| Summary: | CLANG 3.5.0/Kernel Compile Failure | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Base System | Reporter: | Larry Rosenman <ler> | ||||
| Component: | bin | Assignee: | Dimitry Andric <dim> | ||||
| Status: | Closed Unable to Reproduce | ||||||
| Severity: | Affects Only Me | CC: | dim, emaste | ||||
| Priority: | --- | ||||||
| Version: | CURRENT | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
|
Description
Larry Rosenman
2014-11-29 16:53:34 UTC
the .c file is too large to attach, it's at: http://www.lerctr.org/~ler/OCE The file oce_if-216294.c compiles just fine here, using oce_if-216294.sh, both with vanilla clang 3.5.0 (e.g. exactly how llvm.org released it), and patched clang 3.5.0 (as it was in FreeBSD projects/clang350-import r275161). I see no assertion failures. If you repeat the kernel build, does the error occur in the same file each time? If the error keeps occurring in the same file, please rebuild clang with debug information, and generate a backtrace from the core dump. I redid the exact buildkernel, and no failure this time. So. I'm not sure what to do here to help you.... (In reply to Larry Rosenman from comment #3) > I redid the exact buildkernel, and no failure this time. > > So. I'm not sure what to do here to help you.... Even if you delete the oce_if.o file, or if you do a clean build? If this happens just "once in a while", I would suspect hardware issues, e.g. bit errors in RAM. Could very well be, Dimitry. I have *NOT* been able to recreate the crash. I'll keep my eyes open for it again. If you'd like to play on this box, you are more than welcome to an account. Well, let's just close this issue for now. If you see the assertion coming back, and it is consistently reproducible, please re-open it. |