Andre McCurdy [Fri, 6 Jul 2018 05:07:02 +0000 (22:07 -0700)]
ca-certificates: avoid using += with an over-ride
Using += with an over-ride can be a source of confusion so try to
avoid the construct in core recipes.
In this case, the commit which added the over-ride seems to have been
buggy - the commit message mentions "add to SYSROOT_DIRS" rather than
a correct description of what the change actually did, ie "over-ride
SYSROOT_DIRS":
The commit also appears to have been unnecessary as ${sysconfdir} is
appended to SYSROOT_DIRS for -native recipes by default from within
staging.bbclass.
To workaround the bug introduced by the first commit, a subsequent
commit later added ${datadir}/ca-certificates to the over-ride value
(which would not normally be necessary as ${datadir} is included in
the default value of SYSROOT_DIRS - ie the value which was lost due
to being over-ridden):
Therefore the fix seem to be to remove the SYSROOT_DIRS over-ride
entirely - the default value of SYSROOT_DIRS set by staging.bbclass
includes both ${datadir} and ${sysconfdir} when building for -native.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Fri, 25 May 2018 02:54:46 +0000 (10:54 +0800)]
buildtools-tarball: add nativesdk-libnss-nis
Recent glibc change removed libnss-nis module from glibc and a new
recipe libnss-nis.bb was added.
After this change, we need to make sure nativesdk-libnss-nis is also
included in buildtools-tarball, otherwise, we may encounter the following
error when using 'tar' command from buildtools-tarball.
tar: relocation error: /lib/x86_64-linux-gnu/libnss_nis.so.2: \
symbol _nsl_default_nss version GLIBC_PRIVATE not defined \
in file libnsl.so.1 with link time reference
This error occured on my ubuntu16.04 host with 'nis' configured in
/etc/nssswitch.conf.
So add nativesdk-libnss-nis to buildtools-tarball to fix this problem.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Pablo Saavedra [Fri, 6 Jul 2018 09:50:23 +0000 (11:50 +0200)]
patch: Jail patch_task_patch_prefunc in classes/patch into the workdir
With PATCHTOOL=git patches the changes in the patch_task_postfunc
of the classes/patch. This works OK when the S dir is a Git repo
but doesn't if the source is a tarball.
The while condition in the patch_task_patch_prefunc must be
jailed into the WORKDIR. In the opposite, when you are executing
the recipe out of a Git subtree the function simply fails but when
your recipes are into a Git repo the patch_task_postfunc execute a
commit over your BSP local Git repo adding the changes in an
arbitrary Git repo found in the path from the SOURCE directory to
the '/'. This situation is highly probable in cases like ~home
directories under the control of a .git repo or Yocto BSP which
manage the meta layers as git submodules.
This patch fix the changes introduced in
classes/patch: when PATCHTOOL = "git" double-check the repository
commit: 86ab56b55164393924b5e688b20e8f3f3f8fc578
Author: Paul Eggleton <paul.eggleton@linux.intel.com>
Date: Tue Dec 5 14:36:58 2017 +1300
classes/patch: when PATCHTOOL = "git" double-check the repository
If a bug is present or the user has set PATCHTOOL = "git" on a source
tree that isn't git, if we try to perform git operations (such as
committing or changing branches) when extracting source, then we might
in fact be running those operations on the metadata repository if the
build directory is underneath, say, poky or OE-Core, and that could
make a mess. Check if the source tree is a git repository and refuse
to continue if it isn't.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 5 Jul 2018 12:29:29 +0000 (13:29 +0100)]
distutils: clean the build tree in do_configure
base_do_configure() tries to do "make clean" if there is a Makefile present.
For most recipes using distutils there is not a Makefile, but we do know that
"setup.py clean" will work so call that instead.
Signed-off-by: Ross Burton <ross.burton@intel.com>
util-linux.inc: add fallocate & unshare to alternatives
These binaries can be provided by busybox triggering a conflict in
do_rootfs so update-alternatives needs to know about them to properly
create the symlinks.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
with bind 9.11.2+ when the build host has lmdb installed, bind configure looks into
host headers and wrongly interprets that it should be enabling lmdb
disable lmdb to fix
| configure: error: found lmdb include but not library.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It was breaking quite common use case that the dtb files are in
some subdirectory and then kernel build fails to build them.
As reported by khem:
http://lists.openembedded.org/pipermail/openembedded-core/2018-July/152578.html
me:
http://lists.openembedded.org/pipermail/openembedded-core/2018-July/152579.html
on raspberrypi3 build:
make[3]: *** No rule to make target 'arch/arm/boot/dts/dwc2.dtbo'. Stop.
arch/arm/Makefile:345: recipe for target 'dwc2.dtbo' failed
make[2]: *** [dwc2.dtbo] Error 2
Makefile:146: recipe for target 'sub-make' failed
and trevor on the IRC:
20:35:49 < tlwoerner> the recent 2e7f3b2b9318d1e5395ad58131eafb873f614326 commit in oe-core seems to cause dragonboard-410c's kernel to fail to build
20:36:26 < tlwoerner> for the dragonboard-410c, KERNEL_DEVICETREE is set to "qcom/apq8016-sbc.dtb" but the build failure is:
20:36:37 < tlwoerner> *** No rule to make target 'arch/arm64/boot/dts/dts/qcom/apq8016-sbc.dtb'. Stop.
20:36:44 < tlwoerner> i.e. the "qcom/" is getting removed
20:37:08 < tlwoerner> oops!!
20:37:33 < tlwoerner> wrong copy&paste, the actual error is:
20:37:36 < tlwoerner> *** No rule to make target 'arch/arm64/boot/dts/apq8016-sbc.dtb'. Stop.
20:37:53 < tlwoerner> i.e., the "qcom/" is being stripped out
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
oeqa/runtime/cases/parselog.py: ignore a message from weston
The following error message when starting core-image-weston is not
critical as long as the image could start up correctly. So extend
the common_errors list for parselog.py test case to ignore this
message.
logind: cannot setup systemd-logind helper (-61), using legacy fallback
[YOCTO #12835]
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
The htmlParseTryOrFinish function in HTMLparser.c in libxml2 2.9.4
allows attackers to cause a denial of service (buffer over-read) or
information disclosure.
https://bugzilla.gnome.org/show_bug.cgi?id=775200
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Tue, 26 Jun 2018 07:59:00 +0000 (15:59 +0800)]
nfs-utils: 2.1.1 -> 2.3.1
* Removed 001-configure-Allow-to-explicitly-disable-nfsidmap.patch,
the nfsidmap is enabled when --enable-nfsv4, so I added a
PACKAGECONFIG[nfsv4], and default is no since keyutils is not in oe-core by
default.
* Removed 0001-include-stdint.h-for-UINT16_MAX-definition.patch and
nfs-utils-1.2.3-sm-notify-res_init.patch since they are already in the
source.
* Taken two patches from git://git.alpinelinux.org/aports to fix build
with musl, and the nfs-utils-musl-res_querydomain.patch is for musl only.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Wed, 4 Jul 2018 09:15:12 +0000 (10:15 +0100)]
elfutils: clean up patches
In the upgrade a large number of Upstream-Status tags were dropped, so add them
back. I'm taking the stand that copying a patch Debian is carrying doesn't
count as a backport.
Remove two Debian-specific patches (one for Hurd, one for kfreebsd) so
we're not carrying useless patches.
Remove two patches that are no longer applied.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Otavio Salvador [Tue, 3 Jul 2018 12:34:34 +0000 (09:34 -0300)]
go: Update 1.10.2 -> 1.10.3
go1.10.3 (released 2018/06/05) includes fixes to the go command, and
the crypto/tls, crypto/x509, and strings packages. In particular, it
adds minimal support to the go command for the vgo transition.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Tue, 3 Jul 2018 12:34:33 +0000 (09:34 -0300)]
go: Update 1.9.6 -> 1.9.7
go1.9.7 (released 2018/06/05) includes fixes to the go command, and
the crypto/x509, and strings packages. In particular, it adds minimal
support to the go command for the vgo transition.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Olof Johansson [Tue, 3 Jul 2018 12:00:08 +0000 (14:00 +0200)]
rpm: Avoid leaking temporary scriplet files
RPM writes each package scriptlet (post-/preinstall) to
/var/tmp/rpm-tmp.XXXXXX --- a lot of files potentially gets created.
When debugging is enabled, these temporary scriptlet files aren't
cleaned up at all and after a while this results in the filesystem
resources are eaten up (like running out of available inodes).
Normally, the temporary files would have been written to the tmp
directory of the target sysroot (which we can easily clean up), but in
this tree, you can't necessarily run the scriptlets.
Fixes [YOCTO #12792]
Signed-off-by: Olof Johansson <olofjn@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Tue, 3 Jul 2018 04:11:57 +0000 (12:11 +0800)]
ccache: 3.3.5 -> 3.4.2
* The license file is changed to LICENSE.adoc, it is still GPLv3+.
* Removed Revert-Create-man-page-in-the-make-install-from-git-.patch, it was
used for fixing a build failure of no asciidoc, but now there is no such a
failure, so remove it.
* Refreshed 0002-dev.mk.in-fix-file-name-too-long.patch
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Tue, 3 Jul 2018 04:11:54 +0000 (12:11 +0800)]
e2fsprogs: 1.43.8 -> 1.44.2
- Rebased Revert-mke2fs-enable-the-metadata_csum-and-64bit-fea.patch
- Removed backport patch 0001-libext2fs-fix-build-failure-in-swapfs.c-on-big-endia.patch
- The LIC_FILES_CHKSUM changed because it updated the address, others are the same
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 3 Jul 2018 13:04:11 +0000 (14:04 +0100)]
meson: validate cpu_family
Meson has a defined list of known CPU families but these are not currently
validated, so mistakes in cross files or new architectures are not noticed.
Backport a patch from upstream which warns on unknown architectures, but tweak
it to fatally error instead. When we upgrade to Meson 0.47 the first half of
this patch can be dropped.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 3 Jul 2018 13:04:10 +0000 (14:04 +0100)]
meson: map architecture to correct values in cross file
The cross file specifies the host/target cpu_family, which should be one of a
defined set of values[1] but if it isn't Meson won't complain and instead
recipes may behave unexpectedly.
removed:
dont-test-on-host.patch, no longer implemented
drop use-python3-and-fix-install-lib-path.patch, they added the ability to pass in lib dir loctions
drop bind-confgen-build-unix.o-once.patch, fix included in update
Refresh other patches:
add python3 flag for PACKAGECONFIG to pull in python
add new config option --with-eddsa=no (needs openssl support not released)
Python support is disaled by default now.
Acked-by: Martin Hundebøll <mnhu@prevas.dk> Signed-off-by: Armin Kuster <akuster808@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
include several CVE fixes.
CVE: CVE-2018-5733
CVE: CVE-2018-5732
LIC_CHKSUM_FILE updated to SPFX format
https://kb.isc.org/article/AA-01571
remove several patches now included in update.
Shared libarary support is now enabled in configure+lt, use it
and revert to autotools-brokensep
Refresh patches
Aligns support with bind 9.11.x
Add libxml2 support to configure.ac+lt
Signed-off-by: Armin Kuster <akuster808@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paulo Neves [Mon, 2 Jul 2018 15:03:44 +0000 (17:03 +0200)]
kernel-devicetree: Corrected normalize_dtb
The normalize_dtb function was buggy because
it only converted from .dts suffix to .dtb
suffix if the user passed a full source path to
KERNEL_DEVICETREE containing the /dts/ path.
The problem is that if the user did that there
would be a warning.
On the othet hand if user just set the variable
KERNEL_DEVICETREE="file.dts" the bbclass translation
to the respective .dtb target did not occur and
make would fail saying it has no rule to make target
file.dts
This patch decouples the logic of having /dts/ in the
path from the target translation.
Signed-off-by: Paulo Neves <ptsneves@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kai Kang [Mon, 2 Jul 2018 01:15:34 +0000 (09:15 +0800)]
webkitgtk: 2.20.2 -> 2.20.3
Upgrade webkitgtk from 2.20.2 to 2.20.3.
* update context of 0001-WebKitMacros-Append-to-I-and-not-to-isystem.patch
* remove detect-atomics-during-configure.patch that webkitgtk 2.20.3
contains the commit of better solution, see
https://bugs.webkit.org/show_bug.cgi?id=161900#c9
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alex Kiernan [Wed, 20 Jun 2018 04:23:19 +0000 (04:23 +0000)]
kernel-fitimage: Make DTB key insertion optional
If UBOOT_DTB_BINARY is empty, then don't try inserting the U-Boot
signing keys into the DTB. In this configuration the keys are expected
to be already present in U-Boot's DTB.
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
This is a function much like shutil.which or bb.utils.which, retaining
shutil.which-like function semantics, bb.utils.which's support for
returning available candidates for signatures, and most importantly,
supports wildcards, returning only the first occurrance of each found
pathname in the search path.
Signed-off-by: Christopher Larson <chris_larson@mentor.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Results in gcc-cross-canadian-mips failing to build due to the use
of an incorrect sysroot, fix this. All nativesdk pieces should be in
the same sysroot (unprefixed).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 2 Jul 2018 12:43:50 +0000 (12:43 +0000)]
staging: Improve fixup processing code
With the fixes to other parts of multilib, it was found that the fixup code's
assumptions about the recipe sysroot were incorrect. We need to use the value
calculated earlier in the function.
It turns out there is a rather neat way to do this which cleans up the code
as an added bonus.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 2 Jul 2018 09:05:17 +0000 (09:05 +0000)]
lib/oe/sstatesig: Fix task mappings from multilib<->non-multilib contexts
If we're in a multilib context already and want a non-multilib context
this function returned incorrect values.
Try and retain optimisations for the common case not needing to request
a datastore but allow the different multilib/non-multilib combinations
to work too.
This fixes bugs where rootfs generation of a multilib image would
write into incorrect locations, or be unable to find sstate manifest
files due to incorrect data stores being used to expand data.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 2 Jul 2018 08:59:34 +0000 (08:59 +0000)]
classes/utils: Improve all_multilib_tune_values
Currently there is duplication in the code, we can clean this up
by extending the multilib variants list.
This code also currently fails its its called from an existing multilib
context since its assumes the data store passed in is the non-multilib
case. When building an image, say lib32-core-image-sato, this leads to
incorrect PATH values.
To fix this, we also request a data store for the "" variant allowing
this corner case to be fixed and for the function to correctly return
values for each multilib even when a multilib is already selected.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 29 Jun 2018 16:33:26 +0000 (16:33 +0000)]
staging/image: Fix multilib recipe sysroot issues
Currently if you enable multilib, then build an image, the multilib
recipe sysroot is build in the wrong WORKDIR. If you then clean and
rebuild the image you see "file exists" errors.
This patch ensures the real WORKDIR is used consistently and then
cleans/rebuilds also work correctly.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 2 Jul 2018 09:03:12 +0000 (09:03 +0000)]
lib/oe/utils: Improve get_multilib_datastore
Currently this function assumes that no multilib is applied and that
we're applying a multilib. This means if we're in multilib context
and want the non-multilib context we can't obtain it (and no other
function exists for this either).
Improve the function to allow this to be requested.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joe Slater [Thu, 21 Jun 2018 18:22:17 +0000 (11:22 -0700)]
postinst-intercepts: do not execute any variant of delay_to_first_boot
As of commit 2c5c6e3ff we create multilib variants of intercept
hooks but we did not account for delay_to_first_boot variants.
This was covered up until commit a335e7867, but will now cause
an error.
Signed-off-by: Joe Slater <joe.slater@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alp Özmert [Fri, 29 Jun 2018 15:41:02 +0000 (17:41 +0200)]
license: Fix and extend recommendations for license packages.
Changed package recommendations so that the license package of a
recipe is recommended for all packages of a recipe instead of for one
package given by the recipe name.
Pre-patch behaviour results in a missing recommendation when a recipe
does not have a package with the same name.
Signed-off-by: Alp Özmert <info@ib-oezmert.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alistair Francis [Fri, 25 May 2018 15:56:36 +0000 (08:56 -0700)]
recipes-kernel/linux: Enable NUMA Kconfig from MACHINE_FEATURES
If the user has set numa in their MACHINE_FEATURES we should enable
NUMA support in the kernel config.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com> Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hongzhi.Song [Fri, 29 Jun 2018 06:16:20 +0000 (02:16 -0400)]
oeqa/runtime: Add testcases for kernel sample
We are going to let runtime test support kernel tests. Now we just add
kernel self-contained sample tests. And we plan to add overall kernel
tests in the future.
This patch is just add kernel samples test which contains about 13 tests
enabled by kernel-sample.scc. So it needs statement,
KERNEL_FEATURES_append += " features/kernel-sample/kernel-sample.scc" in
local.conf.
Signed-off-by: Hongzhi.Song <hongzhi.song@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 29 Jun 2018 08:02:16 +0000 (08:02 +0000)]
xtrans: Fix multilib .pc file conflict
Error: Transaction check error:
file /usr/share/pkgconfig/xtrans.pc from install of lib32-xtrans-dev-1:1.3.5-r0.core2_32 conflicts with file from package xtrans-dev-1:1.3.5-r0.core2_64
[YOCTO #12511]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Arsalan H. Awan [Thu, 28 Jun 2018 14:16:36 +0000 (19:16 +0500)]
initrdscripts/init-live.sh: fix mounts w/ spaces fail to move to real rootfs
When there are spaces in the mount points of devices e.g.:
a partition mounted at "/run/media/My Root Partition-sda1",
the initrd fails to move such mount points over to the
corresponding directories at /media under the real root filesystem,
and the mount points would appear at the same location as they were
mounted on when detected by initrd, for example:
here: "/run/media/My Root Partition-sda1"
instead of here: "/media/My Root Partition-sda1"
This causes issues such as:
* The disks/partitions cannot be formated with any filesystem
using e.g. mkfs.ext4 or mke2fs in general. When tried to do so
by making sure the device is not mounted, it failed with
errors such as:
> /dev/sda1 is apparently in use by the system; will not make a
filesystem here!
> /dev/sda1: Device or resource busy while setting up superblock
* The read/write operations become extremely slow. e.g. Under testing,
it took approx. 2 hours just to copy 700 MB of data to the partition,
and it took more than 40 minutes to delete that data from it.
Same operations took under 5 minutes on a partition that had no
spaces in its mount point (or that was successfully moved to real
root by initrd and appeared under /media instead of /run/media).
This commit fixes such issues by quoting the arguments of failing mount
move commands and by parsing OCT or HEX encoded special characters
such as spaces to ASCII charecters in the mount points as kernel
populates the procfs like so.
Signed-off-by: Arsalan H. Awan <Arsalan_Awan@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Thu, 28 Jun 2018 17:53:41 +0000 (12:53 -0500)]
bitbake-bblayers/create: Fix layer name generation
The path to where the layer was being created was taken verbatim as the
name of the layer when generating the layer.conf and README files from
templates. This causes problems in the layer.conf file because it would
result in strangely named variables like
BBFILE_PATTERN_../my-layer = "..."
Instead of blindly taking the path, use the name of the last component
of the path as the layer name.
Additionally, rework the template files to use python format strings
with named parameters so that the same argument doesn't have to be
repeated multiple times.
[YOCTO #12808]
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 28 Jun 2018 17:47:45 +0000 (18:47 +0100)]
libsdl2: fix build race
There's an occasional build race from headers being generated in parallel with
other files which include the headers being compiled. Solve this by adding more
dependencies.
[ YOCTO #12815 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
wic: isoimage-isohybrid: debloat image of redundant rootfs
There's no reason to have that rootfs.img filesystem in the image:
it's not used for anything because both the EFI and legacy boot paths
use the /initrd which contains the same contents as the rootfs.img,
only compressed. It was probably forgotten in there :)
My iso went down from 224 to 94 mb.
Tested using UEFI/legacy boots on CD-roms, usb dongle and qemu VM's.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It's really good that OE supports multiple EFI_PROVIDERs and that
commit 9a1709278de87 ("wic: isoimage-isohybrid: use grub-efi from
deploy dir") makes re-use of the grub-efi built image, but we should
still respect the standard otherwise the ISO will not boot, so install
grub images as boot[x64|ia32].efi not ${PN}-boot[x64|ia32].efi.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Daniel Díaz [Thu, 28 Jun 2018 13:56:28 +0000 (08:56 -0500)]
kmscube: Update to master from 2018-06-17
Changes from 0d8de4ce: c2d4ba8 drm-legacy: fix poll for flip event, actually exit on user input 56c3917 formats: use weston's egl config matching logic, centralize format aac3788 Rework default modifier handling 4f7cec0 Use weak functions to handle lack of gbm modifiers 98f31bf cube-tex: make use of modifiers 063ce5c gbm: fix fallback for drivers that don't support modifiers 9dcce71 add MSAA
Also refresh gbm_bo_map/_unmap patch.
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 28 Jun 2018 11:25:12 +0000 (12:25 +0100)]
bitbake.conf: Allow BBINCLUDED to be unset
For some reason the layer index is expanding HOSTTOOLS before BBINCLUDED is
set so recent changes break it. This adds in a simple workaround to stop it b
reaking allowing the index to function correctly again.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andre McCurdy [Tue, 26 Jun 2018 19:48:11 +0000 (12:48 -0700)]
bitbake.conf: handle cmake -dev files packaging with default rules
Move packaging rules for cmake -dev files from cmake.bbclass into
bitbake.conf to handle recipes (e.g. harfbuzz 1.8.1) which build with
autotools but also install cmake -dev files.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Samuli Piippo [Tue, 26 Jun 2018 07:59:51 +0000 (10:59 +0300)]
glib: remove unnecessary dependency to DISTRO_FEATURES
Since DISTRO_FEATURES was expanded in the comments, it created
task dependency to the full content of DISTRO_FEATURES, instead
of just the x11 used below. This prevented reuse of sstate-cache
when unrelated feature flags were changed.
Signed-off-by: Samuli Piippo <samuli.piippo@qt.io> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andre McCurdy [Fri, 18 May 2018 22:50:40 +0000 (15:50 -0700)]
arch-armv7a.inc: default to Thumb2 instruction set for armv7a and above
Although there may still be specific cases which can benefit from the
ARM instruction set, the Thumb2 instruction set is generally a better
default for armv7a class CPUs. Distros such as Debian and Fedora have
been targeting Thumb2 by default for some time.
Note that setting ARM_INSTRUCTION_SET has no effect unless
TUNE_FEATURES contains "thumb" (which is controlled by the "t" suffix
in DEFAULTTUNE, e.g. armv7vehf-neon -vs- armv7vethf-neon, etc) so out
of tree machine configs may need to update their DEFAULTTUNE to take
advantage of this change.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Olof Johansson [Mon, 25 Jun 2018 11:34:46 +0000 (13:34 +0200)]
insane.bbclass: Don't let warnings make previous errors non-fatal
package_qa_handle_error() returns True on non-fatal issues and False on
fatal issues. But the current usage has been to do
sane = package_qa_handle_error(...)
which would always reset sanity status to be that of the last issue
identified. This change the assignments to use the &= operator instead:
sane &= package_qa_handle_error(...)
As far as I can tell, this is not a real problem in practice, because
warnings of different levels (WARN_QA, ERROR_QA) does not seem to have
been mixed in a way that triggered this issue.
Signed-off-by: Olof Johansson <olofjn@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
