Convert to SPDX license tags

This includes marking all documentation files CC-BY-SA-4.0.

python-capdl-tool: fix quirk in get_object_size

Don't return 1 when the value returned is 0.

capdl: support 40-bit PA size on aarch64-hyp

Update vspace structure when starting at level 1

Co-authored-by: Anna Lyons <anna@gh.st>

python-capdl-tool: add untyped allocator

This change adds an offline allocator for calculating untyped
derivations, which will be used in future to specify exact derivations
from boot info untyped objects. This allows the allocation logic to be
removed from the loader, this simplifying the loader model for
verificaion while allowing for policy freedom in terms of allocation.

The initial allocator implementation is a best-fit allocator.

untyped_gen: remove code for dealing with rootserver

We now expect the elfloader and kernel boot to place the rootserver at
the end of available memory. This means that we can ignore them when
allocating from the start of memory (as long as they do not conflict
somewhere in the middle).

cdl_utils: get untyped_gen to run

python-capdl: Allow page_sizes to take Arch type

If page_sizes is called with an actual Arch object type (or an object
that implements the get_pages method, then don't lookup the architecture
again. This is to support cases where an Arch object instance is
available, but the string that was used to obtain it is not in scope.

python-capdl-tool: add clz implementation

Add clz for doing bit magic

python-capdl-tool: expose levels

Expose the levels function so that it can be used by other tools.
Make Arch be an abstract class with default methods.

python-capdl-tool: expose list of valid archs

Expose the keys of the dictionary representing valid architectures. This
can then be used as part of the arguments to scripts.

trivial: python-capdl-tool: style util.py

python-capdl-tool: Add RISCV architecture layouts

The kernel currently assumes SV32 memory layout for rv32 and SV39 layout
for rv64 which is reflected here.

python-capdl-tool: Split normalise_architecture

Normalising architecture would normalise aarch32 and arm_hyp
architectures to aarch32 as it went through the CapDL spec arch which
may not have per architecture names

python-capdl-tool: add helpers to utils

- ctz
- round_up
- is_aligned

python-capdl-tool: merge Aarch32Arch and ArmHypArch

The only difference now is the level size, so add a flag for this rather
than duplicate the code.

python-capdl-tool: use get_object_size for frame sizes

This change removes the hardcoded frame sizes from util.c and uses
get_object_size to query the sizes of frames.

python-capdl-tool: reduce duplicated object constants

This change removes duplicated object types for types that are shared
across architecutes. For example, rather than have both
seL4_IA32_PageTableObject and seL4_ARM_PageTableObject, just use
seL4_PageTableObject.

This reduces code complexity.

python-capdl-tool: Add aarch64 support

python: create and use ObjectType enum

Move the object type (seL4_.*Object) constants into a new enum, ObjectType.
This allows the types to be iterable and reduces the number of imports,
resulting in less boilerplate.

Fix whitespace

- Remove trialing whitespace
- Remove duplicate blank lines
- Remove blank lines at end of file

Fix licenses

Replace xrange with six.moves.range

python-capdl: Define x86_64 types

python-capdl-tool: Add IPC buffer size definition

python-capdl-tool: Define a real architecture definition

This replaces and expands upon the architecture abstractions in util.py.
Previously the architecture just abstracted some sizes of pages and
layouts of a two level paging hierarchy. Now this is generalized both
for arbitrary vspace hierarchies, as well as other architecture
information.

The original interface is left and emulated.

Support aarch32 as a way of specifying the internal arm11 architecture

This change supports the new naming scheme for the 'arm' architecture,
which has been disambiguated into aarch32 (arms official name), in
prepartion for supporting aarch64

python-capdl: Remove hyp parameters from paging functions.

This commit marks the final removal of special casing of ARM's Hyp mode. From
this point on, Hyp mode should be considered an entirely separate architecture.

python-capdl: Stop using the hyp parameter as a way to detect Hyp mode.

Now indicated via the "arm_hyp" architecture string alone.

python-capdl: Use normalise_architecture for increased ARM Hyp mode compat.

This commit allows the util function to recognise the architecture string
"arm_hyp" as a way of indicating ARM's Hyp mode. That is, you can now specify
Hyp mode in two ways:

1. architecture "arm" and hyp=True (legacy); or
2. architecture "arm_hyp".

The purpose of this commit is movement towards removing the `hyp` argument
entirely and having a single, consolidated way of indicating the target
architecture. Note that none of this new functionality is currently used yet.

python-capdl: Add a function for normalising architecture strings.

This is currently unused but supports a future commit.

python-capdl: Enable Python 3 compatibility settings from `__future__`.

This change enables a variety of Python 3 compatibility settings from
`__future__` when running Python 2. Essentially this forces all existing code
under Python 2 to conform to the Python 3 semantics. This makes bilingual
compatibility easier.

python-capdl: Use `//` whenever performing a division that assumes floor.

The `//` and `/` operators are equivalent in standard Python 2, but Python 3
changes the `/` operator to return a float, rather than flooring the result.
These changes are part of movement towards switching on division compatibility
from `__future__` within Python 2.

Moved python-capdl-tool into subdirectory