LATX, fix: Release AOT merge and invalidated segment resources

[LaurenIsACoder]

Scope

Based on current upstream bfe17ae7c1. Fix AOT resources whose owning operation has finished:

• Free both merge-input buffers and their vector on success, duplicate input and errors.
• Destroy non-TU merge indexes before their borrowed input data is freed; free owned TB-index values on replacement and destruction.
• Free directory-scan allocations when opening the cache directory fails.
• Invalidate every overlapping segment on successful unmap/remap/replacement, not only the first match. Segment-owned address-specific PE/cache mappings are released; reusable ELF library-cache entries retain their existing policy.

This does not duplicate upstream's AOT reader fix (17b479de7877) or change the cache format/capacity. Thread ownership changes in #472 are separate; this PR does not depend on them.

Author self-review and combined validation

The actual published diffs were reviewed individually and checked against the local candidates. Follow-up corrections were folded into their existing commits; author/sign-off metadata updates preserved the per-PR source trees. The series remains six commits across #472, #473, #474, #475 and #476.

All five branches merge without conflicts on bfe17ae7c1. The combined source tree 80aaaaf13688387640f528d1c0bccb6a800f8d6e builds and passes 33/33 lat-pr-fast tests. On that combined binary, the 500-cycle KZT unload and retained-handle checks, 520-cycle SMC/fork checks, and 512 successful TSYNC operations alongside thread creation also complete successfully. A tests-disabled configuration registers no tests.

The integration suite still reports 27 skips, including after selecting the available guest sysroot with LATX_X86_64_SYSROOT: 25 require clang and two require the x86 cross compiler. These are not passes. This author self-review does not substitute for the upstream CI matrix, maintainer review or real-application acceptance.

Validation

LoongArch64, GCC 15.3, x86_64-linux-user, LAT O1, KZT compiled in, release build:

./configure --target-list=x86_64-linux-user --enable-latx --optimize-O1 \
  --extra-ldflags=-ldl --enable-kzt --enable-tests --disable-docs \
  --disable-werror --meson=meson
ninja -C build -j2 all meson-test-prereq
env -u LAT_LD_PREFIX meson test -C build --no-rebuild \
  --suite lat-pr-fast --print-errorlogs --num-processes 2
env -u LAT_LD_PREFIX meson test -C build --no-rebuild \
  --suite latx-integration --print-errorlogs --num-processes 2

• Isolated PR branch: 27/27 lat-pr-fast passed.
• New TU/non-TU merge regressions read controlled real cache files through the production merger and assert balanced allocations, including replacement and failure paths. TU translation/publication endpoints are stubbed; this is ownership-path evidence, not execution of generated AOT code.
• New unmap regression invokes production target_munmap with real host mappings: a single range removes three segment/cache owners while preserving both neighbors; mincore verifies the released mappings. Invalid and repeated unmap are covered.
• Earlier RED runs on the unfixed upstream production bodies fail the ownership checks.
• Integration suite: 27 skipped (exit 77) due to the unavailable clang/guest-building toolchain. These are not passes. Broader guest mapping/AOT compatibility remains an integration gate.

No installed runtime changes or claim of full application memory acceptance.