Linux-NFS Archive on lore.kernel.org
 help / color / Atom feed
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
To: trond.myklebust@hammerspace.com, anna.schumaker@netapp.com,
	bfields@redhat.com
Cc: linux-nfs@vger.kernel.org
Subject: [PATCH v8 11/20] NFS based on file size issue sync copy or fallback to generic copy offload
Date: Thu, 10 Oct 2019 08:46:13 -0400
Message-ID: <20191010124622.27812-12-olga.kornievskaia@gmail.com> (raw)
In-Reply-To: <20191010124622.27812-1-olga.kornievskaia@gmail.com>

From: Olga Kornievskaia <kolga@netapp.com>

For small file sizes, it make sense to issue a synchronous copy (and
save an RPC callback operation). Also, for the inter copy offload,
copy len must be larger than the cost of doing a mount between the
destination and source server (14RPCs are sent during 4.x mount).

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
---
 fs/nfs/nfs42.h     |  2 +-
 fs/nfs/nfs42proc.c |  4 ++--
 fs/nfs/nfs4file.c  | 16 +++++++++++++++-
 3 files changed, 18 insertions(+), 4 deletions(-)

diff --git a/fs/nfs/nfs42.h b/fs/nfs/nfs42.h
index 02e3810..c891af9 100644
--- a/fs/nfs/nfs42.h
+++ b/fs/nfs/nfs42.h
@@ -16,7 +16,7 @@
 #ifdef CONFIG_NFS_V4_2
 int nfs42_proc_allocate(struct file *, loff_t, loff_t);
 ssize_t nfs42_proc_copy(struct file *, loff_t, struct file *, loff_t, size_t,
-			struct nl4_server *, nfs4_stateid *);
+			struct nl4_server *, nfs4_stateid *, bool);
 int nfs42_proc_deallocate(struct file *, loff_t, loff_t);
 loff_t nfs42_proc_llseek(struct file *, loff_t, int);
 int nfs42_proc_layoutstats_generic(struct nfs_server *,
diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c
index 9c7feac..aab6b7b 100644
--- a/fs/nfs/nfs42proc.c
+++ b/fs/nfs/nfs42proc.c
@@ -357,7 +357,7 @@ static ssize_t _nfs42_proc_copy(struct file *src,
 ssize_t nfs42_proc_copy(struct file *src, loff_t pos_src,
 			struct file *dst, loff_t pos_dst, size_t count,
 			struct nl4_server *nss,
-			nfs4_stateid *cnr_stateid)
+			nfs4_stateid *cnr_stateid, bool sync)
 {
 	struct nfs_server *server = NFS_SERVER(file_inode(dst));
 	struct nfs_lock_context *src_lock;
@@ -368,7 +368,7 @@ ssize_t nfs42_proc_copy(struct file *src, loff_t pos_src,
 		.dst_fh		= NFS_FH(file_inode(dst)),
 		.dst_pos	= pos_dst,
 		.count		= count,
-		.sync		= false,
+		.sync		= sync,
 	};
 	struct nfs42_copy_res res;
 	struct nfs4_exception src_exception = {
diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c
index 2af30b7..8978325 100644
--- a/fs/nfs/nfs4file.c
+++ b/fs/nfs/nfs4file.c
@@ -138,6 +138,7 @@ static ssize_t __nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
 	struct nl4_server *nss = NULL;
 	nfs4_stateid *cnrs = NULL;
 	ssize_t ret;
+	bool sync = false;
 
 	/* Only offload copy if superblock is the same */
 	if (file_inode(file_in)->i_sb != file_inode(file_out)->i_sb)
@@ -146,8 +147,21 @@ static ssize_t __nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
 		return -EOPNOTSUPP;
 	if (file_inode(file_in) == file_inode(file_out))
 		return -EOPNOTSUPP;
+	/* if the copy size if smaller than 2 RPC payloads, make it
+	 * synchronous
+	 */
+	if (count <= 2 * NFS_SERVER(file_inode(file_in))->rsize)
+		sync = true;
 retry:
 	if (!nfs42_files_from_same_server(file_in, file_out)) {
+		/* for inter copy, if copy size if smaller than 12 RPC
+		 * payloads, fallback to traditional copy. There are
+		 * 14 RPCs during an NFSv4.x mount between source/dest
+		 * servers.
+		 */
+		if (sync ||
+			count <= 14 * NFS_SERVER(file_inode(file_in))->rsize)
+			return -EOPNOTSUPP;
 		cn_resp = kzalloc(sizeof(struct nfs42_copy_notify_res),
 				GFP_NOFS);
 		if (unlikely(cn_resp == NULL))
@@ -162,7 +176,7 @@ static ssize_t __nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
 		cnrs = &cn_resp->cnr_stateid;
 	}
 	ret = nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count,
-				nss, cnrs);
+				nss, cnrs, sync);
 out:
 	kfree(cn_resp);
 	if (ret == -EAGAIN)
-- 
1.8.3.1


  parent reply index

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-10 12:46 [PATCH v8 00/20] client and server support for "inter" SSC copy Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 01/20] NFS NFSD: defining nl4_servers structure needed by both Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 02/20] NFS: add COPY_NOTIFY operation Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 03/20] NFS: add ca_source_server<> to COPY Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 04/20] NFS: inter ssc open Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 05/20] NFS: skip recovery of copy open on dest server Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 06/20] NFS: for "inter" copy treat ESTALE as ENOTSUPP Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 07/20] NFS: COPY handle ERR_OFFLOAD_DENIED Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 08/20] NFS: also send OFFLOAD_CANCEL to source server Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 09/20] NFS handle NFS4ERR_PARTNER_NO_AUTH error Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 10/20] NFS: handle source server reboot Olga Kornievskaia
2019-10-10 12:46 ` Olga Kornievskaia [this message]
2019-10-10 12:46 ` [PATCH v8 12/20] NFS: replace cross device check in copy_file_range Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 13/20] NFSD fill-in netloc4 structure Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 14/20] NFSD add ca_source_server<> to COPY Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 15/20] NFSD COPY_NOTIFY xdr Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 16/20] NFSD add COPY_NOTIFY operation Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 17/20] NFSD check stateids against copy stateids Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 18/20] NFSD generalize nfsd4_compound_state flag names Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 19/20] NFSD: allow inter server COPY to have a STALE source server fh Olga Kornievskaia
2019-10-10 12:46 ` [PATCH v8 20/20] NFSD add nfs4 inter ssc to nfsd4_copy Olga Kornievskaia

Reply instructions:

You may reply publically 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=20191010124622.27812-12-olga.kornievskaia@gmail.com \
    --to=olga.kornievskaia@gmail.com \
    --cc=anna.schumaker@netapp.com \
    --cc=bfields@redhat.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=trond.myklebust@hammerspace.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

Linux-NFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-nfs/0 linux-nfs/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-nfs linux-nfs/ https://lore.kernel.org/linux-nfs \
		linux-nfs@vger.kernel.org
	public-inbox-index linux-nfs

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-nfs


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git