Gave the gnu ports a CMake build, which most of them lacked - #607

Problem

The project guidelines ask for CMake and Ninja, but only 15 of the 59 gnu port
directories had a CMakeLists.txt. None of the 27 AArch64 ports had one, so the
architecture whose examples were repaired over #597, #599 and #602 still could not
be built the way the project says to build it, and nothing in CI could compile it.

Change

44 new CMakeLists.txt. Three of them are templates in ports_arch, because
34 of the 44 are generated. The ARMv7-A and AArch64 source lists are uniform
within each family — checked across every core — so one template per family serves
all of them and update.sh distributes it:

The other 10 ports have no generator and get their own file.

30 new toolchain files, following the shape of cmake/cortex_a9.cmake.
AArch64 needs a base file of its own rather than a variant of
arm-none-eabi.cmake: it has no -marm/-mthumb to choose between and no
-mfloat-abi, and aarch64-none-elf-gcc rejects -mlong-calls outright, so that
flag cannot simply be carried across.

The tools are named without a path, unlike cmake/cortex_r52.cmake, which pins
one. Pinning 30 files to a single machine's directory layout is exactly what #603
removed from the launch configurations, so the majority convention wins here;
override CMAKE_C_COMPILER for a toolchain that is not on PATH.

Three of those files cover ports that already had a CMakeLists.txt but no way
to select it — the Armv8-M mainline gnu ports cortex_m33, cortex_m55 and
cortex_m85. Without a cmake/<arch>.cmake, the documented invocation cannot
reach them.

One top-level fix. CMakeLists.txt derives the SMP port directory as
<arch>_smp, but ports_smp/linux and ports_smp/win64 predate that convention
and carry no suffix, so neither could ever be configured:

set(TX_ARCH_DIR "${THREADX_ARCH}_smp")
if(NOT IS_DIRECTORY "${CMAKE_CURRENT_LIST_DIR}/${TX_PORT_DIR}/${TX_ARCH_DIR}")
    set(TX_ARCH_DIR "${THREADX_ARCH}")
endif()

The fallback only fires when the suffixed directory is absent, so no port that
already resolved changes behaviour, and ports_smp/win64's existing
CMakeLists.txt becomes reachable as a side effect.

Verification

Every port was configured and built, not merely written.

53 of 53 static libraries build with cmake -G Ninja, using Arm GNU Toolchain
14.3.Rel1 for both arm-none-eabi and aarch64-none-elf. That is the 44 new
ports plus the 9 that already worked, and it includes ports_smp/linux, which
failed before the fallback and now produces a library. Sizes land where you would
expect: ARMv7-A around 267–270 KB, AArch64 ThreadX around 333 KB, AArch64 SMP
around 400 KB.

• scripts/check_ports.sh passes, so the three templates and their 34 generated
  copies agree.
• scripts/check_clang.sh passes unchanged — 42 of 42 script-driven examples and
  5 of 5 Cortex-R52 CMake images. That last one matters, because the R52 CMake
  stage runs through the top-level file this change edits.

All three were re-run after rebasing onto #606, with identical results.

Still outside the CMake build

Six gnu ports, every one for want of a compiler rather than a CMakeLists.txt:

• rxv1, rxv2, rxv3 need the Renesas RX GNU toolchain, and
  mips32_interaptiv_smp needs a MIPS one. Neither is available here, and writing
  toolchain files for them would mean shipping untested guesses.
• risc-v32 and risc-v64 already build, through their own differently named
  toolchain files.

Suggested follow-up

Nothing in CI builds any of this yet, which is how the AArch64 ports came to sit
two releases behind in the first place. A workflow running the 53
configure-and-build steps would need the aarch64-none-elf toolchain downloaded
and cached alongside the existing arm-none-eabi one, in the manner
clang_check.yml already caches ATfE.