Re: [PATCH V11 1/4] rust: Fix "multiple candidates for rmeta dependency core" error

[Link Mauve <linkmauve@linkmauve.fr>]

On Tue, Apr 21, 2026 at 05:49:32PM +0530, Mukesh Kumar Chaurasiya wrote:
> On Tue, Apr 21, 2026 at 12:26:51PM +0200, Link Mauve wrote:
> > On Tue, Apr 21, 2026 at 03:25:22PM +0530, Mukesh Kumar Chaurasiya wrote:
> > > On Tue, Apr 21, 2026 at 11:38:08AM +0200, Link Mauve wrote:
> > > > Hi Mukesh,
> > > > 
> > > > This patch doesn’t apply on top of mainline, which tree did you base it
> > > > off?
> > > > 
> > > It was on mainline v7.0 tag.
> > 
> > Great thanks, they do apply there!
> > 
> > I needed three more patches for the kernel to build on PPC32, I’ve
> > attached them but they are absolutely not patches which could go into
> > the kernel (except for the second, enabling asm_experimental_arch).
> > 
> > What do you think we should do about them?
> > 
> Regarding the 2nd patch you sent, As the support is experimental as of
> now, Let's wait till we get everything stablized, till then we'll use
> nightly build for all variants of powerpc. Once we mark it as
> maintained, then we can try to push that patch.

Enabling the feature is mandatory even on nightly or with
RUSTC_BOOTSTRAP=1, otherwise no asm!() macro can be called and so the
kernel crate can’t be built.

> 
> Maybe Miguel can comment better on this.
> 
> Regarding other patches, I am planning to work on ppc32 once ppc64le
> stabilizes.

Thanks!

> 
> Regards,
> Mukesh
> 
> > > On latest there is a confilict. I'll send out a rebased version.
> > > 
> > > > On Fri, Apr 17, 2026 at 08:52:50PM +0530, Mukesh Kumar Chaurasiya (IBM) wrote:
> > > > > When building Rust code for powerpc64le with LLVM=1 and -j1, rustc
> > > > > encounters an error: "multiple candidates for `rmeta` dependency `core`
> > > > > found", with two candidates:
> > > > > 1. The host's standard library from the rustup toolchain
> > > > > 2. The kernel's custom libcore.rmeta in the rust/ directory
> > > > > 
> > > > > This occurs because the build system uses `-L$(objtree)/rust` for host
> > > > > library builds (proc_macro2, quote, syn), which causes rustc to search
> > > > > the rust/ directory. During this search, rustc finds both the kernel's
> > > > > custom libcore.rmeta and gains access to the host's standard library,
> > > > > creating a conflict.
> > > > > 
> > > > > The solution is to separate host libraries into a dedicated rust/host/
> > > > > subdirectory and use `-L$(objtree)/rust/host` for host builds instead
> > > > > of `-L$(objtree)/rust`. This ensures that:
> > > > > 
> > > > > 1. Host library builds (proc_macro2, quote, syn) only search rust/host/
> > > > >    and never encounter the kernel's libcore.rmeta
> > > > > 2. Proc macro builds use `-L$(objtree)/rust/host` to find their
> > > > >    dependencies
> > > > > 
> > > > > Special handling is added for rustdoc-pin_init, which is a host build
> > > > > (to access the alloc crate) but depends on proc macros from the main
> > > > > rust/ directory. It uses explicit `--extern` paths to reference the
> > > > > proc macros without adding `-L$(objtree)/rust`, which would reintroduce
> > > > > the conflict.
> > > > > 
> > > > > The rust/host/ directory is added to clean-files to ensure it's removed
> > > > > during `make clean`.
> > > > > 
> > > > > Link: https://github.com/Rust-for-Linux/linux/issues/105
> > > > > Link: https://github.com/linuxppc/issues/issues/451
> > > > > Signed-off-by: Mukesh Kumar Chaurasiya (IBM) <mkchauras@gmail.com>
> > > > > ---
> > > > >  rust/Makefile | 38 +++++++++++++++++++++-----------------
> > > > >  1 file changed, 21 insertions(+), 17 deletions(-)
> > > > > 
> > > > > diff --git a/rust/Makefile b/rust/Makefile
> > > > > index 9801af2e1e02..e234b8a39358 100644
> > > > > --- a/rust/Makefile
> > > > > +++ b/rust/Makefile
> > > > > @@ -3,6 +3,9 @@
> > > > >  # Where to place rustdoc generated documentation
> > > > >  rustdoc_output := $(objtree)/Documentation/output/rust/rustdoc
> > > > >  
> > > > > +# Clean generated host directory
> > > > > +clean-files := host/
> > > > > +
> > > > >  obj-$(CONFIG_RUST) += core.o compiler_builtins.o ffi.o
> > > > >  always-$(CONFIG_RUST) += exports_core_generated.h
> > > > >  
> > > > > @@ -27,7 +30,7 @@ endif
> > > > >  
> > > > >  obj-$(CONFIG_RUST) += exports.o
> > > > >  
> > > > > -always-$(CONFIG_RUST) += libproc_macro2.rlib libquote.rlib libsyn.rlib
> > > > > +always-$(CONFIG_RUST) += host/libproc_macro2.rlib host/libquote.rlib host/libsyn.rlib
> > > > >  
> > > > >  always-$(CONFIG_RUST_KERNEL_DOCTESTS) += doctests_kernel_generated.rs
> > > > >  always-$(CONFIG_RUST_KERNEL_DOCTESTS) += doctests_kernel_generated_kunit.c
> > > > > @@ -150,7 +153,7 @@ quiet_cmd_rustdoc = RUSTDOC $(if $(rustdoc_host),H, ) $<
> > > > >  	OBJTREE=$(abspath $(objtree)) \
> > > > >  	$(RUSTDOC) $(filter-out $(skip_flags) --remap-path-prefix=% --remap-path-scope=%, \
> > > > >  			$(if $(rustdoc_host),$(rust_common_flags),$(rust_flags))) \
> > > > 
> > > > The issue is here ↑, --remap-path-prefix=% got removed and the two
> > > > previous lines got merged into one.
> > > > 
> > > Yeah will fix this.
> > > > > -		$(rustc_target_flags) -L$(objtree)/$(obj) \
> > > > > +		$(rustc_target_flags) -L$(objtree)/$(obj)$(if $(rustdoc_host),/host) \
> > > > >  		-Zunstable-options --generate-link-to-definition \
> > > > >  		--output $(rustdoc_output) \
> > > > >  		--crate-name $(subst rustdoc-,,$@) \
> > > > […]
> > > > 
> > > > Thanks anyway for iterating on this series!  I’ve started writing a DRM
> > > > driver based off a previous version. :)
> > > > 
> > > That's cool. Thanks.
> > > 
> > > Regards,
> > > Mukesh
> > > > -- 
> > > > Link Mauve
> > > > 
> > 
> > -- 
> > Link Mauve
> 
> > From 79b614dd7be16a1108e1f0176a7bc235fa163a4e Mon Sep 17 00:00:00 2001
> > From: Link Mauve <linkmauve@linkmauve.fr>
> > Date: Wed, 8 Apr 2026 14:16:04 +0200
> > Subject: [PATCH 1/3] XXX: Workaround sstep.c
> > 
> > These variables are used only on CONFIG_PPC64, which causes compilation
> > to fail on PPC32.
> > ---
> >  arch/powerpc/lib/sstep.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/arch/powerpc/lib/sstep.c b/arch/powerpc/lib/sstep.c
> > index ac3ee19531d8..dc129f0cb6a0 100644
> > --- a/arch/powerpc/lib/sstep.c
> > +++ b/arch/powerpc/lib/sstep.c
> > @@ -1354,15 +1354,15 @@ int analyse_instr(struct instruction_op *op, const struct pt_regs *regs,
> >  #ifdef CONFIG_PPC64
> >  	unsigned int suffixopcode, prefixtype, prefix_r;
> >  #endif
> > -	unsigned int opcode, ra, rb, rc, rd, spr, u;
> > +	unsigned int opcode, ra, rb/*, rc*/, rd, spr, u;
> >  	unsigned long int imm;
> >  	unsigned long int val, val2;
> >  	unsigned int mb, me, sh;
> > -	unsigned int word, suffix;
> > +	unsigned int word/*, suffix*/;
> >  	long ival;
> >  
> >  	word = ppc_inst_val(instr);
> > -	suffix = ppc_inst_suffix(instr);
> > +	//suffix = ppc_inst_suffix(instr);
> >  
> >  	op->type = COMPUTE;
> >  
> > @@ -1480,7 +1480,7 @@ int analyse_instr(struct instruction_op *op, const struct pt_regs *regs,
> >  	rd = (word >> 21) & 0x1f;
> >  	ra = (word >> 16) & 0x1f;
> >  	rb = (word >> 11) & 0x1f;
> > -	rc = (word >> 6) & 0x1f;
> > +	//rc = (word >> 6) & 0x1f;
> >  
> >  	switch (opcode) {
> >  #ifdef __powerpc64__
> > -- 
> > 2.54.0
> > 
> 
> > From d4e72fad43175bde99287b8efdece900e5b53444 Mon Sep 17 00:00:00 2001
> > From: Link Mauve <linkmauve@linkmauve.fr>
> > Date: Fri, 10 Apr 2026 13:05:24 +0200
> > Subject: [PATCH 2/3] XXX: Enable asm_experimental_arch for PowerPC asm!()
> > 
> > This is needed to compile the kernel crate, otherwise this error
> > happens:
> > 
> > error[E0658]: inline assembly is not stable yet on this architecture
> >   --> ../rust/kernel/sync/barrier.rs:19:14
> >    |
> > 19 |     unsafe { core::arch::asm!("") };
> >    |              ^^^^^^^^^^^^^^^^^^^^
> >    |
> >    = note: see issue #93335 <https://github.com/rust-lang/rust/issues/93335> for more information
> >    = help: add `#![feature(asm_experimental_arch)]` to the crate attributes to enable
> >    = note: this compiler was built on 2026-03-25; consider upgrading it if it is out of date
> > 
> > error: aborting due to 1 previous error
> > 
> > For more information about this error, try `rustc --explain E0658`.
> > ---
> >  rust/kernel/lib.rs     | 3 +++
> >  scripts/Makefile.build | 2 +-
> >  2 files changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
> > index d93292d47420..92ccd47dc3ee 100644
> > --- a/rust/kernel/lib.rs
> > +++ b/rust/kernel/lib.rs
> > @@ -47,6 +47,9 @@
> >  // To be determined.
> >  #![feature(used_with_arg)]
> >  //
> > +// Needed for PowerPC inline assembly.
> > +#![feature(asm_experimental_arch)]
> > +//
> >  // `feature(derive_coerce_pointee)` is expected to become stable. Before Rust
> >  // 1.84.0, it did not exist, so enable the predecessor features.
> >  #![cfg_attr(CONFIG_RUSTC_HAS_COERCE_POINTEE, feature(derive_coerce_pointee))]
> > diff --git a/scripts/Makefile.build b/scripts/Makefile.build
> > index 3652b85be545..c7cc21994c5a 100644
> > --- a/scripts/Makefile.build
> > +++ b/scripts/Makefile.build
> > @@ -321,7 +321,7 @@ $(obj)/%.lst: $(obj)/%.c FORCE
> >  #
> >  # Please see https://github.com/Rust-for-Linux/linux/issues/2 for details on
> >  # the unstable features in use.
> > -rust_allowed_features := asm_const,asm_goto,arbitrary_self_types,lint_reasons,offset_of_nested,raw_ref_op,slice_ptr_len,strict_provenance,used_with_arg
> > +rust_allowed_features := asm_const,asm_goto,asm_experimental_arch,arbitrary_self_types,lint_reasons,offset_of_nested,raw_ref_op,slice_ptr_len,strict_provenance,used_with_arg
> >  
> >  # `--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
> > -- 
> > 2.54.0
> > 
> 
> > From 2c0a3ec3da6fa1f0151225c05159f7a812317d32 Mon Sep 17 00:00:00 2001
> > From: Link Mauve <linkmauve@linkmauve.fr>
> > Date: Fri, 10 Apr 2026 13:51:24 +0200
> > Subject: [PATCH 3/3] XXX: Workaround for __udivdi3() and __umoddi3()
> > MIME-Version: 1.0
> > Content-Type: text/plain; charset=UTF-8
> > Content-Transfer-Encoding: 8bit
> > 
> > The core crate currently depends on these two functions for i64/u64/
> > i128/u128/core::time::Duration formatting, but we shouldn’t use that in
> > the kernel so let’s panic if they are ever called.
> > ---
> >  rust/exports.c | 9 +++++++++
> >  1 file changed, 9 insertions(+)
> > 
> > diff --git a/rust/exports.c b/rust/exports.c
> > index 587f0e776aba..5f1cdf13882e 100644
> > --- a/rust/exports.c
> > +++ b/rust/exports.c
> > @@ -12,6 +12,7 @@
> >   */
> >  
> >  #include <linux/export.h>
> > +#include <linux/panic.h>
> >  
> >  #define EXPORT_SYMBOL_RUST_GPL(sym) extern int sym; EXPORT_SYMBOL_GPL(sym)
> >  
> > @@ -20,6 +21,14 @@
> >  #include "exports_bindings_generated.h"
> >  #include "exports_kernel_generated.h"
> >  
> > +void __udivdi3(void) {
> > +	panic("__udivdi3() called but shouldn’t be made available on this architecture!\n");
> > +}
> > +
> > +void __umoddi3(void) {
> > +	panic("__umoddi3() called but shouldn’t be made available on this architecture!\n");
> > +}
> > +
> >  // For modules using `rust/build_error.rs`.
> >  #ifdef CONFIG_RUST_BUILD_ASSERT_ALLOW
> >  EXPORT_SYMBOL_RUST_GPL(rust_build_error);
> > -- 
> > 2.54.0
> > 
> 
> 

-- 
Link Mauve