Firmware b208f8c7ce1bc3279b8bb89a3ec47c11e87c09bf onwards kernel corruption for kernels > 512KB with kernel_old=1 and arm_64bit=1

[petemoore]

From firmware revision b208f8c onwards, with kernel_old=1 and arm_64bit=1, kernels are getting loaded at physical address 0 as required, but are also getting overlaid at 0x80000. Control passes to the ARM and execution begins at address 0x0, but if the kernel is larger than 512KB it is corrupted due to the second copy which is overlaid into RAM at 0x80000.

I ran a full git bisection of the firmware, to determine that this is the first commit where the issue happens, and can confirm it is still happening on the current tip of the master branch.

[petemoore]

I've created https://github.com/petemoore/issue1979 as a full working demonstration of the issue.

[petemoore]

Hi, just following up on this issue. Could you confirm whether this issue is in or out of scope, or whether there are any open questions? Many thanks for your time and all the work you do on the firmware.