GitHub repo for XDFs

[Gampy]

Hell I ain't in no hurry, biz definitely comes first ... Besides I wouldn't leave the dyno either just to argue with some internet asshole!

[edit]
And lets be clear, I have not asked you to do anything, there is no request to satisfy.

[ShorTuning]

Oh I know, but I'm genuinely curious now myself.

[antus]

I think it'll work, but lets proove it the developers are years ahead of what theyre releasing in development. The move to amd would have been known in advance. Maybe they just couldnt get the chips in time for the first p59s or wanted to manage risk and prove the p59 in the wild for a year before they made too many changes to the flash process. Releasing cars in volume that brick occasionally is a huge commercial failure.

[Gampy]

Uh Oh, That's scary ...

I do agree they are far smarter then I.

[ShorTuning]

Okay so I flashed a 2500 Savana VIN into the PCM over top of the 1500 Silverado tune that I had already flashed into it. The VIN stuck and here is a read back out of the file with the new VIN but with the same parameters from the prior 4.8 Silverado.

[antus]

Cool! Compatibility confirmed.

[Gampy]

Okay so I flashed a 2500 Savana VIN into the PCM over top of the 1500 Silverado tune that I had already flashed into it. The VIN stuck and here is a read back out of the file with the new VIN but with the same parameters from the prior 4.8 Silverado.

It appears I may have shot myself in the foot, HPT might well be using it's own kernel to write the VIN.

I agreed to that, I'll eat the crow!

I will still state blatantly that there is no AMD code in that OS so it is impossible for that Os to write to AMD flash.

[NSFW]

I wonder what would happen if you ran that OS on AMD hardware and it tried to store a fault code.

[Gampy]

I wonder what would happen if you ran that OS on AMD hardware and it tried to store a fault code.

It'll choke and puke!

I stand pat ... To the grave!

That is an interesting point, what would it do if it failed ... Obviously if it works it works.
Would it crash the Os or silently ignore it, it can't speak out about it, no one to talk to, can't write about it, that's what's failing!
How would one know it failed ...

I believe using PcmHammer to do the VIN change will make the difference. IIRC PcmHammer just requests the OS to make the actual change, it doesn't write it until the ignition is turned off, that is also when the Os does it's parameter and fault writes.
Thus any el cheapo elm device and PuTTY could do it with some carefully crafted commands.

I want to try this myself, I can't get PcmHammer to even upload the kernel, it's the first time I've tried PcmHammer at commit 01f9de3. I see there is more and some words like "Parameter block writes disabled", I'll pull again and try again, then debug if I need to.

[antus]

Well this was very interesting. Same results as above - I was able to flash the older OS in to an AMD P59. I was also able to update the VIN with ls1 flash free which definitely uses the factory flash routines in the bin. The change did persist over a key off. The changes did also persist over a short complete power off, but not over a longer (1 minute) power off. So the flash is NOT updated... as can be seen in the readback bin in the left hand side of the screen shot, and by doing a longer complete power off test.