Below is part of the contents of /var/crash/textdump.tar.0 It occurs if I run: poudriere bulk -T -f ~/pkg/one -j 140 It can be fixed by USE_TMPFS=no in poudriere.conf I thought this bug was fixed some time ago ? uname -a: FreeBSD fc.opsec.eu 14.0-CURRENT FreeBSD 14.0-CURRENT #0 main-n260810-cbac1f346495: Sat Feb 11 21:26:27 CET 2023 pi@fc.opsec.eu:/usr/obj/usr/src/amd64.amd64/sys/GENERIC amd64 panic: Assertion VOP_ISLOCKED(vp) failed at /usr/src/sys/fs/tmpfs/tmpfs_vnops.c:388 cpuid = 9 time = 1676197346 KDB: stack backtrace: db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe0259d3d9e0 vpanic() at vpanic+0x152/frame 0xfffffe0259d3da30 panic() at panic+0x43/frame 0xfffffe0259d3da90 tmpfs_access() at tmpfs_access+0x137/frame 0xfffffe0259d3dad0 tmpfs_rename() at tmpfs_rename+0x89f/frame 0xfffffe0259d3dba0 VOP_RENAME_APV() at VOP_RENAME_APV+0x3c/frame 0xfffffe0259d3dbc0 kern_renameat() at kern_renameat+0x416/frame 0xfffffe0259d3de00 amd64_syscall() at amd64_syscall+0x143/frame 0xfffffe0259d3df30 fast_syscall_common() at fast_syscall_common+0xf8/frame 0xfffffe0259d3df30 --- syscall (128, FreeBSD ELF64, rename), rip = 0x1c4a29a30aaa, rsp = 0x1c4a273787d8, rbp = 0x1c4a27379660 ---
Adding kib@ in cc: as he may have knowledge about tmpfs
Adding fsu@ in cc: too as he recently touched tmpfs
this should be fixed as of 3a3450eda6d4616df51a30f84a872d9d43669d78
I've upgraded twice since I opened this PR. It looks better now, but as I'm only running a newer version since 2nd of January, I have not yet enough work done on that version to really confirm fix. Will reopen, if it re-occurs.