new(acpica.org): iasl ACPI compiler

[tannevaled]

Adds acpica.org, packaging Intel ACPICA's iasl — the ASL+ optimizing compiler/disassembler.

Why

EDK II (#13239) shells out to iasl to compile .asl ACPI tables into .aml during firmware builds (RamDiskDxe's NFIT table, etc.). iasl isn't currently in the pantry, so the edk2 build fails with iasl: command not found (exit 127). This packages it as a standalone build-time dependency.

Prerequisite for #13239 — that PR build-depends on acpica.org, so this should land first.

Notes

• Tags switched from R20YY_MM_DD to a clean YYYYMMDD scheme around mid-2025; we match only the latter.
• darwin/arm64 links with -Wl,-no_fixup_chains: ACPICA keeps unaligned function pointers in __DATA,__const tables that the mach-o chained-fixups format rejects (ld: pointer not aligned in '_AcpiExDumpNode'; lld emits the same advice).

Test plan

Built iasl locally on darwin/arm64 (pantry flex/bison/clang) from the resolved tag 20260408; iasl -v reports the version and it compiles a trivial SSDT to .aml.

[tannevaled]

Hey @jhheider 👋 — sorry to ping you directly, no rush at all!

When you have a spare moment, would you mind taking a look at this one? It's a small standalone recipe for iasl (ACPICA's ASL compiler). It's the prerequisite for #13239 (the EDK II / OVMF firmware blobs) — that build shells out to iasl to compile the ACPI tables, so it can't go green until this lands in main.

CI is green on all four platforms here. The only platform-specific bit is the -Wl,-no_fixup_chains on darwin/arm64 (ACPICA keeps unaligned pointers in __DATA,__const that the mach-o chained-fixups format rejects).

Thanks a lot for everything you do on the pantry! 🙏