nfsd: don't call nfsd_file_put from client states seqfile display
authorJeff Layton <jlayton@kernel.org>
Fri, 28 Oct 2022 12:13:53 +0000 (08:13 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 31 Dec 2022 12:31:58 +0000 (13:31 +0100)
commit23ac79e755dfa10d8d317f8b7a3ab1c8467c79a7
tree8b6443719ea730e38bc5d67d4ddab81401c3743c
parentb04556ef19ed92720c9fad6c9fb28fcab4a576e3
nfsd: don't call nfsd_file_put from client states seqfile display

[ Upstream commit e0aa651068bfd520afcd357af8ecd2de005fc83d ]

We had a report of this:

    BUG: sleeping function called from invalid context at fs/nfsd/filecache.c:440

...with a stack trace showing nfsd_file_put being called from
nfs4_show_open. This code has always tried to call fput while holding a
spinlock, but we recently changed this to use the filecache, and that
started triggering the might_sleep() in nfsd_file_put.

states_start takes and holds the cl_lock while iterating over the
client's states, and we can't sleep with that held.

Have the various nfs4_show_* functions instead hold the fi_lock instead
of taking a nfsd_file reference.

Fixes: 78599c42ae3c ("nfsd4: add file to display list of client's opens")
Link: https://bugzilla.redhat.com/show_bug.cgi?id=2138357
Reported-by: Zhi Li <yieli@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
fs/nfsd/nfs4state.c