[PATCH] rust: kbuild: skip gendwarfksyms in `bindings.o` for Rust >= 1.91.0

[PATCH] rust: kbuild: skip gendwarfksyms in `bindings.o` for Rust >= 1.91.0

[Miguel Ojeda]

Starting with Rust 1.91.0 (released 2025-10-30), in upstream commit
ab91a63d403b ("Ignore intrinsic calls in cross-crate-inlining cost model")
[1][2], `bindings.o` stops containing DWARF debug information because the
`Default` implementations contained `write_bytes()` calls which are now
ignored in that cost model (note that `CLIPPY=1` does not reproduce it).

This means `gendwarfksyms` complains:

      RUSTC L rust/bindings.o
    error: gendwarfksyms: process_module: dwarf_get_units failed: no debugging information?

For the moment, conditionally skip `gendwarfksyms` for Rust >= 1.91.0.

Cc: stable@vger.kernel.org # Needed in 6.12.y and later (Rust is pinned in older LTSs).
Reported-by: Haiyue Wang <haiyuewa@163.com>
Closes: https://lore.kernel.org/rust-for-linux/b8c1c73d-bf8b-4bf2-beb1-84ffdcd60547@163.com/
Link: https://github.com/rust-lang/rust/commit/ab91a63d403b0105cacd72809cd292a72984ed99 [1]
Link: https://github.com/rust-lang/rust/pull/145910 [2]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
---
 rust/Makefile | 1 +
 1 file changed, 1 insertion(+)

diff --git a/rust/Makefile b/rust/Makefile
index 3e545c1a0ff4..269bf7cf5b97 100644
--- a/rust/Makefile
+++ b/rust/Makefile
@@ -543,6 +543,7 @@ $(obj)/ffi.o: private skip_gendwarfksyms = 1
 $(obj)/ffi.o: $(src)/ffi.rs $(obj)/compiler_builtins.o FORCE
 	+$(call if_changed_rule,rustc_library)
 
+$(obj)/bindings.o: private skip_gendwarfksyms := $(if $(call rustc-min-version,109100),1)
 $(obj)/bindings.o: private rustc_target_flags = --extern ffi --extern pin_init
 $(obj)/bindings.o: $(src)/bindings/lib.rs \
     $(obj)/ffi.o \

base-commit: dc77806cf3b4788d328fddf245e86c5b529f31a2
-- 
2.51.2

Re: [PATCH] rust: kbuild: skip gendwarfksyms in `bindings.o` for Rust >= 1.91.0

[Miguel Ojeda]

On Sat, Nov 8, 2025 at 2:44 AM Miguel Ojeda <ojeda@kernel.org> wrote:
>
> note that `CLIPPY=1` does not reproduce it

And obviously this means we have to special case this one too...

    $(obj)/bindings.o: private skip_gendwarfksyms := $(if $(call
rustc-min-version,109100),$(if $(KBUILD_CLIPPY),,1))

There may be other cases that disable the optimization or similar,
plus we may have other small crates in the future that could also
suffer from this, so it may be best to simply force to generate the
DWARF with a dummy symbol from that crate for the time being as the
fix:

    #[expect(unused)]
    static DUMMY_SYMBOL_FOR_DWARF_DEBUGINFO_GENERATION_FOR_GENDWARFKSYMS:
() = ();

With `#[no_mangle]` may be more reliable and it also gives an actual
exported symbol.

And then later do something in the `cmd` command itself or teaching
`genkallksyms` to auto-skip in cases like this.

What do you think?

Cheers,
Miguel

Re: [PATCH] rust: kbuild: skip gendwarfksyms in `bindings.o` for Rust >= 1.91.0

[Miguel Ojeda]

On Sat, Nov 8, 2025 at 3:30 AM Miguel Ojeda
<miguel.ojeda.sandonis@gmail.com> wrote:
>
> With `#[no_mangle]` may be more reliable and it also gives an actual
> exported symbol.

Or `#[used]` to keep it mangled since we don't care.

Cheers,
Miguel

Re: [PATCH] rust: kbuild: skip gendwarfksyms in `bindings.o` for Rust >= 1.91.0

[Sami Tolvanen]

On Fri, Nov 7, 2025 at 6:31 PM Miguel Ojeda
<miguel.ojeda.sandonis@gmail.com> wrote:
>
> On Sat, Nov 8, 2025 at 2:44 AM Miguel Ojeda <ojeda@kernel.org> wrote:
> >
> > note that `CLIPPY=1` does not reproduce it
>
> And obviously this means we have to special case this one too...
>
>     $(obj)/bindings.o: private skip_gendwarfksyms := $(if $(call
> rustc-min-version,109100),$(if $(KBUILD_CLIPPY),,1))
>
> There may be other cases that disable the optimization or similar,
> plus we may have other small crates in the future that could also
> suffer from this, so it may be best to simply force to generate the
> DWARF with a dummy symbol from that crate for the time being as the
> fix:
>
>     #[expect(unused)]
>     static DUMMY_SYMBOL_FOR_DWARF_DEBUGINFO_GENERATION_FOR_GENDWARFKSYMS:
> () = ();

This looks reasonable to me, and indeed seems to fix the problem.

> And then later do something in the `cmd` command itself or teaching
> `genkallksyms` to auto-skip in cases like this.
>
> What do you think?

Like I mentioned in the other thread, we ideally wouldn't silently
skip symbol versioning because the exported symbols won't be usable in
this case.

Sami