Tags · cx-home/v

cx-patched-v

cx patched V pin — CX-free fork tip (eviction complete)

Jun 18, 2026
e8a9921
zip
tar.gz

archive/v0.7.0-cx-patches

fix(libgc): P0 vlang#14 — flip bundled amalgamation thread-local-allo…

…c to ON

The bundled gc.c amalgamation embedded `/* #undef THREAD_LOCAL_ALLOC */`
in its autoheader config — the one alloc-lock mitigation left OFF. Flip it
to `#define THREAD_LOCAL_ALLOC 1` (the canonical configure-equivalent of
`--enable-thread-local-alloc=yes`); the TLA implementation is already present
in the amalgamation (guarded by `#ifdef THREAD_LOCAL_ALLOC`), so this activates
real code. Compiles clean.

This affects the Linux / `-prod`-bundled `gc.o` path only. The macOS `cx`
build links the prebuilt `thirdparty/tcc/lib/libgc.a` (built by
thirdparty-macos-arm64_bdwgc.sh), which already defaults TLA on — so the flip
brings the bundled config into parity with what macOS already ships.

Companion runtime lever (already present at this pin): the macOS marker-pin
`GC_set_markers_count(1)` before `GC_INIT()` in cmain.v::gen_boehm_gc_init().
Together: ~2.4-5x on MARK-bound multi-thread work. Neither removes the
GC_allocate_ml alloc-lock, so alloc-heavy `[?map [par]]` (cx-private vlang#14) stays
~1.3x slower than serial — partial relief by design. Full acceptance +
measurements in cx-private bench/parallel-alloc/P0-ACCEPTANCE.md.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Jun 11, 2026
0807dd1
zip
tar.gz

archive/v0.7.0-cx-patches-master

builtin: skip precompiled libgc gc.o on macOS even under -prod

The precompiled `thirdparty/libgc/gc.o` was built with parallel-mark
trampolines enabled, which require rwx pages via mprotect(PROT_EXEC).
macOS hardened runtime denies rwx pages by default, so any V `-shared`
+ `-prod` build on macOS aborts at first allocation with "Cannot
allocate executable pages".

The fix is to skip the precompiled fast-path on macOS and force the
source-compile path (via `thirdparty/tcc/lib/libgc.a` archive), which
produces a libgc without the trampoline allocator. Linux is unaffected
and keeps the precompiled fast-path for `-prod` builds.

Companion to the vlang#27178 cgen patch in this fork. With both patches
applied, `v -prod -shared` produces a libcx.dylib that dlopen's
cleanly under macOS hardened runtime AND auto-registers GC threads
for host runtimes that need it.