[Bug 289219] lock order reversal between vtnet1-rx0 (if_vtnet.c:2213) and in6_multi_sx (in6_mcast.c:1217)

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=289219

            Bug ID: 289219
           Summary: lock order reversal between vtnet1-rx0
                    (if_vtnet.c:2213) and in6_multi_sx (in6_mcast.c:1217)
           Product: Base System
           Version: 15.0-CURRENT
          Hardware: Any
                OS: Any
            Status: New
          Severity: Affects Some People
          Priority: ---
         Component: kern
          Assignee: [email protected]
          Reporter: [email protected]
                CC: [email protected]

I see this lock order reversal on 15-PRERELEASE
(778bfd4e1033715de6cf21e2e2454bdea2a0f1ab) as a guest vm inside bhyve.

Here is the log:
```
vtnet0: link state changed to UP
vtnet1: link state changed to UP
Invoking IPv6 network device address event may sleep with the following
non-sleepable locks held:
exclusive sleep mutex vtnet1-rx0 (vtnet1-rx0) r = 0 (0xfffff80003680480) locked
@ /usr/src/sys/dev/virtio/network/if_vtnet.c:2213
stack backtrace:
#0 0xffffffff80bff62c at witness_debugger+0x6c
#1 0xffffffff80c00840 at witness_warn+0x430
#2 0xffffffff80dc7c97 at in6_update_ifa+0xc97
#3 0xffffffff80df45b0 at in6_ifadd+0x1e0
#4 0xffffffff80df0c83 at nd6_ra_input+0x1023
#5 0xffffffff80dc1986 at icmp6_input+0x5b6
#6 0xffffffff80ddb3a8 at ip6_input+0xc98
#7 0xffffffff80cf0ea4 at netisr_dispatch_src+0xb4
#8 0xffffffff80cd319a at ether_demux+0x16a
#9 0xffffffff80cd483e at ether_nh_input+0x3ce
#10 0xffffffff80cf0ea4 at netisr_dispatch_src+0xb4
#11 0xffffffff80cd35f5 at ether_input+0xd5
#12 0xffffffff809a40ca at vtnet_rxq_eof+0x74a
#13 0xffffffff809a38cc at vtnet_rx_vq_process+0x9c
#14 0xffffffff80b3d326 at ithread_loop+0x266
#15 0xffffffff80b39602 at fork_exit+0x82
#16 0xffffffff810abd9e at fork_trampoline+0xe
lock order reversal: (sleepable after non-sleepable)
 1st 0xfffff80003680480 vtnet1-rx0 (vtnet1-rx0, sleep mutex) @
/usr/src/sys/dev/virtio/network/if_vtnet.c:2213
 2nd 0xffffffff81edd170 in6_multi_sx (in6_multi_sx, sx) @
/usr/src/sys/netinet6/in6_mcast.c:1217
lock order vtnet1-rx0 -> in6_multi_sx attempted at:
#0 0xffffffff80bff201 at witness_checkorder+0xbe1
#1 0xffffffff80b925a0 at _sx_xlock+0x60
#2 0xffffffff80dcffe1 at in6_joingroup+0x31
#3 0xffffffff80dc805a at in6_update_ifa+0x105a
#4 0xffffffff80df45b0 at in6_ifadd+0x1e0
#5 0xffffffff80df0c83 at nd6_ra_input+0x1023
#6 0xffffffff80dc1986 at icmp6_input+0x5b6
#7 0xffffffff80ddb3a8 at ip6_input+0xc98
#8 0xffffffff80cf0ea4 at netisr_dispatch_src+0xb4
#9 0xffffffff80cd319a at ether_demux+0x16a
#10 0xffffffff80cd483e at ether_nh_input+0x3ce
#11 0xffffffff80cf0ea4 at netisr_dispatch_src+0xb4
#12 0xffffffff80cd35f5 at ether_input+0xd5
#13 0xffffffff809a40ca at vtnet_rxq_eof+0x74a
#14 0xffffffff809a38cc at vtnet_rx_vq_process+0x9c
#15 0xffffffff80b3d326 at ithread_loop+0x266
#16 0xffffffff80b39602 at fork_exit+0x82
#17 0xffffffff810abd9e at fork_trampoline+0xe
```

With the configuration below (especially accept_rtadv), it happens immediately
after boot:
ipv6_activate_all_interfaces="YES"
ifconfig_DEFAULT="inet6 accept_rtadv"
ifconfig_vtnet0_ipv6="inet6 fdb5:c59b:114e::a"

-- 
You are receiving this mail because:
You are on the CC list for the bug.