All of lore.kernel.org
 help / color / mirror / Atom feed
From: Trond Myklebust <trond.myklebust@primarydata.com>
To: linux-nfs@vger.kernel.org
Cc: Olga.Kornievskaia@netapp.com
Subject: [PATCH 2/2] NFSv4.1: Ask for no delegation on OPEN if already holding one
Date: Tue,  3 Feb 2015 16:59:44 -0500	[thread overview]
Message-ID: <1423000784-93180-2-git-send-email-trond.myklebust@primarydata.com> (raw)
In-Reply-To: <1423000784-93180-1-git-send-email-trond.myklebust@primarydata.com>

If we already hold a delegation, there should be no reason for the
server to issue it to us again. Unfortunately, there appear to be
servers out there that engage in this practice. While it is often
harmless to do so, there is one case where this creates a problem
and that is when the client is in the process of returning that
delegation.
This patch uses the NFSv4.1 NFS4_SHARE_WANT_NO_DELEG flag to inform
the server not to return a delegation in these cases.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
---
 fs/nfs/nfs4proc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index cd4295d84d54..fb41624bafc9 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -942,8 +942,10 @@ static bool nfs4_clear_cap_atomic_open_v1(struct nfs_server *server,
 
 static u32
 nfs4_map_atomic_open_share(struct nfs_server *server,
+		struct inode *inode,
 		fmode_t fmode, int openflags)
 {
+	struct nfs_delegation *delegation;
 	u32 res = 0;
 
 	switch (fmode & (FMODE_READ | FMODE_WRITE)) {
@@ -959,8 +961,25 @@ nfs4_map_atomic_open_share(struct nfs_server *server,
 	if (!(server->caps & NFS_CAP_ATOMIC_OPEN_V1))
 		goto out;
 	/* Want no delegation if we're using O_DIRECT */
-	if (openflags & O_DIRECT)
+	if (openflags & O_DIRECT) {
 		res |= NFS4_SHARE_WANT_NO_DELEG;
+		goto out;
+	}
+	if (inode == NULL)
+		goto out;
+	rcu_read_lock();
+	delegation = rcu_dereference(NFS_I(inode)->delegation);
+	/*
+	 * If we have a delegation, either ask for an upgrade or ask for
+	 * no delegation
+	 */
+	if (delegation) {
+		if ((fmode & FMODE_WRITE) && !(delegation->type & FMODE_WRITE))
+			res |= NFS4_SHARE_WANT_WRITE_DELEG;
+		else
+			res |= NFS4_SHARE_WANT_NO_DELEG;
+	}
+	rcu_read_unlock();
 out:
 	return res;
 }
@@ -1028,7 +1047,7 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
 	p->o_arg.open_flags = flags;
 	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
 	p->o_arg.share_access = nfs4_map_atomic_open_share(server,
-			fmode, flags);
+			dentry->d_inode, fmode, flags);
 	/* don't put an ACCESS op in OPEN compound if O_EXCL, because ACCESS
 	 * will return permission denied for all bits until close */
 	if (!(flags & O_EXCL)) {
@@ -2724,7 +2743,7 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data)
 	}
 	calldata->arg.share_access =
 		nfs4_map_atomic_open_share(NFS_SERVER(inode),
-				calldata->arg.fmode, 0);
+				NULL, calldata->arg.fmode, 0);
 
 	nfs_fattr_init(calldata->res.fattr);
 	calldata->timestamp = jiffies;
-- 
2.1.0


  reply	other threads:[~2015-02-03 21:59 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-03 21:59 [PATCH 1/2] NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT Trond Myklebust
2015-02-03 21:59 ` Trond Myklebust [this message]
2015-02-03 22:47   ` [PATCH 2/2] NFSv4.1: Ask for no delegation on OPEN if already holding one Kornievskaia, Olga
2015-02-04  1:04     ` Trond Myklebust
2015-02-05 14:07   ` Christoph Hellwig
2015-02-05 14:49     ` Trond Myklebust
2015-02-04  8:16 ` [PATCH 1/2] NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT Christoph Hellwig
2015-02-04 12:32   ` Trond Myklebust
2015-02-05 11:27     ` Christoph Hellwig

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=1423000784-93180-2-git-send-email-trond.myklebust@primarydata.com \
    --to=trond.myklebust@primarydata.com \
    --cc=Olga.Kornievskaia@netapp.com \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.