A couple of updates:
* Tim fixed ThinLTO on ppc64el via fix-clang-selection.patch, added in
the 123.0.6312.86 release.
* Thanks to bug #1072299, I've built chromium 126.0.6478.126 for
bookworm for my own usage/testing, and in addition to the allocator
tweak suggested in that thread, I enabled ThinLTO. Here are the tweaks
currently needed on stable to avoid subsequent build breakage:
--- a/build/config/compiler/BUILD.gn
+++ b/build/config/compiler/BUILD.gn
@@ -772,10 +772,6 @@ config("compiler") {
if (is_apple) {
ldflags += [ "-Wcrl,object_path_lto" ]
}
-
- # We only use one version of LLVM within a build so there's no need to
- # upgrade debug info, which can be expensive since it runs the verifier.
- ldflags += [ "-Wl,-mllvm,-disable-auto-upgrade-debug-info" ]
}
# TODO(crbug.com/335365324): Enable on other platforms.
--- a/build/config/rust.gni
+++ b/build/config/rust.gni
@@ -75,7 +75,7 @@ declare_args() {
#
# TODO(crbug.com/40281834): Re-enable ThinLTO for Rust on LaCrOS
# TODO(b/300937673): Re-enable ThinLTO for Rust on ash-chrome
- toolchain_supports_rust_thin_lto = !is_chromeos
+ toolchain_supports_rust_thin_lto = false
# Any extra std rlibs in your Rust toolchain, relative to the standard
# Rust toolchain. Typically used with 'rust_sysroot_absolute'
(The stable version of Rust appears not capable of LTO.)
So far, I haven't noticed any ill effects in my "production" home
environment... a laptop with 3 GB RAM, and chrome://discards showing 691
tabs :] (But I can't say it feels noticeably snappier, either. Some
kind of browser-performance benchmarking is probably in order...)
