The irony is that the first version of android contains gplv3 software : https://android.git.kernel.org/?p=platform/external/dosfstools.git;a=summary .

But it was quickly replaced by a bsd clone : fsck_msdos.git

I don't know if there are any product with the gplv3 fsck. May be some cupcake versions.

First arm only does the cpu, everything that it is around it (timer, interrupt controller, memory controller, uart ,...) can be anything.

And there is no easy way to discover them (like pci bus, acpi, ...). That what device tree want to do.
But that won't really solve the code size. Embedded company want to reduce cost and often design simple soc block (gpio, uart, ...), and this make the number of driver very big.
Also there no standard interface like ehci, ahci, ...
Even the same vendor can change the controller across the chip generation.

So I think it will be very difficult to unify things. What the advantage of arm for soc maker. There are free to do what they want.

The update is managed by gsm carrier/phone maker and lot's of phone don't get any update.

That's a broken model.

With apple the device is close, but all devices get new update.

All version of android will got some local exploit bugs (from kernel, app running as root, ...).
This means people can create valid application (without any specific perm), that can :
- run exploit and become root
- destroy your phone (erase bootloader)
- steal your information (spy your location, your call)
- make your phone a spam relay, ...

However, many people insist on using more than 1GB of physical memory on such a 32 bit system. This makes it necessary for the Linux kernel to jump through some interesting hoops... Basically the system uses the following tactics: * Memory above the physical address of 896MB are temporarily mapped into kernel virtual memory whenever the kernel needs to access that memory. * Data which the kernel frequently needs to access is allocated in the lower 896MB of memory (ZONE_NORMAL) and can be immediately accessed by the kernel (see Temporary mapping). * Data which the kernel only needs to access occasionally, including page cache, process memory and page tables, are preferentially allocated from ZONE_HIGHMEM. * The system can have additional physical memory zones to deal with devices that can only perform DMA to a limited amount of physical memory, ZONE_DMA and ZONE_DMA32. * Allocations and pageout pressure on the various memory zones need to be balanced (see Memory Balancing).

Cortex A15 uses 40-bit physical addressing is really useless : it is x86 PAE

But 64 bits is really interesting for the kernel. With Linux you can only map 1G of map without using segmentation (highmem).

Using 64 bits kernel with 32 bit programs solve the issue. And that's what is done on sparc, ppc, mips...

Also some algorithm really need 64 bits arithmetic to be faster.

No they should have construct IP adress like phone number : an extend it by adding prefix/suffix.

It should have simplified router job, because they could have done the routing only on some part of the number