linux.git/scripts/Makefile.build, branch v5.16.1

kbuild: reuse $(cmd_objtool) for cmd_cc_lto_link_modules

2021-09-22T00:26:25+00:00

For CONFIG_LTO_CLANG=y, the objtool processing is not possible at the
compilation, hence postponed by the link time.

Reuse $(cmd_objtool) for CONFIG_LTO_CLANG=y by defining objtool-enabled
properly.

For CONFIG_LTO_CLANG=y:

  objtool-enabled is off for %.o compilation
  objtool-enabled is on  for %.lto link

For CONFIG_LTO_CLANG=n:

  objtool-enabled is on for %.o compilation
      (but, it depends on OBJECT_FILE_NON_STANDARD)

Set part-of-module := y for %.lto.o to avoid repeating --module.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Kees Cook 
Acked-by: Josh Poimboeuf

kbuild: detect objtool update without using .SECONDEXPANSION

2021-09-22T00:26:25+00:00

Redo commit 8852c5524029 ("kbuild: Fix objtool dependency for
'OBJECT_FILES_NON_STANDARD_ := n'") to add the objtool
dependency in a cleaner way.

Using .SECONDEXPANSION ends up with unreadable code due to escaped
dollars. Also, it is not efficient because the second half of
Makefile.build is parsed twice every time.

Append the objtool dependency to the *.cmd files at the build time.

This is what fixdep and gen_ksymdeps.sh already do. So, following the
same pattern seems a natural solution.

This allows us to drop $$(objtool_dep) entirely.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Kees Cook 
Reviewed-by: Nick Desaulniers 
Acked-by: Josh Poimboeuf

kbuild: factor out OBJECT_FILES_NON_STANDARD check into a macro

2021-09-22T00:26:25+00:00

The OBJECT_FILES_NON_STANDARD check is quite long.

Factor it out into a new macro, objtool-enabled, to not repeat it.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Nick Desaulniers 
Reviewed-by: Kees Cook 
Acked-by: Josh Poimboeuf

kbuild: store the objtool command in *.cmd files

2021-09-22T00:26:25+00:00

objtool_dep includes include/config/{ORC_UNWINDER,STACK_VALIDATION}
so that all the objects are rebuilt when CONFIG_ORC_UNWINDER or
CONFIG_STACK_VALIDATION is toggled.

BTW, the correct option name is not CONFIG_ORC_UNWINDER, but
CONFIG_UNWINDER_ORC. Commit 11af847446ed ("x86/unwind: Rename
unwinder config options to 'CONFIG_UNWINDER_*'") missed to
adjust this part. So, this dependency has been broken for a
long time.

As you can see in 'objtool_args', there are more CONFIG options
that affect the objtool command line.

Adding more and more include/config/* is ugly and unmaintainable.

Another issue is that non-standard objects are needlessly rebuilt.
Objects specified as OBJECT_FILES_NON_STANDARD is not processed by
objtool, but they are rebuilt anyway when CONFIG_STACK_VALIDATION
is toggled. This is not a big deal, but better to fix.

A cleaner and more precise fix is to include the objtool command in
*.cmd files so any command change is naturally detected by if_change.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Nick Desaulniers 
Reviewed-by: Kees Cook 
Acked-by: Josh Poimboeuf

kbuild: rename __objtool_obj and reuse it for cmd_cc_lto_link_modules

2021-09-22T00:26:25+00:00

Rename __objtool_obj to objtool, and move it out of the
'ifndef CONFIG_LTO_CLANG' conditional, so it can be used for
cmd_cc_lto_link_modules as well.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Nick Desaulniers 
Reviewed-by: Kees Cook 
Acked-by: Josh Poimboeuf

kbuild: move objtool_args back to scripts/Makefile.build

2021-09-22T00:26:25+00:00

Commit b1a1a1a09b46 ("kbuild: lto: postpone objtool") moved objtool_args
to Makefile.lib, so the arguments can be used in Makefile.modfinal as
well as Makefile.build.

With commit 850ded46c642 ("kbuild: Fix TRIM_UNUSED_KSYMS with
LTO_CLANG"), module LTO linking came back to scripts/Makefile.build
again.

So, there is no more reason to keep objtool_args in a separate file.

Get it back to the original place, close to the objtool command.

Remove the stale comment too.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Kees Cook 
Acked-by: Josh Poimboeuf

kbuild: remove stale *.symversions

2021-09-02T23:17:20+00:00

cmd_update_lto_symversions merges all the existing *.symversions, but
some of them might be stale.

If the last EXPORT_SYMBOL is removed from a C file, the *.symversions
file is not deleted or updated. It contains stale CRCs, but still they
will be used for linking the vmlinux or modules.

It is not a big deal when the EXPORT_SYMBOL is really removed. However,
when the EXPORT_SYMBOL is moved to another file, the same __crc_
will appear twice in the merged *.symversions, possibly with different
CRCs if the function argument is changed at the same time. It would
confuse module versioning.

If no EXPORT_SYMBOL is found, let's remove *.symversions explicitly.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Kees Cook

kbuild: remove unused quiet_cmd_update_lto_symversions

2021-09-02T23:17:20+00:00

This is not used anywhere because the short log is displayed when
it is used through a $(call cmd,...) invocation.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Kees Cook

kbuild: Fix TRIM_UNUSED_KSYMS with LTO_CLANG

2021-09-02T23:12:39+00:00

With CONFIG_LTO_CLANG, we currently link modules into native
code just before modpost, which means with TRIM_UNUSED_KSYMS
enabled, we still look at the LLVM bitcode in the .o files when
generating the list of used symbols. As the bitcode doesn't
yet have calls to compiler intrinsics and llvm-nm doesn't see
function references that only exist in function-level inline
assembly, we currently need a whitelist for TRIM_UNUSED_KSYMS to
work with LTO.

This change moves module LTO linking to happen earlier, and
thus avoids the issue with LLVM bitcode and TRIM_UNUSED_KSYMS
entirely, allowing us to also drop the whitelist from
gen_autoksyms.sh.

Link: https://github.com/ClangBuiltLinux/linux/issues/1369
Signed-off-by: Sami Tolvanen 
Reviewed-by: Alexander Lobakin 
Tested-by: Alexander Lobakin 
Signed-off-by: Masahiro Yamada

Kbuild: lto: fix module versionings mismatch in GNU make 3.X

2021-07-18T14:48:14+00:00

When building modules(CONFIG_...=m), I found some of module versions
are incorrect and set to 0.
This can be found in build log for first clean build which shows

WARNING: EXPORT symbol "XXXX" [drivers/XXX/XXX.ko] version generation failed,
symbol will not be versioned.

But in second build(incremental build), the WARNING disappeared and the
module version becomes valid CRC and make someone who want to change
modules without updating kernel image can't insert their modules.

The problematic code is
+	$(foreach n, $(filter-out FORCE,$^),				\
+		$(if $(wildcard $(n).symversions),			\
+			; cat $(n).symversions >> $@.symversions))

For example:
  rm -f fs/notify/built-in.a.symversions    ; rm -f fs/notify/built-in.a; \
llvm-ar cDPrST fs/notify/built-in.a fs/notify/fsnotify.o \
fs/notify/notification.o fs/notify/group.o ...

`foreach n` shows nothing to `cat` into $(n).symversions because
`if $(wildcard $(n).symversions)` return nothing, but actually
they do exist during this line was executed.

-rw-r--r-- 1 root root 168580 Jun 13 19:10 fs/notify/fsnotify.o
-rw-r--r-- 1 root root    111 Jun 13 19:10 fs/notify/fsnotify.o.symversions

The reason is the $(n).symversions are generated at runtime, but
Makefile wildcard function expends and checks the file exist or not
during parsing the Makefile.

Thus fix this by use `test` shell command to check the file
existence in runtime.

Rebase from both:
1. [https://lore.kernel.org/lkml/20210616080252.32046-1-lecopzer.chen@mediatek.com/]
2. [https://lore.kernel.org/lkml/20210702032943.7865-1-lecopzer.chen@mediatek.com/]

Fixes: 38e891849003 ("kbuild: lto: fix module versioning")
Co-developed-by: Sami Tolvanen 
Signed-off-by: Lecopzer Chen 
Signed-off-by: Masahiro Yamada