[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAKwvOdmsNwszu4Vk1K7e39hencNcjEmjr5q7EbTJCjdY5TDsFA@mail.gmail.com>
Date: Tue, 19 Sep 2023 12:44:14 -0700
From: Nick Desaulniers <ndesaulniers@...gle.com>
To: Matthew Maurer <mmaurer@...gle.com>
Cc: Miguel Ojeda <ojeda@...nel.org>,
Alex Gaynor <alex.gaynor@...il.com>,
Wedson Almeida Filho <wedsonaf@...il.com>,
Masahiro Yamada <masahiroy@...nel.org>,
Boqun Feng <boqun.feng@...il.com>, Gary Guo <gary@...yguo.net>,
Björn Roy Baron <bjorn3_gh@...tonmail.com>,
Benno Lossin <benno.lossin@...ton.me>,
Andreas Hindborg <a.hindborg@...sung.com>,
Alice Ryhl <aliceryhl@...gle.com>,
Nathan Chancellor <nathan@...nel.org>,
Nicolas Schier <nicolas@...sle.eu>,
rust-for-linux@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-kbuild@...r.kernel.org
Subject: Re: [PATCH v2] rust: Respect HOSTCC when linking for host
On Mon, Sep 18, 2023 at 4:44 PM Matthew Maurer <mmaurer@...gle.com> wrote:
>
> Currently, rustc defaults to invoking `cc`, even if `HOSTCC` is defined,
> resulting in build failures in hermetic environments where `cc` does not
> exist. This includes both hostprogs and proc-macros.
>
> Since we are setting the linker to `HOSTCC`, we set the linker flavor to
> `gcc` explicitly.
>
> Signed-off-by: Matthew Maurer <mmaurer@...gle.com>
> ---
>
> Updated the patch to reflect Nick's comment that KBUILD_HOSTLDFLAGS
> should be respected as well.
>
> I did not switch it to use HOSTLD for two reasons:
> * That variable is not globally defined - it is only available in two
> subdirectories of tools/
> * C host scripts are linked by HOSTCC as well, even when linking a
> collection of object files. It *prints* HOSTLD, but invokes HOSTCC.
> See scripts/Makefile.host cmd_host-cmulti for an example.
Sure, that makes sense to me, thanks for pointing that out.
>
> rust/Makefile | 4 ++++
> scripts/Makefile.host | 4 ++++
> 2 files changed, 8 insertions(+)
>
> diff --git a/rust/Makefile b/rust/Makefile
> index 87958e864be0..b60b7eb8c5a0 100644
> --- a/rust/Makefile
> +++ b/rust/Makefile
> @@ -380,9 +380,13 @@ $(obj)/exports_bindings_generated.h: $(obj)/bindings.o FORCE
> $(obj)/exports_kernel_generated.h: $(obj)/kernel.o FORCE
> $(call if_changed,exports)
>
> +KBUILD_HOSTLDFLAGS_SQ = '$(subst ','\'',$(KBUILD_HOSTLDFLAGS))'
I don't think we need to do this kind of escaping. If a linker flag is
passed to the linker directly as the driver has spaces such as `-z
relro`, if it is instead passed to the compiler as the driver will not
contain spaces (`-Wl,-z,relro`). As such, I don't think we need this
escaping (famous last words). Mind submitting a v3 without it?
> +
> quiet_cmd_rustc_procmacro = $(RUSTC_OR_CLIPPY_QUIET) P $@
> cmd_rustc_procmacro = \
> $(RUSTC_OR_CLIPPY) $(rust_common_flags) \
> + -Clinker-flavor=gcc -Clinker=$(HOSTCC) \
> + -Clink-args=$(KBUILD_HOSTLDFLAGS_SQ) \
> --emit=dep-info=$(depfile) --emit=link=$@ --extern proc_macro \
> --crate-type proc-macro \
> --crate-name $(patsubst lib%.so,%,$(notdir $@)) $<
> diff --git a/scripts/Makefile.host b/scripts/Makefile.host
> index 8f7f842b54f9..dc0410cae5ca 100644
> --- a/scripts/Makefile.host
> +++ b/scripts/Makefile.host
> @@ -87,10 +87,14 @@ hostcxx_flags = -Wp,-MMD,$(depfile) \
> $(KBUILD_HOSTCXXFLAGS) $(HOST_EXTRACXXFLAGS) \
> $(HOSTCXXFLAGS_$(target-stem).o)
>
> +KBUILD_HOSTLDFLAGS_SQ = '$(subst ','\'',$(KBUILD_HOSTLDFLAGS))'
> +
> # `--out-dir` is required to avoid temporaries being created by `rustc` in the
> # current working directory, which may be not accessible in the out-of-tree
> # modules case.
> hostrust_flags = --out-dir $(dir $@) --emit=dep-info=$(depfile) \
> + -Clinker-flavor=gcc -Clinker=$(HOSTCC) \
> + -Clink-args=$(KBUILD_HOSTLDFLAGS_SQ) \
> $(KBUILD_HOSTRUSTFLAGS) $(HOST_EXTRARUSTFLAGS) \
> $(HOSTRUSTFLAGS_$(target-stem))
>
> --
> 2.42.0.459.ge4e396fd5e-goog
>
--
Thanks,
~Nick Desaulniers
Powered by blists - more mailing lists