Warning options are enabled and disabled in inconsistent ways and
inconsistent locations. Start rearranging those by moving all options
into Makefile.extrawarn.

This should not change any behavior, but makes sure we can group them
in a way that ensures that each warning that got temporarily disabled
is turned back on at an appropriate W=1 level later on.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Tested with Qt5 5.15 and Qt6 6.4. Note that earlier versions of Qt5
are no longer guaranteed to work.
Signed-off-by: Boris Kolpackov <boris@codesynthesis.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Menuconfig has a feature where you can "press the key in the (#) prefix
to jump directly to that location. You will be returned to the current
search results after exiting this new menu."

This feature is poorly documented,
so add it to the kconfig.rst documentation.
Signed-off-by: Jesse Taube <Mr.Bossman075@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Menuconfig has a feature where you can "press the key in the (#) prefix
to jump directly to that location. You will be returned to the current
search results after exiting this new menu."

This commit adds this feature to nconfig, with almost identical code.
Signed-off-by: Jesse Taube <Mr.Bossman075@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Commit 6ab7e1f9 ("setlocalversion: use only the correct release
tag for git-describe") was absolutely correct to limit which annotated
tags would be used to compute the -01234-gabcdef suffix. Otherwise, if
some random annotated tag exists closer to HEAD than the vX.Y.Z one,
the commit count would be too low.

However, since the version string always includes the
${file_localversion} part, now the problem is that the count can be
too high. For example, building an 6.4.6-rt8 kernel with a few patches
on top, I currently get

$ make -s kernelrelease
6.4.6-rt8-00128-gd78b7f406397

But those 128 commits include the 100 commits that are in
v6.4.6..v6.4.6-rt8, so this is somewhat misleading.

Amend the logic so that, in addition to the linux-next consideration,
the script also looks for a tag corresponding to the 6.4.6-rt8 part of
what will become the `uname -r` string. With this patch (so 29 patches
on top of v6.4.6-rt8), one instead gets

$ make -s kernelrelease
6.4.6-rt8-00029-gd533209291a2

While there, note that the line

  git describe --exact-match --match=$tag $tag 2>/dev/null

obviously asks if $tag is an annotated tag, but it does not actually
tell if the commit pointed to has any relation to HEAD. So remove both
uses of --exact-match, and instead just ask if the description
generated is identical to the tag we provided. Since we then already
have the result of

  git describe --match=$tag

we also end up reducing the number of times we invoke "git describe".
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Nobody has complained since 2a73cce2 ("scripts/setlocalversion:
remove mercurial, svn and git-svn supports"), so let's also clean up
the header comment accordingly.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

debian/rules is generated by shell, but the escape sequence (\$) is
unreadable.

debian/rules embeds only two variables (ARCH and KERNELRELEASE).

Split them out to debian/rules.vars, and check-in the rest of Makefile
code to scripts/package/debian/rules.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>

Debian Policy "4.9. Main building script: debian/rules" requires
"debian/rules must start with the line #!/usr/bin/make -f". [1]

Currently, Kbuild does not follow this policy.

When Kbuild generates debian/rules, "#!$(command -v $MAKE) -f" is
expanded by shell. The resuling string may not be "#!/usr/bin/make -f".

There was a reason to opt out the Debian policy.

If you run '/path/to/my/custom/make deb-pkg', debian/rules must also be
invoked by the same Make program. If #!/usr/bin/make were hard-coded in
debian/rules, the sub-make would be executed by a possibly different
Make version.

This is problematic due to the MAKEFLAGS incompatibility, especially the
job server flag. Old Make versions used --jobserver-fds to propagate job
server file descriptors, but Make >= 4.2 uses --jobserver-auth. The flag
disagreement between the parent/child Makes would result in a process
fork explosion.

However, having a non-standard path in the shebang causes another issue;
the generated source package is not portable as such a path does not
exist in other build environments.

This commit solves those conflicting demands.

Hard-code '#!/usr/bin/make -f' in debian/rules to create a portable and
Debian-compliant source package.

Pass '--rules-file=$(MAKE) -f debian/rules' when dpkg-buildpackage is
invoked from Makefile so that debian/rules is executed by the same Make
program as used to start Kbuild.

[1] https://www.debian.org/doc/debian-policy/ch-source.html#main-building-script-debian-rulesSigned-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>

Remove the Elf_Rela variables used in the for-loop in section_rel().

This makes the code consistent; section_rel() only uses Elf_Rel,
section_rela() only uses Elf_Rela.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

MIPS64 little endian target has an odd encoding of r_info.

This commit makes the special handling less ugly. It is still ugly,
but #if conditionals will go away, at least.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

All of addend_*_rel() need the Elf_Rela pointer just for calculating
ELF_R_TYPE(r->r_info).

You can do it on the caller to de-duplicate the code.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Now that none of addend_*_rel() returns a meaningful value (the return
value is always 0), change all of them to return the value of r_addend.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Commit 8818039f ("kbuild: add ability to make source rpm buildable
using koji") added the BuildRequires: field.

Checking the build dependency is fine, but one annoyance is that
'make (bin)rpm-pkg' fails on non-rpm systems [1]. For example, Debian
provides rpmbuild via 'apt install rpm', but of course cannot meet the
requirement listed in the BuildRequires: field.

It is possible to pass RPMOPTS=--nodeps to work around it, but it is
reasonable to do it automatically.

If 'rpm -q rpm' fails, it is not an RPM-managed system. (The command
'rpm' is not installed at all, or was installed by other means.)

In that case, pass --nodeps to skip the build dependency check.

[1]: https://lore.kernel.org/linux-kbuild/Y6mkdYQYmjUz7bqV@li-4a3a4a4c-28e5-11b2-a85c-a8d192c6f089.ibm.com/Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Merge the similar build targets.

Also, make the output location consistent.

Previously, source packages were created in the build directory,
while binary packages under ~/rpmbuild/RPMS/.

Now, Kbuild creates the rpmbuild/ directory in the build directory,
and saves all packages under it.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Currently, 'make rpm-pkg' always builds the kernel from the pristine
source tree in the ~/rpmbuild/BUILD/ directory.

Build the kernel incrementally just like 'make binrpm-pkg'.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Now kernel.spec and binkernel.spec have the exactly same contents.

Use kernel.spec for binrpm-pkg as well.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Most of the lines in the spec file are independent of any build
condition.

Split the body of the spec file into scripts/package/kernel.spec.
scripts/package/mkspec will prepend some env-dependent variables.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

scripts/package/mkspec preprocesses the spec file by sed, but it is
unreadable. This commit removes the last portion of the sed scripting.

Remove the $S$M prefixes from the conditionally generated lines.
Instead, surround the code with %if %{with_devel} ... %endif.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

For the same reason as commit 4243afdb ("kbuild: builddeb: always
make modules_install, to install modules.builtin*"), run modules_install
even when CONFIG_MODULES=n to install modules.builtin*.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

To reduce the preprocess of the spec file, invoke the kernel build
from rpmbuild.

Run init/build-version to increment the release number not only for
binrpm-pkg but also for srcrpm-pkg and rpm-pkg.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

If this affects only %{buildroot}, it should be enough to use a fixed
string for _arch when it is undefined.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Avoid hard-coding the Version field in the generated spec file.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The kernel-devel RPM package and the linux-headers Debian package
provide headers and scripts needed for building external modules.

They copy the necessary files in slightly different ways - the RPM
copies almost everything except some exclude patterns, while the Debian
copies less number of files. There is no need to maintain different code
to do the same thing.

Split the Debian code out to scripts/package/install-extmod-build, which
is called from both of the packages.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

There are some cases where we want to run a command with the same
environment variables as Kbuild uses. For example, 'make coccicheck'
invokes scripts/coccicheck from the top Makefile so that the script can
reference to ${LINUXINCLUDE}, ${KBUILD_EXTMOD}, etc. The top Makefile
defines several phony targets that run a script.

We do it also for an internally used script, which results in a somewhat
complex call graph.

One example:

 debian/rules binary-arch
   -> make intdeb-pkg
      -> scripts/package/builddeb

It is also tedious to add a dedicated target like 'intdeb-pkg' for each
use case.

Add a generic target 'run-command' to run an arbitrary command in an
environment with all Kbuild variables set.

The usage is:

  $ make run-command KBUILD_RUN_COMMAND=<command>

The concept is similar to:

  $ dpkg-architecture -c <command>

This executes <command> in an environment which has all DEB_* variables
defined.

Convert the existing 'make intdeb-pkg'.

Another possible usage is to interrogate a Make variable.

  $ make run-command KBUILD_RUN_COMMAND='echo $(KBUILD_CFLAGS)'

might be useful to see KBUILD_CFLAGS set by the top Makefile.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Avoid hard-coding the value of KERNELRELEASE in the generated spec file.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

${version} will be replaced with the value of the Version field.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Currently, we rely on the top Makefile defining ARCH option when we
run 'make rpm-pkg' or 'make binrpm-pkg'.

It does not apply when we run 'make srcrpm-pkg', and separately run
'rpmbuild' for the generated SRPM. This is a problem for cross-build.

Just like the Debian package, save the value of ARCH in the spec file.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

This is useful to pass more common Make options.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Currently, $MAKE will expand to the GNU Make program that created the
source RPM. This is problematic if you carry it to a different build
host to run 'rpmbuild' there.

Consider this command:

  $ /path/to/my/custom/make srcrpm-pkg

The spec file in the SRPM will record '/path/to/my/custom/make', which
exists only on that build environment.

To create a portable SRPM, the spec file should avoid hard-coding $MAKE.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

This is unneeded because the Makefile in the output directory wraps
the top-level Makefile in the srctree.

Just run $MAKE irrespective of the build location.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Commit 3089b2be ("kbuild: rpm-pkg: fix build error when _arch is
undefined") does not work as intended; _arch is always defined as
$UTS_MACHINE.

The intention was to define _arch to $UTS_MACHINE only when it is not
defined.

Fixes: 3089b2be ("kbuild: rpm-pkg: fix build error when _arch is undefined")
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Remove hack for ancient version of module-init-tools that was added in
Linux 3.0.

Since then module-init-tools was replaced with kmod.

This hack adds an additional indirection, and causes confusing errors
to be printed when depmod fails.

Reverts commit 8fc62e59 ("kbuild: Do not write to builddir in modules_install")
Reverts commit bfe5424a ("kbuild: Hack for depmod not handling X.Y versions")

Link: https://lore.kernel.org/linux-modules/CAK7LNAQMs3QBYfWcLkmOQdbbq7cj=7wWbK=AWhdTC2rAsKHXzQ@mail.gmail.com/Signed-off-by: Michal Suchanek <msuchanek@suse.de>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

You do not need to remember the index of each jump key because you can
count it up after a key is pressed.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Jesse Taube <Mr.Bossman075@gmail.com>

Commit 95ac9b3b ("menuconfig: Assign jump keys per-page instead
of globally") injected a lot of hacks to the bottom of the textbox
infrastructure.

I reverted many of them without changing the behavior. (almost)
Now, the key markers are inserted when constructing the search result
instead of updating the text buffer on-the-fly.

The buffer passed to the textbox got back to a constant string.
The ugly casts from (const char *) to (char *) went away.

A disadvantage is that the same key numbers might be displayed multiple
times in the dialog if you use a huge window (but I believe it is
unlikely to happen).
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Jesse Taube <Mr.Bossman075@gmail.com>

Currently, all files with EXPORT_SYMBOL() are rebuilt when CONFIG_MODULES
is flipped due to <linux/export.h> depending on CONFIG_MODULES.

Now that modpost can make a final decision about export symbols,
<linux/export.h> does not need to make EXPORT_SYMBOL() no-op.
Instead, modpost can skip emitting KSYMTAB when CONFIG_MODULES is unset.

This commit will reduce the number of recompilation when CONFIG_MODULES
is toggled.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

This diagnostic checks whether there is a type mismatch when
converting enums (assign an enum of type A to an enum of type B, for
example) and it caught a legit issue recently. The reason it didn't show
is because that warning is enabled only with -Wextra with GCC. Clang,
however, enables it by default.

GCC folks were considering enabling it by default but it was too noisy
back then:

  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78736

Now that due to clang all those warnings have been fixed, enable it with
GCC too.

allmodconfig tests done with: x86, arm{,64}, powerpc{,64}, riscv
crossbuilds.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Pull tracing fixes from Steven Rostedt:

 - Swapping the ring buffer for snapshotting (for things like irqsoff)
   can crash if the ring buffer is being resized. Disable swapping when
   this happens. The missed swap will be reported to the tracer

 - Report error if the histogram fails to be created due to an error in
   adding a histogram variable, in event_hist_trigger_parse()

 - Remove unused declaration of tracing_map_set_field_descr()

* tag 'trace-v6.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  tracing/histograms: Return an error if we fail to add histogram to hist_vars list
  ring-buffer: Do not swap cpu_buffer during resize process
  tracing: Remove unused extern declaration tracing_map_set_field_descr()

Merge tag 'kbuild-fixes-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix stale help text in gconfig

 - Support *.S files in compile_commands.json

 - Flatten KBUILD_CFLAGS

 - Fix external module builds with Rust so that temporary files are
   created in the modules directories instead of the kernel tree

* tag 'kbuild-fixes-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kbuild: rust: avoid creating temporary files
  kbuild: flatten KBUILD_CFLAGS
  gen_compile_commands: add assembly files to compilation database
  kconfig: gconfig: correct program name in help text
  kconfig: gconfig: drop the Show Debug Info help text

`rustc` outputs by default the temporary files (i.e. the ones saved
by `-Csave-temps`, such as `*.rcgu*` files) in the current working
directory when `-o` and `--out-dir` are not given (even if
`--emit=x=path` is given, i.e. it does not use those for temporaries).

Since out-of-tree modules are compiled from the `linux` tree,
`rustc` then tries to create them there, which may not be accessible.

Thus pass `--out-dir` explicitly, even if it is just for the temporary
files.

Similarly, do so for Rust host programs too.
Reported-by: Raphael Nestler <raphael.nestler@gmail.com>
Closes: https://github.com/Rust-for-Linux/linux/issues/1015Reported-by: Andrea Righi <andrea.righi@canonical.com>
Tested-by: Raphael Nestler <raphael.nestler@gmail.com> # non-hostprogs
Tested-by: Andrea Righi <andrea.righi@canonical.com> # non-hostprogs
Fixes: 295d8398 ("kbuild: specify output names separately for each emission type from rustc")
Cc: stable@vger.kernel.org
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Tested-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>