Re: Panic after update main-n269202-4e7aa03b7076 -> n269230-f6f67f58c19d

On Tue, Apr 9, 2024 at 4:47 AM David Wolfskill <david@catwhisker.org> wrote:
>
> Machine had been running:
>
> FreeBSD 15.0-CURRENT #43 main-n269202-4e7aa03b7076: Mon Apr  8 11:19:58 UTC 2024     root@freebeast.catwhisker.org:/common/S4/obj/usr/src/amd64.amd64/sys/GENERIC amd64 1500018 1500018
>
> This was an in-place source update, after updating sources to
> main-n269230-f6f67f58c19d.  On reboot (after "make installworld"
> completed, I see this on the serial console (copy/pasted):
>
> ...
> Starting lockd.
I'd guess this is caused by some recent change to AF_UNIX socket
creation. The crash appears to be either the SOCK_LOCK() or
SOCKBUF_LOCK(&so->so_rcv) not being initialized.
If you can find out what source line# corresponds to
clnt_vc_create+0x4f4 you can probably tell which one it is.

All local_rpcb() does is a
  error = socreate(AF_LOCAL, &so, SOCK_STREAM, 0, curthread->td_ucred,
curthread);
  and then calls clnt_vc_create(..so..) with the socket.

I think that socreate() is not initializing one of those two mutexes
for some reason.

rick

>
>
> Fatal trap 12: page fault while in kernel mode
> cpuid = 9; apic id = 09
> fault virtual address   = 0x18
> fault code              = supervisor read data, page not present
> instruction pointer     = 0x20:0xffffffff80b208c5
> stack pointer           = 0x28:0xfffffe048c204920
> frame pointer           = 0x28:0xfffffe048c204960
> code segment            = base 0x0, limit 0xfffff, type 0x1b
>                         = DPL 0, pres 1, long 1, def32 0, gran 1
> processor eflags        = interrupt enabled, resume, IOPL = 0
> current process         = 1208 (rpc.Starting automountd.
> lockd)
> rdi: 0000000000000000 rsi: fffff801078b0740 rdx: 0000000000000000
> rcx: 000000000000010a  r8: ffffffff818d30f0  r9: 0000000000000000
> rax: 0000000000000000 rbx: 00000000Starting powerd.00000018 rbp: fffffe048c204960
> r10: 0000000000010000 r11: 0000000000000001 r12: fffff80274e32c18
> r13: 000000000000010a r14: fffff80274e32c00 r15: ffffffff812ae38a
> trap number             = 12
> panic: page fault
> cpuid = 9
> time = 1712662362
> KDB: stack backtrace:
> db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe048c2045f0
> vpanic() at vpanic+0x135/frame 0xfffffe048c204720
> panic() at panic+0x43/frame 0xfffffe048c204780
> trap_fatal() at trap_fatal+0x40b/frame 0xfffffe048c2047e0
> trap_pfault() at trap_pfault+0xa0/frame 0xfffffe048c204850
> calltrap() at calltrap+0x8/frame 0xfffffe048c204850
> --- trap 0xc, rip = 0xffffffff80b208c5, rsp = 0xfffffe048c204920, rbp = 0xfffffe
> 048c204960 ---
> __mtx_lock_flags() at __mtx_lock_flags+0x45/frame 0xfffffe048c204960
> clnt_vc_create() at clnt_vc_create+0x4f4/frame 0xfffffe048c204ab0
> local_rpcb() at local_rpcb+0x11b/frame 0xfffffe048c204b50
> rpcb_unset() at rpcb_unset+0x24/frame 0xfffffe048c204bb0
> svc_tp_create() at svc_tp_create+0xee/frame 0xfffffe048c204c90
> sys_nlm_syscall() at sys_nlm_syscall+0x3d0/frame 0xfffffe048c204e00
> amd64_syscall() at amd64_syscall+0x158/frame 0xfffffe048c204f30
> fast_syscall_common() at fast_syscall_common+0xf8/frame 0xfffffe048c204f30
> --- syscall (154, FreeBSD ELF64, nlm_syscall), rip = 0x3f00a2dfd2a, rsp = 0x3f00
> 96f7168, rbp = 0x3f0096f7230 ---
> KDB: enter: panic
> [ thread pid 1208 tid 101107 ]
> Stopped at      kdb_enter+0x33: movq    $0,0x104eb92(%rip)
> db>
>
>
> Given suitable clues, I can poke at it a bit -- this is my "build
> machine," so it doesn't have critical work to do at the moment.  (I
> would normally have powered it down for the day: here's no need for
> it to be wasting energy.)
>
> Laptops are still building ports under stable/14 -- something seems
> to want the llvm17 port, and they have firefox to build, so they
> won't be testing CURRENT/head for a while, yet.
>
> Peace,
> david
> --
> David H. Wolfskill                              david@catwhisker.org
> Alexey Navalny was a courageous man; Putin has made him a martyr.
>
> See https://www.catwhisker.org/~david/publickey.gpg for my public key.