Uhh, no, not at all.
The iOS kernel is based on the OSX kernel which has already gone through two 32->64 bit transitions, on both big and little-endian hardware. I would imagine there's very little of it that is not wrapped in macros which will automatically compile appropriately. There'll be the usual necessity for some manual intervention for VM and atomics, nothing that hasn't been done before.
The ARM A8 manuals describe how to run existing 32-bit binaries, and it doesn't involve emulation, any more than running 32-bit x86 binaries requires emulation. Finally, Apple is, you know, a SECRETIVE company. If they want to keep this secret until launch, they will not have this into the SDK. After all, it's not necessary right now.
The advantages of going to 64 bit for Apple are
- boasting. (Not technically important, but a way to impress the crowd who are excited by Moto's X8 architecture and Samsung's Octacore)
- allows them to move to >4GB of RAM EASILY and without a period of frustration. Apple has handled pretty much all these transitions well --- because they planned EARLY. MS and the Intel world have NOT handled them well (32->64 bit, various jumps in HD sizes, 4K disk sectors) because they DON'T plan years in advance.
- the sooner they go 64-bit the sooner they can use the A8 architecture. This is the real win, more so than a larger address space. The A8 architecture was ARM's chance to rethink everything they've done over the past ten years or so, improve what works well, and drop what was a bad idea.
