<feed xmlns='http://www.w3.org/2005/Atom'>
<title>user/sven/linux.git/scripts/Makefile.build, branch v6.0.18</title>
<subtitle>Linux Kernel
</subtitle>
<id>https://git.stealer.net/cgit.cgi/user/sven/linux.git/atom?h=v6.0.18</id>
<link rel='self' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/atom?h=v6.0.18'/>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/'/>
<updated>2022-07-27T12:17:59Z</updated>
<entry>
<title>kbuild: remove sed command from cmd_ar_builtin</title>
<updated>2022-07-27T12:17:59Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-06-14T05:51:49Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=5439d4d4dcee548b1ad984c293225c1668045ad2'/>
<id>urn:sha1:5439d4d4dcee548b1ad984c293225c1668045ad2</id>
<content type='text'>
Replace a pipeline of echo and sed with printf to decrease process forks.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Nick Desaulniers &lt;ndesaulniers@google.com&gt;
</content>
</entry>
<entry>
<title>kbuild: avoid regex RS for POSIX awk</title>
<updated>2022-06-07T16:27:26Z</updated>
<author>
<name>Kevin Locke</name>
<email>kevin@kevinlocke.name</email>
</author>
<published>2022-06-07T02:42:54Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=7bf179de5b2dfae54a6839eaf7caba44a888ee2e'/>
<id>urn:sha1:7bf179de5b2dfae54a6839eaf7caba44a888ee2e</id>
<content type='text'>
In 22f26f21774f8 awk was added to deduplicate *.mod files.  The awk
invocation passes -v RS='( |\n)' to match a space or newline character
as the record separator.  Unfortunately, POSIX states[1]

&gt; If RS contains more than one character, the results are unspecified.

Some implementations (such as the One True Awk[2] used by the BSDs) do
not treat RS as a regular expression.  When awk does not support regex
RS, build failures such as the following are produced (first error using
allmodconfig):

      CC [M]  arch/x86/events/intel/uncore.o
      CC [M]  arch/x86/events/intel/uncore_nhmex.o
      CC [M]  arch/x86/events/intel/uncore_snb.o
      CC [M]  arch/x86/events/intel/uncore_snbep.o
      CC [M]  arch/x86/events/intel/uncore_discovery.o
      LD [M]  arch/x86/events/intel/intel-uncore.o
    ld: cannot find uncore_nhmex.o: No such file or directory
    ld: cannot find uncore_snb.o: No such file or directory
    ld: cannot find uncore_snbep.o: No such file or directory
    ld: cannot find uncore_discovery.o: No such file or directory
    make[3]: *** [scripts/Makefile.build:422: arch/x86/events/intel/intel-uncore.o] Error 1
    make[2]: *** [scripts/Makefile.build:487: arch/x86/events/intel] Error 2
    make[1]: *** [scripts/Makefile.build:487: arch/x86/events] Error 2
    make: *** [Makefile:1839: arch/x86] Error 2

To avoid this, use printf(1) to produce a newline between each object
path, instead of the space produced by echo(1), so that the default RS
can be used by awk.

[1]: https://pubs.opengroup.org/onlinepubs/9699919799/utilities/awk.html
[2]: https://github.com/onetrueawk/awk

Fixes: 22f26f21774f ("kbuild: get rid of duplication in *.mod files")
Signed-off-by: Kevin Locke &lt;kevin@kevinlocke.name&gt;
Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
</content>
</entry>
<entry>
<title>kbuild: factor out the common objtool arguments</title>
<updated>2022-06-04T21:20:57Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-28T15:47:04Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=b42d2306502419688190aa6dd4dab4a6def24b3d'/>
<id>urn:sha1:b42d2306502419688190aa6dd4dab4a6def24b3d</id>
<content type='text'>
scripts/Makefile.build and scripts/link-vmlinux.sh have similar setups
for the objtool arguments.

It was difficult to factor out them because all the vmlinux build rules
were written in a shell script. It is somewhat tedious to touch the two
files every time a new objtool option is supported.

To reduce the code duplication, move the objtool for vmlinux.o into
scripts/Makefile.vmlinux_o. Then, move the common macros to Makefile.lib
so they are shared between Makefile.build and Makefile.vmlinux_o.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
</content>
</entry>
<entry>
<title>kbuild: rebuild multi-object modules when objtool is updated</title>
<updated>2022-06-01T14:07:29Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:55Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=f6b66ca4f38b1169313383aec7fa0a8446205ebb'/>
<id>urn:sha1:f6b66ca4f38b1169313383aec7fa0a8446205ebb</id>
<content type='text'>
When CONFIG_LTO_CLANG or CONFIG_X86_KERNEL_IBT is enabled, objtool for
multi-object modules is postponed until the objects are linked together.

Make sure to re-run objtool and re-link multi-object modules when
objtool is updated.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Kees Cook &lt;keescook@chromium.org&gt;
Acked-by: Josh Poimboeuf &lt;jpoimboe@redhat.com&gt;
Tested-by: Nathan Chancellor &lt;nathan@kernel.org&gt;
Reviewed-by: Nicolas Schier &lt;n.schier@avm.de&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
</content>
</entry>
<entry>
<title>kbuild: make *.mod rule robust against too long argument error</title>
<updated>2022-06-01T14:07:29Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:53Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=c6031b1dbbbfec03891bf1baefa2e0803d705601'/>
<id>urn:sha1:c6031b1dbbbfec03891bf1baefa2e0803d705601</id>
<content type='text'>
Like built-in.a, the command length of the *.mod rule scales with
the depth of the directory times the number of objects in the Makefile.

Add $(obj)/ by the shell command (awk) instead of by Make's builtin
function.

In-tree modules still have some room to the limit (ARG_MAX=2097152),
but this is more future-proof for big modules in a deep directory.

For example, you can build i915 as a module (CONFIG_DRM_I915=m) and
compare drivers/gpu/drm/i915/.i915.mod.cmd with/without this commit.

The issue is more critical for external modules because the M= path
can be very long as Jeff Johnson reported before [1].

[1] https://lore.kernel.org/linux-kbuild/4c02050c4e95e4cb8cc04282695f8404@codeaurora.org/

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Nicolas Schier &lt;nicolas@fjasle.eu&gt;
Tested-by: Nathan Chancellor &lt;nathan@kernel.org&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
</content>
</entry>
<entry>
<title>kbuild: make built-in.a rule robust against too long argument error</title>
<updated>2022-06-01T14:07:29Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:52Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=cd968b97c49214e6557381bddddacbd0e0fb696e'/>
<id>urn:sha1:cd968b97c49214e6557381bddddacbd0e0fb696e</id>
<content type='text'>
Kbuild runs at the top of objtree instead of changing the working
directory to subdirectories. I think this design is nice overall but
some commands have a scalability issue.

The build command of built-in.a is one of them whose length scales with:

    O(D * N)

Here, D is the length of the directory path (i.e. $(obj)/ prefix),
N is the number of objects in the Makefile, O() is the big O notation.

The deeper directory the Makefile directory is located, the more easily
it will hit the too long argument error.

We can make it better. Trim the $(obj)/ by Make's builtin function, and
restore it by a shell command (sed).

With this, the command length scales with:

    O(D + N)

In-tree modules still have some room to the limit (ARG_MAX=2097152),
but this is more future-proof for big modules in a deep directory.

For example, you can build i915 as builtin (CONFIG_DRM_I915=y) and
compare drivers/gpu/drm/i915/.built-in.a.cmd with/without this commit.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Nicolas Schier &lt;nicolas@fjasle.eu&gt;
Tested-by: Nathan Chancellor &lt;nathan@kernel.org&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
</content>
</entry>
<entry>
<title>kbuild: check static EXPORT_SYMBOL* by script instead of modpost</title>
<updated>2022-06-01T14:07:02Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:51Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=31cb50b5590fe911077b8463ad01144fac8fa4f3'/>
<id>urn:sha1:31cb50b5590fe911077b8463ad01144fac8fa4f3</id>
<content type='text'>
The 'static' specifier and EXPORT_SYMBOL() are an odd combination.

Commit 15bfc2348d54 ("modpost: check for static EXPORT_SYMBOL*
functions") tried to detect it, but this check has false negatives.

Here is the sample code.

  Makefile:

    obj-y += foo1.o foo2.o

  foo1.c:

    #include &lt;linux/export.h&gt;
    static void foo(void) {}
    EXPORT_SYMBOL(foo);

  foo2.c:

    void foo(void) {}

foo1.c exports the static symbol 'foo', but modpost cannot catch it
because it is fooled by foo2.c, which has a global symbol with the
same name.

s-&gt;is_static is cleared if a global symbol with the same name is found
somewhere, but EXPORT_SYMBOL() and the global symbol do not necessarily
belong to the same compilation unit.

This check should be done per compilation unit, but I do not know how
to do it in modpost. modpost runs against vmlinux.o or modules, which
merges multiple objects, then forgets their origin.

modpost cannot parse individual objects because they may not be ELF but
LLVM IR when CONFIG_LTO_CLANG=y.

Add a simple bash script to parse the output from ${NM}. This works for
CONFIG_LTO_CLANG=y because llvm-nm can dump symbols of LLVM IR files.

Revert 15bfc2348d54.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Nick Desaulniers &lt;ndesaulniers@google.com&gt;
Tested-by: Nathan Chancellor &lt;nathan@kernel.org&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
</content>
</entry>
<entry>
<title>kbuild: do not create *.prelink.o for Clang LTO or IBT</title>
<updated>2022-05-29T09:39:35Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:49Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=c25e1c55822f9b3b53ccbf88b85644317a525752'/>
<id>urn:sha1:c25e1c55822f9b3b53ccbf88b85644317a525752</id>
<content type='text'>
When CONFIG_LTO_CLANG=y, additional intermediate *.prelink.o is created
for each module. Also, objtool is postponed until LLVM IR is converted
to ELF.

CONFIG_X86_KERNEL_IBT works in a similar way to postpone objtool until
objects are merged together.

This commit stops generating *.prelink.o, so the build flow will look
similar with/without LTO.

The following figures show how the LTO build currently works, and
how this commit is changing it.

Current build flow
==================

 [1] single-object module

                                      $(LD)
           $(CC)                     +objtool              $(LD)
    foo.c --------------------&gt; foo.o -----&gt; foo.prelink.o -----&gt; foo.ko
                              (LLVM IR)          (ELF)       |    (ELF)
                                                             |
                                                 foo.mod.o --/
                                                 (LLVM IR)

 [2] multi-object module
                                      $(LD)
           $(CC)         $(AR)       +objtool               $(LD)
    foo1.c -----&gt; foo1.o -----&gt; foo.o -----&gt; foo.prelink.o -----&gt; foo.ko
                           |  (archive)          (ELF)       |    (ELF)
    foo2.c -----&gt; foo2.o --/                                 |
                 (LLVM IR)                       foo.mod.o --/
                                                 (LLVM IR)

  One confusion is that foo.o in multi-object module is an archive
  despite of its suffix.

New build flow
==============

 [1] single-object module

  Since there is only one object, there is no need to keep the LLVM IR.
  Use $(CC)+$(LD) to generate an ELF object in one build rule. When LTO
  is disabled, $(LD) is unneeded because $(CC) produces an ELF object.

               $(CC)+$(LD)+objtool              $(LD)
    foo.c ----------------------------&gt; foo.o ---------&gt; foo.ko
                                        (ELF)     |      (ELF)
                                                  |
                                      foo.mod.o --/
                                      (LLVM IR)

 [2] multi-object module

  Previously, $(AR) was used to combine LLVM IR files into an archive,
  but there was no technical reason to do so. Use $(LD) to merge them
  into a single ELF object.

                               $(LD)
             $(CC)            +objtool          $(LD)
    foo1.c ---------&gt; foo1.o ---------&gt; foo.o ---------&gt; foo.ko
                                 |      (ELF)     |      (ELF)
    foo2.c ---------&gt; foo2.o ----/                |
                     (LLVM IR)        foo.mod.o --/
                                      (LLVM IR)

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Reviewed-by: Nicolas Schier &lt;nicolas@fjasle.eu&gt;
Tested-by: Nathan Chancellor &lt;nathan@kernel.org&gt;
Reviewed-by: Sami Tolvanen &lt;samitolvanen@google.com&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14 (x86-64)
Acked-by: Josh Poimboeuf &lt;jpoimboe@kernel.org&gt;
</content>
</entry>
<entry>
<title>kbuild: replace $(linked-object) with CONFIG options</title>
<updated>2022-05-29T09:39:35Z</updated>
<author>
<name>Masahiro Yamada</name>
<email>masahiroy@kernel.org</email>
</author>
<published>2022-05-27T10:01:48Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=0cfd90060d714e3d1e0d1b30cfd1c46b71f4e721'/>
<id>urn:sha1:0cfd90060d714e3d1e0d1b30cfd1c46b71f4e721</id>
<content type='text'>
*.prelink.o is created when CONFIG_LTO_CLANG or CONFIG_X86_KERNEL_IBT
is enabled.

Replace $(linked-object) with $(CONFIG_LTO_CLANG)$(CONFIG_X86_KERNEL_IBT)
so you will get a quick idea of when the --link option is passed.

No functional change is intended.

Signed-off-by: Masahiro Yamada &lt;masahiroy@kernel.org&gt;
Acked-by: Josh Poimboeuf &lt;jpoimboe@kernel.org&gt;
Reviewed-by: Nick Desaulniers &lt;ndesaulniers@google.com&gt;
Tested-by: Sedat Dilek &lt;sedat.dilek@gmail.com&gt; # LLVM-14
</content>
</entry>
<entry>
<title>Merge tag 'kbuild-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild</title>
<updated>2022-05-26T19:09:50Z</updated>
<author>
<name>Linus Torvalds</name>
<email>torvalds@linux-foundation.org</email>
</author>
<published>2022-05-26T19:09:50Z</published>
<link rel='alternate' type='text/html' href='https://git.stealer.net/cgit.cgi/user/sven/linux.git/commit/?id=df202b452fe6c6d6f1351bad485e2367ef1e644e'/>
<id>urn:sha1:df202b452fe6c6d6f1351bad485e2367ef1e644e</id>
<content type='text'>
Pull Kbuild updates from Masahiro Yamada:

 - Add HOSTPKG_CONFIG env variable to allow users to override pkg-config

 - Support W=e as a shorthand for KCFLAGS=-Werror

 - Fix CONFIG_IKHEADERS build to support toybox cpio

 - Add scripts/dummy-tools/pahole to ease distro packagers' life

 - Suppress false-positive warnings from checksyscalls.sh for W=2 build

 - Factor out the common code of arch/*/boot/install.sh into
   scripts/install.sh

 - Support 'kernel-install' tool in scripts/prune-kernel

 - Refactor module-versioning to link the symbol versions at the final
   link of vmlinux and modules

 - Remove CONFIG_MODULE_REL_CRCS because module-versioning now works in
   an arch-agnostic way

 - Refactor modpost, Makefiles

* tag 'kbuild-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: (56 commits)
  genksyms: adjust the output format to modpost
  kbuild: stop merging *.symversions
  kbuild: link symbol CRCs at final link, removing CONFIG_MODULE_REL_CRCS
  modpost: extract symbol versions from *.cmd files
  modpost: add sym_find_with_module() helper
  modpost: change the license of EXPORT_SYMBOL to bool type
  modpost: remove left-over cross_compile declaration
  kbuild: record symbol versions in *.cmd files
  kbuild: generate a list of objects in vmlinux
  modpost: move *.mod.c generation to write_mod_c_files()
  modpost: merge add_{intree_flag,retpoline,staging_flag} to add_header
  scripts/prune-kernel: Use kernel-install if available
  kbuild: factor out the common installation code into scripts/install.sh
  modpost: split new_symbol() to symbol allocation and hash table addition
  modpost: make sym_add_exported() always allocate a new symbol
  modpost: make multiple export error
  modpost: dump Module.symvers in the same order of modules.order
  modpost: traverse the namespace_list in order
  modpost: use doubly linked list for dump_lists
  modpost: traverse unresolved symbols in order
  ...
</content>
</entry>
</feed>
