I am finding myself confused by the various descriptions of the watchdogs behaviour.
A couple items
Reset - There appears to be several ways to reset the WDT. The obvious one is the interrupt acknowledge which reloads the time from WLOAD. Presumabley this could be done even w/o the interrupt having fired. However, there seems to be some indication you can reload the running timer directly. This is rather worrisome actually since it provides a very effective way of making the watchdog ineffective by simply making the watchdog period too long to be effective.
Lockout - I am used to being able to lock out access to a watchdog except to be able periodically reset it (usually by a encoded sequence). There is a lock register but it's not clear what it affects. The TIVAware documentation says it locks out the configuaration registers (without saying which are considered configuration), the chip documentation says it locks out all registers. So which is correct? It would be nice if it was the latter since that would prevent direct reload of the timer and presumably the interrupt clear so that the reset could be part of a protected sequence. If it is configuration registers though, which registers are considered configuration and which ones are not?
The best result for me would be if the locking protected all write access and I could reset the timer at will with the interrupt clear. This would allow the reset to be a protected sequence (so accidental writes won't reset the watchdog), let me disable the interrupt (so I get a reset on double the watchdog period w/o having to provide an empty interrupt service routine) and protects the watchdog period from modification. Even better would be if I could also provide and upper limit to the watchdog reset frequency (i.e. if there were resets following each other too closely the watchdog would fires, just as it does if they are too far apart) but that is a bit much to expect.
Robert