| Summary: | Tune file problems | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] Meta-yocto | Reporter: | Mark Hatle <mark.hatle> |
| Component: | configuration | Assignee: | Mark Hatle <mark.hatle> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | major | ||
| Priority: | High | CC: | jessica.zhang, poky.bs.watcher, poky.watcher |
| Version: | 1.2 | ||
| Target Milestone: | 1.2 M4 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | (1.2) | ||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | --- | |
|
Description
Mark Hatle
2012-04-03 15:22:32 UTC
Additional information: This started with a seemingly small PowerPC tuning fix from Chris Larson and my response to the patch... http://lists.linuxtogo.org/pipermail/openembedded-core/2012-March/020127.html The resulting discussion lead to: http://lists.linuxtogo.org/pipermail/openembedded-core/2012-March/020259.html Where the tunings were investigated and discovered to be at the best inconsistent, and worst very broken. The patch being work on adds the missing README files for the generic tuning and per-arch readmes. It also attempts to address various broken tunes as a result of the testing to resolve the PowerPC and other problems. It does -not- fix non-fatal issues, such as the various ARM architecture namings. (More details on the potential issues there once the patch has been posted for review.) Preliminary test results: All AVAILTUNES are able to build through core-image-minimal, except for: ARM: following fail to build due to toolchain issues armv5eb armv5e armv5eb-vfp armv5ehfb-vfp armv5ehf-vfp armv5e-vfp cortexm3 cortexr4 ep9312 MIPS64: not yet supported mips64 mips64el mips64el-n32 mips64el-nf mips64el-nf-n32 mips64-n32 mips64-nf mips64-nf-n32 x32: Need the layer to build core2-64-x32 x86-64-x32 (SH was existing, but didn't follow the tune format, it now does) SH: Still experimental -- SH3 requires a missing patch, "eb" not supported sh3 sh3eb sh5aeb sh4eb -- Multilibs are currently being tested -- Patch sent to oe-core mailing list. Multilibs worked, only failures were in individual tunings. See the following for more info: http://lists.linuxtogo.org/pipermail/openembedded-core/2012-April/020497.html Patches have been accepted into oe-core: http://git.openembedded.org/openembedded-core/commit/?id=603a15bf4c838e4b6352e31f70a958d93f91138f http://git.openembedded.org/openembedded-core/commit/?id=0e71abea5458122188d5eddef2c17147f61ff895 http://git.openembedded.org/openembedded-core/commit/?id=f81f71bcff4bb1032b034b068efe6065113ca9e7 http://git.openembedded.org/openembedded-core/commit/?id=efbfa2ace3362393a20340af93e8dcab17a8619a http://git.openembedded.org/openembedded-core/commit/?id=ab77d3401908964f3249c761969600b5ec1bfbd0 http://git.openembedded.org/openembedded-core/commit/?id=e138f9f7e48e0af94c5c88045c4f0581cc68248d |