panic in zfs_znode_free() during lx_chdir()

[wiedi]

Had a system panic with the following stack trace:

> ::status   
debugging crash dump vmcore.0 (64-bit) from hv0401
operating system: 5.11 joyent_20180427T013714Z (i86pc)
image uuid: (not set)
panic message: BAD TRAP: type=d (#gp General protection) rp=fffffe0133904990 addr=0
dump content: kernel pages only
> $C
fffffe0133904af0 mutex_owner_running+0xd()
fffffe0133904b30 zfs_znode_free+0x37(fffffe9c70f4c008)
fffffe0133904b90 zfs_inactive+0x16e(fffffe9c70f33d80, ffffff78d2798198, 0)
fffffe0133904bf0 fop_inactive+0x76(fffffe9c70f33d80, ffffff78d2798198, 0)
fffffe0133904c20 vn_rele+0x8a(fffffe9c70f33d80)
fffffe0133904db0 chdirec_common+0x56(fffffe9c2c43b008, fffffe9c70f33d80, 0, 1)
fffffe0133904de0 chdirec+0x33(fffffe9c70f33d80, 0, 1)
fffffe0133904e30 chdir+0x71(61bf0)
fffffe0133904e70 lx_chdir+0x50(61bf0)
fffffe0133904ef0 lx_syscall_enter+0x16f()
fffffe0133904f10 sys_syscall+0x142()

I can make the dump available.

[ikozhukhov]

i saw some similar panics on DilOS with mutex_owner_running() and it was fixed by reverting changes with dependency to libfakekernel - where we can overlap of mutexes