bitbake: bitbake: toaster: do not trap SIGCHLD

We remove trapping SIGCHLD due to a weird interaction
with the bash version used with Ubuntu 14.04 LTS.

(Bitbake rev: 59f2f33440449c586c23dd3a192698a37aaf0595)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: prserv: Fix exit race issues

We shouldn't immediately remove the pid file when stopping the server, if we do, this
causes a traceback within the server itself which can then hang. Fix this by removing
the stale pid file as the last thing we do.

Also:

* don't printing a new "waiting" line every 0.5 seconds.
* make the loop more granular since the user can 'feel' the 0.5 seconds

[YOCTO #5984]

(Bitbake rev: 81f41a806aeddcc38992163557672e296bcbc967)

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

bitbake: bin/bitbake/cooker: Ensure initial featureset is optimal

If the featureset didn't match the defaults, we'd pay the price of two
base configuration parses which showed up adversely in the performance
benchmarks. This also passes the feature set into the cooker creation so
in the common case we don't have to reset the server. This speeds up
both knotty and hob startup. If the featureset doesn't match, the system
will reset as before, this just streamlines the common case.

(Bitbake rev: 1249543c4dbf3edeac033d888497864cfc807a4e)

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

bitbake: cooker: Only change self.data if it exists

With the change to more optimal default featureset behaviour, a race was
exposed by hob where the code may try and change self.data before it
exists. This change avoids that.

When the datastore is created, the cooker configuration is used so
data tracking is correctly handled regardless.

(Bitbake rev: 9d8f7efbc39d64124936ccaeb3c47a112e595d78)

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

zisofs-tools-native: Add missing dependency on zlib-native

(From OE-Core rev: c581059df88d5801cc20ab24a096e4a67b737d49)

Signed-off-by: Drew Moseley <drew_moseley@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

e2fsprogs: Add e2fsprogs-mke2fs and e2fsprogs-e2fsck as recommend packages

The mke2fs and e2fsck commands are regular tools of e2fsprogs, so they
should be installed.

(From OE-Core rev: a2b6610dceb0c27484d3958d84d7f491aa759a04)

Signed-off-by: Yue Tao <Yue.Tao@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-ust: add python to lttng-ust's RDEPENDS

| Note: adding Smart RPM DB channel
|
| Note: to be installed:  run-postinsts@x86_64 kernel-modules@qemux86_64 packagegroup-core-boot@qemux86_64 lttng-ust@x86_64
| Loading cache...
| Updating cache...               ######################################## [100%]
|
| Computing transaction...error: Can't install lttng-ust-2:2.3.0-r0.0@x86_64: no package provides /usr/bin/python
|
| Saving cache...
|
| WARNING: exit code 1 from a shell command.
| ERROR: Function failed: do_rootfs (log file is located at tmp/work/qemux86_64-wrs-linux/wrlinux-image-glibc-small/1.0-r1/temp/do_rootfs/log.do_rootfs.13619)

(From OE-Core rev: 0360a49f0c216b3f35c4bffabab3406c31589fd3)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpcap: add PACKAGECONFIG for libnl1

Add --with-libnl autoconfig parameter and dependency
between libpcap and libnl1.

Disable libnl1 by default to avoid libpcap build error
when libnl1 is involved.

(From OE-Core rev: 52f16a5a56868137e17cf52fa7b664047ec7bcaf)

Signed-off-by: Hu Yadi <Yadi.hu@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

syslinux.bbclass: Enable additional kernel parameters for syslinux

Add additional parameter 'SYSLINUX_KERNEL_ARGS' in order to allow
for specific kernel parameters to be set when using syslinux.

The extra kernel parameters are added to btype[1] and then written out
as part of the APPEND field.

(From OE-Core rev: d78c4d51ed266c14b0425f6abf553392c6ebe408)

Signed-off-by: Konrad Scherer <Konrad.Scherer@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

weston : refactor to identify EGL, cairoglesv2 support

(From OE-Core rev: faa86f0cda47f3ea6eeaf4c3fd3cf95a286f4571)

Signed-off-by: Prabhu Sundararaj <prabhu.sundararaj@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate-cache-management: rm_by_stamps - discover all suffixes like remove_duplicated does

* instead of hard coding list of sstate enabled task use
  the same function as remove_duplicated to find them in
  sstate-cache directory

(From OE-Core rev: dfdb397db5865c3287cd9ccb5ea9a336eb77ca90)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate-cache-management: skip populate_lic archives when removing duplicates

* there are 2 issues

* all *gcc* recipes produce "duplicate" files and we cannot keep just the newest one:
  OE qemux86@ ~/build/oe-core $ find sstate-cache/ | grep gcc.*populate_lic
  sstate-cache/cd/sstate:gcc::4.8.2:r0::3:cd8d1a33b832662b624f5e4bd9f55154_populate_lic.tgz
  sstate-cache/cd/sstate:gcc::4.8.2:r0::3:cd8d1a33b832662b624f5e4bd9f55154_populate_lic.tgz.siginfo
  sstate-cache/d5/sstate:gcc::4.8.2:r0::3:d591af812395cc386e017fa2c8fa0722_populate_lic.tgz
  sstate-cache/d5/sstate:gcc::4.8.2:r0::3:d591af812395cc386e017fa2c8fa0722_populate_lic.tgz.siginfo
  sstate-cache/f7/sstate:gcc::4.8.2:r0::3:f74cfd4982fd7c349a5418f07297f955_populate_lic.tgz
  sstate-cache/f7/sstate:gcc::4.8.2:r0::3:f74cfd4982fd7c349a5418f07297f955_populate_lic.tgz.siginfo
  sstate-cache/74/sstate:gcc::4.8.2:r0::3:7489e4502cb446506b9829cda13c7630_populate_lic.tgz
  sstate-cache/74/sstate:gcc::4.8.2:r0::3:7489e4502cb446506b9829cda13c7630_populate_lic.tgz.siginfo

  OE qemux86@ ~/build/oe-core $ find tmp-eglibc/stamps/ | grep gcc.*populate_lic
  tmp-eglibc/stamps/i586-oe-linux/gcc-cross-initial/4.8.2-r0.do_populate_lic_setscene.cd8d1a33b832662b624f5e4bd9f55154
  tmp-eglibc/stamps/i586-oe-linux/libgcc/4.8.2-r0.do_populate_lic_setscene.f74cfd4982fd7c349a5418f07297f955
  tmp-eglibc/stamps/i586-oe-linux/gcc-cross/4.8.2-r0.do_populate_lic_setscene.d591af812395cc386e017fa2c8fa0722
  tmp-eglibc/stamps/i586-oe-linux/gcc-runtime/4.8.2-r0.do_populate_lic_setscene.7489e4502cb446506b9829cda13c7630

* the same for target and native versions of the same recipe:
  OE qemux86@ ~/build/oe-core $ find tmp-eglibc/stamps/ | grep opkg-utils.*populate_lic
  tmp-eglibc/stamps/x86_64-linux/opkg-utils-native/0.1.8+gitAUTOINC+c33b217016-r0.do_populate_lic_setscene.dd21a3c5444482ce90be4c9a33d806f1
  tmp-eglibc/stamps/i586-oe-linux/opkg-utils/0.1.8+gitAUTOINC+c33b217016-r0.do_populate_lic_setscene.8571422f9e311dc41bb6b21e71a09bc0

  OE qemux86@ ~/build/oe-core $ find sstate-cache/ | grep opkg-utils.*populate_lic
  sstate-cache/dd/sstate:opkg-utils::0.1.8+gitAUTOINC+c33b217016:r0::3:dd21a3c5444482ce90be4c9a33d806f1_populate_lic.tgz.siginfo
  sstate-cache/dd/sstate:opkg-utils::0.1.8+gitAUTOINC+c33b217016:r0::3:dd21a3c5444482ce90be4c9a33d806f1_populate_lic.tgz
  sstate-cache/85/sstate:opkg-utils::0.1.8+gitAUTOINC+c33b217016:r0::3:8571422f9e311dc41bb6b21e71a09bc0_populate_lic.tgz.siginfo
  sstate-cache/85/sstate:opkg-utils::0.1.8+gitAUTOINC+c33b217016:r0::3:8571422f9e311dc41bb6b21e71a09bc0_populate_lic.tgz

* similar problem for .siginfo file of do_patch:
  OE qemux86@ ~/build/oe-core $ find sstate-cache/ | grep eglibc.*patch
  sstate-cache/99/sstate:eglibc::2.19:r0::3:99b7e1d688e4c6c659b458c9c57611df_patch.tgz.siginfo
  sstate-cache/11/sstate:eglibc::2.19:r0::3:11401ddd208a753fee17696b14dc1e52_patch.tgz.siginfo

  OE qemux86@ ~/build/oe-core $ find tmp-eglibc/stamps/  | grep /eglibc.*patch
  tmp-eglibc/stamps/i586-oe-linux/eglibc-initial/2.19-r0.do_patch.sigdata.99b7e1d688e4c6c659b458c9c57611df
  tmp-eglibc/stamps/i586-oe-linux/eglibc/2.19-r0.do_patch.sigdata.11401ddd208a753fee17696b14dc1e52

* maybe these should have ideally identical signatures, but before it's fixed lets
  just skip removing them

(From OE-Core rev: 9ae16469e70714acf7cb1e6c5238934a6b20d58a)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate-cache-management: rm_by_stamps - remove .siginfo and .done files

* it's possible that corresponding .tgz files were already removed
  (e.g. with -d option and older version of this script) and this
  won't find orphaned .siginfo or .done files to remove
* add sort -u to count files found multiple times only once

(From OE-Core rev: 5a58cd2474ed96d6f58576203bedf8211d309c99)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate-cache-management: rm_by_stamps - include signatures of .sigdata. files

* there are .siginfo files in sstate-cache, but STAMPS_DIR calls them
  .sigdata, make sure that such signatures are kept, because some tasks
  like do_package, don't have _setscene or main task entry and are
  removed:

  stamps:
  tmp-eglibc/stamps/i586-oe-linux/busybox/1.22.1-r0.do_package.sigdata.cd6f625471ef1b20a9379e90519db6f1
  tmp-eglibc/stamps/i586-oe-linux/busybox/1.22.1-r0.do_package_write_ipk.f20fe66285219f23373fc64d5de1d412
  tmp-eglibc/stamps/i586-oe-linux/busybox/1.22.1-r0.do_package_write_ipk.sigdata.f20fe66285219f23373fc64d5de1d412
  tmp-eglibc/stamps/i586-oe-linux/busybox/1.22.1-r0.do_packagedata.sigdata.c55a3dbe90010c15aa3294753fbd402e
  tmp-eglibc/stamps/i586-oe-linux/busybox/1.22.1-r0.do_packagedata_setscene.c55a3dbe90010c15aa3294753fbd402e.qemux86
  sstate (we want to keep all in this case):
  sstate-cache/c5/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:c55a3dbe90010c15aa3294753fbd402e_packagedata.tgz
  sstate-cache/c5/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:c55a3dbe90010c15aa3294753fbd402e_packagedata.tgz.siginfo
  sstate-cache/cd/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:cd6f625471ef1b20a9379e90519db6f1_package.tgz
  sstate-cache/cd/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:cd6f625471ef1b20a9379e90519db6f1_package.tgz.siginfo
  sstate-cache/f2/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:f20fe66285219f23373fc64d5de1d412_package_write_ipk.tgz
  sstate-cache/f2/sstate:busybox:i586-oe-linux:1.22.1:r0:i586:3:f20fe66285219f23373fc64d5de1d412_package_write_ipk.tgz.siginfo

(From OE-Core rev: c2e5c0b6bdc432449ad1792176aa28667c3d34b9)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Extend --rootfs-dir to connect rootfs-dirs

The wic command-line param --rootfs-dir gets generalized to support
multiple directories. Each '--rootfs-dir' could be connected using a
special string, that should be present in .wks. I.e:

wic create ... --rootfs-dir rootfs1=/some/rootfs/dir \
  --rootfs-dir rootfs2=/some/other/rootfs/dir

  part / --source rootfs --rootfs-dir="rootfs1" --ondisk sda --fstype=ext3 \
    --label primary --align 1024

  part /standby --source rootfs --rootfs-dir="rootfs2" \
    --ondisk sda --fstype=ext3 --label secondary --align 1024

The user could use harded-code directory instead of connectors. Like this:

  wic create ... hard-coded-path.wks -r /some/rootfs/dir

  part / --source rootfs --ondisk sda --fstype=ext3 --label primary --align 1024

  part /standby --source rootfs --rootfs-dir=/some/rootfs/dir \
    --ondisk sda --fstype=ext3 --label secondary --align 1024

(From OE-Core rev: 719d093c40e4c259a4c97d6c8a5efb5aeef5fd38)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Report all ROOTFS_DIR artifacts

When a .wks has more than one ROOTFS_DIR it's better to report
all ROOTFS_DIR that was used to create the image.

(From OE-Core rev: a8762f3be215678a6806cabe49647083f42323a8)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Add option --rootfs-dir to --source

The '--rootfs-dir' option is optional and only takes efect is a
partition is set up like this:

  part /standby --source rootfs --rootfs-dir=<special rootfs> ...

So '--rootfs-dir' is used instead of bitbake ROOTFS_DIR variable or
'-r' param.

(From OE-Core rev: d486db593e6643bd10b8fe90257d547a9f341043)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Use partition label to be part of rootfs filename

When a partition from .wks file is set up like this:

  part /standby --source rootfs --rootfs-dir=<special rootfs> ... --label \
    --label secondary

This means that 'rootfs' must use '<special rootfs>' as rootfs and
the default partition filename in /var/tmp/wic/build/ will be create
using the '--label' as part of the name. E.g:

  /var/tmp/wic/build/rootfs_secondary.ext3

(From OE-Core rev: c7efb3a21618ce3069811042279a0d898237ac0f)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Add rootfs_dir argument to do_prepare_partition() method

The do_prepare_partition() method from RootfsPlugin class need
to know what will be the rootfs_dir. This makes sense when .wks
file has a partition set up like this:

  part /standby --source rootfs --rootfs-dir=<special rootfs> ...

then do_prepare_partition() will work with the correct rootfs.

(From OE-Core rev: 6042b097a8fc24f2b85eb9848fb007a3c6c090a9)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Hook up RootfsPlugin plugin

Remove the 'rootfs' case when internal call code is used and
replace to call the general-purpose plugin.

For now RootfsPluing class continues to invoke prepare_rootfs()
method from Wic_PartData. However RootfsPlugin could implement them.

(From OE-Core rev: 26cd93b79318cbfaebb971d1e728041904e015f1)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Add RootfsPlugin

Implement RootfsPlugin class. The do_prepare_partition() method
is implemented using code in Wic_PartData class.

This class have 'rootfs' name, which is the name that should
be used in the --source parameters of the .wks partition commands.

(From OE-Core rev: 68dd66849bbaca6e3a0cf00beec0dba1c08e9070)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-lsb: Limit mips64 issues to qt4 packages only

The COMPATIBLE_HOST setting was only there for mips64 issues. Move that
restriction to the qt4 packages themselves so the rest of the lsb images can
be built.

(From OE-Core rev: ef7968dbded62cfce91e4f44bc96e8d04b076f15)

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

libcap 2.22: fix, disable gperf detection

gperf straight invoke is not suitable for cross environment (gperf-native should be used instead).
Formal patch has been submited to the upstream.
As libcap 2.24 is currently available, I prefer doing this quick fix.

(From OE-Core rev: 5c7946f790b5e8da8f30e3493c04c07b30b89d17)

Signed-off-by: Matthieu Crapet <Matthieu.Crapet@ingenico.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Fix bitbake_env_command for 'None' case

bitbake_env_command will choke if it isn't given an image, make sure
it does the right thing in that case.

(From OE-Core rev: a17f879cd5bc7401597ccee908801f8e3efa34c0)

Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-toolset-native: remove it

Remove it since it seems that it is not widely used by oe.

(From OE-Core rev: a56ad23ecf3fbb6d3085a856622e2c5f0018934b)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

core-image-sato-sdk: Drop qt4 on mips64

qt4 doesn't build on mips64 so don't include it in sato-sdk images.

(From OE-Core rev: 900584946698b5bb2c459ad9555709665843be2c)

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

make-3.81: Fix build failures with newer makeinfo

(From OE-Core rev: f81925316323b662f9b4d368976f4205fb96dc67)

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

linux-yocto/3.8: remove versioned recipes

Yocto 1.6 will support the LTSI 3.4/3.10 kernels and the 3.14 kernel. As
such, we remove the 3.8 linux-yocto recipes to keep our number of supported
kernels at three.

(From OE-Core rev: 940137ed36e1274bcb4e6b246b69c8c9172cabf7)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: fix drm build failure

Andrea Adami reported the following build failure:

    .../drm/drm_mm.h:105:2: error: implicit declaration of function
    'BUG_ON' [-Werror=implicit-function-declaration]
    |   BUG_ON(!hole_node->hole_follows);
    |   ^
    |   CC      drivers/pci/setup-res.o
    |   CC      drivers/gpu/drm/i915/i915_drv.o
    | cc1: some warnings being treated as errors
    | make[6]: *** [drivers/gpu/drm/ttm/ttm_agp_backend.o] Error 1
    | make[5]: *** [drivers/gpu/drm/ttm] Error 2

Cherry picking mainline commit 86e81f0e6 [drm/mm: include required headers in drm_mm.h]
fixes the build problems.

cc: Andrea Adami <andrea.adami@gmail.com>
(From OE-Core rev: 42c0eba4fac6b8bd28b58ec04574d04b0ab0c457)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: update intel common meta data

Bumping the meta branch SRCREV for the following two commits:

  df3aa753c882 intel-common: Add media-all to the standard builds
  4b0d57269dae intel-common: Add mohonpeak BSP

(From OE-Core rev: 821b1c03db3793609e3ae564358de10e23591604)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: update to v3.10.34

Updating to the latest korg -stable release.

(From OE-Core rev: 4d636396435f1165481cb103512026a8660ca8b2)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: add valleyisland io

Updating the 3.10 meta SRCREV to include the valleyisland IO .scc and
configuration files.

(From OE-Core rev: 5adccadee00ac4408d3b3d4e6a0b7ee7c84cba97)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: update to v3.10.33

Updating to the latest -stable release.

(From OE-Core rev: 9acc956c23ea161d0af691ff4685b472d3eff086)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: update EFI configuration

Importing the following two meta data changes for EFI configuration:

  284e9589436a meta: efi.cfg/efi-ext.cfg: add EFIVAR_FS to default efi fragment
  0a8c4971e2d9 meta: update efi config fragment to include EFI_STUB by default

(From OE-Core rev: ec6ff275ba44ec183c00910f47e5b8916e58ab14)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.4: update mohonpeak.cfg for SATA, SMBus, LPC, WDT, crypto & highmem64g

Updating the meta SRCREV with the latest configuration updates.

(From OE-Core rev: 8e98af96a4b2b725724cd97276168c03e95aa99d)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.10: intel-core*/common configuration updates

Importing the following configuration changes to support the latest intel
common BSPs:

  6e0e756d5137 intel-common: Remove GMA500 support
  226c3b7a2b82 intel-core*: Add baytrail soc support
  25df7acf2cc4 baytrail: Add feature/soc/baytrail
  8715856ab617 meta: input: add CONFIG_INPUT dependency

(From OE-Core rev: 88d2bb9c5959a9b1e744bb517c26a322cd537023)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.py: check file exists before deleting

When RM_OLD_IMAGE = "1", we delete old images but we didn't check they
actually exist...

[YOCTO #6029]

(From OE-Core rev: 8910d3cc94899ab4d509e681b438ae96218fa777)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_*.bbclass: Simplify addtask

The package_write task was previously removed. Remove a remaining superfluous
reference to it.

(From OE-Core rev: 76bbf9e8f07f3e6f20c890dd4c82c72641e2ca88)

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

populate_sdk_*: Drop now unneeded recrdeptask flags

Now populate_sdk_base has the appropriate flags, we can drop these from the individual
classes.

(From OE-Core rev: 388bfe2dc168d31ba3c5c85684f3c96d2ae13800)

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

populate_sdk_base: add dependency of do_package_write_* tasks

nativesdk packages were created only for the first backend listed in
PACKAGE_CLASSES. Hence, if one had it set to "package_rpm package_ipk"
and did a 'bitbake -c populate_sdk core-image-something', the nativesdk
packages were created only for rpm.

This is particularily bad for adt-installer which is based on opkg
repos.

Credits go to richard.purdie@linuxfoundation.org who suggested me this
fix.

[YOCTO #5900]

(From OE-Core rev: 85c3238ee713bc27e99a2e393e3bf8438ed4d91f)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: add new cleanup method

This commit adds a new _cleanup() internal method that will be called at
the end of rootfs creation, so that each backend can delete various
files that were probably generated during rootfs postprocess execution,
etc.

[YOCTO #6049]

(From OE-Core rev: 6151d69875f3f4f097b6e2fdef2a0f3ab391e2fd)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_manager.py: leave the __db.00* files in place

Do not delete the __db.00* files in the PackageManager class. Leave this
operation up to the client classes. One side effect of this deletion was
the following message appearing in the output of the next rpm command
executed:

rpmdb: BDB1540 configured environment flags incompatible with existing
environment

We might also gain some time here by not deleting/creating those files
very often.

[YOCTO #6049]

(From OE-Core rev: 12e300f0af2a27c15d80298d3fbb27b092c35154)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nss-3.15.1: fix CVE-2013-5605

Mozilla Network Security Services (NSS) 3.14 before 3.14.5 and
3.15 before 3.15.3 allows remote attackers to cause a denial
of service or possibly have unspecified other impact via
invalid handshake packets.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-5605
(From OE-Core rev: 09e8cd6f09284ad3faf0bc05d623a43e2b174866)

Signed-off-by: yanjun.zhu <yanjun.zhu@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nss-3.15.1: fix CVE-2013-1741

Integer overflow in Mozilla Network Security Services (NSS)
3.15 before 3.15.3 allows remote attackers to cause a denial
of service or possibly have unspecified other impact via a
large size value.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-1741
(From OE-Core rev: b666d173ff0ba213bf81e2c035a605a28e5395ea)

Signed-off-by: yanjun.zhu <yanjun.zhu@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libtiff: fix CVE-2013-4244

cve description:
The LZW decompressor in the gif2tiff tool in libtiff 4.0.3 and earlier
allows context-dependent attackers to cause a denial of service
(out-of-bounds write and crash) or possibly execute arbitrary code via
a crafted GIF image.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-4244
(From OE-Core rev: 4eec8fae3f972a27bfb986066f5b3603599ebc25)

Signed-off-by: Baogen Shang <baogen.shang@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libarchive: fix CVE-2013-0211

CVE description:
Integer signedness error in the archive_write_zip_data function in
archive_write_set_format_zip.c in libarchive 3.1.2 and earlier, when running
on 64-bit machines, allows context-dependent attackers to cause a denial of
service (crash) via unspecified vectors, which triggers an improper conversion
between unsigned and signed types, leading to a buffer overflow.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-0211
(From OE-Core rev: 355a8086637b859a469e1f2dc717b4ccec00b970)

Signed-off-by: Baogen Shang <baogen.shang@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

autoconf: new autotest/testsuite option to enable automake test result format

lib/autotest/general.m4: added "--am-fmt | -A" command line parameter
for testsuite script to enable "RESULT: testname" output format; to be
used by yocto ptest packages directly or with autoconf TESTSUITEFLAGS.

(From OE-Core rev: 9d288e5afd6696cc519574470c7d47ca55403d27)

Signed-off-by: Radu Patriu <radu.patriu@enea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: filter tasks with cache attempts for all attempts

Adjust the filter for task cache attempts to include all attempts except
SSTATE_NA.

[YOCTO #5923]

(Bitbake rev: b1f52a87e367b8dea9bd974bc5a886d84d839c45)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: fix dirinfo empty dir expansion

[YOCTO 6036]

In the page that shows the directory layout for an image, if the
directory is empty, then the directory folder icon should be black and
the table row should not be 'expandable' on a mouse click.  That
behavior depends on the directory's child entry count calculated in the
view function controlling that page.

Two images in the database with the same directory path in the target
image, but one with path having entries and the other not having
any entries caused the path without entries to be clickable; the
query for a directory's count of entries, didn't filter on the image id,
only on the path.

(Bitbake rev: 964d2d6efe9a2cfa7cd8760cda4453c3d69b2e27)

Signed-off-by: Dave Lerner <dave.lerner@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: added file types to the Outputs column in the build

The file types are displayed in the Outputs column in the build page. The file types
are derived from the target image filenames.

[YOCTO #5947]

(Bitbake rev: 842abf6759894690d5bc770f4ea2ac15b127e5e2)

Signed-off-by: Farrell Wymore <farrell.wymore@windriver.com>
Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: add support for empty states in pages

Add support for empty states in the top build page,
the all packages page, and the all tasks page.

[YOCTO #4865]

(Bitbake rev: eaff7b50d7102c97b75df185b9ef917970319d59)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: show installed package name

[YOCTO #5922]

Implement changes that show the installed package name after the
official 'recipe-named' package name.  If the alias exists and
is different than the package name, then the alias is shown as a
'muted' string after the package name in the form 'as some-alias-name'.
This formatting appears in the included package pages in the elements:
* local breadcrumbs at the top of package included pages,
* <h1> title headings along with a help bubble that is not hovering,
and
* package lists where the help bubble appears when the mouse hovers
over the row.

The changes in detail in this patch per file are:
views.py
- added function that tests whether the package object's installed_name
should be shown,
- added function that appends package name with version and revision to
encapsulate package name formatting in one place and referred to as
package.fullpackagespec,
- changed package_built* and package_included* functions to use both
of the above new formatting functions, passing the formatted values to
templates, and
- adhered to django coding styles by renaming  module local
'get_package*' functions with "_" prefix.

package_detail_base.html
- added display of package aliases for included package page,
- refactored to use package.fullpackagespec, formatted by view function,
- added javascript function to format package alias with help, and
- removed trailing whitespace.

package_included_detail.html
- used javascript function above to format package alias, and
- refactored to use package.fullpackagespec.

package_included_dependencies.html
- used javascript function above to format package alias,
- refactored to use package.fullpackagespec,
- forced empty data cells following hover-help to draw borders
by appending space, and
- removed trailing whitespace.

package_included_reverse_dependencies.html
- use javascript function above to format package alias,
- refactor to use views fullpackagespe, and
- force empty data cells following hover-help to draw borders
by appending space.

package_built_detail.html
- refactored to use package.fullpackagespec, and
- removed trailing whitespace.

package_built_dependencies.html
- refactored to use package.fullpackagespec, and
- removed trailing whitespace.

projecttags.py
- removed unused filter to handle installed name
- removed extra spaces around "title = " in format_vpackage_namehelp

(Bitbake rev: c604e14df8cdb1f47535f093d7044955d4c2057d)

Signed-off-by: Dave Lerner <dave.lerner@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: Fix typo in heading code

Fix typo in the code that handles the changes
in the h1 text in order to display the number of results
returned by a search.

[YOCTO #6001]

(Bitbake rev: 65bdd6cf8d0cc2af6cd424de735a5e3f2e54fa99)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: insure _get_query returns distinct records

The '_get_query' can return duplicate records if a search term appears
multiple times in the same row, so the queryset must be made
distinct before returning.
This commit also removes the initial special case for configvars in
favor of this general solution.

[YOCTO #6012]

(Bitbake rev: d21b64bad8a6a5e23eab552868d555f6e004f4c7)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: blocks for custom/highlighted navigation and breadcrumb links

Implement the navigation links as blocks so that each page can override and have
its respective link appear highlighted. Make the build breadcrumb a block so that
it is customizable to not be a link for the dashboard page. Reorder the page headers
to be consistent order for extends, projecttags, localbreadcrumb, nav-links.

[YOCTO #5916]
[YOCTO #4258]

(Bitbake rev: cb26c4df04170143babd6c9fd60600bfb31486ed)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: toaster: add Image detail and multiple targets to dashboard

Filled in the Image section detail information and allow for multiple targets.
Each target has a separate section. Added license manifest display. Changed the
target of the license manifest link. Added Tasks failed in the build summary.
The target lists required filters to create sorted lists.

[YOCTO #4258]
[YOCTO #5936]

(Bitbake rev: 09b099903bdf51bfb277b9a8f922255cfe83ab96)

Signed-off-by: Farrell Wymore <farrell.wymore@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-toolset-native: Update after ocf-linux -> cryptodev-linux change

(From OE-Core rev: b16de9af5c7108396a347e7107c73608a2f8d74f)

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

ocf-linux: remove recipe

ocf-linux only provides header file and no kernel module is built. We
can't use ocf-linux without its implementation. And linux-yocto uses an
alternative project cryptodev-linux, so we remove ocf-linux and use
cryptodev-linux instead.

(From OE-Core rev: 45f1659f49edbceed0b75c0319880151161fdc8e)

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssl: replace dependency ocf-linux with cryptodev-linux

ocf-linux only provides header files but no implementation in kernel.
And Yocto kernel linux-yocto use cryptodev-linux to implement
/dev/crypto interface. So replace dependency ocf-linux with
cryptodev-linux for openssl.

(From OE-Core rev: b36b15cddbe52e6770b96e06af2959cea0e2436f)

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cryptodev-linux: add recipe

Yocto kernel linux-yocto uses cryptodev-linux to use device /dev/crypto.
So add cryptodev-linux which is one alternative of ocf-linux and then
remove ocf-linux later.

(From OE-Core rev: 6b6c24eccdb0030ecccadefe94c1c5b4387e46d1)

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

run-postinsts: fix issue with checking IMAGE_FEATURES

The old implementation was wrong. It was not very generic and it checked
IMAGE_FEATURES while building the recipe, which led to various issues
with the generation of the final script. That is, the run-postinsts
script was generated once, while building the package for the first
time. Hence, any other changes to IMAGE_FEATURES, like removing/adding
'package-management' did not reflect in the final script.

This commit makes run-postinsts script autodetect the backend used for
creating the image, making it generic.

[YOCTO #5666]
[YOCTO #5972]

(From OE-Core rev: 44902f7550e490a9d4d2e2bcdf8c577329b4af75)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

build-appliance-image: update to head ae938eba92b2c89a9fd91161e57c5dbc594ad4ad

Build Appliance includes a more recent version
of poky, up to commit:
ae938eba92b2c89a9fd91161e57c5dbc594ad4ad.

Adds bitbake fixes/optimizations;

(From OE-Core rev: f41f1b263438e19e2209876798bbcbbcee646f34)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gummiboot: fix the installed but not shipped warning

Fix the warning:
WARNING: QA Issue: gummiboot: Files/directories were installed but not shipped
  /usr/lib
  /usr/lib/gummiboot
  /usr/lib/gummiboot/gummibootx64.efi

This is because it uses "/usr/lib" in gummibootlibdir, use ${libdir} to
fix it.

(From OE-Core rev: ee587332f2ac9d88d4a300732645b0e2f793ce5f)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gummiboot: use objcopy from the env

Fix the error:
[snip]
| nm -D -u src/efi/gummiboot.so | grep ' U ' && exit 1 || :
| objcopy -j .text -j .sdata -j .data -j .dynamic \
|   -j .dynsym -j .rel -j .rela -j .reloc -j .eh_frame \
|   --target=efi-app-x86_64 src/efi/gummiboot.so gummibootx64.efi
| objcopy: src/efi/gummiboot.so: Invalid bfd target
| make[1]: *** [gummibootx64.efi] Error 1
| make[1]: *** Waiting for unfinished jobs....
| make: *** [all] Error 2
[snip]

It uses the "objcopy" directly, which is not suitable for cross compile.

[YOCTO #6054]

(From OE-Core rev: 393fe40c14e7adeadbccb72953027b63b6f8030c)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gummiboot: add COMPATIBLE_HOST

The gummiboot depends on gnu-efi which had set:

COMPATIBLE_HOST = "(x86_64.*|i.86.*)-linux"

We also need set this for gummiboot, otherwise there would be build
failures for other non-x86 archs.

(From OE-Core rev: f1b23a32d0c823577cec532e3646c2f78e81ccda)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnu-efi: fix the LIBDIR

Its LIBDIR in Makefile is:

LIBDIR = $(PREFIX)/lib

This is incorrect for 64 bit bsp, thus will cause build failures on gummiboot:

ld: cannot open linker script file /path/to/usr/lib64/elf_x86_64_efi.lds: No such file or directory

[YOCTO #6053]

(From OE-Core rev: a18e4bef5f284c5b940007e60c7be28128a94c44)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: knotty: Show a link to the logfile for failed setscene tasks

Its not immediately obvious to the user that a logfile exists for a failed setscene
task. Add code to knotty to display where that logfile is in those cases.

[YOCTO #6055]

(Bitbake rev: 0664fa15597785dd90cf205531a9801e6da6ba47)

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

bitbake: knotty: Split error output onto stderr

When using bitbake -e in scripts, it would be helpful if the error
output appeared on stderr, not stdout. This change enables that building
upon the new bb.msg filters.

[YOCTO #5274]

(Bitbake rev: ebb797fc5c37d729e3cc8b2dc7156287d385c13b)

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

bitbake: knotty: Ensure the progress bar shows on stdout

I can understand why some programs would want the progress on stderr so
that real output can be captured on stdout. This is confusing for bitbake
since we don't show a progress bar at all in non-interactive cases.

Therefore make sure the progress bar goes to stdout, not the stderr default.

(Bitbake rev: 0529aa9966df5c56b07affe865efce18852efe5a)

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

bitbake: msg: Add stdout/stderr filters

Add logging filters which can be used to split output between stdout and
stderr, ERROR messages and above as passed by the Err filter, anything
below ERROR is passed by the Out filter. This is useful when trying to make
stderr more useful.

(Bitbake rev: d3e1419ef76be5e9ec976597361a5e14a7b6bcb6)

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

bitbake.conf: Adds bitbake qemu option for ppc e6500 & ppc e6500-64b.

This patch adds the bitbake qemu option for the ppc e6500 &
ppc e6500-64b architectures.

(From OE-Core rev: 62b0f09c13aa8e9c75ddea286586d1a2385a80be)

Signed-off-by: Valentin Cobelea <valentin.cobelea@enea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: cooker/event: Overhaul sanity test mechanism

Sanity tests are currently a pain as its hard to control when they run. This results
in issues where for example the bitbake -e output is not useful as the sanity tests
prevent it from executing. The sanity tests should run later than the base configuration.

This patch changes the sanity tests to always be event triggered with the option of
returning either events on the status, or raising errors. A new cooker feature is used
to change the behaviour depending on the controlling UI.

This does need a change to sanity.bbclass in the OE metadata but its worth the pain
for the increased flexibility and control this offers UIs and the improvement to the
user experience.

(Bitbake rev: 32e171bcc92c6e27fefee971e8714ddf8e1a8ac1)

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

bitbake: runqueue/siggen: Pass in commandline options to dump_sigs()

This allows the commandline options to be processed in the dump signature
code.

(Bitbake rev: ef8537a2e9b48f4fe065a165c102935aee2c9029)

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

bitbake: bitbake: Force -S option to take a parameter

There is no easy way to make this change. We really need parameters for the -S
(dump signatures) handling code. Such a parameter can then be used within the
codebase to handle the signatures in different ways.

For now, "none" is the recommended default and "printdiff" will execute the
new (and more expensive) comparison algorithms.

(Bitbake rev: b9873588696507dfb6aade6821f6f75cb9a19e0a)

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

selftest/bbtests.py: Fixed regex and added bitbake output to test_warnings_errors

The test failed when more than 1 error or 1 warning is present.
Also pasting the bitbake output when the test fails.

(From OE-Core rev: abc691026592b406e69f8bf9e4fffe2e6a17fffc)

Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lsb: fix lsb_log_msg() implementation

LSB lsb_log_message calls a begin()
function that should be implemented
in /etc/init.d/functions.
The aforementioned script does not
implement the begin() function, as
such there is a small issue related to logging.

This fix implements a local version of
the function, while cleaning up the
troublesome previous implementation.

Fix [YOCTO #5795]

(From OE-Core rev: 365ab9118b6c68aedb2e79129202b385329a8abb)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass: replace rootfs with /dev/root

Replace 'rootfs' with '/dev/root' in read_only_rootfs_hook function
to match the latest change in fstab file from the base-files recipe.

The related commit is as follows.

commit e8bc7a136a81a0d8df2d32dfba0920c1b2835141
    base-files: use /dev/root in /etc/fstab for systemd support

(From OE-Core rev: 31b5aeb5a0b82842e1dd8545bf5d43778d8c218b)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-serialgetty: update to match systemd 211

(From OE-Core rev: c48a2827a5494983dcefa70d8bebcc50e8a5fc3b)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

git: Fix perl paths in scripts and population of the perltools package

References to "perl-native" were slipping into the target packages. These
changes ensure those references are cleaned up and that tools using perl
are packaged in the correct perltools package. The same issues affected
the nativesdk-git output so are also applied there.

[YOCTO #5918]

(From OE-Core rev: fd4a6b0cd275931e552cd23233c178e9ec54bdbb)

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

base.bbclass: Merge two ConfigParsed event handlers

There were two ConfigParsed event handlers in base.bbclass, this merges
them together for small efficiency wins.

(From OE-Core rev: ff919ed132b543f70e9635be7a31f799aafcf8d6)

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

sanity.bbclass: Update against bitbake sanity event changes

Bitbake will now trigger sanity events when it needs the checks to run in all cases
so we can drop the ConfigParsed hook. We now control whether events are generated
or errors are raised from the event itself.

(From OE-Core rev: 97108a5647f9278280c923ef69d2b0b945a26eef)

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

scripts: Update after addtion of parameter to bitbake -S

(From OE-Core rev: fe2692c1ad1fcdbcdbc0ba419db8ff9ddb390fa7)

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

base.bbclass: Run oe_import before other INHERITs

Its possible for classes listed in INHERIT directives to use things like
the oe.utils functions. If that happens the user sees a traceback since
the modules don't become available until the ConfigParsed event.

This change to use immediate expansion means that the oe modules become available
much sooner and can be used in the core classes, including within base.bbclass.

(From OE-Core rev: a9ecad713f37f2703e99c6b856207abeb6c5ad1f)

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

meta-skeleton: Add name attribute to SRC_URI

The SRCREV_machine line does not work without having name=machine
attribute in SRC_URI.

This error is seen if the custom kernel recipe is used without
the name attribute:

NOTE: Error during finalise of .../linux-yocto-custom.bb
ERROR: ExpansionError during parsing .../linux-yocto-custom.bb: Failure expanding variable do_patch: ExpansionError: Failure expanding variable SRCPV, expression was ${@bb.fetch2.get_srcrev(d)} which triggered exception FetchError: Fetcher failure for URL: 'git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git;protocol=git;nocheckout=1'. Please set a valid SRCREV for url ['SRCREV_default_pn-linux-yocto-custom', 'SRCREV_default', 'SRCREV_pn-linux-yocto-custom', 'SRCREV'] (possible key names are git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git;protocol=git;nocheckout=1, or use a ;rev=X URL parameter)

(From OE-Core rev: 760ae021fe1714d04c34bc00d472e2d756b3823a)

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: Add SD/MMC-Cards support to '--ondisk' param

The special case when wic is set up to use SD/MMC-Cards in place
of sdX disks is not handled properly.

Append 'p' to the rootdev when disk is SD/MMC-Cards fix this situation.

(From OE-Core rev: 3a95c4549f743aa47456c76e687a863c64c7a7f4)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

apt/package_mamager: Ensure WORKDIR is used for lists directory

The native sysroot should not be used as a store for the lists files since
multiple images running at once would conflict over this. Instead redirect
this to WORKDIR. This means some extra directories need to be created.

Also create apt.conf.d to silence some warnings.

(From OE-Core rev: dc4abfc8f99c08e0c1ac9d098ce17838d0eda028)

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

package_manager: Fix apt-ftparchive index creation problems

a) There were missing parameters to the release and package commands (".")

b) The commands need to be executed as one block since they build upon each other

(From OE-Core rev: a3965b76ed4361455c89c982761263be03e1a8e5)

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

package_manager: Ensure we don't process directories twice

Processing directories twice is both pointless and introduces a race condition.
When building the list, ensure duplicates (like "all" and "noarch") are handled
correctly.

(From OE-Core rev: 4c487543422ae471a01a573bab44e3f6a6d2497a)

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

mmc-utils: Add user space mmc utilities for eMMC

* The mmc-utils are useful userspace utilities for configuring and
  working with MMC devices.  These are particularly useful when
  working with eMMC devices to do the initial programming of the
  device.

(From OE-Core rev: 6ce4010951a291aec72a3e4997cd7c523a22ac87)

Signed-off-by: Chase Maupin <Chase.Maupin@ti.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base-files: do_install.sigdata: remove the depends on DATE

If we run "bitbake -S base-files" today, and re-run it tomorrow with
nothing changed, we would see that the do_install.sigdata changes
because of:

do_intall -> do_install_basefilesissue -> DISTRO_VERSION -> DATE

We had set:
IMAGE_NAME[vardepsexclude] += "DATETIME"
in meta/conf/bitbake.conf, we can set a similar line in
base-files_3.0.14.bb to fix the problem.

[YOCTO #6032]

(From OE-Core rev: cd06824bda76a9d08a3318e0621e31c0e8c39f74)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

meta/conf/bitbake.conf: add STAMPCLEAN to BB_HASHBASE_WHITELIST

The problem is that do_configure.sigdata depends on STAMPS_DIR because:

do_configure -> STAMPCLEAN -> STAMPS_DIR

this will make the sigdata generated by "STAMPS_DIR=/tmp/stps bitbake -S
recipe" doesn't match the ones in our build dir, but it should. We can
add STAMPS_DIR or STAMPCLEAN to BB_HASHBASE_WHITELIST to fix the
problem, but we can't add  STAMPS_DIR since once it is in
BB_HASHBASE_WHITELIST, the "STAMPS_DIR=/tmp/stps bitbake -S recipe"
would not run again.

[YOCTO $6031]

(From OE-Core rev: faf3e74d5c488a66fdabd485eb916f555d7353fd)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

util-linux-native: fix qsort_r for CentOS 5.10

The qsort_r() was added to glibc in version 2.8, so there is no qsort_r() on
the host like CentOS 5.x, use qsort() to fix it since they are nearly
identical.

(From OE-Core rev: cda5310e32ce05bc54602d4c18ee2d28a53be57f)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cpio rootfs build: Avoid modifying rootfs dir

The Linux kernel requires that initrd images contain a /init file for
the image to be used as an initrd, even if it is empty. Adding it into
the rootfs directory creates a race, that can upset tar when building
both a .tar and .cpio image file ("tar: .: file changed as we read it").
Additionally, whether or not the tar file will contain the /init file is
also up to the race condition.

To avoid this problem, move the /init addition out from the rootfs
directory, and thus only include it in the .cpio image.

(From OE-Core rev: 706055503f493a5af73240c0983c46dbe31d8fe9)

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
Signed-off-by: Josep Puigdemont <josep.puigdemont@enea.com>
Cc: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnupg: CVE-2013-4576

GnuPG 1.x before 1.4.16 generates RSA keys using sequences of introductions
with certain patterns that introduce a side channel, which allows physically
proximate attackers to extract RSA keys via a chosen-ciphertext attack and
acoustic cryptanalysis during decryption. NOTE: applications are not typically
expected to protect themselves from acoustic side-channel attacks, since this
is arguably the responsibility of the physical device. Accordingly, issues of
this type would not normally receive a CVE identifier. However, for this
issue, the developer has specified a security policy in which GnuPG should
offer side-channel resistance, and developer-specified security-policy
violations are within the scope of CVE.

(From OE-Core rev: 46b80c80b0e008820b34f4360054e1697df2650d)

Signed-off-by: Yong Zhang <yong.zhang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnupg: CVE-2013-4351

GnuPG 1.4.x, 2.0.x, and 2.1.x treats a key flags subpacket with all bits
cleared (no usage permitted) as if it has all bits set (all usage permitted),
which might allow remote attackers to bypass intended cryptographic protection
mechanisms by leveraging the subkey.

(From OE-Core rev: 259aebc9dbcaeb1587aaaab849942f55fa321724)

Signed-off-by: Ming Liu <ming.liu@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Security Advisory - openssl - CVE-2013-6449

The ssl_get_algorithm2 function in ssl/s3_lib.c in OpenSSL before 1.0.2
obtains a certain version number from an incorrect data structure, which
allows remote attackers to cause a denial of service (daemon crash) via
crafted traffic from a TLS 1.2 client.

(From OE-Core rev: 3e0ac7357a962e3ef6595d21ec4843b078a764dd)

Signed-off-by: Yue Tao <Yue.Tao@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Security Advisory - openssl - CVE-2013-6450

The DTLS retransmission implementation in OpenSSL through 0.9.8y and 1.x
through 1.0.1e does not properly maintain data structures for digest and
encryption contexts, which might allow man-in-the-middle attackers to
trigger the use of a different context by interfering with packet delivery,
related to ssl/d1_both.c and ssl/t1_enc.c.

(From OE-Core rev: 94352e694cd828aa84abd846149712535f48ab0f)

Signed-off-by: Yue Tao <Yue.Tao@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Security Advisory - openssl - CVE-2013-4353

The ssl3_take_mac function in ssl/s3_both.c in OpenSSL 1.0.1 before
1.0.1f allows remote TLS servers to cause a denial of service (NULL
pointer dereference and application crash) via a crafted Next Protocol
Negotiation record in a TLS handshake.

(From OE-Core rev: 35ccce7002188c8270d2fead35f9763b22776877)

Signed-off-by: Yue Tao <Yue.Tao@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: data_smart: Fix caching issue for double remove references

FOO = "foo bar"
FOO_remove = "bar"
FOO_FOO = "${FOO} ${FOO}"

would show FOO_FOO = "foo foo bar" rather than the expected "foo foo".

This is actually a cache bug, this patch ensures the right value is
put into the cache. The preceeding patch adds a test case to ensure
we don't regress in future.

[YOCTO #6037]

(Bitbake rev: 2a80735183e8faa110b4c6d8d85c4707f28e03a1)

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

bitbake: test/data: Add in test for append/prepend/remove override operations

We currently don't have test cases for _append, _prepend and _remove. This
patch adds some basic tests and includes a test case for a recently reported
issue with the _remove operator.

(Bitbake rev: 93291bd90e18808c7a1c8c692949396bbc7e4348)

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