| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
In v5.4 74d5f3d06f70 ("tools build: Add capability-related feature
detection") introduced libcap support. Disable it to keep perf's
features consistent and the number of dependencies small.
Signed-off-by: Bastian Krause <bst@pengutronix.de>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
PTXCONF_KERNEL_ARCH_STRING is used for cases when for a given ptxdist
architecture multiple kernel architectures exist. This only happens on
more than 10 year old Kernels, it's time to remove this compatibility
cruft. We can now use GENERIC_KERNEL_ARCH which has the same purpose and
is used in other places already, but doesn't handle the multiple kernel
architecture case.
PTXCONF_KERNEL_ARCH_STRING was useful on PowerPC where the kernel
architectures "ppc" and "powerpc" existed and on x86 where we had "i386"
and "x86_64". These are stories from the old days, since v2.6.24 we only
have "x86" for all x86 machines and since v2.6.27 we only have "powerpc"
for all powerpc machines.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
| |
It's the upstream, so we can always get a recent enough version.
And it provides more tools that will be needed to create fit images.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
| |
The path for the dummy initramfs was accidentally changes during the OOT
rework. As a result, it could not be created due to a missing directly.
Change it back to what it is supposed to be.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
This should really be the same anyways and some packages tend to be
forgotten when new options appear. So use one variable for all.
This adds missing options, e.g. for the kernel template.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
| |
The value should be the same. Always use PTXDIST_SYSROOT_* for
consistency.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This file has been touched 2 times in the last 7 years, and is no longer
up to date. Remove it to carry less unmaintained stuff around.
Also remove all references to it from file headers using
for file in `g grep -l 'See CREDITS for details about who has'`; do
sed -i '/See CREDITS for details about who has/,+1d' $file
done
as well as the different wording from rules/other/Toplevel.make.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
| |
This reverts commit 8f40148f4e9ee7199d213f37f84529a90d3f7cc8.
Implicit fallthrough errors are now avoided my the toolchain wrapper.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
| |
These are just a few files and tools/objtool/parse-options.c fails with
-Werror=implicit-fallthrough= errors.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The old implementation is broken, when a initramfs is used.
Changing the kconfig symbol back to '# Automatically set by PTXDist' is
broken because '$(<)' is not the kernel config any more (Broken since
cd920ab716bab0c190666e0273c10b19c7cdaeec "barebox/kernel: drop explicit
dependency").
Also, world/kconfig-sync compares .config and KERNEL_CONFIG and this does
not work because CONFIG_INITRAMFS_SOURCE in .config has been modified.
Instead, make sure that KERNEL_CONFIG contains the path independent dummy
text before copying it to .config.
Then run the default prepare stage that checks or updates the config.
Insert the path for the empty initramfs after all checks are done.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Tested-by: Jon Ringle <jringle@gridpoint.com>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It fails with:
In file included from [...]/usr/include/fcntl.h:313:0,
from iio_event_monitor.c:25:
In function 'open',
inlined from 'main' at iio_event_monitor.c:297:5:
[...]/fcntl2.h:44:5: error: call to '__open_too_many_args' declared with
attribute error: open can be called either with 2 or 3 arguments, not more
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
| |
it is compressed ELF image. Mostly interesting for MIPS platforms.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
With linux-4.19 and gcc-8.x oldconfig will prompt for extra options that
are not available with e.g. gcc-7.x. This causes problems when a BSP is
built with different host compilers.
Add a workaround to optionally disable these extra options by forcing the
compiler check to fail.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use _CONF_OPT/_MAKE_OPT/_INSTALL_OPT. This makes it possible to use more of
the generic build stages.
KBUILD_BUILD_* is now handled by the generic build stages.
Add some magic to detect external kernel modules that still use
KERNEL_MAKEVARS.
Drop the patch handling in clean. Patches are not applied if a local_src/
symlink is used.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
| |
This is now autogenerated by ptxd_lib_dgen.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
| |
It's obsolete and broken. dev packages should be used instead.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
| |
kernel-headers url is by default the same as kernel one, however
it does not reflect -rc tarballs are no longer provided. Fix that.
Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
| |
It's broken for some kernel versions (v4.15 and maybe others).
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Starting with 4.12-rc1, Linus no longer provides signed tarballs for
pre-release ("-rc") kernels. Download the version automatically generated
by cgit. For details, see:
https://www.kernel.org/rc-tarballs-and-patches-starting-with-412-rc1.html
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
"-C" breaks building with some toolchains. Building can fail with:
[...]/usr/include/bits/mathinline.h: In function 'floor':
[...]/usr/include/bits/mathinline.h:746:380: error: expected ':' or ')' before string constant
[...]
Instead disable -Werror to avoid build errors caused by icecc and
-Wimplicit-fallthrough.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Icecc splits preprocessing and compiling. As a result comments are no
longer available at compile time and building fails with:
parse-options.c: In function 'get_value':
parse-options.c:101:7: error: this statement may fall through [-Werror=implicit-fallthrough=]
parse-options.c:104:3: note: here
parse-options.c: In function 'print_option_help':
parse-options.c:716:6: error: this statement may fall through [-Werror=implicit-fallthrough=]
parse-options.c:719:2: note: here
Compile with '-C' to preserve comments. This can have side effects, so it
cannot be enabled globally.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Recent debugging tools support separate binaries and debug information
files. Providing a unique 'build-id' embedded in the binary enables the
debugging tools to select the correct and matching debug information
files from the system's root filesystem.
The 'build-id' is a 40 character calculated SHA1 from important sections in
the binary. The debug information files get stored at a central direcory at
"/usr/lib/debug/", with a special path based on the 'build-id'. If the
'build-id' is "123456789abcdef", the path and filename will be
".build-id/12/3456789abcdef.debug".
Its possible to have more than one 'build-id' in the binary. This
implementation prefers the first one in the list, which might be the
wrong one. It's unclear yet how the common debugging tools will behave in
this case.
Signed-off-by: Juergen Borleis <jbe@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
| |
With this change strip=y will no longer be specified for non elf files.
This will be needed for later changes, when any explicitly specified strip
argument is enforced.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
| |
Not-Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
|
|
|
|
|
|
|
|
| |
This symbol controls if a kernel xpkg is created. All kernel tool should
select this so that the xpkg is created. Kernel modules will end up in a
different xpkg anyways.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
search & replace for:
- GPL-1.0
- GPL-2.0
- GPL-3.0
- LGPL-2.0
- LGPL-3.0
- LGPL-2.1
- AGPL-3.0
- AFL-2.1
- Autoconf-exception-3.0
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Even if the developer enables the usage of prebuild packages, the Linux kernel
should not be part of it. The kernel is a too sensitive part and its build must
always be forced.
Signed-off-by: Juergen Beisert <jbe@pengutronix.de>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
| |
and the host wrapper will call the correct HOSTCC
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- TARGET_HARDEN_STACK:
the kernel has its own option CC_STACKPROTECTOR for this
- TARGET_HARDEN_FORTIFY:
this makes no sense for the kernel
- TARGET_HARDEN_RELRO, TARGET_HARDEN_BINDNOW:
seems to break some builds and is not useful for the kernel
- TARGET_HARDEN_PIE:
ASLR is not relevant for the kernel
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|