WDOG Register Address

[makotokatsukura]

Dear Forum members,

In the "Vybrid Referense Manual(Rev 5)",  "26.6 WDOG Memory Map/Register Difinition",

it says WDOG_WCR register absolute address is 0x0.

However, I can't access to the register and I think that address is odd.

Because at "3.1 System memory map", it says 0x0 to 0x7fffff is memory map for Boot ROM.

Dose anyone know the absolute address of WDOG_WCR is really 0x0?

(or I'm misunderstanding the reference manual?)

Regards,

Makoto Katsukura

[juangutierrez · Accepted solution]

Definitely the WDOG base address should not be 0u. Replace that WDOG_BASE definition with something like this:

#ifdef ARMCC_A5

#define WDOG_BASE                                (0x4003E000u)

#else /* M4 */

#define WDOG_BASE                                (0x4004F000u)

#endif

You can check Section 3.2 Peripheral Bridge 0 (AIPS-Lite 0) Memory Map in RM

[kubiznak_petr]

Hi Juan,

there is actually a typo in your code, it should be as follows (notice M4 address):

#ifdef ARMCC_A5

#define WDOG_BASE                                (0x4003E000u)

#else /* M4 */

#define WDOG_BASE                                (0x4003F000u)

#endif

Why isn't this issue fixed in MQX 4.1.0? Wasn't it reported?

[juangutierrez]

I was not aware this was also in MQX. This was corrected or at list notified for our Vybrid Sample code but not for MQX.

Are you sure this is a problem for MQX. I mean is not handled different in MQX?

[kubiznak_petr]

Yes, in mqx\source\psp\MVF50GS10MK50.h (in both MQX 4.0.2 and MQX 4.1.0), there is the following:

/* WDOG - Peripheral instance base addresses */

/** Peripheral WDOG base pointer */

#define WDOG_BASE_PTR                            ((WDOG_MemMapPtr)0u)

Besides the fact the address is absurd, there is also no conditioning for A5/M4 core. And in RM Rev.5, it is the same (see section 26.6). After the address fix, everything works as expected.

Obviously Freescale prefers SW watchdog, but I don't consider that to be a good solution, definitely not on multicore where some incautious operation of SW on one core can immediately kill the SW on second core.

[juangutierrez]

Hi Petr

Would you have a simple sample that show the WDOG functionality, that you can share, please?

[kubiznak_petr]

Hi Juan,

sorry for the long delay. Please see attached code snippet extracted from my application. It should well demonstrate WDOG functionality and workaround on Vybrid M4.

Petr

[juangutierrez]

Great!

Thanks Petr :smileyhappy:

[juangutierrez]

Ok, thanks Petr

I will be checking and reporting this to the MQX team.