Pull Kbuild/Kconfig updates from Nicolas Schier:
 "Kbuild updates:

   - Use --force-group-allocation when linking modules

     Have the linker resolve the COMDAT groups and place their members
     as regular sections instead of possibly leaving multiple copies in
     the resulting modules and unnecessary group metadata.

   - UAPI header files: Canonicalize __ASSEMBLER__ / __ASSEMBLY__ mixed
     use to __ASSEMBLER__

     There is an ongoing effort to change __ASSEMBLY__ to __ASSEMBLER__
     treewide. For consistency, UAPI headers are normalised to use
     __ASSEMBLER__ only. Normalisation is done in two subsequent patches
     to simplify a revert in the unexpected case of a regression report.

   - link-vmlinux.sh: Improve detection of third pass requirement

   - modpost: Canonicalize format of warnings and errors

   - Minor changes:
      - Remove srctree path from CHECK output
      - Set the initial value of subdir-rustflags-y
      - Remove broken and unused modules.builtin(.modinfo) targets from
        the top-level Makefile
      - Add symbol size for kallsyms symbols that can change size
      - modpost: Prevent leak when early return no suffix .o in
        read_symbols()
      - scripts/config: Update usage of POSIX sed
      - 'make tags': Add support for rust source files and prevent
        binary files from being analysed
      - Several spelling mistakes and rephrasing

  Kconfig updates:

   - Add Julian Braha as Kconfig reviewer

   - Fix submenu rendering of negative dependencies

   - Minor changes:
      - merge_config.sh: Keep temp file in the output dir
      - Abort rather than loop for ever on EOF"

* tag 'kbuild-7.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux: (23 commits)
  modpost: use mod_warn() and mod_error(), clean up logging
  modpost: add module as parameter to modpost_log()
  kconfig: fix submenu rendering of negative dependencies
  kbuild: link-vmlinux.sh: improve detection of third pass requirement
  kallsyms: add symbol size for kallsyms symbols that can change size
  kbuild: fix modules.builtin(.modinfo) targets in the top-level Makefile
  kbuild: set the initial value of subdir-rustflags-y
  scripts/config: Use in-place editing (-i) in sed portably
  scripts/config: Use POSIX standard ERE (-E) in sed
  modpost: prevent leak when early return no suffix .o in read_symbols()
  usr: Correct a spelling by changing a letter
  fixdep: make gendered language gender-neutral
  kconfig: fix minor typos in comments
  scripts: fix spelling mistakes
  kconfig: abort rather than loop for ever on EOF
  scripts/tags.sh: Add support for rust source files
  scripts/tags.sh: Prevent binary files appearing in cscope.files
  MAINTAINERS: add Julian Braha as Kconfig reviewer
  scripts: headers_install.sh: Normalize __ASSEMBLY__ to __ASSEMBLER__
  scripts: headers_install.sh: Normalize __ASSEMBLER__ to __ASSEMBLY__
  ...

Commit 7a342e6c7735 ("kbuild: move modules.builtin(.modinfo) rules to
Makefile.vmlinux_o") moved the modules.builtin(.modinfo) rules from
link-vmlinux.sh to Makefile.vmlinux_o, and added the corresponding
targets to the top-level Makefile.

Commit 39cfd5b12160 ("kbuild: extract modules.builtin.modinfo from
vmlinux.unstripped") later moved these rules from Makefile.vmlinux_o to
Makefile.vmlinux, but left the corresponding targets in the top-level
Makefile unchanged.

These modules.builtin(.modinfo) targets in the top-level Makefile should
be moved alongside the vmlinux target, since they are now generated by
Makefile.vmlinux.

However, simply removing these trivial targets might be a better choice,
as it makes the Makefile cleaner and avoids the need to keep them in sync
across multiple files, reducing the chance of future mistakes.

Fixes: 39cfd5b12160 ("kbuild: extract modules.builtin.modinfo from vmlinux.unstripped")
Signed-off-by: Yuntao Wang <yuntao.wang@linux.dev>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Link: https://patch.msgid.link/20260729071737.818007-1-yuntao.wang@linux.dev
Signed-off-by: Nicolas Schier <nsc@kernel.org>

Pull Kbuild fixes from Nathan Chancellor:

 - Fix regression with MO= when building out of tree kernel modules due
   to incorrectly overwriting build tree's Makefile

 - Avoid stripping .BTF sections from modules when building debug .rpm
   packages

* tag 'kbuild-fixes-7.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux:
  kbuild: rpm-pkg: Preserve BTF sections in kernel modules during debuginfo stripping
  kbuild: Stop modifying $(objtree)/Makefile when building oot-kmods oos

Pull rust fixes from Miguel Ojeda:
 "Toolchain and infrastructure:

   - 'zerocopy' crates: update to v0.8.54 to fix a modpost error under
     'CONFIG_CC_OPTIMIZE_FOR_SIZE=y'.

     There are actually two updates in the PR: the one to v0.8.52 is
     fairly large and was originally not intended for a fixes PR, but the
     actual fix landed in the v0.8.54 one. Thus I included both here.

     The v0.8.52 update includes two things upstream added for us:
     '--cfg no_fp_fmt_parse' to avoid a local workaround, and the new
     'most_traits' feature.

     The good news is that, after these updates, the delta with upstream
     is now trivial: only an identifier prefix change and the SPDX
     parentheses.

   - Fix an objtool warning by adding one more 'noreturn' function for
     Rust 1.99.0 (expected 2026-10-01).

   - Clean up new 'semicolon_in_expressions_from_macros' lint errors for
     Rust 1.99.0 (expected 2026-10-01). The lint can be allowed, but it
     will be a hard error at some point in the future anyway, so clean it
     up now.

   - Locally allow new 'suspicious_runtime_symbol_definitions' lint for
     Rust 1.98.0 (expected 2026-08-20).

   - Globally allow 'clippy::unwrap_or_default' lint since it relies on
     optimizations -- under 'CONFIG_CC_OPTIMIZE_FOR_SIZE=y' it does not
     work well.

  'kernel' crate:

   - 'time' module: fix 'Delta::as_micros_ceil()' to round negative values
     correctly"

* tag 'rust-fixes-7.2-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux:
  rust: time: fix as_micros_ceil() to round correctly for negative Delta
  rust: device: avoid trailing ; in printing macros
  objtool/rust: add one more `noreturn` Rust function for Rust 1.99.0
  rust: zerocopy: update to v0.8.54
  rust: zerocopy: update to v0.8.52
  rust: allow `clippy::unwrap_or_default` globally
  rust: allow `suspicious_runtime_symbol_definitions` lint for Rust >= 1.98

Update output Makefile in the corresponding tree only: for out-of-source
in-tree builds update $(objtree)/Makefile, for out-of-source out-of-tree
module builds update $(KBUILD_EXTMOD_OUTPUT)/Makefile instead.

In-source builds are not affected.

Since commit c9bb03ac2c66 ("kbuild: reduce output spam when building out
of tree"), building out-of-tree kernel modules out-of-source (make M=...
MO=...) causes a rewrite of $(objtree)/Makefile with KBUILD_EXTMOD and
KBUILD_EXTMOD_OUTPUT being set.  That is problematic:

 * $(objtree)/ must not be changed in any way when building out-of-tree
   modules as it breaks other uses of $(objtree).

 * Setting KBUILD_EXTMOD and KBUILD_EXTMOD_OUTPUT in $(objtree)/Makefile
   kills the tree for incremental builds that start right there
   ('make -C $(objtree)'); builds starting in $(srctree) reset
   $(objtree)/Makefile to its original content.

Further, $(KBUILD_EXTMOD_OUTPUT)/Makefile was not generated any more at
all.

This commit restores the previous kbuild behaviour prior to commit
c9bb03ac2c66 ("kbuild: reduce output spam when building out of tree")
but leaves in-place the use of filechk for output spam reduction.

Fixes: c9bb03ac2c66 ("kbuild: reduce output spam when building out of tree")
Reported-by: Anish Rashinkar <rashinkar.anish@gmail.com>
Closes: https://lore.kernel.org/r/CAOESE2Q2-0KUDaM0mUo+c_F-tMaUsBZ-gpnhdoe0rmYdgnnuJQ@mail.gmail.com
Cc: stable@vger.kernel.org
Tested-by: Philipp Hahn <p.hahn@avm.de>
Reviewed-by: Philipp Hahn <p.hahn@avm.de>
Signed-off-by: Nicolas Schier <n.schier@fritz.com>
Signed-off-by: Nicolas Schier <nsc@kernel.org>
Link: https://patch.msgid.link/20260723105845.1704689-2-nsc@kernel.org
Signed-off-by: Nathan Chancellor <nathan@kernel.org>