Re: kern/57371: Automated test fs/vfs/t_vnops:nfs_rename_reg_nodir fails intermittently on Rpi3 and Rpi4

[David Holland <dholland-bugs%netbsd.org@localhost>]

The following reply was made to PR kern/57371; it has been noted by GNATS.

From: David Holland <dholland-bugs%netbsd.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/57371: Automated test fs/vfs/t_vnops:nfs_rename_reg_nodir
 fails intermittently on Rpi3 and Rpi4
Date: Sun, 14 May 2023 01:32:20 +0000

 On Wed, Apr 19, 2023 at 09:30:00PM +0000, jspath55%gmail.com@localhost wrote:
  > tc-end: 1681852322.816669, nfs_rename_reg_nodir, failed, /usr/src/tests/fs/vfs/t_vnops.c:486: sb.st_nlink != 1
 
 so it seems what this test does is:
  
    touch file1 file2
    link file1 file_link
    rename file1 file3
    rename file3 file2
    rename file2 file_link
 
 that is, after hardlinking file1 it successively renames it (a) to a
 fresh name, (b) over file2, then (c) over the hard link to itself.
 
 It would be nice to see the state of the filesystem when it fails, and
 in particular whether the nfs sillyrename logic has (incorrectly)
 fired. However, I realize that between atf hiding everything and the
 difficulty of inspecting things happening inside rump that's not going
 to happen, so someone's going to have to set up a test VM and
 experiment.
 
 Do you know if the prints mentioning ZFS are because it's nfs
 exporting a ZFS volume or if they're just random prints? I can't tell
 from the maze of driver logic.
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost