From: Peng Tao <bergwolf@gmail.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-kernel@vger.kernel.org, Ned Bass <bass6@llnl.gov>,
Peng Tao <bergwolf@gmail.com>,
Andreas Dilger <andreas.dilger@intel.com>
Subject: [PATCH 22/39] staging/lustre/mdc: document mdc_rpc_lock
Date: Fri, 15 Nov 2013 00:32:45 +0800 [thread overview]
Message-ID: <1384446782-13741-23-git-send-email-bergwolf@gmail.com> (raw)
In-Reply-To: <1384446782-13741-1-git-send-email-bergwolf@gmail.com>
From: Ned Bass <bass6@llnl.gov>
As this lock can be a bottleneck, clarifying why it is needed may be
helpful to those working on client performance.
Lustre-change: http://review.whamcloud.com/6593
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3443
Signed-off-by: Ned Bass <bass6@llnl.gov>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Keith Mannthey <keith.mannthey@intel.com>
Signed-off-by: Peng Tao <bergwolf@gmail.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
---
drivers/staging/lustre/lustre/include/lustre_mdc.h | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h
index 1900025..3bb2f83 100644
--- a/drivers/staging/lustre/lustre/include/lustre_mdc.h
+++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h
@@ -67,9 +67,27 @@ 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;
};
--
1.7.9.5
next prev parent reply other threads:[~2013-11-14 16:54 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-11-14 16:32 [PATCH 00/39] staging/lustre: patch bomb 2 Peng Tao
2013-11-14 16:32 ` [PATCH 01/39] staging/lustre/hsm: remove hsm_nl proc file Peng Tao
2013-11-15 4:26 ` Greg Kroah-Hartman
2013-11-15 10:32 ` Peng Tao
2013-11-14 16:32 ` [PATCH 02/39] staging/lustre/clio: wrong cl_lock usage Peng Tao
2013-11-14 16:32 ` [PATCH 03/39] staging/lustre/build: fix 'deadcode' errors Peng Tao
2013-11-14 16:32 ` [PATCH 04/39] staging/lustre/build: fix 'integer handling' issues Peng Tao
2013-11-14 16:32 ` [PATCH 05/39] staging/lustre/ptlrpc: add rpc_cache Peng Tao
2013-11-14 16:32 ` [PATCH 06/39] staging/lustre/lnet: build Lustre with Intel OFED for Xeon Phi Peng Tao
2013-11-14 16:32 ` [PATCH 07/39] staging/lustre/osd-ldiskfs: remove dependency on mdd module Peng Tao
2013-11-14 16:32 ` [PATCH 08/39] staging/lustre/libcfs: improve validity test for valid file descriptors Peng Tao
2013-11-14 16:32 ` [PATCH 09/39] staging/lustre/llite: use 64bits flags in ll_lov_setea() Peng Tao
2013-11-14 16:32 ` [PATCH 10/39] staging/lustre/hsm: small fixes for HSM Peng Tao
2013-11-14 16:32 ` [PATCH 11/39] staging/lustre/mdt: CDT cleanup follow-on patch Peng Tao
2013-11-14 16:32 ` [PATCH 12/39] staging/lustre/utils: Posix copytool fixes Peng Tao
2013-11-14 16:32 ` [PATCH 13/39] staging/lustre/lmv: update coding style Peng Tao
2013-11-14 16:32 ` [PATCH 14/39] staging/lustre/hsm: copy start error should set HP_FLAG_COMPLETED Peng Tao
2013-11-14 16:32 ` [PATCH 15/39] staging/lustre/lov: Get the correct address of lmm_objects Peng Tao
2013-11-14 16:32 ` [PATCH 16/39] staging/lustre/lvfs: remove llog_lvfs.c and other lvfs code from llog Peng Tao
2013-11-14 16:32 ` [PATCH 17/39] staging/lustre/llite: reset writeback index in ll_writepages Peng Tao
2013-11-14 16:32 ` [PATCH 18/39] staging/lustre/llite: Delaying creation of client side proc entries Peng Tao
2013-11-14 16:32 ` [PATCH 19/39] staging/lustre/libcfs: Add relocation function to libcfs heap Peng Tao
2013-11-14 16:32 ` [PATCH 20/39] staging/lustre/build: fix 'data race condition' issues Peng Tao
2013-11-14 16:32 ` [PATCH 21/39] staging/lustre/autoconf: remove LC_LOCK_MAP_ACQUIRE test Peng Tao
2013-11-14 16:32 ` Peng Tao [this message]
2013-11-14 16:32 ` [PATCH 23/39] staging/lustre/autoconf: remove LC_FS_STRUCT_USE_PATH Peng Tao
2013-11-14 16:32 ` [PATCH 24/39] staging/lustre/obdclass: fix wrong device put in case of race Peng Tao
2013-11-14 16:32 ` [PATCH 25/39] staging/lustre/lmv: choose right MDT for open by FID Peng Tao
2013-11-14 16:32 ` [PATCH 26/39] staging/lustre/osd: remove fld lookup during configuration Peng Tao
2013-11-14 16:32 ` [PATCH 27/39] staging/lustre/mdt: HSM EXIST event not triggered at last rm/mv Peng Tao
2013-11-14 16:32 ` [PATCH 28/39] staging/lustre/ldlm: ldlm_flock_deadlock() ASSERTION( req != lock ) failed Peng Tao
2013-11-14 16:32 ` [PATCH 29/39] staging/lustre/changelogs: Account for changelog_ext_rec in CR_MAXSIZE Peng Tao
2013-11-14 16:32 ` [PATCH 30/39] staging/lustre/build: build error with gcc 4.7.0 20110509 Peng Tao
2013-11-14 16:32 ` [PATCH 31/39] staging/lustre/build: fix 'NULL pointer dereference' errors Peng Tao
2013-11-14 16:32 ` [PATCH 32/39] staging/lustre/clio: honor O_NOATIME Peng Tao
2013-11-14 16:32 ` [PATCH 33/39] staging/lustre/llite: return compatible fsid for statfs Peng Tao
2013-11-14 16:32 ` [PATCH 34/39] staging/lustre/llite: cancel open lock before closing file Peng Tao
2013-11-14 16:32 ` [PATCH 35/39] staging/lustre/hsm: Add support to drop all pages for ll_data_version Peng Tao
2013-11-14 16:32 ` [PATCH 36/39] staging/lustre/build: fix 'no effect' errors Peng Tao
2013-11-14 16:33 ` [PATCH 37/39] staging/lustre/obd: add md_stats to MDC and LMV devices Peng Tao
2013-11-14 16:33 ` [PATCH 38/39] staging/lustre/autoconf: remove LC_SECURITY_PLUG test Peng Tao
2013-11-14 16:33 ` [PATCH 39/39] staging/lustre/osc: Lustre returns EINTR from writes when SA_RESTART is set Peng Tao
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=1384446782-13741-23-git-send-email-bergwolf@gmail.com \
--to=bergwolf@gmail.com \
--cc=andreas.dilger@intel.com \
--cc=bass6@llnl.gov \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
/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).