linux/poky.git/meta/classes/archiver.bbclass, branch yocto-3.1.28

archiver: don't use machine variables in shared recipes

2022-06-22T22:46:32+00:00

When using multiconfig with the same TMP folder we can have races because the shared recipes like gcc-source run twice. ARCHIVER_OUTDIR = ${ARCHIVER_TOPDIR}/${TARGET_SYS}/${PF}/ which includes TARGET_SYS and between the two different MACHINE values, this changes from 'arm-poky-linux-gnueabi' to 'aarch64-poky-linux'. This leads to the task running twice, once for each multiconfig. To solve this we need to store the shared output in a common place for all machines and in this way the stamps will be the same for each machine so the gcc-source will on run once regardless of the machine used. (From OE-Core rev: 135adeb82c9303c26193bb6f6bd3bc696793aa62) Signed-off-by: Jose Quaresma Signed-off-by: Alexandre Belloni Signed-off-by: Richard Purdie (cherry picked from commit 5abe497aad39a6ce8d72556fcdda1938a0f8c1bc) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

archiver: use bb.note instead of echo

2022-06-22T22:46:32+00:00

(From OE-Core rev: fd1ed873ec00fda2f58370770ad7aed039bdc470) Signed-off-by: Jose Quaresma Signed-off-by: Luca Ceresoli Signed-off-by: Richard Purdie (cherry picked from commit 6420c8a6a8143f53ccad7ab2d56b2ba06db83099) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

archiver.bbclass: fix do_ar_configured failure for kernel

2021-08-10T10:14:11+00:00

(From OE-Core rev: 224913ff067e6c26cabe0df0ecbf50fb03b85165) Signed-off-by: Changqing Li Signed-off-by: Alexandre Belloni Signed-off-by: Richard Purdie (cherry picked from commit d7776a23cbea836ddb8ac5ec77012af2449ab875) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

archiver.bbclass: Fix patch error for recipes that inherit dos2unix.

2021-08-10T10:14:11+00:00

do_unpack_and_patch error happens for these recipes inherit dos2unix. (From OE-Core rev: 3e8cb257546e00a988136dfa6d7f5e6b7f7f16d8) Signed-off-by: Lei Maohui Signed-off-by: Alexandre Belloni Signed-off-by: Richard Purdie (cherry picked from commit 2ceda7c90c0087f52693c54d5ccab143b27f4d21) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

base/bitbake.conf: Enable pseudo path filtering

2021-02-03T14:13:53+00:00

This is a pretty big change to the way pseudo operates when used in OpenEmbedded. Normally, pseudo monitors and logs (adds to its database) any file created or modified whilst in a fakeroot environment. There are large numbers of files we simply don't care about the permissions of whilst in that fakeroot context, for example ${S}, ${B}, ${T}, ${SSTATE_DIR}, the central sstate control directories, This change uses new functionality in pseudo to ignore these directory trees, resulting in a cleaner database with less chance of "stray" mismatches if files are modified outside pseudo context. It also should reduce some overhead from pseudo as the interprocess round trip to the server is avoided. There is a possible complication where some existing recipe may break, for example, we found a recipe which was writing to "${B}/install" for "make install" in do_install and since we listed ${B} as not to be tracked, there were errors trying to chown root for files in this location. This patch fixes a few corner cases in OE-Core when used with this new ignore list: * The archiver directory matched a "${WORKDIR}/deploy*" pattern so was renamed to something else since that directory does need its root permissions * The ${S} and ${B} ignoring is conditional on them being different to ${WORKDIR} * package_write_* task output (the debs/rpms/ipks) are now owned by the build user so we don't want the file ownership information in the hashequiv outhash calculation even if they are built under pseudo. * The fontcache postinstall intercept is run under qemu outside of pseudo context so delete files it may delete up front where pseudo can see this. * SSTATE_DIR is in PSEUDO_PATHS_IGNORE, which is in FAKEROOTENV which is cached by bitbake. We therefore need to trigger reparsing if this changes, which means SSTATE_DIR can be in BB_HASHBASE_WHITELIST but not BB_HASHCONFIG_WHITELIST. Rework the variables to handle this. This otherwise breaks some of our sstate tests in oe-selftest. * Ignore the temp directory wic uses for rebuilding rootfs. (From OE-Core rev: 61f8cf0940f8b05e9399a062eddb8055ea69bc5e) Signed-off-by: Richard Purdie (cherry picked from commit ad8f5532ffaead9a5ad13e1034fe9e5e1b7979f4) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

archiver.bbclass: Fix --runall=deploy_archives for images

2021-01-09T09:17:16+00:00

Fixed: INHERIT += "archiver" COPYLEFT_LICENSE_INCLUDE = "*" $ bitbake core-image-minimal --runall=deploy_archives [snip] KeyError: '/path/to/meta/recipes-core/images/core-image-minimal.bb:do_ar_original' [snip] (From OE-Core rev: 1625481cdfc6560bc8e49aea7f7e9175cbaaa15e) Signed-off-by: Robert Yang Signed-off-by: Richard Purdie (cherry picked from commit 59785a51110c450c7629218f6042f1d9d309618e) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

classes/archive: do_configure should not depend on do_ar_patched

2020-07-16T21:28:40+00:00

The commit d4be264061 ("classes/archiver: run do_unpack_and_patch after do_preconfigure") correctly moved do_unpack_and_patch to depend on do_preconfigure, but left do_ar_patched as a dependency of do_configure introduced from 8a7c779487 ("classes/archiver: Create patched archive before configuring"). Having do_configure depend on do_ar_patched is blatantly wrong. Firstly, doing so causes the taskhash of do_configure to change which is undesirable. Secondly, the anon python that sets up the tasks carefully skips GCC recipes that depend on gcc-source to provide their source code since running do_unpack_and_patch in them would delete the source code. Adding in the dependency effectively bypasses this and makes these recipes delete the shared gcc-source This fixes errors (for real this time) like: cat: .../gcc/defaults.h: No such file or directory when making certain configuration changes. For posterity, you could easily reproduce gcc-source being erased by running the commands (replacing aarch64 with your correct target arch): bitbake -c cleanall gcc-source gcc-cross-aarch64 bitbake -c do_deploy_archives gcc-source # The following would fail because # gcc-cross-aarch64:do_unpack_and_patch erases shared source, so # do_configure after that fails. bitbake gcc-cross-aarch64 (From OE-Core rev: 2399fdf98a6e6e4245c2e7ed1f72404ebece3eca) Signed-off-by: Joshua Watt Signed-off-by: Richard Purdie (cherry picked from commit f02c1904330da47a25cf85f0077712f584f4ee5b) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

classes/archiver: run do_unpack_and_patch after do_preconfigure

2020-07-16T21:28:40+00:00

The commit 8a7c779487 ("classes/archiver: Create patched archive before configuring") fixed a race condition when using the archiver, but introduced a bug where the GCC source would occasionally be deleted due to the archiving code running before do_preconfigure. Instead, make sure the archiving code runs after do_preconfigure by making do_unpack_and_patch depend on it. This makes more sense anyway since do_preconfigure is effectively an extension of do_patch. This fixes errors like: cat: .../gcc/defaults.h: No such file or directory when making certain configuration changes (From OE-Core rev: b6d4176f82b9728af91a4b982ca25fbbee5f9890) Signed-off-by: Joshua Watt Signed-off-by: Richard Purdie (cherry picked from commit d4be264061023ae0ea6e023d82a7b99605bfc22c) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

classes/archiver: Create patched archive before configuring

2020-06-26T17:26:49+00:00

do_configure and do_preconfigure can modify source files, which causes race conditions if these tasks run in parallel with do_ar_patched. Add explicit task dependencies to ensure that do_ar_patched finishes before these tasks start. Specifically, this fixes a race condition with gcc-source where do_ar_patched races with do_preconfigure deleting gcc/gengtype-lex.c (From OE-Core rev: fdd06278dc67e8c13468b89746c562d298a2d63b) Signed-off-by: Joshua Watt Signed-off-by: Richard Purdie (cherry picked from commit 8a7c7794870815030239e90b03e37ed302b7e885) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie

archiver.bbclass: Fix duplicated SRC_URIs for do_ar_original

2020-06-05T20:36:30+00:00

The argument urls of bb.fetch2.Fetch(urls, d) are duplicated to SRC_URI, which caused errors like: bb.data_smart.ExpansionError: Failure expanding variable SRCPV, expression was ${@bb.fetch2.get_srcrev(d)} which triggered exception FetchError: Fetcher failure: The SRCREV_FORMAT variable must be set when multiple SCMs are used. The SCMs are: git://github.com/docker/notary.git;destsuffix=git/src/github.com/docker/notary git://github.com/docker/notary.git The first one is from original SRC_URI, the second one is from the variable 'urls', so cleanup SRC_URI before call bb.fetch2.Fetch() can fix the problem. (From OE-Core rev: a7f50876f95a9be9fe045af1e4efddfe53a983f5) Signed-off-by: Robert Yang Signed-off-by: Richard Purdie (cherry picked from commit b244c4f3427cd07376d4b8f7d27e38735bcc90e7) Signed-off-by: Steve Sakoman Signed-off-by: Richard Purdie