[Cluster-devel] [GFS2 PATCH 0/5] gfs2; jdata io deadlocks
Bob Peterson
rpeterso at redhat.com
Thu Mar 26 18:40:15 UTC 2020
This patch set addresses several problems I encountered while testing writes
to journaled data (jdata) files.
Bob Peterson (5):
gfs2: don't lock sd_log_flush_lock in try_rgrp_unlink
gfs2: instrumentation wrt ail1 stuck
gfs2: change from write to read lock for sd_log_flush_lock in journal
replay
gfs2: special log flush sequence to protect jdata writes
gfs2: implement special transaction type for jdata sync writes
fs/gfs2/aops.c | 25 ++++++--
fs/gfs2/glops.c | 2 +-
fs/gfs2/incore.h | 1 +
fs/gfs2/log.c | 137 +++++++++++++++++++++++++++++++++----------
fs/gfs2/log.h | 6 +-
fs/gfs2/ops_fstype.c | 1 +
fs/gfs2/recovery.c | 4 +-
fs/gfs2/rgrp.c | 2 -
fs/gfs2/trans.c | 39 +++++++++---
fs/gfs2/trans.h | 4 ++
10 files changed, 171 insertions(+), 50 deletions(-)
--
2.25.1
More information about the Cluster-devel
mailing list