Cristian Iorga [Fri, 30 May 2014 17:11:11 +0000 (20:11 +0300)]
v86d: fix for SeaBIOS VGA BIOS compatibility
This fixes a compatibility issue between v86d
and SeaBIOS VGA BIOS where the leal instruction
is not decoded properly. Read the
http://thread.gmane.org/gmane.comp.emulators.qemu/271806
thread for more details.
Patch taken from upstream.
Signed-off-by: Cristian Iorga <cristian.iorga@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Sat, 31 May 2014 00:11:55 +0000 (01:11 +0100)]
socat: upgrade to 1.7.2.4
socat manually maintains config.h.in instead of using autoheader, so exclude
autoheader from autoreconf and remove all patches that are intended to enable
use of autoheader.
The license checksum changed, because the company address changed.
Based on a upgrade by Hongxu Jia <hongxu.jia@windriver.com>, with cleanup after
noticing that config.h.in was hand-maintained.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Saul Wold [Thu, 29 May 2014 22:00:54 +0000 (15:00 -0700)]
util-linux: add sulogin to alternatives list
sulogin is provided by both busybox in /sbin/sulogin and util-linux provides one
in /usr/sbin/sulogin, so move util-linux's to sbin and setup ALTERNATIVE_LINK.
[YOCTO #6384]
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dan McGregor [Fri, 30 May 2014 21:39:48 +0000 (15:39 -0600)]
gzip: Avoid host contamination during build
On systems with /bin and /usr/bin merged the gzip recipe may find grep
in /usr/bin, while OE puts it in /bin. Force the recipe to find grep in
the correct place.
Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 30 May 2014 12:32:54 +0000 (13:32 +0100)]
gcc: Clean up configure_prepend and fix for mingw
The do_configure_prepend was duplicated in gcc-4.X.inc and
gcc-configure-common.inc leading to confusion when reading the resulting
do_configure task where the file was processed twice.
The only difference was the removal of the include line for gcc 4.8/4.9.
On mingw were were seeing two issues, firstly that the if statements meant
the values we wanted weren't being set, the second that the include
paths were still wrong as there was no header path set.
To fix the first issue, the #ifdef conditionals were removed, we want
to set these things unconditionally. The second issue is addressed by
setting the NATIVE_SYSTEM_HEADER_DIR variable here (it was already
set in t-oe).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 30 May 2014 12:31:47 +0000 (13:31 +0100)]
classextend: Fix crosssdk remapping for multilib
Multilib builds only require one crosssdk toolchain. We therefore shouldn't
be remapping crosssdk names. This resolves build failures looking for
weird multilib crosssdk toolchains.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
combo-layer: fix revlist taking into account file_filter
If file_filter is set, git format-patch takes account but git rev-list
does not. So revlist is going to get with wrong revisions. And last_revision
will be updated with wrong revision. The next time that user run
combo-layer it complain about applying patches.
So ensure that 'git rev-list' are using file_filter as 'git format-patch'.
Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Thu, 29 May 2014 01:50:53 +0000 (09:50 +0800)]
update-rc.d: fix to handle priority numbers correctly
The update-rc.d treated priority numbers begin with '0' as octal
numbers. This led to problems of update-rc.d being unable to handle
priorities like '08' or '09' correctly.
This patch fixes the above problem.
[YOCTO #6376]
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 29 May 2014 07:01:03 +0000 (00:01 -0700)]
multilib_header: Do not install mutlilib headers for musl
musl is not multilib and this creates trouble. eg. when
util-linux probes for ncurses it does not find it because
ncurses has installed the multilibbed header and this
header includes bits/wordsize.h and this header does not
exist on musl systems. If and when musl adds multilib
support we will revisit it.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 29 May 2014 07:01:01 +0000 (00:01 -0700)]
siteinfo, insane: Recognize musl specific triplets
We will use '-musl' to identify musl based systems
this patch lays the foundation for recognising those
and map them to internal variable representations
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch adds toaster tests using the oe-selftest infrastructure.
You need to have builds done - the tests will verify data integrity
after the toaster collection phase.
Once you have your toaster builds done, to run the automated backend
tests via oe-selftest do the followings:
1. Update builddir/conf/bblayers.conf to contain the meta-selftest
layer
2. From the builddir run:
'oe-selftest toaster'
or if you just want to run a single test:
'oe-selftest toaster.Toaster_DB_Tests.testname'
This first part adds the meta/lib/oeqa toaster file.
Signed-off-by: Ionut Chisanovici <ionutx.chisanovici@intel.com> Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tudor Florea [Wed, 28 May 2014 16:59:54 +0000 (18:59 +0200)]
ethtool: use serial-tests config needed by ptest.
buildtest-TESTS and runtest-TESTS targets are required by ptest.
In order to have those targets in automake 1.13.4 serial-tests
should be specified since parallel-tests is assumed by default
and serial-tests is optional.
Richard Purdie [Wed, 28 May 2014 16:38:36 +0000 (17:38 +0100)]
sstatesig: Move saferecipedeps handling to be earlier
We want to use the saferecipedeps handling code to allow gcc-cross-* to
work on multiple different tunes. Its currently in target only code
so it needs to be earlier to allow it to work on native-> target
dependencies.
This change has no effect on existing uses but makes gcc-cross become
shared as desired.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Roy Li [Thu, 22 May 2014 06:42:51 +0000 (14:42 +0800)]
openssl: add openssl-CVE-2010-5298.patch SRC_URI
make openssl-CVE-2010-5298.patch truely work
Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 28 May 2014 11:56:25 +0000 (12:56 +0100)]
readline: use upstream's aclocal.m4 as acinclude.m4
Instead of shipping a fork of the upstream aclocal.m4, simply rename it to
acinclude.m4 at configure time. We don't need the fork now that autoheader is
excluded.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 28 May 2014 11:56:24 +0000 (12:56 +0100)]
readline: exclude autoheader from autoreconf
readline ships a hand-maintained config.h, instead of letting autoheader
generate one from configure.ac. The required arguments to AC_DEFINE are not in
configure.ac so autoheader will produce warnings and the generated code will not
behave as expected.
Solve this by excluding autoheader from autoreconf, so the upstream config.h.in
is used.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Tue, 27 May 2014 19:58:59 +0000 (12:58 -0700)]
i2c-tools: Add i2c-tools to the core
i2c-tools has been sitting outside of oe-core for long enough now. It is
a required tool for board validation, and many people are pulling it
into their builds and their own layers. Let's add it to the core.
This patch includes the i2c-tools recipe from meta-oe as of:
Darren Hart [Fri, 23 May 2014 16:57:59 +0000 (09:57 -0700)]
linux-yocto-dev: Dynamic SRCREV update
The current implementation would result in the default SRCREVs being
used by the fetcher, even though the anonymous python would update them
to AUTOREV. This appears to be something to do with early parsing
bitbake black magic.
This patch ensures the default is never assigned if we are actually
building the recipe by using a function to assign it in the first place.
The USE_DEFAULT* variables are removed as they are not necessary to
allow for overriding the SRCREVs.
The anonymous python parse check is moved closer to the top of the
recipe to be a bit more logically representative of its intended
purpose.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 26 May 2014 17:40:09 +0000 (13:40 -0400)]
beaglebone: enable the nowayout option for the watchdog
Bumping the meta SRCREV for the following fix:
[
The default watchdog behaviour is to stop the timer if the process
managing it closes the file /dev/watchdog. The system would not reboot
if watchdog daemon crashes due to a bug in it or get killed by other
malicious code. So we prefer to enable nowayout option for the
watchdong. With this enabled, there is no way of disabling the watchdog
once it has been started. This option is also enabled in the predecessor
of this BSP (beagleboard)
]
[YOCTO: 3937]
Signed-off-by: Kevin Hao <kexin.hao@windriver.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 22 May 2014 09:49:47 +0000 (10:49 +0100)]
libgpg-error: Extend pkgconfig support to m4 macros
Whilst there is currently .pc file pkgconfig support, it was unused by the
m4 macros. This extends the support so they're used instead of the -config
scripts.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 27 May 2014 15:09:16 +0000 (16:09 +0100)]
devshell: Add interactive python shell
Being able to interact with the python context in the Bitbake task execution
environment has long been desireable. This patch introduces such a
mechanism. Executing "bitbake X -c devpyshell" will open a terminal connected
to a python interactive interpretor in the task context so for example you can
run commands like "d.getVar('WORKDIR')"
This version now includes readline support for command history and various other
bug fixes such as exiting cleanly compared to previous versions.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Seebach [Wed, 28 May 2014 00:14:46 +0000 (19:14 -0500)]
pseudo: Honor umask again
The fchmodat-permissions patch was fine for the fchmod case, but
had the unintended side effect of disregarding umask settings for
open, mknod, mkdir, and their close relatives. Start tracking umask
and masking the umask bits out where appropriate.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mike Crowe [Tue, 27 May 2014 16:28:43 +0000 (17:28 +0100)]
cmake: Avoid accidentally including libacl.h
The cmake recipe doesn't depend on libacl yet cmake will detect libacl.h
and use it by default. This risks build failures if libacl.h is unstaged
during the build and it also means that the build cmake will sometimes
support ACLs and sometimes not.
This can be avoided by setting ENABLE_ACL=0 but until the fix for
http://cmake.org/Bug/view.php?id=14866 is released we also need to set
HAVE_ACL_LIBACL_H=0.
Signed-off-by: Mike Crowe <mac@mcrowe.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Tue, 27 May 2014 19:47:45 +0000 (12:47 -0700)]
ddimage: Support Mac OS
Update the ddimage script to allow it to work on Mac OS too. The biggest
difference is sysfs vs diskutil and in the syntax of the stat command
between Mac OS and Linux, unfortunately. Workarounds using ls, cut, and
columns got really fragile really quickly. Relying on stat and switching
on uname seemed the more robust solution.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Cc: Koen Kooi <koen@dominion.thruhere.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tim Orling [Thu, 22 May 2014 16:02:19 +0000 (09:02 -0700)]
wayland: upgrade to 1.5.0
* update disable-macro-checks-not-used-for-scanner.patch
= trivial change to non-patched text (+ posix_fallocate)
* drop just-scanner.patch, no longer needed
Signed-off-by: Tim Orling <TicoTimo@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Ross Burton [Thu, 22 May 2014 14:28:32 +0000 (15:28 +0100)]
perl: stop perl-modules recommending perl-ptest
Change the logic that generates the perl-modules recommends to be an include
filter instead of an exclude filter, so that new sub-packages don't become
dependants of perl-modules (such as perl-ptest).
[ YOCTO #6203 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Cristian Iorga [Thu, 22 May 2014 09:33:51 +0000 (12:33 +0300)]
bluez5: upgrade to 5.19
- Fixes to OBEX, AVRCP browsing, HID over GATT
and handling of device unpaired events for dual-mode devices.
- New features: user space based HID host implementation (for BR/EDR).
yanjun.zhu [Tue, 20 May 2014 01:27:47 +0000 (09:27 +0800)]
perl: fix for CVE-2010-4777
The Perl_reg_numbered_buff_fetch function in Perl 5.10.0, 5.12.0,
5.14.0, and other versions, when running with debugging enabled,
allows context-dependent attackers to cause a denial of service
(assertion failure and application exit) via crafted input that
is not properly handled when using certain regular expressions,
as demonstrated by causing SpamAssassin and OCSInventory to
crash.
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2010-4777 Signed-off-by: yanjun.zhu <yanjun.zhu@windriver.com> Signed-off-by: Roy Li <rongqing.li@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Peter Seebach [Wed, 21 May 2014 23:12:33 +0000 (18:12 -0500)]
pseudo: handle fchmodat better, mask out unwanted write bits
It turns out that pseudo's decision not to report errors from
the host system's fchmodat() can break GNU tar in a very strange
way, resulting in directories being mode 0700 instead of whatever
they should have been.
Additionally, it turns out that if you make directories in your
rootfs mode 777, that results in the local copies being mode 777,
which could allow a hypothetical attacker with access to the
machine to add files to your rootfs image. We should mask out
the 022 bits when making actual mode changes in the rootfs.
This patch represents a backport to the 1.5.1 branch of three
patches from the 1.6 branch, because it took a couple of tries
to get this quite right.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Robert Yang [Tue, 20 May 2014 02:49:47 +0000 (10:49 +0800)]
initramfs-live-install: avoid using grub.d/40_custom
We have this in recipes-bsp/grub/grub/40_custom:
[snip]
menuentry "Linux" {
set root=(hd0,1)
linux /vmlinuz root=__ROOTFS__ rw __CONSOLE__ __VIDEO_MODE__ __VGA_MODE__ quiet
}
[snip]
These lines are only for initrdscripts/files/init-install.sh, the side
effect is that it would make the target's grub-mkconfig doesn't work
well since the 40_custom will be installed to /etc/grub.d/40_custom, the
grub-mkconfig will run the 40_custom, and there will always be a
'menuentry "Linux"' menu in grub.cfg no matter it is valid or not, we
can do this in init-install.sh rather than grub to fix the problem,
which is also much simpler.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Richard Purdie [Wed, 21 May 2014 16:50:17 +0000 (17:50 +0100)]
taglib: Force a disable of the floating dependency on boost
taglib appears to depend on boost if it finds it in the sysroot. Force
it not to do this. Someone with better cmake skills may be able to
do this in a neater way.
Yue Tao [Mon, 28 Apr 2014 03:12:34 +0000 (11:12 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0849
The roq_decode_init function in libavcodec/roqvideodec.c in FFmpeg
before 1.1 allows remote attackers to have an unspecified impact via a
crafted (1) width or (2) height dimension that is not a multiple of
sixteen in id RoQ video data.
Yue Tao [Sun, 27 Apr 2014 13:02:51 +0000 (21:02 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0850
The decode_slice_header function in libavcodec/h264.c in FFmpeg before
1.1 allows remote attackers to have an unspecified impact via crafted
H.264 data, which triggers an out-of-bounds array access.
Yue Tao [Sun, 27 Apr 2014 12:24:18 +0000 (20:24 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0856
The lpc_prediction function in libavcodec/alac.c in FFmpeg before 1.1
allows remote attackers to have an unspecified impact via crafted Apple
Lossless Audio Codec (ALAC) data, related to a large nb_samples value.
Yue Tao [Sun, 27 Apr 2014 12:04:19 +0000 (20:04 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0854
The mjpeg_decode_scan_progressive_ac function in libavcodec/mjpegdec.c
in FFmpeg before 1.1 allows remote attackers to have an unspecified
impact via crafted MJPEG data.
Yue Tao [Sun, 27 Apr 2014 11:51:12 +0000 (19:51 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0851
The decode_frame function in libavcodec/eamad.c in FFmpeg before 1.1
allows remote attackers to have an unspecified impact via crafted
Electronic Arts Madcow video data, which triggers an out-of-bounds array
access.
Yue Tao [Sun, 27 Apr 2014 11:44:28 +0000 (19:44 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0858
The atrac3_decode_init function in libavcodec/atrac3.c in FFmpeg before
1.0.4 allows remote attackers to have an unspecified impact via ATRAC3
data with the joint stereo coding mode set and fewer than two channels.
Yue Tao [Sun, 27 Apr 2014 07:37:10 +0000 (15:37 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0852
The parse_picture_segment function in libavcodec/pgssubdec.c in FFmpeg
before 1.1 allows remote attackers to have an unspecified impact via
crafted RLE data, which triggers an out-of-bounds array access.
Yue Tao [Sun, 27 Apr 2014 07:10:15 +0000 (15:10 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0845
libavcodec/alsdec.c in FFmpeg before 1.0.4 allows remote attackers to
have an unspecified impact via a crafted block length, which triggers an
out-of-bounds write.
Yue Tao [Sun, 27 Apr 2014 03:56:19 +0000 (11:56 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0868
libavcodec/huffyuvdec.c in FFmpeg before 1.1.2 allows remote attackers
to have an unspecified impact via crafted Huffyuv data, related to an
out-of-bounds write and (1) unchecked return codes from the init_vlc
function and (2) len==0 cases.
Yue Tao [Fri, 25 Apr 2014 08:26:00 +0000 (16:26 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2014-2099
The msrle_decode_frame function in libavcodec/msrle.c in FFmpeg before
2.1.4 does not properly calculate line sizes, which allows remote
attackers to cause a denial of service (out-of-bounds array access) or
possibly have unspecified other impact via crafted Microsoft RLE video
data.
Yue Tao [Mon, 14 Apr 2014 10:58:29 +0000 (18:58 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2013-0865
The vqa_decode_chunk function in libavcodec/vqavideo.c in FFmpeg before
1.0.4 and 1.1.x before 1.1.2 allows remote attackers to have an
unspecified impact via a large (1) cbp0 or (2) cbpz chunk in Westwood
Studios VQA Video file, which triggers an out-of-bounds write.
Yue Tao [Mon, 14 Apr 2014 10:38:34 +0000 (18:38 +0800)]
gst-ffmpeg: fix for Security Advisory CVE-2014-2263
The mpegts_write_pmt function in the MPEG2 transport stream (aka DVB)
muxer (libavformat/mpegtsenc.c) in FFmpeg, possibly 2.1 and earlier,
allows remote attackers to have unspecified impact and vectors, which
trigger an out-of-bounds write.
wic: add support to look in all layers and get plugins
Plugins are looked in 'scripts/lib/mic/plugins/[type]/' directory on all
BBLAYERS variable returned by bitbake environment. If found, it will
be load at runtime.
The user could create your own plugin and keep it inside its layers. For
now the path must be <layer-dir>/scripts/lib/mic/plugins/[type]/. Where
'type' could be 'imager' or 'source'.
Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Chen Qi [Mon, 19 May 2014 08:03:28 +0000 (16:03 +0800)]
runqemu-internal: add "console=ttyS0" to ramfs image kernel parameters
We need this kernel command parameter so that when we start a ramfs
image, we can actually get some output. Although we can make this
happen by specifying the 'bootparams' for the 'runqemu' command, it's
better to make this the default behaviour.
