 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : How to intentionally crash MMBASIC on the PM....
Page 1 of 2   |
|||||
| Author | Message | ||||
Grogster Admin Group  Joined: 31/12/2012 Location: New ZealandPosts: 9063 |
This is an interesting one, and probably not a request that is seen very often - if at all. In my other thread about the HW watchdog, I am looking to use that, to get around a problem I have with some PM boards, where MMBASIC crashes for some reason I have yet to discover. This is a hard-crash. The interpreter is totally done for, and you cannot talk to the module's VCP - AT ALL. There simply IS no VCP to connect to, to try to CTRL-C in and play about with the code. You can only ever get things going again, by pressing the RESET button on the module, or cycling the power. I now want to CREATE that situation, so I can test if the HW watchdog on the 2040 chip, can recover the situation. The software watchdog cannot, cos the MMBASIC interpreter has crashed, so it no longer resets the system to get things going again.  So, I need a way to actually hard-crash MMBASIC on the PM module, to see if the HW watchdog gets me around this issue - which I am hopeful it will.  Can I use POKE and a memory location or something, for force a hard-crash to test this idea? As I said - this is probably not a request that you will hear of very often. Most people WANT the interpreter to stay alive at all times!  Smoke makes things work. When the smoke gets out, it stops! |
||||
disco4now Guru  Joined: 18/12/2014 Location: AustraliaPosts: 844 |
Call the CSUB GONE should do it. Remove the WatchDog HW and I think you need a hard reset to get out of it. Print MM.Info(BOOT COUNT) Print MM.Info(AUTORUN) WatchDog HW 8000 GONE Print " I did not DIE" 'File LOG.bas written 08-03-2024 13:23:49 v1.44 CSub GONE 00000000 'main AF00B580 0000E7FE End CSub C code void main() { while (1){ } } Edited 2024-03-08 12:32 by disco4now Latest F4 Latest H7 |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9063 |
EXCELLENT!!!  Thanks. I will try that with b6 over the weekend. Smoke makes things work. When the smoke gets out, it stops! |
||||
LeoNicolas Guru Joined: 07/10/2020 Location: CanadaPosts: 444 |
Can you just use an infinite function recursion to force a stack overflow? |
||||
| JohnS Guru Joined: 18/11/2011 Location: United KingdomPosts: 3659 |
May well be detected by MMBasic with an error and back to the command prompt. John |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9063 |
Yes, I need something that can FORCE a total MMBASIC hard-crash. Assume when that happens, that MMBASIC no longer exists. The only way to recover, is a hard-reset via a tact-button, or cycling the power. VERY keen to try out disco4now's crash-code during the weekend. EDIT: Yes, the timeout is fixed. Times greater then 201ms now work in this beta. I will now try the crash-code to see what happens, but I am hopeful it will recover fine. EDIT: disco4now's code hard-crashes MMBASIC beautifully - no CTRL-C access, no nothing!  The HW watchdog then DOES work as I hoped, and it kicks MMBASIC back into life. I expect this will fix my issue completely. Hopefully, one day, I will be able to find out WHAT causes the PM to hard-crash like this, but this effectively becomes my "Get out of jail free" card - for now. Edited 2024-03-09 10:13 by Grogster Smoke makes things work. When the smoke gets out, it stops! |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9063 |
ADDITIONAL: Well, the crash-code DOES work, but then, so does the standard software watchdog if I set that before calling the GONE routine, so perhaps still not quite the hard-crash that I need. I expect this is simply cos the software watchdog must still be running in the background, even if you call a CSUB that never returns, so it then kicks the system to bring it back to life. Progress, but as the software watchdog still works during a call to a CSUB, I don't think that the MMBASIC interpreter is actually hard-crashed - it is just stuck in a C routine that never returns to MMBASIC, so can anyone think of something else I can try to totally kill MMBASIC? It is that aspect that I need to test, cos that is what is happening. I currently have software watchdog commands in the code, but as the PM is locking up so bad, the software watchdog never comes into effect. That's what I need to simulate.  Smoke makes things work. When the smoke gets out, it stops! |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 8582 |
The club needs to disable interrupts to do what you need |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 8582 |
The club needs to disable interrupts to do what you need |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 8582 |
The club needs to disable interrupts to do what you need |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 8582 |
The csub needs to disable interrupts to do what you need Edited 2024-03-09 17:31 by matherp |
||||
Bryan1 Guru Joined: 22/02/2006 Location: AustraliaPosts: 1209 |
ah stuck in a never ending loop mate just aswell you weren't in windows as it would be forever |
||||
Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 5727 |
Peter finally managed to use his club to disable the interrupts by the looks of things. :) Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
| zeitfest Guru Joined: 31/07/2019 Location: AustraliaPosts: 381 |
Using an illegal address will force a reset in many micros. But many microcontrollers will automatically reset and then restart after a serious error condition, it is a deliberate design strategy and one of the differences between microcontrollers and microprocessors. Don't know about PM / STM. It sound like you need to watch it with a debugger as it runs, catch it and see what the status registers tell. |
||||
| LouisG Senior Member  Joined: 19/03/2016 Location: AustraliaPosts: 121 |
As a humble alternative, have you considered an external transistor pulling down the reset pin upon loss of, say, a square wave from one of the Picomite's pins? Might get away with just a diode and R-C circuit? Can't get a harder hardware reset than that. |
||||
| Mixtel90 Guru Joined: 05/10/2019 Location: United KingdomPosts: 5727 |
555 as a watchdog chip. Cheap and nasty but they work. :) Or - and prettier - an ATTiny as a watchdog. :) Mick Zilog Inside! nascom.info for Nascom & Gemini Preliminary MMBasic docs & my PCB designs |
||||
Geoffg Guru Joined: 06/06/2011 Location: AustraliaPosts: 3165 |
I think that Grogster is looking for a way to totally crash the CPU, not reset/restart it. So far disco4now's CSUB GONE with interrupts disabled sounds the best. Geoff Graham - http://geoffg.net |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 8582 |
Sorry about the multiple posts, bloody phones. Press save, nothing happens, repeat etc. then they all post  back to the watchdog issue now I'm on a proper computer: The S/W watchdog will work in all circumstances except when the H/W timer interrupt has been disabled. The H/W watchdog will work in ALL circumstances. The test CSub did not disable interrupts so the S/W watchdog still worked. To test the H/W watchdog the CSUB needs to disable all interrupts. I can't think of a simple way of doing this from a CSub because you need to link the RP2040 sdk to get the register definition but I can guarantee the H/W watchdog WILL WORK in all circumstances |
||||
| Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9063 |
Excellent, thanks chums. It was a bit of an odd request. I mean - who WANTS to intentionally crash the FW running on an MCU? (rhetorical!) I will simply adopt the HW watchdog now that I can set a time greater then 201ms. BTW: @ Peter - Do you know what was going on there? (the 201ms issue) It seems to be corrected in b6, so it's all good, I am just curious. Also @ Peter - do you want me to delete the duplicate posts via your phone, or are you happy that I leave them there? It gave us all a chuckle, so I think they should stay, but I can delete them on your request. Smoke makes things work. When the smoke gets out, it stops! |
||||
| JohnS Guru Joined: 18/11/2011 Location: United KingdomPosts: 3659 |
Don't you need to carry out a test which disables interrupts - i.e. GONE with something like cpsid if (maybe __asm() can do that?) John |
||||
| Page 1 of 2 |
|||||
 Print this page |
