linux-yocto: Enable kernel-sample features for runtime tests

Enable kernel-sample features by default with the machine of qemu.

Signed-off-by: Hongzhi.Song <hongzhi.song@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

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>

python3: fix file conflict when multilib enabled

Config file python3.5m-config conflicts between 32 and 64 bit packages.
Use update-alternatives to add base_libdir as suffix to avoid it.

[YOCTO #12511]

Signed-off-by: Zhang Xiao <xiao.zhang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

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>

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>

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>

libsoup: fix CVE-2018-12910

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssl: disable ccache usage

ccache and openssl don't get on:

| make[1]: Entering directory '/home/prj/yocto/build/tmp/work/core2-64-poky-linux/openssl/1.0.2o-r0/openssl-1.0.2o/crypto'
| ccache: invalid option -- 'D'

Disable the use of ccache in the openssl recipe until someone root-causes this.

[ YOCTO #12810 ]

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>

wic: isoimage-isohybrid: fix UEFI spec breakage

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>

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>

mesa: Fix parallel make race

Builds keep failing with a race over the generated header file, fix it!

[YOCTO #12828]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnu-efi: update to 3.0.8

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

debianutils: update to 4.8.6

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libbsd: update to 0.9.1

License-Update: update COPYING file format

Drop 0001-Replace-__BEGIN_DECLS-and-__END_DECLS.patch and
0002-Remove-funopen.patch since they are already fixed upstream.

Backport 0001-flopen-Add-missing-fcntl.h-include.patch to fix build with
musl.

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

file: update to 5.33

Drop 0001-Add-P-prompt-into-Usage-info.patch since it is already fixed
upstream.

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

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>

maintainers.inc: recipes cleanup

Cleanup the recipes which had been removed from oe-core.

Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

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>

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>

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>

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>

insane.bbclass: Make missing license file fatal

If a license file referenced from LIC_FILES_CHKSUM doesn't exist,
insane.bbclass would output an error message, but would continue the
build. This change makes this error fatal (as I suspect has been the
intention).

Signed-off-by: Olof Johansson <olofjn@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glibc: fix CVE-2018-11237

glibc: fix CVE-2018-11237

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-8: define GLIBC_DYNAMIC_LINKER relative to SYSTEMLIBS_DIR on riscv

Refresh 0014-Define-GLIBC_DYNAMIC_LINKER-and-UCLIBC_DYNAMIC_LINKE.patch
to also define GLIBC_DYNAMIC_LINKER relative to SYSTEMLIBS_DIR on riscv.

Signed-off-by: Ricardo Salveti <ricardo@opensourcefoundries.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

maintainers: assign systemtap to Victor Kanensky

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Set SOURCE_DATE_EPOCH to kernel git timestamp if not set

If SOURCE_DATE_EPOCH is unset (in addition to the existing "0" behaviour)
parse out the top most commit timestamp from the kernel tree to use as the
timestamp.

Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image: Set COREBASE as the git directory for timestamp

When REPRODUCIBLE_TIMESTAMP_ROOTFS is unset and we want to parse one
from git, use COREBASE as the base for the git command so we have a
known repository which we're using. Without this the build may fail
if the current directory is not part of a git repository.

Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

db: drop legacy ARM specific SWP based mutex optimisation

Although the ARM SWP instruction may exist for ARMv6 and above, it's
not guaranteed to work, especially on SMP systems where it's use may
lead to instability at runtime, etc:

  https://community.arm.com/processors/b/blog/posts/locks-swps-and-two-smoking-barriers

Keeping the optimisation for architecture levels which pre-date SMP
(ie <= ARMv5) may be safe, however other distros (Buildroot, Debian,
Fedora, etc) are not doing so and mutex contention is likely to be
less of an issue on uniprocessor systems anyway, so the benefits of
this micro optimisations are not clear. Since OE uses ARMv5 qemu as
a proxy for testing all 32bit ARM architecture levels, it's desirable
to keep the ARMv5 builds aligned with later ARM architecture levels
wherever possible.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mdadm: fix one more issue when building with gcc8

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: escape paths passed to shell

Systemd mount configuration file must have a name that match the mount
point directory they control. So for instance, if a mount file contains

    [Mount]
    ...
    Where=/mnt/my-data

The file must be named `mnt-my\x2ddata.mount`, or systemd will refuse to
honour it.

If this config file contains an [Install] section, it will silently fail
because the unit file is not escaped properly when systemctl is called.
To fix that, make sure paths are escaped through `shlex.quote`.

Signed-off-by: Damien Riegel <damien.riegel@savoirfairelinux.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

util-linux: Enable unshare utility

It allows to run program with some namespaces unshared from parent.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nss: update to 3.37.1

remove Fix-compilation-for-X32.patch as a solution simular is included in update.

notable changes:
The TLS 1.3 implementation was updated to Draft 28.
The CA certificates list was updated to version 2.24.

refresh patches
fix 32 bit build error nss bug: https://bugzilla.mozilla.org/show_bug.cgi?format=default&id=1459739

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

findutils: Refresh patches with devtool

We get fuzz warnings when applying these patches and devtool
reports it

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

update-rc.d: Add nativesdk

Add nativesdk for update-rc.d.

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

createrepo-c: Add nativesdk

Add nativesdk for createrepo-c.

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libnewt-python: Add nativesdk

Add nativesdk for libnewt-python.

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libnewt: Add nativesdk

Add nativesdk for libnewt.

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

slang: add nativesdk

Add nativesdk for slang.

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rng-tools: start earlier in the boot process

Entropy is needed earlier in the boot process in some cases, for example
connman can require it, and rgd doesn't require much, so move it earlier
in the boot process, 03 for sysvinit, and before sysinit for systemd.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rng-tools: drop unnecessary INHIBIT_UPDATERCD_BBCLASS

This was added as the recipe didn't inherit systemd, but it does inherit
systemd now.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dmidecode: correct docdir

Without this, the package clutters up the root of /usr/share/doc.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ovmf: set PARALLEL_MAKE for target as well

This can fail for target, not just native.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

console-tools: add missing flex/bison deps

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pcmciautils: depend on bison-native

pcmciautils needs yacc as well as lex.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oe.scriptutils.run_editor: ditch the error-prone argument quoting

Rather than trying to construct a string by quoting the files in an
error-prone way, parse $EDITOR to pass a list to subprocess rather than
a string.

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa: add Git mirror test

Add an automated test to exercise that fetching a git: recipe...
1) works over git without mirrors
2) fails without git connectivity or mirrors
3) works without git connectivity but with a mirror

(2) is done by setting GIT_PROXY_COMMAND to 'false' which should break any git
network operations.

[ YOCTO #12805 ]

Signed-off-by: Ross Burton <ross.burton@intel.com>

cryptodev: Fix build errors with v4.17+

Backport from upstream to update internal syscall function usage.
https://github.com/cryptodev-linux/cryptodev-linux
f60aa08c63fc02780554a0a12180a478ca27d49f

Signed-off-by: He Zhe <zhe.he@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

piglit: upgrade to latest revision

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

avahi: fix error at boot time for avahi-daemon.service

The following error messages appear now and then at boot time.

  avahi-daemon/chroot.c: open() failed: No such file or directory
  Failed to open /etc/resolv.conf: Invalid argument

The problem is about /etc/resolv.conf. In Yocto's systemd based
systems, it's a symlink to /etc/resolv-conf.systemd which in turn
is a symlink to /run/systemd/resolve/resolv.conf. The systemd-resolved
service handles creation of /run/systemd/resolve/resolv.conf file.

So if avahi-daemon is started before systemd-resolved, the error messages
appear.

Fix this problem by making avahi-daemon start after systemd-resolved.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

avahi: reorganize codes

This patch does nothing but reorganizing codes.

avahi.inc is shared by avahi and avahi-ui recipes. Move common things
into it, and move uncommon things out of it.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

libepoxy: upgrade 1.5.0 -> 1.5.2

Remove upstreamed patch:
        no-tests.patch

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

ofono: upgrade 1.23 -> 1.24

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

iproute2: upgrade to release 4.17

Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>

nss: Fix build error for aarch64be.

Signed-off-by: Lei Maohui <leimaohui@cn.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

mtd-utils: upgrade 2.0.1 -> 2.0.2

Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

systemd: remove the group 'lock'

The upstream commit 61f32bff6130a44d077886d38cff89ad161bf177 included in
the release v229 removed the use of the group:

    commit 61f32bff6130a44d077886d38cff89ad161bf177
    Author: Martin Pitt <martin.pitt@ubuntu.com>
    Date:   Mon Feb 1 12:09:34 2016 +0100

        tmpfiles: drop /run/lock/lockdev

        Hardly any software uses that any more, and better locking mechanisms like
        flock() have been available for many years.

        Also drop the corresponding "lock" group from sysusers.d/basic.conf.in, as
        nothing else is using this.

    [...]
    diff --git a/sysusers.d/basic.conf.in b/sysusers.d/basic.conf.in
    index 823d6cb20..b2dc5ebd4 100644
    --- a/sysusers.d/basic.conf.in
    +++ b/sysusers.d/basic.conf.in
    @@ -19,7 +19,6 @@ g wheel   -     -            -

     # Access to certain kernel and userspace facilities
     g kmem    -     -            -
    -g lock    -     -            -
     g tty     @TTY_GID@     -            -
     g utmp    -     -            -
    [...]

The upstream documentation doc/UIDS-GIDS.md says that basic.conf.in is "the
precise list of the currently defined groups":

    ## Special `systemd` GIDs

    `systemd` defines no special UIDs beyond what Linux already defines (see
    above). However, it does define some special group/GID assignments, which are
    primarily used for `systemd-udevd`'s device management. The precise list of the
    currently defined groups is found in this `sysusers.d` snippet:
    [basic.conf](https://raw.githubusercontent.com/systemd/systemd/master/sysusers.d/basic.conf.in)

    It's strongly recommended that downstream distributions include these groups in
    their default group databases.

Removing the creation of the group also avoids the need to define a GID
for it when using static ids.

Signed-off-by: Hannu Lounento <hannu.lounento@vaisala.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

packagegroup-full-cmdline: remove popt

popt is a library, so it makes no sense for this to be in a "command line"
packagegroup.

Signed-off-by: Ross Burton <ross.burton@intel.com>

lsbinitscripts: remove spurious build dependencies

A recipe that doesn't configure, doesn't compile, and just ships a single shell
script certainly does not need to build depend on popt and glib-2.0.

Signed-off-by: Ross Burton <ross.burton@intel.com>

minicom: change SRC_URI to use Debian mirrors

Alioth is dead and the minicom project on Salsa (the replacement for Alioth)
doesn't actually have any files in, so just use the Debian mirror instead.

Signed-off-by: Ross Burton <ross.burton@intel.com>

mirrors: add Debian snapshot mirror for 2018

Add a Debian snapshot mirror from 2018 (specifically, 10th March, the date 9.4
was released) to DEBIAN_MIRRORS.

Signed-off-by: Ross Burton <ross.burton@intel.com>

shadow: update SRC_URI now Alioth is down

Alioth is dead so we can't use it for SRC_URI anymore.  There is a shadow
repository on GitHub which is the new upstream, but for some reason it is
missing the 4.2.1 tag and tarball that we use.  Also 4.2.1 was never uploaded
into Debian itself, so we can't use their mirror network either.

For now point SRC_URI at the Yocto Project source mirror and set
UPSTREAM_CHECK_URI so that we get nagged to upgrade to 4.6.

Signed-off-by: Ross Burton <ross.burton@intel.com>

chrpath: Alioth is dead, use DEBIAN_MIRROR

The previous host of chrpath, Alioth, is dead.  chrpath hasn't yet moved to
Salsa, so download the tarball from the Debian mirrors.

Signed-off-by: Ross Burton <ross.burton@intel.com>

iso-codes: update to 3.79 and fix SRC_URI

Updated package to version 3.79.

The original FTP server, alioth.debian.org has been permanently taken
down by the Debian Project and replaced with salsa.debian.org.

Source: https://wiki.debian.org/Alioth

Also added HOMEPAGE, and BUGTRACKER.

[ tweak URL to work with checkpkg - RB ]

Signed-off-by: Oleksandr Kravchuk <dev@sashko.rv.ua>
Signed-off-by: Ross Burton <ross.burton@intel.com>

atk: Correct multilib conflict in header.

Signed-off-by: Jeremy Puhlman <jpuhlman@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

qemu: CVE-2018-11806 slirp heap buffer overflow

Signed-off-by: Jeremy Puhlman <jpuhlman@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

trace-cmd: remove

Perf is superior in most ways and is preferred.

Signed-off-by: Ross Burton <ross.burton@intel.com>

libunistring: upgrade to 0.9.10

License-Update: Checksum changed due to updation in documentation. There
are no changes in the license terms.

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

maintainers: Assign ownership of newlib and libgloss to Alejandro Hernandez

Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandr@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-8: Enable fp arch extention when fpu is available

This is especially needed when defaulting to hard-float ABI
Fixes errors e.g.
cc1: error: -mfloat-abi=hard: selected processor lacks an FPU

Fixes [YOCTO #12795]

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oe/types: Allow boolean to accept an existing boolean

Exception: TypeError: boolean accepts a string, not '<class 'bool'>

is a bit annoying if you pass in True/False. Tweak the function
to make it forgive that situation.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

maintainers: update Intel owners

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: add lost Upstream-Status

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf/testimage: Add in compatiblity hacks

The testimage changes break the autobuilder. The plan is to revert these hacks
but due to various changes happening with the autobuilder, we need these for now
to keep things working until we can move to the new autobuilder codebase.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

subversion: Update HOMEPAGE

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mktemp: remove

Both busybox and coreutils provide mktemp, and the only difference between those
(and standalone mktemp) is that coreutils supports --suffix.

Also mktemp.org has disappeared, so it's fair to assume that the standalone
mktemp (last released in 2010) is dead.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: Update 18.1.1 -> 18.1.2

This includes bug fixes for 18.1.x series. For list of changes check:

https://www.mesa3d.org/relnotes/18.1.2.html

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-firmware: upgrade to d114732 revision

Following changes are applied:

d114732 Merge branch 'for-upstreaming-v1.7.2-vsw' of https://github.com/felix-cavium/linux-firmware
fe4a9d4 brcm: update firmware for bcm4356 pcie
a292181 brcm: update firmware for bcm4354 sdio
be1c535 brcm: update firmware for bcm43362 sdio
5ac5ad5 brcm: update firmware for bcm43340 sdio
ec734a3 brcm: update firmware for bcm43430 sdio
ba51e86 Update Cypress license termination clause
1fa9ce3 amdgpu: update vega10 VCE firmware to version 55.3
9489f55 linux-firmware: Update firmware patch for Intel Bluetooth 7265 (D0)
6be4747 linux-firmware:Update firmware patch for Intel Bluetooth 7265 (D1)
4949d88 Merge branch 'master' of https://github.com/vgarodia/linux-firmware
7602644 qcom: add venus firmware files for v5.2
7518922 Update AMD cpu microcode for family 15h
3f8243c linux-firmware: Update firmware patch for Intel Bluetooth 8260
2c4f7c1 qed: Add firmware 8.37.2.0
0e193ca linux-firmware: liquidio: update vswitch firmware to v1.7.2
2a9b2cf Merge branch 'for-upstreaming-v1.7.2' of https://github.com/felix-cavium/linux-firmware
112d300 linux-firmware: Update firmware file for Intel Bluetooth,8265
629a5e0 linux-firmware: Update firmware patch for Intel Bluetooth 7260 (B5/B6)
7035335 linux-firmware: Update firmware patch for Intel Bluetooth 7260 (B3/B4)
c84bd66 linux-firmware: Update firmware file for Intel Bluetooth,9260
897330f linux-firmware:Update firmware patch for Intel Bluetooth 7265 (D1)
e94352c linux-firmware: Update firmware file for Intel Bluetooth,9560
7710151 linux-firmware: Update AMD cpu microcode
be1e8c5 Merge commit 'a6327c5a5f74c3d11da76d0dcb51c6f5db7f8a35' of https://github.com/Netronome/linux-firmware
86d39f8 amdgpu: sync up polaris12 firmware with 18.10 release
b5320a9 amdgpu: sync up polaris11 firmware with 18.10 release
152c924 amdgpu: sync up polaris10 firmware with 18.10 release
0d672f7 amdgpu: sync up vega10 firmware with 18.10 release
30447b1 amdgpu: sync up carrizo firmware with 18.10 release
dd6f936 amdgpu: sync up topaz firmware with 18.10 release
de9c175 amdgpu: sync up stoney firmware with 18.10 release
5e22609 amdgpu: sync up tonga firmware with 18.10 release
499042e amdgpu: sync up fiji firmware with 18.10 release
1db3eec amdgpu: sync up raven firmware with 18.10 release
a6327c5 nfp: Add symlink for Agilio CX 1x40GbE flower firmware
8b78780 nfp: update Agilio SmartNIC flower firmware to rev AOTC-2.9.A.16
d3b6941 linux-firmware: liquidio: update firmware to v1.7.2

License-Update: copyright years updated and filelist updated.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

findutils: make -delete honour the -ignore_readdir_race option

the '-delete' action no longer complains about disappeared files
when the '-ignore_readdir_race' option is given, too. That action
will also returns true in such a case now.

Signed-off-by: Zhixiong Chi <zhixiong.chi@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xkbcomp: update to 1.4.2

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-vmware: update to 13.3.0

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-input-evdev: update 2.10.6

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xkeyboard-config: update to 2.24

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

subversion: 1.9.7 -> 1.10.0

License changed since licenses for the bundled linenoise and lz4 codebases
were added. We don't build either of them. Add MIT since utf8proc is MIT
licensed.

Configure to use the internal utf8proc codebase since we have no copy of that in
OE-Core, nor any need to add one.

Add a dependency on lz4 which is now required rather than using the internal
codebase within subversion.

Drop a patch merged upstream.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-modules: upgrade 2.10.5 -> 2.10.6

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image/testimage: Rework auto image test execution

The TEST_IMAGE interface has never particularly worked and image testing currently
gets configured manually.

This reworks the interface to better serve the needs of its users, replacing it
with TESTIMAGE_AUTO.

This does away with the need for the separate class due to better bitbake APIs for
changing tasks.

TESTIMAGE_AUTO will automatically boot under qemu any image that is built. It
also adds in dependencies so that any SDK for which testing is requested will
automatically be built first.

The code in bitbake.conf was error prone (e.g. testsdk wasn't considered), this
improves it to standardise on IMAGE_CLASSES as the standard configuration mechanism.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image/image-live: Improve handling of live/iso/hddimg types (drop NOISO/NOHDD)

The logic can be improved and the historical NOISO/NOHDD variables
moved into the class and out of common code.

The variables are also then removed in favour of directly controlling
the behaviour from IMAGE_FSTYPES in line with all the other image types.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image: Combine all the class handling code into a single section

Right now the code handling class inherits is spread all over and
its hard to get an idea of what is happening overall. Combine all
the code together to make it clearer. There shoould be no functionality
changes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rm_work: Simplify looping code

The current looping structure is confusing, simplify it a bit
to improve readability. Should be no functionality changes.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rm_work: Improve code comments

This function is a little obtuse, add more comments about what its
doing and why. Also combine some of the statements where possible
to improve clarity.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rm_work: The do_setscene task has not existed for years

Once, there were do_setscene tasks but this hasn't been the case for years,
drop the old code.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rm_work: Stop appending _setscene to do_image_complete_setscene stamps

This is a reworked version of the issue which 5479654eeaaa0f81bfff54ca49369c87f1658705
attempted to fix.

"""
Each time I build my image after the first, I end up with a
do_image_complete_setscene stamp file with an extra _setscene appended to
the name. Eventually, the filenames end up being so long that mv complains
and the build fails.

It looks like this behaviour was introduced when the special handling was
added for do_image_complete in 2ff9d40dc88d43567472218cf3d3faf414398c71.
"""

Instead of the original approach which broke do_package_setscene, add
an entry to explictly stop the stacking _setscene pieces on do_image_complete.

It's not straightforward to just move *do_image_complete* after the
*_setscene* pattern because do_image_complete stamps would then match
do_image*.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Revert "rm_work: Stop appending _setscene to do_image_complete_setscene stamps"

Whilst this fixes the do_image_complete_setscene append problem, it creates a
new problem since the code can no longer reach the *do_package_setscene.*
code block below it. This breaks builds as per [YOCTO #12765]. Revert this
change in search of a better fix.

This reverts commit 5479654eeaaa0f81bfff54ca49369c87f1658705.

xf86-video-intel: update dependencies for the tools

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-intel: disable DRI 1, enable DRI 3

DRI 1 is only needed for UMS on i810[1], and modern X servers (1.20 onwards)
don't support DRI 1, so disable DRI 1.

However, DRI 3 isn't new and unreliable anymore, so enable support for it.

Also, now that the *proto packages are all part of xorgproto which is a required
DEPENDS, we don't need to mention it over and over.

[1] https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=359477215092ac1b602ad1e2f17a28963d9224c2

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-intel: patch cleanup

always_include_xorg_server.h.patch is merged upstream, so drop it.

disable-x11-dri3.patch was purely to solve determinism issues with the shared sysroot, which
we don't have anymore.

Also remove redundant tarball checksums as this recipe fetches from git.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-omapfb: remove

We were shipping the Debian fork of the original code, but that has disappeared
now that Alioth is down.  The driver didn't provide anything better than the
kernel and xserver's modesetting driver, so remove it.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-omap: remove

This driver is dead upstream and doesn't appear to provide any advantage over
upstream kernel drivers and xserver's modesetting support.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

musl: Upgrade to latest

Changelog:
https://git.musl-libc.org/cgit/musl/log/?qt=range&q=941bd884cc0221d051840ce6d21650339e711863..d5e55ba3320c30310ca1d8938925d5424a652422

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: Update 18.0.2 -> 18.1.1

This includes all bugfixes of 18.0.x series and major new
features. For list of changes check:

https://www.mesa3d.org/relnotes/18.1.0.html
https://www.mesa3d.org/relnotes/18.1.1.html

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>