| Summary: | [multilib] Finish core-image-sato-sdk multilib enabling | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] Meta-yocto | Reporter: | Dongxiao Xu <dongxiao.xu> |
| Component: | meta-yocto | Assignee: | Bogdan Marinescu <bogdan.a.marinescu> |
| Status: | VERIFIED WORKSFORME | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium+ | CC: | alexandru.c.georgescu, alexandru.damian, dongxiao.xu, jiajun.xu, mihai.lindner, poky.bs.watcher, poky.watcher, sgw, shane.wang, song.liu |
| Version: | unspecified | ||
| Target Milestone: | 1.3 M4 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | (scheduled) Depends on Jiajun's full test | ||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 1563 | ||
|
Description
Dongxiao Xu
2011-11-03 23:07:16 UTC
I am going to delay multilib new features (satosdk and world) into M3, since we don't have more time on multilib development. I have passed core-image-sato-sdk building and booting. But early this week, qt4-x11-free build fail. Fix is on going. I noticed sato-sdk was enabled but only qt broke now. I will open another bug for qt breaks. Mark the bug as "QA Testing Completed" since we already cover the testing for the feature in 1.2 M4. Edwin/Shane, Is this already completed in 1.2 with the bug for broken qt opened? If so, please mark this verified and change the milestone to 1.2. Otherwise, let's keep it this way for 1.3. (In reply to comment #5) > Edwin/Shane, > > Is this already completed in 1.2 with the bug for broken qt opened? If so, > please mark this verified and change the milestone to 1.2. Otherwise, let's > keep it this way for 1.3. Song, In fact, many commits in master break it from time to time. When I made core-image-sato-sdk build and run well in 1.2 development phase, but it got broken later. Dongxiao test and make sure it works well recently. I suggest to close it and open new bug if needed in future. Dongxiao, Do you agree? (In reply to comment #6) Song, In fact, many commits in master break it from time to time. When > I made core-image-sato-sdk build and run well in 1.2 development phase, but > it got broken later. Dongxiao test and make sure it works well recently. I > suggest to close it and open new bug if needed in future. Dongxiao, Do you > agree? I ever did a brief test for the core-image-sato-sdk multilib build and it passed. However we need QA's full tests to make sure it indeed works. I think Jiajun will perform the full multilib functionality test recently, and we can close the bug at that time. What do you think, Edwin and Song? (In reply to comment #7) > > I ever did a brief test for the core-image-sato-sdk multilib build and it > passed. However we need QA's full tests to make sure it indeed works. I > think Jiajun will perform the full multilib functionality test recently, and > we can close the bug at that time. > > What do you think, Edwin and Song? I agree. Song, I move it to M3 since we put https://bugzilla.yoctoproject.org/show_bug.cgi?id=2038 ahead from M3 to M2. Tried bitbake core-image-sato-sdk Build Configuration: BB_VERSION = "1.15.3" TARGET_ARCH = "x86_64" TARGET_OS = "linux" MACHINE = "qemux86-64" DISTRO = "poky" DISTRO_VERSION = "1.2+snapshot-20120830" TUNE_FEATURES = "m64" TARGET_FPU = "" meta meta-yocto = "master:b4c5725af4cd85d5644f0373e2674e903c4eab2b" I had the following in my conf/local.conf MACHINE ??= "qemux86-64" require conf/multilib.conf MULTILIBS = "multilib:lib32" DEFAULTTUNE_virtclass-multilib-lib32 = "x86" The build finishes with some QA Issues but no errors. Please let me know if I missed something, the multilib wiki page seems a little outdated, Bogdan submitted a few changes. tried bitbake lib32-core-image-sato-sdk as well, but it crashes at do_rootfs | Processing task-core-x11-base... | Processing kernel-dev... | Multiple matches for kernel-dev: | unable to find package kernel-dev | ERROR: some packages were missing | ERROR: Function failed: do_rootfs (see /home/radu/Documents/Development/yocto/build/tmp/work/qemux86_64-poky-linux/core-image-sato-sdk-1.0-r0/temp/log.do_rootfs.11604 for further information) ERROR: Task 7 (/home/radu/Documents/Development/yocto/poky/meta/recipes-sato/images/core-image-sato-sdk.bb, do_rootfs) failed with exit code '1' NOTE: Tasks Summary: Attempted 6229 tasks of which 5056 didn't need to be rerun and 1 failed. same error I get when building for quemux86-64 without multilib enabled, so I guess is not multilib related. Will continue to keep an eye on this. the correct log, the previous one was with multilib disabled. | Processing lib32-task-core-x11-base... | Processing lib32-kernel-dev... | unable to find package kernel-dev | ERROR: some packages were missing | ERROR: Function failed: do_rootfs (see /home/radu/Documents/Development/yocto2/build/tmp/work/qemux86_64-poky-linux/lib32-core-image-sato-sdk-1.0-r0/temp/log.do_rootfs.21128 for further information) ERROR: Task 7 (virtual:multilib:lib32:/home/radu/Documents/Development/yocto2/poky/meta/recipes-sato/images/core-image-sato-sdk.bb, do_rootfs) failed with exit code '1' NOTE: Tasks Summary: Attempted 6447 tasks of which 1545 didn't need to be rerun and 1 failed. lib32-core-image-* has been building fine for me for quite a while now, so I'm closing this bug. What Radu has experienced is most likely a temporary issue with Yocto, not a multilib problem. 3168 must be fixed in order to verify this. since 3168 is duplicate of 2918, then 2918 have to be fixed first in order to verify this bug. (In reply to comment #13) > lib32-core-image-* has been building fine for me for quite a while now, so > I'm closing this bug. What Radu has experienced is most likely a temporary > issue with Yocto, not a multilib problem. Are we sure that the resolution for this is FIXED? I don't see a commit for this. Maybe it's WORKSFORME or NEEDINFO (to establish the steps to produce _this_ issue)? Either way, with the current resolution, this issue cannot go towards a VERIFIED status. Thanks. (In reply to comment #13) > lib32-core-image-* has been building fine for me for quite a while now, so > I'm closing this bug. What Radu has experienced is most likely a temporary > issue with Yocto, not a multilib problem. |