Just on question regarding the behavior of WD flag and the rebootdevice method.
WD flag is used in my compny to know how many times the platform crashes during a given period. Its a kind of indicator on how our platform is reliable or not.
However in some cases users needs to reboot the platform when application parameters have changed. We are using PowerState.RebootDevice() method to perform that.
The problem is that rebooting the device by software like that overwrite the WD flag and indicates the watchdog has fire which is wrong in our scenario.
Do you observe the same behavior on your system? how to manage that? I can use BRAM to store if the reboot action is a user request and merge it with the WD registry value but to my opinion it’s a workaround and a big portion of my applpication now is focused on turnaround of undesired feature…
@ leforban - this is normal as we use the watchdog to reset the system.
@ andre.m - I handle the reset watchdog properly! Each time WD fires the last reset causes is watchdog. If the board is swicth off/on the last reset cause is not the watchdog…
@ Gus, this explain why the last reset cause is watchdog. To my opinion rebooting the device by using the watchdog lead to possible misunderstanding. This should be described in the documentation. To me and this was the case in all my previous design but I may be wrong, a watchdog that fires indicates mainly a stall situation of the platform (crash, too much time to process data…) and not a restart asked by the user.
@ leforban - the processor does not have a reset feature so using watchdog is required.
Ok so the only thing I can do is to put a variable in bram that store a reset request by user and manage that at start up.
Thanks for these info
@ leforban - yes that a good solution.