Post by thread:[PICLIST] PIC RTOS

You will find very little information concerning RTOS for PIC's mostly because the PIC architecture is ill-suited to RTOS implementation. In order to guarantee execution time for a high-priority task, an RTOS needs to be able to interrupt the execution of a lower-priority task at any time and force the processor to begin working on the high-priority task. This means that the current state of the low-priority task must be saved and later restored. Saving the state of a task means saving away the contents of all the relevant hardware registers. If you study the examples of interrupt routines in the PIC manuals, you will see that this is rather clumsy but possible for most of the registers, such as W, PCLATH, FSR, etc. What you won't see is storing away the current stack pointer. Since this represents the execution history of the interrupted task, it is an absolutely vital part of that task's state. For an interrupt routine, it is not necessary to alter the stack, since there can only be one interrupted background task, and the interrupt routine will always return to the same task. Any RTOS, however, depends on a hardware interrupt to allow it to gain control of the CPU when the time comes to switch to a high-priority task. That interrupt must be able to alter the stack, so that it returns not to the interrupted task, but to the high-priority task. If the low-priority task is ever to be resumed, then all of its processor state must be preserved somewhere. The PIC16xxx and PIC17xxx processors have absolutely no way for an RTOS to access the return stack. There is simply no hardware data path available. The PIC18xxx processors are minimally able to support an RTOS, since they have a set of instructions that are able to remove and insert items to the return stack. It is a relatively slow and inefficient mechanism, but there is one. As much as I like the PIC processors for many reasons, for a research project, I think you would find it easier going to use a less-integrated processor in which the return stack is implemented in general-purpose RAM rather than in special registers as they are in the PIC's. Don Hyde dhyde(atsign)axonn.com Senior Software Engineer Axonn Corporation http://www.axonn.com -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

Someone was asking recently about RTOS for PIC. I've just had a flyer arrive from a local UK distributor, Unique Memec, which mentions what they call the "First RTOS support for PICMicro architecture". It's called Real Time Architect and is for the 18Cxxx micros. It is developed by Realogy: http://www.realogy.com/. The same distributor is now sampling 18C658/858 CAN and 16C745/765 USB devices and they are flagging the new 16F73/74/76/77 parts as "Flash micros priced lower than OTP", so it looks like Microchip are seriously starting to migrate. Cheers, Andy. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! use .....listservKILLspam@spam@mitvma.mit.edu?body=SET%20PICList%20DIGEST

Reposted due to usual tagging error: ----- Original Message ----- From: "Andy Howard" <musicaKILLspamUKONLINE.CO.UK> To: <.....PICLISTKILLspam.....MITVMA.MIT.EDU> Sent: Monday, November 06, 2000 3:56 PM Subject: PIC RTOS > Someone was asking recently about RTOS for PIC. > > I've just had a flyer arrive from a local UK distributor, Unique Memec, > which mentions what they call the "First RTOS support for PICMicro > architecture". It's called Real Time Architect and is for the 18Cxxx micros. > It is developed by Realogy: http://www.realogy.com/. > > The same distributor is now sampling 18C658/858 CAN and 16C745/765 USB > devices and they are flagging the new 16F73/74/76/77 parts as "Flash micros > priced lower than OTP", so it looks like Microchip are seriously starting to {Quote hidden}> migrate. > > > Cheers, > > Andy. > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > use EraseMElistservspam_OUTTakeThisOuTmitvma.mit.edu?body=SET%20PICList%20DIGEST > > > > > -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! use listservspam_OUTmitvma.mit.edu?body=SET%20PICList%20DIGEST

There is another RTOS vendor on the scene. Pumpkin Software "http://www.pumpkin.com". They shared the booth with HiTech at the Embedded Systems Conference in San Jose, CA a few weeks ago. It is a cooperative multi-tasking scheme with an extremely small memory footprint. RTOS calls can only be made from the top level of tasks (no RTOS calls from functions called by tasks). It's pretty clever, I believe that the RTOS calls are really macros that "return" to the RTOS after saving some context info (the PC?). I have no direct experience with the software other than looking at their demo on the CD they were giving away, but it looks interesting. They claim to do all the right stuff - wake up on flags, wake up on semaphores, wake up after delay, task priority, rotate tasks with the same priority, etc. Regards, Jim Ham At 07:56 AM 11/6/2000 , you wrote: {Quote hidden}>Someone was asking recently about RTOS for PIC. > >I've just had a flyer arrive from a local UK distributor, Unique Memec, >which mentions what they call the "First RTOS support for PICMicro >architecture". It's called Real Time Architect and is for the 18Cxxx micros. >It is developed by Realogy: http://www.realogy.com/. > >The same distributor is now sampling 18C658/858 CAN and 16C745/765 USB >devices and they are flagging the new 16F73/74/76/77 parts as "Flash micros >priced lower than OTP", so it looks like Microchip are seriously starting to >migrate. > > >Cheers, > >Andy. > >-- >http://www.piclist.com#nomail Going offline? Don't AutoReply us! >use @spam@listservKILLspammitvma.mit.edu?body=SET%20PICList%20DIGEST > > Jim Ham, Porcine Associates (650)326-2669 fax(650)326-1071 "http://www.porcine.com" -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! use KILLspamlistservKILLspammitvma.mit.edu?body=SET%20PICList%20DIGEST