priority Run state Task A Ready state Task B Ready state Task C Low User's Manual U13773EJ3V0UM 83 CHAPTER 8 SCHEDULER (2) Cyclically activated handler starts when the predetermined period of time has passed and issues the rot_rdq system call. In this way, task A is queued at the tail end of the ready queue in accordance with its priority level and changes from the run state to ready state. The ready queue changes to the state shown in Figure 8-2. Figure 8-2. Ready Queue State (2) Processing execution state High Ready queue Handler Priority Ready state Task B Ready state Task C Ready state Task A Low (3) Task A changes from the run state to the ready state and task B changes from the ready state to the run state. Figure 8-3 shows the ready queue state at this time. Figure 8-3. Ready Queue State (3) Activation wait state High Ready queue Handler Priority Run state Task B Ready state Task C Ready state Task A Low 84 User's Manual U13773EJ3V0UM CHAPTER 8 SCHEDULER (4) By issuing the rot_rdq system call from the cyclically activated handler, which is started at constant intervals, the scheduling method (round-robin method) in which tasks are switched every time the specified period (T) elapses is implemented. Figure 8-4 shows the processing flow when the round-robin method is used. Figure 8-4. Flow of Processing by Using Round-Robin Method Task A Task B Task C Cyclically activated handler rot_rdq return T rot_rdq return T rot_rdq return T rot_rdq return User's Manual U13773EJ3V0UM 85 CHAPTER 8 SCHEDULER 8.5 Scheduling Lock Function In the RX850 Pro a function is offered which drives the scheduler from a user processing program (task) and which disables or resumes dispatch processing (task scheduling processing). This function is implemented by issuing the following system calls from within a task. * dis_dsp system call Disables dispatching (task scheduling). If this system call is issued, control is not passed to another task until the ena_dsp system call is issued. * ena_dsp system call Resumes dispatching (task scheduling). When the dis_dsp system call has been issued, if a system call that requires task scheduling (such as chg_pri or sig_sem) is issued, the RX850 Pro merely executes processing such as wait queue operation until the ena_dsp system call is issued. Actual scheduling is delayed and batch-executed upon the issuance of the ena_dsp system call. * loc_cpu system call Disables the acknowledgement of maskable interrupts, then disables dispatching (task scheduling). If this system call is issued, control will not be passed to another task or handler until the unl_cpu system call is issued. * unl_cpu system call Enables the acknowledgement of maskable interrupts, then restarts dispatching (task scheduling). If a maskable interrupt has occurred between the issuance of the loc_cpu system call and that of the unl_cpu system call, transfer of control to the corresponding interrupt handling (processing of the interrupt handler) is delayed until unl_cpu system call is issued. Also, if a system call which is necessary for task scheduling processing (such as chg_pri or sig_sem) is issued during the interval after the loc_cpu system call is issued and until the unl_cpu system call is issued, only processing of wait queue operations is delayed until the unl_cpu system call is issued, being performed by batch processing. The flow of control if scheduling processing is not delayed (normal) is shown in Figure 8-5 and the flow of control if the dis_dsp and loc_cpu system calls are issued is shown in Figure 8-6 and Figure 8-7. Figure 8-5. Flow of Control if Scheduling Processing Is Not Delayed (Normal) Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk Occurrence of interrupt ret_int wup_tsk 86 User's Manual U13773EJ3V0UM CHAPTER 8 SCHEDULER Figure 8-6. Flow of Control if dis_dsp System Call Is Issued Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk dis_dsp Occurrence of interrupt ret_int wup_tsk ena_dsp Figure 8-7. Flow of Control if loc_cpu System Call Is Issued Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk loc_cpu Occurrence of interrupt wup_tsk unl_cpu ret_int User's Manual U13773EJ3V0UM 87 CHAPTER 8 SCHEDULER 8.6 Scheduling While Handler Is Operating To quickly terminate handlers (interrupt handlers and cyclically activated handlers), the RX850 Pro delays the driving of the scheduler until processing within the handler terminates. Therefore, if a system call that requires task scheduling (such as chg_pri or sig_sem) is issued, the RX850 Pro merely executes processing such as wait queue operation until the completion of return processing from the handler (such as ret_int system call or the issue of return instruction). Actual scheduling is delayed and then batch-executed upon the completion of return processing. Figure 8-8 shows the control flow when a handler issues a system call that requires scheduling. Figure 8-8. Flow of Control if wup_tsk System Call Is Issued Task A Priority: Low Task B Priority: High Interrupt handler slp_tsk Occurrence of interrupt wup_tsk ret_int 8.7 Idle Handler 8.7.1 Idle handler The idle handler is started from the scheduler if all the tasks (user defined tasks) are not in the run state or not in the ready state, that is, if there is not even one task which is an object of RX850 Pro scheduling in the system. The processing of the idle handler is to switch the CPU to the HALT state. Therefore, if there is not even one task in the system, the RX850 Pro switches the CPU to the HALT state. However, this idle handler cannot switch the CPU to the IDLE or STOP state. To switch to the IDLE or STOP state, or to describe idle processing, create a task with the lowest priority and use it as an idle task. This realizes processing identical to the idle handler. However, since the HALT, IDLE, or STOP state is released by an interrupt, be sure not to leave interrupts in a disabled state in the idle task. 88 User's Manual U13773EJ3V0UM CHAPTER 9 SYSTEM INITIALIZATION This chapter explains the system initialization performed by the RX850 Pro. For details of the system initialization, refer to the RX850 Pro Installation User's Manual (U13774E). 9.1 Overview System initialization consists of initializing the hardware required by the RX850 Pro, as well as initializing the software. In other words, in the RX850 Pro, the processing performed immediately after the system has been started is system initialization. Figure 9-1 shows the flow of system initialization. Figure 9-1. Flow of System Initialization V850 Series reset entry Hardware initialization section Boot processing Nucleus Nucleus initialization section Software initialization section Scheduler Initial task User's Manual U13773EJ3V0UM 89 CHAPTER 9 SYSTEM INITIALIZATION 9.2 Boot Processing Boot processing is the function assigned to the V850 Series reset entry (handler address: 0x0) and the first function executed in system initialization. The files boot.s (NEC Electronics version) and boot.850 (GHS version) are used in the sample boot processing (function name: _ _boot). Boot processing involves the following. * Setting of the gp, tp, and ep registers * Initialization of a memory area without initial values * Calling of the hardware initialization section * Transfer of control to the nucleus initialization section In the sample boot processing, the processing can be rewritten to adapt to user needs. 9.3 Hardware Initialization Section The hardware initialization section is a function called from the boot processing and it is prepared for initializing the hardware in the execution environment (target system). The file init.c is used in the sample initialization (function name: reset). In this hardware initialization section, the following processing is performed. * Initialization of the internal unit * Initialization of an interrupt controller * Initialization of a clock controller * Initialization of a peripheral controller * Returns control to boot processing The hardware initialization section depends on the hardware configuration of the execution environment. Designing this section into the LSI improves portability to the target system and simplifies customization. Rewrite in accordance with the user execution environment. 90 User's Manual U13773EJ3V0UM CHAPTER 9 SYSTEM INITIALIZATION 9.4 Nucleus Initialization Section The nucleus initialization section is a function called after the boot processing completion and it generates and initializes the management objects based on the information (such as task information or semaphore information) described in the information files (system information table and system information header file). The RX850 Pro is activated after completion of this processing. This processing section is included in the nucleus library. The nucleus initialization section performs the following processing. * Generation/initialization of management objects * Task generation * Generation/initialization of a semaphore * Generation/initialization of event flags * Generation/initialization of a memory pool * Registration of the indirectly activated interrupt handler * Registration of the cyclically activated handler * Registration of the extended SVC handler * Activation of an initial task * Activation of the system task (idle task) * Calling of the software initialization section * Transfer of control to the scheduler 9.5 Software Initialization Section The software initialization section is a function called from the nucleus initialization section and used if some processing is to be executed before the activation of the RX850 Pro. The file varfunc.c is used in the sample processing (function name: varfunc). The software initialization section performs the following processing. * Copying of an initialization data * Returns control to the nucleus initialization section User's Manual U13773EJ3V0UM 91 CHAPTER 10 INTERFACE LIBRARY This chapter explains the interface library. For details of the interface library, refer to the RX850 Pro Installation User's Manual (U13774E). 10.1 Overview In the RX850 Pro, an interface library is provided which is positioned midway between the user processing program and the RX850 Pro nucleus. The interface library has a function for transferring control after performing setting of each type of necessary information, etc. for enabling processing by the nucleus. When a processing program (task/non-task) is written in C language, an external function format is used to issue a system call or to call an extended SVC handler. The issuance format that the nucleus can understand (nucleus issuance format), however, differs from the external function format. It is therefore necessary to convert the system call issue format or expanded SVC handler calling format from the external function format to the nucleus issuance format (interfacing). There is an interface which performs the role of intermediary between the processing program and the nucleus for each system call. All these interfaces collected together are called the interface library. Figure 10-1 shows the positioning of the interface library in the RX850 Pro. Figure 10-1. Positioning of Interface Library Task Issue of a system call External function format Interface library Interface processing Nucleus issuance format Nucleus System call processing 92 User's Manual U13773EJ3V0UM CHAPTER 10 INTERFACE LIBRARY 10.2 Processing in the Interface Library The following processing is performed in the interface library. * Setting of the necessary information in tables managed by the nucleus. * Setting the necessary data in registers. * After setting system call error values (with the exception of errors set in the nucleus), it returns to the processing program. By providing an interface library, it becomes easy to separate the nucleus and the user processing program. For example, even if it becomes necessary to change the user's processing program after the nucleus body has been loaded in ROM, it becomes unnecessary to change the ROM where the nucleus body is stored. It also becomes possible to create it with the load module divided. 10.3 Types of Interface Libraries There are two types of interface libraries offered with RX850 Pro, one with a function for checking system call parameters, and one without this function. The type of interface library which will be incorporated is specified at linking. The use of library with the parameter check function always return return values, if the parameters specified when a system call is issued are incorrect. On the other hand, the use of library without the parameter check function may not return return values, if the parameters specified when a system call is issued are incorrect. Utilization of these two library types can be divided in accordance with the use. For example, during debugging, by use of the library with the parameter check function and by use of the library without the parameter check function during the actual build-in, improvements in program performance and capacity reductions can be realized. Remark Errors in which return values are returned with the library which does not have a parameter check function are marked by "*" in the system call return value column in CHAPTER 11 SYSTEM CALLS. Caution When the library without the parameter check function is used, if errors occur in which return values are not returned, the operation of the application system cannot be guaranteed. 10.4 Supported Interface Libraries The RX850 Pro supports the following two interface libraries: * For NEC Electronics V850 Series C compiler CA850 * For C cross V800 compiler CCV850 manufactured by Green Hills Software, Inc. Remark To use other compilers, the interface library must be rewritten in accordance with the register used in its compiler. For the addresses where the interface libraries are stored, refer to the RX850 Pro Installation User's Manual (U13774E). User's Manual U13773EJ3V0UM 93 CHAPTER 11 SYSTEM CALLS This chapter describes the system calls supported by the RX850 Pro. 11.1 Overview A system call is a procedure or function for invoking RX850 Pro service routines from the user's processing programs (tasks/non-tasks). The user can use system calls to indirectly manipulate those resources (such as counters and queues) that are managed directly by the RX850 Pro. The RX850 Pro supports its own seven system calls as well as the 65 defined in the ITRON3.0 specifications, thus enhancing the versatility of application systems. System calls can be classified into the following seven groups, according to their functions. (1) Task management system calls (14) These system calls are used to manipulate the status of a task. This group provides functions for creating, activating, terminating, and deleting a task, a function for disabling and resuming dispatch processing, a function for changing the task priority, a function for rotating a task ready queue, a function for forcibly releasing a task from the wait state, and a function for referencing the task status. cre_tsk ter_tsk rel_wai del_tsk dis_dsp get_tid sta_tsk ena_dsp ref_tsk ext_tsk chg_pri vget_tid exd_tsk rot_rdq (2) Task-associated synchronization system calls (7) These system calls perform synchronous operations associated with tasks. This group provides a function for placing a task in the suspend state and restarting a suspended task, a function for placing a task in the wake-up wait state and waking up a task currently in the wake-up wait state, and another function for canceling a task wake-up request. sus_tsk wup_tsk rsm_tsk can_wup frsm_tsk slp_tsk tslp_tsk (3) Synchronous communication system calls (25) These system calls are used for the synchronization (exclusive control and queuing) and communication between tasks. This group provides a function for manipulating semaphores, a function for manipulating events and flags, and a function for manipulating mailboxes. cre_sem twai_sem set_flg ref_flg rcv_msg del_sem ref_sem clr_flg vget_flg prcv_msg sig_sem vget_sid wai_flg cre_mbx trcv_msg wai_sem cre_flg pol_flg del_mbx ref_mbx preq_sem del_flg twai_flg snd_msg vget_mid 94 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS (4) Interrupt management system calls (9) These system calls perform processing that is dependent on the maskable interrupts. This group provides a function for registering an indirectly activated interrupt handler and subsequently canceling the registration, a function for returning from a directly activated interrupt handler, and a function for changing or referencing an interrupt-enabled level. def_int loc_cpu ret_int unl_cpu ret_wup chg_icr ena_int ref_icr dis_int (5) Memory pool management system calls (8) These system calls allocate memory. This group provides a function for creating and deleting a memory pool, a function for acquiring and returning a memory block, and a function for referencing the status of a memory pool. cre_mpl rel_blk del_mpl ref_mpl get_blk vget_pid pget_blk tget_blk (6) Time management system calls (6) These system calls perform processing that is dependent on time. This group provides a function for setting or referencing the system clock, a function for placing a task in the timeout wait state, a function for registering a cyclically activated handler and subsequently canceling the registration, and a function for controlling and referencing the state of a cyclically activated handler. set_tim ref_cyc (7) System management system calls (4) These system calls perform processing that varies with the system. This group provides a function for acquiring version information, a function for referencing the system status, a function for registering an extended SVC handler and subsequently canceling the registration, and a function for calling an extended SVC handler. get_ver ref_sys def_svc viss_svc get_tim dly_tsk def_cyc act_cyc User's Manual U13773EJ3V0UM 95 CHAPTER 11 SYSTEM CALLS 11.2 Calling System Calls System calls issued from processing programs (task/non-task) written in C language are called as C language functions. Their parameters are passed as arguments. When issuing system calls from processing programs written in assembly language, set parameters and a return address according to the function calling rules of the C compiler, used before calling them with the jarl instruction. Caution The RX850 Pro declares the prototype of a system call in the stdrx85p.h file. Accordingly, when issuing a system call from a processing program, the following must be coded to include the header file: #include 11.3 System Call Function Codes The system calls supported by the RX850 Pro are assigned function codes conforming to the ITRON3.0 specifications. Table 11-1 lists the function codes assigned to system calls. In the RX850 Pro, a value of 1 or greater is used when registering an extended SVC handler described by the user. Table 11-1. System Call Function Codes Function Code -256 to -225 -224 to -5 -4 to 0 1 or more Classification RX850 Pro original system calls System calls conforming to the ITRON3.0 specifications Reserved by the system Extended SVC handler 96 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.4 Data Types of Parameters The system calls supported by the RX850 Pro have parameters that are defined based on data types that conform to the ITRON3.0 specifications. Table 11-2 lists the data types of the parameters specified upon the issuance of a system call. Table 11-2. Data Types of Parameters Macro B H INT W UB UH UINT UW VB VH VW *VP (*FP) () BOOL FN ID BOOL_ID HNO ATR ER PRI TMO CYCTIME DLYTIME char short int long unsigned char unsigned short unsigned int unsigned long char short long void void short short short short short unsigned short long short long long long Data Type Signed 8-bit integer Signed 16-bit integer Signed 32-bit integer Signed 32-bit integer Unsigned 8-bit integer Unsigned 16-bit integer Unsigned 32-bit integer Unsigned 32-bit integer Variable data type value (8 bits) Variable data type value (16 bits) Variable data type value (32 bits) Variable data type value (pointer) Processing program start address Boolean value Function code Object ID number Wait task available or not Cyclically activated handler specification number Object attribute Error code Task priority Wait time Cyclically activated time interval (residual time) Delay time Description User's Manual U13773EJ3V0UM 97 CHAPTER 11 SYSTEM CALLS 11.5 Parameter Value Range Some of the system call parameters supported by the RX850 Pro have a range of permissible values, while others allow the use of only system reserved specific values. Table 11-3 lists the ranges of parameter values that can be specified upon the issuance of a system call. Table 11-3. Ranges of Parameter Values Parameter Type Object ID number Object key ID number Interrupt handler interrupt level Specification number of cyclically activated handler Extended function code of extended SVC handler Object priority Maximum number of semaphore resources Interrupt enable level of maskable interrupt System clock time Wait time Delay time Activation time interval of cyclically activated handler Task stack size Memory pool size Memory block size Message priority level Value Range 0x0 to max_cntNote 1 -0x8000 to 0x7FFFNote 2 0x0 to 0xF 0x1 to max_cnt 0x1 to max_cnt 0x1 to max_cnt 0x1 to max_cnt 0x0 to 0xF 0x0 to 0x7FFF FFFF FFFF -0x1 to 0x7FFF FFFF 0x0 to 0x7FFF FFFF 0x1 to 0x7FFF FFFF 0x0 to 0x7FFF FFFF 0x1 to 0x7FFF FFFF 0x1 to 0x7FFF FFFF 0x1 to 0x7FFF Notes 1. 2. max_cnt: Maximum number of objects specified during system configuration "0x0" cannot be specified for the object key ID number. 98 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.6 System Call Return Values The system call return values supported by the RX850 Pro are based on the ITRON3.0 specifications. Table 11-4 lists the system call return values. Table 11-4. System Call Return Values Macro E_OK E_NOMEM E_NOSPT Value 0 -10 -17 Normal termination An area for objects cannot be allocated. A system call with the CF not defined, or an unregistered extended SVC handler was called. Invalid object attribute specification Invalid parameter specification Invalid ID number specification No relevant object exists. The status of the specified object is invalid. An unauthorized ID number was specified. The state in which the system call is issued is invalid. The count exceeded 127. The target object was deleted. Timeout A wait state was forcibly canceled by the rel_wai system call. Description E_RSATR E_PAR E_ID E_NOEXS E_OBJ E_OACV E_CTX E_QOVR E_DLT E_TMOUT E_RLWAI -24 -33 -35 -52 -63 -66 -69 -73 -81 -85 -86 11.7 System Call Extension The RX850 Pro supports the extension of system calls (functions coded by users are registered in the nucleus as extended system calls). No limitations are imposed on those functions registered as extended system calls; standard system calls (system calls supported by the RX850 Pro) can also be included. If, however, standard system calls that can be issued only in the task state are included, the issuance state of the extended system calls is limited to "issuable only from task." Extended system calls are positioned as user-defined system calls, despite their having properties similar to tasks. That is, like standard system calls, the scheduler is started upon the termination of processing and an optimum task is selected. If a standard system call is included in extended system calls, note that control may pass to another task that is currently processing an extended system call because the scheduler is also started upon the termination of a standard system call. User's Manual U13773EJ3V0UM 99 CHAPTER 11 SYSTEM CALLS 11.8 Explanation of System Calls The following explains the system calls supported by the RX850 Pro, in the format shown below. Figure 11-1. System Call Description Format 1 2 3 ( ) 5 Overview 4 6 C format 7 Parameter(s) I/O Parameter Description 8 Explanation 9 Return value 100 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS (1) Name Indicates the name of the system call. (2) Semantics Indicates the source of the name of the system call. (3) Function code Indicates the function code of the system call. (4) Origin of system call Indicates where the system call can be issued. Task: Non-task: The system call can only be issued from a task. The system call can only be issued from a non-task (directly activated interrupt handler, indirectly activated interrupt handler and cyclically activated handler). The system call can be issued from both a task and a non-task. The system call can only be issued from a directly activated interrupt handler. Task/non-task: Directly activated interrupt handler: Cyclically activated handler: The system call can only be issued from a cyclically activated handler. (5) Overview Outlines the functions of the system call. (6) C format Indicates the format to be used when describing a system call to be issued in C language. (7) Parameter(s) System call parameters are explained in the following format. I/O A Parameter B Description C A: Parameter classification I ... Parameter input to RX850 Pro O ... Parameter output from RX850 Pro B: Parameter data type C: Description of parameter (8) Explanation Explains the function of a system call. (9) Return value Indicates a system call's return value using a macro and value. Value returned by both RX850 Pro having and that not having the parameter check function Return value not marked with an asterisk (*): Value returned only by RX850 Pro having the parameter check function Return value marked with an asterisk (*): User's Manual U13773EJ3V0UM 101 CHAPTER 11 SYSTEM CALLS 11.8.1 Task management system calls This section explains the group of system calls that are used to manipulate the task status (task management system calls). Table 11-5 lists the task management system calls. Table 11-5. Task Management System Calls System Call cre_tsk del_tsk sta_tsk ext_tsk exd_tsk ter_tsk dis_dsp ena_dsp chg_pri rot_rdq rel_wai get_tid ref_tsk vget_tid Creates another task. Deletes another task. Activates another task. Terminates the task which issued the system call. Terminates the task which issued the system call, then deletes it. Forcibly terminates another task. Disables dispatch processing. Resumes dispatch processing. Changes the priority of a task. Rotates a task ready queue. Forcibly releases another task from a wait state. Acquires the ID number of the task that issued the system call. Acquires task information. Acquires the task ID number. Function 102 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Create Task (-17) cre_tsk Task Overview Creates a task. C format * When an ID number is specified #include ER * When an ID number is not specified #include ER Parameters I/O I I O ID T_CTSK ID Parameter tskid; *pk_ctsk; *p_tskid; Task ID number Description Start address of packet storing task creation information Address of area used to store ID number * Structure of task creation information T_CTSK typedef struct VP ATR FP PRI INT VP VP ID } T_CTSK; t_ctsk { exinf; tskatr; task; itskpri; stksz; gp; tp; keyid; /* /* /* /* /* /* /* /* Extended information Task attribute Task activation address Task priority at activation (initial priority) Task stack size Specific GP register value for task Specific TP register value for task Task key ID number */ */ */ */ */ */ */ */ User's Manual U13773EJ3V0UM 103 CHAPTER 11 SYSTEM CALLS Explanation The RX850 Pro supports two types of interfaces for task creation: one in which an ID number is specified for task creation, and another in which an ID number is not specified. * When an ID number is specified A task having the ID number specified by tskid is created based on the information specified by pk_ctsk. The specified task changes from the non-existent state to the dormant state, in which it is managed by the RX850 Pro. * When an ID number is not specified A task is created based on the information specified by pk_ctsk. The specified task changes from the non-existent state to the dormant state, in which it is managed by the RX850 Pro. An ID number is allocated by the RX850 Pro and the allocated ID number is stored in the area specified by p_tskid. The following describes task creation information in detail. exinf ... Extended information exinf is an area for storing user-specific information on a specified task. It can be used as necessary by the user. Information set in exinf can be acquired dynamically by issuing the ref_tsk system call from a processing program (task/non-task). tskatr ... Task attribute Bit 0 .. Task language TA_ASM(0): TA_HLNG(1): Bit 8 .. Assembly language C language Existence of key ID number specification TA_KEYID(1): Specifies key ID number Bit 9 .. Memory area specification TA_SPOL0(0): TA_SPOL1(1): Secures the stack area from system memory area 0. Secures the stack area from system memory area 1. Bit 10 .. Existence of specific GP register value specification TA_DPID(1): Specifies a specific GP register value. Bit 11 .. Existence of specific TP register value specification TA_DPIC(1): Specifies a specific TP register value. Bit 12 .. Maskable interrupt acknowledgement enabled or disabled TA_ENAINT(0): When a task is activated, the acknowledgement of maskable interrupts is enabled. TA_DISINT(1): When a task is activated, the acknowledgement of maskable interrupts is disabled. 104 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 15 tskatr 8 7 0 Task language Existence of key ID number specification Memory area specification Existence of specific GP register value specification Existence of specific TP register value specification Maskable interrupt acknowledgement enabled/disabled task itskpri stksz gp tp keyid Remark ... ... ... ... ... ... Task activation address Task initial priority (assigned upon activation) Stack size of task (unit: bytes) Specific GP register value for task Specific TP register value for task Task key ID number If the value of Bit 8 is not 1 (TA_KEYID), the contents of keyid are meaningless. If the value of Bit 10 is not 1 (TA_DPID), the contents of gp are meaningless. If the value of Bit 11 is not 1 (TA_DPIC), the contents of tp are meaningless. Return value *E_OK *E_NOMEM *E_NOSPT E_RSATR E_PAR 0 -10 -17 -24 -33 Normal termination An area for task management block cannot be allocated. The cre_tsk system call is not defined as CF. Invalid specification of attribute tskatr Invalid parameter specification * The start address of the packet storing task creation information is invalid (pk_ctsk = 0). * Invalid activation address specification (task = 0) * Invalid initial priority specification (itskpri 0, maximum priority < itskpri) * Invalid key ID number specification (keyid = 0) (at TA_KEYID attribute specification) * The address of the area used to store the ID number is invalid (p_tskid = 0) (When a task is created with no ID number specified) E_ID *E_OBJ E_OACV E_CTX -35 -63 -66 -69 Invalid ID number specification (maximum number of tasks created < tskid) A task having the specified ID number has already been created. An unauthorized ID number (tskid 0) was specified. The cre_tsk system call was issued from a non-task. User's Manual U13773EJ3V0UM 105 CHAPTER 11 SYSTEM CALLS Delete Task (-18) del_tsk Task Overview Deletes another task. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call changes the task specified by tskid from the dormant state to the non-existent state. This releases the target task from the control of the RX850 Pro. Note that the exd_tsk system call is used when it is necessary for a task to delete itself. Caution This system call does not queue delete requests. Accordingly, if the target task is not in the dormant state, this system call returns E_OBJ as the return value. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ E_OACV E_CTX 0 -17 -35 -52 -63 -66 -69 Normal termination The del_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is not in the dormant state. An unauthorized ID number (tskid 0) was specified. The del_tsk system call was issued from a non-task. 106 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Start Task (-23) sta_tsk Task/non-task Overview Activates another task. C format #include ER Parameters I/O I I ID INT Parameter tskid; stacd; Task ID number Activation code Description Explanation This system call changes the task specified by tskid from the dormant state to the ready state. The target task is scheduled by the RX850 Pro. For stacd, specify the activation code to be passed to the target task. The target task can be manipulated by handling the activation code as if it were a function parameter. Caution This system call does not queue activation requests. Accordingly, when a target task is not in the dormant state, this system call returns E_OBJ as the return value. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ E_OACV 0 -17 -35 -52 -63 -66 Normal termination The sta_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is not in the dormant state. An unauthorized ID number (tskid 0) was specified. User's Manual U13773EJ3V0UM 107 CHAPTER 11 SYSTEM CALLS Exit Task (-21) ext_tsk Task Overview Terminates the task that issued the system call. C format #include void Parameter None. Explanation This system call changes the state of the task from the run state to the dormant state. The task is excluded from RX850 Pro scheduling. Remarks 1. This system call initializes the "task creation information" specified at task creation (at configuration or upon the issuance of the cre_tsk system call). 2. If a task is coded in assembly language, perform coding as follows to terminate the issuing task. jr _ext_tsk Cautions 1. If this system call is issued from a non-task or in the dispatch disabled state, its operation is not guaranteed. 2. This system call does not release those resources (memory block, semaphore count, etc.) that were acquired before the termination of the issuing task. Accordingly, the user has to release such resources before issuing this system call. Return value None. 108 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Exit and Delete Task (-22) exd_tsk Task Overview Terminates the task that issued the system call, then deletes it. C format #include void Parameter None. Explanation This system call changes the task from the run state to the non-existent state. This releases the task from the control of the RX850 Pro. Remark If a task is coded in assembly language, perform coding as follows to terminate or delete the issuing task. jr _exd_tsk Cautions 1. If this system call is issued from a non-task or in the dispatch disabled state, its operation is not guaranteed. 2. This system call does not release those resources (memory block, semaphore count, etc.) that were acquired before the termination of the issuing task. Accordingly, the user has to release such resources before issuing this system call. Return value None. User's Manual U13773EJ3V0UM 109 CHAPTER 11 SYSTEM CALLS Terminate Task (-25) ter_tsk Task Overview Forcibly terminates another task. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call forcibly changes the state of the task specified by tskid to the dormant state. Remark This system call initializes the "task creation information" specified at task creation (at configuration or upon the issuance of the cre_tsk system call). Cautions 1. This system call does not queue termination requests. Accordingly, if a target task is not in the ready, wait, suspend, or wait-suspend state, this system call returns E_NOEXS or E_OBJ as the return value. 2. This system call does not release those resources (memory block, semaphore count, etc.) that were acquired before the termination of the issuing task. Accordingly, the user has to release such resources before issuing this system call. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ 0 -17 -35 -52 -63 Normal termination The ter_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is the task that issued this system call, or the task is in the dormant state. An unauthorized ID number (tskid 0) was specified. The ter_tsk system call was issued from a non-task. E_OACV E_CTX -66 -69 110 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Disable Dispatch (-30) dis_dsp Task Overview Disables dispatch processing. C format #include ER Parameter None. Explanation This system call disables dispatch processing (task scheduling). Dispatch processing is disabled until the ena_dsp system call is issued after this system call has been issued. If a system call such as chg_pri or sig_sem is issued to schedule tasks after the dis_dsp system call is issued but before the ena_dsp system call is issued, the RX850 Pro merely performs operations on a wait queue and delays actual scheduling until the ena_dsp system call is issued, at which time the processing is performed in batch. Cautions 1. This system call does not queue disable requests. Accordingly, if the dis_dsp system call has already been issued and dispatch processing has been disabled, no processing is performed and a disable request is not handled as an error. 2. If a system call such as wait_sem and wai_flg is issued, causing the state of the task to change to the wait state after the dis_dsp system call is issued but before the ena_dsp system call is issued, the RX850 Pro returns E_CTX as the return value, regardless of whether the wait conditions are satisfied. Return value *E_OK *E_NOSPT *E_CTX 0 -17 -69 Normal termination The dis_dsp system call is not defined as CF. Context error * The dis_dsp system call was issued from a non-task. * The dis_dsp system call was issued after the loc_cpu system call was issued. User's Manual U13773EJ3V0UM 111 CHAPTER 11 SYSTEM CALLS Enable Dispatch (-29) ena_dsp Task Overview Enables dispatch processing. C format #include ER Parameter None. Explanation This system call enables dispatch processing (task scheduling). If a system call such as chg_pri and sig_sem is issued to schedule tasks after the dis_dsp system call is issued but before the ena_dsp system call is issued, the RX850 Pro merely performs operations on a wait queue and delays actual scheduling until the ena_dsp system call is issued, at which time the processing is performed in batch. Caution This system call does not queue resume requests. Accordingly, if the ena_dsp system call has already been issued and dispatch processing has been resumed, no processing is performed. The resume request is not handled as an error. Return value *E_OK *E_NOSPT *E_CTX 0 -17 -69 Normal termination The ena_dsp system call is not defined as CF. Context error * The ena_dsp system call was issued from a non-task. * The ena_dsp system call was issued after the loc_cpu system call was issued. 112 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Change Priority (-27) chg_pri Task/non-task Overview Changes the priority of a task. C format #include ER Parameters I/O I ID Parameter tskid; Task ID number TSK_SELF(0): Value: Description Local task Task ID number I PRI tskpri; Task priority TPRI_INI(0): Value: Task initial priority Task priority Explanation This system call changes the value of the task priority specified by tskid to that specified by tskpri. If the target task is in the run state or the ready state, this system call executes priority change processing and queues the target task at the tail end of the ready queue in accordance with its priority. Remarks 1. If the specified task is queued in a wait queue according to its priority, the issue of the chg_pri system call may change the wait order. Example When three tasks (task A: priority 10, task B: priority 11, task C: priority 12) are placed in a semaphore wait queue according to their priority, and if the priority of task B is changed from 11 to 9, then the wait order of the wait queue changes as shown below. Semaphore Task A Priority: 10 Task B Priority: 11 Task C Priority: 12 chg_pri (Task B, 9) Task B Priority: 9 Task A Priority: 10 Task C Priority: 12 Semaphore User's Manual U13773EJ3V0UM 113 CHAPTER 11 SYSTEM CALLS Remarks 2. The value specified by tskpri is active until the next chg_pri system call is issued, or until the target task changes to the dormant state. 3. Return value The task priority in the RX850 Pro becomes higher as its value decreases. *E_OK *E_NOSPT E_PAR E_ID 0 -17 -33 -35 Normal termination The chg_pri system call is not defined as CF. Invalid priority specification (tskpri < 0, maximum priority < tskpri) Invalid ID number specification * Maximum number of tasks created < tskid * When the chg_pri system call was issued from a non-task, TSK_SELF was specified in tskid. *E_NOEXS *E_OBJ E_OACV -52 -63 -66 The target task does not exist. The target task is in the dormant state. An unauthorized ID number (tskid < 0) was specified. 114 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Rotate Ready Queue (-28) rot_rdq Task/non-task Overview Rotates a task ready queue. C format #include ER Parameter I/O I PRI Parameter tskpri; Task priority TPRI_RUN(0): Value: Description Priority of task in run state Task priority Explanation This system call queues the first task in a ready queue to the end of the queue according to the priority specified by tskpri. Remarks 1. If no task of the specified priority exists in a ready queue, this system call performs no processing. This is not regarded as an error. 2. By issuing the rot_rdq system call at regular intervals, round-robin scheduling can be achieved. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The rot_rdq system call is not defined as CF. Invalid priority specification (tskpri < 0, maximum priority < tskpri) User's Manual U13773EJ3V0UM 115 CHAPTER 11 SYSTEM CALLS Release Wait (-31) rel_wai Task/non-task Overview Forcibly releases another task from the wait state. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call forcibly releases the task specified by tskid from the wait state. The target task is excluded from a wait queue, and its state changes from the wait state to the ready state, or from the wait-suspend state to the suspend state. For a task released from the wait state by the rel_wai system call, E_RLWAI is returned as the return value of the system call (slp_tsk, wai_sem, etc.) that caused transition to the wait state. Caution The rel_wai system call does not release the suspend state. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ E_OACV 0 -17 -35 -52 -63 -66 Normal termination The rel_wai system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is in neither the wait nor wait-suspend state. An unauthorized ID number (tskid 0) was specified. 116 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Get Task Identifier (-24) get_tid Task/non-task Overview Acquires a task ID number. C format #include ER Parameter I/O O ID Parameter *p_tskid; Description Address of area used to store ID number Explanation This system call stores the ID number of the task that issued this system call in the area specified by p_tskid. Caution If this system call is issued from a non-task, FALSE (0) is stored in the area specified by p_tskid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The get_tid system call is not defined as CF. The address of the area used to store the ID number is invalid (p_tskid = 0). User's Manual U13773EJ3V0UM 117 CHAPTER 11 SYSTEM CALLS Refer Task Status (-20) ref_tsk Task/non-task Overview Acquires task information. C format #include ER Parameters I/O O I T_RTSK ID Parameter *pk_rtsk; tskid; Description Start address of packet used to store task information Task ID number TSK_SELF(0): Value: Local task Task ID number * Structure of task information T_RTSK typedef struct VP PRI UINT UINT ID INT INT ID } T_RTSK; t_rtsk { exinf; tskpri; tskstat; tskwait; wid; wupcnt; suscnt; keyid; /* /* /* /* /* /* /* /* Extended information Current priority Task status Wait cause ID number of wait object Number of wake-up requests Number of suspend requests Key ID number */ */ */ */ */ */ */ */ Explanation This system call stores the task information (extended information, current priority, etc.) specified by tskid in the packet specified by pk_rtsk. The following describes the task information in detail. exinf tskpri ... ... Extended information Current priority 118 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS tskstat ... Task state TTS_RUN(H'01): TTS_RDY(H'02): TTS_WAI(H'04): TTS_SUS(H'08): TTS_WAS(H'0c): TTS_DMT(H'10): Run state Ready state Wait state Suspend state Wait-suspend state Dormant state tskwait ... Type of wait state TTW_SLP(H'0001): TTW_DLY(H'0002): TTW_FLG(H'0010): TTW_SEM(H'0020): TTW_MBX(H'0040): TTW_MPL(H'1000): Wake-up wait state Timeout wait state Event flag wait state Resource wait state Message wait state Memory block wait state wid wupcnt suscnt keyid ... ... ... ... ID number of wait object (semaphore, event, flag, etc.) Number of wake-up requests Number of suspend requests Key ID number FALSE(0): Value: No key ID number specified at creation Key ID number Remarks 1. When the value of tskstat is other than TTS_WAI or TTS_WAS, the contents of tskwait will be undefined. 2. When the value of tskwait is other than TTW_FLG, TTW_SEM, TTW_MBX, or TTW_MPF, the contents of wid will be undefined. Return value *E_OK *E_NOSPT E_PAR E_ID 0 -17 -33 -35 Normal termination The ref_tsk system call is not defined as CF. The start address of the packet used to store task information is invalid (pk_rtsk = 0) Invalid ID number specification * Maximum number of tasks created < tskid * When the ref_tsk system call was issued from a non-task, TSK_SELF was specified in tskid. *E_NOEXS E_OACV -52 -66 The target task does not exist. An unauthorized ID number (tskid < 0) was specified. User's Manual U13773EJ3V0UM 119 CHAPTER 11 SYSTEM CALLS Get Task Identifier (-248) vget_tid Task/non-task Overview Acquires a task ID number. C format #include ER Parameters I/O O I ID ID Parameter *p_tskid; keyid; Description Address of area used to store ID number Task key ID number Explanation This system call stores the task ID number specified by keyid in the area specified by p_tskid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The vget_tid system call is not defined as CF. Invalid parameter specification * The address of the area used to store the ID number is invalid (p_tskid = 0). * Invalid key ID number specification (keyid = 0) *E_NOEXS -52 The target task does not exist. 120 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.8.2 Task-associated synchronization system calls This section explains the group of system calls that perform the synchronous operations associated with tasks (task-associated synchronization system calls). Table 11-6 lists the task-associated synchronization system calls. Table 11-6. Task-Associated Synchronization System Calls System Call sus_tsk rsm_tsk frsm_tsk slp_tsk tslp_tsk wup_tsk can_wup Function Places another task in the suspend state. Restarts a task in the suspend state. Forcibly restarts a task in the suspend state. Places the task that issued this system call into the wake-up wait state. Places the task that issued this system call into the wake-up wait state (with timeout). Wakes up another task. Invalidates a request to wake up a task. User's Manual U13773EJ3V0UM 121 CHAPTER 11 SYSTEM CALLS Suspend Task (-33) sus_tsk Task/non-task Overview Places another task in the suspend state. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call issues a suspend request to the task specified by tskid (the suspend request counter is incremented by 0x1). If the target task is in the ready or wait state when this system call is issued, this system call changes the target task from the ready state to the suspend state or from the wait state to the wait-suspend state, and also issues a suspend request (increments the suspend request counter). Caution The suspend request counter managed by the RX850 Pro consists of seven bits. Therefore, once the number of suspend requests exceeds 127, the sus_tsk system call returns E_QOVR as the return value without incrementing the suspend request counter. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ 0 -17 -35 -52 -63 Normal termination The sus_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. Invalid state of the specified task * The target task is in the dormant state. * The issuing task is specified as the target task when the sus_tsk system call is issued from a task. E_OACV *E_QOVR -66 -73 An unauthorized ID number (tskid 0) was specified. The number of suspend requests exceeded 127. 122 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Resume Task (-35) rsm_tsk Task/non-task Overview Restarts a task in the suspend state. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call cancels only one of the suspend requests that are issued to the task specified by tskid (the suspend request counter is decremented by 0x1). If the issuance of this system call causes the suspend request counter for the target task to be 0x0, this system call changes the task from the suspend state to the ready state or from the wait-suspend state to the wait state. Caution This system call does not queue cancel requests. Accordingly, if a target task is not in the suspend or wait-suspend state, this system call returns E_OBJ as the return value without decrementing the suspend request counter. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ E_OACV 0 -17 -35 -52 -63 -66 Normal termination The rsm_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is not in the suspend or wait-suspend state. An unauthorized ID number (tskid 0) was specified. User's Manual U13773EJ3V0UM 123 CHAPTER 11 SYSTEM CALLS Force Resume Task (-36) frsm_tsk Task/non-task Overview Forcibly restarts a task in the suspend state. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call cancels all the suspend requests issued to the task specified by tskid (the suspend request counter is set to 0x0). The target task changes from the suspend state to the read state or from the wait-suspend state to the wait state. Caution This system call does not queue cancel requests. Accordingly, if a target task is not in the suspend or wait-suspend state, this system call returns E_OBJ as the return value without setting the suspend request counter. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ E_OACV 0 -17 -35 -52 -63 -66 Normal termination The frsm_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. The target task is not in the suspend or wait-suspend state. An unauthorized ID number (tskid 0) was specified. 124 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Sleep Task (-38) slp_tsk Task Overview Places the task that issued this system call into the wake-up wait state. C format #include ER Parameter None. Explanation This system call cancels only one of the wake-up requests issued to the task (the wake-up request counter is decremented by 0x1). If the wake-up request counter for the task is 0x0 when this system call is issued, this system call changes the state of the task from the run state to the wait state (wake-up wait state) without canceling a wake-up request (decrementing the wake-up request counter). The wake-up wait state is released when the wup_tsk, ret_wup, or rel_wai system call is issued. changes from the wake-up wait state to the ready state. Return value The task *E_OK *E_NOSPT E_CTX 0 -17 -69 Normal termination The slp_tsk system call is not defined as CF. Context error * The slp_tsk system call was issued from a non-task. * The slp_tsk system call was issued in the dispatch disabled state. *E_RLWAI -86 The wake-up wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 125 CHAPTER 11 SYSTEM CALLS Sleep Task with Timeout (-37) tslp_tsk Task Overview Places the task that issued this system call into the wake-up wait state (with timeout). C format #include ER Parameter I/O I TMO Parameter tmout; Wait time (unit: ms) TMO_POL(0): TMO_FEVR(-1): Value: Description Quick return Permanent wait Wait time Explanation This system call cancels only one of the wake-up requests issued to the task (the wake-up request counter is decremented by 0x1). If the wake-up request counter for the task is 0x0 when this system call is issued, this system call changes the task from the run state to the wait state (wake-up wait state) without canceling a wake-up request (decrementing the wake-up request counter). Note that the wake-up wait state is canceled if the wait time specified by tmout elapses or if the wup_tsk, ret_wup, or rel_wai system call is issued, and the issuing task changes to the ready state. Return value *E_OK *E_NOSPT E_PAR E_CTX 0 -17 -33 -69 Normal termination The tslp_tsk system call is not defined as CF. Invalid wait time specification (tmout < TMO_FEVR) Context error * The tslp_tsk system call was issued from a non-task. * The tslp_tsk system call was issued in the dispatch disabled state. *E_TMOUT *E_RLWAI -85 -86 The wait time has elapsed. The wake-up wait state was forcibly released by the rel_wai system call. 126 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Wakeup Task (-39) wup_tsk Task/non-task Overview Wakes up another task. C format #include ER Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call issues a wake-up request to the task specified by tskid (the wake-up request counter is incremented by 0x1). If the target task is in the wait state (wake-up wait state) when this system call is issued, this system call changes the task from the wake-up wait state to the ready state without issuing a wake-up request (incrementing the wakeup request counter). Caution The wake-up request counter managed by the RX850 Pro consists of 7-bits. Therefore, when the number of wake-up requests exceeds 127, the wup_tsk system call returns E_QOVR as the return value without incrementing the wake-up request counter. Return value *E_OK *E_NOSPT E_ID *E_NOEXS *E_OBJ 0 -17 -35 -52 -63 Normal termination The wup_tsk system call is not defined as CF. Invalid ID number specification (maximum number of tasks created < tskid) The target task does not exist. Invalid state of the specified task * The target task is in the dormant state. * The issuing task is specified as the target task when the wup_tsk system call is issued from a task. E_OACV *E_QOVR -66 -73 An unauthorized ID number (tskid 0) was specified. The number of wake-up requests exceeded 127. User's Manual U13773EJ3V0UM 127 CHAPTER 11 SYSTEM CALLS Cancel Wakeup Task (-40) can_wup Task/non-task Overview Invalidates a request to wake up a task. C format #include ER Parameters I/O O I INT ID Parameter *p_wupcnt; tskid; Description Address of area used to store the number of wake-up requests Task ID number TSK_SELF(0): Value: Local task Task ID number Explanation This system call cancels all the wake-up requests issued to the task specified by tskid (the wake-up request counter is set to 0x0). The number of wake-up requests canceled by this system call is stored in the area specified by p_wupcnt. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The can_wup system call is not defined as CF. The address of the area used to store the number of wake-up requests is invalid (p_wupcnt = 0). E_ID -35 Invalid ID number specification * Maximum number of tasks created < tskid * When the can_wup system call was issued from a non-task, TSK_SELF was specified for tskid. *E_NOEXS *E_OBJ E_OACV -52 -63 -66 The target task does not exist. The target task is in the dormant state. An unauthorized ID number (tskid < 0) was specified. 128 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.8.3 Synchronous communication system calls This section explains the group of system calls that are used for synchronization (exclusive control and queuing) and communication between tasks (synchronous communication system calls). Table 11-7 lists the synchronous communication system calls. Table 11-7. Synchronous Communication System Calls System Call cre_sem del_sem sig_sem wai_sem preq_sem twai_sem ref_sem vget_sid cre_flg del_flg set_flg clr_flg wai_flg pol_flg twai_flg ref_flg vget_fid cre_mbx del_mbx snd_msg rcv_msg prcv_msg trcv_msg ref_mbx vget_mid Creates a semaphore. Deletes a semaphore. Returns resources. Acquires resources. Acquires resources (polling). Acquires resources (with timeout). Acquires semaphore information. Acquires a semaphore ID number. Creates an event flag. Deletes an event flag. Sets a bit pattern. Clears a bit pattern. Checks a bit pattern. Checks a bit pattern (polling). Checks a bit pattern (with timeout). Acquires event flag information. Acquires an event flag ID number. Creates a mailbox. Deletes a mailbox. Transmits a message. Receives a message. Receives a message (polling). Receives a message (with timeout). Acquires mailbox information. Acquires a mailbox ID number. Function User's Manual U13773EJ3V0UM 129 CHAPTER 11 SYSTEM CALLS Create Semaphore (-49) cre_sem Task Overview Creates a semaphore. C format * When an ID number is specified #include ER * When an ID number is not specified #include ER Parameters I/O I I O ID T_CSEM ID Parameter semid; *pk_csem; *p_semid; Semaphore ID number Description Start address of packet containing semaphore creation information Address of area used to store ID number * Structure of semaphore creation information T_CSEM typedef struct VP ATR INT INT ID } T_CSEM; t_csem { exinf; sematr; isemcnt; maxsem; keyid; /* /* /* /* /* Extended information Semaphore attribute Initial semaphore resource count Maximum semaphore resource count Semaphore key ID number */ */ */ */ */ 130 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Explanation The RX850 Pro provides two types of interfaces for semaphore creation: one in which an ID number must be specified, and one in which an ID number is not specified. * When an ID number is specified A semaphore having an ID number specified by semid is created based on the information specified by pk_csem. * When an ID number is not specified A semaphore is created based on the information specified by pk_csem. An ID number is allocated by the RX850 Pro and the allocated ID number is stored in the area specified by p_semid. Semaphore creation information is described in detail below. exinf ... Extended information An area for storing user-specific information on a target semaphore. The user can use this area as required. Information set in exinf can be dynamically acquired by issuing the ref_sem system call from a processing program (tasks and non-tasks). sematr ... Semaphore attribute Bit 0 .. Method of queuing into a wait queue TA_TPRI(0): TA_TFIFO(1): Bit 8 .. Priority order FIFO order Existence of specifying the key ID number TA_KEYID(1): Specifies the key ID number 15 sematr 8 0 Method of queuing into a wait queue Existence of key ID number specification isemcnt maxsem keyid ... ... ... Initial semaphore resource count Maximum semaphore resource count Semaphore key ID number Remark If the value of bit 8 is not TA_KEYID in sematr, the contents of keyid are meaningless. User's Manual U13773EJ3V0UM 131 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOMEM *E_NOSPT E_RSATR E_PAR 0 -10 -17 -24 -33 Normal termination The semaphore management block area cannot be secured. The cre_sem system call is not defined as CF. Invalid specification of attribute sematr Invalid parameter specification * The start address of a packet storing semaphore creation information is invalid (pk_csem = 0). * The initial resource count is invalid (isemcnt < 0). * The maximum resource count is invalid (maxsem 0, maxsem < isemcnt). * Invalid key ID number specification (keyid = 0) (when TA_KEYID attribute specified) * The address of the area used to store an ID number is invalid (p_semid = 0). (When a semaphore is created without an ID number specified) E_ID *E_OBJ E_OACV E_CTX -35 -63 -66 -69 Invalid ID number specification (maximum number of semaphores created < semid) A semaphore having the specified ID number has already been created. An unauthorized ID number (semid 0) was specified. The cre_sem system call was issued from a non-task. 132 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Delete Semaphore (-50) del_sem Task Overview Deletes a semaphore. C format #include ER Parameter I/O I ID Parameter semid; Semaphore ID number Description Explanation This system call deletes the semaphore specified by semid. The target semaphore is released from the control of the RX850 Pro. The task released from the wait state (resource wait state) by the del_sem system call has E_DLT returned as the return value of the system call (wai_sem or twai_sem) that initiated transition to the wait state. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV E_CTX 0 -17 -35 -52 -66 -69 Normal termination The del_sem system call is not defined as CF. Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. The del_sem system call was issued from a non-task. User's Manual U13773EJ3V0UM 133 CHAPTER 11 SYSTEM CALLS Signal Semaphore (-55) sig_sem Task/non-task Overview Returns resources. C format #include ER Parameter I/O I ID Parameter semid; Semaphore ID number Description Explanation This system call returns resources to the semaphore specified by semid (the semaphore counter is incremented by 0x1). If tasks are queued in the wait queue of the target semaphore when this system call is issued, this system call passes the resources to the relevant task (the first task in the wait queue) without returning the resources (incrementing the semaphore counter). Consequently, the relevant task is removed from the wait queue, and its state changes from the wait state (resource wait state) to the ready state, or from the wait-suspend state to the suspend state. Caution The semaphore counter managed by the RX850 Pro counts up to the maximum number of resources that can be acquired as specified at the time it is created. Therefore, when the number of resources exceeds the maximum number of resources, by issuing the sig_sem system call, E_QOVR is returned as the return value without incrementing the semaphore counter. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV *E_QOVR 0 -17 -35 -52 -66 -73 Normal termination The sig_sem system call is not defined as CF. Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. The resource count exceeded the maximum resource count specified at creation. 134 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Wait on Semaphore (-53) wai_sem Task Overview Acquires resources. C format #include ER Parameter I/O I ID Parameter semid; Semaphore ID number Description Explanation This system call acquires resources from the semaphore specified by semid (the semaphore counter is decremented by 0x1). When this system call is issued, if no resource can be acquired from a target semaphore (when there are no free resources), this system call places the task in the wait queue of the specified semaphore, then changes it from the run state to the wait state (resource wait state). The resource wait state is released upon the issuance of the sig_sem, del_sem, or rel_wai system call, and the task returns to the ready state. Remark When a task queues in the wait queue of the target semaphore, it is executed in the order (FIFO order or priority order) specified when that semaphore was created (at configuration or when a cre_sem system call was issued). Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV E_CTX 0 -17 -35 -52 -66 -69 Normal termination The wai_sem system call is not defined as CF. Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. Context error * The wai_sem system call was issued from a non-task. * The wai_sem system call was issued in the dispatch disabled state. *E_DLT *E_RLWAI -81 -86 The specified semaphore was deleted by the del_sem system call. The resource wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 135 CHAPTER 11 SYSTEM CALLS Poll and Request Semaphore (-107) preq_sem Task/non-task Overview Acquires resources (polling). C format #include ER Parameter I/O I ID Parameter semid; Semaphore ID number Description Explanation This system call acquires resources from the semaphore specified by semid (the semaphore counter is decremented by 0x1). When this system call is issued, if no resource can be acquired from a target semaphore (when there are no free resources), this system call returns E_TMOUT as the return value. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV *E_TMOUT 0 -17 -35 -52 -66 -85 Normal termination The preq_sem system call is not defined as CF. Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. The resource count for the target semaphore is 0x0. 136 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Wait on Semaphore with Timeout (-171) twai_sem Task Overview Acquires resources (with timeout). C format #include ER Parameters I/O I I ID TMO Parameter semid; tmout; Semaphore ID number Wait time (unit: ms) TMO_POL(0): TMO_FEVR(-1): Value: Quick return Permanent wait Wait time Description Explanation This system call acquires resources from the semaphore specified by semid (the semaphore counter is decremented by 0x1). When this system call is issued, if no resource can be acquired from a target semaphore (when there are no free resources), this system call places the task in the wait queue of the target semaphore, then changes it from the run state to the wait state (resource wait state). The resource wait state is released when the wait time specified by tmout elapses or when the sig_sem, del_sem, or rel_wai system call is issued, at which time it changes to the ready state. Remark The task is queued into the wait queue of a target semaphore in the order (FIFO order or priority order) specified when the semaphore was created (at configuration or upon the issuance of the cre_sem system call). Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV E_CTX 0 -17 -33 -35 -52 -66 -69 Normal termination The twai_sem system call is not defined as CF. Invalid wait time specification (tmout < TMO_FEVR) Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. Context error * The twai_sem system call was issued from a non-task. * The twai_sem system call was issued in the dispatch disabled state. A target semaphore was deleted by the del_sem system call. Wait time elapsed. The resource wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM *E_DLT *E_TMOUT *E_RLWAI -81 -85 -86 137 CHAPTER 11 SYSTEM CALLS Refer Semaphore Status (-52) ref_sem Task/non-task Overview Acquires semaphore information. C format #include ER Parameters I/O O I T_RSEM ID Parameter *pk_rsem; semid; Description Start address of packet used to store semaphore information Semaphore ID number * Structure of semaphore information T_RSEM typedef struct VP BOOL_ID INT INT ID } T_RSEM; Explanation This system call stores the semaphore information (extended information, existence of waiting task, etc.) for the semaphore specified by semid in the packet specified by pk_rsem. Semaphore information is described in detail below. exinf wtsk ... ... Extended information Existence of waiting task FALSE(0): There is no waiting task Value: ID number of first task in wait queue Current resource count Maximum resource count specified at creation Key ID number FALSE(0): No key ID number specified at creation Value: Key ID number t_rsem { exinf; wtsk; semcnt; maxsem; keyid; /* /* /* /* /* Extended information Existence of waiting task Current resource count Maximum resource count Key ID number */ */ */ */ */ semcnt maxsem keyid ... ... ... 138 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV 0 -17 -33 -35 -52 -66 Normal termination The ref_sem system call is not defined as CF. The start address of the packet used to store semaphore information is invalid (pk_rsem = 0). Invalid ID number specification (maximum number of semaphores created < semid) The target semaphore does not exist. An unauthorized ID number (semid 0) was specified. User's Manual U13773EJ3V0UM 139 CHAPTER 11 SYSTEM CALLS Get Semaphore Identifier (-246) vget_sid Task/non-task Overview Acquires the semaphore ID number. C format #include ER Parameters I/O O I ID ID Parameter *p_semid; keyid; Description Address of an area used to store an ID number Semaphore key ID number Explanation This system call stores the semaphore ID number specified by keyid in the area specified by p_semid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The vget_sid system call is not defined as CF. The start address of the packet used to store semaphore information is invalid (pk_rsem = 0). * Invalid key ID number specification (keyid = 0) * The address of the area used to store the ID number is invalid (p_semid = 0). *E_NOEXS -52 The target semaphore does not exist. 140 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Create Event Flag (-41) cre_flg Task Overview Creates an event flag. C format * When an ID number is specified #include ER * When an ID number is not specified #include ER Parameters I/O I I O ID T_CFLG ID Parameter flgid; *pk_cflg; *p_flgid; Event flag ID number Description Start address of packet storing event flag creation information Address of area used to store ID number * Structure of event flag creation information T_CFLG typedef struct VP ATR UINT ID } T_CFLG; t_cflg { exinf; flgatr; iflgptn; keyid; /* /* /* /* Extended information Event flag attribute Initial bit pattern of event flag Event flag key ID number */ */ */ */ Explanation The RX850 Pro provides two types of interfaces for event flag creation: one in which an ID number must be specified and one in which an ID number is not specified. * When an ID number is specified An event flag having the ID number specified by flgid is created based on the information specified by pk_cflg. * When an ID number is not specified An event flag is created based on the information specified by pk_cflg. An ID number is allocated by the RX850 Pro and the allocated ID number is stored in the area specified by p_flgid. User's Manual U13773EJ3V0UM 141 CHAPTER 11 SYSTEM CALLS Event flag creation information is described in detail below. exinf ... Extended information exinf is an area used for storing user-specific information on a target event flag. The user can use this area as required. Information set in exinf can be dynamically acquired by issuing the ref_flg system call from a processing program (task or non-task). flgatr ... Event flag attribute Bit 3 .. Number of tasks that can be queued into a wait queue TA_WSGL(0): TA_WMUL(1): Bit 8 One task only Two or more tasks .. Existence of key ID number specification TA_KEYID(1): Key ID number specified 15 flgatr 8 3 0 Number of tasks that can be queued into a wait queue Existence of key ID number specification iflgptn keyid Remark ... ... Initial bit pattern of event flag Event flag key ID number If the value of bit 8 is not TA_KEYID in flgatr, the contents of keyid are meaningless. Return value *E_OK *E_NOMEM *E_NOSPT E_RSATR E_PAR 0 -10 -17 -24 -33 Normal termination The event flag management block area cannot be secured. The cre_flg system call is not defined as CF. Invalid specification of attribute flgatr Invalid parameter specification * The start address of the packet storing event flag creation information is invalid (pk_cflg = 0). * Invalid key ID number specification (keyid = 0) (when TA_KEYID attribute specified) * The address of the area used to store the ID number is invalid (p_flgid = 0). (When an event flag is created with no ID number specified) E_ID *E_OBJ E_OACV E_CTX -35 -63 -66 -69 Invalid ID number specification (maximum number of event flags created < flgid) An event flag having the specified ID number has already been created. An unauthorized ID number (flgid 0) was specified. The cre_flg system call was issued from a non-task. 142 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Delete Event Flag (-42) del_flg Task Overview Deletes an event flag. C format #include ER Parameter I/O I ID Parameter flgid; Event flag ID number Description Explanation This system call deletes the event flag specified by flgid. The target event flag is released from the control of the RX850 Pro. The task released from the wait state (event flag wait state) by this system call has E_DLT returned as the return value of the system call (wai_flg or twai_flg) that initiated transition to the wait state. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV E_CTX 0 -17 -35 -52 -66 -69 Normal termination The del_flg system call is not defined as CF. Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. An unauthorized ID number (flgid 0) was specified. The del_flg system call was issued from a non-task. User's Manual U13773EJ3V0UM 143 CHAPTER 11 SYSTEM CALLS Set Event Flag (-48) set_flg Task/non-task Overview Sets a bit pattern. C format #include ER Parameters I/O I I ID UINT Parameter flgid; setptn; Event flag ID number Bit pattern to be set (32-bit width) Description Explanation This system call executes a logical OR between the bit pattern specified by flgid and that specified by setptn, and sets the result in the specified event flag. For example, when this system call is issued, if the target event flag's bit pattern is B'1100 and the bit pattern specified by setptn is B'1010, the bit pattern of the target event flag becomes B'1110. When this system call is issued, if the wait condition for a task queued in the wait queue of the target event flag is satisfied, the task is removed from the wait queue. Consequently, the relevant task changes from the wait state (event flag wait state) to the ready state, or from the wait-suspend state to the suspend state. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV 0 -17 -35 -52 -66 Normal termination The set_flg system call is not defined as CF. Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. An unauthorized ID number (flgid 0) was specified. 144 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Clear Event Flag (-47) clr_flg Task/non-task Overview Clears a bit pattern. C format #include ER Parameters I/O I I ID UINT Parameter flgid; clrptn; Event flag ID number Description Bit pattern to be cleared (32-bit width) Explanation This system call executes a logical AND between the bit pattern specified by flgid and that specified by clrptn, and sets the result in the specified event flag. For example, when this system call is issued, if the target event flag's bit pattern is B'1100 and the bit pattern specified by clrptn is B'1010, the target event flag's bit pattern becomes B'1000. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV 0 -17 -35 -52 -66 Normal termination The clr_flg system call is not defined as CF. Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. An unauthorized ID number (flgid 0) was specified. User's Manual U13773EJ3V0UM 145 CHAPTER 11 SYSTEM CALLS Wait Event Flag (-46) wai_flg Task Overview Checks a bit pattern. C format #include ER Parameters I/O O I I I UINT ID UINT UINT Parameter *p_flgptn; flgid; waiptn; wfmode; Description Address of area used to store bit pattern when condition is satisfied Event flag ID number Request bit pattern (32-bit width) Wait condition or condition satisfaction TWF_ANDW(0): TWF_ORW(2): TWF_CLR(1): AND wait OR wait Bit pattern is cleared Explanation This system call checks whether a bit pattern that satisfies the request bit pattern specified by waiptn, as well as the wait condition specified by wfmode, is set in the event flag specified by flgid. If a bit pattern satisfying the wait condition is set in the target event flag, this system call stores the bit pattern of the event flag in the area specified by p_flgptn. When this system call is issued, if the bit pattern of the target event flag does not satisfy the wait condition, this system call queues the task at the end of the wait queue for the target event flag, then changes it from the run state to the wait state (event flag wait state). The event flag wait state is released when a bit pattern satisfying the wait condition is set by the set_flg system call, or when the del_flg or rel_wai system call is issued, at which time it changes to the ready state. The specification format for wfmode is shown below. * wfmode = TWF_ANDW This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. * wfmode = (TWF_ANDW|TWF_CLR) This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. If the wait condition is satisfied, the bit pattern for the target event flag is cleared (B'0000 is set). 146 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS * wfmode = TWF_ORW This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. * wfmode = (TWF_ORW|TWF_CLR) This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. If the wait condition is satisfied, the bit pattern of the target event flag is cleared (B'0000 is set). Cautions 1. The RX850 Pro specifies the number of tasks that can be queued into the wait queue of an event flag at creation (at configuration or upon the issuance of the cre_flg system call). TA_WSGL attribute: Only one task can be queued. TA_WMUL attribute: Two or more tasks can be queued. For this reason, if this system call is issued for the event flag having the TA_WSGL attribute for which waiting tasks are already queued, the wai_flg system call returns E_OBJ as the return value without performing bit pattern checking. 2. If the event flag wait state is forcibly released by issuing the del_flg or rel_wai system call, the contents of the area specified by p_flgptn will be undefined. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The wai_flg system call is not defined as CF. Invalid parameter specification * The address of the area used to store a bit pattern when a condition is satisfied is invalid (p_flgptn = 0). * Invalid specification of request bit pattern (waiptn = 0). * Invalid specification of wait condition or condition satisfaction parameter wfmode. E_ID *E_NOEXS *E_OBJ -35 -52 -63 Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. The wai_flg system call was issued for the event flag having the TA_WSGL attribute for which waiting tasks were already queued. An unauthorized ID number (flgid 0) was specified. Context error * The wai_flg system call was issued from a non-task. * The wai_flg system call was issued from the dispatch disabled state. E_OACV E_CTX -66 -69 *E_DLT *E_RLWAI -81 -86 The target event flag was deleted by the del_flg system call. The event flag wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 147 CHAPTER 11 SYSTEM CALLS Poll Event Flag (-106) pol_flg Task/non-task Overview Checks a bit pattern (polling). C format #include ER Parameters I/O O I I I UINT ID UINT UINT Parameter *p_flgptn; flgid; waiptn; wfmode; Description Address of area used to store bit pattern when condition is satisfied Event flag ID number Request bit pattern (32-bit width) Wait condition or condition satisfaction TWF_ANDW(0): TWF_ORW(2): TWF_CLR(1): AND wait OR wait Bit pattern is cleared. Explanation This system call checks whether a bit pattern satisfying both the request bit pattern specified by waiptn and the wait condition specified by wfmode is set in the event flag specified by flgid. If a bit pattern satisfying the wait condition is set in the target event flag, this system call stores the bit pattern of the event flag into the area specified by p_flgptn. When this system call is issued, if the bit pattern of the target event flag does not satisfy the wait condition, this system call returns E_TMOUT as the return value. The wfmode specification format is shown below. * wfmode = TWF_ANDW This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. * wfmode = (TWF_ANDW|TWF_CLR) This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. If the wait condition is satisfied, the bit pattern for the specified event flag is cleared (B'0000 is set). 148 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS * wfmode = TWF_ORW This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. * wfmode = (TWF_ORW|TWF_CLR) This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. If the wait condition is satisfied, the bit pattern for the target event flag is cleared (B'0000 is set). Caution The RX850 Pro specifies the number of tasks that can be queued into the wait queue of an event flag at creation (at configuration or upon the issuance of the cre_flg system call). TA_WSGL attribute: Only one task can be queued. TA_WMUL attribute: Two or more tasks can be queued. For this reason, if this system call is issued for an event flag having the TA_WSGL attribute for which waiting tasks are already queued, the wai_flg system call returns E_OBJ as the return value without performing bit pattern checking. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The pol_flg system call is not defined as CF. Invalid parameter specification * The address of the area used to store a bit pattern when a condition is satisfied is invalid (p_flgptn = 0). * Invalid specification of request bit pattern (waiptn = 0). * Invalid specification of wait condition or condition satisfaction parameter wfmode. E_ID *E_NOEXS *E_OBJ -35 -52 -63 Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. This pol_flg system call was issued for the event flag of TA_WSGL attribute for which waiting tasks are already queued. An unauthorized ID number (flgid 0) was specified. The bit pattern of the target event flag does not satisfy the wait condition. E_OACV *E_TMOUT -66 -85 User's Manual U13773EJ3V0UM 149 CHAPTER 11 SYSTEM CALLS Wait Event Flag with Timeout (-170) twai_flg Task Overview Checks a bit pattern (with timeout). C format #include ER Parameters I/O O I I I UINT ID UINT UINT Parameter *p_flgptn; flgid; waiptn; wfmode; Description Address of area used to store bit pattern when condition is satisfied Event flag ID number Request bit pattern (32-bit width) Wait condition or condition satisfaction TWF_ANDW(0): TWF_ORW(2): TWF_CLR(1): AND wait OR wait Bit pattern is cleared. I TMO tmout; Wait time (unit: ms) TMO_POL(0): TMO_FEVR(-1): Value: Quick return Permanent wait Wait time Explanation This system call checks whether a bit pattern satisfying both the request bit pattern specified by waiptn and the wait condition specified by wfmode is set in the event flag specified by flgid. If a bit pattern satisfying the wait condition is set in the target event flag, this system call stores the bit pattern of the event flag in the area specified by p_flgptn. Upon the issuance of this system call, if the bit pattern of the target event flag does not satisfy the wait condition, this system call queues the task at the end of the wait queue for the target event flag, then changes it from the run state to the wait state (event flag wait state). The event flag wait state is released upon the elapse of the wait time specified by tmout, when a bit pattern satisfying the wait condition is set by the set_flg system call, or when the del_flg or rel_wai system call is issued, at which time the task returns to the ready state. The wfmode specification format is shown below. 150 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS * wfmode = TWF_ANDW This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. * wfmode = (TWF_ANDW|TWF_CLR) This system call checks whether all the bits of waiptn that are set to 1 are set in the target event flag. If the wait condition is satisfied, the bit pattern for the target event flag is cleared (B'0000 is set). * wfmode = TWF_ORW This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. * wfmode = (TWF_ORW|TWF_CLR) This system call checks whether at least one of the bits of waiptn that are set to 1 is set in the target event flag. If the wait condition is satisfied, the bit pattern of the target event flag is cleared (B'0000 is set). Cautions 1. The RX850 Pro specifies the number of tasks that can be queued into the wait queue of the event flag at creation (at configuration or upon the issuance of the cre_flg system call). TA_WSGL attribute: Only one task can be queued. TA_WMUL attribute: Two or more tasks can be queued. For this reason, if this system call is issued for an event flag having the TA_WSGL attribute for which waiting tasks are already queued, this system call returns E_OBJ as the return value without performing bit pattern checking. 2. If the event flag wait state is forcibly released by the del_flg or rel_wai system call, the contents of the area specified by p_flgptn will be undefined. User's Manual U13773EJ3V0UM 151 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The twai_flg system call is not defined as CF. Invalid parameter specification * The address of the area used to store the bit pattern when the condition is satisfied is invalid (p_flgptn = 0). * The specification of the request bit pattern is invalid (waiptn = 0). * The specification of the wait condition or condition satisfaction parameter wfmode is invalid. * Invalid wait time specification (tmout < TMO_FEVR) E_ID *E_NOEXS *E_OBJ -35 -52 -63 Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. This twai_flg system call was issued for the event flag having the TA_WSGL attribute in which waiting tasks were already queued. An unauthorized ID number (flgid 0) was specified. Context error * The twai_flg system call was issued from a non-task. * The twai_flg system call was issued from the dispatch disabled state. E_OACV E_CTX -66 -69 *E_DLT *E_TMOUT *E_RLWAI -81 -85 -86 The specified event flag was deleted by the del_flg system call. Wait time elapsed. The event flag wait state was forcibly released by the rel_wai system call. 152 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Refer Event Flag Status (-44) ref_flg Task/non-task Overview Acquires event flag information. C format #include ER Parameters I/O O I T_RFLG ID Parameter *pk_rflg; flgid; Description Start address of packet used to store event flag information Event flag ID number * Structure of event flag information T_RFLG typedef struct VP BOOL_ID UINT ID } T_RFLG; Explanation This system call stores the event flag information (extended information, existence of waiting task, etc.) for the event flag specified by flgid in the packet specified by pk_rflg. Event flag information is described in detail below. exinf wtsk ... ... Extended information Existence of waiting task FALSE(0): Value: flgptn keyid ... ... There is no waiting task. ID number of first task in wait queue t_rflg { exinf; wtsk; flgptn; keyid; /* /* /* /* Extended information Existence of waiting task Current bit pattern Key ID number */ */ */ */ Current bit pattern Key ID number FALSE(0): Value: No key ID number specified at generation Key ID number User's Manual U13773EJ3V0UM 153 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV 0 -17 -33 -35 -52 -66 Normal termination The ref_flg system call is not defined as CF. The start address of the packet used to store event flag information is invalid (pk_rflg = 0). Invalid ID number specification (maximum number of event flags created < flgid) The target event flag does not exist. An unauthorized ID number (flgid 0) was specified. 154 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Get Event Flag Identifier (-247) vget_fid Task/non-task Overview Acquires the event flag ID number. C format #include ER Parameters I/O O I ID ID Parameter *p_flgid; keyid; Description Address of area used to store ID number Event flag key ID number Explanation This system call stores the event flag ID number specified by keyid in the area specified by p_flgid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The vget_fid system call is not defined as CF. Invalid parameter specification * Invalid key ID number specification (keyid = 0) * The address of the area used to store the ID number is invalid (p_flgid = 0). *E_NOEXS -52 The target event flag does not exist. User's Manual U13773EJ3V0UM 155 CHAPTER 11 SYSTEM CALLS Create Mailbox (-57) cre_mbx Task Overview Creates a mailbox. C format * When an ID number is specified #include ER * When an ID number is not specified #include ER Parameters I/O I I O ID T_CMBX ID Parameter mbxid; *pk_cmbx; *p_mbxid; Mailbox ID number Description Start address of packet used to store mailbox creation information Address of area used to store ID number * Structure of mailbox creation information T_CMBX typedef struct VP ATR ID } T_CMBX; t_cmbx { exinf; mbxatr; keyid; /* /* /* Extended information Mailbox attribute Mailbox key ID number */ */ */ Explanation The RX850 Pro provides two types of interfaces for mailbox creation: one in which an ID number must be specified for mailbox creation, and one in which an ID number is not specified. * When an ID number is specified A mailbox having the ID number specified by mbxid is created based on the information specified by pk_cmbx. * When an ID number is not specified A mailbox is created based on the information specified by pk_cmbx. An ID number is allocated by the RX850 Pro. The allocated ID number is stored in the area specified by p_mbxid. 156 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Mailbox creation information is described in detail below. exinf ... Extended information exinf is an area used for storing user-specific information on the target mailbox. The user can use this area as required. Information set in exinf can be dynamically acquired by issuing the ref_mbx system call from a processing program (task/non-task). mbxatr ... Mailbox attribute Bit 0 .. Method of queuing into a task wait queue TA_TPRI(0): TA_TFIFO(1): Bit 1 Priority order FIFO order .. Method of queuing into a message wait queue TA_MPRI(0): TA_MFIFO(1): Priority order FIFO order Bit 8 .. Existence of key ID number specification TA_KEYID(1): Key ID number specified 15 mbxatr 8 0 Method of queuing into a task wait queue Existence of key ID number specification Method of queuing into a message wait queue keyid ... Mailbox key ID number Remark If the value of bit 8 is not TA_KEYID in mbxatr, the contents of keyid are meaningless. User's Manual U13773EJ3V0UM 157 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOMEM *E_NOSPT E_RSATR E_PAR 0 -10 -17 -24 -33 Normal termination The mailbox management block area cannot be secured. The cre_mbx system call is not defined as CF. Invalid specification of attribute mbxatr Invalid parameter specification * The start address of the packet storing the mailbox creation information is invalid (pk_cmbx = 0). * The specification of the key ID number is invalid (keyid = 0) (when TA_KEYID specified). * The address of the area used to store the ID number is invalid (p_mbxid = 0). (When a mailbox is created without an ID number specified). E_ID *E_OBJ E_OACV E_CTX -35 -63 -66 -69 Invalid ID number specification (maximum number of mailboxes created < mbxid) A mailbox having the specified ID number has already been created. An unauthorized ID number (mbxid 0) was specified. The cre_mbx system call was issued from a non-task. 158 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Delete Mailbox (-58) del_mbx Task Overview Deletes a mailbox. C format #include ER Parameter I/O I ID Parameter mbxid; Mailbox ID number Description Explanation This system call deletes the mailbox specified by mbxid. The target mailbox is released from the control of the RX850 Pro. The task released from the wait state (message wait state) by this system call has E_DLT returned as the return value of the system call (rcv_msg or trcv_msg) that instigated the transition to the wait state. Remark When this system call is issued, any message using a memory block acquired from a memory pool is queued into the message wait queue of the target mailbox, and the message (memory block) is then returned to the memory pool. For this reason, if this system call uses an area other than memory blocks acquired from the memory pool, operation is not guaranteed. This system call should therefore not be issued in the above case. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV E_CTX 0 -17 -35 -52 -66 -69 Normal termination The del_mbx system call is not defined as CF. Invalid ID number specification (maximum number of mailboxes created < mbxid) The target mailbox does not exist. An unauthorized ID number (mbxid 0) was specified. The del_mbx system call was issued from a non-task. User's Manual U13773EJ3V0UM 159 CHAPTER 11 SYSTEM CALLS Send Message (-63) snd_msg Task/non-task Overview Transmits a message. C format #include ER Parameters I/O I I ID T_MSG Parameter mbxid; *pk_msg; Mailbox ID number Description Start address of packet used to store a message * Structure of message T_MSG typedef struct VW PRI VB } T_MSG; t_msg { msgrfu; msgpri; msgcont[]; /* /* /* Message management area Message priority Message body */ */ */ Explanation This system call transmits the message specified in pk_msg to the mailbox specified in mbxid (queues the message into a message wait queue). When this system call is issued, if a task is queued into the task wait queue of the target mailbox, this system call passes the message to the task (first task in the task wait queue) without performing message queuing. Consequently, the relevant task is removed from the task wait queue, and its state changes from the wait state (message wait state) to the ready state, or from the wait-suspend state to the suspend state. Remark When a message queues in the message wait queue of the target mailbox, it is executed in the order (FIFO order or priority order) specified when that mailbox was generated (at configuration or when the cre_mbx system call was issued). Caution The RX850 Pro uses the first four bytes (message management area msgrfu) of a message as a link area for enabling queuing into a message wait queue. Accordingly, transmitting a message to the target mailbox requires that 0x0 be set in msgrfu before issuing the snd_msg system call. If a value other than 0x0 is set in msgrfu when the snd_msg system call is issued, the RX850 Pro recognizes that the relevant message is already queued into a message wait queue, and this system call returns E_OBJ as the return value without transmitting the message. 160 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OBJ E_OACV 0 -17 -33 -35 -52 -63 -66 Normal termination The snd_msg system call is not defined as CF. The start address of the packet used to store the message is invalid (pk_msg = 0). Invalid ID number specification (maximum number of mailboxes created < mbxid) The target mailbox does not exist. The area specified for a message is already being used for messages. An unauthorized ID number (mbxid 0) was specified. User's Manual U13773EJ3V0UM 161 CHAPTER 11 SYSTEM CALLS Receive Message from Mailbox (-61) rcv_msg Task Overview Receives a message. C format #include ER Parameters I/O O I T_MSG ID Parameter **ppk_msg; mbxid; Description Address of area used to store start address of message Mailbox ID number Explanation This system call receives a message from the mailbox specified by mbxid and stores its start address in the area specified by ppk_msg. When this system call is issued, if a message cannot be received from the target mailbox (when no message exists in a message wait queue), this system call queues the task into the task wait queue of the target mailbox, then changes its state from the run state to the wait state (message wait state). The message wait state is released when the snd_msg, del_mbx, or rel_wai system call is issued, and the task returns to the ready state. Remark When a task queues in the task wait queue of the target mailbox, it is executed in the order (FIFO order or priority order) specified when that mailbox was created (at configuration or when the cre_mbx system call was issued). Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV E_CTX 0 -17 -33 -35 -52 -66 -69 Normal termination The rcv_msg system call is not defined as CF. The address of the area used to store the start address of a message is invalid (ppk_msg = 0). Invalid ID number specification (maximum number of mailboxes created < mbxid) The target mailbox does not exist. An unauthorized ID number (mbxid 0) was specified. Context error * The rcv_msg system call was issued from a non-task. * The rcv_msg system call was issued from the dispatch disabled state. *E_DLT *E_RLWAI -81 -86 The target mailbox was deleted by a del_mbx system call. The message wait state was forcibly released by the rel_wai system call. 162 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Poll and Receive Message from Mailbox (-108) prcv_msg Task/non-task Overview Receives a message (polling). C format #include ER Parameters I/O O I T_MSG ID Parameter **ppk_msg; mbxid; Description Address of area used to store the start address of a message Mailbox ID number Explanation This system call receives a message from the mailbox specified by mbxid and stores its start address in the area specified by ppk_msg. When this system call is issued, if a message cannot be received from the target mailbox (when no message exists in the message wait queue), E_TMOUT is returned as the return value. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The prcv_msg system call is not defined as CF. The address of the area used to store the start address of the message is invalid (ppk_msg = 0). E_ID *E_NOEXS E_OACV *E_TMOUT -35 -52 -66 -85 Invalid ID number specification (maximum number of mailboxes created < mbxid) A target mailbox does not exist. An unauthorized ID number (mbxid 0) was specified. No message exists in the target mailbox. User's Manual U13773EJ3V0UM 163 CHAPTER 11 SYSTEM CALLS Receive Message from Mailbox with Timeout (-172) trcv_msg Task Overview Receives a message (with timeout). C format #include ER Parameters I/O O I I T_MSG ID TMO Parameter **ppk_msg; mbxid; tmout; Description Address of area used to store start address of message Mailbox ID number Wait time (unit: basic clock cycles) TMO_POL(0): TMO_FEVR(-1): Value: Quick return Permanent wait Wait time Explanation This system call receives a message from the mailbox specified by mbxid and stores its start address in the area specified by ppk_msg. When this system call is issued, if a message cannot be received from the target mailbox (when no message exists in the message wait queue), this system call queues the task into the task wait queue of the target mailbox, then changes its state from the run state to the wait state (message wait state). The message wait state is released when the wait time specified by tmout elapses or when the snd_msg, del_mbx, or rel_wai system call is issued, and the task returns to the ready state. Remark When a task queues in the task wait queue of the target mailbox, it is executed in the order (FIFO order or priority order) specified when that mailbox was created (at configuration or when the cre_mbx system call was issued). 164 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The trcv_msg system call is not defined as CF. Invalid parameter specification * The address of the area used to store the start address of a message is invalid (ppk_msg = 0). E_ID *E_NOEXS E_OACV E_CTX -35 -52 -66 -69 Invalid ID number specification (maximum number of mailboxes created < mbxid) The target mailbox does not exist. An unauthorized ID number (mbxid 0) was specified. Context error * The trcv_msg system call was issued from a non-task. * The trcv_msg system call was issued from the dispatch disabled state. *E_DLT *E_TMOUT *E_RLWAI -81 -85 -86 The specified mailbox was deleted by the del_mbx system call. The wait time has elapsed. The message wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 165 CHAPTER 11 SYSTEM CALLS Refer Mailbox Status (-60) ref_mbx Task/non-task Overview Acquires mailbox information. C format #include ER Parameters I/O O I T_RMBX ID Parameter *pk_rmbx; mbxid; Description Start address of packet used to store mailbox information Mailbox ID number * Structure of mailbox information T_RMBX typedef struct VP BOOL_ID T_MSG ID } T_RMBX; t_rmbx { exinf; wtsk; *pk_msg; keyid; /* /* /* /* Extended information Existence of waiting task Existence of waiting message Key ID number */ */ */ */ Explanation This system call stores mailbox information (extended information, existence of waiting task, etc.) for the mailbox specified by mbxid into the packet specified by pk_rmbx. Mailbox information is described in detail below. exinf wtsk ... ... Extended information Existence of waiting task FALSE(0): Value: pk_msg ... No waiting task ID number of the first task of wait queue Existence of waiting message NADR(-1): Value: No waiting message Address of the first message of wait queue keyid ... Key ID number FALSE(0): Value: No key ID number specified at creation Key ID number 166 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV 0 -17 -33 -35 -52 -66 Normal termination The ref_mbx system call is not defined as CF. The start address of the packet used to store mailbox information is invalid (pk_rmbx = 0). Invalid ID number specification (maximum number of mailboxes created < mbxid) The target mailbox does not exist. An unauthorized ID number (mbxid 0) was specified. User's Manual U13773EJ3V0UM 167 CHAPTER 11 SYSTEM CALLS Get Mailbox Identifier (-245) vget_mid Task/non-task Overview Acquires the mailbox ID number. C format #include ER Parameters I/O O I ID ID Parameter *p_mbxid; keyid; Description Address of area used to store ID number Mailbox key ID number Explanation This system call stores the mailbox ID number specified by keyid in the area specified by p_mbxid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The vget_mid system call is not defined as CF. Invalid parameter specification * Invalid key ID number specification (keyid = 0) * The address of the area used to store the ID number is invalid (p_mbxid = 0). *E_NOEXS -52 The target mailbox does not exist. 168 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.8.4 Interrupt management system calls This section explains the group of system calls that perform processing that depends on maskable interrupts (interrupt management system calls). Table 11-8 lists the interrupt management system calls. Table 11-8. Interrupt Management System Calls System Call def_int ret_int ret_wup ena_int dis_int loc_cpu unl_cpu chg_icr ref_icr Function Registers an indirectly activated interrupt handler and cancels its registration. Returns from a directly activated interrupt handler. Wakes up another task and returns from a directly activated interrupt handler. Enables the acknowledgement of maskable interrupts. Disables the acknowledgement of maskable interrupts. Disables the acknowledgement of maskable interrupts and dispatch processing. Enables the acknowledgement of maskable interrupts and dispatch processing. Changes the interrupt control register. Acquires the interrupt control register. User's Manual U13773EJ3V0UM 169 CHAPTER 11 SYSTEM CALLS Define Interrupt Handler (-65) def_int Task/non-task Overview Registers an indirectly activated interrupt handler and cancels its registration. C format #include ER Parameters I/O I I UINT T_DINT Parameter eintno; *pk_dint; Description Interrupt request number of indirectly activated interrupt handler Start address of packet storing indirectly activated interrupt handler registration information * Structure of indirectly activated interrupt handler registration information T_DINT typedef struct t_dint { ATR FP VP VP } T_DINT; Explanation This system call uses the information specified by pk_dint to register the indirectly activated interrupt handler activated upon the occurrence of the maskable interrupt with the interrupt request number specified by eintno. Indirectly activated interrupt handler registration information is described in detail below. intatr ... Attribute of indirectly activated interrupt handler Bit 0 .. Language in which an indirectly activated interrupt handler is coded TA_ASM(0): TA_HLNG(1): Bit 10 Bit 11 .. .. TA_DPID(1): TA_DPIC(1): Assembly language C language Specific GP register value specified. Specific TP register value specified. intatr; inthdr; gp; tp; /* /* /* /* Attribute of indirectly activated interrupt handler Activation address of indirectly activated interrupt handler Specific GP register value for indirectly activated interrupt handler Specific TP register value for indirectly activated interrupt handler */ */ */ */ Existence of a specific GP register value specification Existence of a specific TP register value specification 170 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 15 intatr 8 7 0 Language in which an indirectly activated interrupt handler is coded Existence of a specific GP register value specification Existence of a specific TP register value specification inthdr gp tp ... ... ... Activation address of indirectly activated interrupt handler Specific GP register value for indirectly activated interrupt handler Specific TP register value for indirectly activated interrupt handler When this system call is issued, if an indirectly activated interrupt handler corresponding to the specified interrupt request number has already been registered, this system call does not handle this as an error and newly registers the specified indirectly activated interrupt handler. When this system call is issued, if NADR(-1) is set in the area specified by pk_dint, the registration of the interrupt handler specified by eintno is canceled. Remarks 1. If the value of bit 10 is not 1 (TA_DPID), the contents of gp are meaningless. 2. If the value of bit 11 is not 1 (TA_DPIC), the contents of tp are meaningless. Return value *E_OK *E_NOSPT E_RSATR E_PAR 0 -17 -24 -33 Normal termination The def_int system call is not defined as CF. Invalid specification of attribute intatr Invalid parameter specification * Invalid interrupt request number specification (eintno < 0, maximum interrupt source number < eintno) * The start address of the packet storing indirectly activated interrupt handler registration information is invalid (pk_dint = 0). * Invalid specification of the activation address (inthdr = 0) User's Manual U13773EJ3V0UM 171 CHAPTER 11 SYSTEM CALLS Return from Interrupt Handler (-69) ret_int Directly activated interrupt handler Overview Returns from a directly activated interrupt handler. C format #include void Parameter None. Explanation This system call returns from a directly activated interrupt handler. If a system call (chg_pri, sig_sem, etc.) requiring task scheduling is issued from a directly activated interrupt handler, the RX850 Pro merely queues the tasks into the wait queue and delays actual scheduling until a system call (ret_int or ret_wup) is issued to return from the directly activated interrupt handler, at which point the queued tasks are all processed in batch. Cautions 1. This system call does not notify the external interrupt controller of the termination of processing (issue of EOI command). Accordingly, for return from the directly activated interrupt handler activated by an external interrupt request, the external interrupt controller must be notified of termination before the issuance of this system call. 2. When describing a directly activated interrupt handler in assembly language, describe the following line for return from the directly activated interrupt handler: jr _ret_int However, for return using the macro RTOS_IntReturn provided in the RX850 Pro, no description is required because the ret_int system call is issued within the macro. 3. When describing an indirectly activated interrupt handler in C language, describe the following line for return from the indirectly activated interrupt handler: return (TSK_NULL); 4. When describing an indirectly activated interrupt handler in assembly language, describe the following line for return from the indirectly activated interrupt handler: mov jmp Return value None. TSK_NULL, r10 [lp] 172 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return and Wakeup Task (-70) ret_wup Directly activated interrupt handler Overview Wakes up another task and returns from a directly activated interrupt handler. C format #include void Parameter I/O I ID Parameter tskid; Task ID number Description Explanation This system call returns from a directly activated interrupt handler after the issuance of a wake-up request to the task specified by tskid (the wake-up request counter is incremented by 0x1). When this system call is issued, if the target task is in the wait state (wake-up wait state), this system call changes the target task from the wake-up wait state to the ready state without issuing a wake-up request (incrementing the wake-up request counter). If a system call (chg_pri, sig_sem, etc.) requiring task scheduling is issued from a directly activated interrupt handler, the RX850 Pro merely queues the tasks into a wait queue and delays the actual scheduling until a system call (issuing ret_wup or ret_int system call) is issued to return from the directly activated interrupt handler, at which point the queued tasks are all processed in batch. Cautions 1. This system call does not notify the external interrupt controller of processing termination (issue of the EOI command). Accordingly, for return from the directly activated interrupt handler activated by an external interrupt request, the external interrupt controller must be notified of processing termination before the issuance of this system macro. 2. In this system call, if the following types of errors occur, only processing for returning from the directly activated interrupt handler is performed. * Invalid ID number specification (maximum number of tasks created < tskid) * The target task does not exist. * The target task is in the dormant state. * The number of wake-up requests exceeded 127. User's Manual U13773EJ3V0UM 173 CHAPTER 11 SYSTEM CALLS 3. When describing a directly activated interrupt handler in assembly language, describe the following line for return from the directly activated interrupt handler: mov jr tskid, r10 _ret_wup However, for return using the macro RTOS_IntReturnWakeup provided in the RX850 Pro, no description is required because the ret_wup system call is issued within the macro. If this macro is used, describe as follows. RTOS_IntReturnWakeup r10 (r10 = task ID) 4. When describing an indirectly activated interrupt handler in C language, describe the following line for return from the indirectly activated interrupt handler: return (ID tskid); 5. When describing an indirectly activated interrupt handler in assembly language, describe the following line for return from a wake-up and the indirectly activated interrupt handler: mov jmp Return value None. tskid, r10 [lp] 174 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Enable Interrupt (-71) ena_int Task/non-task Overview Enables acknowledgement of maskable interrupts. C format #include ER Parameter None. Explanation This system call allows the resumption of acknowledgement of maskable interrupts that were disabled by issuing the dis_int system call. If a maskable interrupt occurs after the dis_int system call is issued before this ena_int system call is issued, the RX850 Pro delays switching to interrupt processing (a directly activated interrupt handler or an indirectly activated interrupt handler) until the ena_int system call is issued. Caution This system call does not queue resume requests. Therefore, if this system call has been issued already and acknowledgement of maskable interrupts has been enabled, no processing is executed and it is not treated as an error. Return value *E_OK *E_NOSPT 0 -17 Normal termination The ena_int system call is not defined as CF. User's Manual U13773EJ3V0UM 175 CHAPTER 11 SYSTEM CALLS Disable Interrupt (-72) dis_int Task/non-task Overview Disables acknowledgement of maskable interrupts. C format #include ER Parameter None. Explanation This system call disables the acknowledgement of maskable interrupts. This disables the acknowledgement of maskable interrupts before the ena_int system call is issued. If a maskable interrupt occurs after the dis_int system call is issued before this ena_int system call is issued, the RX850 Pro delays switching to interrupt processing (a directly activated interrupt handler or an indirectly activated interrupt handler) until the ena_int system call is issued. Caution This system call does not queue disable requests. Therefore, if this system call has been issued already and acknowledgement of maskable interrupts has been disabled, no processing is executed and it is not treated as an error. Return value *E_OK *E_NOSPT 0 -17 Normal termination The dis_int system call is not defined as CF. 176 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Lock CPU (-8) loc_cpu Task Overview Disables the acknowledgement of maskable interrupts and dispatch processing. C format #include ER Parameter None. Explanation This system call disables the acknowledgement of maskable interrupts and dispatch processing (task scheduling). Therefore, for the period of time from the issuance of this system call to the issuance of the unl_cpu, there is no transfer of control to another handler or task. If a maskable interrupt occurs after this system call is issued but before the unl_cpu system call is issued, the RX850 Pro delays processing for the interrupt (interrupt handler) until the unl_cpu system call is issued. If a system call (chg_pri, sig_sem, etc.) requiring task scheduling is issued, the RX850 Pro merely queues the tasks into a wait queue and delays the actual scheduling until the unl_cpu system call is issued, at which point all the tasks are processed in batch. Caution This system call does not queue disable requests. Therefore, if this system call has been issued already and acknowledgement of maskable interrupts and dispatch processing has been disabled, no processing is executed and it is not treated as an error. Return value *E_OK *E_NOSPT E_CTX 0 -17 -69 Normal termination The loc_cpu system call is not defined as CF. The loc_cpu system call was issued from a non-task. User's Manual U13773EJ3V0UM 177 CHAPTER 11 SYSTEM CALLS Unlock CPU (-7) unl_cpu Task Overview Enables the acknowledgement of maskable interrupts and dispatch processing. C format #include ER Parameter None. Explanation This system call allows the resumption of acknowledgement of maskable interrupts and dispatch processing (task scheduling) disabled by issuing the loc_cpu system call. If a maskable interrupt occurs after the loc_cpu system call is issued but before this system call is issued, the RX850 Pro delays processing for the interrupt (interrupt handler) until this system call is issued. If a system call (chg_pri, sig_sem, etc.) requiring task scheduling is issued, the RX850 Pro merely queues the tasks into a wait queue and delays actual scheduling until the unl_cpu system call is issued, at which point all the tasks are processed in batch. Remark Dispatch processing that was disabled by the issuance of the dis_dsp system call is reenabled by this system call. Caution This system call does not queue resume requests. Therefore, if this system call has been issued already and acknowledgement of maskable interrupts and dispatch processing has resumed, no processing is executed and it is not treated as an error. Return value *E_OK *E_NOSPT E_CTX 0 -17 -69 Normal termination The unl_cpu system call is not defined as CF. The unl_cpu system call was issued from a non-task. 178 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Change Interrupt Control Register (-67) chg_icr Task/non-task Overview Changes the contents of the interrupt control register. C format ER chg_icr(UINT eintno, UB icrcmd); Parameters I/O I I UINT UB Parameter eintno; icrcmd; Interrupt request number Description Specification of interrupt request flag ICR_CLRINT (0x20): No interrupt request Specification of interrupt mask flag ICR_CLRMSK (0x10): Enables interrupt processing ICR_SETMSK (0x40): Disables interrupt processing Specification of changing interrupt priority order ICR_CHGLVL (0x08): Changes interrupt priority order Specification of interrupt priority order Value (0 to 7): Interrupt priority order Explanation This system call changes the contents of the interrupt control register specified by eintno to the value specified by icrcmd. The icrcmd specification formats are as follows: * icrcmd = ICR_CLRINT Changes the interrupt request flag of the interrupt control register to 0. * icrcmd = ICR_CLRMSK Changes the interrupt mask flag of the interrupt control register to 0. * icrcmd = ICR_SETMSK Changes the interrupt mask flag of the interrupt control register to 1. * icrcmd = (ICR_CHGLVL | value) Changes the interrupt priority order of the interrupt control register to the value specified by "Value". The value "0" corresponds to level 0 and value "7" to level 7. Remark Specify the value calculated by [(the exceptional code of the specified interrupt request number - 0x80) / 0x10] for the interrupt request number eintno. User's Manual U13773EJ3V0UM 179 CHAPTER 11 SYSTEM CALLS Caution When the RX850 Pro is operated on the V850E core, even if the system call chg_icr is issued, the desired interrupt control register may not operate. In the RX850 Pro, the interrupt control register address is calculated from the interrupt source number. However, in the V850E core, the correct register address cannot be obtained since the alignment of the interrupt source numbers and interrupt control registers differs from other V850 Series products. Therefore, use of the system call chg_icr is restricted. For manipulating the interrupt control register via an application, directly manipulate the register without using this system call. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The chg_icr system call is not defined as CF. Invalid parameter specification * Invalid specification of interrupt request number (eintno < 0, maximum interrupt source number < eintno) * Invalid specification of interrupt request flag (eintno < 0, maximum interrupt source number < eintno) * (ICR_CLRMSK ICR_SETMSK) is specified as an interrupt request flag. 180 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Refer Interrupt Control Register Status (-68) ref_icr Task/non-task Overview Acquires the contents of the interrupt control register. C format ER ref_icr(UB *p_regptn, UINT eintno); Parameters I/O O I UB UINT Parameter *p_regptn; eintno; Description Address of area used to store contents of interrupt control register Interrupt request number Explanation This system call stores the contents of the interrupt control register specified by eintno in the area specified by p_regptn. The following figure shows the contents of the acquired interrupt control register: Bit 7 0 0 0 Bit 0 Interrupt mask flag 0: Enables interrupt processing 1: Disables interrupt processing (pending) Interrupt request flag 0: No interrupt request 1: Interrupt request exists Interrupt priority order 000: Specifies level 0 (highest) 001: Specifies level 1 010: Specifies level 2 011: Specifies level 3 100: Specifies level 4 101: Specifies level 5 110: Specifies level 6 111: Specifies level 7 (lowest) Remark Specify the value calculated by [(the exceptional code of the specified interrupt request number - 0x80) / 0x10] for the interrupt request number eintno. Caution When the RX850 Pro is operated on the V850E core, even if the system call ref_icr is issued, the desired interrupt control register may not operate. In the RX850 Pro, the interrupt control register address is calculated from the interrupt source number. However, in the V850E core, the correct register address cannot be obtained since the alignment of the interrupt source numbers and interrupt control registers differs from other V850 Series products. Therefore, use of the system call ref_icr is restricted. For manipulating the interrupt control register via an application, directly manipulate the register without using this system call. User's Manual U13773EJ3V0UM 181 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The ref_icr system call is not defined as CF. Invalid parameter specification * The address of the area used to store the contents of interrupt control register (p_regptn) is 0. * Invalid specification of interrupt request number (eintno < 0, maximum interrupt source number < eintno) 182 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS 11.8.5 Memory pool management system calls This section explains the group of system calls that allocate memory blocks (memory pool management system calls). Table 11-9 lists the memory pool management system calls. Table 11-9. Memory Pool Management System Calls System Call cre_mpl del_mpl get_blk pget_blk tget_blk rel_blk ref_mpl vget_pid Creates a memory pool. Deletes a memory pool. Acquires a memory block. Acquires a memory block (polling). Acquires a memory block (with timeout). Returns a memory block. Acquires memory pool information. Acquires memory pool ID number. Function User's Manual U13773EJ3V0UM 183 CHAPTER 11 SYSTEM CALLS Create Variable-size Memory Pool (-137) cre_mpl Task Overview Creates a memory pool. C format * When an ID number is specified #include ER * When an ID number is not specified #include ER Parameters I/O I I O ID T_CMPL ID Parameter mplid; *pk_cmpl; *p_mplid; Memory pool ID number Description Start address of packet containing memory pool creation information Address of area used to store ID number * Structure of memory pool creation information T_CMPL typedef struct VP ATR INT ID } T_CMPL; t_cmpl { exinf; mplatr; mplsz; keyid; /* /* /* /* Extended information Memory pool attribute Memory pool size Memory pool key ID number */ */ */ */ Explanation The RX850 Pro provides two types of interfaces for memory pool creation: one in which an ID number must be specified for memory pool creation, and one in which an ID number is not specified. * When an ID number is specified A memory pool having the ID number specified by mplid is created based on the information specified by pk_cmpl. * When an ID number is not specified A memory pool is created based on the information specified by pk_cmpl. An ID number is allocated by the RX850 Pro and the allocated ID number is stored in the area specified by p_mplid. 184 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Memory pool creation information is described in detail below. exinf ... Extended information exinf is an area used for storing user-specific information for the specified memory pool. The user can use this area as required. Information set in exinf can be dynamically acquired by issuing the ref_mpl system call from a processing program (task/non-task). mplatr ... Memory pool attribute Bit 0 .. Method of queuing to a wait queue TA_TPRI(0): TA_TFIFO(1): Bit 8 .. Priority order FIFO order Existence of key ID number specification TA_KEYID(1): Key ID number specified Bit 9 .. Memory area specification TA_UPOL0(0): Secures the memory pool area from user memory area 0. TA_UPOL1(1): Secures the memory pool area from user memory area 1. 15 mplatr 9 8 0 Method of queuing to a wait queue Existence of key ID number specification Memory area specification mplsz keyid ... ... Memory pool size (unit: byte) Memory pool key ID number Remark If the value of bit 8 is not 1 (TA_KEYID) in mplatr, the contents of keyid are meaningless. User's Manual U13773EJ3V0UM 185 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT *E_NOMEM E_RSATR E_PAR 0 -17 -10 -24 -33 Normal termination The cre_mpl system call is not defined as CF. A memory pool management block or memory pool area cannot be allocated. Invalid specification of attribute mplatr Invalid parameter specification * The start address of the packet storing memory pool creation information is invalid (p_mplid = 0). * Invalid size specification (mplsz 0) * Invalid key ID number specification (keyid = 0) (at TA_KEYID attribute specified) * The address of the area used to store the ID number is invalid (p_mplid = 0) (for creation without specifying the ID number) E_ID *E_OBJ E_OACV E_CTX -35 -63 -66 -69 Invalid ID number specification (maximum number of memory pools created < mplid) A memory pool having the specified ID number has already been created. An unauthorized ID number (mplid 0) was specified. The cre_mpl system call was issued from a non-task. 186 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Delete Variable-size Memory Pool (-138) del_mpl Task Overview Deletes a memory pool. C format #include ER Parameter I/O I ID Parameter mplid; Memory pool ID number Description Explanation This system call deletes the memory pool specified by mplid. The target memory pool is released from the control of the RX850 Pro. The task released from the wait state (memory block wait state) by this system call has E_DLT returned as the return value of the system call (get_blk or tget_blk) that instigated the transition to the wait state. If this system call is issued when the task acquires a memory block that the target memory pool manages, the memory block is also released from the control of the RX850 Pro. Accordingly, the contents of the acquired memory block are undefined. Return value *E_OK *E_NOSPT E_ID *E_NOEXS E_OACV E_CTX 0 -17 -35 -52 -66 -69 Normal termination The del_mpl system call is not defined as CF. Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. An unauthorized ID number (mplid 0) was specified. The del_mpl system call was issued from a non-task. User's Manual U13773EJ3V0UM 187 CHAPTER 11 SYSTEM CALLS Get Variable-size Memory Block (-141) get_blk Task Overview Acquires a memory block. C format #include ER Parameters I/O O I I VP ID INT Parameter *p_blk; mplid; blksz; Description Address of area used to store start address of memory block Memory pool ID number Memory block size (unit: bytes) Explanation This system call acquires a memory block of the size specified by blksz from the memory pool specified by mplid and stores its start address in the area specified by p_blk. If no memory block can be acquired from the target memory pool (when there is no free area of the requested size) upon the issuance of this system call, this system call places the task in the wait queue of the target memory pool before changing its state from the run state to the wait state (memory block wait state). The memory block wait state is released when a memory block that satisfies the requested size is released by the rel_blk system call or upon the issuance of the del_mpl or rel_wai system call, and the task returns to the ready state. Caution The RX850 Pro does not clear the memory upon acquiring a memory block. Accordingly, the contents of the acquired memory block are undefined. Remark When a task queues in the wait queue of the target memory pool, it is executed in the order (FIFO order or priority order) specified when that memory pool was generated (at configuration or when the cre_mpl system call was issued). 188 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The get_blk system call is not defined as CF. Invalid parameter specification * The address of the area used to store the start address of the memory block is invalid (p_blk = 0). * Invalid specification of memory block size (blksz 0) E_ID *E_NOEXS E_OACV E_CTX -35 -52 -66 -69 Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. An unauthorized ID number (mplid 0) was specified. Context error * The get_blk system call was issued from a non-task. * The get_blk system call was issued in the dispatch disabled state. *E_DLT *E_RLWAI -81 -86 The target memory pool was deleted by the del_mpl system call. The memory block wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 189 CHAPTER 11 SYSTEM CALLS Poll and Get Variable-size Memory Block (-104) pget_blk Task/non-task Overview Acquires a memory block (polling). C format #include ER Parameters I/O O I I VP ID INT Parameter *p_blk; mplid; blksz; Description Address of area used to store start address of memory block Memory pool ID number Memory block size (unit: bytes) Explanation This system call acquires a memory block of the size specified by blksz from the memory pool specified by mplid and stores its start address in the area specified by p_blk. When this system call is issued, if no memory block can be acquired from the target memory pool (when there is no free area of the requested size), this system call returns E_TMOUT as the return value. Caution The RX850 Pro does not clear the memory upon acquiring a memory block. Accordingly, the contents of the acquired memory block are undefined. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The pget_blk system call is not defined as CF. Invalid parameter specification * The address of the area used to store the start address of a memory block is invalid (p_blk = 0). * Invalid specification of memory block size (blksz 0) E_ID *E_NOEXS E_OACV *E_TMOUT -35 -52 -66 -85 Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. An unauthorized ID number (mplid 0) was specified. There is no free space in the target memory pool. 190 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Get Variable-size Memory Block with Timeout (-168) tget_blk Task Overview Acquires a memory block (with timeout). C format #include ER Parameters I/O O I I I VP ID INT TMO Parameter *p_blk; mplid; blksz; tmout; Description Address of area used to store start address of memory block Memory pool ID number Memory block size (unit: bytes) Wait time (unit: ms) TMO_POL(0): TMO_FEVR(-1): Value: Quick return Permanent wait Wait time Explanation This system call acquires a memory block of the size specified by blksz from the memory pool specified by mplid and stores its start address in the area specified by p_blk. If a memory block cannot be acquired from the target memory pool (when there is no free area of the requested size) when this system call is issued, this system call places the task in the wait queue of the target memory pool before changing it from the run state to the wait state (memory block wait state). The memory block wait state is released when the wait time specified by tmout elapses, when a memory block that satisfies the requested size is released by the rel_blk system call, or when the del_mpl or rel_wai system call is issued. The task then returns to the ready state. Caution The RX850 Pro does not clear the memory upon acquiring a memory block. Accordingly, the contents of the acquired memory block are undefined. Remark When a task queues in the wait queue of the target memory pool, it is executed in the order (FIFO order or priority order) specified when that memory pool was generated (at configuration or when the cre_mpl system call was issued). User's Manual U13773EJ3V0UM 191 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The tget_blk system call is not defined as CF. Invalid parameter specification * The address of the area used to store the start address of the memory block is invalid (p_blk = 0). * Invalid specification of memory block size (blksz 0) E_ID *E_NOEXS E_OACV E_CTX -35 -52 -66 -69 Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. An unauthorized ID number (mplid 0) was specified. Context error * The tget_blk system call was issued from a non-task. * The tget_blk system call was issued in the dispatch disabled state. *E_DLT *E_TMOUT *E_RLWAI -81 -85 -86 The target memory pool was deleted by the del_mpl system call. Timeout elapsed. The memory block wait state was forcibly released by the rel_wai system call. 192 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Release Variable-size Memory Block (-143) rel_blk Task/non-task Overview Returns a memory block. C format #include ER Parameters I/O I I ID VP Parameter mplid; blk; Memory pool ID number Start address of memory block Description Explanation This system call returns the memory block specified by blk to the memory pool specified by mplid. If the size of the returned memory block satisfies the size requested by the (first) task queuing in the target memory pool's wait queue when this system call is issued, the memory block is transferred to that task. The relevant task is consequently removed from the wait queue, and changes from the wait state (memory block wait state) to the ready state, or from the wait-suspend state to the suspend state. Cautions 1. The contents of a returned memory block are not cleared by the RX850 Pro. Thus, the contents of a memory block may be undefined when that memory block is returned. 2. The RX850 Pro includes two different specifications for the rel_blk system call. (1) When a memory block is returned by a rel_blk system call, if the first four bytes of the memory block are not filled with zeros, the return value E_OBJ is used for termination instead of returning the memory block. (2) When the rel_blk system call is issued, the memory block is returned even if the first four bytes of the memory block are not filled with zeros (return value = E_OK). The first specification applies when the memory block is used as a mailbox's message area, and this is the specification that has been used for the rel_blk system call as it has been implemented thus far in the RX850 Pro. User's Manual U13773EJ3V0UM 193 CHAPTER 11 SYSTEM CALLS When the memory block is used as a mailbox's message area, the first four bytes serve as the link area for the message's wait queue. In other words, if messages are queued in the mailbox, when the rel_blk system call is issued and the memory block must be returned, in which case it is the message area linked to the queue that is returned. To prevent this, the specification requires the first four bytes that comprise the link area to be filled with zeros, otherwise it will be recognized as the memory block used as the message area and the return value E_OBJ will be used for termination instead of returning the memory block. Under this specification, the first four bytes must be cleared to zero in order to use rel_blk to return the memory block. These specifications of rel_blk are stored in separate libraries so that one or the other rel_blk specification can be used. Link to the library of the rel_blk specification to be used. (1) Library containing rel_blk that requires zero-clearing of first four bytes of memory block librxp.a (2) Library containing rel_blk that does not require zero-clearing of first four bytes of memory block librxpm.a 3. Treat a memory pool that returns a memory block the same as a memory pool specified when issuing the get_blk, pget_blk, or tget_blk system call. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The rel_blk system call is not defined as CF. Invalid parameter specification * Invalid specification of the start address of a memory block (blk = 0) * The memory pool specified when acquired differs from that specified upon the issuance of the rel_blk system call. E_ID *E_NOEXS E_OBJ -35 -52 -63 Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. A value other than 0x0 is placed in the first four bytes of the memory block to be returned. * This return value is returned when returning the memory block used as a message area. E_OACV -66 An unauthorized ID number (mplid 0) was specified. 194 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Refer Variable-size Memory Pool Status (-140) ref_mpl Task/non-task Overview Acquires memory pool information. C format #include ER Parameters I/O O I T_RMPL ID Parameter *pk_rmpl; mplid; Description Start address of packet used to store memory pool information Memory pool ID number * Structure of memory pool information T_RMPL typedef struct VP BOOL_ID INT INT ID } T_RMPL; Explanation This system call stores the memory pool information (extended information, existence of waiting tasks, etc.) for the memory pool specified by mplid in the packet specified by pk_rmpl. Memory pool information is described in detail below. exinf wtsk ... ... Extended information Existence of waiting task FALSE(0): Value: frsz maxsz keyid ... ... ... No waiting task ID number of first task in the wait queue t_rmpl { exinf; wtsk; frsz; maxsz; keyid; /* /* /* /* /* Extended information Existence of waiting task Total size of free area Maximum memory block size that can be acquired Key ID number */ */ */ */ */ Total size of free area (unit: bytes) Maximum memory block size that can be acquired (unit: bytes) Key ID number FALSE(0): Value: No specification for key ID number at generation Key ID number User's Manual U13773EJ3V0UM 195 CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR E_ID *E_NOEXS E_OACV 0 -17 -33 -35 -52 -66 Normal termination The ref_mpl system call is not defined as CF. The start address of the packet used to store memory pool information is invalid (pk_rmpl = 0). Invalid ID number specification (maximum number of memory pools created < mplid) The target memory pool does not exist. An unauthorized ID number (mplid 0) was specified. 196 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Get Variable-size Memory Pool Identifier (-242) vget_pid Task/non-task Overview Acquires the memory pool ID number. C format #include ER Parameters I/O O I ID ID Parameter *p_mplid; keyid; Description Address of area used to store ID number Memory pool key ID number Explanation This system call stores the memory pool ID number specified by keyid in the area specified by p_mplid. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The vget_pid system call is not defined as CF. Invalid parameter specification * Invalid key ID number specification (keyid = 0) * The address of the area used to store the ID number is invalid (p_mplid = 0). *E_NOEXS -52 The target memory pool does not exist. User's Manual U13773EJ3V0UM 197 CHAPTER 11 SYSTEM CALLS 11.8.6 Time management system calls This section explains the group of system calls that perform processing dependent on time (time management system calls). Table 11-10 lists the time management system calls. Table 11-10. Time Management System Calls System Call set_tim get_tim dly_tsk def_cyc act_cyc ref_cyc Sets the system clock. Acquires the time from the system clock. Changes the task to the timeout wait state. Registers a cyclically activated handler or cancels its registration. Controls the activity state of a cyclically activated handler. Acquires cyclically activated handler information. Function 198 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Set Time (-83) set_tim Task/non-task Overview Sets the system clock. C format #include ER Parameter I/O I SYSTIME Parameter *pk_tim; Description Start address of packet storing time * Structure of system clock SYSTIME typedef struct UW H } SYSTIME; t_systime { ltime; utime; /* /* Time (lower 32 bits) Time (higher 16 bits) */ */ Explanation This system call sets the system clock to the time specified by pk_tim. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The set_tim system call is not defined as CF. The start address of the packet storing time is invalid (pk_tim = 0). User's Manual U13773EJ3V0UM 199 CHAPTER 11 SYSTEM CALLS Get Time (-84) get_tim Task/non-task Overview Acquires the time from the system clock. C format #include ER Parameter I/O O SYSTIME Parameter *pk_tim; Description Start address of packet storing time * Structure of system clock SYSTIME typedef struct UW H } SYSTIME; t_systime { ltime; utime; /* /* Time (lower 32 bits) Time (higher 16 bits) */ */ Explanation This system call sets the current system clock time in the packet specified by pk_tim. Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The get_tim system call is not defined as CF. The start address of the packet storing time is invalid (pk_tim = 0). 200 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Delay Task (-85) dly_tsk Task Overview Changes the task to the timeout wait state. C format #include ER Parameter I/O I DLYTIME Parameter dlytim; Delay time (unit: ms) Description Explanation This system call changes the state of the task from the run state to the wait state (timeout wait state) for the delay time specified by dlytim. The timeout wait state is released upon the elapse of the delay specified by dlytim or when the rel_wai system call is issued. The task then returns to the ready state. Caution The timeout wait state is released by neither the wup_tsk or ret_wup system call. Return value *E_OK *E_NOSPT E_PAR E_CTX 0 -17 -33 -69 Normal termination The dly_tsk system call is not defined as CF. Invalid specification of delay time (dlytim < 0) Context error * The dly_tsk system call was issued from a non-task. * The dly_tsk system call was issued in the dispatch disabled state. *E_RLWAI -86 The timeout wait state was forcibly released by the rel_wai system call. User's Manual U13773EJ3V0UM 201 CHAPTER 11 SYSTEM CALLS Define Cyclic Handler (-90) def_cyc Task/non-task Overview Registers a cyclically activated handler or cancels its registration. C format #include ER Parameters I/O I I HNO T_DCYC Parameter cycno; *pk_dcyc; Description Specification number of cyclically activated handler Start address of packet storing cyclically activated handler registration information * Structure of cyclically activated handler registration information T_DCYC typedef struct VP ATR FP UINT CYCTIME VP VP } T_DCYC; Explanation This system call uses the information specified by pk_dcyc to register the cyclically activated handler having the specification number specified by cycno. Cyclically activated handler registration information is described in detail below. exinf ... Extended information exinf is an area used for storing user-specific information on the specified cyclically activated handler. The user can use this area as required. Information set in exinf can be dynamically acquired by issuing the ref_cyc system call from a processing program (task/non-task). cycatr ... Attribute of cyclically activated handler Bit 0 .. Language in which the cyclically activated handler is coded TA_ASM(0): TA_HLNG(1): Assembly language C language t_dcyc { exinf; cycatr; cychdr; cycact; cyctim; gp; tp; /* /* /* /* /* /* /* Extended information Attribute of cyclically activated handler Activation address of cyclically activated handler Initial activity state of cyclically activated handler Activation time interval of cyclically activated handler Specific GP register value of cyclically activated handler Specific TP register value of cyclically activated handler */ */ */ */ */ */ */ 202 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Bit 10 Bit 11 .. .. Existence of specific GP register value specification TA_DPID(1): TA_DPIC(1): Specific GP register value specified. Specific TP register value specified. Existence of specific TP register value specification 15 cycatr 10 0 Language in which a cyclically activated handler is coded Existence of specific GP register value specification Existence of specific TP register value specification cychdr cycact ... ... Activation address of cyclically activated handler Initial activity state of cyclically activated handler TCY_OFF(0): TCY_ON(1): The initial activity state is OFF The initial activity state is ON cyctim gp tp ... ... ... Activation time interval of cyclically activated handler (unit: basic clock cycles) Specific GP register value for cyclically activated handler Specific TP register value for cyclically activated handler When this system call is issued, if a cyclically activated handler corresponding to the target specification number is already registered, this system call does not handle this as an error and newly registers the specified cyclically activated handler. If this system call is issued with NADR(-1) set in the area specified by pk_dcyc, the registration of the cyclically activated handler specified by cycno is canceled. Remarks 1. If the value of bit 10 is not 1 (TA_DPID) in cycatr, the contents of gp are meaningless. 2. If the value of bit 11 is not 1 (TA_DPIC) in cycatr, the contents of tp are meaningless. Return value *E_OK *E_NOSPT E_RSATR E_PAR 0 -17 -24 -33 Normal termination The def_cyc system call is not defined as CF. Invalid specification of attribute cycatr Invalid parameter specification * Invalid specification of specification number (cycno 0, maximum number of cyclically activated handlers that can be registered < cycno) * The start address of the packet storing cyclically activated handler registration information is invalid (pk_dcyc = 0). * Invalid specification of activation address (cychdr = 0) * Invalid specification of initial activity state cycact * Invalid specification of activation time interval (cyctim 0) User's Manual U13773EJ3V0UM 203 CHAPTER 11 SYSTEM CALLS Activate Cyclic Handler (-94) act_cyc Task/non-task Overview Controls the activity state of a cyclically activated handler. C format #include ER Parameters I/O I I HNO UINT Parameter cycno; cycact; Description Specification number of cyclically activated handler Specification of activity state and cycle counter TCY_OFF(0): TCY_ON(1): TCY_INI(2): Changes the activity state to the OFF state. Changes the activity state to the ON state. Initializes the cycle counter. Explanation This system call changes the activity state of the cyclically activated handler specified by cycno to the state specified by cycact. The specification format of cycact is described below. * cycact = TCY_OFF Changes the activity state of the target cyclically activated handler to the OFF state. Even when the activation time is reached, the target cyclically activated handler is not activated. Caution Even when the activity state of the cyclically activated handler is OFF, the RX850 Pro increments the cycle counter. * cycact = TCY_ON Changes the activity state of the target cyclically activated handler to the ON state. When the activation time is reached, the target cyclically activated handler is activated. * cycact = TCY_INI Initializes the cycle counter of the target cyclically activated handler. * cycact = (TCY_ON|TCY_INI) Changes the activity state of the target cyclically activated handler to the ON state before initializing the cycle counter. When the activation time is reached, the target cyclically activated handler is activated. 204 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The act_cyc system call is not defined as CF. Invalid parameter specification * The specification number of the cyclically activated handler is invalid (cycno 0, maximum number of cyclically activated handlers that can be registered < cycno). * Invalid specification of activity state or cycle counter cycact *E_NOEXS -52 The target cyclically activated handler is not registered. User's Manual U13773EJ3V0UM 205 CHAPTER 11 SYSTEM CALLS Refer Cyclic Handler Status (-92) ref_cyc Task/non-task Overview Acquires cyclically activated handler information. C format #include ER Parameters I/O O I T_RCYC HNO Parameter *pk_rcyc; cycno; Description Start address of packet used to store cyclically activated handler information Specification number of cyclically activated handler * Structure of cyclically activated handler information T_RCYC typedef struct VP CYCTIME UINT } T_RCYC; Explanation This system call stores the cyclically activated handler information (extended information, remaining time, etc.) of the cyclically activated handler specified by cycno in the packet specified by pk_rcyc. Cyclically activated handler information is described in detail below. exinf lfttim cycact ... ... ... Extended information Time remaining until the cyclically activated handler is next activated (unit: basic clock cycles) Current activity state TCY_OFF(0): TCY_ON(1): Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The ref_cyc system call is not defined as CF. Invalid parameter specification * The start address of the packet used to store cyclically activated handler information is invalid (pk_rcyc = 0). * The specification number of the cyclically activated handler is invalid (cycno 0, maximum number of cyclically activated handlers that can be registered < cycno). The target cyclically activated handler is not registered. User's Manual U13773EJ3V0UM t_rcyc { exinf; lfttim; cycact; /* /* /* Extended information Remaining time Current activity state */ */ */ Activity state is OFF. Activity state is ON. *E_NOEXS -52 206 CHAPTER 11 SYSTEM CALLS 11.8.7 System management system calls This section explains the group of system calls that perform processing dependent on the system (system management system calls). Table 11-11 lists the system management system calls. Table 11-11. System Management System Calls System Call get_ver ref_sys def_svc viss_svc Function Acquires RX850 Pro version information. Acquires system information. Registers an extended SVC handler or cancels its registration. Calls an extended SVC handler. User's Manual U13773EJ3V0UM 207 CHAPTER 11 SYSTEM CALLS Get Version Information (-16) get_ver Task/non-task Overview Acquires RX850 Pro version information. C format #include ER Parameter I/O O T_VER Parameter *pk_ver; Description Start address of packet used to store version information * Structure of version information T_VER typedef struct UH UH UH UH UH UH UH } T_VER; t_ver { maker; id; spver; prver; prno[4]; cpu; var; /* /* /* /* /* /* /* OS maker OS format Specification version OS version Product number, production management information CPU information Variation descriptor */ */ */ */ */ */ */ Explanation This system call stores the RX850 Pro version information (OS maker, OS format, etc.) in the packet specified by pk_ver. Version information is described in detail below. maker ... OS maker H'000d: id ... OS format H'0000: spver ... Not used NEC Specification version H'5302: ITRON3.0 Ver. 3.02 prver ... OS product version H'0300: RX850 Pro Ver. 3.00 208 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS prno[4] ... Product number/product management information Undefined: Serial number of delivery product (each unit has a unique number) cpu ... CPU information H'0d37: PD703100 var ... Variation descriptor H'c000: ITRON level E, file not supported Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The get_ver system call is not defined as CF. The start address of the packet used to store version information is invalid (pk_ver = 0). User's Manual U13773EJ3V0UM 209 CHAPTER 11 SYSTEM CALLS Refer System Status (-12) ref_sys Task/non-task Overview Acquires system information. C format #include ER Parameter I/O O T_RSYS Parameter *pk_rsys; Description Start address of packet used to store system information * Structure of system information T_RSYS typedef struct INT } T_RSYS; t_rsys { sysstat; /* System state */ Explanation This system call stores the current value of dynamically-changing system information (system state) in the packet specified by pk_rsys. System information is described in detail below. sysstat ... System state TSS_TSK(0): Task processing is being performed. enabled. TSS_DDSP(1): Task processing is being performed. disabled. TSS_LOC(3): Task processing is being performed. The acknowledgement of maskable interrupts and dispatch processing is disabled. TSS_INDP(4): Processing of a non-task (interrupt handler, cyclically activated handler, etc.) is being performed. Dispatch processing is Dispatch processing is Return value *E_OK *E_NOSPT E_PAR 0 -17 -33 Normal termination The ref_sys system call is not defined as CF. The start address of the packet used to store system information is invalid (pk_rsys = 0). 210 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Define Supervisor Call Handler (-9) def_svc Task/non-task Overview Registers an extended SVC handler or cancels its registration. C format #include ER Parameters I/O I I FN T_DSVC Parameter s_fncd; *pk_dsvc; Description Extended function code of extended SVC handler Start address of packet storing the extended SVC handler registration information * Structure of extended SVC handler registration information T_DSVC typedef struct ATR FP VP VP } T_DSVC; Explanation This system call uses information specified by pk_dsvc to register the extended SVC handler having the extended function code specified by s_fncd. Extended SVC handler registration information is described in detail below. svcatr ... Attribute of extended SVC handler Language in which the extended SVC handler is coded Bit 0 ... TA_ASM(0): TA_HLNG(1): Bit 10 ... Bit 11 ... Assembly language C language t_dsvc { svcatr; svchdr; gp; tp; /* /* /* /* Attribute of extended SVC handler Activation address of extended SVC handler Specific GP register value for extended SVC handler Specific TP register value for extended SVC handler */ */ */ */ Existence of specific GP register value specification TA_DPID(1): Specific GP register value specified. Existence of specific TP register value specification TA_DPIC(1): Specific TP register value specified. User's Manual U13773EJ3V0UM 211 CHAPTER 11 SYSTEM CALLS 15 svcatr 10 0 Language in which an extended SVC handler is coded Existence of specific GP register value specification Existence of specific TP register value specification svchdr gp tp ... ... ... Activation address of extended SVC handler Specific GP register value of extended SVC handler Specific TP register value of extended SVC handler When this system call is issued, if an extended SVC handler corresponding to the target extended function code has already been registered, this system call does not handle this as an error and newly registers the specified extended SVC handler. When this system call is issued, if NADR(-1) is set in the area specified by pk_dsvc, the registration of the extended SVC handler specified by s_fncd is canceled. Remarks 1. If the value of bit 10 is not 1 (TA_DPID) in svcatr, the contents of gp are meaningless. 2. If the value of bit 11 is not 1 (TA_DPIC) in svcatr, the contents of tp are meaningless. Return value *E_OK *E_NOSPT E_RSATR E_PAR 0 -17 -24 -33 Normal termination The def_svc system call is not defined as CF. Invalid specification of attribute svcatr Invalid parameter specification * Invalid specification of extended function code (s_fncd 0, maximum number of extended SVC handlers that can be registered < s_fncd) * The start address of the packet storing the extended SVC handler registration information is invalid (pk_dsvc = 0). * Invalid specification of activation address (svchdr = 0) 212 User's Manual U13773EJ3V0UM CHAPTER 11 SYSTEM CALLS Issued Supervisor Call Handler (-250) viss_svc Task/non-task Overview Calls an extended SVC handler. C format #include ER Parameters I/O I I I I FN VW VW VW Parameter s_fncd; prm1; prm2; prm3; Description Extended function code of extended SVC handler Parameter 1 passed to extended SVC handler Parameter 2 passed to extended SVC handler Parameter 3 passed to extended SVC handler Explanation This system call calls the extended SVC handler having the extended function code specified by s_fncd. Remark When this system call is used to call an extended SVC handler, the interface library for the extended SVC handlers need not be coded. Return value *E_OK *E_NOSPT 0 -17 Normal termination The viss_svc system call is not defined as CF, or this system call calls an extended SVC handler that is not registered. Invalid specification of extended function code (s_fncd 0, maximum number of extended SVC handlers that can be registered < s_fncd) Return value from extended SVC handler E_PAR -33 Others User's Manual U13773EJ3V0UM 213 APPENDIX A PROGRAMMING METHODS This appendix explains how to describe processing programs when using the NEC Electronics V850 Series C compiler CA850 or the C cross V800 compiler CCV850 manufactured by Green Hills Software, Inc. A.1 Overview In the RX850 Pro, processing programs are classified according to purpose, as shown below. * Task The minimum unit of a processing program that can be executed by the RX850 Pro. * Directly activated interrupt handler A routine dedicated to interrupt processing. When an interrupt occurs, this handler is activated without using the RX850 Pro. Accordingly, a high-speed response as close as the hardware limitation can be expected. * Indirectly activated interrupt handler A routine dedicated to interrupt processing. When an interrupt occurs, this handler is activated upon the completion of the interrupt preprocessing by the RX850 Pro (such as saving the contents of the registers or switching the stack). Compared with the directly activated interrupt handler, the response is slower. However, this handler has the advantage of simplicity for the processing within the handler because of the interrupt preprocessing by the RX850 Pro. * Cyclically activated handler A routine dedicated to cyclic processing. Every time the specified time elapses, this handler is activated immediately. This routine is handled independently of tasks. When the activation time has been reached, therefore, the processing of the task currently being executed is canceled even if that task has the highest priority relative to all other tasks in the system, and control is passed to the cyclically activated handler. A cyclically activated handler incurs a smaller overhead before the start of execution, relative to any other cyclic processing programs written by the user. * Extended SVC handler A function registered by the user as an extended system call. These processing programs have their own basic formats according to the general conventions or conventions to be applied when the RX850 Pro is used. 214 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS A.2 Keywords The character strings listed below are reserved as keywords for the configurator. These strings cannot, therefore, be used for other purposes. clkhdr flgsvc maxcyc maxmpl mbx no_use prttsk semsvc svc TA_HLNG TA_WMUL tsk UPOL1 clktim ini maxflg maxpri mbxsvc prtflg RX850PRO ser_def syssvc TA_MFIFO TA_WSGL tsksvc cyc inthdr maxint maxsem mem prtmbx rxsers sit_def TA_ASM TA_MPRI TCY_OFF TTS_DMT defstk intstk maxintfactor maxsvc mpl prtmpl sct_def SPOL0 TA_DISINT TA_TFIFO TCY_ON TTS_RDY flg intsvc maxmbx maxtsk mplsvc prtsem sem SPOL1 TA_ENAINT TA_TPRI timsvc UPOL0 A.3 Reserved Words The character strings listed below are reserved as external symbols for the RX850 Pro. These strings cannot, therefore, be used for other purposes. _x_ _f_ _e_ _rx_ Remark The use of these character strings is prohibited when a single load module is created. There is no problem if a symbol starting with any of these character strings is used when a load module that separates the RX850 Pro and application is created. User's Manual U13773EJ3V0UM 215 APPENDIX A PROGRAMMING METHODS A.4 Tasks A.4.1 CA850-supported version When describing a task in C language, describe it as a void-type function having one INT-type argument after function declaration by pragma directive. The activation code specified when the sta_tsk system call is issued is set for the argument (stacd). Figure A-1 shows the task description format (in C language) when the CA850 is used. Figure A-1. Task Description Format When Using CA850 (C Language) #include #pragma void func_task(INT stacd) { /* Processing of task func_task */ ............................. ............................. ............................. /* Termination of task func_task */ ext_tsk(); } Remark For details about the function declaration by pragma directive, refer to the CA850 C Language User's Manual (U16054E). 216 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing a task in assembly language, describe it as a function conforming to the function call conventions of the CA850. The activation code specified when the sta_tsk system call is issued is set for the argument (r6 register). Figure A-2 shows the task description format (in assembly language) when the CA850 is used. Figure A-2. Task Description Format When Using CA850 (Assembly Language) .include "stdrx85p.inc" .text .align .globl 4 _func_task _func_task: # Processing of task func_task ........................... ........................... ........................... # Termination of task func_task jr _ext_tsk User's Manual U13773EJ3V0UM 217 APPENDIX A PROGRAMMING METHODS A.4.2 CCV850-supported version When describing a task in C language, describe it as a void-type function having one INT-type argument. The activation code specified upon the issuance of the sta_tsk system call is set for the argument (stacd). Figure A-3 shows the task description format (in C language) when the CCV850 is used. Figure A-3. Task Description Format When Using CCV850 (C Language) #include void func_task(INT stacd) { /* Processing of task func_task */ ............................. ............................. ............................. /* Termination of task func_task */ ext_tsk(); } 218 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing a task in assembly language, describe it as a function conforming to the function call conventions of the CCV850. The activation code specified upon the issuance of the sta_tsk system call is set for the argument (r6 register). Figure A-4 shows the task description format (in assembly language) when CCV850 is used. Figure A-4. Task Description Format When Using CCV850 (Assembly Language) #include .text .align .globl _func_task: 4 _func_task # Processing of task func_task ......................... ......................... ......................... # Termination of task func_task jr _ext_tsk Cautions 1. When describing a task in assembly language, specify ".850" as the file extension. 2. When compiling a task in assembly language, specify "-D_ _asm_ _" as the option at compilation. User's Manual U13773EJ3V0UM 219 APPENDIX A PROGRAMMING METHODS A.5 Directly Activated Interrupt Handler A.5.1 CA850-supported version Use assembly language when describing a directly activated interrupt handler. However, the processing itself can be described in C language and called by the jarl instruction. The registers must be saved before the directly activated interrupt handler processing and restored after the processing. However, the RX850 Pro provides a macro that describes the process of saving and restoring the registers in order to mitigate the workload of the user in describing a handler in assembly language. The following figure shows the description format (in assembly language) of the directly activated interrupt handler when the CA850 is used. Figure A-5. Description Format of Directly Activated Interrupt Handler When Using CA850 (Assembly Language) .include "stdrx85p.inc" .section "int_name", text jr _func_inthdr .text .align .globl _func_inthdr: 4 _func_inthdr /* Saving registers, switching stack */ RTOS_IntEntry /* Main processing of directly activated interrupt handler */ .extern _inthdr_body jarl _inthdr_body, lp /* r10: ID of task to be woken up after returning from handler */ /* Switching stack, restoring registers */ /* Return from directly activated interrupt handler and waking up task */ RTOS_IntReturnWakeup r10 220 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS #include ID inthdr_body( ) { __asm ("mov #__tp_TEXT, tp"); __asm ("mov #__gp_DATA, gp"); /* Processing of directly activated interrupt handler func_inthdr */ .......................................................... .......................................................... .......................................................... /* Return from directly activated interrupt handler func_inthdr */ return tskid } First, describe the interrupt handler entry processing (jr instruction) at the handler address. Refer to the second and third rows in this example. Next, describe the interrupt handler main unit processing. The macro RTOS_IntEntry notifies the RX850 Pro of the activation of the handler, the saving of the temporary register and lp, and the switching of the task. The other registers (r20 to r30) are then saved, and control is transferred to the handler. In the above example, the C function, inthdr_body, of the handler is called. Before the execution of the handler main unit processing, set the TP (text pointer) and GP (global pointer) used by the handler. As described in 5.3 Directly Activated Interrupt Handler, the values of the GP and TP become undefined. Since this setting must be described in assembly language, use the _ _asm instruction as in the above example or the #pragma asm to pragma endasm directives to describe the handler in C language. In the handler, "the system calls that can be issued from the handler" explained in the user's manual can be issued. When the issuance processing of the handler is completed, the registers saved by the user must be restored and execution must return from the interrupt handler. To wake up a task specified after execution has returned from an interrupt, the ID of the task to be woken up must be set to register r10. In the above example, a task ID is returned as a return value when execution returns from inthdr_body, and its value is copied to r10. This operation is performed with the code output from the CA850. To return from the handler and wake up other tasks, describe the macro RTOS_IntReturnWakeup r10. To simply return from the handler, use the macro RTOS_IntReturn. The handler can be also terminated by using the system call jr_ret_int or jr_ret_wup. In this case, however, the registers that have been saved must be restored before the system call is issued. It is easier to return from the handler by using the macros. To return from the handler after performing only simple processing, the reti instruction can be also used. In this case, the registers must be restored before issuing the instruction. The macro RTOS_IntExit is equivalent to this instruction and restores the registers. When returning from the handler by using these instructions and macros, however, do not issue a system call in the handler. If a system call is used in the handler, use RTOS_IntReturnWakeup or RTOS_IntReturn to return from the handler. To enable multiple interrupts during the servicing of the handler, execute ei/di after the processing of RTOS_IntEntry and before RTOS_IntReturnWakeup, RTOS_IntReturn, or RTOS_IntExit. User's Manual U13773EJ3V0UM 221 APPENDIX A PROGRAMMING METHODS Remark Set a branch instruction that branches to the directly activated interrupt handler at the handler address to which the processor transfers control if an interrupt occurs. This is done by the .section quasi directive in Figure A-5. For details of the .section quasi directive, refer to the CA850 C Compiler Package Assembly Language User's Manual (U16042E). Specify an interrupt request name defined in the device file as "int_name". 222 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS A.5.2 CCV850-supported version Use assembly language when describing a directly activated interrupt handler. However, the processing itself can be described in C language and called by the jarl instruction. The registers must be saved before the directly activated interrupt handler processing and restored after the processing. However, the RX850 Pro provides a macro that describes the process of saving and restoring the registers in order to mitigate the workload of the user in describing a handler in an assembly language. The following figure shows the description format (in assembly language) of the directly activated interrupt handler when the CCV850 is used. Figure A-6. Description Format of Directly Activated Interrupt Handler When Using CCV850 (Assembly Language) .include "stdrx85p.inc" .org handler_address_number jr _func_inthdr .text .align .globl _func_inthdr: 4 _func_inthdr /* Saving registers, switching stack */ RTOS_IntEntry /* Main processing of directly activated interrupt handler */ .extern _inthdr_body jarl _inthdr_body, lp /* r10: ID of task to be woken up after returning from handler */ /* Switching stack, restoring registers */ /* Return from directly activated interrupt handler and waking up task */ jr RTOS_IntReturnWakeup r10 User's Manual U13773EJ3V0UM 223 APPENDIX A PROGRAMMING METHODS #include ID inthdr_body( ) { __asm ("__tp, tp"); __asm ("__gp, gp"); /* Processing of directly activated interrupt handler func_inthdr */ .......................................................... .......................................................... .......................................................... /* Return from directly activated interrupt handler func_inthdr */ return tskid } First, describe the interrupt handler entry processing (jr instruction) at the handler address. Refer to the second and third rows in this example. Next, describe the interrupt handler main unit processing. The macro RTOS_IntEntry notifies the RX850 Pro of the activation of the handler, the saving of the temporary register and lp, and the switching of the task. The other registers (r20 to r30) are then saved, and control is transferred to the handler. In the above example, the C function, inthdr_body, of the handler is called. Before the execution of the handler main unit processing, set the TP (text pointer) and GP (global pointer) used by the handler. As described in 5.3 Directly Activated Interrupt Handler, the values of the GP and TP become undefined. Since this setting must be described in assembly language, use the _ _asm instruction as in the above example or the #pragma asm to pragma endasm directives to describe the handler in C language. In the handler, "the system calls that can be issued from the handler" explained in the user's manual can be issued. When the issuance processing of the handler is completed, the registers saved by the user must be restored and execution must return from the interrupt handler. To wake up a task specified after execution has returned from an interrupt, the ID of the task to be woken up must be set to register r10. In the above example, a task ID is returned as a return value when execution returns from inthdr_body, and its value is copied to r10. This operation is performed with the code output from the CCV850. To return from the handler and wake up other tasks, describe the macro RTOS_IntReturnWakeup r10. To simply return from the handler, use the macro RTOS_IntReturn. The handler can be also terminated by using the system call jr_ret_int or jr_ret_wup. In this case, however, the registers that have been saved must be restored before the system call is issued. It is easier to return from the handler by using the macros. To return from the handler after performing only simple processing, the reti instruction can be also used. In this case, the registers must be restored before issuing the instruction. The macro RTOS_IntExit is equivalent to this instruction and restores the registers. When returning from the handler by using these instructions and macros, however, do not issue a system call in the handler. If a system call is used in the handler, use RTOS_IntReturnWakeup or RTOS_IntReturn to return from the handler. To enable multiple interrupts during the servicing of the handler, execute ei/di after the processing of RTOS_IntEntry and before RTOS_IntReturnWakeup, RTOS_IntReturn, or RTOS_IntExit. 224 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS Remark Set a branch instruction that branches to the directly activated interrupt handler at the handler address to which the processor transfers control if an interrupt occurs. This is done by the .org instruction in Figure A-6. Specify the handler address of an interrupt as handler_address_number. Caution When describing a directly activated interrupt handler in assembly language, specify ".850" as the file extension. User's Manual U13773EJ3V0UM 225 APPENDIX A PROGRAMMING METHODS A.6 Indirectly Activated Interrupt Handler A.6.1 CA850-supported version When describing an indirectly activated interrupt handler in C language, describe it as an ID-type function having no argument. Figure A-7 shows the description format of an indirectly activated interrupt handler (in C language) when the CA850 is used. Figure A-7. Description Format of Indirectly Activated Interrupt Handler When Using CA850 (C Language) #include ID func_inthdr() { /* Processing of indirectly activated interrupt handler func_inthdr */ ............................................................ ............................................................ ............................................................ /* Return processing from indirectly activated interrupt handler func_inthdr */ return(TSK_NULL); } Remark An indirectly activated interrupt handler is a subroutine called by interrupt processing in the nucleus. Therefore, when an indirectly activated interrupt handler is described, an instruction for branching to the indirectly activated interrupt handler needs to be set for the handler address to which the processor passes control upon the occurrence of an interrupt. language. However, because the RX850 Pro provides the processing that should be described as the branch instruction in the form of a macro, this macro should be used. For example, to use the INTP100 (address: 0x100) maskable interrupt as an indirectly activated interrupt handler, describe as follows. .section "INTP100" RTOS_IntEntry_Indirect The same description is required for timer interrupts since they are handled as indirectly activated interrupt handlers. This setting must be described in assembly 226 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing an indirectly activated interrupt handler in assembly language, describe it as a function conforming to the function call conventions of the CA850. Figure A-8 shows the description format of an indirectly activated interrupt handler (in assembly language) when the CA850 is used. Figure A-8. Description Format of Indirectly Activated Interrupt Handler When Using CA850 (Assembly Language) .include "stdrx85p.inc" .text .align .globl 4 _func_inthdr _func_inthdr: # Processing of indirectly activated interrupt handler func_inthdr .......................................................... .......................................................... .......................................................... # Return processing from indirectly activated interrupt handler func_inthdr mov jmp TSK_NULL, r10 [lp] Remark An indirectly activated interrupt handler is a subroutine called by interrupt processing in the nucleus. Therefore, when an indirectly activated interrupt handler is described, an instruction for branching to the indirectly activated interrupt handler needs to be set for the handler address to which the processor passes control upon the occurrence of an interrupt. language. However, because the RX850 Pro provides the processing that should be described as the branch instruction in the form of a macro, this macro should be used. For example, to use the INTP100 (address: 0x100) maskable interrupt as an indirectly activated interrupt handler, describe as follows. .section "INTP100" RTOS_IntEntry_Indirect The same description is required for timer interrupts since they are handled as indirectly activated interrupt handlers. This setting must be described in assembly User's Manual U13773EJ3V0UM 227 APPENDIX A PROGRAMMING METHODS A.6.2 CCV850-supported version When describing an indirectly activated interrupt handler in C language, describe it as an ID-type function having no argument. Figure A-9 shows the description format of an indirectly activated interrupt handler (in C language) when the CCV850 is used. Figure A-9. Description Format of Indirectly Activated Interrupt Handler When Using CCV850 (C Language) #include Remark An indirectly activated interrupt handler is a subroutine called by interrupt processing in the nucleus. Therefore, when an indirectly activated interrupt handler is described, an instruction for branching to the indirectly activated interrupt handler needs to be set for the handler address to which the processor passes control upon the occurrence of an interrupt. language. However, because the RX850 Pro provides the processing that should be described as the branch instruction in the form of a macro, this macro should be used. For example, to use the INTP100 (address: 0x100) maskable interrupt as an indirectly activated interrupt handler, describe as follows. .org 00000100 RTOS_IntEntry_Indirect The same description is required for timer interrupts since they are handled as indirectly activated interrupt handlers. This setting must be described in assembly 228 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing an indirectly activated interrupt handler in assembly language, describe it as a function conforming to the function call conventions of the CCV850. Figure A-10 shows the description format of an indirectly activated interrupt handler (in assembly language) when the CCV850 is used. Figure A-10. Description Format of Indirectly Activated Interrupt Handler When Using CCV850 (Assembly Language) #include _func_inthdr: # Processing of indirectly activated interrupt handler func_inthdr ........................................................... ........................................................... ........................................................... # Return processing from indirectly activated interrupt handler func_inthdr mov jmp TSK_NULL, r10 [lp] Remark An indirectly activated interrupt handler is a subroutine called by interrupt processing in the nucleus. Therefore, when an indirectly activated interrupt handler is described, an instruction for branching to the indirectly activated interrupt handler needs to be set for the handler address to which the processor passes control upon the occurrence of an interrupt. language. However, because the RX850 Pro provides the processing that should be described as the branch instruction in the form of a macro, this macro should be used. For example, to use the INTP100 (address: 0x100) maskable interrupt as an indirectly activated interrupt handler, describe as follows. .org 00000100 RTOS_IntEntry_Indirect The same description is required for timer interrupts since they are handled as indirectly activated interrupt handlers. This setting must be described in assembly Caution When describing an indirectly activated interrupt handler in assembly language, specify ".850" as the file extension. User's Manual U13773EJ3V0UM 229 APPENDIX A PROGRAMMING METHODS A.7 Cyclically Activated Handler A.7.1 CA850-supported version When describing a cyclically activated handler in C language, describe it as a void-type function having no argument. Figure A-11 shows the description format of a cyclically activated handler (in C language) when the CA850 is used. Figure A-11. Description Format of Cyclically Activated Handler When Using CA850 (C Language) #include void func_cychdr() { /* Processing of cyclically activated handler func_cychdr */ ...................................................... ...................................................... ...................................................... /* Return processing from cyclically activated handler func_cychdr */ return; } Remark A cyclically activated handler is a subroutine called by system clock processing in the nucleus. 230 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing a cyclically activated handler in assembly language, describe it as a function conforming to the function call conventions of the CA850. Figure A-12 shows the description format of a cyclically activated handler (in assembly language) when the CA850 is used. Figure A-12. Description Format of Cyclically Activated Handler When Using CA850 (Assembly Language) .include "stdrx85p.inc" .text .align .globl 4 _func_cychdr _func_cychdr: # Processing of cyclically activated handler func_cychdr .................................................... .................................................... .................................................... # Return processing from cyclically activated handler func_cychdr jmp [lp] Remark A cyclically activated handler is a subroutine called by system clock processing in the nucleus. User's Manual U13773EJ3V0UM 231 APPENDIX A PROGRAMMING METHODS A.7.2 CCV850-supported version When describing a cyclically activated handler in C language, describe it as a void-type function having no argument. Figure A-13 shows the description format of a cyclically activated handler (in C language) when the CCV850 is used. Figure A-13. Description Format of Cyclically Activated Handler When Using CCV850 (C Language) #include void func_cychdr() { /* Processing of cyclically activated handler func_cychdr */ ...................................................... ...................................................... ...................................................... /* Return processing from cyclically activated handler func_cychdr */ return; } Remark A cyclically activated handler is a subroutine called by system clock processing in the nucleus. 232 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing a cyclically activated handler in assembly language, describe it as a function conforming to the function call conventions of the CCV850. Figure A-14 shows the description format of a cyclically activated handler (in assembly language) when the CCV850 is used. Figure A-14. Description Format of Cyclically Activated Handler When Using CCV850 (Assembly Language) #include _func_cychdr: # Processing of cyclically activated handler func_cychdr ................................................... ................................................... ................................................... # Return processing from cyclically activated handler func_cychdr jmp [lp] Remark A cyclically activated handler is a subroutine called by system clock processing in the nucleus. Caution When describing a cyclically activated handler in assembly language, specify ".850" as the file extension. User's Manual U13773EJ3V0UM 233 APPENDIX A PROGRAMMING METHODS A.8 Extended SVC Handler A.8.1 CA850-supported version When describing an extended SVC handler in C language, describe it as an INT-type function. Figure A-15 shows the description format of an extended SVC handler (in C language) when the CA850 is used. Figure A-15. Description Format of Extended SVC Handler When Using CA850 (C Language) #include INT func_svchdr(VW prml, VW prm2, VW prm3) { int ret; /* Processing of extended SVC handler func_svchdr */ ................................................... ................................................... ................................................... /* Return processing from extended SVC handler func_svchdr */ return(INT ret); } 234 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing an extended SVC handler in assembly language, describe it as a function conforming to the function call conventions of the CA850. Figure A-16 shows the description format of an extended SVC handler (in assembly language) when the CA850 is used. Figure A-16. Description Format of Extended SVC Handler When Using CA850 (Assembly Language) .include "stdrx85p.inc" .text .align .globl 4 _func_svchdr _func_svchdr: # Processing of extended SVC handler func_svchdr ................................................. ................................................. ................................................. # Return processing from extended SVC handler func_svchdr mov jmp ret, r10 [lp] User's Manual U13773EJ3V0UM 235 APPENDIX A PROGRAMMING METHODS A.8.2 CCV850-supported version When describing an extended SVC handler in C language, describe it as an INT-type function. Figure A-17 shows the description format of an extended SVC handler (in C language) when the CCV850 is used. Figure A-17. Description Format of Extended SVC Handler When Using CCV850 (C Language) #include INT func_svchdr(VW prml, VW prm2, VW prm3) { int ret; /* Processing of extended SVC handler func_svchdr */ ................................................... ................................................... ................................................... /* Return processing from extended SVC handler func_svchdr */ return(INT ret); } 236 User's Manual U13773EJ3V0UM APPENDIX A PROGRAMMING METHODS When describing an extended SVC handler in assembly language, describe it as a function conforming to the function call conventions of the CCV850. Figure A-18 shows the description format of an extended SVC handler (in assembly language) when CCV850 is used. Figure A-18. Description Format of Extended SVC Handler When Using CCV850 (Assembly Language) #include _func_svchdr: # Processing of extended SVC handler func_svchdr ................................................. ................................................. ................................................. # Return processing from extended SVC handler func_svchdr mov jmp ret, r10 [lp] Caution When describing an extended SVC handler in assembly language, specify ".850" as the file extension. User's Manual U13773EJ3V0UM 237 APPENDIX B Q & A [Product Overview] Q.1 Is the RX850 Pro compliant with either the ITRON or ITRON specification? A.1 The RX850 Pro is compliant with the ITRON3 specification. Q.2 What is the RX850 Pro's nucleus (kernel) size? A.2 It ranges from 5 to 13 KB. This varies depending on the number of system calls used. When all system calls are used, the size is 13 KB. For details of size estimates, see the "RX850 Pro Memory Size Estimate page" in the FAQ at the NEC Electronics Microcomputer website. Q.3 What RAM size does the RX850 Pro use? A.3 The RAM size used by the RX850 Pro depends on the application. The number of control blocks varies depending on the number of created tasks and other resources. For description of control block sizes and methods for estimating the RAM sizes used, see CHAPTER 5 MEMORY AND MEMORY CAPACITY ESTIMATION in the RX850 Pro Installation User's Manual (U13774E) or use the "RX850 Pro Memory Size Estimate page" in the FAQ at the NEC Electronics Microcomputer website. Q.4 What language is the RX850 Pro coded in? A.4 It is coded entirely in assembly language. 238 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.5 How is the RX850 Pro incorporated into applications? A.5 The RX850 Pro provides libraries (librxp.a/librxpm.a) containing system call processing blocks and also provides common processing blocks as objects (rxcore.o/rxtmcore.o). In other words, the RX850 Pro is incorporated into applications by system call processing blocks that reference libraries when linked and common processing blocks that link to objects. See Q.65 for description of differences between librxp.a and librxpm.a, and Q.111 for description of differences between rxcore.o and rxtmcore.o. Q.6 When should the RX850 be used and when should the RX850 Pro be used? A.6 Although both the RX850 and the RX850 Pro comply with the ITRON3 specification, they differ in some ways. Their main differences are described below. * The RX850 enables only static generation of resources while the RX850 Pro enables both static and active generation. * The RX850's memory pools include both variable-length and fixed-length memory pools while the RX850 Pro has only variable-length memory pools. The RX850 is positioned as a subset of the RX850 Pro, with a smaller code size and less RAM capacity usage than the RX850 Pro. The RX850 is also slightly faster than the RX850 Pro. However, since it uses r0 relative instructions, certain restrictions apply as to how memory is allocated. By contrast, the RX850 Pro has a larger code size and more RAM capacity usage than the RX850, and no such restrictions apply to how memory is allocated. Also, since it enables resources to be generated actively, it allows more leeway in creating applications. Either RX version is able to operate in V850 Series devices, but, in the view of the differences described above, we recommend the RX850 when using V850 Series devices with smaller memory resources and the RX850 Pro when using V850 Series devices with more memory and higher MIPS values. It may be useful to refer to a document for more detailed specification of these differences (RX850/RX850 Pro specification comparison, document number: SUD-T-4961). To obtain this document, contact an NEC Electronics sales representative. Q.7 Which compilers does the RX850 Pro support? A.7 Currently, the RX850 Pro supports the NEC Electronics CA850 compiler and the GHS (Green Hills Software) CCV850E compiler. The RX850 Pro product is packaged with support for both of these compilers, and one or the other must be selected at the time of installation. User's Manual U13773EJ3V0UM 239 APPENDIX B Q & A Q.8 I'm interested in using the RD850 Pro task debugger. Is that sold separately? A.8 The RD850 Pro is standard-equipped with the RX850 Pro. For use instructions, see the RD850 Pro User's Manual. Q.9 I'm interested in using the AZ850 system performance analyzer. Is that sold separately? A.9 The AZ850 is not equipped with the RX850 Pro (it is sold separately). However, it is included in the SP850 package. The AZ850 can be used to obtain information such as task transitions and CPU usage rates. Q.10 What caution points should I note when migrating from RX850 Ver. 3.13 to RX850 Pro? A.10 At the system call level, compatibility is maintained, so there is no need to modify the programs per se. However, the argument types used in system calls differ slightly. The layout and location of links in system information, etc., and the memory management methods differ significantly, and so those aspects must be reviewed and modified accordingly. The migration procedures are described in a document (document excerpted from the RX850 Pro RX850 (Ver. 3.1x), document number: SUD-T-4816), so refer to this document. Electronics sales representative. Q.11 What is a CF (configuration) definition file? To obtain this document, contact an NEC A.11 A CF definition file is a file that contains various data that is provided to the nucleus (kernel). The user describes various information (task information, etc.) using a specified format, and the file is then processed by a configurator (cf850pro) and converted to an assembly language file. 240 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.12 What is defined by the RX850 Pro's CF definition files? A.12 The configuration files define system information (SIT) and system call information (SCT). The SIT (System Information Table) information includes the following: * System information * System maximum value information * System memory information * Task information * Event flag information * Semaphore information * Mailbox information * Memory pool information * Cyclic handler information * Priority information * Indirectly activated interrupt handler information * Clock interrupt information * Extended SVC handler information The SCT (System Call Table) information includes the following: * Task management/task-associated synchronization management function system call information * Synchronous communication management function system call information * Memory pool management function system call information * Time management function system call information * System management function system call information Q.13 How are CF definition files created? A.13 They are text files that are created using a specified format. For description of the format, see CHAPTER 6 CONFIGURATION FILE in the RX850 Pro Installation User's Manual (U13774E). User's Manual U13773EJ3V0UM 241 APPENDIX B Q & A Q.14 What is the configurator (cf850pro)? A.14 The configurator is an application that uses user-generated information from CF definition files to create table information and other information needed by the real-time OS (RX850 Pro). The configurator is an MS-DOS application and the executable file (for the RX850 Pro) is called cf850pro.exe. The configurator can be started by entering a command at the DOS prompt. It can also be started from NEC Electronics PM plus (Ver. 3.15 or later). In other words, when the configuration file is updated, cf850pro is automatically started in order to create SIT and SCT information. Q.15 What are the options when starting the configurator (cf850pro)? A.15 The options when starting the configurator are as follows. cf850pro -i sit_file -c sct_file -d h_file cf_file [return] sit_file: System information table (.sit) sct_file: System call table (.sct) h_file: cf_file: Q.16 What is the "key ID" that is specified when used to create resources in CF definition files? System information header file (.h) CF definition file (.cf) A.16 A key ID is required if the setting "automatically assign ID numbers" is entered when creating resources. Users do not need to pay attention to key IDs when manually assigning ID numbers. Specify "0" to enter the setting "do not use key ID." When auto creation of ID numbers has been specified, ID numbers are automatically assigned by the configurator and the user is not able to tell what the ID numbers of resources are. That is why key IDs are needed. A unique key ID is assigned to each resource when the resource is created. Once the user has the specified key ID number, the user can use a system call (vget_*id) to acquire the ID number of the corresponding resource. 242 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.17 When the following is set and executed in a CF definition file: mem SPOL0 0xffffd200 0x0000fe0 the configurator (CF850 Pro) outputs the following error message. E2259: enough system memorypool "SPOL0" block size The memory estimates are described below. Although the memory pools are not used, what areas must be allocated other than for the task stacks, interrupt stacks, and management objects? Task A stack Task B stack Task C stack Sub total Interrupt stack Management objects System management table Task management block Semaphore management block Interrupt handler Cyclically activated time management Sub total Total 504 + 4 + 16 + 32 = 556 bytes 3 x 56 1 x 20 6 x 16 + 104 3 x 40 = 168 bytes = 20 bytes = 200 bytes = 120 bytes 1064 bytes (0x428) 0x9c4 + 0x1f4 + 0x428 = 0xfe0 0x1f4 bytes 0x3e8 bytes 0x3e8 bytes 0x9c4 bytes 0x1f4 bytes ----------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------- A.17 The task stack size requires that 100 bytes (28 bytes for the task stack management table and 72 bytes for the context area) be added to the size specified by the CF definition file. For details, see 5.3 Capacity of Task Stack in the RX850 Pro Installation User's Manual (U13774E) or use the "RX850 Pro Memory Size Estimate page" in the FAQ at the NEC Electronics Microcomputer website. Task A stack Task B stack Task C stack Sub total 0x1f4 + 100 0x3e8 + 100 0x3e8 + 100 = 600 bytes = 1100 bytes = 1100 bytes 2800 bytes ---------------------------------------------------------------------------- User's Manual U13773EJ3V0UM 243 APPENDIX B Q & A In addition, the interrupt handler stack frame size (idle task = 80 bytes) must be added to the interrupt handler stack area. For details, see 5.4 Capacity of Stack for Interrupt Handler in the RX850 Pro Installation User's Manual (U13774E). Interrupt stack Total 0x1f4 + 80 = 580 bytes 2800 + 580 + 1064 = 4444 bytes (0x115c) In such cases, a value of 0x115c or more must be specified to SPOL0 in the CF definition file. Q.18 How are tasks created? A.18 They are created statically by CF definition files or when the "cre_tsk system call" is issued from an application. Settings that are entered when a task is created include the task name (ID), task activation address, task stack size, task initial priority, task initial status, the task activation code, interrupt status at activation, values specific to the GP and TP registers, and the key ID number. Q.19 How many tasks can be created? A.19 The number of tasks specified by the maxtsk (maximum number of tasks) value can be created. However, the highest value that can be specified for maxtsk as this maximum number is 32767, so up to 32767 tasks can be created. Q.20 What is the initialization handler? A.20 The initialization handler is the "handler that is executed before the scheduler is activated." The scheduler is activated as part of the RX850 Pro's processing that occurs after a startup routine activates the RX850 Pro's initialization processing. The initialization handler's address can be specified by the CF definition file (it is specified by the "ini" setting). In the sample, this is the address called "varfunc." Thus, the initialization handler provides a convenient means for coding processing that precedes system activation, such as task activation and hardware initialization. In cases where use of the initialization handler is not desired, it should not be specified in the CF definition file. If this specification is omitted, the RX850 Pro will operate without an initialization handler, and will therefore have a slightly smaller code size. 244 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.21 Does the RX850 Pro include any idle tasks? A.21 The default setting is that it does not include idle tasks, but the user is able to create tasks that have the lowest priority in the system and use them as idle tasks. In cases where there are neither idle tasks nor other tasks to be executed, the OS issues a HALT instruction to stop the CPU. An idle task must be created in order to use the "STOP" and "IDLE" modes that are available as low power modes in the V850 Series devices. In other words, this can be implemented by writing code that sets "STOP" or "IDLE" into an idle task. When this is done, the idle task must be created as the lowest-priority task in the system. Q.22 What happens when there are no more tasks to be executed? A.22 The CPU enters HALT mode. In other words, HALT processing is executed in the RX850 Pro. Q.23 What is the task priority range in the RX850 Pro? A.23 This range must be within the "task priority range (maxpri)" specified in the CF definition file. This priority specification range is from 1 to 252. Q.24 Are there any caution points to note when specifying the task priority range? A.24 Only set a range within the task priority range that will actually be used. The reason for doing this is that it affects the amount of time the RX850 Pro needs to search the ready queue. Only the range of "task priority numbers" set in the CF definition file will be used to create the ready queue (which requires four bytes per priority number). In other words, a larger range of task priority numbers increases the RAM size. In addition, when searching for the task used to activate the RX850 Pro scheduler, problems may occur if the target task is not found quickly and, in the worst case, all numbers from the highest to lowest priority may have to be searched. This results in longer processing times. That is why it is best to restrict the priority numbers to the range actually needed for use. User's Manual U13773EJ3V0UM 245 APPENDIX B Q & A Q.25 How is a task stack reserved in memory? A.25 Only the amount of memory set by the "task stack size" specified in the CF definition file's task information is reserved from the stack area. The stack area can be specified in either the SPOL0 or SPOL1 system memory area. These addresses are set by the "system memory information (mem)" specified in the CF definition file. Note that no problems occur when the task stack is set only to the SPOL0 or only to the SPOL1 area. Q.26 How large should be task stack be? A.26 The size of the task stack depends on the application. The stack is used for the following purposes. * To save the task context when a task is preempted * To save register information when an interrupt occurs * To save local variables The task context size and stack frame size during interrupts are clearly specified. For details of these, see CHAPTER 5 MEMORY AND MEMORY CAPACITY ESTIMATION in the RX850 Pro Installation User's Manual. The local variables also differ, depending on how many are used and their sizes. When using an array that has a large number of elements, a fairly large stack is utilized. During development, or at other times stack size estimates are only vague, a large stack should be reserved. The stack size can then be reduced if possible later, after a clearer estimate has been made. Also, another means of estimating is to use the RD850 Pro to compare the initial task's stack pointer value (SP value) with the stack pointer value when the application is operating (i.e., when switching to another task). 246 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.27 When several tasks are being created, are task stacks allocated separately to internal RAM and external RAM? A.27 Yes, they are. Task stack areas are defined in CF definition files at the same time as the tasks themselves. Although task stacks are allocated to one of the system memory pools (SPOL0 or SPOL1), and when SPOL0 and SPOL1 are defined, they should be allocated either to internal RAM or external RAM. When a task is created statically in a CF definition file is given as an example. For example, when SPOL0 is allocated to external RAM and SPOL1 to internal RAM, SPOL0 and SPOL1 are defined as follows. mem SPOL0 0x200000 0x1000 mem SPOL1 0xffffe000 0x1000 Furthermore, when TASK01 stack (0x200 bytes) is allocated to external RAM and TASK02 stack (0x100 bytes) is allocated to internal RAM, these tasks are defined as follows. tsk TASK01 TTS_RDY 0x0 0x0 TA_ASM _task01 0x8 TA_ENAINT 0x200:SPOL1 no_use no_use 0x0 tsk TASK02 TTS_RDY 0x0 0x0 TA_ASM _task02 0x6 TA_ENAINT 0x100:SPOL0 no_use no_use 0x0 Also, when tasks are created actively by the cre_tsk system call, they are specified by the arguments in cre_tsk. Q.28 When the rel_wai system call (forced wake-up wait release) is issued to forcibly release the wait status of other tasks, are the target number of task wake-up requests cleared? A.28 The rel_wai system call does not clear the number of task wake-up requests. In other words, when there have been multiple wake-up requests, the request count remains the same. The can_wup system call is used to clear the number of wake-up requests to 0. Q.29 Can the #pragma rtos_task instruction that is included in the NEC Electronics compiler (CA850) be used? A.29 Yes, it can be used. When this instruction is used, the compiler recognizes the specified function type as being for a real-time OS task and outputs code accordingly. With an ordinary function, prologue processing and epilogue processing of the function are required and the corresponding code is output by the compiler. However, this code is not needed in the real-time OS. When the #pragma rtos_task instruction is used to specify function type as being a task, there is no output of the code for prologue processing and epilogue processing of the function and the code size is therefore reduced. User's Manual U13773EJ3V0UM 247 APPENDIX B Q & A Q.30 How is the mailbox created? A.30 It can be created statically by a CF definition file or by issuing a cre_mbx system call from an application. Q.31 How many mailboxes can be created? A.31 The number of mailboxes up to the specified maximum number can be created. However, the highest value that can be specified as the maximum number is 32767, so up to 32767 mailboxes can be created. Q.32 Are there any restrictions on the sizes of messages when using mailboxes? A.32 There are no such restrictions. Any size allowed by the available memory can be used. Q.33 Are the messages themselves being sent to mailboxes? A.33 No, only the messages' start addresses (pointers) are being sent. Q.34 What area should be used for writing messages? A.34 For the sake of facilitating memory area management, we recommend using memory blocks that are managed by the RX850 Pro (see Q.54). The specific sequence is to obtain (with get_blk) the memory area to be used for messages, write messages to that area, and then send the message (with snd_msg) to the mailbox. The receiving side receives the message (with rcv_msg), reads its contents, and then releases the memory area (with rel_blk) that was used by the message. However, when using del_mbx to actively delete the mailbox where messages are queued, that area can be released by the RX850 Pro if the message area is a memory block, but if it is an area not controlled by the RX850 Pro, the delete operation cannot be guaranteed. 248 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.35 Can priorities be assigned to messages? A.35 Yes, they can. When a prioritized message is sent to the mailbox that has a message priority (as specified by the TA_MPRI attribute), the message is queued according to its priority order. The priority number is stored in the message itself. It is stored in a two-byte area that is five bytes from the start of the message. When handling prioritized messages, writing to the target area needs to be kept organized. Generally, the area for prioritized messages is written 8 bytes ahead of the start in order to maintain alignment. Q.36 How are messages queued in mailboxes? A.36 There are two methods for queuing messages in mailboxes. One method is called "order of message admission (FIFO order)" and other is called "order of message priority (PRI order)". One of these queuing methods is specified as a mailbox attribute in a CF definition file. In other words, the queuing method is determined separately for each mailbox. The TA_MFIFO attribute is used to specify the first method and the TA_MPRI attribute is used to specify the second method. Q.37 Can messages be sent for specified tasks? A.37 There is no function that specifies the task when sending a message. However, this sort of operation can be done by creating a dedicated mailbox to receive a particular message (see Q.30). Q.38 If several tasks are waiting for messages, in what order are the messages transferred? A.38 They are transferred either in order of waiting message or according to the tasks' priority. The order to be used is specified by an attribute for the target mailbox. When creating a mailbox, if the specified attribute is TA_TFIFO, messages are queued in the waiting order and if the specified attribute is TA_TPRI, they are queued in order of task priority. Reference example: -- Mailbox information mbx mbx 0x1 0x2 0x0 0x0 TA_TFIFO TA_MFIFO TA_TPRI 0x1 TA_MFIFO(or TA_MPRI) 0x2 User's Manual U13773EJ3V0UM 249 APPENDIX B Q & A Q.39 0xffffffab was the result when referencing the value returned by the debugger in response to the prcv_msg system call. When this value is referenced as a hexadecimal value, it is expected to be "-55" but there is no such return value of "-55" in the user's manual. How should this return value be referenced? A.39 The system call return value described in the user's manual is a decimal value. When 0xffffffab is expressed as a signed decimal number, it is "-85." The prcv_msg return value "-85" is returned by E_TMOUT when there are no messages in the target mailbox. Q.40 How are event flags created? A.40 They are created either by a CF definition file or by using the cre_flg system call in a program. Q.41 How many event flags can be created? A.41 The number of event flags up to the specified maximum number can be created. However, the highest value that can be specified as the maximum number is 32767, so up to 32767 event flags can be created. Q.42 What is size (in bits) of event flags? A.42 32 bits Q.43 With the RX850 Pro, can several tasks wait for an event that uses just one event flag? A.43 Yes, they can. When event flags are created, the setting can be either "enable only single task (TA_WSGL attribute)" or "enable multiple tasks (TA_WMUL attribute)." 250 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.44 In the RX850 Ver. 3.13, there are 1-bit event flags, but are there any of these in the RX850 Pro? A.44 No, there are not any 1-bit event flags in the RX850 Pro. 1-bit event flags are used in the RX850 Ver. 3.13, but when the program is upgraded to the RX850 Pro, all event flags must be 32-bit event flags. Likewise, any system calls related to 1-bit event flags should be converted to 32-bit event flags. For details of how to make these changes, see the document (SUD-T-4816-2) describing migration from the RX850 (Ver. 3.1x). To obtain this document, contact an NEC Electronics sales representative. Q.45 What is the timing of the clear processing related to the "event flag clear specification" (TWF_CLR) for the wai_flg, twai_flg, and pol_flg arguments? A.45 After set_flg, a separate routine is used to clear event flags; not the routine within wai_flg, twai_flg, or pol_flg. Next, a check is done to determine whether or not any tasks are waiting for an event flag. If there are any such tasks, their wait conditions are checked, the wait status is cleared, and then the flag is cleared. Since the contents of the separate routine operate during interrupt disable mode set by the di instruction, any set_flg specification within an interrupt will not change flag settings. Also, with regard to task transitions, a scheduler activation request is issued following completion of processing such as clearing of wait statuses, so flag settings are not changed by other tasks. Q.46 In cases where multiple tasks are waiting for an event set by wai_flg to occur in relation to a particular bit in a single event flag, how are tasks activated once the event has occurred? A.46 When multiple tasks are waiting for an event set by a single event flag, if the condition for clearing the wait status is the same, once the wait status is cleared all of the waiting tasks are set to ready state. Afterward, the scheduler is activated and the highest-priority task is set to run state. In other words, at the time when the set_flg system call is issued, it is not just the highest-priority task that gets activated; all of the waiting tasks are activated. After all of these tasks are activated, the activated task that has the highest priority is set to run state. However, it is not absolutely necessary that the task that has the highest priority among the tasks that had been in event flag wait status will be set to run state. User's Manual U13773EJ3V0UM 251 APPENDIX B Q & A Q.47 When the set_flg system call is issued, if the wai_flg system call was used to specify bit pattern clearing (TWF_CLR) for even one of the tasks that was waiting for the event flag set by that system call, is the flag actually cleared? A.47 Clearing of an event flag wait state occurs after the event flag wait queue (FIFO) is checked starting from the start of the queue, when the task that establishes the wait condition is activated. At that time, if clearing of the task has been specified, the bit pattern is cleared immediately. Accordingly, if clearing has been specified for the task that establishes the wait condition, then the subsequent tasks that are still in the wait queue are not activated when this wait condition is established. An example in which one task is waiting is shown in 4.4.5 Messages. Q.48 How are semaphores created? A.48 They are created either by a CF definition file or by using the cre_sem system call in a program. Q.49 How many semaphores can be created? A.49 The number of semaphores up to the specified maximum number can be created. However, the highest value that can be specified as the maximum number is 32767, so up to 32767 semaphores can be created. Q.50 Where is the number of initial semaphore resources set? A.50 It is set by a CF definition file or by the cre_sem system call when creating semaphores. Q.51 Is there an upper limit on the number of semaphore resources? A.51 Yes, there is an upper limit. The upper limit is 0x7fffffff (2147483647). Q.52 When semaphore resources exist, what happens if an operation is performed to return the resources? A.52 Semaphores include resource counters that are incremented when such an operation is performed. 252 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.53 Can management objects related to resources such as semaphores be assigned to specific resources in SPOL0 or SPOL1? Also, if both SPOL0 and SPOL1 are defined by a CF definition file, which one is used first when they are activated? A.53 All resources including semaphores are managed by the system. This means that the system maintains management tables for all resources in SPOL0 and they are not assigned to SPOL1. Before the RX850 Pro is activated, there is no concept of SPOL0 or SPOL1. In other words, in order to use stacks, a stack area must be reserved as part of the boot processing. See the sample programs provided with this product for further description. After the RX850 Pro is activated, management tables are created as the system area in SPOL0. Note that SPOL0 must be defined. Be careful to note that it is not possible to define only SPOL1. Q.54 How is memory management set up? A.54 In the RX850 Pro, memory is managed in two main categories: memory pools and memory blocks. There are several memory blocks in each memory pool. Requests to obtain and return memory areas for these memory pools are issued by tasks or the interrupt handler. These obtained or returned memory areas are called memory blocks. Q.55 How are memory pools created? A.55 They are created by a CF definition file or by issuing a cre_mpl system call. The memory pool name (ID), task queuing method, memory pool size, and key ID are set in a CF definition file. Q.56 How many memory pools can be created? A.56 The number of memory pools up to the specified maximum number can be created. However, the highest value that can be specified as the maximum number is 32767, so up to 32767 memory pools can be created. Q.57 Are there any fixed-length memory pools? A.57 No, all are variable-length memory pools. If fixed-length memory pools are used when migrating an application from the RX850, they must all be replaced by variable-length memory pools. User's Manual U13773EJ3V0UM 253 APPENDIX B Q & A Q.58 When the system memory pools (SPOL0 and SPOL1) and/or user memory pools (UPOL0 and UPOL1) are defined by a CF definition file, is it necessary to avoid placing them in the same area as the .bss or .data section? Also, are there any memory pools that must be placed in internal RAM? A.58 SPOL0, SPOL1, UPOL0, and UPOL1 all specify direct addresses and they occupy an area whose size is specified starting from the specified address. Although variables are placed in the .bss or .data section by a compiler or linker, compilers and linkers are not recognized in relation to the above-mentioned memory areas that are used by the RX850 Pro. Consequently, layout of .bss or .data section in the SPOL0, SPOL1, UPOL0, and UPOL1 areas should be avoided. When the linker uses link directives to reserve in advance segments to be used by the RX850 Pro, it tends to avoid problems caused by overlapping with other areas (for a relevant sample, see Q.63). However, it is not possible to detect problems such as overflows. Also, there are no restrictions for the RX850 Pro with regard to having to place the above memory areas in either internal or external memory space. Q.59 In variable-length memory pools, is any "garbage collection" done to organize the unused areas that are left after repeated acquisition and release of memory blocks? A.59 No, it is not done. This is because the C language which is used in applications uses pointers, etc., to handle direct addresses. If the RX850 Pro were to perform such garbage collection and move entities (memory blocks), contradictions would occur in the application between the addresses held by pointers and the actual addresses of such entities, which could cause operation faults in the application. With the RX850 Pro, when rel_blk is used to return memory blocks, the spaces before and after each returned memory block are checked and if either (or both) of these spaces is empty (unused), it is merged to create a large empty space and minimize fragmentation of memory pool. However, if the empty spaces are not contiguous, any memory block that is larger than the largest empty space cannot be obtained. 254 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.60 What are the meanings of SPOL0, SPOL1, UPOL0, and UPOL1? A.60 Their meanings are listed as follows. * SPOL0 ... System Memory Pool 0 * SPOL1 ... System Memory Pool 1 * UPOL0 ... User Memory Pool 0 * UPOL1 ... User Memory Pool 1 Memory areas that are used as resources for the RX850 Pro are assigned to one of the above memory pools. Specifically: * Interrupt stack area: SPOL0 or SPOL1 * Task stack area: SPOL0 or SPOL1 * Memory pool area: UPOL0 or UPOL1 The start address and size of SPOL0, SPOL1, UPOL0, and UPOL1 are set by CF definition files. Setting example: -- Memory information mem mem mem mem mem Q.61 Would assigned stacks only to SPOL0 (System Pool 0) cause any problems? SPOL0 SPOL1 UPOL0 UPOL0 UPOL1 0x1000 0x2000 0x3000 0x20000 0x30000 0x0000 0x1000 0x7000 0x2500 0x1500 A.61 No, it would not cause problems. User's Manual U13773EJ3V0UM 255 APPENDIX B Q & A Q.62 When a startup routine (boot.s or boot.850) is used to set a stack pointer (sp), how should the size be determined? A.62 Before starting up the RX850 Pro, if there is any use of the stack, then the corresponding stack pointer is used. For example, when a function call is executed and the lp value at that time is retained, the target stack is the one specified by the stack pointer (sp). This is also the case when a function call occurs within a function. In the startup routine that is included with the product as a sample, a stack is not used in the NEC Electronics version before the RX850 Pro is started up, i.e., up until the jump to kernel initialization by jmp [lp]. However, in the GHS version, there is a function call for memory initialization, so a stack is used to save the lp for that section (see Q.66). Still, as is shown in the sample, a size as large as 0x28000 bytes is not used. The reason why 0x28000 is specified in the sample is that the link directive file overlaps this area with the memory information used by the RX850 Pro so that, after the RX850 Pro is started, this section is used as a memory area (specified by mem). The total size specified by mem in the sample's CF definition file is 0x28000. In the sample, if the stack is not regulated, a user who is using a modified version of the sample might issue a function call without first setting up a stack, which can result in operation faults. That is why an ample stack size is used. One way to determine the stack size without having to use a startup routine, is to use the debugger to check the approximate stack value and then use that value as the stack size. Alternatively, if the size specified by mem is used, there will not be much space wasted, so that value can also be used. Q.63 With the RX850 Pro, since the linker cannot be used to detect memory areas specified by mem in a CF definition file, there is no restriction against overlapping with memory areas used by a user program. What should be done to detect an "overlap" error when linking? A.63 Since the CF definition file and link directive file are not linked, the link directive file must be modified. A method for doing this is described below, using the RX850 Pro sample program as an example. The system memory pools and user memory pools defined by the sample CF definition file (sys.cf) are as follows. -- memory information mem SPOL0 0x00110000 0x00010000 mem SPOL1 0x00120000 0x00010000 mem UPOL0 0x00130000 0x00008000 256 User's Manual U13773EJ3V0UM APPENDIX B Q & A This works as long as this area is reserved as a dummy section. [NEC Electronics version] The following code is added to the startup routine (boot.s). ------------------------------------------ /*** SPOL0 area specifying ***/ .section ".spol0", .lcomm .lcomm __spol0_end, bss 0x10000, 0, 4 4 __spol0_head, -- /*** SPOL1 area specifying ***/ .section ".spol1", .lcomm .lcomm __spol1_end, bss 0x10000, 0, 4 4 __spol1_head, -- /*** UPOL0 area specifying ***/ .section ".upol0", .lcomm .lcomm __upol0_end, bss 0x8000, 0, 4 4 __upol0_head, ---------------------------------------Next, the following code is added to the link directive. -----------------------------------------------------------SYSPOL0 }; SYSPOL1 }; USERPOL0 : !LOAD ?RW .upol0 }; -----------------------------------------------------------V0x00130000 { = $NOBITS ?AW .upol0; : !LOAD ?RW .spol1 V0x00120000 { = $NOBITS ?AW .spol1; : !LOAD ?RW .spol0 V0x00110000 { = $NOBITS ?AW .spol0; User's Manual U13773EJ3V0UM 257 APPENDIX B Q & A [NEC Electronics version] The following code is added to the startup routine (boot.850). ----------------------------------------- /*** SPOL0 area specifying ***/ .section ".spol0", bss -- /*** SPOL1 area specifying ***/ .section ".spol1", bss -- /*** UPOL0 area specifying ***/ .section ".upol0", bss ---------------------------------------Next, the following code is added to the link directive. ---------------------------------------: : .spol0 .spol1 .upol0 0x00110000 0x00120000 0x00130000 : : ---------------------------------------This enables the area to be reserved by sys.cf to be recognized by the linker. Consequently, when the application is linked, an error can be detected if this area overlaps with another area. However, certain other errors such as overflows cannot be detected. : : : 258 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.64 The return value for the rel_blk system call was a negative value. A negative value should not occur. Why did this happen? A.64 The RX850 Pro complies with the ITRON3.0 standard, and when an error occurs during ITRON3.0-compliant system call processing, the return value is set to be a negative value. Q.65 When the rel_blk system call is executed, E_OBJ (-63) is returned each time, and the memory block is never returned. Why? A.65 This phenomenon is unique to the RX850 Pro. When the rel_blk system call is used to return a memory block, if the first four bytes of the memory block are not filled with zeros, the return value E_OBJ is used for termination instead of returning the memory block. This specification was created in consideration of cases where a memory block is used as the message area for a mailbox. When a memory block is used as the message area for a mailbox, the first four bytes become the message's wait queue link area. In other words, when a message is queued in a mailbox and the rel_blk system call is issued, if the specification required that a memory block be returned, what would actually be returned would instead be the queue-linked message area. To prevent this, the link area (first four bytes) must be all zeros in order for it to be regarded as a memory block that is being used as a message area, in which case the return value E_OBJ is used for termination instead of returning the memory block. Therefore, when returning a memory block, the first four bytes must be cleared to zero. However, starting in Ver. 3.15, in consideration of cases where the memory block is not used as a message area, a version of rel_blk that is able to return the memory block even when the first four bytes are not all zeros has been added as a separate library. Function Library containing rel_blk that requires zero-clearing of first four bytes of memory block (same as previous specification for rel_blk) Library containing rel_blk that does not require zero-clearing of first four bytes of memory block (new specification for rel_blk) Library Name librxp.a librxpm.a These specifications of rel_blk are stored in separate libraries so that one or the other rel_blk specification can be used. Link to the library of the rel_blk specification to be used. User's Manual U13773EJ3V0UM 259 APPENDIX B Q & A Q.66 In the GHS version, an endless loop occurs after the sample's memory initialization processing. Why? A.66 In the GHS version's sample, there is a place where meminit.c is called from boot.850. In meminit.c, RAM initialization processing is performed, but if processing cannot exit from that point or returning from initialization processing, operations are repeated from the start of boot.850. This is caused by a damaged stack. Since the stack area that is used in the startup routine is also zero-cleared in meminit.c, this type of phenomenon occurs. Here are two fixes: * Do not initialize stack areas that are used in startup routines. * Temporarily save stack areas that are used in startup routines to a separate location. For further description of stacks in startup routines, see Q.62. Q.67 Where should the processing that is executed when an interrupt occurs be coded? A.67 It should be coded in the interrupt handler. The interrupt handler is a dedicated interrupt handling routine that is activated as soon as an interrupt occurs, and as such it is handled independently from tasks. In the RX850 Pro, there are two types of interrupt handlers: a directly activated interrupt handler and an indirectly activated interrupt handler. Q.68 What is the difference between the directly activated interrupt handler and the indirectly activated interrupt handler? A.68 Directly activated interrupt handler: When an interrupt has occurred, this interrupt handler is activated without going via the RX850 Pro. However, the user is responsible for coding the register save processing and stack switching processing (a macro is provided for saving and restoring registers). Indirectly activated interrupt handler: When an interrupt has occurred, this interrupt handler is activated only after the RX850 Pro has performed interrupt preprocessing such as register save processing and stack switching processing. 260 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.69 Which is better to use: the directly activated interrupt handler or the indirectly activated interrupt handler? A.69 Since the directly activated interrupt handler starts handler processing as soon as an interrupt occurs, it can be expected to provide faster processing. However, since users must themselves perform the processing to save contents of registers and stacks, this makes coding of the handler more complicated. On the other hand, when an interrupt occurs, the indirectly activated interrupt handler shifts processing to the RX850 Pro, which saves the contents of registers and stacks, before passing processing to the handler. Although the indirectly activated interrupt handler has somewhat slower responsiveness compared to the directly activated interrupt handler, it is simpler since it spares the user from having to code only the handler processing. The user should take these factors into consideration when selecting which type of handler to use. Q.70 How should the directly activated interrupt handler be registered? A.70 The directly activated interrupt handler can be registered by assigning it to the handler address where control is passed by the processor when an interrupt occurs, or by setting a branch instruction to the directly activated interrupt handler. For details of this programming, see A.5 Directly Activated Interrupt Handler. Q.71 How should the indirectly activated interrupt handler be registered? A.71 It is registered by being programmed in the CF definition file. It can also be registered actively by using the def_int system call. User's Manual U13773EJ3V0UM 261 APPENDIX B Q & A Q.72 What is the method for registering a timer interrupt? A.72 Each timer is started as an indirectly activated interrupt handler. For example, the source code appears as follows. (CA850) ------------------------.section "INTCMD" RTOS_IntEntry_Indirect ------------------------(CCV850) ------------------------.org 0x00000240 RTOS_IntEntry_Indirect ------------------------RTOS_IntEntry_Indirect should be used as shown above. This enables insertion of a timer interrupt. Q.73 Are multiple interrupts supported? A.73 Yes, they are. Multiple interrupts are when one interrupt is inserted while another interrupt handler is being processed. However, in the RX850 Pro, when an interrupt handler is started, its initial setting is "interrupt disabled," so interrupts must be enabled within the interrupt handler in order to acknowledge multiple interrupts. In other words, the user must exercise control as to whether interrupts will be enabled (set via EI or the ena_int system call) or disabled (set via DI or the dis_int system call). Q.74 Can the #pragma rtos_interrupt in the NEC Electronics compiler (CA850) be used? A.74 That is a pragma directive that is provided in order to simplify the coding of the real-time OS's directly activated interrupt handler. Currently, however, normal operation is not possible when that pragma directive is used to program a directly activated interrupt handler. Therefore, due to this restriction, it should not be used. For details of programming the directly activated interrupt handler, see A.5 Directly Activated Interrupt Handler. 262 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.75 What is the method for returning from an interrupt handler? A.75 The return method differs for directly activated and indirectly activated interrupt handlers. To return from a directly activated interrupt handler, use the macro that is provided in the handler's end section (the interrupt return processing is coded in this macro). For description of the macro name and use method, see A.5 Directly Activated Interrupt Handler. To return from an indirectly activated interrupt handler, use "return (TSK_NULL)" in the handler's end section if for ordinary return from an indirectly activated interrupt handler. Use "return (task ID number)" if a wake-up request is being issued for a task that is specified by a parameter. Q.76 Although the following type of return value is required for indirectly activated interrupt handlers, what happens when registering a function in which "return (TSK_NULL)" has been deleted and there is no return value, as a handler? ID func_inthdr() { /* Processing of indirectly activated interrupt handler */ ................................................ ................................................ ................................................ /* Return processing from indirectly activated interrupt handler */ return(TSK_NULL); } A.76 When registering a function that has no return value as a handler, there is a risk of a runaway condition. Although return values are returned within the RX850 Pro when stored to r10, the r10 value is checked, and if it is not TSK_NULL (= 0), then the task that has the corresponding ID number is woken up. In other words, operation is normal as long as 0 has been stored to r10, but when that is not the case an unexpected task may be woken up and linked to the ready queue and may actually be activated if it is a high-priority task. Consequently, the desired operation may not occur unless return from an indirectly activated interrupt handler has been programmed. Therefore, be sure to include either "return (TSK_NULL)" or "return (ID tskid)" at the end of the interrupt handler. User's Manual U13773EJ3V0UM 263 APPENDIX B Q & A Q.77 Where is the interrupt handler's stack area reserved? A.77 The interrupt handler's stack area is reserved at the location specified by the CF definition file. The CF definition file specifies the stack size and reserved area (SPOL0 or SPOL1). Code example: -- System information clktim clkhdr defstk intstk prttsk prtsem prtflg prtmbx prtmpl 0x1 0x7 0x100 0x100:SPOL0 Stack information for interrupt handler 0x1 0x1 0x1 0x1 0x1 However, the interrupt handler's stack is not used as soon as the interrupt has been inserted. If the interrupt is inserted while a task is being performed, the register contents to be saved are pushed onto the stack of the interrupted task. Afterward, the stack pointer (SP: r3) switches to the interrupt handler's stack (to be precise, the task's lp and sp values are pushed onto the interrupt stack). In the case of multiple interrupts, since an interrupt is inserted during processing of another interrupt, the interrupt handler's stack is used. In other words, the more multiple interrupts are inserted, the more the interrupt handler's stack gets consumed. 264 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.78 When returning from a directly activated interrupt handler, the global pointer (GP) value becomes incorrect and a runaway condition occurs. Why? A.78 When returning from a directly activated interrupt handler, it is possible that the GP value may get corrupted, which can then cause operation problems. This is because directly activated interrupt handlers do not guarantee TP and GP values, so whenever these values must be used within the handler, they must be set at the start of the handler. Code example for version that supports CA850: -----------------------------------------------------------#include ID inthdr_body( ) { __asm("mov #__tp_TEXT, tp"); __asm("mov #__gp_DATA, gp"); /* Processing of directly activated interrupt handler func_inthdr */ ....................................... ....................................... ....................................... /* Return processing from directly activated interrupt handler func_inthdr */ return tskid } -----------------------------------------------------------For details, see A.5 Directly Activated Interrupt Handler. Using macros (RTOS_IntEntry and RTOS_IntReturn) at the start and end of the directly activated interrupt handler can save and restore the GP value, but the value is not set. This is because, due to the characteristics of the directly activated interrupt handler, it is not an interrupt handler that can be reliably controlled by the RX850 Pro. When a timer handler is being activated, the GP (r4) value is used as temporary register. If a direct interrupt is inserted at that time, the correct GP value will not be set and operation will be abnormal. However, in the case of an indirectly activated interrupt handler, setting of a specific GP and specific TP are both done by the RX850 Pro. For detailed description of this coding, see A.6 Indirectly Activated Interrupt Handler. User's Manual U13773EJ3V0UM 265 APPENDIX B Q & A Q.79 If a system call is issued during non-maskable interrupt processing, what kind of operation occurs? A.79 Since non-maskable interrupts are not subject to interrupt prioritization, they are acknowledged with higher priority than all other interrupts. Consequently, NMIs are acknowledged even when operating in interrupt disabled mode. In the RX850 Pro, management information and queues may be overwritten when system calls are issued. If an NMI occurs while the RX850 Pro is processing management information and if a system call is issued during this processing, the management information may not be updated correctly and subsequent operations may lead to a runaway condition. Therefore, in the RX850 Pro, if a system call is issued within a non-maskable interrupt handler, subsequent operations are not guaranteed. Q.80 Interrupt handlers cannot be activated. Why? A.80 The following causes are possible. * Has an interrupt actually been inserted? Use the debugger to set a break point at "interrupt handler address (vector)" and check whether or not the target interrupt has been inserted. If it has not been inserted, it means there has been no physical insertion of an interrupt. In such cases, it may be necessary to check how hardware is being initialized. Check the interrupt control register settings (if the interrupt mask is open, etc.) and make sure an interrupt trigger has been set. After doing the above, if the interrupt still doesn't occur, it may be due to a problem in the target device. * Does the interrupt source name (interrupt source number) in the CF definition file match? When using an indirectly activated interrupt handler, the interrupt source name is specified in the CF definition file. Unlike in the RX850, this specification is not made as an interrupt source name per se but rather as a value calculated using the formula "target interrupt exception code - 0x80/0x10". Make sure that this value is correct. Also, be sure to check the start address (start of function) for the interrupt handler that is activated when the interrupt is inserted. This does not have to be done when using a directly activated interrupt handler. When using a directly activated interrupt handler, directly code a jr instruction at the interrupt handler address so that processing will branch to the handler at that point. If none of the above solves the problem, it could be that the application did not download correctly or that the link was not established correctly. 266 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.81 When a timer interrupt occurs, if a UART transmit or receive interrupt occurs for a directly activated interrupt handler, the program enters a runaway condition. At the time when the UART transmit or receive interrupt occurs, the TP setting is expected to be TP = 0x00000400, but when checked it was found to be TP = 0xFFFFFFFF. As a result, an illegal branch occurred to the interrupt vector area and a runaway condition ensued. Why did this happen? A.81 The following cause is possible. When using a directly activated interrupt handler in the RX850 Pro, the TP and GP values must be set at the start of the handler. In other words, the TP and GP values to be used by the handler must be determined before starting the handler. [Reference] Q.78 Q.82 During linkage, the message "Warning : register r1 used as source register" appears when using the RTOS_IntEntry macro in the directly activated interrupt handler. Why? A.82 The RTOS_IntEntryIndirect macro that is provided with the RX850 Pro uses reserve register r1, and that causes this warning message to be output. However, there is actually no problem and in this case the warning message can be ignored. Output of this warning message can be prevented by specifying the -w option in as850 to suppress warning messages. However, this option will suppress all warning messages, so it is better to overwrite part of macro.h as described below. In macro.h, line 28: ---------------------------------------/* go indirect interrupt handler */ ld.w jmp sbt_intent[r2], r1 [r1] ---------------------------------------Overwrite as follows. ---------------------------------------.option nowarning ld.w jmp sbt_intent[r2], r1 [r1] .option warning ---------------------------------------- User's Manual U13773EJ3V0UM 267 APPENDIX B Q & A Q.83 What is a cyclic handler? A.83 A cyclic handler is a cyclical processing program available to users. Among all cyclical processing programs, the cyclic handler has the least amount of overhead prior to start of execution. Q.84 How is a cyclic handler created? A.84 It can be created and registered by a CF definition file or a def_cyc system call. Q.85 How many cyclic handlers can be created? A.85 The number of cyclic handlers up to the specified maximum number can be created. However, the highest value that can be specified as the maximum number is 32767, so up to 32767 cyclic handlers can be created. Q.86 How are cyclic handlers started? A.86 In a CF definition file, it is started when "activate" is specified as the initial status. Also, it can be started or stopped by issuing an act_cyc system call during certain tasks. Q.87 Does the cyclically activated handler get activated during interrupt enabled mode? Does it get activated during interrupt disabled mode? A.87 The RX850 Pro's cyclically activated handler does get activated during interrupt enabled mode. To set interrupt disabled mode within the cyclically activated handler, a DI instruction (to issue the dis_int system call) must be used at the start of the handler. However, if an interrupt occurs after the cyclically activated handler is activated but before the DI instruction is executed, that interrupt will be inserted. The inserted interrupt is subject to prioritization, however. For details, see Q.111. 268 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.88 What is the method for returning from a cyclically activated handler? A.88 When using C language, return by issuing the return instruction in the last section of the cyclically activated handler. When using assembly language, return by issuing a jmp [lp] instruction. Q.89 Where are clock interrupt sources specified? A.89 They are specified in the CF definition file. However, specifying a clock interrupt in a CF definition file only registers the clock interrupt. To actually initialize the timer hardware, an initialization handler, initialization task, or startup routine must be programmed. Code example: -- System information clktim clkhdr defstk intstk prttsk prtsem prtflg prtmbx prtmpl Q.90 What is the shortest possible cyclic activation interval for the cyclically activated handler? 0x1 0x7 0x100 0x100:SPOL0 0x1 0x1 0x1 0x1 0x1 Basic clock cycle (unit: [ms]) Clock interrupt source number: (interrupt exception code - 0x80)/0x10 A.90 Although there is no specific restrictions, it is not possible to specify a value that is equal to or less than the "timer's basic clock cycle (clktim)" as specified in the CF definition file. In cases where the cyclically activated handler is activated at a short interval or a lot of processing is performed within the handler, a clock interrupt may be inserted before the cyclic handler's main processing has been completed. In such cases, there is a chance that the cyclically activated handler's processing may get nested for continuation, which would keep other task processing from occurring and or otherwise not operate as expected. Also, in cases where multiple cyclically activated handlers are activated at the same time, it should be noted with caution that their processing may not be completed before the next timer interrupt is inserted. User's Manual U13773EJ3V0UM 269 APPENDIX B Q & A Q.91 The minimum unit for the cyclically activated handler's processing is 1 ms, so what can be done to set up processing that is completed within 100 s? A.91 The RX850 Pro's cyclically activated handler was designed to run on the basic clock, which is input at a frequency of about 1 ms. For processing that is completed in less than 1 ms, use the directly activated interrupt handler instead. However, if interrupt processing were to be performed at a cyclical rate of 100 s, it would drastically impede the progress of the RX850 Pro's processing. Q.92 In a cyclically activated handler that is activated at an interval of 10 ms, what happens when processing that requires more than 10 ms is executed? The interrupts are enabled in the cyclically activated handler. A.92 In the RX850 Pro Ver. 3.13, when the timing requires the next cyclically activated handler be executed while the cyclically activated handler is still being executed, nesting is used to execute both cyclically activated handlers. A transition diagram of this nesting is shown below. Task being executed Cyclic handler 1 Cyclic handler 2 Similar nesting for execution is used whether the cyclically activated handlers are different or the same. Q.93 Is it correct that the interrupt processing is interrupted in a cyclically activated handler? A.93 A cyclically activated handler can be activated during interrupt handler processing in the following cases. * When the target interrupt handler is an interrupt that is activated with a lower priority than a timer interrupt * When an EI section exists in the target interrupt handler * When a timer interrupt is inserted before reaching the EI section, or when a timer interrupt is inserted during the EI section * When the inserted timer interrupt is at the timing where a cyclically activated handler is activated In cases such as these, the priority of the timer interrupt should be considered as dependent on the priority of the cyclically activated handler. 270 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.94 When cyclically activated handlers are created and activated using the def_cyc system call, is the first activation of the cyclically activated handler timed to occur immediately after the def_cyc system call is issued? Or does this happen only after a specified amount of time has elapsed from when the def_cyc system call is issued? A.94 It happens after a specified amount of time has elapsed from when the def_cyc system call is issued. Similarly, the timing for cyclic handler activation by act_cyc also requires a specified amount of time to elapse after issuing the act_cyc. Q.95 When multiple cyclically activated handlers are linked in a cyclically activated handler activation request queue, are tasks required to wait until all of the cyclically activated handler processing has been completed? A.95 When multiple cyclically activated handlers are linked in a cyclically activated handler activation request queue, tasks are required to wait until all of the cyclically activated handler processing has been completed. For details, see 7.6.6 Activation order of cyclically activated handler. Q.96 Is the following information correct with regard to the RX850 Pro interrupt management function's system calls? Maskable Interrupt loc_cpu dis_int unl_cpu ena_int Disabled Disabled Enabled Enabled Scheduler Interrupt Disabled Disabled Enabled Enabled Dispatch Processing Disabled Status before call is retained Enabled Status before call is retained Maskable interrupt: Interrupt connected to the interrupt controller Scheduler interrupt: Interrupt that triggers scheduling (periodic interrupt) Dispatch processing: Scheduling that is triggered by a system call other than a scheduler interrupt A.96 Yes, the above information is correct. User's Manual U13773EJ3V0UM 271 APPENDIX B Q & A Q.97 When a timeout period has elapsed before a task is woken up due to fulfillment of a wake-up condition such as a twai_flg system call, is the return value E_OK? Or is it E_TMOUT? Also, what happens if the fulfillment of a wake-up condition and the timeout happen in reverse order? A.97 If the timeout occurs before the wake-up condition is met, the timeout results in a return value of E_TMOUT. In other words, this reflects the sequence of events. Accordingly, if the wake-up condition is met before the timeout occurs, the return value will be E_OK. Q.98 What is the relation between clktim as defined in a CF definition file and the delay time of dly_tsk (dlytim)? When 10 is set for clktim and dly_tsk (100) is set, the delay becomes one second, and the expected operation does not occur. A.98 The value specified for clktim is actually expressed as the interval of the interrupt inserted by hardware, and the RX850 Pro uses this as the basis for calculating the time for the delay specified by dly_tsk. In other words, if dly_tsk (1000) is specified, the task operation is extended by 1000 ms, but this does not indicate the timer interrupt interval in the RX850 Pro. That is why clktim is used to specify the timer interrupt interval as a number of ms. This enables the calculation of the number of interrupts needed for a 1000 ms delay. Therefore, the expected delay time is not obtained when the actually input timer interval differs from the value specified by clktim. Q.99 What is the RX850 Pro's timer precision? A.99 The RX850 Pro's timer precision is at least 1 ms. Although the timer precision itself depends on hardware settings, the minimum value that can be specified as the "basic clock cycle (clktim)" that is specified in the CF definition file is 1 ms. This value must be combined with the timer cycle that is actually set by hardware. When the RX850 Pro performs time-related processing, this processing is based on the numbers (in ms units) that are specified by system call arguments. For example, when dly_tsk (1000) is specified, it sets a 1000 ms delay for task processing, and this 1000 ms value was calculated based on the value specified by clktim. In other words, if "1" is specified as "clktim", the RX850 Pro interprets that as meaning that 1000 timer interrupts need to be inserted to create a 1000 ms delay. If "5" is specified as "clktim", 200 timer interrupts need to be inserted to create a 1000 ms delay. Note with caution when migrating from the RX850 to the RX850 Pro that this calculation method differs from that used in the RX850. In the RX850, the value specified as the argument is the "number of the timer interrupts inserted (tick)". 272 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.100 This question concerns the clock handler cycle. In the AZ850's Analyze Window, the cycle of the output INTCM40 was approximately 0.113 ms. The reset.c timer was set as follows. TMC40 = 0x86; CM40 = 10; CMIC40 = 0x0; In the V850E/MS1, the expected settings are m/16 for the internal count clock value, /8 for the intermediate clock value, and 10 for the count value, but the output result did not match the calculated result. How should that be interpreted? A.100 It means there is an error in a register setting. In the V850E/MS1, when TMC40 = 0x86 is set via a register, the internal count clock value becomes /32. The internal system clock is calculated as 25 MHz, so that the cycle of INTCM40 (timer 4 interval) becomes approximately 0.113 ms. The specific calculation steps are described below. 1. The timer control register (TMC40) settings are determined based on the selected internal count clock and intermediate clock. When TMC40 = 0x86 (10000110), the internal count clock setting is m/32 and the intermediate clock setting has been selected as /8. User's Manual U13773EJ3V0UM 273 APPENDIX B Q & A [Overview of timer control register (TMC40)] 7 TMC40 CE40 6 0 5 0 4 0 3 0 2 1 0 PRS400 PRM401 PRM400 Bit position Bit name 7 CE40 Controls timer operation 0: No operation 1: Count operation 2 PRS400 Meaning Internal count clock ( m: intermediate clock) 0: m/16 1: m/32 Intermediate clock m ( : internal system clock) PRM401 0 0 1 1 PRM400 0 1 0 1 1, 0 PRM401, PRM400 m /2 /4 /8 RFU (reserved) 2. The interval time is determined using the following formula. (Interval time) = (Compare register value + 1) x (Count clock cycle) The count clock cycle is determined using the following formula. (Count clock cycle) = 1/(Internal count clock) When the internal system clock = 25 [MHz] = 25 x 10 [Hz], the count clock cycle is as shown below. 6 1/(/8 x 1/32) = 1/{25 x 10 x 1/8 x 1/32} 6 = (8 x 32)/(25 x 10 ) 6 Consequently, when the compare register CM40 = 10, the interval time is determined as follows. (10 + 1) x {(8 x 32)/(25 x 10 ) [Hz]} 6 0.113 [ms] [Reference] Q.101 274 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.101 What settings should be made in order to enable use of timer 4 as an interval timer? A.101 The following description uses the V850E/MS1 as an example. When the internal system clock is 25 MHz, the timer control register (TMC40) and compare register (CM40) values are set so that the interval time becomes approximately 0.1 ms. 1. Select internal count clock and intermediate clock and set the timer control register (TMC40) value. Example: Internal count clock m/32 Intermediate clock /8 7 TMC40 1 6 0 5 0 4 0 3 0 2 1 1 1 0 0 [Overview of timer control register (TMC40)] 7 TMC40 CE40 6 0 5 0 4 0 3 0 2 1 0 PRS400 PRM401 PRM400 Bit position Bit name 7 CE40 Controls timer operation 0: No operation 1: Count operation 2 PRS400 Meaning Internal count clock ( m: intermediate clock) 0: m/16 1: m/32 1, 0 PRM401, Intermediate clock m ( : internal system clock) PRM400 PRM401 0 0 1 1 PRM400 0 1 0 1 m /2 /4 /8 RFU (reserved) User's Manual U13773EJ3V0UM 275 APPENDIX B Q & A 2. Set the compare register (CM40) value. The interval time is determined using the following formula. (Interval time) = (Compare register value + 1) x (Count clock cycle) 1) Determine the count clock cycle. The count clock cycle is determined using the following formula. (Count clock cycle) = 1/(Internal count clock) When the internal system clock = 25 [MHz] = 25 x 10 [Hz], the count clock cycle is as shown below. 6 1/(/8 x 1/32) = 1/{25 x 10 x 1/8 x 1/32} 6 = (8 x 32)/(25 x 10 ) 6 2) Determine the compare register value (compare register value: n). When the interval time is 0.1 [ms] = 1 x 10 [s], the compare register value is calculated as follows. -4 1 x 10 [s] = (n + 1) x {(8 x 32)/(25 x 10 ) [Hz]} -4 6 n = 1 x 10 [s] x {(25 x 10 ) [Hz]/(8 x 32)} - 1 -4 6 9 Consequently, to set the interval time as a value of approximately 0.1 ms, 9 should be set as the CM40 register value. Since the interval time is 0.1024 ms when CM40 = 9, the value to be expected from the calculation is approximately 0.1 ms. 276 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.102 The error message "fa01 (F): PC location's line information not found" appears. Why? A.102 When the Source Window appears after a program has been stopped, it displays the source file that corresponds to the program counter (PC) value at the time the program was stopped. This error message indicates that the source file corresponding to that program counter (PC) value cannot be found. The following causes are possible. (1) Source file exists somewhere outside of the source path. (2) The program was stopped at a location where there is no corresponding source file, such as at a library or real-time OS system call. (3) The program had entered a runaway condition, and the execution had stopped after branching to an address not used by the program. (4) No build was done in debug mode, so the debug information is not included in the object. If the cause is (1) above, the source path is the directory where the load module files being downloaded are stored by default. If the source file is located in any other directory, go to the [Option] menu and select [Debugger options] to specify the source path. If the cause is (2) above, and if display of this message cannot be avoided due to the structure of the program, use the Console Window to turn off display of error dialog boxes. [Reference] Q.114 Q.103 During linkage, the message "multiple inclusion of same file attempted, ignored." appears. Why? A.103 This message is displayed when the source for an object that is registered as a startup file has been registered as the source to be assembled. In other words, start.o has been registered as the startup file and its source file, start.s, is to be registered as the object to be assembled (for the RX850 Pro, these files are boot.o instead of start.o and boot.s instead of start.s). In the dialog box used to set up projects for PM, delete start.s from the "Source file name" list and then use the Linker Options Setup dialog box to change start.s to start.o as the "Startup file" setting. Also, before creating objects for ROM programming, the same phenomenon occurs when creating code that is used to reserve ROM programming space. As with start.o files, the method for fixing this is to build without registering the object as a source file. This type of operation is especially needed for applications that use a real-time OS, when it is essential to use a particular startup module. User's Manual U13773EJ3V0UM 277 APPENDIX B Q & A Q.104 During linkage, the error message "undefined: '__e_sysfnc' referenced in 'c:\nectools32\lib850e_ghs\r32\rxcore.o'" appears. Why? A.104 This message was displayed because svc.o, which is used to assemble svc.s, was not linked. This file contains a system call table that was created by the configurator cf850pro. This table's start address is __e_sysfunc. Note with caution that this file is easily overlooked, since it is not required by the RX850 V3.1x versions. System calls to be used must be specified in a CF definition file. For description of the specification method and configurator startup method, see 6.6 Specification Format for SCT Information and CHAPTERS 7 and 8 OPERATING CONFIGURATOR in the RX850 Pro Installation User's Manual. Q.105 The message "Warning:address is too long" appears after using PM plus to build when hx850 is executed. Why? A.105 This is due to a limitation of the Intel hexadecimal format. This message is always displayed when the Intel hexadecimal format has been specified and the program to be run extends over 1 MB of space. During the actual development, this should be considered when converting load modules to hex format, but otherwise this message can be ignored. This message will not appear if Motorola hexadecimal format is selected instead of Intel hexadecimal format. 278 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.106 Why has the system entered a runaway condition? A.106 The following causes are possible. * Are the task stack area and interrupt stack area (system stack area) large enough? Insufficient stack areas account for most causes of poor operation among applications that use a real-time OS. Specifically, when a task stack or interrupt stack exceeds its specified size, other task's stacks can be damaged, or the RX850 Pro's system management area can be damaged, and in either case a runaway condition may result. Note also that, with the RX850 Pro, incorrect values included in addresses are not questioned in order to ensure reliable operation of the address information. * Is the indirectly activated interrupt handler's termination processing coded correctly? If the indirectly activated interrupt handler's termination processing, which means the setup for the return value, is not correct, processing after an interrupt may become unreliable. This is not always the case: sometimes processing happens to continue correctly. When returning from the indirectly activated interrupt handler, be sure to specify "TSK_NULL" as an argument, such as "return (TSK_NULL);". Or, if another task should be woken up after returning from the indirectly activated interrupt handler, specify the task ID of the target task as an argument, such as "return (TASK_ID);". * If the task does not have an endless loop specification (for ( ; ; ), while (1)), is an ext_tsk system call being issued to perform the termination processing? This is not a problem if the task has an endless loop specification, but if it does not then the task must be terminated by issuing the ext_tsk system call at the end. When a system call is not specified for a task with no endless loop specification, then the RX850 Pro is unable to determine whether or not the task has been terminated. This will definitely result in a runaway condition. Q.107 The boot processing itself works well, and control is passed to the RX850 Pro after boot processing is completed, but afterward no tasks can be performed, and the system gets stuck in HALT mode. Why? A.107 This can happen when problems occur in the RX850 Pro's initialization program. When control is passed to the RX850 Pro after boot processing, other initialization tasks are started, such as the creation of a system base table (SBT). At that point, various management tables and memory pools are created, but the system may be set to HALT mode if a failure occurs when attempting to make them. Causes for such failures include not having a reserved system memory pool (SPOL0) available for creating the SBT and management tables, which can occur when memory space is either write-protected or insufficient. User's Manual U13773EJ3V0UM 279 APPENDIX B Q & A Q.108 Why doesn't the expected operation occur after issuing a system call? A.108 The following causes are possible. * Causes related to task priority When a wait mode is cleared and a system call is issued, if the priority of the task released from wait mode is lower than that of other tasks, the released task may not be performed immediately as expected. * Causes related to stack damage If wait-related system calls such as wai_flg and wai_sem are issued but problems occur when their wait modes are cleared, this can result in damage to task stack or interrupt stack areas. If either task stack or interrupt stack areas grow beyond their set values, they can damage the RX850 Pro management area and the wait information it contains. Q.109 The return value for a system call was "-17". What does this error value mean? A.109 In this case, E_NOSPT is indicated as the type of error. It means that the system call that was issued has not been registered in the system call table. Q.110 I understand about linking the RX850 Pro's nucleus object rxcore.o (rxtmcore.o), but I have not been able to register the object to the file list using PM plus. How can I link this file? A.110 When specifying an object to be linked using PM plus, go to the "Other" part of the Linker Options Setup dialog box and enter the object name where it says "Other option (Y)". For safety's sake, also enter the file path. For example, to link rxcore.o, specify the following (when using rxtmcore.o, specify "c:nectools32\lib850e\r32\rxtmcore.o"). c:\nectools32\lib850e\r32\rxcore.o Use a space to separate multiple objects. Do not use a ": (colon)" or "; (semicolon)" as a separator. 280 User's Manual U13773EJ3V0UM APPENDIX B Q & A Q.111 What is the difference between nucleus objects rxcore.o and rxtmcore.o? A.111 They differ in the way they acknowledge interrupts during processing of timer interrupts. Object Name rxcore.o rxtmcore.o Description Nucleus kernel that can acknowledge interrupts at all interrupt levels during cyclically activated handler processing Nucleus kernel that can acknowledge only interrupts with a higher priority than timer interrupts during cyclically activated handler processing The cyclically activated handler is called from the timer handler. With rxcore.o, interrupt termination processing (reti) occurs once during execution of the timer handler. Therefore, the interrupts at all interrupt levels can be acknowledged during processing of the cyclically activated handler. However, with rxtmcore.o, the cyclically activated handler is called during execution of the timer handler, so only interrupts with a higher priority than timer interrupts can be acknowledged. Q.112 In the RX850 Ver. 3.1, the .sit section had to be placed within 32 KB of address 0. Is this also the case in the RX850 Pro? A.112 This restriction does not exist in the RX850 Pro. There are no restrictions on placement locations. Q.113 How can I program ROM or FLASH ROM using only the RX850 Pro? A.113 One method is to first program the RX850 Pro software, then make necessary revisions in user applications only. This is enabled by separating ROM-programmed sections. The sections available for ROM programming in the RX850 Pro are listed below. * .system * .system_int * .system_cmn Software that can be placed in the .system section includes the following. * Routines commonly used by RX in rxcore.o * svc.o (system call table) * System call main programs (cretsk.o, etc.) * Routines commonly used with system calls (f_memget.o, etc.) User's Manual U13773EJ3V0UM 281 APPENDIX B Q & A Software that can be placed in the .system_int section includes the following. * Interrupt processing programs within rxcore.o Software that can be placed in the .system_cmn section includes the following. * Scheduler processing programs within rxcore.o However, the system call main programs that may be used in the future even if they are currently not used must also be placed. For information on which objects are actually used from among those in the libraries (librxp.a and librxpm.a), use the linker options to specify the link map output (NEC Electronics version: -m, GHS version: -map) and acquire the link information, then refer to the object names that are output. Once the above countermeasures are performed, the ROM-programmed area in the RX850 Pro is no longer affected by changes in user programs. The interface library and .sit section should be linked to the user application side. The user application side's boot sections (boot.s and boot.850) includes the following code. mov #__rx_start, lp jmp [lp] The __rx_start symbol is on the ROM-programmed RX850 Pro's side, so a jmp instruction is required after setting the actual address as #__rx_start. The following shows examples of link directives for these sections (NEC Electronics version). [Example 1] Method for setting separate output sections for objects in libraries SYSTEM : !LOAD ?RX { .system_svc = $PROGBITS ?AX .system { svc.o }; .system_core = $PROGBITS ?AX .system { ..\..\..\lib850e\r32\rxcore.o }; .os_lib1 = $PROGBITS ?AX .system {udfsys.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib2 = $PROGBITS ?AX .system {relblk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib3 = $PROGBITS ?AX .system {getblk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib4 = $PROGBITS ?AX .system {gettim.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib5 = $PROGBITS ?AX .system {sndmsg.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib6 = $PROGBITS ?AX .system {rcvmsg.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib7 = $PROGBITS ?AX .system {sigsem.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib8 = $PROGBITS ?AX .system {waisem.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib9 = $PROGBITS ?AX .system {setflg.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib10 = $PROGBITS ?AX .system {waiflg.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib11 = $PROGBITS ?AX .system {wuptsk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib12 = $PROGBITS ?AX .system {statsk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib13 = $PROGBITS ?AX .system {exdtsk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib14 = $PROGBITS ?AX .system {exttsk.o(c:\nectools32\lib850e\r32\librxp.a)}; .os_lib15 = $PROGBITS ?AX .system {cretsk.o(c:\nectools32\lib850e\r32\librxp.a)}; .system_cmn = $PROGBITS ?AX .system_cmn; .system_int = $PROGBITS ?AX .system_int; }; TEXT : !LOAD ?RX { .text = $PROGBITS ?AX .text; }; 282 User's Manual U13773EJ3V0UM APPENDIX B Q & A [Example 2] Method for setting one section for all objects in libraries SYSTEM : !LOAD ?RX { .system_svc = $PROGBITS ?AX .system { svc.o }; .system_core = $PROGBITS ?AX .system { ..\..\..\lib850e\r32\rxcore.o }; .system = $PROGBITS ?AX .system { svc.o ..\..\..\lib850e\r32\rxcore.o udfsys.o(c:\nectools32\lib850e\r32\librxp.a) relblk.o(c:\nectools32\lib850e\r32\librxp.a) getblk.o(c:\nectools32\lib850e\r32\librxp.a) gettim.o(c:\nectools32\lib850e\r32\librxp.a) sndmsg.o(c:\nectools32\lib850e\r32\librxp.a) rcvmsg.o(c:\nectools32\lib850e\r32\librxp.a) sigsem.o(c:\nectools32\lib850e\r32\librxp.a) waisem.o(c:\nectools32\lib850e\r32\librxp.a) setflg.o(c:\nectools32\lib850e\r32\librxp.a) waiflg.o(c:\nectools32\lib850e\r32\librxp.a) wuptsk.o(c:\nectools32\lib850e\r32\librxp.a) statsk.o(c:\nectools32\lib850e\r32\librxp.a) exdtsk.o(c:\nectools32\lib850e\r32\librxp.a) exttsk.o(c:\nectools32\lib850e\r32\librxp.a) cretsk.o(c:\nectools32\lib850e\r32\librxp.a) }; .system_cmn = $PROGBITS ?AX .system_cmn; .system_int = $PROGBITS ?AX .system_int; }; TEXT : !LOAD ?RX { .text = $PROGBITS ?AX .text; }; Both of the above examples are just general examples. During linkage, be sure to add an option to reference librxp.a (or librxpm.a). Q.114 The external RAM that is implemented in our target system is accessible in both 16-bit and 32-bit units but cannot be accessed in 8-bit (1-byte) units. Can the RX850 Pro be used with this system? A.114 The SPOL0 area assigned to the RX850 Pro's management area requires 8-bit (1-byte) accessibility for data access. In other words, in cases where 8-bit access is not enabled in the external RAM area, the SPOL0 area cannot be placed in that area. If it is placed there, data being operated on would be lost and normal operations could not be guaranteed. However, since there is no 8-bit access to stacks or memory pools, there would be no problem using the SPOL1 area. Nevertheless, any code that accesses data in memory blocks created within the SPOL1 area would have to have 8-bit access suppressed, such as via a compile option. This type of problem can be resolved by placing the SPOL0 area within the internal RAM of any V850 Series device. User's Manual U13773EJ3V0UM 283 APPENDIX C INDEX [A] act_cyc ............................................................. 77, 204 def_svc .................................................................. 211 del_flg .............................................................. 44, 143 del_mbx ........................................................... 49, 159 del_mpl ............................................................ 67, 187 del_sem ........................................................... 38, 133 Basic clock cycle ...................................................... 72 Boot processing ....................................................... 90 del_tsk ............................................................. 34, 106 Delayed wake-up ..................................................... 73 dly_tsk.......................................................... 73, 201 Development environment....................................... 21 Directly activated interrupt handler .............54, 55, 220 Description format....................................... 220, 223 Flow of operation.................................................. 55 Internal processing performed by the handler ...... 55 Limitation imposed on system calls...................... 56 Registering ........................................................... 55 Return processing ................................................ 57 Saving/restoring the registers............................... 55 Stack switching .................................................... 55 dis_dsp ............................................................ 86, 111 dis_int .................................................................... 176 Dispatching........................................................ 61, 86 Disabling ............................................................ 111 Resuming........................................................... 112 dly_tsk ............................................................. 73, 201 Dormant state .......................................................... 30 Drive method ........................................................... 81 [B] [C] CA850 ................................................................ 21, 93 can_wup................................................................. 128 CCV850.............................................................. 21, 93 CF850 Pro................................................................ 19 chg_icr ............................................................. 62, 179 chg_pri ................................................................... 113 Clock interrupt .................................................... 63, 72 clr_flg ............................................................... 44, 145 Communication function..................................... 28, 37 Mailbox ..................................................... 28, 37, 48 Configurator ....................................................... 18, 19 cre_flg .................................................................... 141 cre_mbx ................................................................. 156 cre_mpl .................................................................. 184 cre_sem ................................................................. 130 cre_tsk ................................................................... 103 Cross tool ................................................................. 21 Cyclically activated handler .............................. 76, 230 Acquiring cyclically activated handler information ................................................ 79, 206 Activation order..................................................... 80 Activity state ......................................................... 77 Controlling the activity state................................ 204 Description format ...................... 230, 231, 232, 233 Internal processing performed by the handler ...... 78 Interrupt ................................................................ 80 Limitation imposed on system calls ...................... 79 Registering ........................................................... 76 Registering/canceling registration....................... 202 Return processing ................................................ 79 Saving/restoring the registers ............................... 78 Stack switching..................................................... 78 [E] ena_dsp........................................................... 86, 112 ena_int................................................................... 175 Event flag..................................................... 28, 37, 43 Acquiring an ID number ..................................... 155 Acquiring event flag information ......................... 153 Checking a bit pattern ...................44, 146, 148, 150 Clearing a bit pattern.................................... 44, 145 Deleting........................................................ 44, 143 Event flag information .......................................... 45 Generating ................................................... 43, 141 Setting a bit pattern ...................................... 44, 144 Event flag wait state................................................. 31 Event-driven technique ............................................ 81 Exclusive control function .................................. 28, 37 exd_tsk .......................................................33, 34, 109 Execution environment ............................................ 20 ext_tsk ............................................................. 33, 108 Extended SVC handler .......................................... 234 [D] Data type.................................................................. 97 Debugger ................................................................. 21 def_cyc................................................................... 202 def_int .................................................................... 170 284 User's Manual U13773EJ3V0UM APPENDIX C INDEX Calling ................................................................ 213 Description format ...................... 234, 235, 236, 237 Registering/canceling registration ...................... 211 External RAM........................................................... 20 def_int .................................................................170 dis_int .................................................................176 ena_int ................................................................175 loc_cpu ...................................................61, 86, 177 ref_icr............................................................62, 181 ret_int............................................................57, 172 ret_wup .........................................................57, 173 unl_cpu ...................................................61, 86, 178 I/O board for in-circuit emulator ................................21 [F] FCFS method .................................................... 29, 82 Forced termination ................................................... 33 frsm_tsk ................................................................. 124 [G] get_blk ............................................................. 67, 188 get_tid .................................................................... 117 get_tim ............................................................. 72, 200 get_ver................................................................... 208 [K] Keyword .................................................................215 [L] Level E .....................................................................17 loc_cpu .......................................................61, 86, 177 Lock function ............................................................86 [H] Hardware environment............................................. 21 Hardware initialization section ................................. 90 High-level language interface library........................ 18 Host machine........................................................... 21 [M] ITRON3.0 specification...........................................17 Mailbox .........................................................28, 37, 48 Acquiring an ID number ......................................168 Acquiring mailbox information.............................166 Deleting.........................................................49, 159 Generating ....................................................48, 156 Mailbox information...............................................51 Receiving a message....................50, 162, 163, 164 Sending a message ......................................49, 160 Management object ..................................................65 Typical arrangement .............................................65 Maskable interrupt ....................................................61 Disabling acknowledgement and dispatch ..........177 Resuming acknowledgement and dispatch.........178 Memory block ...........................................................66 Acquiring.......................................67, 188, 190, 191 Returning ......................................................68, 193 Memory block wait state ...........................................31 Memory pool.............................................................66 Acquiring a memory block.............67, 188, 190, 191 Acquiring an ID number ......................................197 Acquiring memory pool information...............69, 195 Deleting.........................................................67, 187 Generating ....................................................66, 184 Memory pool information ......................................69 Returning a memory block ............................68, 193 Memory pool management function ...................28, 64 Memory pool management system call ............95, 183 cre_mpl ...............................................................184 [I] Idle handler .............................................................. 88 In-circuit emulator .................................................... 21 Indirectly activated interrupt handler .......... 54, 58, 226 Description format ...................... 226, 227, 228, 229 Flow of operation.................................................. 58 Internal processing performed by the handler ...... 59 Limitation imposed on system calls ...................... 59 Registering ........................................................... 58 Registering/canceling registration ...................... 170 Return processing ................................................ 60 Saving/restoring the registers............................... 59 Stack switching..................................................... 59 Interface library .................................................. 19, 92 Positioning............................................................ 92 Processing in the library ....................................... 93 Types ................................................................... 93 Internal ROM/RAM................................................... 18 Interrupt control register........................... 62, 179, 181 Acquiring ...................................................... 62, 181 Changing...................................................... 62, 179 Interrupt handler....................................................... 54 Directly activated interrupt handler ......... 54, 55, 220 Indirectly activated interrupt handler....... 54, 58, 226 Interrupt management function .......................... 28, 54 Interrupt management system call ................... 95, 169 chg_icr.......................................................... 62, 179 User's Manual U13773EJ3V0UM 285 APPENDIX C INDEX del_mpl ......................................................... 67, 187 get_blk .......................................................... 67, 188 pget_blk ........................................................ 67, 190 ref_mpl.......................................................... 69, 195 rel_blk ........................................................... 68, 193 tget_blk ................................................... 68, 75, 191 vget_pid ........................................................ 70, 197 Message................................................................... 51 Allocating an area................................................. 51 Composition of messages .................................... 51 Priority .................................................................. 51 Message wait state .................................................. 31 Multiple interrupt....................................................... 63 Flow ...................................................................... 63 Multitask OS............................................................. 17 Multitasking ........................................................ 17, 37 Programming ......................................................... 214 Cyclically activated handler ................................ 230 Directly activated interrupt handler..........54, 55, 220 Extended SVC handler....................................... 234 Indirectly activated interrupt handler .......54, 58, 226 Task ................................................................... 216 [R] rcv_msg ........................................................... 50, 162 Ready state ............................................................. 30 Real-time OS ........................................................... 16 Real-time processing ............................................... 17 ref_cyc ............................................................. 79, 206 ref_flg............................................................... 45, 153 ref_icr............................................................... 62, 181 ref_mbx............................................................ 51, 166 ref_mpl............................................................. 69, 195 ref_sem............................................................ 40, 138 ref_sys ................................................................... 210 ref_tsk.............................................................. 35, 118 rel_blk .............................................................. 68, 193 rel_wai ................................................................... 116 Reserved word....................................................... 215 Resource wait state ................................................. 31 ret_int............................................................... 57, 172 ret_wup............................................................ 57, 173 return ....................................................................... 60 Return value ............................................................ 99 ROMization.............................................................. 18 rot_rdq ................................................................... 115 Rotating a ready queue ......................................... 115 Round-robin method ................................................ 83 rsm_tsk .................................................................. 123 Run state ................................................................. 30 RX850 Pro ............................................................... 29 [N] Non-existent state .................................................... 30 Non-maskable interrupt ............................................ 63 Normal termination................................................... 33 Nucleus .............................................................. 19, 27 Configuration ........................................................ 27 Function................................................................ 28 Nucleus initialization section .................................... 91 [O] Operating system ..................................................... 17 ITRON3.0 specification....................................... 17 Level E.................................................................. 17 OS ........................................................................ 21 [P] Parameter ................................................................ 97 Parameter value range............................................. 98 PC interface board ................................................... 21 Peripheral controller ................................................. 20 pget_blk ........................................................... 67, 190 pol_flg .............................................................. 45, 148 prcv_msg.......................................................... 50, 163 preq_sem ......................................................... 39, 136 Priority method ................................................... 29, 82 Processing program ............................................... 214 Cyclically activated handler .......................... 76, 230 Directly activated interrupt handler ......... 54, 55, 220 Extended SVC handler ....................................... 234 Indirectly activated interrupt handler....... 54, 58, 226 Task.................................................................... 216 [S] Sample source file ................................................... 20 Boot processing ................................................... 90 Hardware initialization section.............................. 90 Software initialization section ............................... 91 System initialization.............................................. 89 Scheduler .......................................................... 29, 81 Drive method........................................................ 81 Lock function .................................................. 17, 86 Scheduling method ........................................ 82, 88 Semaphore ........................................................ 28, 37 Acquiring a resource .....................39, 135, 136, 137 286 User's Manual U13773EJ3V0UM APPENDIX C INDEX Acquiring an ID number...................................... 140 Acquiring semaphore information....................... 138 Deleting ........................................................ 38, 133 Generating ................................................... 38, 130 Returning a resource.................................... 38, 134 Semaphore information ........................................ 40 set_flg .............................................................. 44, 144 set_tim ............................................................. 72, 199 sig_sem ........................................................... 38, 134 slp_tsk.................................................................... 125 snd_msg .......................................................... 49, 160 Software environment .............................................. 21 Software initialization section................................... 91 Software timer.......................................................... 72 sta_tsk ............................................................. 33, 107 sus_tsk................................................................... 122 Suspend state.......................................................... 31 Synchronization function.......................................... 37 Event flag ................................................. 28, 37, 43 Semaphore..................................................... 28, 37 Synchronous communication function ............... 28, 37 Synchronous communication system call ........ 94, 129 clr_flg............................................................ 44, 145 cre_flg ................................................................ 141 cre_mbx ............................................................. 156 cre_sem ............................................................. 130 del_flg........................................................... 44, 143 del_mbx........................................................ 49, 159 del_sem........................................................ 38, 133 pol_flg........................................................... 45, 148 prcv_msg...................................................... 50, 163 preq_sem ..................................................... 39, 136 rcv_msg........................................................ 50, 162 ref_flg ........................................................... 45, 153 ref_mbx ........................................................ 51, 166 ref_sem ........................................................ 40, 138 set_flg........................................................... 44, 144 sig_sem........................................................ 38, 134 snd_msg....................................................... 49, 160 trcv_msg................................................. 50, 75, 164 twai_flg ................................................... 45, 74, 150 twai_sem ................................................ 39, 74, 137 vget_fid......................................................... 46, 155 vget_mid....................................................... 52, 168 vget_sid........................................................ 40, 140 wai_flg .......................................................... 44, 146 wai_sem ....................................................... 39, 135 System call .............................................................. 94 act_cyc..........................................................77, 204 Calling...................................................................96 can_wup .............................................................128 chg_icr ..........................................................62, 179 chg_pri ................................................................113 clr_flg ............................................................44, 145 cre_flg .................................................................141 cre_mbx ..............................................................156 cre_mpl ...............................................................184 cre_sem ..............................................................130 cre_tsk ................................................................103 def_cyc ...............................................................202 def_int .................................................................170 def_svc ...............................................................211 del_flg ...........................................................44, 143 del_mbx ........................................................49, 159 del_mpl .........................................................67, 187 del_sem ........................................................38, 133 del_tsk ..........................................................34, 106 dis_dsp .........................................................86, 111 dis_int .................................................................176 dly_tsk...........................................................73, 201 ena_dsp ........................................................86, 112 ena_int ................................................................175 exd_tsk ...................................................33, 34, 109 ext_tsk ..........................................................33, 108 Extension ..............................................................99 frsm_tsk ..............................................................124 Function code .......................................................96 get_blk ..........................................................67, 188 get_tid .................................................................117 get_tim ..........................................................72, 200 get_ver................................................................208 loc_cpu ...................................................61, 86, 177 Parameter .......................................................97, 98 pget_blk ........................................................67, 190 pol_flg ...........................................................45, 148 prcv_msg ......................................................50, 163 preq_sem......................................................39, 136 rcv_msg ........................................................50, 162 ref_cyc ..........................................................79, 206 ref_flg............................................................45, 153 ref_icr............................................................62, 181 ref_mbx.........................................................51, 166 ref_mpl..........................................................69, 195 ref_sem.........................................................40, 138 ref_sys ................................................................210 ref_tsk ...........................................................35, 118 User's Manual U13773EJ3V0UM 287 APPENDIX C INDEX rel_blk ........................................................... 68, 193 rel_wai ................................................................ 116 ret_int............................................................ 57, 172 ret_wup......................................................... 57, 173 Return value ......................................................... 99 rot_rdq ................................................................ 115 rsm_tsk ............................................................... 123 set_flg ........................................................... 44, 144 set_tim .......................................................... 72, 199 sig_sem ........................................................ 38, 134 slp_tsk ................................................................ 125 snd_msg ....................................................... 49, 160 sta_tsk .......................................................... 33, 107 sus_tsk ............................................................... 122 ter_tsk........................................................... 33, 110 tget_blk ................................................... 68, 75, 191 trcv_msg ................................................. 50, 75, 164 tslp_tsk ......................................................... 74, 126 twai_flg ................................................... 45, 74, 150 twai_sem ................................................ 39, 74, 137 unl_cpu ................................................... 61, 86, 178 vget_fid ......................................................... 46, 155 vget_mid ....................................................... 52, 168 vget_pid ........................................................ 70, 197 vget_sid ........................................................ 40, 140 vget_tid ......................................................... 36, 120 viss_svc .............................................................. 213 wai_flg .......................................................... 44, 146 wai_sem ....................................................... 39, 135 wup_tsk .............................................................. 127 System clock ............................................................ 72 Setting and reading .............................................. 72 System construction procedure................................ 22 System initialization.................................................. 89 Boot processing.................................................... 90 Flow ...................................................................... 89 Hardware initialization section .............................. 90 Nucleus initialization section................................. 91 Sample source file ................................................ 20 Software initialization section................................ 91 System management system call..................... 95, 207 def_svc ............................................................... 211 get_ver................................................................ 208 ref_sys ................................................................ 210 viss_svc .............................................................. 213 System performance analyzer.................................. 21 AZ850 ................................................................... 21 [T] Task................................................................. 17, 216 Acquiring an ID number ..................................... 117 Acquiring task information............................ 35, 118 Activating ............................................................. 33 Canceling a suspend request..................... 123, 124 Canceling a wake-up request..............125, 126, 128 Changing the priority .......................................... 113 Delayed wake-up ................................................. 73 Deleting................................................................ 34 Description format .......................216, 217, 218, 219 Generating ................................................... 33, 103 Internal processing of task ................................... 34 Issuing a suspend request ................................. 122 Issuing a wake-up request ................................. 127 Limitation imposed on system calls...................... 34 Releasing from the wait state ............................. 116 Rotating a ready queue...................................... 115 Saving/restoring the registers............................... 34 Stack switching .................................................... 34 State transition ..................................................... 30 Task context......................................................... 30 Task information................................................... 35 Terminating .......................................................... 33 Timeout wait....................................................... 201 Task debugger......................................................... 21 Task management function ............................... 28, 30 Task management system call ........................ 94, 102 chg_pri ............................................................... 113 cre_tsk ............................................................... 103 del_tsk.......................................................... 34, 106 dis_dsp......................................................... 86, 111 ena_dsp ....................................................... 86, 112 exd_tsk....................................................33, 34, 109 ext_tsk.......................................................... 33, 108 get_tid ................................................................ 117 ref_tsk .......................................................... 35, 118 rel_wai................................................................ 116 rot_rdq................................................................ 115 sta_tsk.......................................................... 33, 107 ter_tsk .......................................................... 33, 110 vget_tid ........................................................ 36, 120 Task-associated synchronization system call .. 94, 121 can_wup............................................................. 128 frsm_tsk ............................................................. 124 rsm_tsk .............................................................. 123 slp_tsk................................................................ 125 sus_tsk............................................................... 122 288 User's Manual U13773EJ3V0UM APPENDIX C INDEX tslp_tsk ......................................................... 74, 126 wup_tsk .............................................................. 127 ter_tsk .............................................................. 33, 110 tget_blk ...................................................... 68, 75, 191 Time management function ............................... 29, 72 Time management system call ........................ 95, 198 act_cyc ......................................................... 77, 204 def_cyc ............................................................... 202 dly_tsk .......................................................... 73, 201 get_tim ......................................................... 72, 200 ref_cyc.......................................................... 79, 206 set_tim.......................................................... 72, 199 Timeout.................................................................... 74 tget_blk................................................... 68, 75, 191 trcv_msg................................................. 50, 75, 164 tslp_tsk ......................................................... 74, 126 twai_flg ................................................... 45, 74, 150 twai_sem ................................................ 39, 74, 137 Timeout wait state.................................................... 31 Changes............................................................. 201 Timer operation........................................................ 72 trcv_msg .................................................... 50, 75, 164 tslp_tsk............................................................. 74, 126 twai_flg....................................................... 45, 74, 150 twai_sem.................................................... 39, 74, 137 [U] unl_cpu.......................................................61, 86, 178 Utility.........................................................................18 Configurator ....................................................18, 19 High-level language interface library.....................18 [V] Version information acquiring .................................208 vget_fid.............................................................46, 155 vget_mid...........................................................52, 168 vget_pid............................................................70, 197 vget_sid ............................................................40, 140 vget_tid.............................................................36, 120 viss_svc..................................................................213 [W] wai_flg ..............................................................44, 146 wai_sem ...........................................................39, 135 Wait function.......................................................28, 37 Event flag..................................................28, 37, 43 Wait state .................................................................31 Forcibly release...................................................116 Wait-suspend state...................................................31 Wake-up wait state ...................................................31 wup_tsk ..................................................................127 User's Manual U13773EJ3V0UM 289 APPENDIX D REVISION HISTORY A history of the revisions up to this edition is shown below. "Applied to" indicates the chapters to which the revision was applied. Edition 2nd Contents Modification of return value E_NOSPT from -11 to -17 Modification of description on target CPU Modification of description on hardware and software environments Addition of explanation on system construction procedure Addition of a caution on return processing from directly activated interrupt handler Addition of a caution on changing/acquiring interrupt control register Modification of description on overview of memory management function Modification of description on management objects Modification of description on memory pool and memory blocks Addition of a caution on returning a memory block Addition of explanation on interrupt during cyclically activated handler Addition of explanation on activation order of cyclically activated handler Addition of explanation on idle handler Addition of explanation on system initialization processing Addition of explanation on directly activated interrupt handler Addition of explanation on indirectly activated interrupt handler 3rd Modification of V850 Family to V850 Series Addition of operating target CPU Modification of description on peripheral controller Addition of target device for in-circuit emulator Addition of I/O board for in-circuit emulator and target device Modification of description on software environment OS and addition of debugger Modification of description of caution on returning a memory block CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION CHAPTER 7 TIME MANAGEMENT FUNCTION CHAPTER 11 SYSTEM CALLS CHAPTER 7 TIME MANAGEMENT FUNCTION CHAPTER 6 MEMORY POOL MANAGEMENT FUNCTION CHAPTER 5 INTERRUPT MANAGEMENT FUNCTION Throughout CHAPTER 1 OVERVIEW Applied to: CHAPTER 8 SCHEDULER CHAPTER 9 SYSTEM INITIALIZATION APPENDIX A PROGRAMMING METHODS Throughout CHAPTER 1 OVERVIEW Modification of description on interrupts in cyclically activated handler Modification of caution in description of rel_blk to match caution on returning a memory block Modification of systime to t_systime in description of structure of system clock SYSTIME for set_tim and get_tim Addition of APPENDIX B Q & A APPENDIX B Q & A 290 User's Manual U13773EJ3V0UM |
Price & Availability of RX703100
 |
|
|
|
|
All Rights Reserved © IC-ON-LINE 2003 - 2022 |
| [Add Bookmark] [Contact Us] [Link exchange] [Privacy policy] |
|
Mirror Sites : [www.datasheet.hk]
[www.maxim4u.com] [www.ic-on-line.cn]
[www.ic-on-line.com] [www.ic-on-line.net]
[www.alldatasheet.com.cn]
[www.gdcy.com]
[www.gdcy.net] |