public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Boaz Harrosh <bharrosh@panasas.com>
To: Fred Isaman <iisaman@umich.edu>
Cc: Benny Halevy <bhalevy@panasas.com>,
	Fred Isaman <iisaman@netapp.com>,
	linux-nfs@vger.kernel.org
Subject: [PATCH] FIXME: pnfs-obj: Short circuit the objlayout_commit to be a no-op
Date: Wed, 09 Jun 2010 18:15:34 +0300	[thread overview]
Message-ID: <4C0FB016.3020208@panasas.com> (raw)
In-Reply-To: <4C0FAF6B.90902@panasas.com>


Do to a bug in generic client: Return of NFS_FILE_SYNC from
write_done will render the system useless. When returning
NFS_UNSTABLE the generic layer then returns with a call to
objlayout_commit. At the outpost a successful nfs_commit_complete()
should be called and PNFS_ATTEMPTED returned.

Since nfs_commit_complete cannot be called from within
objlayout_commit. It is scheduled on an rpc task to be called
asynchronously.

TODO:
  All this is good code, actually needed and missing from obio_osd.
  What's missing is the actual call to osd_flush() and the completion
  call on request return.

Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
---
 fs/nfs/objlayout/objio_osd.c |    2 +-
 fs/nfs/objlayout/objlayout.c |   17 ++++++++++++++++-
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index 315f8c6..4e266a2 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -852,7 +852,7 @@ static ssize_t _write_done(struct objio_state *ios)
 	if (likely(!ret)) {
 		/* FIXME: should be based on the OSD's persistence model
 		 * See OSD2r05 Section 4.13 Data persistence model */
-		ios->ol_state.committed = NFS_FILE_SYNC;
+		ios->ol_state.committed = NFS_UNSTABLE; //NFS_FILE_SYNC;
 		status = ios->length;
 	} else {
 		status = ret;
diff --git a/fs/nfs/objlayout/objlayout.c b/fs/nfs/objlayout/objlayout.c
index 880d987..60f64b7 100644
--- a/fs/nfs/objlayout/objlayout.c
+++ b/fs/nfs/objlayout/objlayout.c
@@ -287,15 +287,30 @@ objlayout_io_set_result(struct objlayout_io_state *state, unsigned index,
 	}
 }
 
+static void _rpc_commit_complete(struct work_struct *work)
+{
+	struct rpc_task *task;
+	struct nfs_write_data *wdata;
+
+	dprintk("%s enter\n", __func__);
+	task = container_of(work, struct rpc_task, u.tk_work);
+	wdata = container_of(task, struct nfs_write_data, task);
+
+	pnfs_client_ops->nfs_commit_complete(wdata);
+}
+
 /*
  * Commit data remotely on OSDs
  */
 enum pnfs_try_status
 objlayout_commit(struct pnfs_layout_type *pnfslay,
 		 int sync,
-		 struct nfs_write_data *data)
+		 struct nfs_write_data *wdata)
 {
 	int status = PNFS_ATTEMPTED;
+
+	INIT_WORK(&wdata->task.u.tk_work, _rpc_commit_complete);
+	schedule_work(&wdata->task.u.tk_work);
 	dprintk("%s: Return %d\n", __func__, status);
 	return status;
 }
-- 
1.6.6.1



  reply	other threads:[~2010-06-09 15:15 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-08  4:18 [PATCH 00/24] LAYOUTGET invocation (rebased) Fred Isaman
2010-06-08  4:18 ` [PATCH 01/24] Revert "pnfs-nonfilelayout: Prelim support for non-file layout O_DIRECT" Fred Isaman
2010-06-08  4:18   ` [PATCH 02/24] Revert "pnfs: Enable O_DIRECT write path." Fred Isaman
2010-06-08  4:19     ` [PATCH 03/24] Revert "pnfs: Enable O_DIRECT read path." Fred Isaman
2010-06-08  4:19       ` [PATCH 04/24] Revert "pnfs: Add function to set up O_DIRECT I/O" Fred Isaman
2010-06-08  4:19         ` [PATCH 05/24] SQUASHME: ensure pnfs_update_lseg clears lsegp on error Fred Isaman
2010-06-08  4:19           ` [PATCH 06/24] pnfs: filelayout: clean and breakup nfs4_pnfs_dserver_get Fred Isaman
2010-06-08  4:19             ` [PATCH 07/24] pnfs: filelayout: remove some dead code from filelayout_commit Fred Isaman
2010-06-08  4:19               ` [PATCH 08/24] pnfs: remove PNFS_LAYOUTGET_ON_OPEN Fred Isaman
2010-06-08  4:19                 ` [PATCH 09/24] pnfs: track the number of outstanding commits Fred Isaman
2010-06-08  4:19                   ` [PATCH 10/24] pnfs_submit: mandate basic io path operations for layout drivers Fred Isaman
2010-06-08  4:19                     ` [PATCH 11/24] pnfs_submit: expose pnfs_update_layout, put_lseg, and get_lseg functions Fred Isaman
2010-06-08  4:19                       ` [PATCH 12/24] pnfs_submit: stash and refcount lseg in read path Fred Isaman
2010-06-08  4:19                         ` [PATCH 13/24] pnfs_submit: read path changeover Fred Isaman
2010-06-08  4:19                           ` [PATCH 14/24] pnfs_submit: use fsdata to pass lseg Fred Isaman
2010-06-08  4:19                             ` [PATCH 15/24] pnfs_submit: stash and refcount lseg in write path Fred Isaman
2010-06-08  4:19                               ` [PATCH 16/24] pnfs_submit: remove pnfs_file_operations Fred Isaman
2010-06-08  4:19                                 ` [PATCH 17/24] pnfs_submit: remove pnfs_update_layout_commit Fred Isaman
2010-06-08  4:19                                   ` [PATCH 18/24] pnfs_submit: remove pnfs_writepages LAYOUTGET invocation Fred Isaman
2010-06-08  4:19                                     ` [PATCH 19/24] pnfs: export some commit error handling for use by layout drivers Fred Isaman
2010-06-08  4:19                                       ` [PATCH 20/24] pnfs_submit: API change: remove pnfs_commit layoutget invocation Fred Isaman
2010-06-08  4:19                                         ` [PATCH 21/24] pnfs_submit: filelayout: rewrite filelayout_commit to use new API Fred Isaman
2010-06-08  4:19                                           ` [PATCH 22/24] pnfs_submit: remove unecessary pnfs_fl_call_data field pnfs_client Fred Isaman
2010-06-08  4:19                                             ` [PATCH 23/24] pnfs_submit: remove unecessary pnfs_fl_call_data field commit_through_mds Fred Isaman
2010-06-08  4:19                                               ` [PATCH 24/24] pnfs_submit: pnfs_update_layout can return void Fred Isaman
2010-06-09  9:09                                         ` [PATCH 20/24] pnfs_submit: API change: remove pnfs_commit layoutget invocation Benny Halevy
2010-06-09 12:21                                           ` Fred Isaman
2010-06-09 15:12                                             ` Boaz Harrosh
2010-06-09 15:15                                               ` Boaz Harrosh [this message]
2010-06-08  7:34                                 ` [PATCH 16/24] pnfs_submit: remove pnfs_file_operations Christoph Hellwig
2010-06-09 10:38                             ` [PATCH 14/24] pnfs_submit: use fsdata to pass lseg Benny Halevy
2010-06-09 12:08                               ` Fred Isaman
2010-06-10 10:33                                 ` Fred Isaman
2010-06-10 12:45                                   ` Benny Halevy
2010-06-10 12:48                                     ` Benny Halevy
2010-06-10 13:09                                       ` Boaz Harrosh
2010-06-09 19:33                             ` Boaz Harrosh
2010-06-09 19:19                           ` [PATCH 13/24] pnfs_submit: read path changeover Boaz Harrosh
2010-06-09 19:29                             ` Fred Isaman
     [not found]                               ` <AANLkTilecdPbSOJCDkGYH-X25gcZB-1fmBmU9mEpFO_y-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2010-06-09 19:39                                 ` Boaz Harrosh
2010-06-09 19:46                                   ` Fred Isaman
2010-06-10  6:26                                     ` Boaz Harrosh
2010-06-09 18:58                       ` [PATCH 11/24] pnfs_submit: expose pnfs_update_layout, put_lseg, and get_lseg functions Boaz Harrosh
2010-06-09 19:20                         ` Fred Isaman
2010-06-09 18:18           ` [PATCH 05/24] SQUASHME: ensure pnfs_update_lseg clears lsegp on error Boaz Harrosh
2010-06-09 18:06   ` [PATCH 01/24] Revert "pnfs-nonfilelayout: Prelim support for non-file layout O_DIRECT" Boaz Harrosh

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=4C0FB016.3020208@panasas.com \
    --to=bharrosh@panasas.com \
    --cc=bhalevy@panasas.com \
    --cc=iisaman@netapp.com \
    --cc=iisaman@umich.edu \
    --cc=linux-nfs@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