Mark Hatle [Wed, 13 May 2020 16:12:48 +0000 (11:12 -0500)]
sstate.bbclass: When siginfo or sig files are missing, stop fetcher errors
Prior to fetching, the system checks if the sstate file is present
either locally or on the mirror. If it is, then it goes to the fetch
stage. Up to three files can be fetched, sstate, sstate.siginfo and
sstate.sig (if signature validation is enabled).
The previous pstaging_fetch function would iterate over these, and if
a download error occurred would spew forth a great amount of fetcher
failure messages as well as stop fetching the next item in the set.
This was resolved by adding a fetcher.checkstatus() call prior to
the download. If the file isn't present, then the exception will
be triggered, and no fetcher failure messages will reach the user.
The exception handler is then modified to be a pass so that it will
loop and pull the rest of the files that that are requested.
Additionally, a check for the existance of the .sig file was added
to the sstate_installpkg to avoid an error trying to load the .sig
if it wasn't downloaded.
Signed-off-by: Mark Hatle <mark.hatle@xilinx.com> Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
this patch sequences the build targets so it can work reliably with
different kernel versions
Divide the target into scripts_basic scripts is not
strictly necessary and was simply what was used for
testing on kernel 4.1, which is quite an old kernel
perhaps just using scripts is sufficient, but it is not tested to not
known will cause the build race as seen above.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Cc: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Konrad Weihmann [Wed, 27 May 2020 18:07:58 +0000 (20:07 +0200)]
qemurunner: fix ip fallback detection
When falling back from detecting ip from /proc/./cmdline the
output of runqemu is acutally
'Network configuration: ip=192.168.7.2::192.168.7.1::255.255.255.0'
which doesn't match the given regex and leading to run failure, although
IP is detectable.
Fix regex by inserting an optional 'ip=' prefix to first IP
Signed-off-by: Konrad Weihmann <kweihmann@outlook.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
qemu: enable virglrenderer and glx options subject to 'opengl' DISTRO_FEATURE
Note that to actually use accelerated GL passthrough, there are two options
1) a suitable frontend need to be also enabled - gtk+ and SDL both seem to work well.
Previously I struggled to make SDL work, but now it seems fine.
2) it is also possible to render off-screen with -display egl-headless option,
and see the output with a VNC viewer (for which, qemu needs to be started
with a VNC server):
$ runqemu kvm egl-headless publicvnc
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
bitbake.conf: propagate 'opengl' DISTRO_FEATURE to native/nativesdk from target
This will allow better control over native virgl/qemu configurations.
Adjust gtk+3/cairo native configurations to actually ignore opengl
when building for -native: we do not need it, and it would cause build
failures as only a limited subset of mesa-native is currently built.
Drop native/nativesdk overrides from virglrenderer/libepoxy recipes
as opengl feature is now correctly set for those variants.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 15 Apr 2020 15:25:12 +0000 (16:25 +0100)]
multilib/recipes: Use new RecipePostKeyExpansion event
There are issues with multilib due to the ordering of events where some
functions see the remapped multilib dependencies and some do not. A significant
problem is that the multilib class needs to make some changes before key expansion
and some afterwards but by using existing event handlers, some code sees things
in a partially translated state, leading to bugs.
This patch changes things to use a new event handler from bitbake which makes the
ordering of the changes explcit.
The challenge in doing this is that it breaks some existing anonymous python and
dyanmic assignments. In some cases these used to be translated and no longer are,
meaning MLPREFIX has to be added. In some cases these are now translated and the
MLPREFIX can be removed.
This change does now make it very clear when MLPREFIX is required and when it is
not, its just the migration path which is harder. The patch changes the small number
of cases where fixes are needed.
In particular, where a variable like RDEPENDS is conditionally extended (e.g.
with an override), MLPREFIX is now required.
This patch also reverts:
base: Revert 'base.bbclass: considering multilib when setting LICENSE_EXCLUSION'
Richard Purdie [Tue, 26 May 2020 20:50:10 +0000 (21:50 +0100)]
ltp: Exclude the memcg_stress tests due to timeout problems
This test runs for 900s, we often see tests killed after 300s without
output which makes the test results unreliable and inconsistent. The
easiest solution for now is to skip this long running test, patching
it out wth sed.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 26 May 2020 14:48:38 +0000 (15:48 +0100)]
ltp: Add missing dependencies on coreutils, bc, e2fsprogs and gdb
When the tests are run we see messages like:
/opt/ltp/testcases/bin/run_cpuctl_stress_test.sh: line 242: nice: command not found
/opt/ltp/testcases/bin/run_cpuctl_test_fj.sh: line 66: tac: command not found
vma05 1 TCONF: 'gdb' not found
memcg_failcnt 1 TCONF: 'bc' not found
Owner=nobody; perms=-rw-------; sudo: lsattr: command not found
so add missing dependencies to avoid these.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Daniel Díaz [Mon, 25 May 2020 17:07:06 +0000 (12:07 -0500)]
ltp: Add net-tools as rdependency
The version of ifconfig provided by Busybox is not enough for
some tests. Here's an example of how it fails with test
netns_breakns_ns_exec_ipv4_ioctl:
ifconfig: bad address '192.168.0.2/24'
ifconfig: bad address 'inet6'
This works with net-tools' ifconfig.
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Fri, 22 May 2020 02:22:34 +0000 (23:22 -0300)]
go-mod.bbclass: Add class for `go mod` support
When using Go Modules, the the current working directory MUST be at or
below the location of the 'go.mod' file when the go tool is used, and
there is no way to tell it to look elsewhere. It will automatically
look upwards for the file, but not downwards.
To support this use case, we provide the `GO_WORKDIR` variable, which
defaults to `GO_IMPORT` but allows for easy override.
[YOCTO #13883]
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Michael Ho [Mon, 25 May 2020 08:23:01 +0000 (10:23 +0200)]
package_rpm.bbclass: add PACKAGE_ADD_METADATA_RPM to sstate variables
Refactor the addition of vardeps for this package class to use RPMEXTRAVARS
and pass the items through gen_packagevar so package name specific overrides
are accounted for. This matches the style used in package_ipk.bbclass. Since
do_package_write_rpm does not have any "extra vars" that affect the rpm
packaging yet, this commit has to add the variable.
This refactoring is needed to add PACKAGE_ADD_METADATA_RPM to RPMEXTRAVARS
so it affects the sstate hash of do_package_write_rpm.
Signed-off-by: Michael Ho <Michael.Ho@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Michael Ho [Mon, 25 May 2020 08:23:00 +0000 (10:23 +0200)]
package_deb.bbclass: add PACKAGE_ADD_METADATA_DEB to sstate variables
Refactor the addition of vardeps for this package class to use DEBEXTRAVARS
and pass the items through gen_packagevar so package name specific overrides
are accounted for. This matches the style used in package_ipk.bbclass.
Additionally add PACKAGE_ADD_METADATA_DEB to the DEBEXTRAVARS list so the
packaging is redone if the variable changes value.
Signed-off-by: Michael Ho <Michael.Ho@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Michael Ho [Mon, 25 May 2020 08:22:58 +0000 (10:22 +0200)]
package.bbclass: add PACKAGE_ADD_METADATA to sstate variables
The PACKAGE_ADD_METADATA can be used to define additional metadata for
packages. Changes to this variable should affect the packaging sstate hash
so packages are re-generated when this variable changes.
This variable is added to both PKGDATA_VARS and PACKAGEVARS. It is needed
in PACKAGEVARS to ensure changes invoke do_package to re-run. It is needed
in PKGDATA_VARS to ensure changes are written into the pkgdata so that the
hash equiv reports a change and does not skip over the package_write tasks.
Signed-off-by: Michael Ho <Michael.Ho@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Michael Ho [Mon, 25 May 2020 07:45:49 +0000 (09:45 +0200)]
package_rpm.bbclass: respect package overrides for the main package
Apply ${PN} to OVERRIDES when determining the base package spec variables.
Without this, there is a mismatch in behaviour where overrides for the base
package have no effect while overrides for subpackages do. The subpackages
should not be affected by this override as they apply their own package
overrides and will by default fetch the base variables and apply them in
the spec file.
This makes the behaviour for package overrides match the implementation
in package_deb.bbclass and package_ipk.bbclass.
Signed-off-by: Michael Ho <Michael.Ho@bmw.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Thu, 21 May 2020 04:05:56 +0000 (21:05 -0700)]
archiver.bbclass: Fix duplicated SRC_URIs for do_ar_original
The argument urls of bb.fetch2.Fetch(urls, d) are duplicated to SRC_URI, which caused errors like:
bb.data_smart.ExpansionError: Failure expanding variable SRCPV, expression was ${@bb.fetch2.get_srcrev(d)} which triggered exception FetchError: Fetcher failure: The SRCREV_FORMAT variable must be set when multiple SCMs are used.
The SCMs are:
git://github.com/docker/notary.git;destsuffix=git/src/github.com/docker/notary
git://github.com/docker/notary.git
The first one is from original SRC_URI, the second one is from the
variable 'urls', so cleanup SRC_URI before call bb.fetch2.Fetch() can fix the
problem.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Fri, 22 May 2020 22:08:26 +0000 (17:08 -0500)]
checklayer: Skip layers without a collection
As in other places in the file, skip layers that don't define a
collection when searching for a layer to resolve a dependency. Fixes
KeyError exceptions when attempting to access the layer collections
later
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It adds meson options 'instrospection' to disable gobject introspection
files build. So no need to check distro feature 'gobject-introspection-data'
any longer.
Keep the spaces adjustment and comment removal.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kai Kang [Fri, 22 May 2020 07:00:22 +0000 (15:00 +0800)]
libsecret: add meson option introspection
Add meson option 'introspection' for libsecret. For bsp which doesn't
support qemu usermode, it could disable gobject introspection build.
Then remove distro feature check for 'gobject-introspection-data' too.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Fri, 22 May 2020 18:23:08 +0000 (02:23 +0800)]
boost: upgrade 1.72.0 -> 1.73.0
refresh 0001-Don-t-set-up-arch-instruction-set-flags-we-do-that-o.patch
remove 0001-revert-cease-dependence-on-range.patch
since it is included in 1.73.0
Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 21 May 2020 15:18:47 +0000 (08:18 -0700)]
cve-check: Run it after do_fetch
Certain recipes e.g. bash readline ( from meta-gplv2 ) download patches instead of having them in
metadata, this could fail cve_check
ERROR: readline-5.2-r9 do_cve_check: File Not found: qemuarm/build/../downloads/readline52-001
This patch ensures that download is done before running CVE scan, even
though these will be external patches and may not contain CVE tags as it
expects, but it will fix the run failures as seen above
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
file: add bzip2-replacement-native to DEPENDS to fix sstate issue
file-native when built on a Debian 10 host will embed a dependency to
'libbz2.so.1.0' (instead of 'libbz2.so.1'). This can cause issues
when sharing the sstate between hosts e.g.:
recipe-sysroot-native/usr/lib/rpm/rpmdeps:
error while loading shared libraries: libbz2.so.1.0: \
cannot open shared object file: No such file or directory
To avoid this situation, let's add the bzip2-replacement-native to the
file recipe's DEPENDS_class-native .
Details in https://bugzilla.yoctoproject.org/show_bug.cgi?id=13915 .
Signed-off-by: Jan-Simon Moeller <dl9pf@gmx.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Wed, 20 May 2020 16:31:09 +0000 (17:31 +0100)]
avahi: Don't advertise example services by default
The example service files are placed into /etc/avahi/services when we
run `make install` for avahi. This results in ssh and sftp-ssh services
being announced by default even if no ssh server is installed in an
image.
These example files should be moved away to another location such as
/usr/share/doc/avahi (taking inspiration from Arch Linux).
Signed-off-by: Paul Barker <pbarker@konsulko.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Inherit also pkgconfig (required for TI-RPC and Netlink based route
tests), to fix:
configure:6045: error: possibly undefined macro: AC_DEFINE
Refreshed patches (some tests have been fixed for musl)
* 0001-build-Add-option-to-select-libc-implementation.patch
(add a note here)
* 0001-Add-more-musl-exclusions.patch
Removed patches (different fix in upstream)
* 0003-Check-if-__GLIBC_PREREQ-is-defined-before-using-it.patch
devtool: do not write md5sums into upgraded recipes
This will drop them md5sums from recipes that still have them,
and will not re-add them for recipes where they're already
removed.
[RP: Added fix for the test recipe] Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
base/insane: Check pkgs lics are subset of recipe lics only once
Move logic checking that all packages licenses are only a subset of
recipe licenses from base.bbclass to the insane.bbclass so that it's
evaluated only once, during do_package_qa.
As explained in the linked bugzilla entry, if a package license is not
part of the recipe license, the warning message gets shown an
unreasonable amount of time because it's evaluated every time a recipe
is parsed.
[YOCTO #10130]
This also makes it possible to silence this error with INSANE_SKIP.
Signed-off-by: Quentin Schulz <quentin.schulz@streamunlimited.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Konrad Weihmann [Mon, 18 May 2020 19:28:07 +0000 (21:28 +0200)]
lib/oe/recipeutils.py: passthrough of FETCHCMD
let FETCHCMD_* settings be passed to upstream check functions from
get_recipe_upgrade_status. It enables using different values for
the fetching tools, as otherwise always the defaults are used.
E.g. use different timeout or retry values for wget.
This is especially useful for remote server that do request-limiting (429)
or that are simply unavailable (500, 504) and can speedup upgrade-check
for a larger set of recipe significantly
Signed-off-by: Konrad Weihmann <kweihmann@outlook.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ovidiu Panait [Tue, 19 May 2020 08:43:09 +0000 (11:43 +0300)]
lttng-tools: upgrade to 2.12.0
Upgrade lttng-tools 2.11.2 -> 2.12.0:
* drop x32 patch, as it was merged upstream
* update ptest file mi-lttng-3.0.xsd -> mi-lttng-4.0.xsd
* add local patch to fix ptest build failure on musl
* License-Update: licenses were moved under LICENSES directory and changed in
upstream commit [1]
Mingli Yu [Mon, 18 May 2020 05:33:36 +0000 (13:33 +0800)]
python3-setuptools: add the missing rdepends
Add the missing rdepends to fix below error:
# python3
[snip]
>>> import setuptools.lib2to3_ex
[snip]
ModuleNotFoundError: No module named 'lib2to3'
ModuleNotFoundError: No module named 'pickle'
Signed-off-by: Mingli Yu <mingli.yu@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Mon, 9 Mar 2020 14:21:36 +0000 (14:21 +0000)]
archiver.bbclass: Make do_deploy_archives a recursive dependency
To ensure that archives are captured for all dependencies of a typical
bitbake build we add do_deploy_archives to the list of recursive
dependencies of do_build. Without this, archives may be missed for
recipes such as gcc-source which do not create packages or populate a
sysroot.
do_deploy_archives is also added to the recursive dependencies of
do_populate_sdk so that all sources required for an SDK can be captured.
Signed-off-by: Paul Barker <pbarker@konsulko.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Diego [Thu, 14 May 2020 22:26:14 +0000 (00:26 +0200)]
bmap-tools: update development snapshot revision
Update from a previous development snapshot to the latest development snapshot
that includes, among other improvements, support for zstd compression format.
Signed-off-by: Diego Rondini <diego.rondini@kynetics.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Diego [Thu, 14 May 2020 22:16:02 +0000 (00:16 +0200)]
image_types: declare support for wic.zst
Declare images in wic.zst format as supported, as bmaptool now supports
zstd:
https://github.com/intel/bmap-tools/commit/1b8437d58447d3796dc11fd0f2c62bec5746e5d6
Given the very fast decompression speed of zstd over, for example, gzip,
writing images with bmaptool can provide big write speed improvements
when decompression speed is the limiting factor (especially with very
sparse images).
Signed-off-by: Diego Rondini <diego.rondini@kynetics.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Sat, 16 May 2020 15:58:39 +0000 (08:58 -0700)]
glibc: Do not synthesize wordsize.h for arm multilibs
This has been constant source of trouble, because it is fundamental file
which sets machine word length and everything else builts on top of that
so when it is sythesized like this, where the sythesize template itself
needs wordsize.h to determine machine word length, it creates the
catch-22 problem, which is seen when building things like bpf, or
running clang-tidy etc. where compiler internal defines may not be used
this ends up in all sorts of problems. Now that glibc provides exact
same header for arm and aarch64, its no longer needed to be multilibbed
here
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Sat, 16 May 2020 00:45:45 +0000 (02:45 +0200)]
webkitgtk, pinentry: require gobject-introspection-data in DISTRO_FEATURES when libsecret PACKAGECONFIG is enabled
* gobject-introspection is mandatory since libsecret upgrade to 0.20.3
* enable libsecret in webkitgtk only when gobject-introspection-data
is in DISTRO_FEATURES
* drop comment about libsecret being in meta-gnome from pinentry
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 14 May 2020 17:30:09 +0000 (10:30 -0700)]
valgrind: Do not use outline-atomics on aarch64
This avoids the __getauxval undefined reference error seen with gcc10 on
doing static linking with -nodefaultlibs, which is uncommon usecase
anyway, disabling outline-atomics is not a huge deal for OE in terms of
performance as we already use -mcpu which is tuned enough to the SOC the
code is being generated for
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
