linux-firmware: upgrade 20200519 -> 20200619

License-Update: new firmwares, all redistributable
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3ae9585efdf5bcef7224c8e752ea76064922ee99)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

kernel.bbclass: Fix Module.symvers support

Starting from v5.8-rc1 commit 269a535ca931 (modpost: generate
vmlinux.symvers and reuse it for the second modpost"), kernel will
generate new vmlinux.symvers instead of dumping all the vmlinux symbols
into Module.symvers in the first pass.

Error log:
    'run.do_shared_workdir.16614' failed with exit code 1:
    DEBUG: cp: cannot stat 'Module.symvers': No such file or directory

This change will check the file Module.symvers existence before copying it.

Signed-off-by: Lili Li <lili.li@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit cd2d62a08a1dfcd890a03ee55132b6d6c65f5ab7)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

wpa-supplicant: Security fix CVE-2020-12695

Source: http://w1.fi/security/
Disposition: Backport from http://w1.fi/security/2020-1/

Affects <= 2.9 wpa-supplicant

Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e9c696397ae1b4344b8329a13076f265980ee74d)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

scripts/install-buildtools: Handle new format checksum files

Autobuilder generated checksum files only have a single space between the sum and the
filename, tweak it to account for this.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d1e71bf5b399372166eb40bb0d99c8fb52231600)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

scripts/install-buildtools: Update to 3.2 M1 buildtools

This fixes issues with openssl certs not working properly which meant error
reporting to an error report server was failing.

Also, all our downloads are now standarised on "sha256sum" so adjust for that.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bc8b44e19a05f499f5cef049eedbed1fede2e765)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

testimage: enable ovmf support

Add support for running wic images with EFI as testimage.
Introduces a variable called QEMU_USE_OVMF for configuration.

Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3af8aaff68ed332d812ea7dc184d392700ad7882)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

runqemu: add QB_ROOTFS_EXTRA_OPT parameter

Content of the optional parameter will be appended to the rootfs-device
in the qemu configuration, in case QB_ROOTFS_OPT is not specified.
By default this is empty.

Example use cases are:

Defining 'readonly' when using squashfs, so multiple instances of qemu
can share the same base image, something that cannot be done by just
specifying 'snapshot'.

Defining 'bootindex=0' which helps to get past the EFI shell in
ovmf-binary. This also enables the use case of running WIC images with
EFI bootloader through the testimage.bbclass.

Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e9b8c194636cb5505774a2a71bf54450580dd5b8)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oeqa/runtime/cases/ptest: Make output content path absolute

The output content is created in current directory, because json content
has no defined absolute path to WORKDIR as in bitbake.

Signed-off-by: Andrej Valek <andrej.valek@siemens.com>
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 14203b2cb9aa62f55cb12230ac8012b3cd995be7)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oeqa/utils/command: Improve stdin handling in runCmd

Occasionally we've been seeing leftover threads from runCmd. The stdin test
assumes we clean up all threads but the code assumes that the daemonic thread
can be left behind.

The issue can be reproduced by adding a time.sleep(10) to the end of
writeThread() which will mean it stays resident past the end of the command.

We may as well add it to the threads list and clean it up properly,
hopefully removing the race in the tests from the autobuilder.

[YOCTO #13055]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9b251dcaffe52d32c1faf41ab57ab414fbc29722)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oeqa/selftest/runcmd: Add better debug for thread count mismatch failures

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d15d0177d328fa3a126b9942bda177f6fae68505)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

pseudo: Fix attr errors due to incorrect library resolution issues

On a tumbleweed system, "install X Y" was showing the error:

pseudo: ENOSYS for 'fsetxattr'.

which was being caused by dlsym() for that function returning NULL. This
appears to be due to it finding an unresolved symbol in libacl for this
symbol in libattr. It hasn't been resolved so its NULL. dlerror() returns
nothing since this is a valid symbol entry, its just not the one we want.

We can add the glibc version string for the symbol we actually want so we get
that version rather than the libattr/libacl one. The calls in libattr are just
wrappers around the libc version so our attaching to the libc versions should
intercept any accesses via these too.

[YOCTO #13952]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 82655cb26ad01de9587ef41eaef155c61c361f67)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

mime.bbclass: fix post install scriptlet error

fix error during post uninstall:
%postun(shared-mime-info-data-2.0-r0.4.corei7_64): execv(/bin/sh) pid 78
+ '[' 0 = 0 ']'
+ set -e
+ '[' x '!=' x ']'
+ echo 'Updating MIME database... this may take a while.'
Updating MIME database... this may take a while.
+ update-mime-database /usr/share/mime
Directory '/usr/share/mime/packages' does not exist!
%postun(shared-mime-info-data-2.0-r0.4.corei7_64): waitpid(78) rc 78 status 100
warning: %postun(shared-mime-info-data-2.0-r0.4.corei7_64) scriptlet failed, exit status 1

when run post uninstall scriptlet, /usr/share/mime/packages has been
removed during unintall, while update-mime-database will check xml under
/usr/share/mime/packages.

workaround by create this dir before update, then remove it

Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 6f262a316d6c32ff9ce96ab4bd95726772b5f20f)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

systemd-serialgetty: do not use BindsTo

This unit was changed from 'BindsTo' in the following commit.
"""
commit f0f359ec9210759f6b4dbfb35d3fba8af208c43a
Author: Jason Wessel <jason.wessel@windriver.com>
Date:   Thu Aug 29 07:00:31 2019 -0700

    serial-getty@.service: Allow device to fast fail if it does not exist
"""

It was changed back to 'BindTo' in the following commit.
"""
commit 63bbff61b78c651339c4b18d8376187379ec3b3c
Author: Otavio Salvador <otavio.salvador@gmail.com>
Date:   Fri Jun 12 14:30:44 2020 -0300

    systemd: Sync systemd-serialgetty@.service with upstream
"""

This is now causing runtime problem for qemuarm64. The default.target
is not reached until a timeout. Output is like below.

"""
root@qemuarm64:~# systemd-analyze
Bootup is not yet finished (org.freedesktop.systemd1.Manager.FinishTimestampMonotonic=0).
Please try again later.
Hint: Use 'systemctl list-jobs' to see active jobs
root@qemuarm64:~# systemctl list-jobs
JOB UNIT TYPE STATE
102 getty.target start waiting
1 multi-user.target start waiting
95 systemd-update-utmp-runlevel.service start waiting
110 serial-getty@hvc0.service start waiting
111 dev-hvc0.device start running
"""

We can see that we are waiting for /dev/hvc0, while in fact there's no /dev/hvc0.
Jason's commit actually solves such problem.

So restore to use Jason's method. Do not use 'BindsTo'.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 43b989c1231d3d867303ccebceda72364a9519ee)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

ell: upgrade 0.31 -> 0.32

This is a bugfix release:

ver 0.32:
Fix issue with handling D-Bus watch removal.

Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 98d97384167a8d297650e49f9cabf4fae823b4a0)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

u-boot: introduce UBOOT_INITIAL_ENV

It defaults to ${PN}-initial-env, no functional changes with current
implementation, but this allows it to be changed in individual u-boot
recipes.

If UBOOT_INITIAL_ENV is empty, then no initial env would be compiled/
installed/deployed, set ALLOW_EMPTY_${PN}-env = "1".

The major purpose for introducing this, is that the users might have
some scripts on targets like:
```
/sbin/fw_setenv -f /etc/u-boot-initial-env
```

and it should be able to run against a identical path generated by
different u-boot recipes.

Signed-off-by: Ming Liu <ming.liu@toradex.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit fb7e8b6b88855f3f523b2176ea2c85a330bfe00b)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

curl: Security fixes for CVE-2020-{8169/8177}

Source: https://curl.haxx.se/
MR: 104472, 104458
Type: Security Fix
Disposition: Backport from https://github.com/curl/curl/commit/{600a8cded447cd/8236aba58542c5f}
ChangeID: 1300924f7a64b22375b4326daeef0b686481e30c
Description:

- Affected versions: curl 7.20.0 to and including 7.70.0
- Not affected versions: curl < 7.20.0 and curl >= 7.71.0

Fixes both CVE-2020-8169 and CVE-2020-8177

Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>

perl: fix CVE-2020-10543 & CVE-2020-10878

Signed-off-by: Lee Chee Yang <chee.yang.lee@intel.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>

dbus: fix CVE-2020-12049

Signed-off-by: Lee Chee Yang <chee.yang.lee@intel.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>

gstreamer1.0-plugins-bad: add support for vdpau

libvdpau support was added to meta-oe by commit bf1de4db2 ("libvdpau:
Add recipe") in 2016. Therefore add a packageconfig for the
corresponding gstreamer plugin.

Based on a patch by Richard Leitner <richard.leitner@skidata.com>

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 878a8561e10017bb91ea42bbbe6d4acfc0985482)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

meson.bbclass: avoid unexpected operating-system names

For example 'linux-gnueabi' is not what meson consumers expect: See canonical
table of OS names for Meson (thanks Ross). Surprisingly this did not pop up
earlier but wrong system-name can break building as described in [1]

[1] https://lists.openembedded.org/g/openembedded-core/topic/meson_host_machine_system/73023862

Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 010202076760329829fbde8dab4f535d6f755873)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

recipetool: create: fix SRCBRANCH not being passed to params

When explicitly passing a branch using --srcbranch in 'devtool add' or
'recipetool create', the branch name is not included in the params of
bb.fetch2.encodeurl and default 'master' branch is used instead.

Signed-off-by: Tuomas Salokanto <tuomas.salokanto@gmail.com>
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 0424df825f1e509faf6cd44403c0736bb91b57c3)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

apr-util: make gdbm optional

This helps with gpl3-free builds.

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3260ad9e8ff185b4799269bbcdd9f638e976c4b4)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

linux-firmware: add ice for Intel E800 series driver

Signed-off-by: Yongxin Liu <yongxin.liu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit cdc65ddba0258be0c82deb4c174cd26005e7c32a)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

go.bbclass: Add `-trimpath` to default build flags

The `-trimpath` option is important for reproducible builds so full
build paths and module paths are not embedded.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e3df027b3697b6a92b417bba8d442a5bd15525fd)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

go-dep: Avoid use of 'go mod' support

dep utility must not use 'go mod' support, so we explicitly disable it.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b34000ae3dd6e0a1d7fc332efb35c5da84cf2275)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

glide: Avoid use of 'go mod' support

Glide utility must not use 'go mod' support, so we explicitly disable
it.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b5a4369ba606677285b0a89a78c040b38f57767f)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

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>
(cherry picked from commit 11c2b06ac98cc5064640705712bffa156519f450)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

net-tools: backport a patch from upstream to use the same ifconfig format as debian/ubuntu

* this is needed for python3-ifcfg to parse it correctly
  https://github.com/ftao/python-ifcfg/issues/43
* backport this single patch, so it can be backported to dunfell
  then for master we should upgrade to new snapshot from debian
  which includes other fixes and improvements as well
* this is already part of net-tools_1.60-26.diff from debian we're using
  but it's first added there and then removed
  $ grep 'sprintf(flags' net-tools_1.60-26.diff
  ++    sprintf(flags, "flags=%d<", ptr->flags);
  +-    sprintf(flags, "flags=%d<", ptr->flags);

* before:
root@qemux86-64:~# ifconfig
eth0      Link encap:Ethernet  HWaddr 52:54:00:12:34:02
          inet addr:192.168.7.2  Bcast:192.168.7.255  Mask:255.255.255.0
          inet6 addr: fe80::5054:ff:fe12:3402/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:6 errors:0 dropped:0 overruns:0 frame:0
          TX packets:106 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:496 (496.0 B)  TX bytes:42832 (41.8 KiB)

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:846 errors:0 dropped:0 overruns:0 frame:0
          TX packets:846 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:53364 (52.1 KiB)  TX bytes:53364 (52.1 KiB)

* after:

root@qemux86-64:~# ifconfig
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500  metric 1
        inet 192.168.7.2  netmask 255.255.255.0  broadcast 192.168.7.255
        inet6 fe80::5054:ff:fe12:3402  prefixlen 64  scopeid 0x20<link>
        ether 52:54:00:12:34:02  txqueuelen 1000  (Ethernet)
        RX packets 6  bytes 496 (496.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 14  bytes 2140 (2.0 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536  metric 1
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1000  (Local Loopback)
        RX packets 80  bytes 6080 (5.9 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 80  bytes 6080 (5.9 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

* for comparison ubuntu-20.04 in docker:

root@dafcbbf25ff2:/# ifconfig
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 172.17.0.2  netmask 255.255.0.0  broadcast 172.17.255.255
        ether 02:42:ac:11:00:02  txqueuelen 0  (Ethernet)
        RX packets 39299  bytes 89614740 (89.6 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 33767  bytes 3807354 (3.8 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        loop  txqueuelen 1000  (Local Loopback)
        RX packets 5658  bytes 1294220 (1.2 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 5658  bytes 1294220 (1.2 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

* python3-ifcfg now parses flags correctly:
root@qemux86-64:~# python3 -m ifcfg.cli | python3 -m json.tool
{
    "eth0": {
        "inet": "192.168.7.2",
        "inet4": [
            "192.168.7.2"
        ],
        "ether": "52:54:00:12:34:02",
        "inet6": [
            "fe80::5054:ff:fe12:3402"
        ],
        "netmask": "255.255.255.0",
        "device": "eth0",
        "flags": "4163<UP,BROADCAST,RUNNING,MULTICAST> ",
        "mtu": "1500",
        "broadcast": "192.168.7.255"
    },
    "lo": {
        "inet": "127.0.0.1",
        "inet4": [
            "127.0.0.1"
        ],
        "ether": null,
        "inet6": [
            "::1"
        ],
        "netmask": "255.0.0.0",
        "device": "lo",
        "flags": "73<UP,LOOPBACK,RUNNING> ",
        "mtu": "65536"
    },
    "sit0": {
        "inet": null,
        "inet4": [],
        "ether": null,
        "inet6": [],
        "netmask": null,
        "device": "sit0",
        "flags": "128<NOARP> ",
        "mtu": "1480"
    }
}

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit dacfd695061b186240a85f8ffc43e6636c56e1fb)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

classes/archiver: Create patched archive before configuring

do_configure and do_preconfigure can modify source files, which causes
race conditions if these tasks run in parallel with do_ar_patched. Add
explicit task dependencies to ensure that do_ar_patched finishes before
these tasks start. Specifically, this fixes a race condition with
gcc-source where do_ar_patched races with do_preconfigure deleting
gcc/gengtype-lex.c

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 8a7c7794870815030239e90b03e37ed302b7e885)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

common-licenses: add BSD-2-Clause-Patent

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d6e8c4a63caefdf36dfbc9d230c9e8db4b43dc5a)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

buildhistory-collect-srcrevs: sort directories

From: Daniel McGregor <daniel.mcgregor@vecima.com>

In order to allow consistent output of buildhistory-collect-srcrevs sort
the list of directories returned by os.walk. Otherwise the list of
SRCREVs is returned in an unspecified order.

We save the output of this command on build, so it creates smaller diffs
between builds.

Signed-off-by: Daniel McGregor <daniel.mcgregor@vecima.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3d7cb207ad4ec3cd5a3064147d3c9b5a1730d0fb)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

perl: Fix host specific modules problems

We were seeing a ton of empty perl modules being created such as
"perl-module-x86-64-linux-encoding" where the name would include
${TARGET_ARCH}-linux. These files were already being filtered in an
earlier do_split_packages() expression so exclude them from the latter
one to remove the pointless empty modules in PACKAGES.

This doesn't explain why some were not deterministic but will recude
the do_package execution time and clean up the build directories
at the very least.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9f1a959d9831f43dda656e3b0c4d059db3363877)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

bind: update to 9.11.19

Bug fix only updates.

suitable for Stable branch updates where applicable.

Drop CVE patches included in update
LIC_FILES_CHKSUM update copyright year to 2020

Full changes found at : https://gitlab.isc.org/isc-projects/bind9/-/blob/v9_11/CHANGES

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit a6ba66cf5e754cdcd41f01d233fbef7b94a10225)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

libpam: Remove option 'obscure' from common-password

libpam does not support 'obscure' checks to password,
there are the same checks in pam_cracklib module.
And this fix can remove the below error message while
updating password with 'passwd':
pam_unix(passwd:chauthtok):unrecognized option[obscure]

Signed-off-by: Haiqing Bai <Haiqing.Bai@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ea761dbac90be77797308666fe1586b05e3df824)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

install-buildtools: remove hardcoded x86-64 architecture

Remove all instances of the hardcoded 'x86_64' and replace with the current
host platform.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 52dc6f671ff67a1149be7ef4c65126ea3c907a3d)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

install-buildtools: fail if an error occurs

Several failure paths were displaying an error message but not returning,
so the install process continued and failed further.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b00e28735b64a781707441ec6187dd7f9240d97a)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

relocatable.bbclass: Avoid an exception if an empty pkgconfig dir exist

Rewrite relocatable_native_pcfiles() so that it can handle that any of
the checked pkgconfig directories are empty without causing an
exception.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f9c5df6dc1c13e9b05ff1b47ad84ad339f6779a4)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

mtd-utils: Fix return value of ubiformat

This changeset fixes a feature regression in ubiformat.  Older versions
of ubiformat, when invoked with a flash-image, would return 0 in the
case no error was encountered.  Upon upgrading to latest, it was
discovered that ubiformat returned 255 even without encountering an
error condition.

This changeset corrects the above issue and causes ubiformat, when given an
image file, to return 0 when no errors are detected.

Backport fix from 2.1.2

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7ebacd9cbaec98fbc406e8ae99c9805a24fdadc6)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

systemd: Sync systemd-serialgetty@.service with upstream

The systemd upstream has been doing some improvements in the service
which were not in sync, with the forked file. Mainly:

 - Use BindsTo to bind the service with the required serial device
 - Add of getty-pre.target as dependency so we can run things before
   getty@.service and serial-getty@.service
 - Add conflicts to rescue to avoid issues with sulogin

We did not change the ExecStart and Environment fields so we don't have
side effects of this change.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 63bbff61b78c651339c4b18d8376187379ec3b3c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

buildtools-tarball: export OPENSSL_CONF in environment setup

The autobuilder has been experiencing SSL: CERTIFICATE_VERIFY_FAILED
errors during error report uploads when using buildtools due to looking
for certs in /opt/poky

Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 197f1d5d14b8e57295f5a81c03c86abba5328614)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

recipetool: Fix list concatenation when using edit

If there are multiple appends, ensure we concatenate compatible things:

Traceback (most recent call last):
  File "/home/akiernan/poky/scripts/recipetool", line 111, in <module>
    ret = main()
  File "/home/akiernan/poky/scripts/recipetool", line 100, in main
    ret = args.func(args)
  File "/home/akiernan/poky/scripts/lib/recipetool/edit.py", line 38, in edit
    return scriptutils.run_editor([recipe_path] + appends, logger)
TypeError: can only concatenate list (not "tuple") to list

Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 4c1e74bdf4922519d168434afd69c9bebcb9bd82)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

qemu: fix CVE-2020-13361

Signed-off-by: Lee Chee Yang <chee.yang.lee@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 2dcef5dfb5c4c57fd793d04ac936a9ff73aae844)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

vim: restore the 'chmod -x' workaround in do_install

These workarounds are removed because a previous patch
solve the host path reference for gawk and perl, and it skips
the do_package_qa issue by setting the INSANE_SKIP. But it
introduces regression for do_rootfs. The dependencies are
calculated and will require packages like python, perl, gawk
and csh. The error is like below.

Error:
 Problem: conflicting requests
    - nothing provides /bin/csh needed by vim-tools-8.2-r0.corei7_64
    - nothing provides /usr/bin/nawk needed by vim-tools-8.2-r0.corei7_64
    - nothing provides /usr/bin/python needed by vim-tools-8.2-r0.corei7_64

So we keep the previous patch which solve the host path reference
problem and restore the long-used 'chmod -x' workaround here.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bf3e799e32b4de300887b844b0b7bce3d60ca379)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

encodings: clear postinst script

Postinst script from xorg-font-common.inc doesn't apply to this recipe.
So clear the postinst script of encodings.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 99ae6dbb7278dfd264453af852c108fa56a0d4e3)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

wpa-supplicant: remove service templates from SYSTEMD_SERVICE

Remove service templates wpa_supplicant-nl80211@.service and
wpa_supplicant-wired@.service from SYSTEMD_SERVICE that they should NOT
be started/stopped by calling 'systemctl' in postinst and prerm scripts.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit fe9b8e50461ab00ab3ad8b065ebd32f0eea2a255)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

mdadm: remove service template from SYSTEMD_SERVICE

Remove service template mdmon@.service from SYSTEMD_SERVICE which should
be not started by systemctl directly. It is hanlded by udev rules.

Replace tab with spaces in SRC_URI as well.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit a55b2b29bb86d98700836de5b7d68965c7cd5ee5)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

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>
(cherry picked from commit faa01e1ae7a92d60699cd8865d1106a13549b096)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

populate_sdk_base.bbclass: fix warning: name not matched

When do populate_sdk with SDK_ARCHIVE_TYPE = "zip", have below warning
in log file, this patch fixes it.

zip warning: name not matched: sysroots/core2-64-poky-linux/etc/tmpfiles.d/etc.conf
zip warning: name not matched: sysroots/core2-64-poky-linux/etc/tmpfiles.d/home.conf
zip warning: name not matched: sysroots/core2-64-poky-linux/etc/resolv-conf.systemd
zip warning: name not matched: sysroots/core2-64-poky-linux/etc/mtab
zip warning: name not matched: sysroots/core2-64-poky-linux/etc/resolv.conf
zip warning: name not matched: sysroots/core2-64-poky-linux/var/lock
zip warning: name not matched: sysroots/core2-64-poky-linux/usr/lib/environment.d/99-environment.conf
zip warning: name not matched: sysroots/core2-64-poky-linux/sbin/klogd
zip warning: name not matched: sysroots/core2-64-poky-linux/sbin/syslogd
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/usr/bin/chfn
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/usr/bin/chsh
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/usr/bin/passwd
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/usr/bin/chpasswd
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/sbin/vipw
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/sbin/vigr
zip warning: name not matched: sysroots/x86_64-pokysdk-linux/sbin/nologin

Signed-off-by: Wenlin Kang <wenlin.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 5f7df08aca81ac269490876fe84c7fc5b1d5739c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

powertop: switch to Arjan's git

01.org seems unmaintained; while the tarball is still there, the
page to download it from is gone.

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bcd6475928dba0a46d196d905cfd449ba7a3ada5)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

patchelf: switch to git

The tarball download website is full of broken links;
while direct tarball download still works, exploring
around and checking for latest versions does not.

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d1c11259464ba2d972e1e1ad5c540050ccc80c15)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

u-boot-tools: backport patch from upstream to fix gcc 10 builds

acpica: Upgrade 20200214 -> 20200430 for gcc-10 fixes

Advance to version 20200430 to avoid multiple definition errors for uninitialized variables.
In this case, it is not enough to add -fcommon to CFLAGS.

Signed-off-by: Joe Slater <joe.slater@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7e1df28fb0534587d4f9305eb163f497852df4ec)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

valgrind: Backport upstream patch to fix __getauxval needs

Drop disabling outline-atomics since that was added to fix this issue in
particular

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 702a1ae29c6a0f920bd67b7a7a41b886ac52a751)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

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>
(cherry picked from commit 3952738d083b888e5b898ed3d63a0ed9a4dd3712)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

syslinux: Fix build with gcc10

Bring in a patch from fedora to fix -fno-common issue

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f635bd89d53cc8c110b18aa593babc55eeffc511)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

squashfs-tools: Backport fix for compiling with gcc 10

Signed-off-by: Jacob Kroon <jacob.kroon@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bae1603a27fadcfa7257801167f28db3ff02dc97)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

wayland: fix condition for strndup detection

current check does not work with gcc10

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ce6c50ea68c34e02b487d54fdd9288e04833cdef)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

glibc: Update to latest on 2.31 branch

There are few fixes specifically for compiling with gcc10 that are good
to have, before hitting them later

Backport build fix from master for aarch64 with gcc10

Drop CVE-2020-10029 patch its already applied on latest 2.31 branch

latest glibc 2.31 added fix for __getauxval/aarch64 issue

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 2c7e0e0bf32eb1ed0b7d8acddb16c0d1e93f2aa1)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

matchbox-wm: Replace -fcommon with fix

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d69a05533048d0e50f5ee47b2c35675bc615728c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

menu-cache: Replace -fcommon with fix

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 97984f171b562f4ccaef527e6766b0e94697cf43)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

at-spi2-atk: Remove -fcommon

This was fixed in upstream version 2.34.0.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9562c93799a37cd18aceb6f24378ba02a69865f0)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

mesa: Remove -fcommon

This was fixed in upstream version 20.0.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 1d269a3f3a84cae92e611e02082150cfd97f8258)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

libxcrypt2: Remove -fcommon

This was fixed in upstream version 4.4.12.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 281d53ebab4c4c1b6c880ab14761d715196fb8d7)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

iproute2: Remove -fcommon

This was fixed in upstream version 5.5.0.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9a9f67b7c50a8c28a75fc48c8abcb8a7bb35f0e6)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

cpio: add patch to fix issues with gcc 10

gcc 10 introduced "-fno-common" as default. Therefore backport the
according cpio patch to fix this issue.

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ca6e5511b59b585de793ff6424c7c6ccfcd8e61e)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

binutils: add patch to fix issues with gcc 10

gcc 10 introduced "-fno-common" as default. Therefore backport the
according binutils patch to fix this issue.

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit fbeb572e1872cf64786d01d5c6408a6765f814ce)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

libcomps: update to 0.1.15

In this update the 0001-Make-__comps_objmrtree_all-static-inline.patch
was mainlined by commit 18f52cb.

Furthermore mainline commit 3237f44 fixes a build problem with the
default "-fno-common" of gcc 10.

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 90b942d01decbcff3b883ee6fb4bba67fb446817)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

dtc: update to 1.6.0

This also fixes issues with gcc 10's default change to "-fno-common".

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3acc1e5d5bce4e229801c9d66391e007390fb373)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

gdbm: add patch to fix link failure against gcc 10

This also makes appending "-fcommon" to CFLAGS obsolete and enables
native builds with a host gcc v10.

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7717b35265e6ef8ef71a4eca4074294dfcc2110c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

libtirpc: remove extra "-fcommon" from CFLAGS

commit daed7eedba ("Avoid multiple-definiton with gcc -fno-common") of
libtirpc fixed the -fno-common issues. Therefore it should be save to
remove the extra "-fcommon" in CFLAGS.

Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9e4546701775fe98532f011f5ad146cfb1b33715)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

libtirpc: upgrade 1.2.5 -> 1.2.6

0001-xdr_float-do-not-include-bits-endian.h.patch
removed since it is included in 1.2.6

Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit c940792e1492923036cae97010f9945e7ad43a32)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

vim: _FORTIFY_SOURCE=2 be gone

vim will abort in many places with this setting.  Replace
it with the benign _FORTIFY_SOURCE=1.

Signed-off-by: Joe Slater <joe.slater@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 18129cbaeddb3278efe9963718556e3765f06c1e)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

build-sysroots: add sysroot paths with native binaries to PATH

staging_populate_sysroot_dir() collects postinsts from the sysroot
and executes them. These postinsts, in turn, may call binaries that
are only available from the sysroot. This works fine with recipe-specific
sysroots, as all necessary paths are already in PATH, but breaks down
in this recipe which imitates the old global sysroot way but doesn't adjust
the PATH to include the binary paths from global sysroot.

To reproduce the failure:

$ bitbake docbook-xml-dtd4-native
$ bitbake -c build_native_sysroot build-sysroots

...

Exception: subprocess.CalledProcessError: Command '/home/akanavin/build/tmp/sysroots/x86_64/usr/bin/postinst-docbook-xml-dtd4-native-xmlcatalog' returned non-zero exit status 127.

Subprocess output:
/home/akanavin/build/tmp/sysroots/x86_64/usr/bin/postinst-docbook-xml-dtd4-native-xmlcatalog: 5: /home/akanavin/build/tmp/sysroots/x86_64/usr/bin/postinst-docbook-xml-dtd4-native-xmlcatalog: xmlcatalog: not found
/home/akanavin/build/tmp/sysroots/x86_64/usr/bin/postinst-docbook-xml-dtd4-native-xmlcatalog: 8: /home/akanavin/build/tmp/sysroots/x86_64/usr/bin/postinst-docbook-xml-dtd4-native-xmlcatalog: xmlcatalog: not found

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 6b5f7bda4204d45cd29670cefcd53dc5da031095)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

linux-firmware: upgrade 20200421 -> 20200519

License-Update: added more entries to WHENCE, all redistributable
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f8d47014e453533b98e37c653177e9541e1deeda)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

db: do not install db_verify if 'verify' is not enabled

On target, when running `db_verify /var/lib/rpm/Packages', we get
the following error.

  db_verify: BDB0571 library build did not include support for database verification

This is because db_verify is installed but 'verify' PACKAGECONFIG is
not enabled. So fix it by not installing do_verify in such case.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b5ffdb296f211403f987b5d5a724a94ee91dc80c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

qemu: force build type to production

qemu will not build for -Og optimization because macros
in lockable.h expect dead-code elimination.  Override DEBUG_BUILD.

Signed-off-by: Joe Slater <joe.slater@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ad404b2e51223625bd95f9a7da4c7a690c37bdfb)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oe-selftest: Recursively patch test case paths

This ensures that builddir is updated correctly to point to the new
selftest build directory when we're given a list of test suites instead
of a list of test cases.

Signed-off-by: Paul Barker <pbarker@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 56e211d0f3c6cb84f5982e5de00faeeed69c1912)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oe-selftest: Allow overriding the build directory used for tests

This may be useful if the parent directory of the original builddir is
not writable, on a lower performance drive, etc.

Signed-off-by: Paul Barker <pbarker@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit dab719e5f1964fd48e9a35e6b9db1508b5414540)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

archiver: Fix test case for srpm archiver mode

We can't assume that rpm packaging will always be enabled when running
oe-selftest. In particular when using nodistro instead of poky this is
not enabled by default.

Signed-off-by: Paul Barker <pbarker@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 23df3024664ec56dbbe8865ef3953d2c33a7a2cb)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

patchelf: Add patch to address corrupt shared library issue

patchelf can corrupt shared libraries if the program headers don't
immediately follow the elf header. Add a patch submitted upstream
to address this.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e7811c787bbe2f5d49b3506309499acc27189988)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

cogl: point to correct HOMEPAGE

As the previousy set wasn't reachable anymore

Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 3ae0aa1df25c8c7511950c6fc965955bdb205cd2)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

cve-check: include epoch in product version output

In the generated cve.log files, include the epoch in the product
version. This better matches how versions are displayed elsewhere,
in particular the bb.warn("Found unpatched CVE...") that appears
on the terminal when CVEs are found.

Signed-off-by: Ralph Siemsen <ralph.siemsen@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e1c3c0b6e5b01304e2127f5058986697e82adf93)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

oeqa/targetcontrol: Rework exception handling to avoid warnings

We're seeing:

WARNING: bitbake/lib/bb/cookerdata.py:136: ResourceWarning: unclosed file <_io.FileIO
name='tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0-r0/testimage/qemurunner_log.20200601181912'
mode='ab' closefd=True

which can only be caused by the qemu.stop() method not being called.
Tweak the error handling to fix the blanket exception handler which
is likely meaning this function isn't getting called.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ee707090848d793e3b2d82dd3861ae22222682c0)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

cairo: Do not try to remove nonexistent directories

Commit 0e1f8fa0 (bitbake.conf: propagate 'opengl' DISTRO_FEATURE to
native/nativesdk from target) changed the default PACKAGECONFIG for
native and nativesdk so that it becomes empty unless "x11" is in
DISTRO_FEATURES since "trace" was also removed (propbably
unintentionally). This highlighted than an empty PACKAGECONFIG would
lead to a build failure since /usr/bin is never created under these
conditions, but the recipe still tried to remove it.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9bb2268677ac8f0c97433bf1f04555abe88028a9)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

logrotate: Drop obsolete setting/comment

The comment applies to an old version and the WORKDIR setting matches
the default so drop it.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7cc4c0e7665111616c7e8a1dc512b8d9022ca664)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

maintainers: Update Ross' email address

Update Ross' address to a valid one.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 22e1605ce45482c49d06a7c56a46b71f76955e31)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

lz4: disable static library

It was found to be built in a non-reproducible way.

[YOCTO #13925]

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f8f3f03e7ffcefdb6a97b747a26276dec4bfa77c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

sysfsutils: rem leftover settings for libsysfs-dev

22af6a2595dbec98ce4a2e3b1324ad8d400390ad removed the PACKAGES
setting, but left the FILES-assignments of libsysfs-dev and -staticdev.
As these have no use anymore they can be safely removed

Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 941cc0cd0a69e95ca571a9a8cb0732bb06fa2b3a)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

util-linux: alternatify rtcwake

This tool can be turned on in busybox which lead into a error during
do_rootfs.

Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit eb22e7a1190a79fd74ae54f23a13ec4932f41b0d)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

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>
(cherry picked from commit 0739a8901140c05d037517ffd89382f151ba627c)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

core-image-minimal-initramfs: keep restriction with initramfs-module-install

Since commit [882ae0dcce core-image-minimal-initramfs: use
initramfs-framework for initialization] applied, it should
keep the restriction with initramfs-module-install

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 295dbabce213855432b6274026e3b45a18b90e80)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

build-appliance-image: Update to dunfell head revision

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

build-appliance: Update branch to point at dunfell

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

build-appliance-image: Update to dunfell head revision

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

layer.conf: Bump OE-Core layer version

The commits 910ffaf5be ("pyelftools: Import from meta-python") and
a96f815c53 ("pycryptodome: Import from meta-python") moved recipes from
meta-python to oe-core. In order for this to be communicated with users,
bump the LAYERVERSION so that meta-python can key of it in its
LAYERDEPENDS.

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 2d503b27e7c88cee9a37c79c4605c77b11f230b6)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

python3-pyelftools: Upgrade 0.25 -> 0.26

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 0323e12624ef45e64e7a8ba6384c06e4d42df064)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

python3-pycryptodome(x): Upgrade 3.9.4 -> 3.9.7

Also splits apart the SRC_URI checksums to make automatic upgrades
easier

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit ae1f210546396b761ea86d9e32bf90c0867ff845)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

pyelftools: Import from meta-python

Imports the pyelftools recipes from meta-python, as of 7c02c7d41
("gnome-themes-extra: correct the recipe name").

This recipe is commonly used by other layers, so moving it into
OE-core helps to cut down on layer dependencies.

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 910ffaf5beed42936588c95b0c7c1b1ad67f99d3)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

pycryptodome: Import from meta-python

Imports the pycryptodome recipes from meta-python, as of 7c02c7d41
("gnome-themes-extra: correct the recipe name").

These recipes are commonly used by other layers, so moving them into
OE-core helps to cut down on layer dependencies.

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit a96f815c53364b119b5743b8b7100eb5588d5cf5)
Signed-off-by: Steve Sakoman <steve@sakoman.com>

linux-yocto/5.4: update to v5.4.43

Updating linux-yocto/5.4 to the latest korg -stable release that comprises
the following commits:

    e0d81ce76004 Linux 5.4.43
    b5100186021a sched/fair: Fix enqueue_task_fair() warning some more
    8b13f5657fa8 sched/fair: Fix reordering of enqueue/dequeue_task_fair()
    a2ad232aa6a9 sched/fair: Reorder enqueue/dequeue_task_fair path
    f4520daa3c5a tpm: check event log version before reading final events
    68b7b8183c12 rxrpc: Fix ack discard
    283eb0016f97 rxrpc: Trace discarded ACKs
    f2da8c0dfe81 iio: adc: stm32-dfsdm: fix device used to request dma
    7b5af65ca246 iio: adc: stm32-dfsdm: Use dma_request_chan() instead dma_request_slave_channel()
    692001a867b6 iio: adc: stm32-adc: fix device used to request dma
    8e8836b2b782 iio: adc: stm32-adc: Use dma_request_chan() instead dma_request_slave_channel()
    1084eee4dc5d x86/unwind/orc: Fix unwind_get_return_address_ptr() for inactive tasks
    860fe59783a9 flow_dissector: Drop BPF flow dissector prog ref on netns cleanup
    bd6f0c799f4d s390/kexec_file: fix initrd location for kdump kernel
    834d24ec3a0a rxrpc: Fix a memory leak in rxkad_verify_response()
    23ae6e3e8aeb rxrpc: Fix the excessive initial retransmission timeout
    c2a26769b43e kasan: disable branch tracing for core runtime
    86217fecc4b7 rapidio: fix an error in get_user_pages_fast() error handling
    689dacb2b09d device-dax: don't leak kernel memory to user space after unloading kmem
    9e451933bba9 s390/kaslr: add support for R_390_JMP_SLOT relocation type
    72f3241508ac s390/pci: Fix s390_mmio_read/write with MIO
    9c84884cd5dc ipack: tpci200: fix error return code in tpci200_register()
    46f47dda27bc mei: release me_cl object reference
    f505a3e24c6a tty: serial: add missing spin_lock_init for SiFive serial console
    a5b4b3f97de7 misc: rtsx: Add short delay after exit from ASPM
    e64b205035fb iio: adc: ti-ads8344: Fix channel selection
    9af65dc54b9b iio: dac: vf610: Fix an error handling path in 'vf610_dac_probe()'
    d54e5a4ff04c iio: sca3000: Remove an erroneous 'get_device()'
    56cff2ac7c9d staging: greybus: Fix uninitialized scalar variable
    a41e02cb4232 staging: kpc2000: fix error return code in kp2000_pcie_probe()
    dee81110a488 staging: iio: ad2s1210: Fix SPI reading
    76296dc723ef media: fdp1: Fix R-Car M3-N naming in debug message
    4adb7a2b3161 Revert "gfs2: Don't demote a glock until its revokes are written"
    bb6524537dc2 kbuild: Remove debug info from kallsyms linking
    ee71c590dd8d bpf: Avoid setting bpf insns pages read-only when prog is jited
    4c732e81bd4d powerpc/64s: Disable STRICT_KERNEL_RWX
    b67da9dbdb89 powerpc: Remove STRICT_KERNEL_RWX incompatibility with RELOCATABLE
    9bcfbd8ba2b5 drm/i915: Propagate error from completed fences
    5e171483e947 drm/i915/gvt: Init DPLL/DDI vreg for virtual display instead of inheritance.
    0e1d5f67253e vsprintf: don't obfuscate NULL and error pointers
    4b1b34621998 dmaengine: owl: Use correct lock in owl_dma_get_pchan()
    0fcbe108b01a dmaengine: dmatest: Restore default for channel
    57c32a52c3fe drm/etnaviv: Fix a leak in submit_pin_objects()
    432b103596bd dmaengine: tegra210-adma: Fix an error handling path in 'tegra_adma_probe()'
    870a45e0b507 apparmor: Fix aa_label refcnt leak in policy_update
    054934aa9faa apparmor: fix potential label refcnt leak in aa_change_profile
    97d817b9ef13 apparmor: Fix use-after-free in aa_audit_rule_init
    3b1e38dfbc9f drm/etnaviv: fix perfmon domain interation
    53683907ef68 arm64: Fix PTRACE_SYSEMU semantics
    96e56055a2f0 scsi: target: Put lun_ref at end of tmr processing
    818657105a0b scsi: qla2xxx: Do not log message when reading port speed via sysfs
    d54c5eff8795 ALSA: hda/realtek - Add more fixup entries for Clevo machines
    80f5822c2bf3 ALSA: hda/realtek - Fix silent output on Gigabyte X570 Aorus Xtreme
    1b17a0f98ad0 ALSA: pcm: fix incorrect hw_base increase
    a44cb2581718 ALSA: iec1712: Initialize STDSP24 properly when using the model=staudio option
    99e392a4979b KVM: x86: Fix pkru save/restore when guest CR4.PKE=0, move it to x86.c
    1c3d707d7d12 ALSA: hda/realtek: Enable headset mic of ASUS UX581LV with ALC295
    26a3a3053332 ALSA: hda/realtek - Enable headset mic of ASUS UX550GE with ALC295
    c5742497dcd2 ALSA: hda/realtek - Enable headset mic of ASUS GL503VM with ALC295
    2523e9010d2b ALSA: hda/realtek: Add quirk for Samsung Notebook
    6cc4dd44e207 ALSA: hda/realtek - Add HP new mute led supported for ALC236
    0d189b31c4d7 ALSA: hda/realtek - Add supported new mute Led for HP
    69d5dc286d05 ALSA: hda: Manage concurrent reg access more properly
    1efaaf74528c ALSA: hda: patch_realtek: fix empty macro usage in if block
    749e58bd2b09 ALSA: hda - constify and cleanup static NodeID tables
    02ebbd1da394 scripts/gdb: repair rb_first() and rb_last()
    9eff404a4382 ARM: futex: Address build warning
    67a5c3104d12 KVM: selftests: Fix build for evmcs.h
    4f48af814798 drm/amd/display: Prevent dpcd reads with passive dongles
    e0bb3075f91b iommu/amd: Call domain_flush_complete() in update_domain()
    e1efb9893bdd platform/x86: asus-nb-wmi: Do not load on Asus T100TA and T200TA
    c8d323578e36 USB: core: Fix misleading driver bug report
    42b32a43529e stmmac: fix pointer check after utilization in stmmac_interrupt
    b68d27c5fffd ceph: fix double unlock in handle_cap_export()
    df0df8ee2ac7 HID: quirks: Add HID_QUIRK_NO_INIT_REPORTS quirk for Dell K12A keyboard-dock
    8a5de4a391e4 gtp: set NLM_F_MULTI flag in gtp_genl_dump_pdp()
    7932168ec08e x86/apic: Move TSC deadline timer debug printk
    1ae9f1a62a50 ftrace/selftest: make unresolved cases cause failure if --fail-unresolved set
    2eac9d3dc23f ibmvnic: Skip fatal error reset after passive init
    f82a3013226e x86/mm/cpa: Flush direct map alias during cpa
    632db044ab99 HID: i2c-hid: reset Synaptics SYNA2393 on resume
    acd3efa17d96 scsi: ibmvscsi: Fix WARN_ON during event pool release
    6ef21295dc20 net/ena: Fix build warning in ena_xdp_set()
    d0db69f9d132 component: Silence bind error on -EPROBE_DEFER
    7a5f60dc3a67 aquantia: Fix the media type of AQC100 ethernet controller in the driver
    445437b417b6 vhost/vsock: fix packet delivery order to monitoring devices
    dcec6678c3b1 configfs: fix config_item refcnt leak in configfs_rmdir()
    2b52a61adb38 scsi: qla2xxx: Delete all sessions before unregister local nvme port
    d2430cb7f2d4 scsi: qla2xxx: Fix hang when issuing nvme disconnect-all in NPIV
    7b481b802a8f HID: alps: ALPS_1657 is too specific; use U1_UNICORN_LEGACY instead
    a08626f6e982 HID: alps: Add AUI1657 device ID
    68988c00b153 HID: multitouch: add eGalaxTouch P80H84 support
    cc6428803d22 gcc-common.h: Update for GCC 10
    3c140d22e3c2 net: drop_monitor: use IS_REACHABLE() to guard net_dm_hw_report()
    87863a7426b2 kbuild: avoid concurrency issue in parallel building dtbs and dtbs_check
    44fd02a3d719 mtd: Fix mtd not registered due to nvmem name collision
    496c7c61bd64 afs: Don't unlock fetched data pages until the op completes successfully
    17c9595cca71 ubi: Fix seq_file usage in detailed_erase_block_info debugfs file
    274cd3c7b5d3 i2c: mux: demux-pinctrl: Fix an error handling path in 'i2c_demux_pinctrl_probe()'
    dd540f2d7c2d evm: Fix a small race in init_desc()
    f96ab0d1f3ec iommu/amd: Fix over-read of ACPI UID from IVRS table
    33769c19feba i2c: fix missing pm_runtime_put_sync in i2c_device_probe
    9f885f17501d ubifs: remove broken lazytime support
    ac6f94d3be65 fix multiplication overflow in copy_fdtable()
    725b0bb0f94d mtd: spinand: Propagate ECC information to the MTD structure
    e3637eb6a351 ACPI: EC: PM: Avoid flushing EC work when EC GPE is inactive
    3be8ece11440 ubifs: fix wrong use of crypto_shash_descsize()
    48bbd44f5fa9 ima: Fix return value of ima_write_policy()
    1066327bf936 evm: Check also if *tfm is an error pointer in init_desc()
    4aedc534b608 ima: Set file->f_mode instead of file->f_flags in ima_calc_file_hash()
    ac46cea606d5 KVM: SVM: Fix potential memory leak in svm_cpu_init()
    1bed86cfe5cb i2c: dev: Fix the race between the release of i2c_dev and cdev

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9cd117dec502f40402ebd3a09ed3e8dba804ce2b)
Signed-off-by: Steve Sakoman <steve@sakoman.com>