[Bug 238789] panic: mutex so_rcv not owned at /usr/src/sys/kern/uipc_socket.c:2359

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=238789

--- Comment #5 from Greg Becker <greg at codeconcepts.com> ---
Sure Kubilay, here's my backtrace.  It's from a kernel module I am working on
that I have not yet published.  I hit this easily with GENERIC, but never hit
it when running with the patch I supplied:

panic: mutex so_rcv not owned at /usr/src/sys/kern/uipc_socket.c:2359
cpuid = 11
time = 1561372118
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2c/frame 0xfffffe012015e5b0
kdb_backtrace() at kdb_backtrace+0x53/frame 0xfffffe012015e660
vpanic() at vpanic+0x265/frame 0xfffffe012015e720
doadump() at doadump/frame 0xfffffe012015e780
__mtx_assert() at __mtx_assert+0x145/frame 0xfffffe012015e800
soreceive_stream() at soreceive_stream+0x963/frame 0xfffffe012015e8b0
soreceive() at soreceive+0x102/frame 0xfffffe012015e910
krpc_recv_tcp() at krpc_recv_tcp+0x55/frame 0xfffffe012015e9a0
svc_rcv_receive() at svc_rcv_receive+0x1c/frame 0xfffffe012015e9e0
tpool_run() at tpool_run+0x92/frame 0xfffffe012015ea30
fork_exit() at fork_exit+0x13b/frame 0xfffffe012015eab0
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe012015eab0
--- trap 0, rip = 0, rsp = 0, rbp = 0 ---
KDB: enter: panic
[ thread pid 0 tid 101260 ]
Stopped at      breakpoint+0x5: popq    %rbp


I think what is happening is that I am post-processing a rcv soupcall, but the
non-blocking call to sblock() in soreceive() loses due to a race with an
asynchronously running call to soshutdown(SHUT_RD) from my module unload code. 
But I have not yet verified that...

I'm using GENERIC with the following options:

options CONSPEED=115200
options BREAK_TO_DEBUGGER
options DDB

options SOCKBUF_DEBUG
options INVARIANTS
options INVARIANT_SUPPORT


FreeBSD sm1.cc.codeconcepts.com 12.0-STABLE FreeBSD 12.0-STABLE #31 r349288M:
Tue Jun 25 07:19:16 CDT 2019    
greg at sm1.cc.codeconcepts.com:/usr/obj/usr/src/amd64.amd64/sys/SM1  amd64

-- 
You are receiving this mail because:
You are the assignee for the bug.