 | | RE: signals / legOS internals
|
|
(...) Ummmmm, I don't want to speak for Markus, but most embedded kernels (and I use a lot of them) use the term signal and semaphore interchangably. I'm willing to bet a couple of bricks that the intent is to implement semaphores to facilitate (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Because they are useful at times, and Markus directed me to implement them. I think the upcoming network code is going to use them, too (at least, that's what I heard). Just 'cause they're there doesn't mean you have to use them. (...) How do (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) This second methed doesn't really solve anything since you still have to communicate with the compute thread from the receive thread. (...) But what can you actually do in the signal handler? Modify the state of the state of the executing (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) This is the same thing as most modern RTOS use, called various things. The terminology I use for this is FLIH/SLIH (first level interrupt handler and second level interrupt handler). Coupled with lock priority inheritance, it makes for a very (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) In OSE interrupt processes (first class processes, that is why we don't call them ISR's) use OS mechanisms (usually messages) to communicate with normal processes. The OS handles all update of shared data structures (such as message queues). A (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) An ISR can certainly modify data that is used by the executing code. For example, if it doesn't save registers that are used the interrupted code will certainly be affected. In fact, the ISR is of little use if it doesn't modify data that can (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) But hardware interrupts schedule a different context (the interrupt process or ISR) from the one already executing. A Unix-style signal asynchronously signal interrupts the executing process and jumps to a different location in the code that (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) An asynchronous signal models a hardware interrupt. There are times when that is the appropriate interface. There are times when other interfaces are less error prone and are easier to write correctly. Frequently using only synchronous (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Why signals?!? Unix-style asynchronous signals is the last feature I would ever want to implement or use. Brain-dead and very dangerous "feature". As you write; "The problem is that a signal can happen at any time". You shouldn't pass this (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Actually, gcc moves the stack pointer before using any variables, so it is always safe to use stack below the current stack pointer. If you are using a different C compiler or doing assembly tricks, then you are on your own. (...) Traditional (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | RE: signals / legOS internals
|
|
(...) Not to put too fine a point on it, and I hope Kekoa agrees, but the "fluff" is the reason why replacement firmware works in the first place. Without the ROM vectors being dispatched through RAM pointers that are initialized by the ROM and then (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Since it sounds like it might matter to you, I should point out that the interrupt situation on the RCX is a bit more complicated than the H8 manual makes it out to be. Interrupt vectors are in ROM, but the ROM does a dispatch to interrupt (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Yes, that is correct. As for passing parameters to a function, they are simply pushed onto the stack before the function call. Therefore, the H8 stack usually looks something like e.g. this: function param sp+10 function param sp+8 return (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) I may have answered my own question here. Since the SP is adjusted before anything is actually done with the local variables, the rule about it always pointing to the last valid datum is not violated even if local variables are placed after (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) I may have down and up reversed, since different machines grow the stack upward or downward, and I've worked with a few. I do have push and pop clear in my mind, though. So, yeah, down. So, by the "always" rule, nothing will ever be below the (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) When you define a bunch of local variables (more than fit in available registers), where are they? Unless you happen to know the answer, I may have to do some more experimenting. Admittedly, it has been awhile, but the last time I was heavy (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) Reading this again, I should mention that pushing things on the stack asynchronously is fine, as long as the stack pointer is always in the right place, which is why it must always always always point to the location I mentioned (and seemingly (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: signals / legOS internals
|
|
(...) The stack pointer must always always always point at the bottom of the stack data area, i.e. to the last valid data item on the stack. As a consequence, push val (or mov.w val,@-r7) never overwrites data. When an interrupt occurs, the first (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | signals / legOS internals
|
|
Peek-a-boo! I've been banging my head against this a bit and thought I'd see if any of you can shove me in the right direction. I am trying to implement signals in legOS. The problem is that a signal can happen at any time. I know how to make that (...) (25 years ago, 23-Jun-99, to lugnet.robotics.rcx.legos)
|
| |
| | Re: IR questions
|
|
(...) Only transmitting keeps the tower alive, at least according to my tests. If absolutely necessary, an occasional 0xff sent by the PC will keep the tower alive; this will result in only a start bit being sent, which ought to occupy very little (...) (25 years ago, 22-Jun-99, to lugnet.robotics.rcx.legos, lugnet.robotics)
|
