| Summary: | AB-INT: gcc 12.3.1 segfault while building ccache - fedora 37 specific | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Alexandre Belloni <alexandre.belloni> |
| Component: | devtools / tool chain | Assignee: | Naveen Gowda <Naveen.Gowda> |
| Status: | RESOLVED OBSOLETE | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium+ | CC: | alexandre.belloni, luca.ceresoli, meta.mr.watcher, meta.watcher, randy.macleod, richard.purdie, ross.burton, sundeep.kokkonda |
| Version: | unspecified | ||
| Target Milestone: | 4.3 M4 | ||
| 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
Alexandre Belloni
2023-05-18 19:53:41 UTC
/tmp/ccCNaot3.out and build-st-2675876 are saved in bug-15123 on fedora37-ty-2 Please report to Fedora and see if they can fix it since it's a host gcc bug. The "good" news is that this replicates on demand. https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5244/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5432/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5338/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5346/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5490/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5314/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5325/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5537/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5542/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5643/steps/14/logs/stdio oe-selftest-fedora on fedora37-ty-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/86/builds/5726/steps/14/logs/stdio oe-selftest-fedora fedora37-ty-3 The issue has been fixed in gcc trunk and gcc-12 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109241 Master branch has gcc-13 and the patch is present in yocto source. Master branch has gcc-13 and the patch is present in yocto source. This is building ccache-native using gcc from the host system so the Yocto Project version of gcc won't help here unfortunately Thanks very much for the clarification. I am trying to reproduce the issue on Fedora-38 without much luck. Linux yow-fedora-builder1 6.5.7-200.fc38.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Oct 11 04:07:58 UTC 2023 x86_64 GNU/Linux Its using the gcc-13 which probably has the fix /usr/bin/gcc --version gcc (GCC) 13.2.1 20231011 (Red Hat 13.2.1-4) As I understand, the solution to the issue is upgrading host gcc to latest version gcc-13 which has the patch. Can you please let me know if its right? Thanks Naveen, no the solution has to only involve changes in bitbake/oe-core or sending a patch to Fedora (37) to fix this issue. Have you tried to reproduce the issue in a F37 docker container? I has never been seen in Fedora 38. F37 is almost at EOL, based on this trend: https://docs.fedoraproject.org/en-US/releases/eol/ but it's possible that it would be the basis of RHEL 10: https://docs.fedoraproject.org/en-US/quick-docs/fedora-and-red-hat-enterprise-linux/index.html See if you can make progress using a F37 container but otherwise, it may be best to just wait this out until the F37 EOL date and resolve the issue as obsolete. Hi Randy, Thanks for the detailed clarification about the status and issue. Naveen Fedora 37 is EOL and this bug was only seen in that distro. |