Richard Purdie [Sun, 20 Dec 2015 13:20:20 +0000 (13:20 +0000)]
bitbake.conf: Add filename and lineno to BB_SIGNATURE_EXCLUDE_FLAGS
This is mainly a performance optimisation. Since we added these flags
to functions, the system spends a lot of time trying to expand these
flags. The values don't really influence checksums and don't need to
be included since if the function content changes, that is will be
detected regardless and is the key detail we care about.
Therefore exclude these from the checksums and gain a signficiant
chunk of parsing speed back.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 14:54:53 +0000 (14:54 +0000)]
toaster: Update for buildstats changes
This updates the toaster class to use the new data format from buildstats.
This does mean it will no longer read IO data from older builds, however
since that data is completely useless anyway, I don't consider that to be
an issue.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 14:54:28 +0000 (14:54 +0000)]
buildstats: Improve to add getrusage data and corrected IO stats
Add IO stats and getrusage() data to the task statistics. We
also drop the CPU percentage calculation since its pretty arbitrary
and not very accurate/useful.
In particular we can now see the user and sys times as well as the
wall clock times.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 14:54:14 +0000 (14:54 +0000)]
buildstats: Separate out the build and task data to allow improvements
The combined build and task data code makes changing things hard, separate
out the functions so that changes can be made to the task data whilst the
build data remains unchanged.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 14:54:01 +0000 (14:54 +0000)]
buildstats: Drop get_bn/set_pn and just use BUILDNAME
The current setting and getting of the "name" to use for buildstats is
convoluted and not particularly interesting. We only need this for the
e.getPkgs()[0] component of the path which is the first target listed
on the commandline. This is pretty arbitrary.
If we drop that piece, we can assume BUILDNAME is common for all events
and simply use this and query it. If BUILDNAME did change, that would
be a bug and it should be fixed elsewhere.
Also take the opportunity to share some common code since the function
now has the eventmask.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 14:53:55 +0000 (14:53 +0000)]
buildstats: Drop disk data from buildstats
The existing diskstats data from buildstats simply isn't useful. It
gives stats on the total IO counts on some random disk within the
system. This means that the count includes data from all other tasks
running at the same time and from any other process running on the
system.
I've been unable to find any use for the data so as a start at
reworking the class, remove the related code.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 12:33:39 +0000 (12:33 +0000)]
nativesdk-buildtools-perl-dummy: Bump PR
Recent changes to this recipe caused automated PR increments
to break, regressing package feeds. The only way to recover
is to bump PR, so do this centrally to fix anyone affected.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 17 Dec 2015 10:55:21 +0000 (10:55 +0000)]
combo-layer: Stop using filterdiff
I ran into an issue where a patch just deleting a single file
within the repository (meta/recipes-devtools/m4/m4/make.patch)
would get skipped by combo-layer.
It turns out this has the patch header (commented to avoid breaking scripts):
and this is classed as > 5 headers in filterdiff. When we piped the path
through filterdiff, the --- line disappears, then the second time we pass
through filterdiff, it shows no lines changed and the patch is assumed
to be empty and skipped.
Changing MAX_HEADERS in filterdiff is one way to fix this, another would
be to grep out "deleted file mode" lines. Instead, we can use new
git syntax to exclude files from the git format-patch instead and avoid
filterdiff entirely.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 15 Dec 2015 16:32:43 +0000 (16:32 +0000)]
meta: more removals of redunant FILES_${PN}-dbg
In some recipes overly-split -dbg packages were merged into PN-dbg. Unless
there's a very good reason, recipes should have a single -dev and -dbg package.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Martin Jansa [Mon, 14 Dec 2015 18:31:46 +0000 (19:31 +0100)]
clutter-gst-3.0: add dependency on libgudev
* it's autodetected from sysroot:
WARN: clutter-gst-3.0: clutter-gst-3.0 rdepends on libcap, but it isn't a build dependency?
WARN: clutter-gst-3.0: clutter-gst-3.0 rdepends on libgudev, but it isn't a build dependency?
WARN: clutter-gst-3.0: clutter-gst-3.0 rdepends on libudev, but it isn't a build dependency?
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Remove configure.patch, it fixed problems with very old versions
of autotools, and testing showed it is no longer needed:
http://git.openembedded.org/openembedded/commit/gmp/gmp-4.1.2/configure.patch?id=43dd0eeb7b7aaf482d3568a0d6a3b99bedbf1cf6
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Drop 0001-hashtable-Remove-duplicate-hashtable_iterator_value-.patch,
the problem has been fixed upstream.
Drop backported 0001-mtd-utils-libfec-use-standard-C-type-instead-of-u_lo.patch
Rebase 0001-Fix-build-with-musl.patch (2 out of 3 fixes have been made upstream)
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
libidn is needed by the new version to compare hostnames.
Openssl compatibility is no longer enabled by default, but is
required by other packages in oe-core.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Maxin B. John [Mon, 14 Dec 2015 14:34:10 +0000 (16:34 +0200)]
libical: Upgrade 1.0.0 -> 1.0.1
* Build system changed to cmake
* Patch build system to not check for perl (we use host
perl which won't be found)
* Patch build system to not break on parallel builds
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Maxin B. John <maxin.john@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Tanu Kaskinen [Sat, 12 Dec 2015 02:23:12 +0000 (04:23 +0200)]
libsndfile1: 1.0.25 -> 1.0.26
Main points from the release announcement:
* Fix for CVE-2014-9496, SD2 buffer read overflow.
* Fix for CVE-2014-9756, file_io.c divide by zero.
* Fix for CVE-2015-7805, AIFF heap write overflow.
* Add support for ALAC encoder in a CAF container.
* Add support for Cart chunks in WAV files.
* Minor bug fixes and improvements.
All patches we had are included in the new release.
Dropped PR from the recipe.
Signed-off-by: Tanu Kaskinen <tanuk@iki.fi> Signed-off-by: Ross Burton <ross.burton@intel.com>
Richard Purdie [Tue, 15 Dec 2015 16:40:36 +0000 (16:40 +0000)]
license: Fix BB_TASKDEPDATA references
We shouldn't try and expand what is a python dict object
and we don't want it to influence the task hashes, task
dependencies are already accounted for.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 15 Dec 2015 16:34:04 +0000 (16:34 +0000)]
coreutils/procps: Revert priority change since coreutils > busybox
In 2ce514a394e1e062172bb097aa92a1792a0f8eaf, "coreutils: fix
ALTERNATIVE_PRIORITY conflict with procps and mktemp" the priority was
lowered from 100 to 30. Unfortunately this lowered coreutils past
busybox which means busybox was ending up in our SDK images. This isn't
what users expect.
Ross Burton [Tue, 15 Dec 2015 16:32:43 +0000 (16:32 +0000)]
meta: more removals of redunant FILES_${PN}-dbg
In some recipes overly-split -dbg packages were merged into PN-dbg. Unless
there's a very good reason, recipes should have a single -dev and -dbg package.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Richard Purdie [Tue, 15 Dec 2015 15:38:54 +0000 (15:38 +0000)]
package: Add auto package splitting of .debug files
Creating FILES_${PN}-dbg is tedious and also pretty pointless. We might as well
assume ".debug" is a special directory name and split into -dbg automatically.
This change does so without changing the rest of the splitting logic too much.
It can be disabled for the cases where we really do want manual control of
the -dbg packages (e.g. qt4) with NOAUTOPACKAGEDEBUG = "1".
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 10 Dec 2015 22:50:56 +0000 (22:50 +0000)]
nopackages: Add class for recipes which don't generate packages
It turns out writing the same list of packaging tasks multiple
times in multiple places is error prone. Move this to a new class
'nopackages", migrate existing users and add glibc-initial and
libgcc-initial since we don't want packages for those recipes.
This means the sstate for those recipes won't be installed, saving
small amounts of build time and bandwidth.
A reference to the old package_write task is also dropped.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 10 Dec 2015 22:49:38 +0000 (22:49 +0000)]
sstate: Ensure populate_lic dependencies are not followed
It was observed that do_fetch dependencies (e.g. subversion-native of tremor)
were being installed even when sstate was available and hence no fetch was
needed. This turned out to be due to the recursive nature of the rootfs
dependencies which include populate_lic.
We can explicitly whitelist these dependencies as being ones we don't
need to follow when installing sstate (the code defaults to being paranoid).
[YOCTO #8746]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 7 Dec 2015 18:28:56 +0000 (18:28 +0000)]
autoconf: Disable macro which causes excessive delays when using dash as sh
At the start of every configure script, the check for solaris 'print' causes
significant problems on a linux machine with dash as /bin/sh since it triggers
the execution of "print" which on some linux systems is a perl script which is
part of mailcap. Worse, this perl script calls "which file" and if successful
ignores the path file was found in and just runs "file" without a path. Each
execution causes PATH to be searched. In something like gettext with multiple
configure scripts, this is worth something like 30,000 syscalls of which 3,000
are execs.
Simply assuming the shell's printf function works cuts out all the fork overhead
and when parallel tasks are running, this overhead appears to be significant.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 7 Dec 2015 18:28:05 +0000 (18:28 +0000)]
automake: Remove delays in configure scripts using automake
By default automake puts "sleep 1" into the start of configure scripts
which adds pointless delays to them. Rather than do this, lets just assume
our systems are sane.
Since this means our patches touch m4 files, we need to stop automake running
autoreconf so we tweak the do_configure to avoid this.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 7 Dec 2015 18:27:29 +0000 (18:27 +0000)]
site/common-linux: Add some macros to avoid sleeps during configure
If you profile gettext do_configure you notice it has some "sleep 1"
type events occurring. This patch ensures we cache the right values to
avoid those pointless delays there and in any other configure scripts
using the same macros.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mike Crowe [Thu, 10 Dec 2015 13:59:46 +0000 (13:59 +0000)]
gstreamer1.0: Split bash completion information into separate package
Many targets won't even have bash installed so the completion
information is completely useless. Follow the lead of other recipes and
put it in a separate -bash-completion package.
Signed-off-by: Mike Crowe <mac@mcrowe.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Andre McCurdy [Wed, 9 Dec 2015 18:52:48 +0000 (10:52 -0800)]
valgrind: import Debian link_tool patch for MIPS
Without this patch, valgrind 3.10.1 fails to run on MIPS:
| root@bcm97425vms:~# valgrind /bin/ls
| valgrind: mmap(0x400000, 700416) failed in UME with error 22 (Invalid argument).
| valgrind: this can be caused by executables with very large text, data or bss segments.
Drop backported CVE fix patches
libtool2.patch has been rebased and renamed to 0001-Fix-libtool-name-in-configure.ac.patch
LICENSE checksum has been updated because more 3rd party attributions have been added to it,
it's otherwise still Apache 2.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
archive.apache.org does not contain current releases, only historical ones,
so upstream checks aren't accurate. It's replaced with official mirrors
containing current versions.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Andre McCurdy [Tue, 8 Dec 2015 21:42:50 +0000 (13:42 -0800)]
dropbear.inc: drop legacy CFLAGS and LD tweaks
The CFLAGS and LD tweaks in dropbear.inc date back to 2005/2006 and
whatever issue they worked around back then seems to have been fixed
in the latest versions of dropbear.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Martin Jansa [Wed, 9 Dec 2015 08:48:13 +0000 (09:48 +0100)]
texinfo: don't create dependency on INHERIT variable
* we don't want the do_package signature depending on INHERIT variable
* e.g. just adding the own-mirrors causes texinfo to rebuild:
# bitbake-diffsigs BUILD/sstate-diff/*/*/texinfo/*do_package.sig*
basehash changed from 015df2fd8e396cc1e15622dbac843301 to 9f1d06c4f238c70a99ccb6d8da348b6a
Variable INHERIT value changed from
' rm_work blacklist blacklist report-error ${PACKAGE_CLASSES} ${USER_CLASSES} ${INHERIT_DISTRO} ${INHERIT_BLACKLIST} sanity'
to
' rm_work own-mirrors blacklist blacklist report-error ${PACKAGE_CLASSES} ${USER_CLASSES} ${INHERIT_DISTRO} ${INHERIT_BLACKLIST} sanity'
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Mon, 30 Nov 2015 11:31:14 +0000 (11:31 +0000)]
libsdl: expand PACKAGECONFIG and enable native builds
Use PACKAGECONFIG instead of using logic in DEPENDS and EXTRA_OECONF, adding new
options for PulseAudio, tslib, DirectFB, OpenGL and X11. Pass
--disable-x11-shared so that it links to the X libraries instead of using
dlopen().
Disable tslib by default as the kernel event input subsystem is generally used.
SDL's OpenGL support requires X11 so check for both x11 and opengl, and merge
the dependencies.
Finally enable native builds, with a minimal PACKAGECONFIG that will build from
oe-core for native and nativesdk.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Juro Bystricky [Thu, 3 Dec 2015 23:09:55 +0000 (15:09 -0800)]
buildtools-tarball.bb: 32bit tools need pseudo 32bit library
"pseudo" does not build 32 bit library by default on 64 bit host, but the
32 bit library is needed when using 32 bit development tools (buildtools-tarball)
running on a 64 bit host. Building of the library can be forced if the
environment variable NO32LIBS is set to "0".
Hence for 32 bit buildtools we export the environment variable NO32LIBS="0"
and NO32LIBS is also added into BB_ENV_EXTRAWHITE (if not already there).
[YOCTO#8581]
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Wed, 2 Dec 2015 11:39:27 +0000 (03:39 -0800)]
rpm: fix file conflicts for MIPS64 N32
The following error can occur:
smart install libc6-2.22-r0.1@lib32_mips32r2octeon3 libc6-dbg-2.22-r0.1@lib32_mips32r2octeon3
error: file /sbin/ldconfig conflicts between attempted installs of libc6-2.22-r0.1.lib32_mips32r2octeon3 and libc6-2.22-r0.1.octeon3_n32
error: file /sbin/.debug/ldconfig conflicts between attempted installs of libc6-dbg-2.22-r0.1.lib32_mips32r2octeon3 and libc6-dbg-2.22-r0.1.octeon3_n32
This was because:
transactions_color = 001 (ELF32) & 010 (ELF64) & 100 (ELF32 N32 MIPS64)
FColor = Current file color (001) & transaction_color (111)
oFcolor = Previous file color (100) & transaction_color (111)
There are two places where the conflict comparisons occur. In both places
the 'else' clause was too restrictive (opposite of the 'positive' clause).
This caused the system to only permit a binary comparison - "new preferred" or
"old preferred". It did not permissing "neither preferred". By removing the
else comparison the system will now perform a 'last-in-wins' resolution when
"neither is preferred".
Note, if _transaction_color is 3, MIPS64 N32 will be skipped (pretend as
installed).
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Hatle [Wed, 2 Dec 2015 11:39:26 +0000 (03:39 -0800)]
rpm: Enable MIPS64 N32 transactions
In a multilib install RPM will break a conflict using the transaction color.
However, by default MIPS64 N32 transactions are not enabled.
This change always enables them. This is effectively a no-op on non-MIPS
systems, and required if we need to support N32 on MIPS.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
