On 22.06.2012 18:04, Eric Nelson wrote:
On 06/21/2012 03:57 PM, Wolfgang Denk wrote:
Dear Eric Nelson,
In message4FCD17A5.6030705@boundarydevices.com you wrote:
The function of the watchdog should be reserved for auch systems who need it to implement security related functions; messing with it in unrelated places (like setting it into unexpected states and or timeouts) may cause undefined behaviour on such systems.
Don't do it.
My comment is that reset_cpu() should probably not return, and at the moment it will. The minimum WDT timeout value is 1/2 second, so execution will continue for some time less than that.
It seems to me that this might be a bad thing, allowing unexpected execution of commands in a script after the execution of the 'reset' command.
If I understand your comment, you're saying that reset_cpu() should not be implemented using the watchdog.
Is that right?
No, that was not my intention. I must have missed the fact that you were talking about reset_cpu() and only this function. Please ignore me.
NP. I'll submit a patch to address the fact that the reset returns.
Ping on this whole patch series ;)
Best regards
Dirk