We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
Hello,
I am dealing with a puzzeling situation: working with an embedded artists LPC1788 board, I have a program that enables the WDT but fails to restart after it resets the system (it crashes hard, disconnecting the debugger). But if I then power cycle or use the reset button - the program starts up correctly! I see no difference between WDT reset and external reset in the user manual. I have to say that it looks as if it is related to memory (?); If I do not start FreeRTOS scheduler the program will not crash, but it is still hard to explain! Any ideas?
I have to say that it looks as if it is related to memory (?)
I mean external memory, but maybe that is not true after all: my bootloader's RAM test passes.
Are there other hardware on the board that are in need of a reset? Will this other electronics get a reset if the processor performs a watchdog reset?
I just managed to find what I _think_ is the cause:
LPC_EMAC->MAC1 = EMAC_MAC1_PASS_ALL;
this code resides in the ethernet driver of lwIP. Somehow it fails _after_ watchdog reset...! I need to check the user manual and the errata sheet of the chip. I see clearly that when this code is executed, the watchdog is operational after a recovery from a watchdog reset (timeout is long) - processor crashes!
I'm not familiar with that chip, but had what sounds like a similar(-ish) situation occur on the LPC3250.
There, the external PHY was not getting a long enough reset when generated by the watchdog timer and it would lock up the whole processor on subsequent access to the MAC.
My 'cure' was to programatically generate an acceptable length reset (RESOUT on the LPC3250) just prior to accessing the MAC.
@IB Shy,
Crap, I just realized it is possibly a silicon issue..! I will download the user manual of the LPC3250 and look into. I hope something similar can be done for the LPC1788 ! Thanks.
"I will download the user manual of the LPC3250 and look into."
There was nothing in the LPC3250 manual that mentions this. I communicated with NXP support about it and they weren't (then) aware of anything.
It was a case of getting the 'scope out and measuring the signals.
When I narrowed the problem down, I told them about it. I never got a follow-up from them regarding the matter.
Anyway ... hope the info helps.
DJR
In other words - the LPC1788 suffers from the same issue, and the EA evaluation board for LPC1788 cannot recover from watchdog resets if the network interface is used. I discusses the situation with one of the hardware people; we are going to use another PHY that requires a much shorter reset pulse!