chore: update Buildroot to 2022.02 #148
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The easrc firmware is used for audio resampling. The firmware for now exists only for imx8mn device Signed-off-by: Michael Trimarchi <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Commit 56d9b88 forgot to add gcc >= 5 dependency to fastboot: Makefile:591: *** libsepol is in the dependency chain of libselinux that has added it to its _DEPENDENCIES variable without selecting it or depending on it from Config.in. Stop. Fixes: - http://autobuild.buildroot.org/results/2430ba4afb40569ed54506201494baeebf6c5d42 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
As suggested by Peter Seiderer package util-linux doesn't need raw support when built for host, so let's remove: HOST_UTIL_LINUX_POST_PATCH_HOOKS += UTIL_LINUX_FIX_DISK_UTILS_COMPILE in favor of: HOST_UTIL_LINUX_CONF_OPTS += --disable-raw This is a better approach than my commit: https://git.buildroot.net/buildroot/commit/?id=d895bd973b35fde10d3dd7db530d9e4782926497 Suggested-by: Peter Seiderer <[email protected]> Tested-by: Peter Seiderer <[email protected]> Reviewed-by: Peter Seiderer <[email protected]> Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Actually genimage fails to create final image with:
```
ERROR: hdimage(disk.img): partition [MBR] (offset 0x1b8, size 0x48) overlaps previous partition boot (offset 0x0, size 0x200)
ERROR: hdimage(disk.img): bootloaders, etc. that overlap with the partition table must declare the overlapping area as a hole.
```
This is because genimage 15 is less permissive than before, so let's add
holes = {"(440; 512)"} to leave room to MBR.
The generated image is identical to the image generated by the previous
version of genimage.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/1865935249
Signed-off-by: Giulio Benetti <[email protected]>
Tested-by: Giulio Benetti <[email protected]>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Changes: * Exclude FAT partition, it is no longer needed for U-Boot. * Switch the rootfs partition from MBR to GPT for partition-uuid to work. * Switch to extlinux, this is convenient for debugging. * Create common scenarios for multiple boards to create an SD image: board/orangepi/common. Currently only orangepi_zero makes use of the common infrastructure. Signed-off-by: Sergey Kuzminov <[email protected]> Reviewed-by: Giulio Benetti <[email protected]> Tested-by: Giulio Benetti <[email protected]> [Arnout: remove rootfs size = 63M and remount rw] Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
…o 5.15.21 and add host openssl for linux Fixes: https://gitlab.com/ymorin/buildroot/-/jobs/2035821098 Signed-off-by: Giulio Benetti <[email protected]> Tested-by: Eugen Hristev <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
…on to 5.15.21 and add host openssl for linux Fixes: https://gitlab.com/ymorin/buildroot/-/jobs/2035821100 Signed-off-by: Giulio Benetti <[email protected]> Tested-by: Eugen Hristev <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
When building package guile with BR2_ENABLE_LOCALE=n
we get the following error:
```
In file included from filesys.c:95:
ports-internal.h:26:10: fatal error: iconv.h: No such file or directory
26 | #include <iconv.h>
| ^~~~~~~~~
compilation terminated.
```
So let's fix it by selecting BR2_PACKAGE_LIBICONV and
adding libiconv package as dependency.
Signed-off-by: Pedro Aguilar <[email protected]>
Reviewed-by: Giulio Benetti <[email protected]>
Tested-by: Giulio Benetti <[email protected]>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Select BR2_TARGET_UBOOT_NEEDS_OPENSSL to fix the following build
error:
include/image.h:1178:12: fatal error: openssl/evp.h: No such file or directory
1178 | # include <openssl/evp.h>
| ^~~~~~~~~~~~~~~
Fixes:
- https://gitlab.com/buildroot.org/buildroot/-/jobs/2076475033
Signed-off-by: Clément Péron <[email protected]>
Signed-off-by: Peter Korsgaard <[email protected]>
Add rng tools. Without this, when booting, prompt will not appear until 'random: crng init done' appears. Avoid this by using rng tools without the libjitterentropy. Libjitterentropy is a jitter source of entropy which does not help for AT91 cpus as the jitter does not have enough variance to generate fast enough entropy. Signed-off-by: Eugen Hristev <[email protected]> Reviewed-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
…120M Default 60M for rootfs are not enough and generate a build failure, so let's expand rootfs size to 120M. Fixes: https://gitlab.com/ymorin/buildroot/-/jobs/2035821088 Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
https://github.com/jpmens/jo/blob/1.6/ChangeLog Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
Now that host-python2 is gone and the corresponding logic in pkg-python. Signed-off-by: Peter Korsgaard <[email protected]>
Now that host-python2 is gone and the corresponding logic in pkg-python. Signed-off-by: Peter Korsgaard <[email protected]>
Update indentation in hash file (two spaces) Gregwar/fatcat@v1.1.0...v1.1.1 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
The doc-asciidoc infra allows document to define sources and resources, and those can only be local (there is no download support for documents). The currently documented way to specify those is to use the $(pkgdir) macro to point to the document sources and resources when they are along the .mk. However, this works only for sources and not for resources. Indeed, sources are used to generate Makefile dependency rules, so $(pkgdir) is expanded and evaluated during the call to $(doc-asciidoc), while resources are expanded and evaluated when the rules are executed, which always happen after all the parsing has be done, at which point $(pkgdir) expands to the last Makefile that was parsed, which is always Buildroot's own manual. This went unnoticed so far because Buildroot's manual is 1) the last to be parsed, and 2) not using $(pkgdir) anyway. Additionally, $(pkgdir) is mostly an internal macro, that is, as this change demonstrates, difficult to use, and thus should not be used, in packages or in documents. We fix that by introducing $(FOO_DOCDIR), similar to $(FOO_PKGDIR), which points to the document directory. We choose not to name the variable FOO_PKGDIR, because it semantically is not a package. As a consequence, the last argument to the doc-asciidoc macro is now superfluous, so it is removed, which causes a little bit of churn in the main macro. As a further consequence, the third argument in the inner macro is no longer needed either, so it too is removed, which causes a bit more churn yet in the inner macro. Signed-off-by: Yann E. MORIN <[email protected]> Cc: Frederic GARDES <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Martin Hicks asked to not maintain cryptsetup anymore: https://lists.buildroot.org/pipermail/buildroot/2022-February/635413.html Also he doesn't mantain any other package, so let's remove him from this file. Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Signed-off-by: Andre Renaud <[email protected]> [Arnout: reorder .mk so conditional parts come after unconditional, and use += within conditions] Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Signed-off-by: James Hilliard <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
This will be needed by qt5webengine. Signed-off-by: James Hilliard <[email protected]> [Peter: explicitly disable gif/gl/jpeg/png/tiff support] Signed-off-by: Peter Korsgaard <[email protected]>
Webp nowadays has explicit configure options for jpeg/png/tiff, so use those to explicitly enable/disable support. Signed-off-by: Peter Korsgaard <[email protected]>
Webp has optional SDL support. Explicitly disable it for now. Signed-off-by: Peter Korsgaard <[email protected]>
Fix the following build failure raised since addition of icu support in libxml2 through commit 49d7c85: In file included from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/ptypes.h:52:0, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/umachine.h:46, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/utypes.h:38, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/ucnv_err.h:88, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/ucnv.h:51, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/bin/../../usr/include/libxml2/libxml/encoding.h:31, from /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/bin/../../usr/include/libxml2/libxml/parser.h:812, from conftest.cpp:61: /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/localpointer.h:224:41: error: 'noexcept' does not name a type LocalPointer(LocalPointer<T> &&src) U_NOEXCEPT : LocalPointerBase<T>(src.ptr) { ^ /home/peko/autobuild/instance-1/output-1/host/mips64el-buildroot-linux-gnu/sysroot/usr/include/unicode/localpointer.h:224:41: note: C++11 'noexcept' only available with -std=c++11 or -std=gnu++11 Fixes: - http://autobuild.buildroot.org/results/be043792e6e5501631517881b10096291a654364 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Add local patch pending upstream[1] to fix the build failure after my suggestion[2] to imitate what I've done with libfuse[3]. [1]: https://gitlab.com/cryptsetup/cryptsetup/-/merge_requests/275?commit_id=5f71b3d63181aa88a68f7f71eab8801f2d8d2cde [2]: https://gitlab.com/cryptsetup/cryptsetup/-/issues/709 [3]: libfuse/libfuse@3aba09a Fixes: http://autobuild.buildroot.net/results/bad/bad4d01fde3191e4d481a8e9dae18df3b1a94537/ Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Building galileo_defconfig throws error: ``` /usr/bin/ld: arch/x86/tools/relocs_64.o:(.bss+0x0): multiple definition of `per_cpu_load_addr'; arch/x86/tools/relocs_32.o:(.bss+0x0): first defined here ``` To fix this let's add an upstreamed patch[1] and add BR2_GLOBAL_PATCH_DIR to galileo_defconfig to point to where patch is. [1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=eeeda4cd06e828b331b15741a204ff9f5874d28d.patch Fixes: https://gitlab.com/ymorin/buildroot/-/jobs/2035821039 Signed-off-by: Giulio Benetti <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Remove the obsolete patch as docutils is now a setuptools package. The package also includes code licensed under BSD-3-Clause licence. COPYING.txt has the following changes: * copyright year * more exceptions from the Public Domain added * BSD-3-Clause license added * emacs mode is GPL-3.0+ instead of just GPL-3.0 * added Python-2.0 for roman.py The "other" license is removed since the list covers all licenses mentioned in COPYING.txt. Also update hash file whitespace to two spaces. Fixes: http://autobuild.buildroot.net/results/02dd94c09dabdbdd8b47673393577a7fd3af3aeb Signed-off-by: Yegor Yefremov <[email protected]> Signed-off-by: James Hilliard <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
Fix the following build failure raised since the addition of fourth patch in commit 8251d8c: 1 out of 22 hunks FAILED -- saving rejects to file CMakeLists.txt.rej Fixes: - http://autobuild.buildroot.org/results/44f6d7c61316e90d22e75cb1fb77c3bc5b31ad66 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
The cmake build appears to be missing features, such as pkg-config generation support which is needed for some packages to find re2, like qt5webengine. Upstream is not keen on fixing the CMake-based buildsystem [0] Instead of patching it on our side, and then have to carry the patch forever, we switch to the Makefile-based generic buildsystem. Note that previously some packages like grpc failed when the generic Makefile was used, because they would not find the cmake-equivalent of pkg-config, but this is now fixed [1]. [0] google/re2#349 [1] grpc/grpc@45e413d Signed-off-by: James Hilliard <[email protected]> [Peter: Reword commit message as suggested by Yann] Signed-off-by: Peter Korsgaard <[email protected]>
Signed-off-by: Angelo Compagnucci <[email protected]> Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <[email protected]>
…for > year 2038 support" This reverts commit 6e33e59. This unfortunately breaks a number of packages, as glibc errors out if 64bit time_t is used without 64bit file offsets, and some packages undefine _FILE_OFFSET_BITS leading to build breakage: # if ! defined (_FILE_OFFSET_BITS) || _FILE_OFFSET_BITS != 64 # error "_TIME_BITS=64 is allowed only with _FILE_OFFSET_BITS=64" .. So revert it for 2022.02. Signed-off-by: Peter Korsgaard <[email protected]>
Drop custom install rules which were added since commit 676797d. Indeed, they result in a broken installation. Especially, they are trying to "guess" what must be installed based on BR2_ARCH but oprofile has its own logic. For example, goldmont microarchitecture files must be installed in i386 directory even if this architecture is 64 bits: https://sourceforge.net/p/oprofile/oprofile/ci/0ad5a9e6af86a88e1dd41180f45bc48b646eba6a This will result in the following runtime failure: oprofile: could not open unit mask description file /usr/share/oprofile//i386/goldmont/unit_masks Unable to find info for event cpu_clk_unhalted Fixes: - https://bugs.buildroot.org/show_bug.cgi?id=14641 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
The default currently is to rely on the package name to decide what to build and install if not specified by the caller. This works nice for target packages, where a 'foo' package will by default build and install a 'foo' executable. However, for host packages, that will build and install a 'host-foo' exzcutable, which is not really, even really not, what would be expected. We fix that by using the package raw name, i.e. the package name with the host- prefix yanked away. It is very improbable that there already are many host-golang packages in the wild (in br2-external trees), but if there are, they would forcibly define those variables to a sane value. This change is not incompatible, as the values provided by packages take precedence; it's just that those packages now carry superfluous, if innocuous, variable assignments. Signed-off-by: Yann E. MORIN <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
This is a patch release for version 2 & 2.4, fixing a regression introduced in 2.4.1. https://github.com/cisco/libsrtp/releases/tag/v2.4.2 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
https://github.com/FreeRDP/FreeRDP/releases/tag/2.6.1 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
Fix the following build failure on armv7 and sparc raised since bump to version 3.7.1 in commit cc27267 and rurban/safeclib@b90c52b: In file included from perf_memcpy32_s.c:7: perf_private.h: In function 'rdtsc': perf_private.h:72:3: error: 'asm' undeclared (first use in this function) 72 | asm volatile("mrc p15, 0, %0, c9, c14, 0" : "=r"(pmuseren)); | ^~~ Fixes: - http://autobuild.buildroot.org/results/ceb13c071b1461eb6d73f5940d6b010095127f41 Signed-off-by: Fabrice Fontaine <[email protected]> Signed-off-by: Peter Korsgaard <[email protected]>
Signed-off-by: Peter Korsgaard <[email protected]>
sfoster1
force-pushed
the
buildroot-update-2022.02
branch
from
744602b to
f2486f1
Compare
RPI_WIFI_FIRMWARE was consolidated in 7527324 to the upstream brcmfmca_sdio-firmware-rpi and friends. Other rpi firmware options were also changed around and we need to accomodate all those changes.
In 2a057339cc124bbf7c6a3e567fd57413bed3dd5d the curl binary was moved out of its own package and into a suboption of libcurl.
The config option for setting a u-boot boot script was moved in db2b684 to host tools implementation. That means we need to move the config options and add host tools.
In glibc 2.35 the default locale C.UTF-8 was merged into source instead of being an external localegen file, so it doesn't have to (and can't) be compiled by localegen (https://sourceware.org/glibc/wiki/Proposals/C.UTF-8)
This has a couple different config elements than 3.7; explicitly enabling openssl lets us compile things like the ssl and hash modules. Since we have to use openssl anyway (and openssl 1.1.1 isn't as much of a horrorshow as it used to be) dump libressl. Some overlays that were into 3.7 now also need to be in 3.10.
Jupyter needs to import itself and many of its dependencies during build time which means those dependencies have to be available at build time. Most of these used to be but got lost in the merge.
Python 3.10 contains some recent PEPs that are intended to make the C API a little more friendly to having multiple instances in the same process; these took the form of making the inc/decref c macros not return lvalues anymore, which broke some extensions like pandas. Update pandas to a recent enough version that it can handle the change.
These were incorrect because the hashes didn't get updated in the merge.
It is not useful.
It's just too big. We can try and find a way to reenable it later, but let's get some builds going.
sfoster1
force-pushed
the
buildroot-update-2022.02
branch
from
f2486f1 to
a6deb98
Compare
This was referenced Apr 4, 2022
1 task
|
Superceded by #166 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There are a lot of changes in here! This updates buildroot to 2022.02, choosing newer versions than ours where conflicts occurred. I'm sure there's a lot more changes and definitely a lot more testing to come, but this is now a clean merge.
Major changes:
Testing
Environments
All below tests should be run on the following environments:
Tests