From 37d19e2a4292ea58976f938abdebca4ace2e2ddf Mon Sep 17 00:00:00 2001
From: Malte Ubl <malte@vercel.com>
Date: Fri, 9 Oct 2026 13:25:54 -0700
Subject: [PATCH] Compile ELF worker runtime units with the initial-exec TLS
 model

Worker runtime variants link only into executables, whose own TLS block
holds every runtime thread-local, but they were compiled with the
default general-dynamic model. The linker relaxes each __tls_get_addr
call to a %fs-relative address, yet the compiler had already treated it
as a call: hot runtime leaves saved and restored extra registers around
every thread-local access (scr_mg_visit pushed five registers instead
of one). With -ftls-model=initial-exec each access is one %fs-relative
operand. Darwin, Windows and wasm variants are unchanged.

tsc-ts self-check, Linux sandbox, Worker build vs non-worker build:
1.04x -> 1.03x (9 interleaved runs).
---
 packages/runtime-pack-common/scripts/build.mjs | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/packages/runtime-pack-common/scripts/build.mjs b/packages/runtime-pack-common/scripts/build.mjs
index f14ea943c..01d7a4db5 100644
--- a/packages/runtime-pack-common/scripts/build.mjs
+++ b/packages/runtime-pack-common/scripts/build.mjs
@@ -229,11 +229,22 @@ async function build() {
                 ? ["-I", zlib]
                 : []),
             ];
+            // Worker variants link only into executables, whose own TLS block
+            // holds every runtime thread-local: on ELF the initial-exec model
+            // makes each access one %fs-relative load, where the default
+            // general-dynamic model compiles to a __tls_get_addr call (the
+            // linker relaxes the call away, but the callers already spilled
+            // their registers around it).
+            const tlsFlags =
+              variant.defines.includes("SCR_WORKERS") && matrix.target.object_format === "elf"
+                ? ["-ftls-model=initial-exec"]
+                : [];
             const unitFlags = [
               ...commonFlags,
               flavorSpec.optimization,
               ...debugFlags,
               ...variant.defines.map((define) => `-D${define}`),
+              ...tlsFlags,
               ...includeFlags,
             ];
             let bitcode;