The compressed LoongArch kernel (vmlinuz.efi) differs between builds in
different directories. The code and data of vmlinux are identical, but
its GNU build ID is not: the debug information of vmlinux holds a
single absolute path, the include directory of the compiler, e.g.
/builder/shared-workdir/build/staging_dir/toolchain-loongarch64_generic_gcc-14.4.0_musl/lib/gcc/loongarch64-openwrt-linux-musl/14.4.0/include
asm/loongarch.h includes the compiler's larchintrin.h, so core files
like arch/loongarch/kernel/cpu-probe.c refer to it. The prefix maps of
the kernel cover the build and the staging directory, but not the
toolchain directory. The build ID hashes all of vmlinux, so it depends
on the build directory, and with it the compressed kernel.
Map the toolchain directory as well, like the staging directory.
On loongarch64/generic, vmlinux built without the map holds this one
absolute path, like the published one; built with it, no absolute path
is left, while .text, .rodata and .data stay byte-identical.
Link: https://github.com/openwrt/openwrt/pull/25648
Signed-off-by: Paul Spooren <mail@aparcar.org>
CompileBPF calls clang with -g but without any prefix map. Stripping debug info
keeps .BTF and .BTF.ext, since libbpf needs them, and their line info carries
the absolute path of each source file. This makes BPF objects like qosify-bpf.o
depend on the build directory and breaks reproducibility:
-/builder/shared-workdir/build/build_dir/.../qosify-bpf.c
+/tmp/rebuilderd-openwrt.juCsuQ/src/build_dir/.../qosify-bpf.c
IREMAP_CFLAGS is not enough here: without CONFIG_REPRODUCIBLE_DEBUG_INFO it
only sets -fmacro-prefix-map, which leaves the line info alone, and it is empty
for external toolchains, which have no say over the host clang. Unlike regular
debug info, BTF ships with the package, so always pass -ffile-prefix-map for
the build and staging directories.
Add an fremap helper next to iremap for that, and let IREMAP_STAGING_DIR and
CompileBPF share the map of the staging directory, so they cannot drift apart.
Link: https://github.com/openwrt/openwrt/pull/25359
Signed-off-by: Paul Spooren <mail@aparcar.org>
Three cases kept a cached answer that no longer held:
A new configure script. A package with an autoreconf fixup regenerates
configure with the staged autoconf, automake and libtool, and a new
version of one of them can change what a cache variable means. The site
script now keeps a checksum of configure beside the cache and drops the
cache when the checksum changes. A new patch to configure is covered too.
A changed precious variable. configure aborts when LIBS, CPP or another
precious variable differs from the value in the cache, and because
nothing removes the cache, it aborts on every later run as well:
configure: error: 'LIBS' was not set in the previous run
configure: error: changes in the environment can compromise the build
The site script now makes the same comparison before configure does and
drops the cache on a difference.
Paths under other names. The libtool answers lt_cv_path_*, acl_cv_path_*
and lt_cv_sys_global_symbol_*, and gmp_cv_prog_m4, hold the path of a
system tool or of the staged sed and m4, and survived a wipe of
staging_dir. The site script now drops them with the ac_cv_path_ answers.
m4 and gmp still configure the same way with a warm cache as they do
without one, and still read it: the m4 configure stage takes 1.5 s warm
against 23 s cold.
Signed-off-by: John Crispin <john@phrozen.org>
iremap is called with BUILD_DIR and with PKG_BUILD_DIR, never with
STAGING_DIR. A header taken from the staging directory therefore reaches
__FILE__ and the debug information with its absolute path, and the result
depends on the directory the tree was built in.
Measured in the published snapshot of lantiq/xway: mt76-connac-lib.ko,
mt76x02-lib.ko and mt7925-common.ko hold
/builder/shared-workdir/build/staging_dir/target-mips_24kc_musl/usr/include/mac80211/net/cfg80211.h
That path is __FILE__ in .rodata, so the strip before the package is built
cannot reach it and the module ships it.
Map STAGING_DIR as well. Keep the staging_dir element in the replacement:
the last element of STAGING_DIR and of BUILD_DIR carry the same name, so
the bare name would make the two roots indistinguishable. The map of
BUILD_DIR does not change, so only a path that leaks today changes.
Verified with the target compiler on a header below a mapped directory:
-ffile-prefix-map removes the absolute path from both __FILE__ and the
debug information, -fmacro-prefix-map removes it from __FILE__ alone. The
shipped half is __FILE__, so both variants cover it.
A local build of mt76 also holds nine staging paths, but those are debug
information: the build directory copy of mt76-connac-lib.ko is 653456
bytes with 13 debug sections and nine such paths, the shipped copy is
84688 bytes with none. They are named here so that the next reader does
not take them for the same defect.
Signed-off-by: John Crispin <john@phrozen.org>
A configure run spends its time on the compiler invocations behind its
checks, not on the shell. tools/m4 runs the compiler 633 times and takes
24 seconds. ccache does not help, because it refuses to cache an
autoconf conftest compile at all; it counts these separately as
"Autoconf compile/link" under its uncacheable calls.
A package that sets PKG_CONFIGURE_CACHE or HOST_CONFIGURE_CACHE keeps a
config.cache under tmp/, so a package that is cleaned and built again
reads the answers instead of running the checks. The configure stage of
tools/m4 drops from 24.5 s to 1.5 s. dirclean removes tmp/, so a full
reset of the tree starts from cold answers.
A package has to opt in, because a cache is only correct for a configure
script that keeps its side effects outside the AC_CACHE_VAL body.
ncurses appends -D_XOPEN_SOURCE and -D_DEFAULT_SOURCE to CPPFLAGS inside
that body, so a cached run restores the value, skips the append, and
then fails to compile wcwidth. ncurses_cfg.h is identical in both cases;
the difference is in the generated Makefile.
The path covers everything else that changes an answer. A target package
is keyed on the toolchain directory, which carries the gcc version. A
host package is keyed on the version string of the host compiler,
because autoconf cannot detect a new compiler behind an unchanged name;
gcc 14 turned an implicit function declaration into an error and changed
many answers that way. Both are keyed on a hash of the configure
arguments and the build flags, and autoconf refuses a cache written with
different flags on top of that.
A cached program path outlives a wipe of staging_dir, so ac_cv_path_ and
ac_cv_prog_ answers are dropped when the cache loads and configure
searches PATH again. Without that, a cache written before staging_dir
held sed pins /usr/bin/sed for good. A search runs no compiler, so this
costs nothing that matters.
The cache arrives through a site script rather than --cache-file, so a
hand written configure script never sees a new argument. The site script
removes the file of the previous arguments, so a different configuration
gets a cold cache instead of an autoconf error. It also unsets the
variable, and make exports it for the configure stage alone, so a nested
configure run keeps its own cache.
Set CONFIGURE_CACHE_DIR to keep the cache elsewhere, or to the empty
string to run every check again.
Host builds now read a site script on Linux as well. Before this change
only Darwin had one.
CONFIG_SITE holds one path name: autoconf before 2.70 reads the whole
value as that one name, so a second path would silence both scripts. It
names the cache script alone, and the cache script loads the site script
of the architecture from CONFIG_SITE_BASE.
Signed-off-by: John Crispin <john@phrozen.org>
On a rebuild, the initrd and the initramfs kernel image usually have
the same content as before, but xz and lzma compress them again on
every build. For the OpenWrt One this takes about 6 s of the build.
Add scripts/cache-run.sh. It runs a command that creates one file from
another, and stores the output together with a hash of the input, the
command line and the tool. If they are the same in the next run, it
copies the stored output instead. The commands are deterministic, so
the result is identical.
On a hit the script compares the stored output with the file that is
already there and copies only when the two differ. The output then keeps
its timestamp, so a later step that compares timestamps does not treat it
as new work.
Use it for the initrd compression and for Build/lzma. bzip2 and lzop
stay as they are: bzip2 writes to stdout, and lzop stores the mtime of
its input. xz and gzip now keep the cpio, as the other tools do, so a
cache hit leaves the same files behind as a normal run.
The slot name follows the path of the output below $(KDIR), because two
targets in different directories can share a basename. cache-run.sh holds
a lock over the slot and swaps the stored output in with a rename, so two
image recipes on one slot never read a half written pair.
Signed-off-by: John Crispin <john@phrozen.org>
buildversion, feedsversion and json_overview_image_info rewrite
version.buildinfo, feeds.buildinfo and profiles.json on every build. The
content is the same, but the new timestamp makes the checksum step hash
every image again, which takes 0.18 s for the OpenWrt One and grows with the
number of devices.
Write the three files to the temporary directory and copy them over the
result of the last build only when the content differs.
This takes the time of the index, image info and checksum step from 0.36 s
to 0.10 s, and the time of a full no-op build from 3.25 s to 3.00 s.
Signed-off-by: John Crispin <john@phrozen.org>
sha256sums hashes every file of a directory on each build. package/index
calls it over the package output of this tree, which holds 1258 files,
and spends 0.7 s of a no-op build on a file that does not change.
Skip the pass while the file exists and nothing in the directory is
newer than it.
The checksum target of a build calls sha256sums as well, but it keeps
writing the file on every build. buildversion and feedsversion rewrite
version.buildinfo and feeds.buildinfo beside it without a guard, so the
directory always holds something newer. Only package/index gains here.
Signed-off-by: John Crispin <john@phrozen.org>
Set BUILD_TIME_LOG=<file> to record a begin and an end event for every
prepare, configure, compile and install stage of every tool, toolchain
component, kernel and package. scripts/build-time-report.pl turns the log
into a report.
The report divides each second of the build between the stages that run
in it, so the numbers show which stages hold up the wall clock rather
than which ones burn the most CPU. Nothing is recorded and no extra
process runs when BUILD_TIME_LOG is unset.
Signed-off-by: John Crispin <john@phrozen.org>
confvar runs a shell and mkhash every time make expands a stamp name that
contains it, 21 times in a make call for a package with a host build,
also if nothing has to be done.
The config symbols do not change during a make call, so keep the hash
of each list of symbols after the first use.
Signed-off-by: John Crispin <john@phrozen.org>
NPROC is a recursive variable, so every use starts a shell. The image
Makefile expands the initramfs rule once per device, which runs nproc
about 200 times and makes the parse of the mediatek image Makefile take
about 1 s. It is parsed twice per build.
Evaluate it on first use and keep the result.
Signed-off-by: John Crispin <john@phrozen.org>
YEAR_2038 writes its test program to $(TMP_DIR)/year2038.c and deletes
it after the test. Packages that are built in parallel run the test at
the same time and use the same file, so one of them can compile a
partly written file or find no file. The test then fails, and the host
build is configured with --disable-year2038.
Pass the test program to the compiler on stdin instead.
Signed-off-by: John Crispin <john@phrozen.org>
Stop overriding CCACHE_COMPILERCHECK from rules.mk and rely on
ccache's default, which hashes the compiler binary's mtime and
size.
For a local tree that is enough: the toolchain only changes when
it is rebuilt, and a rebuild updates mtime and size. Users who
want a stricter check can still set CCACHE_COMPILERCHECK in their
environment.
CI does not depend on this export either. The workflow writes a
secondary ccache.conf with
compiler_check=string:<toolchain-commit-sha>
so the cache key is tied to the toolchain source revision. That
is both stricter than the previous "%compiler% -v -c" setting
and portable across runners, where compiler mtimes would not
match after restoring a cache archive.
Link: https://github.com/openwrt/openwrt/pull/22995
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add DEPRECATION NOTICE for AUTORELEASE to signal that any usage of this
should be dropped and fixed.
The macro is still defined for legacy reason but we at least dropped any
usage of it in both main and feeds.
Link: https://github.com/openwrt/openwrt/pull/20586
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Not hashing CWD is potentially unsafe since it involves deliberately
poisoning the cache in certain situations in exchange for performance
gain. It can lead to debug information pointing out either no longer
existing or much worse incorrect source files, possibly leading
developers onto a false track and wasting a lot of time.
If one wishes to save build time by sharing the cache between multiple
source trees, this can be achieved safely by enabling reproducible
debug information, like this:
CONFIG_CCACHE_DIR="$(HOME)/.ccache"
CONFIG_REPRODUCIBLE_DEBUG_INFO=y
Note that CWD hashing gets disabled implicitly when reproducible debug
information is enabled. The CCACHE_NOHASHDIR option is only for
disabling CWD hashing in cases where it is not safe to do so.
Signed-off-by: Erik Karlsson <erik.karlsson@iopsys.eu>
Link: https://github.com/openwrt/openwrt/pull/20317
Signed-off-by: Robert Marko <robimarko@gmail.com>
Parsing "git log" is fragile. The actual output depends on both global and
local configuration files. Enabling "log.showSignature" makes "git log" prefix
signed commits with multiple lines of gpg verify output, regardless of the
configured log format.
Add "--no-show-signature" to "git log" commands to work around this particular
issue.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Link: https://github.com/openwrt/openwrt/pull/20127
Signed-off-by: Robert Marko <robimarko@gmail.com>
Update the compiler check for ccache so we don't
end up with the wrong binaries. Right now the
compiler check will not be able to correctly
distinguish the compiler used for build
ARMv8 binaries from the one used to build
ARMv7 binaries.
Signed-off-by: Markus Gothe <markus.gothe@genexis.eu>
Link: https://github.com/openwrt/openwrt/pull/16290
Signed-off-by: John Crispin <john@phrozen.org>
- Disable compression for ccache's cached files.
- Disable the hashing of the CWD inside debug information. This
increases the cache hits drastically.
Signed-off-by: Markus Gothe <markus.gothe@genexis.eu>
Add comments to build system makefile functions and variables to help
developers in understanding build system internals and ease the
development process.
This patch adds some documentation examples with proposed doxygen-like
syntax. Hopefully, this would start the discussion and result in
generation of the makefile documentation guidelines.
Signed-off-by: Sergey Matsievskiy <matsievskiysv@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16888
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Permit to overwrite PACKAGE_DIR and PACKAGE_DIR_ALL variables in
rules.mk.
This is to handle a special case with the ImageBuilder where these
variable are overwrite.
The main problem is that any include calling rules.mk again (example
image.mk) will set these variables again dropping the modified value.
To keep the modified value, set the PACKAGE_DIR and PACKAGE_DIR_ALL only
if not already set. This permits the ImageBuilder to use custom
directory instead of the default one defined in rules.mk.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Usage of g++ should mimic the usage of gcc,
otherwise, part of a binary or library may have
optimizations and part of it may not,
unbeknownst to the users or developers working on a build.
This can lead to some features like, for example, FORTIFY_SOURCE,
to have less of an effect or even cause a build error on some hosts.
Therefore, let HOST_CXXFLAGS default to HOST_CFLAGS.
Fixes: 87d489f67 ("build: add HOST_CXXFLAGS for host build")
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Link: https://github.com/openwrt/openwrt/pull/15853
Signed-off-by: Robert Marko <robimarko@gmail.com>
Several GNU tools such as tar, coreutils, and findutils
now build with support for 64-bit time by default
and otherwise require reconfiguring with a flag
--disable-year2038 in order to build without 64-bit time.
Some standard C libraries, for example,
certain older versions of glibc such as 2.31
have large file support but not long time bits support:
checking for ... option to enable large file support... -D_FILE_OFFSET_BITS=64
checking for ... option for timestamps after 2038... support not detected
This test using C code taken from largefile.m4 in gnulib
uses math and casting to check for overflow
with a macro and array pair that can only be defined
when 64-bit time support is present, and otherwise errors.
It is the exact same code used to test for 64-bit time
during the configure stage of building these tools,
so the results of this test before configure takes place
will always be in concordance with the results of
the test that takes place during the configure script.
Based on the test, the configure flag --disable-year2038
is added to every host tool build depending on the host system.
When the year 2038 problem finally comes around,
the effect of the test can be converted
from the toggling of a configure option into a build prerequisite,
requiring it to pass in order to continue building.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Link: https://github.com/openwrt/openwrt/pull/15799
Signed-off-by: Robert Marko <robimarko@gmail.com>
A new option called `USE_APK` is added which generated APK packages
(.apk) instead of OPKG packages (.ipk).
Some features like fstools `snapshot` command are not yet ported
Signed-off-by: Paul Spooren <mail@aparcar.org>
Currently when external toolchain is used, a info.mk is created (with
to-be-filled values) but is never actually filled with real values and
is never actually used in rules.mk.
This info.mk is used down the code with special packages like libgcc to
reference info like gcc version and package special library.
To mimic what is done with internal toolchain, add the include in
rules.mk for info.mk also for external library and in fix
toolchain/wrapper to fill the staging_dir info.mk.
The logic is to check if the external toolchain provide an info.mk and
if it doesn't we at least fill the GCC_VERSION with the value set in the
.config file.
With this special library like libgcc correctly reference and have set the
GCC_VERSION variable.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Make toolchain dirs define more consistent between internal and external
toolchains.
Make use of specific dirs also for intenral toolchain and generilize
include and lib inclusion.
Also set TOOLCHAIN_ROOT_DIR for internal toolchain as this is what
packages should use to reference staging toolchain directory.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
The current minimum OS requirement for OpenWrt is Ubuntu 18.04, which
includes 7 and 8. 8 is necessary for ccache.
gcc and g+++ are now symlinked to staging_dir, similar to Python.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
sstrip only has one functional arg. Make that a bool option, which can
easily depend on other knobs then.
This is required to be disabled for the mold linker.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Add a variable that stores the original value of $PATH
in the host system's shell, before Make alters it.
This can be useful for when it is necessary
to ignore symlinks and programs made by the build system.
Define this new variable before all instances of
'export PATH:=' or similar.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
By having a local copy of gnulib, we can:
import the latest macro fixes into any package,
get rid of some statically stored macros that were otherwise missing,
bootstrap GNU tools with the latest relevant source
without having to wait for a release or rely on git submodules,
and possibly more...
The patch assists in bootstrapping by ignoring
the building of po files using gettext,
and also to allow a user-defined path to a program
to include parameters.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Some uses cases, like with autotools, need a path for 'true'
if we use them to force something to skip.
This will work by default on Linux hosts,
and require MacOS hosts to get coreutils,
which is currently installed in the CI
for 'macos-latest' host.
In the future, prereq stage can be reworked
to search for the actual binary
instead of relying on env.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Some packages using Libtool have the possibility
of using this environment variable if set,
instead of the direct command for "file" using PATH,
or the need to patch a hard-coded path.
This is a new feature of Libtool 2.4.7
Ref: bf261073d ("tools/libtool: bump to 2.4.7")
Signed-off-by: Michael Pratt <mcpratt@pm.me>
These wrappers are not needed as CC doesn't need to be a single word.
a53b084e49 which introduced the wrappers
doesn't explain why they were really needed and why only for the target
and not for the host.
Moreover, name of the wrappers breaks a ccache assumption: since
v4.0-3-g6a92b4cd3a67 it has special handling for "chained" invocation
such as "ccache ccache gcc" where it skips all the "ccache*" names in
the middle and proceeds to run as if it was started as "ccache
gcc"[1][2].
This becomes important when a build system sees ccache in the PATH and
automatically enables it by prepending to CC. An example of such a
system would be autosetup as used by jimtcl. With the wrappers it breaks
as the command line ends up being just "ccache -Os..." because
"ccache_cc" gets skipped as it starts with "ccache".
[1] https://github.com/ccache/ccache/blob/master/src/ccache.cpp#L2105
[2] https://github.com/ccache/ccache/blob/master/src/Util.cpp#L802
Reported-by: Karl Palsson <karlp@etactica.com>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
STAGING_DIR may be provided from command line. We currently hardcoded
STAGING_DIR_HOST and STAGING_DIR_HOSTPKG to the default location but we
currently have some relocatable patch that derive the path from
STAGING_DIR.
Fix this and correctly derive STAGING_DIR_HOST and STAGING_DIR_HOSTPKG
from STAGING_DIR.
The intention is to fix inconsistency from the relocatable patch and the
use of STAGING_DIR_HOST that is always hardcoded.
This with a wrong configuration may end up in broken state with some
host tools expecing a PATH from STAGING_DIR and others using library
from the default staging_dir/host path.
To save downstream project the original implementation is saved while
fixing the inconsistency between patch and .mk.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
We currently enable -Wno-error=unused-but-set-variable and
-Wno-error=unused-result by default on every compile package.
While this is (relatively) unharmful, we should follow other project
direction and starts enforcing good code quality. For example the linux
kernel recently started to enforce Wall by default and clean code is
mandatory for inclusion.
Drop for good these flags and and make it mandatory to correctly handle
return values at least with a warning log if they are not strictly error
condition.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
To generate commitcount we use grep --max-count. This is not present on
alpine grep and cause wrong generation. Use -m as it's just the short
version of --max-count and more portable.
Fixes: #11200
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Move DOWNLOAD_CHECK_CERTIFICATE to include/download.mk as it's a better
place than exporting it in the global rules.mk makefile.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Some packages download a single raw file or set of raw files
instead of an archive.
The filenames of these download files have names and version numbers
that do not directly correspond to the name and version of the package
as it is defined in its Makefile.
To better organize downloaded files, this adds the option of
creating a subdirectory in the download directory for these files
that can be named with package specific variables.
DL_DIR is now a recursively expanded variable
since the value now depends on whether DL_SUBDIR is defined.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Remove flags from wget and curl instructing them to ignore bad server
certificates. Although other mechanisms can protect against malicious
modifications of downloads, other vectors of attack may be available
to an adversary.
TLS certificate verification can be disabled by turning oof the
"Enable TLS certificate verification during package download" option
enabled by default in the "Global build settings" in "make menuconfig"
Signed-off-by: Josh Roys <roysjosh@gmail.com>
[ add additional info on how to disable this option ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Don't add wrapped bin to the TARGET_PATH as it does cause compilation
error.
cmake.mk will use the "command -v" and will use the wrapped bin instead
of the external toolchain bin as they have the same name and command
will select the first result.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
When using the OpenWrt toolchain as an external toolchain the build
failed due to missing LTO support. By choosing the GCC wrappers of
the tools this commit makes sure that the LTO-enabled executables
are being used.
Signed-off-by: Vincent Wiemann <vincent.wiemann@ironai.com>
[ wrap the commit description to 72 char ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Compiler option -no-plt will break kernel builds on some architectures
eg. (x86) Filter this option from the recently introduced handling of
KCFLAGS vs EXTRA_OPTIMISATION
Fixes: 1d42af720c ("kernel: use KCFLAGS for passing EXTRA_OPTIMIZATION flags")
Suggested-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
On recent macOS, /usr/bin/python3 is a wrapper that finds the right python executable
It checks argv[0] to determine if python2 or python3 should be called. Always execute
it as python3 to ensure it calls the right version
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Added HOST_CXXFLAGS to specify CXXFLAGS during host-compile
(e.g. to specify c++ standard: HOST_CXXFLAGS += -std=c++11)
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
ninja is faster at building cmake packages than make, and according to reports
also more reliable at handling parallel builds
This commit includes a patch that adds GNU make jobserver support, in order to
allow more precise control over the number of parallel tasks
Enable parallel build by default for packages using ninja
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Paul Spooren
John Crispin
Hauke Mehrtens
Christian Marangi
Erik Karlsson
Bjørn Mork
Shiji Yang
Konstantin Demin
Felix Fietkau
Markus Gothe
Sergey Matsievskiy
Michael Pratt
Rosen Penev
Andre Heider
Paul Fertser
Josh Roys
Vincent Wiemann
Kevin Darbyshire-Bryant
Sergey V. Lobanov