If the script files we use to construct the SDK installer change then
that really ought to trigger re-execution of the do_populate_sdk(_ext)
task, so add file-checksums varflags to ensure that happens.
(From OE-Core rev: 59fc4bdefb3853d6893b9e3792018d54bb6bf99f)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
At the end of the extensible SDK installation, if we've successfully
prepared the build system then we don't need ext-sdk-prepare.py. I had
thought earlier that this would be used when updating, but a different
mechanism was needed there so this script isn't used for that.
(From OE-Core rev: 037ea78b0c3de01f87bbe06a1b88a162b2b40496)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add the ability to install additional pre-built items (from shared
state) into the extensible SDK. This can already be done implicitly by
adding something to DEPENDS within a recipe you're working on and then
running "devtool build", but it's useful to be able to explicitly
install things particularly if you're using the extensible SDK as a
traditional toolchain.
Note that for this command to be useful you need to have SSTATE_MIRRORS
set in your SDK configuration, and that mirror needs to be populated
with sstate artifacts for recipes you wish to be able to install.
(From OE-Core rev: 3474a42954908d1688fd3a6cb600eed315b27833)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Make the following improvements to the SDK update process:
* Use a manifest file with sha256sums to track files other than sstate
and metadata that we need to update - e.g. conf files. This allows us
to handle where files such as auto.conf may or may not be present,
as well as the configuration changing without affecting task signatures
- we still want the config files copied in that case rather than it
saying nothing needs to be done.
* Write the SSTATE_MIRRORS_append to site.conf rather than local.conf
so that local.conf remains static (since we don't want to trigger an
update every time). Also, If there is an SSTATE_MIRRORS value already
set in the configuration we can skip this and assume it contains the
needed packages.
* Allow the update process to be run in any directory, don't assume
we're already at the base of the SDK
* Where practical, fetch remote files into a temporary location and
then move them to the desired location at the end, to avoid a
failed update leaving the SDK in a broken state.
* Update all installed do_populate_sysroot / do_packagedata tasks
instead of using the SDK targets. This ensures any item installed
through dependencies after installation (e.g. when running
"devtool build") won't go stale.
(From OE-Core rev: 3d35631121f0e030bc8151f5c23d84008d06f44b)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* Use tempfile.mkdtemp() instead of hardcoding temp dir
* Set a variable early for the temp locked sigs file and use that
everywhere
* Delete the temp dir at the end
(From OE-Core rev: bad5d1a8c047a8118d30d9fa708b021d1599e0dc)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When you run devtool build, you need to have the pkgdata written out at
the end, so that if what you're adding is a library and the next thing
you add is something that depends on that library, the necessary
information to map the dependency back to the recipe is present. In
practical terms all this means is we need do_packagedata to run in
addition to do_populate_sysroot.
This does mean that do_package needs to run which wasn't running before,
and that means that the few package QA tests that run within do_package
such as installed-vs-shipped will now be run. This may be a bit
bothersome, and prompted a fix for one of our oe-selftest tests as a
result, but I don't see an easy way around it. Ultimately if you care
about using the recipe in an image you'll need to fix any such errors
anyway.
Fixes [YOCTO #8887].
(From OE-Core rev: 6579c7120ee5a541427ff5b6b07f838d52f9fe7c)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a variable SDK_INCLUDE_PKGDATA which you can set to "1" to include
pkgdata for all recipes in the world target. There are a couple of uses
for this:
1) If you use "devtool add" to add a recipe that builds something which
depends on anything in world, the dependency can then be correctly
mapped to the recipe providing it and that recipe can be added to
DEPENDS, since we have the pkg-config and shared library dependency
data within pkgdata.
2) You'll be able to search for these recipes and any files they
package for the target with "devtool search" since that also uses
pkgdata
This of course assumes you've tailored world through EXCLUDE_FROM_WORLD
to only include recipes you'd want built in your distro, but I think
that's a reasonable assumption; failing that there is a
WORLD_PKGDATA_EXCLUDE variable that you can set to exclude any recipes
you don't want.
Note that this patch relies on functionality implemented in a recent
BitBake patch and will not work without it.
Implements [YOCTO #8600].
(From OE-Core rev: 67149ea097d6fab7496b43e85a40853f40bd527e)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These patches have revealed issues with linux kernel headers which has
been proposed upstream but hasnt been applied yet
(From OE-Core rev: 8d760bd9df101a03f148205e7d79b555b4dfd8c7)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
gettext.m4 macro does not detect musl triplets and falls back to
${libdir}/locale
It ends up with lot of packging warning
"files installed but not packaged"
(From OE-Core rev: 6fae6a92cf3d9810dc68438f3b20913453812437)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Clang finds overflows when comparison is done between an unsigned char
and a integer constant. So explicitly typecast the constant before
comparison
(From OE-Core rev: 2a18273fc74c6493e3d34499a8774e237772f109)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Define _POSIX_C_SOURCE if not defined
Add sockaddr_un* to sockaddr* typecast to make compiler happy
(From OE-Core rev: cdb858caadd81d3bdc9ea3fd596e3eab84ec2d81)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Configure looks for error_t in errno.h and musl does not define error_t
if its not defined then configure emits a definition to be used, now we
also use standalone argp implementation from glibc on musl based systems
which defines this in argp.h as well so now we have it defined twice
and errors like
|
/a/builder/mnt/build-oe/tmp-musl/sysroots/raspberrypi2/usr/include/argp.h:40:13:
error: cannot combine with previous 'int' declaration specifier
| typedef int error_t;
| ^
| ../config.h:257:17: note: expanded from macro 'error_t'
| #define error_t int
| ^
| 1 error generated.
are seen. argp.h definition is conditional on __error_t_defined which is
true here so lets define it on musl
(From OE-Core rev: 29e1edcba39a21057ba5862eef4e79c177e7ef62)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It did differentiate on features in libc but assumed always glibc
this patch make that assumption clear
(From OE-Core rev: c360290f22abfe9e37ca2fc0766b5a08a3276f6b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
changes are here
http://mailman.uclibc-ng.org/pipermail/devel/2016-January/000665.html
Drop upstreamed patch
(From OE-Core rev: 1553adb92bfdcfda9c216a72658ccf327b6ab1aa)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
strol_l is not available on musl
delete charset.alias as well its not needed
Use internal gc function on musl
We get errors on certain functions not being available in boeheme gc
when built with musl. Therefore use the internal versions
e.g.
Undefined gc_set_finalizer_notifier
(From OE-Core rev: 31ee3460c3a0578ed36f3f5a940617ccb873c721)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Its used in several packages
(From OE-Core rev: 8fea65047519a4d84183571e09d3f3f559d9e164)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We have rootfs intercepts which fail on musl because ld.so on musl is an
absolute symlink to /usr/lib/libc.so and then it fails to load when run
with qemu user mode.
Fixes rootfs failures like
WARNING: The postinstall intercept hook 'update_pixbuf_cache' failed
(exit code: 255)! See log for details!
WARNING: The postinstalls for the following packages will be postponed
for first boot: libgdk-pixbuf-2.0-loader-jpeg li
bgdk-pixbuf-2.0-loader-png libgdk-pixbuf-2.0-loader-gif
libgdk-pixbuf-2.0-loader-xpm
WARNING: [log_check] core-image-sato: found a warning message in the
logfile (keyword 'WARNING:'):
[log_check] WARNING: The postinstall intercept hook
'update_pixbuf_cache' failed (exit code: 255)! See log for details!
WARNING: [log_check] core-image-sato: found a warning message in the
logfile (keyword 'WARNING:'):
[log_check] WARNING: The postinstalls for the following packages will be
postponed for first boot: libgdk-pixbuf-2.0-lo
ader-jpeg libgdk-pixbuf-2.0-loader-png libgdk-pixbuf-2.0-loader-gif
libgdk-pixbuf-2.0-loader-xpm
(From OE-Core rev: d48ea3354eff8244e1734fa571bb52c810431b3f)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
pass --hash-style explicitly to linker
and ensure that mips does not use gnu hash style
This inoculates fts when compiler does not pass
hash style option e.g. clang
(From OE-Core rev: a40532a5277e1c0392032ff2d4992046477fadf0)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
enable largefile support if its in DISTRO_FEATURES
(From OE-Core rev: 300defd71205bf350a6e956451e0cc87c4e92902)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
nss's build system assumes that cdefs.h is always available on linux
which is not the case with musl
(From OE-Core rev: c4a5a8c4a6dbdcf735024aaee9e36a7a7b56cb96)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
parallel build patch was removing dependencies on .c files mistakenly
just adding src/yacc_config.h to dependencies in existing rule should
have fixed the original build race
include sys/types.h in lex_config.l for getting u_long definition
(From OE-Core rev: 149a5202d32650775386e166ed06855097256977)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
systemd on musl doesn't yet work even though we have patches to make it
compile it fails to run, therefore lets skip building it for now
(From OE-Core rev: a1986acf66381dee18f5c8deae7cf52490d0f58a)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
let musl depend on these headers so they are staged along with libc
(From OE-Core rev: 2cb184e4abaa69faad9f94631a3906188aa78e96)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This was not only getting included for libc5 but also for not-glibc libc
implementations e.g. musl
(From OE-Core rev: 9d0fed7d3b2e70bc5d6bc2b0e1cc38166df3f0d2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix build with musl
uclibc and glibc dont agree on structure of the struct, musl rightly
adds padding elements, so when doing anonymous initialization struct
elements gets wrongly mapped on 64bit arches
(From OE-Core rev: 3c54b18797eb26d2045fd506d2b0c8b996a0396c)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There are several BSD specific assumptions
and some GNU specific ones.
(From OE-Core rev: 49e8fc6c4511e026dee176422cf3a7b7558c7cfc)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It only considered linux-gnu hosts when cross compiling
here we add linux-musl to the mix as well
Fixes errors e.g.
1.28-r0/tar-1.28/src/tar.c:1351:5: error: 'SAVEDIR_SORT_INODE'
undeclared here (not in a function)
| SAVEDIR_SORT_INODE
| ^
(From OE-Core rev: c17d9a8d7f89b8e855f87d61583075129e4aa72c)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Makery hard sets CC to be some form of gcc, make it set only if CC
variable is absent in environment.
(From OE-Core rev: 8d83c6ad78e1f965a55c0d3f430bad2781a3eb33)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
backtrace is glibc only feature on linux so check for glibc before using
it
Clangs clear_cache has different signature than gcc provided builtin
make it compatible for both
(From OE-Core rev: c3dc346d44c8c5485b3eb1a97f32ba2ed2e76ece)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
wait is not union per posix it is int
remove assumption about glibc is linux
(From OE-Core rev: a590e541338d49130aaae5339856d31badd56719)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Make backtrace() API's use conditional on glibc
this helps make libunwind compile on musl
(From OE-Core rev: acffe22de65fad2f69c702e4fe94b366883a265b)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Turn bzip2 support into a PACKAGECONFIG
zlibs is a must have and therefore add it to DEPENDS will make the build
consistent
(From OE-Core rev: ded03dab901647a7388d219a3a228d4f4f4298e8)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes build failures with musl where gcc5 flags it used before
initialized.
(From OE-Core rev: dda96dbb3e4cd83fe6b8fd8370c8ca2a737cc078)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton
Khem Raj
Ross Burton