I just did a fresh cvs checkout of src, into /usr/src which is on: wd1: <Samsung SSD 850 EVO M.2 500GB> /dev/wd1e on /usr type ffs (log, local) cylgrp dynamic inodes FFSv2 sblock FFSv2 fslevel 5
Without attempting any builds, and just looking at the fresh checkout, according to "file": /usr/src/external/bsd/llvm/dist/llvm/test/MC/X86/intel-syntax-avx512.s: data rather than "ASCII text", and e.g.: Created src/external/bsd/llvm/dist/llvm/test/Object/Inputs/ src/external/bsd/llvm/dist/llvm/test/Object/Inputs/absolute.elf-x86-64 /absolute.elf-x86-64/1.1.1.1/// u=rw,g=rw,o=rw 711 ^?ELF^B^A^A^@^@^@^@^@^@^@^@^@^A^@>^@^A^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@p^@ ^@^@^@^@^@^@^@^@^@^@@^@^@^@^@^@@^@^G^@^D^@^@.symtab^@.strtab^@.shstrtab^@.text^@ .data^@.bss^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ The "general filesystem corruption?" guess of the "current danger?" thread in http://mail-index.netbsd.org/current-users/2017/03/31/msg031433.html seems well founded. The above is with a kernel from today's current/amd64. The kernel is running with DEBUG, DIAGNOSTIC and LOCKDEBUG. (machine was still fine at the beginning of March) Cheers, Patrick
