Re: `tail -F somefile` and ZFS weirdness

On Wed, May 13, 2026 at 03:38:01PM +0000, Shawn Webb wrote:
> Hey all,
> 
> I just noticed this in a bhyve VM running HardenedBSD 16-CURRENT. My
> `uname -a` aligns with FreeBSD commit b76b05c04cc9ed2cd053c7b367c92628447ede6f.
> 
> (so pretend the git commit has in my `uname -a` is FreeBSD's
> b76b05c04cc9ed2cd053c7b367c92628447ede6f commit.)
> 
> I don't recall seeing this message before, so I suspect it might have
> been introduced relatively recently.
> 
> ==== BEGIN LOG ====
> $ uname -a
> FreeBSD hbsd-current-02 16.0-CURRENT FreeBSD 16.0-CURRENT #5  hardened/current/master-n196117-464cdf02df98-dirty: Wed May 13 01:12:14 UTC 2026     shawn@hbsd-current-02:/usr/obj/usr/src/amd64.amd64/sys/HARDENEDBSD amd64
> 
> $ tail -n 21 /var/log/messages
> rms_rlock with the following non-sleepable locks held:
> exclusive sleep mutex kqueue (kqueue) r = 0 (0xfffff80001d0b400) locked @ /usr/src/sys/kern/kern_event.c:2586
> stack backtrace:
> #0 0xffffffff80c921dc at witness_debugger+0x6c
> #1 0xffffffff80c93adc at witness_warn+0x4bc
> #2 0xffffffff80c0e10f at rms_rlock+0x1f
> #3 0xffffffff82799a6b at zfs_freebsd_getattr+0x11b
> #4 0xffffffff8127dba7 at VOP_GETATTR_APV+0x57
> #5 0xffffffff80d2a7db at vn_getsize_locked+0x7b
> #6 0xffffffff80d30b6d at filt_vfsread+0x8d
> #7 0xffffffff80bb4d06 at knote+0xd6
> #8 0xffffffff80d28a48 at vop_read_post+0x48
> #9 0xffffffff8127e07a at VOP_READ_APV+0x9a
> #10 0xffffffff80d450f2 at vn_read+0x242
> #11 0xffffffff80d44df5 at vn_io_fault_doio+0x55
> #12 0xffffffff80d3ff2f at vn_io_fault1+0x19f
> #13 0xffffffff80d3d040 at vn_io_fault+0x190
> #14 0xffffffff80c97e95 at dofileread+0x85
> #15 0xffffffff80c97852 at sys_read+0x132
> #16 0xffffffff811a7974 at amd64_syscall+0x184
> #17 0xffffffff81175f0b at fast_syscall_common+0xf8
> ==== END LOG ====
> 
> Thanks,
> 
> -- 
> Shawn Webb
> Cofounder / Security Engineer
> HardenedBSD
> 
> Signal Username:  shawn_webb.74
> Tor-ified Signal: +1 303-901-1600 / shawn_webb_opsec.50
> https://git.hardenedbsd.org/hardenedbsd/pubkeys/-/raw/master/Shawn_Webb/03A4CBEBB82EA5A67D9F3853FF2E67A277F8E1FA.pub.asc


Try this, untested.  I should have not changed this, but then I do not
understand why it only appeared now.

commit c3a20bdb7caa3628831bf443cfb1e0316e30d3e7
Author: Konstantin Belousov <kib@FreeBSD.org>
Date:   Wed May 13 23:44:45 2026 +0300

    vop_read{,_pgcache}_post(): unlock knote around call into filter
    
    ZFS vop_getattr() needs a sleepable lock.

diff --git a/sys/kern/vfs_subr.c b/sys/kern/vfs_subr.c
index b3a6e8ce2e02..d19776940bc3 100644
--- a/sys/kern/vfs_subr.c
+++ b/sys/kern/vfs_subr.c
@@ -6509,7 +6509,7 @@ vop_read_post(void *ap, int rc)
 	struct vop_read_args *a = ap;
 
 	if (!rc) {
-		VFS_KNOTE_LOCKED(a->a_vp, NOTE_READ);
+		VN_KNOTE_LOCKED(a->a_vp, NOTE_READ);
 		INOTIFY(a->a_vp, IN_ACCESS);
 	}
 }
@@ -6520,7 +6520,7 @@ vop_read_pgcache_post(void *ap, int rc)
 	struct vop_read_pgcache_args *a = ap;
 
 	if (rc == 0) {
-		VFS_KNOTE_LOCKED(a->a_vp, NOTE_READ);
+		VN_KNOTE_UNLOCKED(a->a_vp, NOTE_READ);
 		INOTIFY(a->a_vp, IN_ACCESS);
 	}
 }