gfs2: fix scheduling while atomic bug in glocks
authorBob Peterson <rpeterso@redhat.com>
Tue, 18 May 2021 13:12:10 +0000 (09:12 -0400)
committerAndreas Gruenbacher <agruenba@redhat.com>
Thu, 20 May 2021 11:31:37 +0000 (13:31 +0200)
commit20265d9a67e40eafd39a8884658ca2e36f05985d
treef944efab603c456faf64c95ae23c783ba011dd3e
parent4194dec4b4169e5a9a5171db60c2ec00c4d8cf16
gfs2: fix scheduling while atomic bug in glocks

Before this patch, in the unlikely event that gfs2_glock_dq encountered
a withdraw, it would do a wait_on_bit to wait for its journal to be
recovered, but it never released the glock's spin_lock, which caused a
scheduling-while-atomic error.

This patch unlocks the lockref spin_lock before waiting for recovery.

Fixes: 601ef0d52e96 ("gfs2: Force withdraw to replay journals and wait for it to finish")
Cc: stable@vger.kernel.org # v5.7+
Reported-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
fs/gfs2/glock.c