[llvm] [LLVM][MC][DecoderEmitter] Add support to specialize decoder per bitwidth (PR #154865)

[Sergei Barannikov via llvm-commits]

s-barannikov wrote:

Heads up, pre-merge checks seem to be broken:
```
2025-09-01T21:09:25.5173264Z /home/gha/actions-runner/_work/llvm-project/llvm-project/llvm/lib/Target/RISCV/Disassembler/RISCVDisassembler.cpp:705:13: error: explicit specialization cannot have a storage class [-Werror,-Wexplicit-specialization-storage-class]
2025-09-01T21:09:25.5174352Z   705 | template <> static constexpr uint32_t llvm::MCD::InsnBitWidth<uint16_t> = 16;
2025-09-01T21:09:25.5174684Z       |             ^~~~~~
2025-09-01T21:09:25.5175620Z /home/gha/actions-runner/_work/llvm-project/llvm-project/llvm/lib/Target/RISCV/Disassembler/RISCVDisassembler.cpp:706:13: error: explicit specialization cannot have a storage class [-Werror,-Wexplicit-specialization-storage-class]
2025-09-01T21:09:25.5176769Z   706 | template <> static constexpr uint32_t llvm::MCD::InsnBitWidth<uint32_t> = 32;
2025-09-01T21:09:25.5177142Z       |             ^~~~~~
2025-09-01T21:09:25.5177958Z /home/gha/actions-runner/_work/llvm-project/llvm-project/llvm/lib/Target/RISCV/Disassembler/RISCVDisassembler.cpp:708:13: error: explicit specialization cannot have a storage class [-Werror,-Wexplicit-specialization-storage-class]
2025-09-01T21:09:25.5178922Z   708 | template <> static constexpr uint32_t llvm::MCD::InsnBitWidth<uint64_t> = 48;
2025-09-01T21:09:25.5179256Z       |             ^~~~~~
2025-09-01T21:09:25.5179438Z 3 errors generated.
```
Not sure how this PR did pass the checks (maybe not all bots use -Werror?).


https://github.com/llvm/llvm-project/pull/154865