Re: kern/53928: modules/t_builtin:disable test case randomly fails

[Andreas Gustafsson <gson%gson.org@localhost>]

The following reply was made to PR kern/53928; it has been noted by GNATS.

From: Andreas Gustafsson <gson%gson.org@localhost>
To: "J. Hannken-Illjes" <hannken%eis.cs.tu-bs.de@localhost>
Cc: gnats-bugs%NetBSD.org@localhost
Subject: Re: kern/53928: modules/t_builtin:disable test case randomly fails
Date: Wed, 30 Jan 2019 18:28:07 +0200

 J. Hannken-Illjes wrote:
 >  This is most likely fstrans.
 
 In that case, perhaps the problem started with these commits, which
 were made the day before the first recorded test failure on the
 i386 testbed:
 
   2017.05.07.08.21.08 hannken src/sys/kern/vfs_mount.c 1.59
   2017.05.07.08.21.57 hannken src/sys/miscfs/genfs/layer_vnops.c 1.64
   2017.05.07.08.22.40 hannken src/sys/fs/union/union_vnops.c 1.68
   2017.05.07.08.23.28 hannken src/sys/kern/vfs_trans.c 1.44
   2017.05.07.08.24.20 hannken src/sys/kern/vfs_mount.c 1.60
   2017.05.07.08.24.20 hannken src/sys/kern/vfs_trans.c 1.45
   2017.05.07.08.25.54 hannken src/sys/kern/vfs_syscalls.c 1.514
   2017.05.07.08.25.54 hannken src/sys/miscfs/genfs/genfs_vfsops.c 1.6
   2017.05.07.08.26.58 hannken src/sys/kern/vfs_mount.c 1.61
   2017.05.07.08.26.58 hannken src/sys/kern/vfs_subr.c 1.464
   2017.05.07.08.26.58 hannken src/sys/kern/vfs_syscalls.c 1.515
 
 >  Destroying a "struct mount" and
 >  thus releasing its mnt_op vector is done by the last thread
 >  running fstrans_enter() and seeing a mount with IMNT_GONE.
 >
 >  It may help to "sysctl -w kern.maxvnodes=2", which fails
 >  but may trigger threads holding a reference to the mount.
 
 I'm still confused as to which part of the system you think the
 bug lies in, and what would be the correct fix, because surely
 running "sysctl -w kern.maxvnodes=2" isn't it.
 -- 
 Andreas Gustafsson, gson%gson.org@localhost