| Age | Commit message (Collapse) | Author | Files |
|---|
|
Enhance the logic behind the 'auto' mode a bit by only updating the
SRCREV if the changes are already found upstream. The logic is simple:
update SRCREV only if the current local HEAD commit is found in the
remote branch (i.e. 'origin/<branch_name>'). Otherwise resort to
patching.
This affects a couple of the oe-selftest tests so update those as well.
[YOCTO #7907]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Upgrades a recipe to a particular version and downloads the source code
into a folder. User can avoid patching the source code.
These are the general steps of the upgrade function:
- Extract current recipe source code into srctree and create a branch
- Extract upgrade recipe source code into srctree and rebase with
previous branch. In case the rebase is not correctly applied, source
code will not be deleted, so user correct the patches
- Creates the new recipe under the workspace
[YOCTO #7642]
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If externalsrc is enabled the 'do_unpack' task is run if the recipe has
some local source files. In the case of kernel recipe this caused the
(externalsrc) source tree to be moved/symlinked. This patch prevents the
behaviour, making sure the source tree is not moved around when
externalsrc is enabled. Instead of moving the source tree,
STAGING_KERNEL_DIR will be a symlink to it.
[YOCTO #6658]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Make sure that 'do_unpack' is executed before 'do_kernel_metadata'.
Enabling externalsrc for kernel disables 'do_validate_branches' task
which caused 'do_kernel_metadata' to fail as the dependency chain to
'do_unpack' got broken.
[YOCTO #6658]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Kernel metadata repository needs to be fetched/unpacked into the work
directory in order to build the kernel. Sources pointing to a remote
location are skipped by externalsrc by default which caused kernel build
to fail (because of remote kernel-meta was not made availeble). This
patch will make kernel-meta always available.
[YOCTO #6658]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
- Opkg 0.3.0 deprecated --disable-shave & --withopkglibdir options.
- Add -i option to autoreconf since opkg tarball is missing conf.compile.
- Recreate wget_cache.patch
Signed-off-by: Alejandro del Castillo <alejandro.delcastillo@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
[YOCTO #7898]
[YOCTO #7761]
[YOCTO #7804]
Signed-off-by: Cristina Agurida <cristina-danielax.agurida@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Without this patch, cross compiling for AArch64 with
"-mcpu=thunderx" fails with below error,
-- snip --
Assembler messages:
Error: unknown cpu `thunderx'
Error: unrecognized option -mcpu=thunderx
-- CUT --
Signed-off-by: Jagadeesh Krishnanjanappa <jkrishnanjanappa@mvista.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Upating the SRCREVs to import the following commit:
Author: Kevin Hao <kexin.hao@windriver.com>
Date: Mon Aug 31 19:34:48 2015 +0800
mips: octeon: use ll/sc for the atomic ops for all the predecessor
of octeon2
Even the octeon plus has the support of the 'saa' instruction, but
we
don't have a way to distinguish between octeon and octeon plus at
compile time and pass "-march=octeon" to all the predecessor of
octeon2. So it will cause the following error when trying to
assemble
the "saa" instruction with option "-march=octeon":
{standard input}: Assembler messages:
{standard input}:360: Error: Unrecognized opcode `saa $2,($4)'
scripts/Makefile.build:308: recipe for target
'arch/mips/cavium-octeon/executive/cvmx-lap.o' failed
Forcing to use the "ll/sc" for the atomic ops for all the
predecessor
of octeon2 to fix this issue.
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
killpg should be being called with a pgid, not a pid.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently, we see runqemu and qemu-system-* processes left behind when
bitbake is killed by buildbot. This is due to the use of setpgrp() in
the runqemu subprocess call.
We need the setpgrp call so that all runqemu processes can easily be
killed (by killing their process group). This presents a problem if this
controlling process itself is killed however since those processes don't
notice the death of the parent and merrily continue on.
Rather than hack runqemu to deal with this, we add something to
qemurunner, at least for now to resolve the issue. Basically we fork off
another process which holds an open pipe to the parent and also is
setpgrp. If/when the pipe sees EOF from the parent dieing, it kills the
process group. This is like pctrl's PDEATHSIG but for a process group
rather than a single process.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
* without this the output wasn't shown anywhere even when the bb.warn
says:
"See log for details!"
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fixed when:
PACKAGE_CLASSES = "package_rpm"
IMAGE_INSTALL_append = " perl-module-warnings-register"
$ bitbake core-image-minimal
[snip]
ERROR: perl-module-warnings-register not found in the base feeds
[snip]
And it works well when PACKAGE_CLASSES = "package_ipk" since perl
provides perl-module-warnings-register, the "smart install
perl-module-warnings-register" also works well, this was because
_search_pkg_name_in_feeds() only searched pkg name, but no provides,
this patch fixes the problem.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
update-rc.d is now needed by systemd to interact with sysv
scripts, so if we have a mixed system, then we can not uninstall
update-rc.d as it is going to break systemd functionality
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
getopts is a common applet more so now needed by systemd for working
with sysv scripts
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If a task shows an error, the full log is surpressed since bitbake assumes
the user has been shown what is wrong. In this code path that isn't the
case and its much more helpful to show the user the full error. Therefore
show a warning instead to aid usability.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The smc91c111.c driver appears to have several issues. The can_receive()
function can return that the driver is ready when rx_fifo has not been
freed yet. There is also no sanity check of rx_fifo() in _receive() which
can lead to corruption of the rx_fifo array.
release_packet() can also call qemu_flush_queued_packets() before rx_fifo
has been cleaned up, resulting in cases where packets are submitted
for which there is not yet any space.
This patch therefore:
* fixes the logic in can_receive()
* adds logic to receive() as a sanity check
* moves the flush() calls to the correct places where data is ready
to be received
Its currently undergoing discussion upstream about exactly which pieces
are the correct fix but for now, this stops the segfaults OE is seeing
which has to be an improvement.
[YOCTO #8234]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Enable updating the installed extensible SDK from a local or remote
server, avoiding the need to install it again from scratch when
updating. (This assumes that the updated SDK has been built and then
published somewhere using the oe-publish-sdk script beforehand.)
This plugin is only enabled when devtool is used within the extensible
SDK since it doesn't make sense to use it next to a normal install of
the build system.
E.g.
devtool sdk-update /mnt/sdk-repo/
devtool sdk-update http://mysdkhost/sdk
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When publishing SDK, what we want is basically its metadata and sstate
cache objects. We don't want the SDK to be prepared with running bitbake
as it takes time which reproduces meaningless output for the published SDK.
So this patch adds an option to allow for SDK to be extracted without
preparing the build system.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
These dependencies were deliberately removed because it was assumed that
they were provided by nativesdk packages. On the one hand, nativesdk packages
in extensible SDK don't have these packages; on the other hand, even if we
add these nativesdk packages, they are still not useful because we we need
runqemu to run correctly.
So we don't remove these native qemu dependencies.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Copy the contents of local.conf under TOPDIR into the final generated
local.conf. In this way, custom settings are also made into the final
local.conf like IMAGE_INSTALL, DISTRO_FEATURES, VIRTUAL-RUNTIME_xxx,
etc. Comments and blank lines are filtered out.
Before this change, installing extensible SDK would usually report failure
when preparing the build system if the user has custom configuration for
DISTRO_FEATURES in local.conf. Also, items in IMAGE_INSTALL_append in
local.conf also don't get built correctly.
This patch solves the above problem by making use of bb.utils.edit_metadata.
In addition, we check to avoid any setting that might lead to host paths
bleeding into the SDK's configuration. Basically, variables with values
starting with '/' are removed. A whitelist mechanism is introduced so that
users could specify variables that should not be ignored. The name of the
whitelist is SDK_LOCAL_CONF_WHITELIST.
The SDK_META_CONF_WHITELIST is removed as it's of no use after this
change.
SDK_LOCAL_CONF_BLACKLIST can be used to prevent copying specific
variable settings to the extensible SDK's local.conf; the default is to
exclude PRSERV_HOST (since this is likely to be internal). Similarly,
SDK_INHERIT_BLACKLIST to forbit local.conf in SDK to inherit certain
classes such as 'buildhistory' or 'icecc' that would not normally make
sense in an SDK environment.
[YOCTO #7616]
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The previous code assumes that bitbake/ directory is under the core layer.
This is the case for Yocto project. But users might clone oe-core and bitbake
separately. So we use bb.__file__ to locate the bitbake directory to make sure
it's copied into the extensible SDK.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If we do `bitbake buildtools-tarball' and then after one day do `bitbake
core-image-minimal -c populate_sdk_ext', we would meet errors like below.
| install: cannot stat '/buildarea2/chenqi/poky/build-systemd/tmp/deploy/sdk/
poky-glibc-x86_64-buildtools-tarball-core2-64-buildtools-nativesdk-standalone
-1.8+snapshot-20150429.sh': No such file or directory
The problem is that the output name for buildtools-tarball has ${DATE} in it.
So if populate_sdk_ext task is executed but buildtools-tarball is not rebuilt,
the above error appears.
Instead of hardcoding ${DISTRO_VERSION} which consists of ${DATE} in the
install_tools() function, we should find the latest buildtools-tarball based
on the modification time and install it.
[YOCTO #7674]
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Only poky sets SDK_NAME to include ${IMAGE_BASENAME} (i.e. ${PN}), so we
can't assume the buildtools filename will include it here. Change it to
look for a file with "buildtools-nativesdk-standalone" in the name
(the buildtools-tarball recipe itself sets TOOLCHAIN_OUTPUTNAME to
include this.)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Extensible SDK cannot be installed as root so by default offer to install it in
user's home directory under distro/distro_version replacing the normal SDK
version '+' char with a '_' as that's a restricted character for bitbake
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There are some characters that cannot appear in the installation path, so we
need to check for these
Signed-off-by: Brendan Le Foll <brendan.le.foll@intel.com>
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
In the current state if a SIGTERM is sent to
the testimage worker, the worker will exit but
runqemu and qemu won't exit and the processes
need to be killed manually to free the
bitbake lock.
This allows to catch the SIGTERM signal in
testimage, this way it is possible to stop
runqemu and qemu and allow to free the bitbake lock.
Also this allows to skip the rest of the tests
when running the tests in qemu or real hardware.
This also solves minimal breaks in the setup of the
runtime test when checking if qemu is alive.
[YOCTO #8239]
Signed-off-by: Benjamin Esquivel <benjamin.esquivel@linux.intel.com>
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
* useful when we need to overlay/extend intercept scripts from oe-core
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Pseudo 1.7 adds an experimental feature (which I think needs more testing
before it becomes the default) allowing the pseudo client to store modes
and uid/gid values in extended attributes rather than using the sqlite
database. On most Linux-like systems, this works only if the underlying
file is a plain file or a directory.
Also added is a profiling feature to allow some amount of reporting on
the wall-clock time the client spends in wrappers, processing operations,
or in IPC. This feature is not intendeded to be precisely accurate, but
gives a good overview of where time is going.
Based on the results from the profiling feature, the client now suppresses
OP_OPEN and OP_EXEC messages if the server is not logging messages, and
no longer uses constant dynamic allocation and free cycles for canonicalized
paths.
There's a few other likely-looking optimizations being considered, but
this seemed like a good cutoff for now.
1.7.1 fixes two bugs, one affecting mostly XFS systems with 64-bit
inode values, and one affecting code that called realpath(x, NULL), such
as the RPM backend.
1.7.2 fixes an indirect side-effect of the chmod fixes to deal with
umask 0700, which had no effect with opkg 0.2.4 but appears to cause
failures with 0.3.0.
1.7.3 prevents mkdirat() (and mkfifoat()) from setting errno on success,
because glibc's localedef inexplicably errors out if errno was set, even
if the operation's actual return code (which it tests) indicated
success.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
To provide some element of integrity to sstate archives, allow sstate archives
to be GPG signed with a specified key (detached signature to a sidecar .sig
file), and verify the signatures when sstate archives are unpacked.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Specify a value via PACKAGECONFIG[gui]. The default is
--diable-event-gui.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This changes the default SYSLINUX_ROOTFS such that the rootfs is no
longer expected under a fixed device path. Instead, the UUID is used
to find it. This makes the resulting .hdddirect (and thus also the
vdi/vdmk/qcow2 images derived from that) more flexible.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
As mentioned when introducing the VM images
(https://bugzilla.yoctoproject.org/show_bug.cgi?id=7374), the
resulting images only work when the image is mounted as a disk that
results in the hard-coded path (/dev/sda in the current
default). Using the file system UUID to find the rootfs is more
flexible.
To enable this for boot-direct.bbclass and thus image-vm.bbclass (aka
FSTYPEs vdi/vmdk/qcow2), set SYSLINUX_ROOT =
"root=UUID=<<uuid-of-rootfs>>". The rootfs image must use an ext file
system.
The special string will get replaced in the APPEND line with the
actual UUID when the boot loader (grub-efi, syslinux or gummiboot)
writes the boot loader configuration files. At that time, the rootfs
image has already been created and its UUID can be extracted using
"tune2fs -l", which also should be available because the
e2fsprogs-native tools were needed to create the image in the first
place.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Set INITRD_IMAGE to automatically add a custom initramfs to the boot
partition.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some of the SYSLINUX defaults are the same as in the underlying
syslinux.bbclass. Let's not duplicate them, because that makes
changing the default harder.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Kernel parameters like "uvesafb.mode_option=640x480-32" were turned
into shell variables named "bootparam_uvesafb.mode_option", which
triggered errors from the shell because the name is not valid. Now
points get replaced with underscores, leading to
bootparam_uvesafb_mode_option in this example.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some kernels, for example linux-yocto 3.19 for qemux86, fail to
execute /init in an initramfs unless there is already a /dev/console
char device in the initramfs. Booting then fails with:
Kernel panic - not syncing: /dev/console is missing or not a character device!
Please ensure your rootfs is properly configured
The panic itself comes from a linux-yocto specific patch to
kernel_init_freeable in init/main.c, but even without it, that
function will print an error when /dev/console is missing. The
kernel's Documentation/initrd.txt also mentions creating that device.
It remained unclear why this is not a problem on other machines. On
intel-corei7-64 from meta-intel, something (the kernel?) creates
/dev/console and /dev/[012] before transfering control to the init
script. In that case, creating /dev/console in advance is not
necessary, but does not cause any problem either.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some kernels, for example linux-yocto 3.19 for qemux86, fail to
execute /init in an initramfs unless there is already a /dev/console
char device in the initramfs. Booting then fails with:
Kernel panic - not syncing: /dev/console is missing or not a character device!
Please ensure your rootfs is properly configured
The panic itself comes from a linux-yocto specific patch to
kernel_init_freeable in init/main.c, but even without it, that
function will print an error when /dev/console is missing. The
kernel's Documentation/initrd.txt also mentions creating that device.
It remained unclear why this is not a problem on other machines. On
intel-corei7-64 from meta-intel, something (the kernel?) creates
/dev/console and /dev/[012] before transfering control to the init
script. In that case, creating /dev/console in advance is not
necessary, but does not cause any problem either.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
oprofileui uses gettext during the configuration task so should be inherit
gettext. This issue appears when an older version of gettext is used do to
pinning to the older non-gplv3 version.
[YOCTO #7795]
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When two or more tests had the same name but different classes then
the decorator log whould have the output all wrong. This was because
a comparison which was made only between method names but now it
compares classes too.
[YOCTO #8029]
Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It is not used by anything in oe-core and will be moved to meta-oe
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
* otherwise whole build report submission is rejected because it's too big
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
An rmdir call in do_install_append was trying to remove the
sbindir, however in a system with a merged usr this directory
is not empty and therefore failing to rm it causes an ERROR.
Instead check that sbindir != base_sbindir before trying to
remove the directory.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The Makefile for mount and osd_login utilities forces
/sbin as asbindir, however on a merged /usr system this directory
might not exist. Instead sed in the system sbindir.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When building with a merged /usr dir the udev directory
lives at /usr/lib/udev - update the FILES pattern to also
pick up udev files installed two levels below the / to
ensure a merged /usr works.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When building with a merged /usr dir the udev directory
lives at /usr/lib/udev - update the FILES pattern to also
pick up udev files installed two levels below the / to
ensure a merged /usr works.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When building with a merged /usr dir the udev directory
lives at /usr/lib/udev - update the FILES pattern to also
pick up udev files installed two levels below the / to
ensure a merged /usr works.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When building with a merged /usr dir the udev directory
lives at /usr/lib/udev - update the FILES pattern to also
pick up udev files installed two levels below the / to
ensure a merged /usr works.
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
