Rust: `error: relocation refers to local symbol "" [12], which is defined in a discarded section` only when using ld.gold linker with --gc-sections

[ghost]

tl;dr: https://github.com/rust-lang/rust/issues/59652#issuecomment-482253086

Cannot compile firefox anymore due to:
error: linking with ... cargo-linker
error: relocation refers to local symbol "" [12], which is defined in a discarded section

UPDATE bisect shows it's PR 59401

https://bugzilla.mozilla.org/show_bug.cgi?id=1541214

old info

This issue doesn't happen with
Last good nightly:
nightly-2019-03-28-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (33ef0bad2 2019-03-27)
binary: rustc
commit-hash: 33ef0bad21d6bb646c7c3ab0dbf381ca96c324bf
commit-date: 2019-03-27
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0
nightly-2019-03-29-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (237bf3244 2019-03-28)
binary: rustc
commit-hash: 237bf3244fffef501cf37d4bda00e1fce3fcfb46
commit-date: 2019-03-28
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0

The issue happens with
First bad nightly:
nightly-2019-03-29-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (237bf3244 2019-03-28)
binary: rustc
commit-hash: 237bf3244fffef501cf37d4bda00e1fce3fcfb46
commit-date: 2019-03-28
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0
nightly-2019-03-30-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (e782d790f 2019-03-29)
binary: rustc
commit-hash: e782d790f1b63d82af39248bebe027f92d891bcc
commit-date: 2019-03-29
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0

For each test I was using the same cargo:
cargo 1.35.0-dev (025b01ed 2019-04-01)
release: 1.35.0
commit-hash: 025b01edd0bba49b7e49c1cacc65bb1f6462ee57
commit-date: 2019-04-01

Errors look like this:

0:16.63 = note: /home/user/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-9b00dedc0dda6ebf.rlib(compiler_builtins-9b00dedc0dda6ebf.compiler_builtins.cqcjgied-cgu.0.rcgu.o)(.stack_sizes+0x0): error: relocation refers to local symbol "" [12], which is defined in a discarded section

0:16.84 section group signature: "(null)"
0:16.84 /home/user/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-e4188d26f42911c5.rlib(std-e4188d26f42911c5.std.b8iklw2d-cgu.0.rcgu.o)(.stack_sizes+0x1b): error: relocation refers to local symbol "" [696], which is defined in a discarded section
0:16.84 section group signature: "(null)"

[ghost]

UPDATE: ignore this comment.

I was wrong about first bad nightly: it's not nightly-2019-03-29-x86_64-unknown-linux-gnu but instead it's:

nightly-2019-03-30-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (e782d790f 2019-03-29)
binary: rustc
commit-hash: e782d790f1b63d82af39248bebe027f92d891bcc
commit-date: 2019-03-29
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0

And nightly-2019-03-29-x86_64-unknown-linux-gnu is a good nightly.(confirmed firefox compilation succeeded with it!)

I'm still double-checking but it takes some time...done
I'm also trying to do a rust bisect (even more time) [does anyone know if I have to do a git submodule update after each git bisect good/bad ? answer: not really, ./x.py build does it automatically for changed submodules, but just to be sure I'm doing git submodule update --init --recursive --progress followed by ./x.py clean && time ./x.py build --stage 1 -j 5 then clean firefox build dir(ie. remove it) and retry compiling, error usually appears after 8mins, but to be sure I've to leave it compiling for like 23mins]

[mati865]

You might want to use https://github.com/rust-lang-nursery/cargo-bisect-rustc/blob/master/TUTORIAL.md

[ghost]

8b8488ce8fc047282e7159343f30609417f9fa39 is the first bad commit, made from PR https://github.com/rust-lang/rust/pull/59401

    bootstrap: build compiler-builtins with -Z emit-stack-sizes

$ git bisect log
git bisect start
# good: [237bf3244fffef501cf37d4bda00e1fce3fcfb46] Auto merge of #59478 - Centril:rollup, r=Centril
git bisect good 237bf3244fffef501cf37d4bda00e1fce3fcfb46
# bad: [e782d790f1b63d82af39248bebe027f92d891bcc] Auto merge of #59522 - Centril:rollup, r=Centril
git bisect bad e782d790f1b63d82af39248bebe027f92d891bcc
# bad: [e782d790f1b63d82af39248bebe027f92d891bcc] Auto merge of #59522 - Centril:rollup, r=Centril
git bisect bad e782d790f1b63d82af39248bebe027f92d891bcc
# bad: [3df97f9da8e4a23772f5845bc641adae03e032be] Rollup merge of #59401 - japaric:compiler-builtins-stack-sizes, r=alexcrichton
git bisect bad 3df97f9da8e4a23772f5845bc641adae03e032be
# good: [b75b1655895e69f074936b73394a72e98aa067b9] Rollup merge of #59398 - phansch:rustfix_coverage, r=oli-obk
git bisect good b75b1655895e69f074936b73394a72e98aa067b9
# good: [8794e21ff329d1201d484c015d48e85490a64fa9] Rollup merge of #58019 - Zoxc:combine-late-lints, r=estebank
git bisect good 8794e21ff329d1201d484c015d48e85490a64fa9
# good: [6c8e3a5378e41e08a323139bb7aaa8a8823ec9bc] Remove unused variable
git bisect good 6c8e3a5378e41e08a323139bb7aaa8a8823ec9bc
# good: [f9262afa4d1a88715ff57907bd17eda4d039cea6] Rollup merge of #59394 - mark-i-m:dup-matcher-bindings-2, r=Centril
git bisect good f9262afa4d1a88715ff57907bd17eda4d039cea6
# bad: [7d365cf27f4249fc9b61ba8abfc813abe43f1cb7] compile all crates under test w/ -Zemit-stack-sizes
git bisect bad 7d365cf27f4249fc9b61ba8abfc813abe43f1cb7
# bad: [8b8488ce8fc047282e7159343f30609417f9fa39] bootstrap: build compiler-builtins with -Z emit-stack-sizes
git bisect bad 8b8488ce8fc047282e7159343f30609417f9fa39
# first bad commit: [8b8488ce8fc047282e7159343f30609417f9fa39] bootstrap: build compiler-builtins with 

Just tested commit f8673e0ad85e98997faa76fa7edc99c5825f46ee (HEAD -> master, origin/master, origin/HEAD)
with this patch applied:

diff --git a/src/bootstrap/bin/rustc.rs b/src/bootstrap/bin/rustc.rs
index 7429492f91..d508969d29 100644
--- a/src/bootstrap/bin/rustc.rs
+++ b/src/bootstrap/bin/rustc.rs
@@ -211,7 +211,7 @@ fn main() {
                 || target.contains("-none-eabi")
                 || target.ends_with("-none-elf"))
         {
-            cmd.arg("-Zemit-stack-sizes");
+            //cmd.arg("-Zemit-stack-sizes");
         }

         if let Ok(s) = env::var("RUSTC_CODEGEN_UNITS") {

and can confirm that firefox compiled past the issue.

Will test without the patch next...done: confirmed the issue exists without the above patch!

Will test again with the patch and allow firefox to fully compile&install (done, it worked!)

[mati865]

CC @japaric

[pnkfelix]

triage: P-high. Leaving nominated for now (What is going on here with -Z emit-stack-sizes ...?)

[japaric]

Sorry for the breakage here. If PR #59401 is blocking anything feel free to revert it. It will probably be several days before I have some free time to dig into this (though my guess is that one of the extra, custom linker flags that firefox uses in their builds (see MOZ_CARGO_WRAP_LDFLAGS in the linked bugzilla ticket) is not playing nicely with the new .stack_sizes section)

[ghost]

If I remove -Wl,--gc-sections from the args to cargo-linker then the issue doesn't happen!

--gc-sections is set not by MOZ_CARGO_WRAP_LDFLAGS but rather by lib/librustc_codegen_ssa-6bcbc85a2479ae62.so (I know because it's the only place I hexedited it to --hc-sections and got /usr/bin/ld.gold: --hc-sections: unknown option, and it also shows in the exact same place in the args to cargo-linker)

#!/bin/bash

rustup default master-stage1

export MOZ_CARGO_WRAP_LD='/usr/bin/clang -std=gnu99'
export MOZ_CARGO_WRAP_LDFLAGS='-lpthread -Wl,-O1,--sort-common,--as-needed,-z,relro -Wl,-rpath,/usr/lib/firefox -fuse-ld=gold -Wl,-z,noexecstack -Wl,-z,text -Wl,-z,relro -Wl,-z,nocopyreloc -Wl,-Bsymbolic-functions -Wl,--build-id=sha1 -fstack-protector-strong -Wl,--icf=safe -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/bin -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/lib -fcolor-diagnostics'

# shellcheck disable=SC2016
#CARGO_PKG_REPOSITORY='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION_PATCH=27 CARGO_PKG_VERSION_PRE='' CARGO_TARGET_BINFILE_FULLPATH=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build-script-build CARGO_PKG_VERSION_MAJOR=0 CARGO_PKG_DESCRIPTION='A stable implementation of the upcoming new `proc_macro` API. Comes with an option, off by default, to also reimplement itself in terms of the upstream unstable API.' LD_LIBRARY_PATH='/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps:/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib' CARGO_MANIFEST_DIR=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2 CARGO_PKG_HOMEPAGE='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION=0.4.27 CARGO_PKG_VERSION_MINOR=4 CARGO_PKG_AUTHORS='Alex Crichton <[email protected]>' CARGO_PKG_NAME=proc-macro2 CARGO=/home/user/build/2nonpkgs/rust.stuff/cargo/cargo/target/release/cargo /home/user/bin/rustc --crate-name build_script_build /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2/build.rs --color always --crate-type bin --emit=dep-info,link -C opt-level=2 -C codegen-units=1 --cfg 'feature="default"' --cfg 'feature="proc-macro"' -C metadata=3b19a3eb863545be -C extra-filename=-3b19a3eb863545be --out-dir /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be -C linker=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker -L dependency=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps --cap-lints warn

err="-Wl,--gc-sections"
#err="" #uncomment this line to avoid errors like: /home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-a9bb89774eb1c5f2.rlib(std-a9bb89774eb1c5f2.4ujdo36uwwum3zzu.rcgu.o)(.stack_sizes+0x0): error: relocation refers to local symbol "" [5], which is defined in a discarded section
"/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker" "-Wl,--as-needed" "-Wl,-z,noexecstack" "-m64" "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.build_script_build.ea30uw23-cgu.0.rcgu.o" "-o" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.1zobs1pcbqt3rtjo.rcgu.o" "$err" "-pie" "-Wl,-znow" "-Wl,-O1" "-nodefaultlibs" "-L" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps" "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-Wl,--start-group" "-Wl,-Bstatic" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-a9bb89774eb1c5f2.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libpanic_unwind-01499d126af771f6.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libbacktrace_sys-8b152c95a80afef5.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_demangle-14c23aab33e6c5f4.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libunwind-41eb09b84b2625ab.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liblibc-44a20fe3a684e072.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liballoc-e5275540f2101631.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_core-9f560ba3dfa532fe.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-1f1fadf3472d3239.rlib" "-Wl,--end-group" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-9e3a9887668f9c30.rlib" "-Wl,-Bdynamic" "-ldl" "-lrt" "-lpthread" "-lgcc_s" "-lc" "-lm" "-lrt" "-lpthread" "-lutil" "-lutil"

[ghost]

Apparently one of these:

https://github.com/rust-lang/rust/blob/3750348daff89741e3153e0e120aa70a45ff5b68/src/librustc_codegen_ssa/back/linker.rs#L297-L305

https://github.com/rust-lang/rust/blob/3750348daff89741e3153e0e120aa70a45ff5b68/src/librustc_codegen_ssa/back/linker.rs#L950-L952

[ghost]

By adding a digit to each of those two --gc-sections occurrences, recompiling rust and rerunning the (uhm, commented out line of the)script above(actually see the section at the end of this comment for the exact script) I determined it's exactly the first one, that is, this:

https://github.com/rust-lang/rust/blob/3750348daff89741e3153e0e120aa70a45ff5b68/src/librustc_codegen_ssa/back/linker.rs#L297-L305

Does anyone think that maybe it's --gc-sections in combination with some other arg that's causing the issue? Or wants me to try something else? or can someone provide a simpler reproduction sample? Either way I'm out of ideas ;-)

actual script ran

#!/bin/bash

rustup default master-stage1

export MOZ_CARGO_WRAP_LD='/usr/bin/clang -std=gnu99'
export MOZ_CARGO_WRAP_LDFLAGS='-lpthread -Wl,-O1,--sort-common,--as-needed,-z,relro -Wl,-rpath,/usr/lib/firefox -fuse-ld=gold -Wl,-z,noexecstack -Wl,-z,text -Wl,-z,relro -Wl,-z,nocopyreloc -Wl,-Bsymbolic-functions -Wl,--build-id=sha1 -fstack-protector-strong -Wl,--icf=safe -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/bin -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/lib -fcolor-diagnostics'

# shellcheck disable=SC2016
CARGO_PKG_REPOSITORY='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION_PATCH=27 CARGO_PKG_VERSION_PRE='' CARGO_TARGET_BINFILE_FULLPATH=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build-script-build CARGO_PKG_VERSION_MAJOR=0 CARGO_PKG_DESCRIPTION='A stable implementation of the upcoming new `proc_macro` API. Comes with an option, off by default, to also reimplement itself in terms of the upstream unstable API.' LD_LIBRARY_PATH='/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps:/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib' CARGO_MANIFEST_DIR=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2 CARGO_PKG_HOMEPAGE='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION=0.4.27 CARGO_PKG_VERSION_MINOR=4 CARGO_PKG_AUTHORS='Alex Crichton <[email protected]>' CARGO_PKG_NAME=proc-macro2 CARGO=/home/user/build/2nonpkgs/rust.stuff/cargo/cargo/target/release/cargo /home/user/bin/rustc --crate-name build_script_build /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2/build.rs --color always --crate-type bin --emit=dep-info,link -C opt-level=2 -C codegen-units=1 --cfg 'feature="default"' --cfg 'feature="proc-macro"' -C metadata=3b19a3eb863545be -C extra-filename=-3b19a3eb863545be --out-dir /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be -C linker=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker -L dependency=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps --cap-lints warn

[ghost]

Here's the output with both -Wl,--print-gc-sections(manually added) and -Wl,--gc-sections
https://gist.github.com/howaboutsynergy/065ec6322866b5defa28d1ec5989675d

EDIT:
I have further reduced the script, as long as it gives me the discarded type of errors, output here, script also here:

#!/bin/bash

#rustup default master-stage1

#export MOZ_CARGO_WRAP_LD='/usr/bin/clang -std=gnu99'
#export MOZ_CARGO_WRAP_LDFLAGS='-lpthread -Wl,-rpath,/usr/lib/firefox -fuse-ld=gold -Wl,-z,noexecstack -Wl,-z,text -Wl,-Bsymbolic-functions -Wl,--build-id=sha1 -fstack-protector-strong -Wl,--icf=safe -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/bin -Wl,-rpath-link,/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/dist/lib -fcolor-diagnostics'

# shellcheck disable=SC2016
#CARGO_PKG_REPOSITORY='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION_PATCH=27 CARGO_PKG_VERSION_PRE='' CARGO_TARGET_BINFILE_FULLPATH=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build-script-build CARGO_PKG_VERSION_MAJOR=0 CARGO_PKG_DESCRIPTION='A stable implementation of the upcoming new `proc_macro` API. Comes with an option, off by default, to also reimplement itself in terms of the upstream unstable API.' LD_LIBRARY_PATH='/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps:/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib' CARGO_MANIFEST_DIR=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2 CARGO_PKG_HOMEPAGE='https://github.com/alexcrichton/proc-macro2' CARGO_PKG_VERSION=0.4.27 CARGO_PKG_VERSION_MINOR=4 CARGO_PKG_AUTHORS='Alex Crichton <[email protected]>' CARGO_PKG_NAME=proc-macro2 CARGO=/home/user/build/2nonpkgs/rust.stuff/cargo/cargo/target/release/cargo /home/user/bin/rustc --crate-name build_script_build /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/third_party/rust/proc-macro2/build.rs --color always --crate-type bin --emit=dep-info,link -C opt-level=2 -C codegen-units=1 --cfg 'feature="default"' --cfg 'feature="proc-macro"' -C metadata=3b19a3eb863545be -C extra-filename=-3b19a3eb863545be --out-dir /home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be -C linker=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker -L dependency=/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps --cap-lints warn

#err=("-Wl,--print-gc-sections" "-Wl,--gc-sections")
##err="" #uncomment this line to avoid errors like: /home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-a9bb89774eb1c5f2.rlib(std-a9bb89774eb1c5f2.4ujdo36uwwum3zzu.rcgu.o)(.stack_sizes+0x0): error: relocation refers to local symbol "" [5], which is defined in a discarded section
#"/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker" "-Wl,--as-needed" "-Wl,-z,noexecstack" "-m64" "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.build_script_build.ea30uw23-cgu.0.rcgu.o" "-o" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.1zobs1pcbqt3rtjo.rcgu.o" "${err[@]}" "-pie" "-Wl,-znow" "-Wl,-O1" "-nodefaultlibs" "-L" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps" "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-Wl,--start-group" "-Wl,-Bstatic" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-a9bb89774eb1c5f2.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libpanic_unwind-01499d126af771f6.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libbacktrace_sys-8b152c95a80afef5.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_demangle-14c23aab33e6c5f4.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libunwind-41eb09b84b2625ab.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liblibc-44a20fe3a684e072.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liballoc-e5275540f2101631.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_core-9f560ba3dfa532fe.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-1f1fadf3472d3239.rlib" "-Wl,--end-group" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-9e3a9887668f9c30.rlib" "-Wl,-Bdynamic" "-ldl" "-lrt" "-lpthread" "-lgcc_s" "-lc" "-lm" "-lrt" "-lpthread" "-lutil" "-lutil"
#"/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/build/cargo-linker" 

success=0 #1 if you want to see successful compilation(assuming $causetheissue is not 1), any other value(like 0) otherwise
causetheissue=1 #if 1 you will see 'discarded' type of errors, regardless of the value of $success
if test "$causetheissue" == '1'; then
  err="-Wl,--gc-sections"
else
  err="" #no errors
fi
#noeffect=("-pie" "-m64")
noeffect=("-nopie") #it's -pie by default if -nopie isn't specified
#debug=("-Wl,--no-threads" "-Wl,--debug=all")  #XXX uncomment of lots of seemingly useless(to this issue) information ?!
debug=()

##Won't compile without the following but at least one of them is required to see the 'discarded' type of errors(set $pickone below):
oops1=(
"/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-a9bb89774eb1c5f2.rlib"
"/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libpanic_unwind-01499d126af771f6.rlib"
"/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-1f1fadf3472d3239.rlib"
)
oops2=(
"/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-9e3a9887668f9c30.rlib"
)
if test "$success" != '1'; then
  pickone=1 #0,1,2
  oops1=("${oops1[$pickone]}")
  oops2=()
fi
/usr/bin/clang -v "${noeffect[@]}" -std=gnu99 "${debug[@]}" -lpthread -fuse-ld=gold -fcolor-diagnostics "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.build_script_build.ea30uw23-cgu.0.rcgu.o" "-o" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/build/proc-macro2-3b19a3eb863545be/build_script_build-3b19a3eb863545be.1zobs1pcbqt3rtjo.rcgu.o" "$err" "-nodefaultlibs" "-L" "/home/user/build/1packages/4used/firefox-hg/makepkg_pacman/firefox-hg/src/firefox-hg/obj-x86_64-pc-linux-gnu/release/deps" "-L" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-Wl,--start-group" "-Wl,-Bstatic" "${oops1[@]}" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libbacktrace_sys-8b152c95a80afef5.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_demangle-14c23aab33e6c5f4.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libunwind-41eb09b84b2625ab.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liblibc-44a20fe3a684e072.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liballoc-e5275540f2101631.rlib" "/home/user/build/2nonpkgs/rust.stuff/rust/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_core-9f560ba3dfa532fe.rlib" "-Wl,--end-group" "${oops2[@]}" "-Wl,-Bdynamic" "-ldl" "-lgcc_s" "-lc"

[vorner]

I don't think this is anything specific to firefox. I'm getting this error on today's nightly as well (rustc 1.35.0-nightly (96d700f1b 2019-04-10)). I don't set any linker flags manually as far as I know.

          /home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-40d980a809fdc985.rlib(std-40d980a809fdc985.std.2pwkdt1a-cgu.0.rcgu.o)(.stack_sizes+0x0): error: relocation refers to local symbol "" [1760], which is defined in a discarded section

But maybe some of the dependencies do something?

  = note: "cc" "-Wl,--as-needed" "-Wl,-z,noexecstack" "-m64" "-L" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.0.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.1.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.10.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.11.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.12.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.13.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.14.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.15.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.2.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.3.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.4.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.5.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.6.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.7.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.8.rcgu.o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.build_script_build.f5cx6kaq-cgu.9.rcgu.o" "-o" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc" "/home/vorner/.rust-target/normal/debug/build/num-traits-2ba80e41556ab8cc/build_script_build-2ba80e41556ab8cc.1i7z7qhcjdrul5p9.rcgu.o" "-Wl,--gc-sections" "-pie" "-Wl,-zrelro" "-Wl,-znow" "-nodefaultlibs" "-L" "/home/vorner/.rust-target/normal/debug/deps" "-L" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-Wl,--start-group" "-Wl,-Bstatic" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-40d980a809fdc985.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libpanic_unwind-2133eb2e78115669.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libbacktrace_sys-76a1e895b942f4f1.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_demangle-4423e7ad8fae78b0.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libunwind-121d18364f871b00.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/liblibc-684dc35735333b20.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/liballoc-65ae63f7a5616b63.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_core-c91f0330d41106a6.rlib" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-71db3eaafaac5eb0.rlib" "-Wl,--end-group" "/home/vorner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-138ab271264ca64d.rlib" "-Wl,-Bdynamic" "-ldl" "-lrt" "-lpthread" "-lgcc_s" "-lc" "-lm" "-lrt" "-lpthread" "-lutil" "-lutil"

Should I try finding some more info?

[pnkfelix]

@vorner any other details you can figure out would be appreciated.

Even just a standalone test (so that we didn't have to try to recreate this via a firefox build) would be immensely helpful.

[pnkfelix]

assigning to self to:

1. Attempt to identify a standalone test (or prod other people into providing one)
2. Reproduce the bug locally
3. Revert the offending PR if we can confirm that does indeed fix the problem.

[vorner]

• Create a new bin project
• Add this to Cargo.toml:

[dependencies]
dipstick = "0.7"

• Run cargo +nightly check

It seems to fail on building a C or C++ code somewhere in numtraits. Assuming it is using the cc from the system, this is mine:

$ cc --version
cc (Gentoo 8.3.0-r1 p1.1) 8.3.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

The default ld is:

$ ld --version
GNU gold (Gentoo 2.32 p1 2.32.0) 1.16
Copyright (C) 2019 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License version 3 or (at your option) a later version.
This program has absolutely no warranty.

I expect this problem can be triggered by other things too, dipstick is just the first dependency in my project that caused it.

[ghost]

tl;dr: can't trigger issue with dipstick 0.7.2 but can still trigger it with compiling firefox (used 3 different rustc compiler versions to test both disptick and firefox)

@vorner By using the same rustc compiler(master-stage1) that I used to trigger this issue with firefox, I cannot trigger it with a new project with dipstick = "0.7" and just cargo check. I immediately retried compiling firefox after that(with cleaned/removed build dir), to confirm that it still triggered the issue and it did.

but my cc and ld versions differ from yours:

$ cc --version
cc (GCC) 8.2.1 20181127
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ ld --version
GNU ld (GNU Binutils) 2.31.1
Copyright (C) 2018 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License version 3 or (at your option) a later version.
This program has absolutely no warranty.

cargo +nightly check shows error: no such subcommand: +nightly - not sure what I'm missing here,
but I then tried this instead: rustup default nightly (rustc 1.35.0-nightly (3750348da 2019-04-08)) and then cargo clean && cargo check and still couldn't trigger the issue(Finished dev [unoptimized + debuginfo] target(s) in 19.23s).
I even retried after a rustup update (rustc 1.35.0-nightly (96d700f1b 2019-04-10)) to no avail (Finished dev [unoptimized + debuginfo] target(s) in 19.25s).
Cargo.lock shows dipstick 0.7.2 was used.
I even tried cargo build and cargo run with this nightly.
Then I kept this latest nightly, cleaned/removed firefox build dir and tried recompiling firefox: issue is still triggered(7:29.38 /home/user/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-71db3eaafaac5eb0.rlib(core-71db3eaafaac5eb0.core.bpa42w6m-cgu.0.rcgu.o)(.stack_sizes+0x0): error: relocation refers to local symbol "" [855], which is defined in a discarded section)

In the dipstick console, after having done all of the above, I cannot find any numtraits text, however.
I even renamed ~/.cargo/registry/ (to emulate removal) and did a cargo clean&&cargo check (still using latest nightly): Finished dev [unoptimized + debuginfo] target(s) in 45.50s). Still no mention of numtraits. (EDIT2: firefox compilation with new build dir and this new registry still triggered the issue)
Oh, I just realized it's num_traits and there are plenty of referenced when -v is used, but still no errors and I cannot trigger this issue.

     Running `rustc --crate-name num_traits /home/user/.cargo/registry/src/github.com-1ecc6299db9ec823/num-traits-0.2.6/src/lib.rs --color always --crate-type lib --emit=dep-info,metadata -C debuginfo=2 -C metadata=f4779c1b3a57d428 -C extra-filename=-f4779c1b3a57d428 --out-dir /tmp/pro/target/debug/deps -L dependency=/tmp/pro/target/debug/deps --cap-lints allow -C target-cpu=native --cfg has_i128`

EDIT1: That output made me realize and remember that I've this ~/.cargo/config contents:

[target.'cfg(any(windows, unix))']                                              
rustflags = ["-C", "target-cpu=native"]
# src: https://users.rust-lang.org/t/auto-vectorization-in-rust/24379/4

EDIT3: no effect if I comment out the contents of ~/.cargo/config, on either dipstick 0.7.2 (clean, check, build, run)[doesn't trigger] or firefox[does trigger].

I even tried making sure Cargo.lock has dipstick 0.7.0 by setting =0.7.0 in Cargo.toml, issue won't trigger.

$ rustv
!! LD_LIBRARY_PATH=
!! Executing '/home/user/.cargo/bin/rustc' in pwd='/tmp/pro' with args: '-vV'
!! !MOZ_CARGO_WRAP_LD=''
!! !MOZ_CARGO_WRAP_LDFLAGS=''!
nightly-x86_64-unknown-linux-gnu (default)
rustc 1.35.0-nightly (96d700f1b 2019-04-10)
binary: rustc
commit-hash: 96d700f1b7bc9c53fa0d11567adb1ed2c1c27e79
commit-date: 2019-04-10
host: x86_64-unknown-linux-gnu
release: 1.35.0-nightly
LLVM version: 8.0
!! LD_LIBRARY_PATH=
!! Executing '/home/user/build/2nonpkgs/rust.stuff/cargo/cargo//target/release//cargo' in pwd='/tmp/pro' with args: '-vV'
!! !MOZ_CARGO_WRAP_LD=''
!! !MOZ_CARGO_WRAP_LDFLAGS=''!
nightly-x86_64-unknown-linux-gnu (default)
cargo 1.35.0-dev (f5b60ac5 2019-04-05)
release: 1.35.0
commit-hash: f5b60ac5cc61621f99ba702e8130af49020ea109
commit-date: 2019-04-05

[mati865]

Reproduced with @vorner example using RUSTFLAGS="-C link-arg=-fuse-ld=gold" cargo +nightly check.
Gold linker seems to be the culprit here.

[vorner]

I can confirm. If I switch my linker to ld.bfd, the problem goes away.

I have the vague impression that firefox somehow prefers to link with gold in its build system, but I might be wrong about that one.

[ghost]

Gold linker seems to be the culprit here.

Confirmed! Issue triggered for me too(even though I can't use the cargo +nightly check command), by $ cargo clean && RUSTFLAGS="-C link-arg=-fuse-ld=gold" cargo check
(tested 2 rustc compilers)

This is really cool, thanks @mati865 and @vorner for easy reproduction steps[1,2] that don't involve firefox!

EDIT:

I have the vague impression that firefox somehow prefers to link with gold in its build system, but I might be wrong about that one.

yes, firefox does use ld.gold indeed(but not by default! see edit3 below): ref0 ref1 ref2
EDIT2: Actually it might be my fault for explicitly setting it in mozconfig as ac_add_options --enable-gold
EDIT3 ./configure --help says --enable-gold Enable GNU Gold Linker when it is not already the default (that seems to imply it would be used by default) and commenting out my mozconfig line as #ac_add_options --enable-gold then recompiling firefox with empty build dir ... doesn't seem to cause the issue yet(30minutes into compilation, the issue would be caused around minute 7.5), and I see no mention of gold thus I can only assume it doesn't use gold by default! ergo, it's my fault for explicitly telling it to use it.

readelf output

[ghost]

So when using ld.gold linker with the implied --gc-sections that rust adds(when not a dynamic library crate?), coupled with the .stack_sizes linker section(added by PR https://github.com/rust-lang/rust/pull/59401) then the issue triggers (error: relocation refers to local symbol "" [12], which is defined in a discarded section)

--gc-sections is used when crate_type != config::CrateType::Dylib as seen in this code:
https://github.com/rust-lang/rust/blob/05b4554e77ef32a97060b1bfffd1f621bf15bb88/src/librustc_codegen_llvm/back/link.rs#L908-L909
--gc-sections is added by this code:
https://github.com/rust-lang/rust/blob/3750348daff89741e3153e0e120aa70a45ff5b68/src/librustc_codegen_ssa/back/linker.rs#L297-L304

[ghost]

@vorner How did you have ld.gold set in https://github.com/rust-lang/rust/issues/59652#issuecomment-482141460 ? I assume you had to have had it set since I couldn't reproduce without setting it in RUSTFLAGS. Was it set in ~/.cargo/config or via RUSTFLAGS ? something else?

though you do mention that

The default ld is:

not ld-gold but a normal ld which is a higher version(2.32 p1 2.32.0) than mine(2.31.1). Should I expect then that my future ld version(whenever ArchLinux update binutils package - currently at 2.31.1-4) will break just like ld.gold did and thus trigger this issue without needing ld.gold ?

oh wait, it does say gold in your output(weird that I only see that now!!):

$ ld --version
GNU gold (Gentoo 2.32 p1 2.32.0) 1.16

that was unexpected, here's how it looks on ArchLinux:

$ ld --version
GNU ld (GNU Binutils) 2.31.1
Copyright (C) 2018 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License version 3 or (at your option) a later version.
This program has absolutely no warranty.

$ ld.gold --version
GNU gold (GNU Binutils 2.31.1) 1.16
Copyright (C) 2018 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License version 3 or (at your option) a later version.
This program has absolutely no warranty.

[vorner]

Well, gentoo allows you to tweak the system a lot and you do a lot of compilation with that system. I discovered gold is a bit faster in most cases, so I set it as the my system default linker. It's not the general default if you get a fresh gentoo system.

[ghost]

so it's a ld gold bug: https://sourceware.org/bugzilla/show_bug.cgi?id=23880

other potentially interesting ones:
https://sourceware.org/bugzilla/show_bug.cgi?id=14330
https://sourceware.org/bugzilla/show_bug.cgi?id=14861
https://sourceware.org/bugzilla/show_bug.cgi?id=17670
https://sourceware.org/bugzilla/show_bug.cgi?id=19842
https://sourceware.org/bugzilla/show_bug.cgi?id=10400
https://sourceware.org/bugzilla/show_bug.cgi?id=24191
https://sourceware.org/bugzilla/show_bug.cgi?id=11108
https://sourceware.org/bugzilla/show_bug.cgi?id=12488

[ghost]

I have another workaround, for those who still want to use ld.gold but don't mind recompiling it(binutils), which I've tested to pass dipstick and firefox compilations. I used the following patch. (Then I recompiled binutils without the patch to test that both dipstick and firefox compilations fail without it.)

diff --git a/gold/target-reloc.h b/gold/target-reloc.h
index 97e45da619..20a6fc0542 100644
--- a/gold/target-reloc.h
+++ b/gold/target-reloc.h
@@ -136,6 +136,8 @@ class Default_comdat_behavior
     if (Layout::is_debug_info_section(name))
       return CB_PRETEND;
     if (strcmp(name, ".eh_frame") == 0
+        || strncmp(name, ".stack_sizes", 12+1) == 0
+        //|| strncmp(name, ".gnu.build.attributes", 21) == 0  // FIXME: We should really be checking the section type for ST_NOTE... // NOTE: this line(and inspiration for the above .stack_sizes line) from: https://bugzilla.redhat.com/show_bug.cgi?id=1600431#c6 ) // decided to comment out this line since I cannot repro that bug that it supposedly fixes with GNU gold (GNU Binutils 2.31.1) 1.16 or with GNU gold (GNU Binutils 2.32) 1.16
    || strcmp(name, ".gcc_except_table") == 0)
       return CB_IGNORE;
     return CB_ERROR;

So, is there something that Rust can do here or shall this be solely ld.gold's issue? if the latter, please close the issue, I guess.

I'm personally gonna be using this patch, for fun :))

For completion, other workarounds were:

• https://github.com/rust-lang/rust/issues/59652#issuecomment-479654606
• not using ld.gold linker
• somehow avoiding --gc-sections from being passed to ld.gold linker, code which passes it is here: https://github.com/rust-lang/rust/issues/59652#issuecomment-482253086

[pnkfelix]

Okay, thank you @vorner, that is very helpful.

Just to share knowledge, I don't see the same problem with the dipstick test on my Linux host; here are my versions of cc and ld:

% cc --version
cc (GCC) 8.2.1 20181105 (Red Hat 8.2.1-5)
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

% ld --version
GNU ld version 2.31.1-13.fc29
Copyright (C) 2018 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License version 3 or (at your option) a later version.
This program has absolutely no warranty.

I'll see if I can switch either of those around on this box in order to try to replicate the bug.

Update: Ooops, I should have reloaded the ticket before posting, there's been 16 hours of comments since @vorner posted.

[pnkfelix]

Okay so my next question is what action to take at this point.

• First of all, the change to -Z emit-stack-sizes was part of the recent nightly-to-beta promotion. So this ticket is now a regression from stable-to-beta. I'm going to change the labels accordingly.
• So an obvious change is to just remove the -Z emit-stack-sizes` emission (i.e. revert PR #59401, as suggested by @japaric above), unconditionally. I'm inclined to at least do that on the beta channel right now, to ensure that the regression is limited to nightly, not beta.
• The next question is whether to also remove that flag unconditionally on nightly too, or if we should somehow connect it to the choice of linker.
• Of course another option is to just say "this is a bug in ld.gold; we are not going to try to work around it on our end."
• (We could take an alternative route of trying to avoid injecting --gc-sections when it will break things. That would require, I think, inspecting whether the already-built libtest/libstd has a stack_sizes section when we are building up our linker invocation. That, or assuming it has such a section whenever we are on the target that happens to match the strings as is done in PR #59401's bootstrap changes.)

[pnkfelix]

I've posted a PR to do the revert. Rather than solely cherry-pick it to beta, I have posted it against master (that is, rather than let the issue persist on the nightly builds, I want the behavior on nightly to match that of beta here).

I do welcome a more nuanced implementation, if someone wants to make one, for nightly. But I want to deploy the simplest way to address the regression on both master and beta right now. (And that's what PR #59911 is.)

[ghost]

Closing due to https://github.com/rust-lang/rust/commit/633fc9eef0d26a1c9b59fad8e807632f234a8aae having landed