Hardware: Thinkpad l14 gen2 (i5) Module: rtw8852AE Options: INVARIANTS iNVARIANT_SUPPORT I'm aware that another user has reported this problem, but I didn't see any entry about what the kernel returned to me, and it says "closed." I tried adding the "INVARIANTS" and "INVARIANT_SUPPORT" options to my FreeBSD kernel, and what I got was a kernel panic with this message: panic: Lock &wiphy->mtx not locked /usr/src/sys/kern/kern_sx.c:178 cpuid = 7 time = 1785616894 KDB: stack backtrace: db_trace_self_wrapper() at 0xffffffff883f5c8b = db_trace_self_wrapper+0x2b/frame 0xfffffe01316217e0 vpanic() at 0xffffffff806576d6 = vpanic+0x136/frame 0xfffffe0131621918 panic() at 0xffffffff88657593 = panic+0x43/frame 0xfffffe0131621978 _sx_assert() at 0xffffffff8066208d = _sx_assert+0x11d/frame 0xfffffe0131621988 rtw89_ops_link_info_changed() at 0xffffffff8301233b = rtw89_ops_link_info_changed+0x3b/frame 0xfffffe01316219c0 1kpi_ic_vap_create() at 0xffffffff888b2eба = 1kpi_ic_vap_create+0xa2a/frame 0xfffffe0131621a50 wlan_clone_create() at 0xffffffff887d6e5d = wlan_clone_create+0xed/frame 0xfffffe0131621aa0 if_clone_createif_nl() at 0xffffffff88793e36 = if_clone_createif_nl+0x166/frame 0xfffffe0131621b18 ifc_create_ifp() at 0xffffffff88793cb2 = ifc_create_ifp+8x182/frame 0xfffffe0131621b88 if_clone_create() at 0xffffffff887948d4 = if_clone_create+0x34/frame 0xfffffe0131621bd0 ifioctl() at 0xffffffff887900c8 = ifioctl+0xc18/frame 0xfffffe0131621cd0 kern_ioctl() at 0xffffffff886ccb71 = kern_ioct1+0x2a1/frame 0xfffffe0131621d48 sys_loctl() at 0xffffffff886cc863 = sys_ioct1+0x133/frame 0xfffffe0131621e00 amd64_syscall() at 0xffffffff88a39299 = amd64_syscall+0x169/frame 0xfffffe0131621f30 fast_syscall_common() at 0xffffffff80a0dd9b = fast_syscall_common+8xf8/frame 0xfffffe0131621f30 syscall (54, FreeBSD ELF64, ioctl), rip = 0xf58e8d6926a, rsp = 0xf58e2284e98, rbp = 0xf58e2284ee0 KDB: enter: panic [ thread pid 355 tid 180941 ] Stopped at 0xffffffff886a4a43 = kdb_enter+0x33: movq $0, 0xa6b3c2(Zrip) db> In addition, apparently because of this same issue with MTX, I occasionally experience DNS resolution problems, forcing me to perform a "service netif restart".
The other problem got fixed in stable/15 after 15.1-RELEASE but this looks different. It seems I forgot to push a change; I'll go and find it and rebase. The current code still has the /* XXX-FINISH all callers for lockdep_assert_wiphy(hw->wiphy); */ which I know I did in a branch. Thanks for reporting!