Richard Purdie [Tue, 2 Aug 2011 10:42:21 +0000 (11:42 +0100)]
eglibc: Add missing ${PN} and ${PKGSUFFIX} usages
This fixes warning messages such as:
The providers found were: ['virtual:nativesdk:eglibc_2.13.bb', 'eglibc_2.13.bb']
The PREFERRED_PROVIDER entries resulting in this conflict were: ['PREFERRED_PROVIDER_virtual/libc-nativesdk = eglibc-nativesdk', 'PREFERRED_PROVIDER_virtual/libc = eglibc']
NOTE: multiple providers are available for runtime glibc-utils (eglibc, eglibc-nativesdk, external-csl-toolchain, external-poky-toolchain)
NOTE: consider defining a PREFERRED_PROVIDER entry to match glibc-utils
This highlights how ugly adding to PN is when using BBCLASSEXTEND. We need to do this to
eglibc so the nativesdk class continues to work correctly for now and ensure consistency
in how the various eglibc packages are generated.
Longer term, if we switch nativesdk to use the multilib approach and MLPREFIX, we can
remove all the PKGSUFFIX mess. I'd rather not get into changing nativesdk right at this
time though given all the other changes that are going on.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mark Hatle [Mon, 1 Aug 2011 16:56:37 +0000 (11:56 -0500)]
rootfs_rpm: Cleanup and minor bug fixes
in packages_rpm.bbclass:
Cleanup the way package_update_index_rpm works. Break this out into two
functions, one for common functionality and one for controlling target vs
sdk package processing.
Similar change to package_generate_rpm_conf.
Usage of IMAGE_ROOTFS should actually be "target_rootfs", this fixes a problem
with the SDK case.
in rootfs_rpm.bbclass:
Minor re-order of the values, and avoid multiple uses of IMAGE_ROOTFS in order
to make the merge with the SDK behavior easier.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Dexuan Cui [Tue, 2 Aug 2011 06:08:32 +0000 (14:08 +0800)]
oe-init-build-env, scripts/oe-buildenv-internal: add error detecting for $BDIR
[YOCTO #671]
"readlink -f" in Ubuntu 10.04 is buggy: it doesn't ignore a trailing / (e.g.,
"readlink -f /tmp/non-existent-dir/" returns nothing, but according to
http://www.gnu.org/s/coreutils/manual/coreutils.pdf it should do that --
hence we get bug 671. It seems Ubuntu 10.10 or even later Ubuntu 11.04,
and other Linux distributions(e.g., Open Suse 11.4) haven't such an issue.
So I think we should detect this and ask Ubuntu 10.04 users to avoid supply
a path with trailing slash here.
Moreever, I also add the detection of non-existent path, e.g.,
source oe-init-build-env /non-existent-dir/build
can be detected and we'll print an error msg.
And, if we get errors in oe-buildenv-internal, we should stop the script
and shouldn't further run.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Tue, 2 Aug 2011 01:47:13 +0000 (18:47 -0700)]
scripts/runqemu: grep for line beginning with TMPDIR
Currently the grep regexp matches any occurance of
'TMPDIR=' but if you have another variable defined
e.g. OE_BUILD_TMPDIR=xxx then that gets picked up
too.
Kumar Gala [Thu, 28 Jul 2011 18:09:50 +0000 (13:09 -0500)]
libzypp: Fix variable substitution problem in do_archgen step
The do_archgen step creates a script that utilizes the variable name
${ARCH}. However, we also utilize and define ${ARCH} so instead of
having the following in the script:
COMPAT_WITH="${ARCH},${COMPAT} $COMPAT_WITH"
We get something like:
COMPAT_WITH="powerpc,${COMPAT} $COMPAT_WITH"
Just renaming the variable in the script to not conflict with ${ARCH}
fixes the issue.
[PR bump from Saul Wold] Signed-off-by: Kumar Gala <galak@kernel.crashing.org> Signed-off-by: Saul Wold <sgw@linux.intel.com>
PACKAGE_EXTRA_ARCHS_tune-armv5eb needs to be defined in terms of
the non-e with the same endianness, i.e. PACKAGE_EXTRA_ARCHS_tune-armv5b
not PACKAGE_EXTRA_ARCHS_tune-armv5, otherwise PACKAGE_EXTRA_ARCHS will
end up containing a semi-random mixture of endiannesses and disaster
will ensue. Likewise for the vfp and armv6 variants.
This is all a bit confusing because TUNE_FEATURES is done the opposite
way around, i.e. TUNE_FEATURES_tune-armv5eb is derived by taking the
armv5e version and adding bigendian. But fixing that is probably
a subject for a separate patch.
Signed-off-by: Phil Blundell <philb@gnu.org> Signed-off-by: Khem Raj <raj.khem@gmail.com>
Because of the way BitBake handles ??= under certain circumstances, this
default setting ends up stepping all over the real setting from the arch
include file. Since virtually all arch include files or tune files define
a real value for this we shouldn't need to have a default (or it needs to
be done in a different way).
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Acked-by: Kumar Gala <galak@kernel.crashing.org> Signed-off-by: Khem Raj <raj.khem@gmail.com>
Lianhao Lu [Tue, 5 Jul 2011 05:18:06 +0000 (13:18 +0800)]
package(rootfs)_ipk.bbclass: support multilib in opkg backend.
Support install multiple multilib in opkg backend.
The installation is done in 3 phases.
Phase 1: install normal packages to IMAGE_ROOTFS.
Phase 2: install multilib packages under MULTILIB_TEMP_ROOTFS. Packages
belongs to the same multilib arch would be installed to a unique
directory.
Phase 3: check file confliction between IMAGE_ROOTFS and
MULTILIB_TEMP_ROOTFS, install multilib packages to IMAGE_ROOTFS only if
the sanity check passed.
Yu Ke [Fri, 22 Jul 2011 05:51:33 +0000 (13:51 +0800)]
eglibc: fix for multilib RPROVIDES issue
in eglibc-package.inc, the "PACKAGES" variable use the ${PN} reference,
while RPROVIDES does not use ${PN}. This will lead multilib.bbclass
not mapping its RPROVIDES, and cause "NO RPROVIDER: lib64-glibc-utils"
error.
Yu Ke [Sat, 16 Jul 2011 02:02:53 +0000 (10:02 +0800)]
do_split_packages: revise for multilib case
in multilib case, the PACKAGE_DYNAMIC is overrided with multilib
prefix. Take multilib:lib64-perl as example. the "perl-module-*"
will become "lib64-perl-module-*"
the output_pattern in do_split_packages is designed to work with
PACKAGE_DYNAMIC, so it should be applied with the same logic, i.e.
overriding with multilib prefix. otherwise the do_split_package will
split incorrect files
this patch implements the mulitlib override logic for do_split_packages
We also need to rename the extra_depends to support multilib case
(from Dongxaio Xu).
Signed-off-by: Yu Ke <ke.yu@intel.com> Signed-off-by: Xu Dongxiao <dongxiao.xu@intel.com>
Richard Purdie [Tue, 7 Jun 2011 09:27:03 +0000 (17:27 +0800)]
multilib: Add support for compiling recipes against multiple ABIs
This patch adds the core multilib class which can be used along with a
parameter specifying the mutlilib to use in BBCLASSEXTEND.
The MLPREFIX variable is added and can be used in cases where its too
difficult to dynmaically work out where a mutltilib prefix is needed
to be added to a variable.
This includes:
* SHLIBSDIR and PACKAGE_ARCH fixes from Lianhao Lu.
* PACKAGE_DYNAMIC mapping from Yu Ke
* PACKAGE_INSTALL mapping from Yu Ke
* RPROVIDES mapping from Yu Ke
* TARGET_VENDOR fix from Mark Hatle
* Ignorning *-native-runtime dependnecies as well as *-native from Yu Ke
* Map PKG and ALLOW_EMPTY from Dongxiao Xu
* Ensure RCONFLICTS and PKG field dependencies are remapped (from Dongxiao Xu)
* Ensure PN and MLPREFIX are set at the same time to ensure consistent BPN values (Yu Ke)
Signed-off-by: Yu Ke <ke.yu@intel.com> Signed-off-by: Xu Dongxiao <dongxiao.xu@intel.com> Signed-off-by: Lianhao Lu <lianhao.lu@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Wed, 27 Jul 2011 03:55:12 +0000 (23:55 -0400)]
linux-yocto: update SRCREVs
Updating the linux-yocto SRCREVs to pickup changes from the
continuing uprev to korg 3.0. With this set of update, the
meta/feature audits are complete.
qemumips and qemuppc are fixed with this update and can now
boot to a prompt.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mark Hatle [Wed, 29 Jun 2011 20:41:59 +0000 (15:41 -0500)]
RPM: multilib file class/color
We need to include file class and file color in order to do the conflict
resolution required with a multilib design. This is normally disabled when
we don't use the internal dependency generator, patch enables it in
all cases.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Mark Hatle [Fri, 15 Jul 2011 14:24:41 +0000 (09:24 -0500)]
ncurses: Resolve header and man page multilib conflicts
By default man pages are compressed, the way ncurses compresses
them causes differences (timestamp and name) to be embedded into
the gziped file. So each build will end up with a slightly
different file. Avoid this, by not compressing the man pages.
Avoid header conflict with curses.h using the multilib header
helper.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Mark Hatle [Tue, 26 Jul 2011 19:17:11 +0000 (14:17 -0500)]
eglibc_2.13: Resolve multilib header conflicts
Due to mismatches between various header definitions in i386 and x86_64
we need to overwrite the i386 versions of the headers with their x86_64
equivalents.
This can not be done for the syscall.h as it's dynamically generated,
so use the header conflict resolution helper for that.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
This helper function and associated header will allow us to
resolve two/three header files that conflict due to contents
that change based on wordsize and ABI.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
This creates a live image as an IMAGE_FSTYPES, thus removing the
need to have additional -live.bb recipes. To create a live image
one just needs to add live to the IMAGE_FSTYPES list
siteinfo: Rework the siteinfo for powerpc to fix uclibc fallout
I believe that powerpc-linux is now a common file across 32bit/64bit
linux for powerpc be it uclibc or glibc. I compared the differences
between powerpc-linux-uclibc and powerpc-linux files and it
powerpc-linux was more uptodate and all the new stuff it had was needed
for uclibc anyway so we do not need to keep exact copy of powerpc-linux
as powerpc-linux-uclibc instead we use powerpc-linux for powerpc/uclibc
targets.
Secondly linux specific files were added in archinfo dictionary
which I think logically belongs to targetinfo dictionary therefore
moved them to targetinfo
Tom Rini [Thu, 14 Jul 2011 19:59:11 +0000 (12:59 -0700)]
siteinfo: Move general realloc/malloc values to common-$libc
These are tests for glibc behavior which we have enabled in uclibc.
Note that if we ever disable MALLOC_GLIBC_COMPAT the uclibc tests
will need to be changed (but I believe this would also entail massive
patching to the rest of userspace so this should be unlikely).
Tom Rini [Thu, 14 Jul 2011 19:49:19 +0000 (12:49 -0700)]
siteinfo: Move certain db entries to common-linux
Note that we had a number of unused and ignored settings wrt mutex
support. These have been dropped as they weren't being used and
implied some odd things to boot. This is also a partial resync with
oe.dev which had some, but not all of these changes.
Kang Kai [Wed, 20 Jul 2011 02:16:47 +0000 (10:16 +0800)]
screen: add pam support
According to DISTRO_FEATURES then add screen pam support or not.
Press "Ctrl+a x" will lock the screen terminal, then unlock it will
use pam.
Add pam related configure file sceen.pam from Fedora.
Kang Kai [Tue, 19 Jul 2011 09:08:31 +0000 (17:08 +0800)]
libpam: update to 1.1.4 and add subpackage xtests
Update libpam to 1.1.4, and add dependecy cracklib because run xtexts will
need pam-plugin-cracklib.
There are some additional checks under subdirectory xtests and make it
as a subpackage libpam-xtests.
Bruce Ashfield [Mon, 25 Jul 2011 14:44:55 +0000 (10:44 -0400)]
machine/qemu: set preferred linux-yocto kernel version
The introduction of the linux-yocto-3.0 kernel is taking
precedence over the known working 2.6.37 version. Forcing
2.6.37 until 3.0 is validated on the qemu machines.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
module-init-tools: do not build static utilities for -cross
There is no need to have static module-init-tools for the cross
tools. Building such binaries requires a static libc which is not
required by the rest of OE and which is deprecated and not installed
by distributions like Fedora.
Wildcards in SRC_URI are not supported by oe-core and work only
when last (or first?) FILESPATH element points to the containing
directory. Specifying plain dir works fine for this recipe.
Default to XTERM_TERM* instead of KONSOLE_TERM* since Konsole support
only works for KDE 3.x, and note this in the comments (see Yocto bug 1294
for further details.) Also add a comment about PATCHRESOLVE = "noop".
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Tue, 26 Jul 2011 10:01:38 +0000 (11:01 +0100)]
meta/conf/local.conf.sample: remove oprofileui-native related settings
oprofileui-nativesdk would be a better way to handle these kinds of
requirements with the modern system, and it is not something that most
users will need to configure anyway (nor was it supported.)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
