cluster-devel.redhat.com archive mirror
 help / color / mirror / Atom feed
From: Andreas Gruenbacher <agruenba@redhat.com>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] [PATCH v2 5/6] gfs2: Allow node-wide exclusive glock sharing
Date: Sat,  1 Dec 2018 12:10:18 +0100	[thread overview]
Message-ID: <20181201111019.14363-6-agruenba@redhat.com> (raw)
In-Reply-To: <20181201111019.14363-1-agruenba@redhat.com>

From: Bob Peterson <rpeterso@redhat.com>

Introduce a new LM_FLAG_NODE_SCOPE glock holder flag: when taking a
glock in LM_ST_EXCLUSIVE (EX) mode and with the LM_FLAG_NODE_SCOPE flag
set, the exclusive lock is shared among all local processes who are
holding the glock in EX mode and have the LM_FLAG_NODE_SCOPE flag set as
well.  (From the point of view of other nodes, the lock is still held
exclusively.)

A future patch will start using this flag to improve performance with
rgrp sharing.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/glock.c | 26 ++++++++++++++++++++++----
 fs/gfs2/glock.h |  6 ++++++
 2 files changed, 28 insertions(+), 4 deletions(-)

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 05431324b262b..1c07a1819c1b0 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -279,10 +279,26 @@ void gfs2_glock_put(struct gfs2_glock *gl)
 
 static inline int may_grant(const struct gfs2_glock *gl, const struct gfs2_holder *gh)
 {
-	const struct gfs2_holder *gh_head = list_entry(gl->gl_holders.next, const struct gfs2_holder, gh_list);
-	if ((gh->gh_state == LM_ST_EXCLUSIVE ||
-	     gh_head->gh_state == LM_ST_EXCLUSIVE) && gh != gh_head)
-		return 0;
+	const struct gfs2_holder *gh_head = list_entry(gl->gl_holders.next,
+						       const struct gfs2_holder,
+						       gh_list);
+
+	if (gh != gh_head) {
+		/**
+		 * Here we make a special exception to grant holders who agree
+		 * to share the EX lock with other holders who also have the
+		 * bit set. If the original holder has the LM_FLAG_NODE_SCOPE bit
+		 * is set, we grant more holders with the bit set.
+		 */
+		if (gh_head->gh_state == LM_ST_EXCLUSIVE &&
+		    (gh_head->gh_flags & LM_FLAG_NODE_SCOPE) &&
+		    gh->gh_state == LM_ST_EXCLUSIVE &&
+		    (gh->gh_flags & LM_FLAG_NODE_SCOPE))
+			return 1;
+		if ((gh->gh_state == LM_ST_EXCLUSIVE ||
+		     gh_head->gh_state == LM_ST_EXCLUSIVE))
+			return 0;
+	}
 	if (gl->gl_state == gh->gh_state)
 		return 1;
 	if (gh->gh_flags & GL_EXACT)
@@ -1682,6 +1698,8 @@ static const char *hflags2str(char *buf, u16 flags, unsigned long iflags)
 		*p++ = 'A';
 	if (flags & LM_FLAG_PRIORITY)
 		*p++ = 'p';
+	if (flags & LM_FLAG_NODE_SCOPE)
+		*p++ = 'n';
 	if (flags & GL_ASYNC)
 		*p++ = 'a';
 	if (flags & GL_EXACT)
diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h
index 5e12220cc0c24..3a3c1bd5b2d89 100644
--- a/fs/gfs2/glock.h
+++ b/fs/gfs2/glock.h
@@ -78,6 +78,11 @@ enum {
  * request and directly join the other shared lock.  A shared lock request
  * without the priority flag might be forced to wait until the deferred
  * requested had acquired and released the lock.
+ *
+ * LM_FLAG_NODE_SCOPE
+ * This holder agrees to share the lock within this node. In other words,
+ * the glock is held in EX mode according to DLM, but local holders on the
+ * same node can share it.
  */
 
 #define LM_FLAG_TRY		0x0001
@@ -85,6 +90,7 @@ enum {
 #define LM_FLAG_NOEXP		0x0004
 #define LM_FLAG_ANY		0x0008
 #define LM_FLAG_PRIORITY	0x0010
+#define LM_FLAG_NODE_SCOPE	0x0020
 #define GL_ASYNC		0x0040
 #define GL_EXACT		0x0080
 #define GL_SKIP			0x0100
-- 
2.19.1.546.g028f9c799.dirty



  parent reply	other threads:[~2018-12-01 11:10 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-01 11:10 [Cluster-devel] [PATCH v2 0/6] gfs2: Prepare for resource group glock sharing Andreas Gruenbacher
2018-12-01 11:10 ` [Cluster-devel] [PATCH v2 1/6] gfs2: Only use struct gfs2_rbm for bitmap manipulations Andreas Gruenbacher
2018-12-01 11:10 ` [Cluster-devel] [PATCH v2 2/6] gfs2: Clean up gfs2_adjust_reservation Andreas Gruenbacher
2018-12-03 14:32   ` Bob Peterson
2018-12-03 14:58     ` Andreas Gruenbacher
2018-12-04 22:02   ` Andreas Gruenbacher
2018-12-01 11:10 ` [Cluster-devel] [PATCH v2 3/6] gfs2: Add per-reservation reserved block accounting Andreas Gruenbacher
2018-12-01 11:10 ` [Cluster-devel] [PATCH v2 4/6] gfs2: Add local resource group locking Andreas Gruenbacher
2018-12-01 11:10 ` Andreas Gruenbacher [this message]
2018-12-01 11:10 ` [Cluster-devel] [PATCH v2 6/6] gfs2: Introduce resource group sharing Andreas Gruenbacher

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181201111019.14363-6-agruenba@redhat.com \
    --to=agruenba@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).