Jackie Huang [Wed, 10 Aug 2016 03:19:44 +0000 (11:19 +0800)]
e2fsprogs: Fix missing check for permission denied.
If the path to "ROOT_SYSCONFDIR /mke2fs.conf" has a permission denied problem,
then the get_dirlist() call will return EACCES. But the code in profile_init
will treat that as a fatal error and all executions will fail with:
Couldn't init profile successfully (error: 13).
But the problem should not really be visible for the target package as the path
then will be "/etc/mke2fs.conf", and it is not likely that a user have no
permission to read /etc.
Signed-off-by: Jian Liu <jian.liu@windriver.com> Signed-off-by: Jackie Huang <jackie.huang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
initrdscripts/init-install*: Add rootwait when installing to USB devices
It can take a bit for USB devices to be detected, so if a USB device is
your rootfs and you don't set rootwait you will most likely get a kernel
panic. Fix this by adding rootwait to the kernel command line on
installation.
Fixes [YOCTO #9462].
Signed-off-by: California Sullivan <california.l.sullivan@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mike Looijmans [Tue, 9 Aug 2016 12:19:00 +0000 (14:19 +0200)]
dropbear/init: Allow extra arguments for key generation
This patch adds DROPBEAR_RSAKEY_ARGS and DROPBEAR_DSSKEY_ARGS optional
parameters to /etc/default/dropbear. The contents are simply passed to
the 'dropbearkey' program when generating a host key.
The default keysize for RSA is currently 2048 bits. It takes a CortexA9
running at 700MHz between 4 and 10 seconds to calculate a keypair. The
board boots Linux in about a second, but you have to wait for several
seconds because of the keypair generation. This patch allows one to put
the line DROPBEAR_RSAKEY_ARGS="-s 1024" into /etc/default/dropbear, and
have a host key generated in about 0.2 seconds on the same CPU. This is
particulary useful for read-only rootfs systems which generate a key on
each boot.
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl> Signed-off-by: Ross Burton <ross.burton@intel.com>
initramfs-live-boot: Make sure we kill udev before switching root when live booting
When live booting, we need to make sure the running udev processes are killed
to avoid unexepected behavior, we do this just before switching root,
once we do, a new udev process will be spawned from init and will take care
of whatever work was still missing
[YOCTO #9520]
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Hatle [Mon, 7 Dec 2015 21:36:11 +0000 (15:36 -0600)]
cross-canadian.bbclass: Add BASECANADIANEXTRAOS to specify main extraos
By default the system will expand the extra os entries for uclibc and musl
even if they are not enabled in the build. There was no way to prevent this
behavior while still getting the expansion for things like x32 or spe.
The change adds a new setting which a distribution creator can override
easily, setting the base set of canadianextraos components. The other
expansions are then based on this setting.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Richard Purdie [Mon, 15 Aug 2016 16:56:40 +0000 (17:56 +0100)]
oeqa/oetest: Improve subprocess error reporting
Without this, we get to know the command failed and the exit code but
have no idea how the command failed since we don't get the output by
default.
This makes it much easier to see what went wrong and stand a chance of
fixing it.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 15 Aug 2016 17:05:24 +0000 (18:05 +0100)]
report-error: Fix tracebacks
Currently the code gives tracebacks if there are no recipes to be built in a
BuildStarted event. Parse the list into a string rather than just taking the
first item. There is nothing special about the first time.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 12 Aug 2016 22:00:11 +0000 (23:00 +0100)]
uninative: Update to 1.3
Uninative 1.2 didn't contain the nativesdk locale fix we really needed
to release and update to uninative 1.3 which does contain that fix
and also uses glibc 2.24 final release.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stephano Cetola [Wed, 10 Aug 2016 20:03:16 +0000 (13:03 -0700)]
Allow for simultaneous do_rootfs tasks with rpm
Give each rootfs its own RPM channel to use. This puts the RPM metadata
in a private subdirectory of $WORKDIR, rather than living in DEPLOY_DIR
where other tasks may race with it.
This allows us to reduce the time that the rpm.lock is held to only the
time needed to hardlink the RPMs, allowing the majority of the rootfs
operation to run in parallel.
Also, this fixes the smart tests by generating an index for all packages
at the time of the test, rather than using the one provided by the
rootfs process.
Original credit for the enhancement should go to Steven Walter
stevenrwalter@gmail.com.
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 9 Aug 2016 15:11:44 +0000 (16:11 +0100)]
connman: clean up musl fixes
The upstreamable include fixes have been sent upstream. The patch set adds
AC_USE_SYSTEM_EXTENSIONS so we don't need to explictly define _GNU_SOURCE
anymore.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Tue, 9 Aug 2016 11:12:02 +0000 (12:12 +0100)]
connman: disable version-scripts to fix crashes at startup
With binutils 2.27 on at least MIPS, connmand will crash on startup. This
appears to be due to the symbol visibilty scripts hiding symbols that stdio
looks up at runtime, resulting in it segfaulting.
This certainly appears to be a bug in binutils 2.27 although the problem has
been known about for some time:
Stefan Agner [Mon, 8 Aug 2016 05:47:28 +0000 (22:47 -0700)]
busybox: Fix busybox-init on non-tty consoles
When using non-tty consoles (e.g. VirtIO console /dev/hvc0) the
current init system fails with:
process '/sbin/getty 115200 hvc0' (pid 545) exited. Scheduling for restart.
can't open /dev/ttyhvc0: No such file or directory
The first field needs to be a valid device. The BusyBox inittab example
explains as follows:
"<id>: WARNING: This field has a non-traditional meaning for BusyBox init!
The id field is used by BusyBox init to specify the controlling tty for
the specified process to run on. The contents of this field are
appended to "/dev/" and used as-is."
Signed-off-by: Stefan Agner <stefan@agner.ch> Signed-off-by: Ross Burton <ross.burton@intel.com>
Prior to running oe_runmake make sure $B is the cwd. This is required
due to bitbake commit 67a7b8b021badc17d8fdf447c250e79d291e75f7
"build: don't use $B as the default cwd for functions".
Without this change, do_concat_dtb fails with:
| ERROR: oe_runmake failed
| make: *** No targets specified and no makefile found. Stop.
Signed-off-by: George McCollister <george.mccollister@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
If INITRAMFS_IMAGE is set, build an additional fitImage containing the
initramfs. Copy the additional fitImage and the source (*.its) file, used
to create it to DEPLOYDIR. The fitImage containing the initramfs must be
built before do_deploy and after do_install to avoid circular dependencies.
UBOOT_RD_LOADADDRESS - Specifies the load address used by u-boot for the
initramfs.
UBOOT_RD_ENTRYPOINT - Specifies the entry point used by u-boot for the
initramfs.
Signed-off-by: George McCollister <george.mccollister@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Andrew Bradford [Thu, 4 Aug 2016 17:40:40 +0000 (13:40 -0400)]
kernel-fitimage.bbclass: do_assemble_fitimage(): cd to $B
Prior to assembling the fitimage, ensure that $B is the cwd due to
bitbake commit 67a7b8b021badc17d8fdf447c250e79d291e75f7 "build: don't
use $B as the default cwd for functions".
Without this change, do_assemble_fitimage() fails like:
Log data follows:
| DEBUG: Executing shell function do_assemble_fitimage
| arm-ka-linux-gnueabi-objcopy: 'vmlinux': No such file
| WARNING: exit code 1 from a shell command.
| ERROR: Function failed: do_assemble_fitimage
Signed-off-by: Andrew Bradford <andrew.bradford@kodakalaris.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Wed, 27 Jul 2016 07:51:10 +0000 (00:51 -0700)]
packagefeed-stability.bbclass: fix multilib + rpm
* Fix multilib + rpm since its multilib package name is special.
* Update SSTATE_DUPWHITELIST to avoid shared location conflicted error.
* Fix message when "not copying", now the messages are:
Copying packages for recipe <foo>
Not copying packages for recipe <foo>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
This is a non-functional change, which intends to correct element
names of a tuple returned by Popen.communicate().
Both in python2 and python3 subprocess.Popen.communicate() method
returns a tuple (stdoutdata, stderrdata), thus old assignments and
collateral comments are incorrect from human's point of view, however
formally there is no error in the code.
The change is desired to have to avoid copy-paste errors in future.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
This is a non-functional change, which intends to correct element
names of a tuple returned by Popen.communicate().
Both in python2 and python3 subprocess.Popen.communicate() method
returns a tuple (stdoutdata, stderrdata), thus old assignments and
collateral comments are incorrect from human's point of view, however
formally there is no error in the code.
The change is desired to have to avoid copy-paste errors in future.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Fri, 5 Aug 2016 10:23:52 +0000 (13:23 +0300)]
license: improve handling of license files with identical basenames
Previously, find_license_files() in license.bbclass just blindly assumed
that all different licenses specified in LIC_FILES_CHKSUM have unique
filenames. As a consequence, only the last one of these similarly named
license files was copied and the rest were "lost". This patch changes
the behavior so that all license files get copied. However, if multiple
identically named files are found, they are renamed to <file>.0,
<file>.1 etc.
The patch also changes the handling of NO_GENERIC_LICENSE slightly.
Previously, only basenames of NO_GENERIC_LICENSE and LIC_FILES_CHKSUM
were compared when searching for the correct license file. After this
patch NO_GENERIC_LICENSE must have the full path, matching what is
specified in LIC_FILES_CHKSUM. This is required in order to be able
to handle identical filenames (basenames) consistently. For example, if
you have:
LICENSE = "my-custom-license"
LIC_FILES_CHKSUM = "file://src/LICENCE;md5=d41d8cd98f00b204e9800998ecf8427e"
you must specify:
NO_GENERIC_LICENSE[my-custom-license] = "src/LICENCE"
[YOCTO #9663]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
sysvinit-inittab_2.88dsf.bb: Allow aliasing with SERIAL_CONSOLES_CHECK
With some hardware the name of the device node and the name in
/proc/console differ. This causes SERIAL_CONSOLES_CHECK to not enable
working consoles in these cases. This patch changes SERIAL_CONSOLES_CHECK
to have an optional alias for the checked consoles. The new format is:
<device>:<alias to check(optional)>
Fixes [YOCTO #9440].
Signed-off-by: California Sullivan <california.l.sullivan@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Mon, 8 Aug 2016 13:01:45 +0000 (14:01 +0100)]
buildhistory: output package variables in a repeatable order
This code was outputting variables by iterating a dictionary. In Py2 this
always results in the same iteration order but with Py3 the order changes every
execution, which resulted in buildhistory having to store diffs where fields
were simply re-ordered.
Signed-off-by: Ross Burton <ross.burton@intel.com>
dpkg is using a script (dpkg-architecture.pl) to detect the target
architecture automatically.
Unfortunately, it is using the cross compiler prefix to do the detection
and for ARM, oe-core is using <vendor>-linux-gnueabi for toolchains with
and without call-convention hard. The script then always detects
'armel' and never gets 'armhf' for call-convention hard.
This solves:
dpkg: error processing archive evtest_1.32+0+b8343ec112-r0_armhf.deb (--install):
package architecture (armhf) does not match system (armel)
Errors were encountered while processing:
evtest_1.32+0+b8343ec112-r0_armhf.deb
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Khem Raj [Mon, 8 Aug 2016 06:47:58 +0000 (23:47 -0700)]
world-broken: Add libunwind for musl/arm
gcc and libunwind race and when libunwind wins
results are build fails for gcc as described
this is only seen on musl/arm since on musl
/usr/include is search before gcc fixed headers
and unwind.h is in fixed headers. So it works
ok on glibc but not on musl due to reversed search
order.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mariano Lopez [Fri, 5 Aug 2016 11:25:20 +0000 (11:25 +0000)]
quilt/acl/attc/sed: Fix use of tar's --exclude option for tar >= 1.29
Starting from tar 1.29 the --exclude option won't work
anymore if is not used before the path. There are some
recipes that copy the ptest using tar and --exclude
option. This fixes these for OE-Core recipes.
[YOCTO #9763]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
The bat PACKAGECONFIG does not install the test script correctly. Fix
this by following the packaging used for the other bash scripts. While
at it, fix some tabs.
Signed-off-by: Stefan Müller-Klieser <s.mueller-klieser@phytec.de> Signed-off-by: Ross Burton <ross.burton@intel.com>
Andrej Valek [Fri, 5 Aug 2016 11:16:33 +0000 (13:16 +0200)]
openssl: fix add missing `make depend` command before `make` library
Settings from EXTRA_OECONF like en/disable no-ssl3, are transferred
only into DEPFLAGS. It means that settings have no effect on output files.
DEPFLAGS will be transferred into output files with make depend command.
Markus Lehtonen [Thu, 4 Aug 2016 09:02:11 +0000 (12:02 +0300)]
package.bbclass: better handling of middle-path dir symlinks
For example in a directory structure like this
.
├── symlink -> foo/bar
└── foo
└── bar
└── file
'file' could be referenced by specifying e.g. 'foo/bar/file' or
'symlink/file'. In cases like this populate_packages() might crash if
the file was referenced (in FILES) via the symlinked directory. The
outcome depends on how the user defined FILES_pn. This patch should
make the function behave more consistently. It looks for files which are
referenced via symlinked directories and handles them separately,
failing if their parent directory is a non-existent path. For example,
defining FILES_{PN} = "symlink/file" causes a build failure because
symlinks target 'foo/bar' is not included at all.
[YOCTO #9827]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Pascal Bach [Thu, 4 Aug 2016 13:28:46 +0000 (15:28 +0200)]
util-linux: make prlimit a separate package
Busybox doesn't provide a similar tool so having it in
a separate package allows to us it in addition to busybox without having
to include all of util-linux.
Before it was part of the top level util-linux package.
Now it is a separate package util-linux-prlimit but the top level package
still RRECOMMENDS it so for most users nothing should change.
Signed-off-by: Pascal Bach <pascal.bach@siemens.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
The output format was updated to match yocto ptest rules:
<result>: <testname>
where the result can be PASS, FAIL, or SKIP, and the testname
can be any identifying string.
Signed-off-by: Li Wang <li.wang@windriver.com> Signed-off-by: Mingli Yu <mingli.yu@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Thu, 4 Aug 2016 16:04:51 +0000 (17:04 +0100)]
oeqa/selftest: disable report-error class when builds are expected to fail
Some invocations of bitbake are expected to fail, so we don't want to report the
errors to errors.yoctoproject.org. Also rewrite the messages in
test_invalid_patch so they reflect reality.
[ YOCTO #10052 ]
Signed-off-by: Ross Burton <ross.burton@intel.com>
Davis, Michael [Wed, 3 Aug 2016 21:26:57 +0000 (21:26 +0000)]
pseudo: Add nobody user and group
Nodejs expects the user and group nobody to exist on global install commands.
The target build works as base-passwd contained it, however the fallback passwd did not.
This broke the SDK if nodejs was included.
Signed-off-by: Michael Davis <michael.davis@essvote.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Wed, 3 Aug 2016 08:37:53 +0000 (01:37 -0700)]
package.bbclass: remove unneeded chmod() and chown()
* The mode and owner info are saved in inode, hardlink won't change them,
so remove unneeded chmod() and chown().
* This can avoid the problem that when do_package re-run, the file's mode
maybe different if it is 0444 (changed to 0644 when re-run), this is
caused by pseudo adds 'w' on real file, and doesn't track linked source
when hard link, Peter and Mark may fix pseudo, but the removed code is not
needed, which can avoid the problem.
* To reproduce the problem, for example, version.c from gzip's ${B}:
1) bitbake gzip
2) Edit rpm-native or package.bbclass to make do_package re-run.
3) bitbake gzip
After the first build, build/version.c in gzip-dbg is 0444, but after
the second build, it will be 0644, this because do_package does:
$ ln ${B}/version.c gzip-dbg/version.c,
$ chmod 0444 gzip-dbg/version.c (it runs chmod 0644 on the real filesystem)
And in the second build, the gzip-dbg/version.c will be removed and
created again, so that stat() can't get 0444 but 0644 since
${B}/version.c is not tracked by pseudo.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Fri, 5 Aug 2016 15:17:09 +0000 (16:17 +0100)]
autotools: move aclocal-copy to WORKDIR
To save time move the temporary copy of the autoconf macros, aclocal-copy, from
${B} to ${WORKDIR}. This ensures that it can't conflict with anything in ${S}
and means the pruning code doesn't need to know about it.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 5 Aug 2016 15:17:08 +0000 (16:17 +0100)]
intltool: bump serial for aclocal --install
We modify this macro and need it to be used over local copies in tarballs. It
appears that aclocal doesn't quite want to do the right thing just yet but
increase the version just in case it does in the future.
Upstream typically increments by one, and autoconf handles point versions fine,
so bump it by 0.1.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Aníbal Limón [Fri, 5 Aug 2016 21:42:48 +0000 (16:42 -0500)]
apt: Fix build in musl it was break due to upgrade to 1.2.12
methods/connect.cc: Musl doesn't support AI_IDN flag in netdb.h
header so define it manually.
apt-pkg/contrib/srvrec.h: Add explicity include of sys/types.h
to avoid errors in types u_int_SIZE.
Dengke Du [Wed, 27 Jul 2016 05:02:03 +0000 (01:02 -0400)]
perf: enable man pages for 'help' functionality
When using 'perf help <subcommand>', it can't find 'man' command and pages for
perf.
The perf man pages depends on the xmlto-native and asciidoc-native, so we just
need to add the two packages to the DEPENDS variable and add 'man'to
RDEPENDS_perf-doc.
Signed-off-by: Dengke Du <dengke.du@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Wed, 20 Apr 2016 00:30:03 +0000 (17:30 -0700)]
create-pull-request: add option -a to auto push
Before this patch, we need two steps to create PULL:
* Step 1, create branch:
$ git push <contrib> <local_branch>:<remote_branch>
* Step 2, create PULL:
$ create-pull-request -u <contrib> -l <local_branch> -b <remote_branch> -r <local_branch>~<n>
We can see that the args used in step 1 are in step 2, so we can use
"create-pull-request -a" or set CPR_CONTRIB_AUTO_PUSH in to create the
branch to simplify the steps.
[YOCTO #9408]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Dengke Du [Wed, 27 Jul 2016 05:02:02 +0000 (01:02 -0400)]
Asciidoc: add it
In order to enable perf man pages for basic 'help' functionality,
it needs to produce man pages for perf, which depends on the xmlto
and asciidoc tools.
So add the asciidoc recipe to help produce man pages for perf and
other commands.
Signed-off-by: Dengke Du <dengke.du@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
* When adding new source files from upstream the autogen.sh
script needs to be run
* Rework grub2-remove-sparc64-setup-from-x86-builds.patch
to remove the grub-setup helper program grub-sparc64-setup
in Makefile.util.def instead of the previous Makefile.util.am
to avoid the update for Makefile.util.am in do_patch phase is
overwritten by the autogen.sh in do_configure phase
oeqa/utils/sshcontrol.py: Allows to copy symlinks to target
Currently when copying a symlink to the target it will fail
throwing an exception. This will recreate symlinks from the
system performing the tests to the device under tests.
[YOCTO #9932]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
When using "auto" in TEST_SUITES there is a check
for __init__.py in the <layer>/lib/oeqa/runtime/
directory in all the layers in BBLAYERS.
This check was needed because the way that python 2
import the modules, now that bitbake uses python 3
there is no need these __init__.py files, moreover
these files won't allow to import tests from other
layers.
This patch removes the check.
[YOCTO #9996]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Jeremy Puhlman [Mon, 1 Aug 2016 16:02:55 +0000 (09:02 -0700)]
bitbake.conf: set READELF for cross compilation
In the case of using an external toolchain that supports multilib
compilation with a single binary, TARGET_PREFIX is the same for both main
and multilib abis. Without READELF exported, python3 assumes it is
either the readelf for ${BUILD_SYS}-readelf. Exporting cross readelf
fixes the build issue.
checking LDLIBRARY... libpython$(LDVERSION).so
checking for i586-montavistamllib32-linux-ranlib...
x86_64-montavista-linux-ranlib
checking for i586-montavistamllib32-linux-ar...
x86_64-montavista-linux-ar
checking for i586-montavistamllib32-linux-readelf... no
checking for readelf... readelf
configure: WARNING: using cross tools not prefixed with host triplet
Signed-off-by: Jeremy Puhlman <jpuhlman@mvista.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
