Firmware corruption

[shameem]

I was reading over at msefi - http://www.msefi.com/viewtopic.php?f=38&t=32809

So how do stock systems avoid firmware corruption due to crazy power cycling? Do they have soft on/off method or does it have something to do with the constant 12v?

[Fred]

That's strange, unless he had "AMC" on or something like that it shouldn't be possible to corrupt the flash memory. I guess a suitably bad power supply + alternator keeps producing power and battery is isolated could have forced damage by spiking somehow, but the chip is still behind the reg etc.

Anyway, check this out for a start : an3289

[shameem]

It is strange and doesnt seem like a power issue - rather feels like "windows not shut down properly" kinda issue.......
Could there be similar consequences if the processor is powered down somehow "during" boot up -

Is it possible to make the processor do a check sum verification on boot and indicate if firmware is corrupted (like flashing led or fuel pump cycling or something)

The ideal would be to have another copy of firmware stashed away at some other memory location and load that copy in case checksum verification fails......

[Fred]

Could there be similar consequences if the processor is powered down somehow "during" boot up -

Not unless it is currently writing to flash, which shouldn't happen when it's not being tuned.

Is it possible to make the processor do a check sum verification on boot and indicate if firmware is corrupted (like flashing led or fuel pump cycling or something)

Yes. I have a function header, but have not done anything with it yet. Likely it would only cover code and not tune though. Tune could be done with the external software though.

Fred.