linux/poky.git/meta/classes, branch 5.2_M2

create-spdx: support line numbers

2025-01-23T12:14:29+00:00

LIC_FILES_CHKSUM supports begin-/endline for licenses included in for instance header files. This patch adds support for line numbers to NO_GENERIC_LICENSE, too. (From OE-Core rev: 8e7ee19fc9e74cf042880f4bc317782482ba6f66) Signed-off-by: Denis Osterland-Heim Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Ross Burton

base: Switch virtual/cross-XXX to be under recipe specific providers

2025-01-21T23:09:34+00:00

Currently, providers are set on a global config basis. This change allows for a select set of providers to be configured using BB_RECIPE_VIRTUAL_PROVIDERS on a per recipe basis. This would allow for the selection of virtual/cross-cc as gcc or clang for example. The PROVIDERS are removed from the recipes so that if a version of the dependency accidentally slips through, the build will fail and the user can correct the issue. (From OE-Core rev: 6eeab1a5d7f23917b94c130e417d59afb757b546) Signed-off-by: Richard Purdie

go-vendor: Fix absolute paths issue

2025-01-08T12:37:20+00:00

When building with the class, avoid errors like: "sstate found an absolute path symlink /xxx/vendor pointing at /xxx/import/vendor. Please replace this with a relative link." Use relative symlinks instead. (From OE-Core rev: fc4f589cc7a8e5feba7940ccb244a74bd5f96371) Signed-off-by: Richard Purdie

report-error: Add PN to error report file

2024-12-28T10:57:00+00:00

Extracting the PN out of the PF value is not allways trivial. For example when PE is set the package name will have the PE value before the '_' that clearly makes the division between name and version. Other packages include a number in their naming separated by a '-', for example mozjs-115 and openjdk-17. (From OE-Core rev: a2d54fa0c5ade523edf6be9105c6e943824988fa) Signed-off-by: Angelo Ribeiro Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie

icecc: don't use KERNEL_CC during set_icecc_env

2024-12-13T12:26:01+00:00

The KERNEL_CC variable usage was introduced by commit d2fcaeb153fd ("icecc: Don't generate recipe-sysroot symlinks at recipe-parsing time, 2018-12-20") and is set once a recipe pulls the kernel-arch.bbclass. There are a few packages, e.g. u-boot, barebox, devicetree, perf which pull this class and for such packages the get_cross_kernel_cc() return None because icecc_is_kernel() fail. Fix this for now by replacing the KERNEL_CC check with the icecc_is_kernel() helper, which is also used by the get_cross_kernel_cc(). That said it's still not very robust to determine a compiler based on a variable or inherited class but at least it fixes the issue for now. (From OE-Core rev: 5e3c6a6384fcdc4a2269172012caae14e5f327ca) Signed-off-by: Marco Felsch Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie

icecc: convert set_icecc_env to python prefuncs

2024-12-13T12:26:01+00:00

Since bitbake commit f24bbaaddb36 ("data: Add support for new BB_HASH_CODEPARSER_VALS for cache optimisation") the icecc fails with [1]: ERROR: /Yocto/poky/meta/recipes-core/meta/target-sdk-provides-dummy.bb: no-pn NULL prefix WARNING: /Yocto/poky/meta/recipes-core/meta/target-sdk-provides-dummy.bb: Exception during build_dependencies for set_icecc_env The reason for this is the bb.fatal() within the icecc_version(). icecc_version() is called during the "${@}" python variable expansion while bitbake is running the build_dependencies() for the set_icecc_env() function. To avoid this behaviour set_icecc_env() should be converted into a python function which gets called during task[prefuncs] [2], which is done by this commit. [1] https://lists.yoctoproject.org/g/yocto/topic/icecc_support_broken/103429714 [2] https://lists.openembedded.org/g/openembedded-core/topic/110009272 (From OE-Core rev: 444445c5793aaf831ff0293b62a000f8ab7d40bb) Signed-off-by: Marco Felsch Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie

classes: create-spdx: Fix variable dependencies

2024-12-05T17:07:10+00:00

A few variable dependencies were incorrect from the SPDX tasks, which prevented rebuilds from occurring, or caused them to occur when they should not. (From OE-Core rev: 38a5dd136b3a3713e63d23598011efe1c4b0f459) Signed-off-by: Joshua Watt Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie

yocto-check-layer: expand to cover all required QA checks

2024-12-03T11:42:35+00:00

insane.bbclass now defines CHECKLAYER_REQUIRED_TESTS list with required QA checks that are becoming mandatory for Yocto Project Compatible layers. Update yocto-check-layer.bbclass in order to catch when packages from such layers try to skip any of the required QA checks. (From OE-Core rev: 9c3ba88628853b20fb4c98c99cf3fe8349024016) Signed-off-by: Denys Dmytriyenko Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie

buildhistory: Drop git-gc call as obsolete

2024-12-03T11:21:34+00:00

buildhistory used to result in many individual objects in the repo which could affect performance. This gc call was added over a decade ago to avoid those performance issues. Modern git calls git-gc when needed and this kind of workaround should no longer be needed. There is a small chance this is contributing to problems with the autobuilder buildhistory repositories, so simplifying the code may help that too. (From OE-Core rev: 0a88f9109f7670a6b188463a19392d445af6ff40) Signed-off-by: Richard Purdie

cve-check: fix cvesInRecord

2024-11-28T12:27:12+00:00

Currently flag cvesInRecord is set to false if all CVEs are ignored or patched. This is inconsistent as it shows false if a CVE was fixed via patch and true if this CVE was fixed by upgrade. In both cases the CVE is valid and was fixed. As I understand this flag, it should say if any CVE exists for particular component's product (regardless of how this CVE is handled) and can be used to validate if a product is correctly set. Note that skipping ignored CVEs may make sense in some cases, as ignored may mean that NVD DB is wrong, but in many cases it is ignored for other reasons. Further patch can be done to evaluate ignore subtype but that would be against my understanding of this flag as described above. (From OE-Core rev: c5d499693672ec9619392011b765941cf94aa319) Signed-off-by: Peter Marko Signed-off-by: Mathieu Dubois-Briand Signed-off-by: Richard Purdie