kern/56713: Strange ``tail -f'' behaviour - likely kqueue-related

[paul%whooppee.com@localhost]

>Number:         56713
>Category:       kern
>Synopsis:       Strange ``tail -f'' behaviour - likely kqueue-related
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Feb 14 16:00:00 +0000 2022
>Originator:     Paul Goyette
>Release:        NetBSD 9.99.93
>Organization:
+--------------------+--------------------------+----------------------+
| Paul Goyette       | PGP Key fingerprint:     | E-mail addresses:    |
| (Retired)          | FA29 0E3B 35AF E8AE 6651 | paul%whooppee.com@localhost    |
| Software Developer | 0786 F758 55DE 53BA 7731 | pgoyette%netbsd.org@localhost  |
| & Network Engineer |                          | pgoyette99%gmail.com@localhost |
+--------------------+--------------------------+----------------------+
>Environment:
	
	
System: NetBSD speedy.whooppee.com 9.99.93 NetBSD 9.99.93 (SPEEDY 2022-01-08 03:00:28 UTC) #0: Sat Jan 8 13:04:35 UTC 2022 paul%speedy.whooppee.com@localhost:/build/netbsd-local/obj/amd64/sys/arch/amd64/compile/SPEEDY amd64
Architecture: x86_64
Machine: amd64
>Description:
I've been seeing some very odd behaviour lately, with ``tail -f'',
but I suspect it is a wider issue.  (If someone confirms that this
really is a problem with tail(1) rather than with the kernel, please
feel free to recategorize the PR.)
 
I've got a bunch of packages being built (in a chroot sandbox), and
occassionally I do a ``tail -f'' on the current package's log file
just to make sure it is still making progress.  So the file is open
by both the package-building shell and the tail process.
 
Occassionally the tail process fails to notice any further changes
to the file's size, and just sits there.  The specific cases I've
looked at have all occurred immediately after the first 10-line
display (ie, the initial tail -10 output).  Rather than following
additional changes to the log file, the tail process just hangs.
ps(1) shows the process is stuck in ``wchan=kqueue''.  I've got
one such process that has been stuck for days - for this process,
the 'initial -10' output is more than 2.5kb of output being sent
to my xterm session before hanging!

I don't ever remember seeing this prior to my recent update from
9.99.92 to .93

>How-To-Repeat:
	See above
>Fix:
	please do

>Unformatted: