Paul Eggleton [Tue, 26 Mar 2013 12:43:05 +0000 (12:43 +0000)]
qemuimagetest: collect and print runqemu output
If runqemu (or qemu itself) fails we need to know why, so tee out to a
log file and print it when we can't find the qemu process or determine
its IP address.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Tue, 26 Mar 2013 12:43:04 +0000 (12:43 +0000)]
qemuimagetest: fix erroneous ps errors when qemu couldn't be started
The helper script was printing an error to stdout when it couldn't find
any qemu child processes; output this error to stderr instead and
redirect stderr to /dev/null when running from qemuimage-testlib so that
QEMUPID is actually blank if there are no qemu instances found.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 21 Mar 2013 14:29:57 +0000 (14:29 +0000)]
scripts/qemu-testlib: Add more debugging information
This extra information should allow better forensics if the sanity tests
fail as they're currently doing occasionaly on the autobuilder for unknown
reasons.
The patch also tightens up certain checks to remove pointless noise and
error output from the logs.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 20 Mar 2013 22:58:30 +0000 (22:58 +0000)]
scripts/runqemu-internal: Fix lock races
There are two problems here. Firstly the grep command is unanchored so
pid 345 will match against 12345 and so on.
The second issue is that there are several context switched between attempting
the lock and then writing the pid to it.
Between the two issues, there were issues appearing on the autobuilder due
to these conflicts. This patch replaces the mechanism with flock on fd 8
which should be a safer mechanism to use.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 20 Mar 2013 11:48:53 +0000 (11:48 +0000)]
runqemu: Improve error handling/exit codes
runqemu-internal is sourced so should be returning with an error code in
case of errors. runqemu needs to deal with this.
This patch fixes up the various error paths so we're consistent and get
a sane exit status for runqemu which helps a lot in its use in the qemu
runtime testing on the autobuilder.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 20 Mar 2013 10:58:24 +0000 (10:58 +0000)]
qemuimage-testlib: Add extra debugging and sanity check
Check for a zero IP address since its clearly incorrect if that value
is found. Also add debugging for cases where we can't find the qemu
process. A process listing is handy to help understand what the problem
might be.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 20 Mar 2013 09:27:39 +0000 (09:27 +0000)]
qemuimage-testlib: Add delay to work around races in qemu startup
Qemu changes pid when starting up. On a loaded machine, this can result
in the incorrect pid being returned. Since qemu will take a few seconds to
boot anyway, we might as well delay a short while and allow things to settle
which should fix various race issues being seen on the autobuilder.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 19 Mar 2013 20:54:33 +0000 (20:54 +0000)]
scripts/qemutestlib: Add better process debugging and fix process group issue
In single testing with a shutdown scenario the processes are cleaned up correctly
but the manual cleanup fall back used for a minimal image do not work properly.
This patch fixes the kill commands to revert to non-process groups, fixing
the hung process issues that were occuring.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 19 Mar 2013 11:44:27 +0000 (11:44 +0000)]
qemu-testlib: Add python helper and simplify shell
The current code has a race since it greps for *any* qemu process
running, even if it isn't the one we started. This leads to some sanity
tests potentially failing on machines where multiple sets of sanity tests
are running.
To resovle this and some other ugly code issues, add a python script
to accurately walk the process tree and find the qemu process. We can
then replace all the shell functions attempting this which happen to
work in many cases but not all.
Also clean up some of the error handling so its more legible.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Thu, 21 Feb 2013 17:28:02 +0000 (09:28 -0800)]
install: Look for grub2 files on the initramfs, not rootfs
Fixes [YOCTO #3870] atom-pc - cannot boot image on netbook after install
The problem here is that grub2 is installed but a grub 1 menu.lst is
created at install time. At boot, grub2 doesn't find a grub.cfg file and
drops to the grub shell.
This happens because the installer is looking for 40_custom (a grub2
file) on the rootfs, but grub2 isn't installed on the rootfs. It exists
in the initramfs. Patching the installer to look on the initramfs
resolves the problem.
Note that the problem may have occurred if grub2 used to be installed on
the rootfs but was later removed. In any case, the installer is HORRIBLE
and really needs to be completely redesigned as part of the deployment
effort. For now, this should get the live image installer limping along
again.
Tested on a Toshiba NB-305.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Cc: alexandru.c.georgescu@intel.com Cc: sgw@linux.intel.com Cc: ross.burton@intel.com Cc: richard.purdie@intel.com Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Roy.Li [Sat, 2 Feb 2013 08:18:22 +0000 (16:18 +0800)]
tcp-wrappers: remove size_t.patch
1. it introduces bug in 64bit big endian process with __GLIBC__, At that
condition, size_t is 8byte, and the third parameter of getpeername is
socklen_t which is 4 byte. As a result, getpeername sees third parameter
is always 0, and can not return right value.
The similar program is below, the output is 0, not 9 on PPC64 cpu
main()
{
long aa=9;
printf("%d \n", *((int *)&aa));
}
2. The correct fix is to change getpeername/getsockopt/recvfrom.. last
parameter type from int to socklen_t, but to simplify, we can remove
size_t.patch, since the size of int is same as socklen_t in 32bit/64bit
cpu. and size_t.patch only change three places, there are other places
which uses int, and work well.
2. Fedora, redhat el4 do not use this patch, but Debian uses it, does not
find why this patch is written, maybe it is gcc legency issue which does
not exist.
Hongxu Jia [Thu, 21 Feb 2013 06:22:16 +0000 (14:22 +0800)]
base-passwd.preinst:fix creating passwd and group error
Create files `passwd' and `group' in `$D${sysconfdir}', if `$D${sysconfdir}'
does not exist, there is an error:
...
cannot create $D${sysconfdir}/passwd: Directory nonexistent
cannot create $D${sysconfdir}/group: Directory nonexistent
...
Otavio Salvador [Fri, 8 Feb 2013 14:52:20 +0000 (14:52 +0000)]
tzdata: Simplify code removing not used cases
We shouldn't have an use-case where we'd use 'FUBAR' timezone so
instead of adding postinst handling for this use case we handle it at
install time and keep the Universal as fallback if user did something
wrong.
This also ensure the /etc/localtime file is kept as a symbolic link.
This will make timezone not available when /usr is in separated
partition (and not mounted) however the applications ought to fallback
to GMT timezone in this case and when /usr is made availble timezone
will work fine.
Change-Id: I9a4f05db7a0bdc06511deb5693d1d16569d2fc63 Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Fri, 8 Feb 2013 14:52:19 +0000 (14:52 +0000)]
tzdata: We shouldn't override the localtime if it is valid
The code where mistakenly replacing the localtime file setting so we
end with a copy of file instead of a symbolic link. This fixes it so
now, we'll only do that in case the link is pointing to invalid data.
Change-Id: I16dfa5ea4f293c48bb396f4e23a2ea53e6c9e745 Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The utf-16 decoder in Python 3.1 through 3.3 does not update the
aligned_end variable after calling the unicode_decode_call_errorhandler
function, which allows remote attackers to obtain sensitive information
(process memory) or cause a denial of service (memory corruption and crash)
via unspecified vectors.
Richard Purdie [Thu, 25 Oct 2012 11:19:06 +0000 (11:19 +0000)]
cogl/clutter: Explicitly depend on libdrm for GLX
cogl and clutter explicitly rely on libdrm being present when using the glx
backend. If its not listed in DEPENDS and an alternative to mesa is used, it
may not actually be present. This patch ensures it is and fixes a build
race condition which could see dependencies like clutter-box2d failing to
compile due to missing pkgconfig dependencies.
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>
Paul Eggleton [Fri, 25 Jan 2013 15:17:06 +0000 (15:17 +0000)]
valgrind: explicitly disable MPI2 support
We don't have support for this in OE-Core, so ensure we don't pick this
up from the build host if e.g. openmpi development files happen to be
installed there.
Michael Halstead [Wed, 13 Feb 2013 16:31:03 +0000 (08:31 -0800)]
scripts/qemuimage-testlib: Use wide option to ps calls
Forcing ps to display unlimited column width allows the qemu IP address to
be discovered during sanity testing when the command line is extremely long.
This seems to fix the sanity testing problem on AB05 which was recently updated
to OpenSUSE 12.2. I'm not sure what about qemu or process listing is different
on that distribution but this simpile fix seems to work and my help on other
distro's as well.
Signed-off-by: Michael Halstead <michael@yoctoproject.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Laurentiu Palcu [Mon, 7 Jan 2013 13:07:04 +0000 (15:07 +0200)]
relocate_sdk.py: new interpreter string was not '\0' terminated
The problem: SDK binaries were not properly relocated when the SDK
was installed into a path that had a length less than the default one.
Apparently, there were two problems here: the padding was done wrong
(the size of one program header table entry was used instead of the
program section size) and the new padded string was not used at all.
[YOCTO #3655]
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tom Zanussi [Tue, 5 Feb 2013 14:26:36 +0000 (08:26 -0600)]
crosstap: handle hyphenated x86_64 target arch
systemtap_target_arch() should also translate x86-64 (hyphenated) into
x86_64 for the -a param. Failing to do that causes systemtap to see
an architecture mismatch and create a cloned session with a bogusly
synthesized build directory path, and fails to compile the probe.
Fixes [YOCTO #3756]
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Wed, 6 Feb 2013 14:07:26 +0000 (14:07 +0000)]
classes/rootfs_rpm: fix missed complementary packages with complex arch names
If the package architecture name is complex (e.g. with the meta-yocto-bsp
beagleboard machine and Poky, the architecture-specific package name is
"armv7a-vfp-neon"), rpm reports architecture names that contain
underscores instead of dashes, which when passed to oe-pkgdata-util
during complementary package processing did not match a valid pkginfo
path. Replacing the underscores with dashes again fixes the issue.
Note that this bug affects the danny branch and not master since the
substitution is already performed there in a different manner.
Fixes [YOCTO #3792]
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Ross Burton [Tue, 5 Feb 2013 15:13:17 +0000 (15:13 +0000)]
gst-meta-base: pull in X11 plugins when using X11
If the X11 distro feature is present, make gst-meta-base depend on
gst-meta-x11-base. Also make -x11-base RRECOMMEND xvimagesink for video
playback performance.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Martin Jansa [Fri, 23 Nov 2012 00:47:16 +0000 (01:47 +0100)]
xserver-xorg: disable dri2 too when building without glx PACKAGECONFIG
* it was enabled when dri2proto was built before xserver-xorg
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Tue, 30 Oct 2012 11:30:26 +0000 (11:30 +0000)]
gnomebase: depend on gnome-common-native
gnome-common is a build-only dependency so we should depend on the native
variant. This also resolves an (incorrect) GPLv3 license issue in gnome-common
at build-time.
This will also remove the pointless gnome-common-dev RRECOMMENDS in any -dev
package that uses gnomebase.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Maxin B. John [Fri, 14 Dec 2012 12:00:28 +0000 (13:00 +0100)]
Update the SRC_URI in systemtap recipe
The SRC_URI in systemtap recipe uses 'sources.redhat.com' which
redirects to 'sourceware.org'. This causes random fetch failures.
Updating the recipe to use the direct link.
Signed-off-by: Maxin B. John <Maxin.John@enea.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Noor Ahsan [Thu, 20 Dec 2012 11:37:24 +0000 (16:37 +0500)]
freetype: Update EXTRA_OECONF to use host gcc.
* It uses host gcc in its configure script. When IA32
toolchain is installed it starts using its binaries instead
of native gcc. Modified EXTRA_OECONF so that host gcc is used.
Signed-off-by: Noor Ahsa <noor_ahsan@mentor.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Martin Jansa [Wed, 26 Dec 2012 11:52:52 +0000 (12:52 +0100)]
arch-armv4.inc: add --fix-v4bx to TARGET_LD_KERNEL_ARCH only for armv4 and strongarm1100
* without this patch it does apply --fix-v4bx not only to armv4, but
also all higher (because they also have armv4 in TUNE_FEATURES)
* it causes SIGILL on armv4t
http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-November/042298.html
* someone please test on armv4 device (I tested only bitbake -e output
that it's correctly applied with DEFAULTTUNE == armv4
* maybe we can should fix this in binutils instead (both 2.22 and 2.23
are affected)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Li Wang [Thu, 13 Dec 2012 05:54:21 +0000 (13:54 +0800)]
libxml2 CVE-2012-2871
the patch come from:
http://src.chromium.org/viewvc/chrome/trunk/src/third_party/libxml/src \
/include/libxml/tree.h?r1=56276&r2=149930
libxml2 2.9.0-rc1 and earlier, as used in Google Chrome before 21.0.1180.89,
does not properly support a cast of an unspecified variable during handling
of XSL transforms, which allows remote attackers to cause a denial of service
or possibly have unknown other impact via a crafted document, related to the
_xmlNs data structure in include/libxml/tree.h.
Integer overflow in the queue_init function in unsquashfs.c in
unsquashfs in Squashfs 4.2 and earlier allows remote attackers
to execute arbitrary code via a crafted block_log field in the
superblock of a .sqsh file, leading to a heap-based buffer overflow.
The node name (formerly RsvgNode:type) cannot be used to infer
the sub-type of RsvgNode that we're dealing with, since for unknown
elements we put type = node-name. This lead to a (potentially exploitable)
crash e.g. when the element name started with "fe" which tricked
the old code into considering it as a RsvgFilterPrimitive.
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2011-3146
Signed-off-by: Li Wang <li.wang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Li Wang [Mon, 7 Jan 2013 11:10:01 +0000 (11:10 +0000)]
cups CVE-2011-3170
the patch come from:
http://cups.org/strfiles/3914/str3914.patch
The gif_read_lzw function in filter/image-gif.c in CUPS 1.4.8 and
earlier does not properly handle the first code word in an LZW stream,
which allows remote attackers to trigger a heap-based buffer overflow,
and possibly execute arbitrary code, via a crafted stream, a different
vulnerability than CVE-2011-2896.
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2011-3170
Signed-off-by: Li Wang <li.wang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Li Wang [Mon, 7 Jan 2013 11:10:00 +0000 (11:10 +0000)]
cups - CVE-2011-2896
the patch come from:
http://cups.org/strfiles/3867/str3867.patch
The LZW decompressor in the LWZReadByte function in giftoppm.c
in the David Koblas GIF decoder in PBMPLUS, as used in the
gif_read_lzw function in filter/image-gif.c in CUPS before 1.4.7,
the LZWReadByte function in plug-ins/common/file-gif-load.c
in GIMP 2.6.11 and earlier, the LZWReadByte function in img/gifread.c
in XPCE in SWI-Prolog 5.10.4 and earlier, and other products,
does not properly handle code words that are absent from the
decompression table when encountered, which allows remote attackers to
trigger an infinite loop or a heap-based buffer overflow, and possibly
execute arbitrary code, via a crafted compressed stream, a related
issue to CVE-2006-1168 and CVE-2011-2895.
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2011-2896
Signed-off-by: Li Wang <li.wang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Li Wang <li.wang@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need to fix the "hosts: files dns mdns4" nsswitch.conf line
because for a .local lookup it does a DNS lookup first which will fail.
The recommended solution is:
hosts: files mdns4_minimal [NOTFOUND=return] dns mdns4
Darren Hart [Mon, 7 Jan 2013 11:09:56 +0000 (11:09 +0000)]
bootimg: Use FAT 32 for images larger than 512MB
Fixes [YOCTO #2138]
Commit 217584211625b1c496fe5b78aa4765ccf605d2b9 dropped the forced use
of FAT32 for the hddimg generation as it broke with very small images
(< 32MB). Unfortunately, left to its own devices, mkdosfs appears to select
FAT16 even for very large images, resulting in 2.2GB images being
generated as FAT16:
The result was a runtime boot error from SYSLINUX and a failure to boot
live images greater than 1GB in size.
While strictly speaking it is the cluster count that determines which
FAT size is used, that calculation requires more information than we
have readily available (such as sectors per cluster). If we let mkdosfs
determine sectors per cluster and just set a sane threshold above which
FAT32 is used, we get correct bootable images. With this patch the 2.2GB
core-image-lsb-sdk uses FAT32 and the 21 MB core-image-minimal uses
FAT16, and both boot in qemu successfully:
I have tested and booted core-image-minimal and core-image-lsb-sdk for
atom-pc with qemu-system-i386 using this patch.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Cc: Steve Sakoman <steve@sakoman.com> Cc: Joshua Immanuel <josh@hipro.co.in> Cc: Przemek Czesnowicz <przemyslawx.czesnowicz@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The default value for --with-database is "auto". So, if some package
installs a database library and pulseaudio auto detects, that library
will be added to the dependencies list, causing a dependency error at
do_rootfs time. For example, samba installs libtdb, which is one of
the pulseaudio supported databases.
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br> Signed-off-by: Saul Wold <sgw@linux.intel.com>
Bruce Ashfield [Thu, 18 Oct 2012 14:47:05 +0000 (10:47 -0400)]
kernel.bbclass: remove explicit version.h target
The compilation routine for the kernel has an explicit call to
build version.h, which works fine for most kernels, but the
location of it has recently changes.
commit d183e6f5 [UAPI: Move linux/version.h]
commit 10b63956 [UAPI: Plumb the UAPI Kbuilds into the user
header installation and checking]
moves the file to include/generated/linux/version.h and then to
include/generated/uapi/linux/version.h.
As a result kernel builds of 3.7 or bisection builds of intermediate
kernel commits will fail with:
make[2]: *** No rule to make target `include/linux/version.h'. Stop.
Making the explicit version.h build conditional on the version, or
via a file test would fix the problem, but it introduces some complexity
to the build.
Even without an explicit call to build version.h, it is always produced
by the kernel build, so it can simply be removed.
This extra make line was originally so that the kernel version could be
determined, so that then different instructions could be executed depending
on whether it was a 2.4 or 2.6 kernel. Since we no longer support 2.4, this
code is no longer needed.
[YOCTO: #3293]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Thu, 29 Nov 2012 18:42:22 +0000 (13:42 -0500)]
kernel-yocto: checkout known branch before leaving do_validate_branches
We should always leave the tree on a BSP branch or master when
do_validate_branches completes to avoid modifying version tracked
files are part of the build process. Modifying these files will lead
to errors when changing branches, since the contents would be lost.
This is evident in the case that a the meta branch is reset to a
known SRCREV and the tree was left on the meta branch. This branch
tracks the meta/meta-series, and other artifacts of the original
tree construction. When the build process runs, it updates these same
files, which creates a conflict when switching branches.
This has been fixed in the tree construction scripts to not track
these files, but a secondary fix is also required of not leaving
the build on these branches, to allow arbitrary trees to be built.
[YOCTO #3413]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 27 Nov 2012 16:35:58 +0000 (16:35 +0000)]
autogen: use pkg-config directly instead of guile-config
The autoconf macros in autogen use dpkg (!) and guile-config to determine
what/where Guile is.
If the build host has an installed guile, these can produce conflicting results.
More interestingly, if the Guile library source and compiled form have bad
timestamps (source newer than compiled) the configure scripts knows that Guile
is present but doesn't know what version it is, resulting in compile errors.
[ YOCTO #3370 (partially) ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 28 Nov 2012 17:28:39 +0000 (17:28 +0000)]
guile: touch compiled modules on sstate install
When guile is installed from sstate, if the compiled files get a modified time
older than the source files then guile will produce warnings like this:
;;; note: source file /data/poky-master/tmp/sysroots/x86_64-linux/usr/share/guile/2.0/ice-9/boot-9.scm
;;; newer than compiled /data/poky-master/tmp/sysroots/x86_64-linux/usr/lib/guile/2.0/ccache/ice-9/boot-9.go
Not staging the files and letting guile re-compile them on demand won't work:
;;; compiling /data/poky-master/tmp/sysroots/x86_64-linux/usr/share/guile/2.0/system/base/compile.scm
;;; it seems /data/poky-master/tmp/sysroots/x86_64-linux/usr/share/guile/2.0/system/base/compile.scm
;;; is part of the compiler; skipping auto-compilation
So, use a sstate postinst function to explicitly touch the compiled files after
the extraction to ensure that they are fresher.
[ YOCTO #3370 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix potential stack overflow in get_component() where an individual
pathname component in an extract file (specified on the command line
or in an extract file) could exceed the 1024 byte sized targname
allocated on the stack.
Fix by dynamically allocating targname rather than storing it as
a fixed size on the stack.
[YOCTO #3513]
Signed-off-by: yanjun.zhu <yanjun.zhu@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stack-based buffer overflow in the url::get_pac function in url.cpp
in libproxy 0.4.x before 0.4.9 allows remote servers to have an
unspecified impact via a large proxy.pac file.
Ross Burton [Mon, 12 Nov 2012 16:18:35 +0000 (16:18 +0000)]
initrdscripts: fix udevd in the live boot init scripts
udevd moved location and isn't in $PATH anymore, so use an absolute path to
start it.
The control socket path moved too, so mkdir the directory it's in.
Mounts the new devtmpfs on /dev device tree.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Alexandru Damian <alexandru.damian@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tom Zanussi [Fri, 26 Oct 2012 16:21:16 +0000 (11:21 -0500)]
lttng-modules: remove unused lttng-syscalls patch
commit b7e184508 (lttng-2.0: fix srcrev/pv to match the recipe
filenames) removed the
lttng-sycalls-protect-is_compat_task-from-redefiniti.patch from the
SRC_URI but forgot to remove the patch itself.
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 10 Oct 2012 21:39:36 +0000 (21:39 +0000)]
gdbm: Resolve host contamination issue
The autoconf macros detect the presence of -ldbm or -lndbm on the host
system and add the library to link against, for now good reason I can
explain.
This patch makes the build behave determinstically whether they're
present or not. Other than the extra linkage, there doesn't appear to be
any other change in behaviour from these options and they look like
dead code.
The extra linkage can cause problems where sstate is used on a machine
where the extra librbary isn't present causing build failures.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 10 Oct 2012 16:56:41 +0000 (17:56 +0100)]
pkgconfig: Drop the RREPLACES for pkgconfig-dev
This line causes pkgconfig-dev to replace pkgconfig so the package with all the files
in is replaced by one with no files. This makes no sense and hence we should just
remove this broken line.
At this point in the release, this is the safest way to fix this even if an empty -dev
package is left available.
[YOCTO #2878]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Saul Wold [Wed, 10 Oct 2012 15:55:45 +0000 (08:55 -0700)]
eglibc: force make to use /bin/bash
The eglibc core build still has a number of issues with /bin/dash.
Recently found is both sysdeps/unix/make-syscalls.sh and it's output,
which make calls via SHELL do not play well with /bin/dash. By force
make to use /bin/bash via SHELL setting it works correctly.
Currenly known issues are: make-syscalls.sh line with a bad substitution,
which can be corrected by:
vdso_symver="$(printf '%s\n' ${vdso_symver} | sed 's/\./_/')"
Following that there is an issue with emiting a '\n' through a second
echo and then to the compiler. There maybe more issues beyond that.
[YOCTO #3080]
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tom Zanussi [Wed, 10 Oct 2012 00:48:24 +0000 (19:48 -0500)]
perl: add archlib_exp variable used to generate ARCHLIB_EXP in config.h
perl.c uses an ARCHLIB_EXP define to generate compile-time code that
adds the archlibexp path to @INC during run-time initialization of a
new perl interpreter.
Because we've changed this value in a temporary way to make it
possible to use ExtUtils::Embed in the target build (the temporary
value in config.sh gets re-stripped out during packaging), the
ARCHLIB_EXP value that gets generated still uses the temporary version
instead of the original expected version (i.e. becauses it's in the
generated config.h, it doesn't get stripped out during packaging like
the others in config.sh).
This creates an unmodified version called archlib_exp that gets used
by a modified config_h.SH to get the correct value into config.h
This patch uses an unmodified version of archlibexp called
archlib_exp, introduced to config.sh, which is used to generate the
correct value of ARCHLIB_EXP into config.h
Fixes [YOCTO #3099].
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 8 Oct 2012 13:49:38 +0000 (09:49 -0400)]
linux-yocto/3.4: perf: parallel build and tools fixes
korg commit 42dcd1f4e [perf tools: Fix parallel build] fixes parallel
build issues that are being seen in the autobuilder.
We also have a fix from Tom:
[
perf annotate: replace 'expand' with equivalent sed expression
We don't have 'expand' in our userspace so we need to accomplish the
same thing using 'sed', which we do have.
]
So we apply it to all BSP branches and kernel types.
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Fri, 5 Oct 2012 22:30:48 +0000 (22:30 +0000)]
linux-yocto-tiny: Add tiny recipe for the 3.4 linux-yocto kernel
This recipe builds the "tiny" kernel type defined by the linux-yocto
meta-data. Support is defined for the qemux86 machine via
common-pc-tiny.scc in the linux-yocto meta branch. The resulting
kernel is 1.8 MB and boots to a serial console with with qemux86 and
core-image-minimal using the following command:
Saul Wold [Mon, 8 Oct 2012 01:07:11 +0000 (18:07 -0700)]
nfs-export-root: add explict no_subtree-check to suppress warning
exportfs: /etc/exports [1]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/".
Assuming default behaviour ('no_subtree_check').
NOTE: this default has changed since nfs-utils version 1.0.x
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 8 Oct 2012 04:05:19 +0000 (00:05 -0400)]
linux-yocto/3.4: tiny: Add qemu KMACHINE to common-pc-tiny.scc
Updating the meta SRCREV to pickup the following change:
Ensure the qemux86 machine is defined in common-pc-tiny as it is
for -standard and -rt.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 8 Oct 2012 04:05:18 +0000 (00:05 -0400)]
linux-yocto/3.4: fix kconfig warnings and unnecessary options
Updating the kernel configuration fragments to fix the following
issues:
- remove options that are no longer in the 3.4 kernel
- disable unused, but large kernel modules
- fix kconfig audit warnings for x86 BSPs
- make uprobes reusable by multiple fragments
The following meta branch commits are represented by this update:
3da1172 uprobes: split into enable and patch fragments 17ec51a meta: cleanup invalid/obselete 3.4 CONFIG options b5cee42 meta: disable OCFS2 by default efe937e meta: drm: tag DRM options as 'hardware' 10b5155 meta: emenlow: clean emenlow configuration warnings a907b82 meta: add CONFIG_SHMEM to standard kernel config
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Saul Wold [Fri, 5 Oct 2012 23:59:03 +0000 (16:59 -0700)]
pulseaudio: ensure X11 and consolekit are disabled
When DISTRO_FEATURES does not include X11 ensure that both x11
and gtk are diabled correctly. ConsoleKit also has x11, so ensure
that any RDEPENDS is also excluded.
The flags for x11 changed at somepoint to use enable/disable, but
this recipe was not updated.
Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mark Hatle [Fri, 5 Oct 2012 18:21:51 +0000 (11:21 -0700)]
package_rpm.bbclass: Fix an issue where complementary installs fail
Also ensure that we always cleanup the temporary install manifest files,
some of them will cause problems if they exist in multiple install
attempts.
Finally verify that the lists remain uniquely sorted otherwise the
complementary install may install the same files numerous times,
triggering a failure.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Saul Wold <sgw@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 5 Oct 2012 13:13:13 +0000 (13:13 +0000)]
sstate: Add extra entries to the sstate duplicate files whitelist
This avoids errors where gcc/binutils get installed to the native sysroot
in the same location for multiple package architectures. Ultimately making
these native recipes with ${PACKAGE_ARCH} appended to PN will resolve this
but hide the warnings until this gets sorted out.
Also hide the python and docbook catalog warnings since they're known about,
nothing to worry about and we'll aim to clean them up properly in the 1.4 cycle.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 5 Oct 2012 11:46:05 +0000 (11:46 +0000)]
libdrm: Explicitly disable the cairo dependency
We don't want the cairo dependency. Unfortunately simply checking whether its present
isn't good enough. If its not in DEPENDS, it can disappear half way through building.
We therefore need to explictly disable it.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 4 Oct 2012 23:47:39 +0000 (16:47 -0700)]
toolchain-scripts.bbclass: Export M4
some packages use M4 variable from environment and sometimes
its hardcoded to /usr/bin/m4 if not found in environment. Lets
define it such that it is picked from path
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
