[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1459820223-24754-13-git-send-email-jsimmons@infradead.org>
Date: Mon, 4 Apr 2016 21:36:57 -0400
From: James Simmons <jsimmons@...radead.org>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
devel@...verdev.osuosl.org,
Andreas Dilger <andreas.dilger@...el.com>,
Oleg Drokin <oleg.drokin@...el.com>
Cc: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Lustre Development List <lustre-devel@...ts.lustre.org>,
Ned Bass <bass6@...l.gov>,
James Simmons <jsimmons@...radead.org>
Subject: [PATCH 12/18] staging: lustre: mdc: document mdc_rpc_lock
From: Ned Bass <bass6@...l.gov>
As this lock can be a bottleneck, clarifying why it is needed may be
helpful to those working on client performance.
Signed-off-by: Ned Bass <bass6@...l.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3443
Reviewed-on: http://review.whamcloud.com/6593
Reviewed-by: Andreas Dilger <andreas.dilger@...el.com>
Reviewed-by: Keith Mannthey <keith.mannthey@...el.com>
Signed-off-by: James Simmons <jsimmons@...radead.org>
---
drivers/staging/lustre/lustre/include/lustre_mdc.h | 18 ++++++++++++++++++
1 files changed, 18 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h
index df94f9f..9e54790 100644
--- a/drivers/staging/lustre/lustre/include/lustre_mdc.h
+++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h
@@ -64,9 +64,27 @@ struct obd_export;
struct ptlrpc_request;
struct obd_device;
+/**
+ * Serializes in-flight MDT-modifying RPC requests to preserve idempotency.
+ *
+ * This mutex is used to implement execute-once semantics on the MDT.
+ * The MDT stores the last transaction ID and result for every client in
+ * its last_rcvd file. If the client doesn't get a reply, it can safely
+ * resend the request and the MDT will reconstruct the reply being aware
+ * that the request has already been executed. Without this lock,
+ * execution status of concurrent in-flight requests would be
+ * overwritten.
+ *
+ * This design limits the extent to which we can keep a full pipeline of
+ * in-flight requests from a single client. This limitation could be
+ * overcome by allowing multiple slots per client in the last_rcvd file.
+ */
struct mdc_rpc_lock {
+ /** Lock protecting in-flight RPC concurrency. */
struct mutex rpcl_mutex;
+ /** Intent associated with currently executing request. */
struct lookup_intent *rpcl_it;
+ /** Used for MDS/RPC load testing purposes. */
int rpcl_fakes;
};
--
1.7.1
Powered by blists - more mailing lists