Should DEBUG_VFS_LOCKS messages be reported as bugs?

[Daniel Braniss]

> On 24 Dec 2015, at 09:40, Yuri <yuri at rawbw.com> wrote:
> 
> On 12/23/2015 23:10, Daniel Braniss wrote:
>> I just turned off WITNESS/INVARIANTS:-)
>> the only complain I get is when I do a mountd restart, but it’s harmless.
> 
> Which OS version? I get the messages on 10.2-STABLE with options from my OP.
> 
> Endless messages like this:
> Dec 23 17:59:36 yuri kernel: KDB: stack backtrace:
> Dec 23 17:59:36 yuri kernel: #0 0xffffffff8097c810 at kdb_backtrace+0x60
> Dec 23 17:59:36 yuri kernel: #1 0xffffffff809eae7e at assert_vop_elocked+0x6e
> Dec 23 17:59:36 yuri kernel: #2 0xffffffff809ea6b1 at insmntque1+0x41
> Dec 23 17:59:36 yuri kernel: #3 0xffffffff82e202a1 at unionfs_nodeget+0x281
> Dec 23 17:59:36 yuri kernel: #4 0xffffffff82e23b06 at unionfs_lookup+0x726
> Dec 23 17:59:36 yuri kernel: #5 0xffffffff80e7776f at VOP_CACHEDLOOKUP_APV+0x10f
> Dec 23 17:59:36 yuri kernel: #6 0xffffffff809d92b6 at vfs_cache_lookup+0xd6
> Dec 23 17:59:36 yuri kernel: #7 0xffffffff80e775af at VOP_LOOKUP_APV+0x10f
> Dec 23 17:59:36 yuri kernel: #8 0xffffffff809e1892 at lookup+0x5c2
> Dec 23 17:59:36 yuri kernel: #9 0xffffffff809e0fb4 at namei+0x4e4
> Dec 23 17:59:36 yuri kernel: #10 0xffffffff809f683e at kern_statat_vnhook+0xae
> Dec 23 17:59:36 yuri kernel: #11 0xffffffff809f66cd at sys_stat+0x2d
> Dec 23 17:59:36 yuri kernel: #12 0xffffffff80d4fff4 at amd64_syscall+0x2c4
> Dec 23 17:59:36 yuri kernel: #13 0xffffffff80d339db at Xfast_syscall+0xfb
> Dec 23 17:59:36 yuri kernel: insmntque: non-locked vp: 0xfffff80046d233b0 is not exclusive locked but should be
> 
> Yuri

i have been using unionfs since it appeared, and on 10.2-stable, but i don’t have WITNESS on.
BTW, it’s very slow if you turn WITNESS on, and so should only be used for debugging, and in some
cases, the code actually works only if WITNESS is on :-(

danny