Revert "kheaders: substituting --sort in archive creation"

This reverts commit 700dea5a0bea9f64eba89fae7cb2540326fdfdc1.

The reason for that commit was --sort=ORDER introduced in
tar 1.28 (2014). More than 3 years have passed since then.

Requiring GNU tar 1.28 should be fine now because we require
GCC 5.1 (2015).

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>

kheaders: use standard naming for the temporary directory

If the kheaders archive generation is interrupted then this directory
may be left on disk and not ignored by git.
By using the standard naming schema for temporary files and directories
the default .gitignore and "make clean" rules will apply.

Suggested-by: Nicolas Schier <nicolas@fjasle.eu>
Suggested-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
Tested-by: Nicolas Schier <nicolas@fjasle.eu>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kheaders: explicitly validate existence of cpio command

If the cpio command is not available the error emitted by
gen_kheaders.so is not clear as all output of the call to cpio is
discarded:

GNU make 4.4:

GEN kernel/kheaders_data.tar.xz
find: 'standard output': Broken pipe
find: write error
make[2]: *** [kernel/Makefile:157: kernel/kheaders_data.tar.xz] Error 127
make[1]: *** [scripts/Makefile.build:504: kernel] Error 2

GNU make < 4.4:

GEN kernel/kheaders_data.tar.xz
make[2]: *** [kernel/Makefile:157: kernel/kheaders_data.tar.xz] Error 127
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [scripts/Makefile.build:504: kernel] Error 2

Add an explicit check that will trigger a clear message about the issue:

CHK kernel/kheaders_data.tar.xz
./kernel/gen_kheaders.sh: line 17: type: cpio: not found

The other commands executed by gen_kheaders.sh are part of a standard
installation, so they are not checked.

Reported-by: Amy Parker <apark0006@student.cerritos.edu>
Link: https://lore.kernel.org/lkml/CAPOgqxFva=tOuh1UitCSN38+28q3BNXKq19rEsVNPRzRqKqZ+g@mail.gmail.com/
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Nicolas Schier <n.schier@avm.de>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kbuild: build init/built-in.a just once

Kbuild builds init/built-in.a twice; first during the ordinary
directory descending, second from scripts/link-vmlinux.sh.

We do this because UTS_VERSION contains the build version and the
timestamp. We cannot update it during the normal directory traversal
since we do not yet know if we need to update vmlinux. UTS_VERSION is
temporarily calculated, but omitted from the update check. Otherwise,
vmlinux would be rebuilt every time.

When Kbuild results in running link-vmlinux.sh, it increments the
version number in the .version file and takes the timestamp at that
time to really fix UTS_VERSION.

However, updating the same file twice is a footgun. To avoid nasty
timestamp issues, all build artifacts that depend on init/built-in.a
are atomically generated in link-vmlinux.sh, where some of them do not
need rebuilding.

To fix this issue, this commit changes as follows:

[1] Split UTS_VERSION out to include/generated/utsversion.h from
include/generated/compile.h

include/generated/utsversion.h is generated just before the
vmlinux link. It is generated under include/generated/ because
some decompressors (s390, x86) use UTS_VERSION.

[2] Split init_uts_ns and linux_banner out to init/version-timestamp.c
from init/version.c

init_uts_ns and linux_banner contain UTS_VERSION. During the ordinary
directory descending, they are compiled with __weak and used to
determine if vmlinux needs relinking. Just before the vmlinux link,
they are compiled without __weak to embed the real version and
timestamp.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kheaders: Have cpio unconditionally replace files

For out-of-tree builds, this script invokes cpio twice to copy header
files from the srctree and subsequently from the objtree. According to a
comment in the script, there might be situations in which certain files
already exist in the destination directory when header files are copied
from the objtree:

"The second CPIO can complain if files already exist which can happen
with out of tree builds having stale headers in srctree. Just silence
CPIO for now."

GNU cpio might simply print a warning like "newer or same age version
exists", but toybox cpio exits with a non-zero exit code unless the
command line option "-u" is specified.

To improve compatibility with toybox cpio, add the command line option
"-u" to unconditionally replace existing files in the destination
directory.

Signed-off-by: Daniel Mentz <danielmentz@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kbuild: clean up ${quiet} checks in shell scripts

There were efforts to make 'make -s' really silent when it is a
warning-free build.

The conventional way was to let a shell script check ${quiet}, and if
it is 'silent_', suppress the stdout by itself.

With the previous commit, the 'cmd' takes care of it now. The 'cmd' is
also invoked from if_changed, if_changed_dep, and if_changed_rule.

You can omit ${quiet} checks in shell scripts when they are invoked
from the 'cmd' macro.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kbuild: redo fake deps at include/config/*.h

Make include/config/foo/bar.h fake deps files generation simpler.

* delete .h suffix
those aren't header files, shorten filenames,

* delete tolower()
Linux filesystems can deal with both upper and lowercase
filenames very well,

* put everything in 1 directory
Presumably 'mkdir -p' split is from dark times when filesystems
handled huge directories badly, disks were round adding to
seek times.

x86_64 allmodconfig lists 12364 files in include/config.

../obj/include/config/
├── 104_QUAD_8
├── 60XX_WDT
├── 64BIT
...
├── ZSWAP_DEFAULT_ON
├── ZSWAP_ZPOOL_DEFAULT
└── ZSWAP_ZPOOL_DEFAULT_ZBUD

0 directories, 12364 files

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kbuild: add variables for compression tools

Allow user to use alternative implementations of compression tools,
such as pigz, pbzip2, pxz. For example, multi-threaded tools to
speed up the build:
$ make GZIP=pigz BZIP2=pbzip2

Variables _GZIP, _BZIP2, _LZOP are used internally because original env
vars are reserved by the tools. The use of GZIP in gzip tool is obsolete
since 2015. However, alternative implementations (e.g., pigz) still rely
on it. BZIP2, BZIP, LZOP vars are not obsolescent.

The credit goes to @grsecurity.

As a sidenote, for multi-threaded lzma, xz compression one can use:
$ export XZ_OPT="--threads=0"

Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

kheaders: explain why include/config/autoconf.h is excluded from md5sum

This comment block explains why include/generated/compile.h is omitted,
but nothing about include/generated/autoconf.h, which might be more
difficult to understand. Add more comments.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: remove the last bashism to allow sh to run it

'pushd' ... 'popd' is the last bash-specific code in this script.
One way to avoid it is to run the code in a sub-shell.

With that addressed, you can run this script with sh.

I replaced $(BASH) with $(CONFIG_SHELL), and I changed the hashbang
to #!/bin/sh.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: optimize header copy for in-tree builds

This script copies headers by the cpio command twice; first from
srctree, and then from objtree. However, when we building in-tree,
we know the srctree and the objtree are the same. That is, all the
headers copied by the first cpio are overwritten by the second one.

Skip the first cpio when we are building in-tree.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: optimize md5sum calculation for in-tree builds

This script computes md5sum of headers in srctree and in objtree.
However, when we are building in-tree, we know the srctree and the
objtree are the same. That is, we end up with the same computation
twice. In fact, the first two lines of kernel/kheaders.md5 are always
the same for in-tree builds.

Unify the two md5sum calculations.

For in-tree builds ($building_out_of_srctree is empty), we check
only two directories, "include", and "arch/$SRCARCH/include".

For out-of-tree builds ($building_out_of_srctree is 1), we check
4 directories, "$srctree/include", "$srctree/arch/$SRCARCH/include",
"include", and "arch/$SRCARCH/include" since we know they are all
different.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: remove unneeded 'cat' command piped to 'head' / 'tail'

The 'head' and 'tail' commands can take a file path directly.
So, you do not need to run 'cat'.

cat kernel/kheaders.md5 | head -1

... is equivalent to:

head -1 kernel/kheaders.md5

and the latter saves forking one process.

While I was here, I replaced 'head -1' with 'head -n 1'.

I also replaced '==' with '=' since we do not have a good reason to
use the bashism.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: substituting --sort in archive creation

The option --sort=ORDER was only introduced in tar 1.28 (2014), which
is rather new and might not be available in some setups.

This patch tries to replicate the previous behaviour as closely as
possible to fix the kheaders build for older environments. It does
not produce identical archives compared to the previous version due
to minor sorting differences but produces reproducible results itself
in my tests.

Reported-by: Andreas Schwab <schwab@suse.de>
Signed-off-by: Dmitry Goldin <dgoldin+lkml@protonmail.ch>
Tested-by: Andreas Schwab <schwab@suse.de>
Tested-by: Quentin Perret <qperret@google.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: make headers archive reproducible

In commit 43d8ce9d65a5 ("Provide in-kernel headers to make
extending kernel easier") a new mechanism was introduced, for kernels
>=5.2, which embeds the kernel headers in the kernel image or a module
and exposes them in procfs for use by userland tools.

The archive containing the header files has nondeterminism caused by
header files metadata. This patch normalizes the metadata and utilizes
KBUILD_BUILD_TIMESTAMP if provided and otherwise falls back to the
default behaviour.

In commit f7b101d33046 ("kheaders: Move from proc to sysfs") it was
modified to use sysfs and the script for generation of the archive was
renamed to what is being patched.

Signed-off-by: Dmitry Goldin <dgoldin+lkml@protonmail.ch>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

kheaders: include only headers into kheaders_data.tar.xz

Currently, kheaders_data.tar.xz contains some build scripts as well as
headers. None of them is needed in the header archive.

For ARCH=x86, this commit excludes the following from the archive:

arch/x86/include/asm/Kbuild
arch/x86/include/uapi/asm/Kbuild
include/asm-generic/Kbuild
include/config/auto.conf
include/config/kernel.release
include/config/tristate.conf
include/uapi/asm-generic/Kbuild
include/uapi/Kbuild
kernel/gen_kheaders.sh

This change is actually motivated for the planned header compile-testing
because it will generate more build artifacts, which should not be
included in the archive.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>

kheaders: remove meaningless -R option of 'ls'

The -R option of 'ls' is supposed to be used for directories.

-R, --recursive
list subdirectories recursively

Since 'find ... -type f' only matches to regular files, we do not
expect directories passed to the 'ls' command here.

Giving -R is harmless at least, but unneeded.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>

kheaders: Do not regenerate archive if config is not changed

Linus reported an issue that doing an allmodconfig was causing the
kheaders archive to be regenerated even though the config is the same.
This patch fixes the issue by ignoring the config-related header files
for "knowing when to regenerate based on timestamps". Instead, if the
CONFIG_X_Y option really changes, then we there are the
include/config/X/Y.h which will already tells us "if a config really
changed". So we don't really need these files for regeneration detection
anyway, and ignoring them fixes Linus's issue.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

kheaders: Move from proc to sysfs

The kheaders archive consisting of the kernel headers used for compiling
bpf programs is in /proc. However there is concern that moving it here
will make it permanent. Let us move it to /sys/kernel as discussed [1].

[1] https://lore.kernel.org/patchwork/patch/1067310/#1265969

Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>