On Thu, Jul 31, 2025, at 15:32, Naresh Kamboju wrote:
Regressions while building x86, S390 and riscv64 with hardening.config on the Linux next-20250730 and next-20250731.
Build pass with gcc-13 with hardening.config fails with gcc-8
Regression Analysis:
- New regression? Yes
- Reproducibility? Yes
First seen on the next-20250730 Good: next-20250729 Bad: next-20250730
Build regression: next-20250730 x86 gcc-8 hardening.config vmlinux symbol `.modinfo' required but not present Build regression: next-20250730 S390 gcc-8 hardening.config vmlinux symbol `.modinfo' required but not present Build regression: next-20250730 riscv64 gcc-8 hardening.config vmlinux symbol `__rela_dyn_end' required but not present
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
My guess is that this is caused by 94564d1bb059 ("kbuild: keep .modinfo section in vmlinux.unstripped")
See also https://lore.kernel.org/all/202507310505.tEYm1za7-lkp@intel.com/ https://lore.kernel.org/all/20250728135753.432695A72-agordeev@linux.ibm.com/
Reverting 94564d1bb059 solved the problem for me on arm64.
Arnd