Quote:
Originally Posted by hawhill
I'm not so sure. It reads as if it gets into RAM kernel mode only temporarily, not long enough to really stay there, and then resets.
My guess is on (in order of likeliness to me) a) empty battery, and b) broken hardware (constant resets due to non-software dependent reason).
|
There is a hardware watchdog that must be serviced by whatever has control of the machine (iROM kernel or RAM kernel).
This hardware watchdog can not be disabled or masked out.
So it may just be a timing issue of getting things loaded (between starting the replacement of the iROM kernel to when the RAM kernel takes over).
I haven't read your code, but if you issue a "get info" command as soon as you detect the iROM activity (or just keep sending it blind until the iROM loader replies) - then that will switch the iROM loader into different timing loop.
Depending on the i.MX SoC being addressed, there is either a 30 second or a 90 second window between when the iROM kernel is started and when it gives up, and shuts down the SoC.
(The slower the processor, the longer that time-out in seconds.)
That 30 to 90 second window is independent of the hardware watchdog.
During that window, the iROM code polls the OTG and serial ports for activity (and picks the first one to show activity) - then must receive at least one valid command (the "get info" will do and should be safe on all i.MX SoCs) to switch into "normal" command processing mode.