From: Bob Peterson Date: Thu, 1 Feb 2018 14:00:44 -0500 (EST) Subject: [Cluster-devel] [PATCH] gfs2: Glock dump performance regression fix In-Reply-To: <20180201142216.28719-1-agruenba@redhat.com> References: <20180201142216.28719-1-agruenba@redhat.com> Message-ID: <209702553.6107480.1517511644297.JavaMail.zimbra@redhat.com> List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit ----- Original Message ----- | Restore an optimization removed in commit 7f19449553 "Fix debugfs glocks | dump": keep the glock hash table iterator active while the glock dump | file is held open. This avoids having to rescan the hash table from the | start for each read, with quadratically rising runtime. | | In addition, use rhastable_walk_peek for resuming a glock dump at the | current position: when a glock doesn't fit in the provided buffer | anymore, the next read must revisit the same glock. | | Finally, also restart the dump from the first entry when we notice that | the hash table has been resized in gfs2_glock_seq_start. | | Signed-off-by: Andreas Gruenbacher | --- Hi, Thanks. This is now pushed to the for-next branch of the linux-gfs2 tree: https://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git/commit/fs?h=for-next&id=7ac07fdaf840f9b141c6d5c286805107227c0e68 I did change one small thing: "} else" to "} else {" ... "}" Regards, Bob Peterson Red Hat File Systems