| Summary: | AB-INT: ruby do_compile: Bus Error | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Mathieu Dubois-Briand <mathieu.dubois-briand> |
| Component: | devtools / tool chain | Assignee: | Paul Barker <paul> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium | CC: | meta.mr.watcher, meta.watcher, paul, randy.macleod |
| Version: | unspecified | ||
| Target Milestone: | 6.1 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | AB-INT | ||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |
|
Description
Mathieu Dubois-Briand
2026-05-15 08:31:53 UTC
qemux86-world debian12-vk-7 mathieu/master-next&mathieu/master-next-tests completed at 2026-05-10 09:48:43+00:00 https://autobuilder.yoctoproject.org/valkyrie/#/builders/59/builds/3762/steps/12/logs/stdio could be lack of memory, let's see if it happens again. qemux86-world ubuntu2504-vk-1 master completed at 2026-07-10 03:05:25+00:00 https://autobuilder.yoctoproject.org/valkyrie/#/builders/59/builds/4134/steps/12/logs/stdio qemux86-world debian13-vk-1 mathieu/master-next completed at 2026-07-23 13:11:07+00:00 https://autobuilder.yoctoproject.org/valkyrie/#/builders/59/builds/4220/steps/12/logs/stdio "Bus Error" indicates SIGBUS was received, this is not an out-of-memory error. The build log from 2026-07-23 shows the failing address is 0x00007fee11f02765, inside the memory mapping of the file "/srv/pokybuild/yocto-worker/qemux86-world/build/build/tmp/work/x86-pokymllib32-linux/lib32-ruby/4.0.5/build/i686-linux-gnu-fake.rb". This is likely a race condition or missing dependency in ruby's Makefiles, leading to i686-linux-gnu-fake.rb being re-written by one process while another process is trying to parse it. Partial fix landed upstream: https://github.com/ruby/ruby/commit/bb235bddd4ce7d9e73a80931782f0f0b3971d12b Further testing showed this wasn't a complete fix, so I opened another PR: https://github.com/ruby/ruby/pull/18141 I'm pretty confident that this is now fully solved: - Before applying the above we have 3 instances of "generating i686-linux-gnu-fake.rb" in log.do_compile for ruby. - After applying the above we have just 1 instance :) Upstream are responsive so I'll wait until they've reviewed the second PR before I send patches for OE-core. Fix merged to OE-core in commit 074a5d9c4dd225811f22ea5552029c54bd1423cd. Hopefully that resolves the issue, reopen if the failure happens again. |