From ba8ec7a607e98e8491a1fcf924a2e6c96ac9d413 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Sat, 23 Jul 2022 14:47:28 -0400
Subject: SUNRPC: Shrink size of struct rpc_task

Move the field 'tk_rpc_status' so that we eliminate a 4 byte hole in the
structure.
For x86_64, this shrinks the size of the struct by 8 bytes.

'pahole' output before the change:
        /* size: 232, cachelines: 4, members: 27 */
        /* sum members: 222, holes: 1, sum holes: 4 */
        /* sum bitfield members: 8 bits (1 bytes) */
        /* padding: 5 */
        /* last cacheline: 40 bytes */

'pahole' output after the change:
        /* size: 224, cachelines: 4, members: 27 */
        /* padding: 1 */
        /* last cacheline: 32 bytes */

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/sched.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
index 1d7a3e51b795..acc62647317c 100644
--- a/include/linux/sunrpc/sched.h
+++ b/include/linux/sunrpc/sched.h
@@ -61,8 +61,6 @@ struct rpc_task {
 		struct rpc_wait		tk_wait;	/* RPC wait */
 	} u;
 
-	int			tk_rpc_status;	/* Result of last RPC operation */
-
 	/*
 	 * RPC call state
 	 */
@@ -82,6 +80,8 @@ struct rpc_task {
 	ktime_t			tk_start;	/* RPC task init timestamp */
 
 	pid_t			tk_owner;	/* Process id for batching tasks */
+
+	int			tk_rpc_status;	/* Result of last RPC operation */
 	unsigned short		tk_flags;	/* misc flags */
 	unsigned short		tk_timeouts;	/* maj timeouts */
 
-- 
cgit v1.2.3


From 69d966510d9f5de81588b37d23a9ee8ccc477b23 Mon Sep 17 00:00:00 2001
From: Jeff Layton <jlayton@kernel.org>
Date: Fri, 22 Jul 2022 14:12:20 -0400
Subject: nfs: only issue commit in DIO codepath if we have uncommitted data

Currently, we try to determine whether to issue a commit based on
nfs_write_need_commit which looks at the current verifier. In the case
where we got a short write and then tried to follow it up with one that
failed, the verifier can't be trusted.

What we really want to know is whether the pgio request had any
successful writes that came back as UNSTABLE. Add a new flag to the pgio
request, and use that to indicate that we've had a successful unstable
write. Only issue a commit if that flag is set.

Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/direct.c         |  2 +-
 fs/nfs/write.c          | 48 ++++++++++++++++++++++++++++++------------------
 include/linux/nfs_xdr.h |  1 +
 3 files changed, 32 insertions(+), 19 deletions(-)

(limited to 'include/linux')

diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index a47d13296194..86df66bb14c5 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -690,7 +690,7 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 	}
 
 	nfs_direct_count_bytes(dreq, hdr);
-	if (hdr->good_bytes != 0 && nfs_write_need_commit(hdr)) {
+	if (test_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags)) {
 		if (!dreq->flags)
 			dreq->flags = NFS_ODIRECT_DO_COMMIT;
 		flags = dreq->flags;
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 1c706465d090..16d166bc4099 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1576,25 +1576,37 @@ static int nfs_writeback_done(struct rpc_task *task,
 	nfs_add_stats(inode, NFSIOS_SERVERWRITTENBYTES, hdr->res.count);
 	trace_nfs_writeback_done(task, hdr);
 
-	if (hdr->res.verf->committed < hdr->args.stable &&
-	    task->tk_status >= 0) {
-		/* We tried a write call, but the server did not
-		 * commit data to stable storage even though we
-		 * requested it.
-		 * Note: There is a known bug in Tru64 < 5.0 in which
-		 *	 the server reports NFS_DATA_SYNC, but performs
-		 *	 NFS_FILE_SYNC. We therefore implement this checking
-		 *	 as a dprintk() in order to avoid filling syslog.
-		 */
-		static unsigned long    complain;
+	if (task->tk_status >= 0) {
+		enum nfs3_stable_how committed = hdr->res.verf->committed;
+
+		if (committed == NFS_UNSTABLE) {
+			/*
+			 * We have some uncommitted data on the server at
+			 * this point, so ensure that we keep track of that
+			 * fact irrespective of what later writes do.
+			 */
+			set_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags);
+		}
 
-		/* Note this will print the MDS for a DS write */
-		if (time_before(complain, jiffies)) {
-			dprintk("NFS:       faulty NFS server %s:"
-				" (committed = %d) != (stable = %d)\n",
-				NFS_SERVER(inode)->nfs_client->cl_hostname,
-				hdr->res.verf->committed, hdr->args.stable);
-			complain = jiffies + 300 * HZ;
+		if (committed < hdr->args.stable) {
+			/* We tried a write call, but the server did not
+			 * commit data to stable storage even though we
+			 * requested it.
+			 * Note: There is a known bug in Tru64 < 5.0 in which
+			 *	 the server reports NFS_DATA_SYNC, but performs
+			 *	 NFS_FILE_SYNC. We therefore implement this checking
+			 *	 as a dprintk() in order to avoid filling syslog.
+			 */
+			static unsigned long    complain;
+
+			/* Note this will print the MDS for a DS write */
+			if (time_before(complain, jiffies)) {
+				dprintk("NFS:       faulty NFS server %s:"
+					" (committed = %d) != (stable = %d)\n",
+					NFS_SERVER(inode)->nfs_client->cl_hostname,
+					committed, hdr->args.stable);
+				complain = jiffies + 300 * HZ;
+			}
 		}
 	}
 
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 0e3aa0f5f324..e86cf6642d21 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1600,6 +1600,7 @@ enum {
 	NFS_IOHDR_STAT,
 	NFS_IOHDR_RESEND_PNFS,
 	NFS_IOHDR_RESEND_MDS,
+	NFS_IOHDR_UNSTABLE_WRITES,
 };
 
 struct nfs_io_completion;
-- 
cgit v1.2.3


From 4f5f3b6028343d687d0533329b130e4b8280ab32 Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 21 Jul 2022 14:21:31 -0400
Subject: SUNRPC: Introduce xdr_stream_move_subsegment()

I do this by creating an xdr subsegment for the range we will be
operating over. This lets me shift data to the correct place without
potentially overwriting anything already there.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xdr.h |  2 ++
 net/sunrpc/xdr.c           | 59 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 61 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index 5860f32e3958..7dcc6c31fe29 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -262,6 +262,8 @@ extern unsigned int xdr_align_data(struct xdr_stream *, unsigned int offset, uns
 extern unsigned int xdr_expand_hole(struct xdr_stream *, unsigned int offset, unsigned int length);
 extern bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
 				  unsigned int len);
+extern unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
+					       unsigned int target, unsigned int length);
 
 /**
  * xdr_set_scratch_buffer - Attach a scratch buffer for decoding data.
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index 5d2b3e6979fb..8ba11a754297 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -775,6 +775,34 @@ static void xdr_buf_pages_shift_left(const struct xdr_buf *buf,
 	xdr_buf_tail_copy_left(buf, 0, len - buf->page_len, shift);
 }
 
+static void xdr_buf_head_shift_left(const struct xdr_buf *buf,
+				    unsigned int base, unsigned int len,
+				    unsigned int shift)
+{
+	const struct kvec *head = buf->head;
+	unsigned int bytes;
+
+	if (!shift || !len)
+		return;
+
+	if (shift > base) {
+		bytes = (shift - base);
+		if (bytes >= len)
+			return;
+		base += bytes;
+		len -= bytes;
+	}
+
+	if (base < head->iov_len) {
+		bytes = min_t(unsigned int, len, head->iov_len - base);
+		memmove(head->iov_base + (base - shift),
+			head->iov_base + base, bytes);
+		base += bytes;
+		len -= bytes;
+	}
+	xdr_buf_pages_shift_left(buf, base - head->iov_len, len, shift);
+}
+
 /**
  * xdr_shrink_bufhead
  * @buf: xdr_buf
@@ -1680,6 +1708,37 @@ bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
 }
 EXPORT_SYMBOL_GPL(xdr_stream_subsegment);
 
+/**
+ * xdr_stream_move_subsegment - Move part of a stream to another position
+ * @xdr: the source xdr_stream
+ * @offset: the source offset of the segment
+ * @target: the target offset of the segment
+ * @length: the number of bytes to move
+ *
+ * Moves @length bytes from @offset to @target in the xdr_stream, overwriting
+ * anything in its space. Returns the number of bytes in the segment.
+ */
+unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
+					unsigned int target, unsigned int length)
+{
+	struct xdr_buf buf;
+	unsigned int shift;
+
+	if (offset < target) {
+		shift = target - offset;
+		if (xdr_buf_subsegment(xdr->buf, &buf, offset, shift + length) < 0)
+			return 0;
+		xdr_buf_head_shift_right(&buf, 0, length, shift);
+	} else if (offset > target) {
+		shift = offset - target;
+		if (xdr_buf_subsegment(xdr->buf, &buf, target, shift + length) < 0)
+			return 0;
+		xdr_buf_head_shift_left(&buf, shift, length, shift);
+	}
+	return length;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_move_subsegment);
+
 /**
  * xdr_buf_trim - lop at most "len" bytes off the end of "buf"
  * @buf: buf to be trimmed
-- 
cgit v1.2.3


From 7c4cd5f4d2dd4a028a46bfb696b0cd387caadf33 Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 21 Jul 2022 14:21:32 -0400
Subject: SUNRPC: Add a function for directly setting the xdr page len

We need to do this step during READ_PLUS decoding so that we know pages
are the right length and any extra data has been preserved in the tail.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xdr.h |  1 +
 net/sunrpc/xdr.c           | 30 ++++++++++++++++++++++++++++++
 2 files changed, 31 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index 7dcc6c31fe29..8cd38a9994ca 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -258,6 +258,7 @@ extern __be32 *xdr_inline_decode(struct xdr_stream *xdr, size_t nbytes);
 extern unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len);
 extern void xdr_enter_page(struct xdr_stream *xdr, unsigned int len);
 extern int xdr_process_buf(const struct xdr_buf *buf, unsigned int offset, unsigned int len, int (*actor)(struct scatterlist *, void *), void *data);
+extern void xdr_set_pagelen(struct xdr_stream *, unsigned int len);
 extern unsigned int xdr_align_data(struct xdr_stream *, unsigned int offset, unsigned int length);
 extern unsigned int xdr_expand_hole(struct xdr_stream *, unsigned int offset, unsigned int length);
 extern bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index 8ba11a754297..e4ac700ca554 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -1500,6 +1500,36 @@ unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len)
 }
 EXPORT_SYMBOL_GPL(xdr_read_pages);
 
+/**
+ * xdr_set_pagelen - Sets the length of the XDR pages
+ * @xdr: pointer to xdr_stream struct
+ * @len: new length of the XDR page data
+ *
+ * Either grows or shrinks the length of the xdr pages by setting pagelen to
+ * @len bytes. When shrinking, any extra data is moved into buf->tail, whereas
+ * when growing any data beyond the current pointer is moved into the tail.
+ *
+ * Returns True if the operation was successful, and False otherwise.
+ */
+void xdr_set_pagelen(struct xdr_stream *xdr, unsigned int len)
+{
+	struct xdr_buf *buf = xdr->buf;
+	size_t remaining = xdr_stream_remaining(xdr);
+	size_t base = 0;
+
+	if (len < buf->page_len) {
+		base = buf->page_len - len;
+		xdr_shrink_pagelen(buf, len);
+	} else {
+		xdr_buf_head_shift_right(buf, xdr_stream_pos(xdr),
+					 buf->page_len, remaining);
+		if (len > buf->page_len)
+			xdr_buf_try_expand(buf, len - buf->page_len);
+	}
+	xdr_set_tail_base(xdr, base, remaining);
+}
+EXPORT_SYMBOL_GPL(xdr_set_pagelen);
+
 unsigned int xdr_align_data(struct xdr_stream *xdr, unsigned int offset,
 			    unsigned int length)
 {
-- 
cgit v1.2.3


From e1bd87608d4b6f87813f79b91e834de610f1049b Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 21 Jul 2022 14:21:33 -0400
Subject: SUNRPC: Add a function for zeroing out a portion of an xdr_stream

This will be used during READ_PLUS decoding for handling HOLE segments.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xdr.h |  2 ++
 net/sunrpc/xdr.c           | 23 +++++++++++++++++++++++
 2 files changed, 25 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index 8cd38a9994ca..f0ab06acab61 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -265,6 +265,8 @@ extern bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf
 				  unsigned int len);
 extern unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
 					       unsigned int target, unsigned int length);
+extern unsigned int xdr_stream_zero(struct xdr_stream *xdr, unsigned int offset,
+				    unsigned int length);
 
 /**
  * xdr_set_scratch_buffer - Attach a scratch buffer for decoding data.
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index e4ac700ca554..f09a7ab1a82b 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -1769,6 +1769,29 @@ unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int off
 }
 EXPORT_SYMBOL_GPL(xdr_stream_move_subsegment);
 
+/**
+ * xdr_stream_zero - zero out a portion of an xdr_stream
+ * @xdr: an xdr_stream to zero out
+ * @offset: the starting point in the stream
+ * @length: the number of bytes to zero
+ */
+unsigned int xdr_stream_zero(struct xdr_stream *xdr, unsigned int offset,
+			     unsigned int length)
+{
+	struct xdr_buf buf;
+
+	if (xdr_buf_subsegment(xdr->buf, &buf, offset, length) < 0)
+		return 0;
+	if (buf.head[0].iov_len)
+		xdr_buf_iov_zero(buf.head, 0, buf.head[0].iov_len);
+	if (buf.page_len > 0)
+		xdr_buf_pages_zero(&buf, 0, buf.page_len);
+	if (buf.tail[0].iov_len)
+		xdr_buf_iov_zero(buf.tail, 0, buf.tail[0].iov_len);
+	return length;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_zero);
+
 /**
  * xdr_buf_trim - lop at most "len" bytes off the end of "buf"
  * @buf: buf to be trimmed
-- 
cgit v1.2.3


From 29946fbcb2c31a2a367887dc58a2e7e5b012e285 Mon Sep 17 00:00:00 2001
From: Anna Schumaker <Anna.Schumaker@Netapp.com>
Date: Thu, 21 Jul 2022 14:21:35 -0400
Subject: SUNRPC: Remove xdr_align_data() and xdr_expand_hole()

These functions are no longer needed now that the NFS client places data
and hole segments directly.

Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xdr.h |  2 --
 net/sunrpc/xdr.c           | 66 ----------------------------------------------
 2 files changed, 68 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index f0ab06acab61..f38c97f45354 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -259,8 +259,6 @@ extern unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len);
 extern void xdr_enter_page(struct xdr_stream *xdr, unsigned int len);
 extern int xdr_process_buf(const struct xdr_buf *buf, unsigned int offset, unsigned int len, int (*actor)(struct scatterlist *, void *), void *data);
 extern void xdr_set_pagelen(struct xdr_stream *, unsigned int len);
-extern unsigned int xdr_align_data(struct xdr_stream *, unsigned int offset, unsigned int length);
-extern unsigned int xdr_expand_hole(struct xdr_stream *, unsigned int offset, unsigned int length);
 extern bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
 				  unsigned int len);
 extern unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index f09a7ab1a82b..482586c23fdd 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -1530,72 +1530,6 @@ void xdr_set_pagelen(struct xdr_stream *xdr, unsigned int len)
 }
 EXPORT_SYMBOL_GPL(xdr_set_pagelen);
 
-unsigned int xdr_align_data(struct xdr_stream *xdr, unsigned int offset,
-			    unsigned int length)
-{
-	struct xdr_buf *buf = xdr->buf;
-	unsigned int from, bytes, len;
-	unsigned int shift;
-
-	xdr_realign_pages(xdr);
-	from = xdr_page_pos(xdr);
-
-	if (from >= buf->page_len + buf->tail->iov_len)
-		return 0;
-	if (from + buf->head->iov_len >= buf->len)
-		return 0;
-
-	len = buf->len - buf->head->iov_len;
-
-	/* We only shift data left! */
-	if (WARN_ONCE(from < offset, "SUNRPC: misaligned data src=%u dst=%u\n",
-		      from, offset))
-		return 0;
-	if (WARN_ONCE(offset > buf->page_len,
-		      "SUNRPC: buffer overflow. offset=%u, page_len=%u\n",
-		      offset, buf->page_len))
-		return 0;
-
-	/* Move page data to the left */
-	shift = from - offset;
-	xdr_buf_pages_shift_left(buf, from, len, shift);
-
-	bytes = xdr_stream_remaining(xdr);
-	if (length > bytes)
-		length = bytes;
-	bytes -= length;
-
-	xdr->buf->len -= shift;
-	xdr_set_page(xdr, offset + length, bytes);
-	return length;
-}
-EXPORT_SYMBOL_GPL(xdr_align_data);
-
-unsigned int xdr_expand_hole(struct xdr_stream *xdr, unsigned int offset,
-			     unsigned int length)
-{
-	struct xdr_buf *buf = xdr->buf;
-	unsigned int from, to, shift;
-
-	xdr_realign_pages(xdr);
-	from = xdr_page_pos(xdr);
-	to = xdr_align_size(offset + length);
-
-	/* Could the hole be behind us? */
-	if (to > from) {
-		unsigned int buflen = buf->len - buf->head->iov_len;
-		shift = to - from;
-		xdr_buf_try_expand(buf, shift);
-		xdr_buf_pages_shift_right(buf, from, buflen, shift);
-		xdr_set_page(xdr, to, xdr_stream_remaining(xdr));
-	} else if (to != from)
-		xdr_align_data(xdr, to, 0);
-	xdr_buf_pages_zero(buf, offset, length);
-
-	return length;
-}
-EXPORT_SYMBOL_GPL(xdr_expand_hole);
-
 /**
  * xdr_enter_page - decode data from the XDR page
  * @xdr: pointer to xdr_stream struct
-- 
cgit v1.2.3


From 7ffcdaa670164a2ad3844a5ef6df5423782ba290 Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:21 -0400
Subject: SUNRPC expose functions for offline remote xprt functionality

Re-arrange the code that make offline transport and delete transport
callable functions.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xprt.h |  3 +++
 net/sunrpc/sysfs.c          | 28 +++++-----------------------
 net/sunrpc/xprt.c           | 32 ++++++++++++++++++++++++++++++++
 3 files changed, 40 insertions(+), 23 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index 522bbf937957..0d51b9f9ea37 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -505,4 +505,7 @@ static inline int xprt_test_and_set_binding(struct rpc_xprt *xprt)
 	return test_and_set_bit(XPRT_BINDING, &xprt->state);
 }
 
+void xprt_set_offline_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
+void xprt_set_online_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
+void xprt_delete_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
 #endif /* _LINUX_SUNRPC_XPRT_H */
diff --git a/net/sunrpc/sysfs.c b/net/sunrpc/sysfs.c
index a3a2f8aeb80e..7330eb9a70cf 100644
--- a/net/sunrpc/sysfs.c
+++ b/net/sunrpc/sysfs.c
@@ -314,32 +314,14 @@ static ssize_t rpc_sysfs_xprt_state_change(struct kobject *kobj,
 		goto release_tasks;
 	}
 	if (offline) {
-		if (!test_and_set_bit(XPRT_OFFLINE, &xprt->state)) {
-			spin_lock(&xps->xps_lock);
-			xps->xps_nactive--;
-			spin_unlock(&xps->xps_lock);
-		}
+		xprt_set_offline_locked(xprt, xps);
 	} else if (online) {
-		if (test_and_clear_bit(XPRT_OFFLINE, &xprt->state)) {
-			spin_lock(&xps->xps_lock);
-			xps->xps_nactive++;
-			spin_unlock(&xps->xps_lock);
-		}
+		xprt_set_online_locked(xprt, xps);
 	} else if (remove) {
-		if (test_bit(XPRT_OFFLINE, &xprt->state)) {
-			if (!test_and_set_bit(XPRT_REMOVE, &xprt->state)) {
-				xprt_force_disconnect(xprt);
-				if (test_bit(XPRT_CONNECTED, &xprt->state)) {
-					if (!xprt->sending.qlen &&
-					    !xprt->pending.qlen &&
-					    !xprt->backlog.qlen &&
-					    !atomic_long_read(&xprt->queuelen))
-						rpc_xprt_switch_remove_xprt(xps, xprt);
-				}
-			}
-		} else {
+		if (test_bit(XPRT_OFFLINE, &xprt->state))
+			xprt_delete_locked(xprt, xps);
+		else
 			count = -EINVAL;
-		}
 	}
 
 release_tasks:
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 86d62cffba0d..8f8e3c952f24 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -2152,3 +2152,35 @@ void xprt_put(struct rpc_xprt *xprt)
 		kref_put(&xprt->kref, xprt_destroy_kref);
 }
 EXPORT_SYMBOL_GPL(xprt_put);
+
+void xprt_set_offline_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (!test_and_set_bit(XPRT_OFFLINE, &xprt->state)) {
+		spin_lock(&xps->xps_lock);
+		xps->xps_nactive--;
+		spin_unlock(&xps->xps_lock);
+	}
+}
+
+void xprt_set_online_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (test_and_clear_bit(XPRT_OFFLINE, &xprt->state)) {
+		spin_lock(&xps->xps_lock);
+		xps->xps_nactive++;
+		spin_unlock(&xps->xps_lock);
+	}
+}
+
+void xprt_delete_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (test_and_set_bit(XPRT_REMOVE, &xprt->state))
+		return;
+
+	xprt_force_disconnect(xprt);
+	if (!test_bit(XPRT_CONNECTED, &xprt->state))
+		return;
+
+	if (!xprt->sending.qlen && !xprt->pending.qlen &&
+	    !xprt->backlog.qlen && !atomic_long_read(&xprt->queuelen))
+		rpc_xprt_switch_remove_xprt(xps, xprt);
+}
-- 
cgit v1.2.3


From 895245ccea251ff54ea19bc364c9a49007918115 Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:22 -0400
Subject: SUNRPC add function to offline remove trunkable transports

Iterate thru available transports in the xprt_switch for all
trunkable transports offline and possibly remote them as well.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/clnt.h |  1 +
 net/sunrpc/clnt.c           | 46 +++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 47 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 90501404fa49..d14333f4947a 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -234,6 +234,7 @@ int		rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *,
 			struct rpc_xprt_switch *,
 			struct rpc_xprt *,
 			void *);
+void		rpc_clnt_manage_trunked_xprts(struct rpc_clnt *);
 
 const char *rpc_proc_name(const struct rpc_task *task);
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index a97d4e06cae3..2b079c4d8af1 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -3000,6 +3000,52 @@ out_put_switch:
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_add_xprt);
 
+static int rpc_xprt_offline(struct rpc_clnt *clnt,
+			    struct rpc_xprt *xprt,
+			    void *data)
+{
+	struct rpc_xprt *main_xprt;
+	struct rpc_xprt_switch *xps;
+	int err = 0;
+
+	xprt_get(xprt);
+
+	rcu_read_lock();
+	main_xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
+	xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+	err = rpc_cmp_addr_port((struct sockaddr *)&xprt->addr,
+				(struct sockaddr *)&main_xprt->addr);
+	rcu_read_unlock();
+	xprt_put(main_xprt);
+	if (err)
+		goto out;
+
+	if (wait_on_bit_lock(&xprt->state, XPRT_LOCKED, TASK_KILLABLE)) {
+		err = -EINTR;
+		goto out;
+	}
+	xprt_set_offline_locked(xprt, xps);
+
+	xprt_release_write(xprt, NULL);
+out:
+	xprt_put(xprt);
+	xprt_switch_put(xps);
+	return err;
+}
+
+/* rpc_clnt_manage_trunked_xprts -- offline trunked transports
+ * @clnt rpc_clnt structure
+ *
+ * For each active transport found in the rpc_clnt structure call
+ * the function rpc_xprt_offline() which will identify trunked transports
+ * and will mark them offline.
+ */
+void rpc_clnt_manage_trunked_xprts(struct rpc_clnt *clnt)
+{
+	rpc_clnt_iterate_for_each_xprt(clnt, rpc_xprt_offline, NULL);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_manage_trunked_xprts);
+
 struct connect_timeout_data {
 	unsigned long connect_timeout;
 	unsigned long reconnect_timeout;
-- 
cgit v1.2.3


From 95d0d30c66b855f614e677b8cd0455eed0765a6f Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:24 -0400
Subject: SUNRPC create an iterator to list only OFFLINE xprts

Create a new iterator helper that will go thru the all the transports
in the switch and return transports that are marked OFFLINE.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xprtmultipath.h |  3 ++
 net/sunrpc/clnt.c                    | 11 +++-
 net/sunrpc/xprtmultipath.c           | 99 ++++++++++++++++++++++++++++++++----
 3 files changed, 101 insertions(+), 12 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xprtmultipath.h b/include/linux/sunrpc/xprtmultipath.h
index bbb8a5fa0816..688ca7eb1d01 100644
--- a/include/linux/sunrpc/xprtmultipath.h
+++ b/include/linux/sunrpc/xprtmultipath.h
@@ -63,6 +63,9 @@ extern void xprt_iter_init(struct rpc_xprt_iter *xpi,
 extern void xprt_iter_init_listall(struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *xps);
 
+extern void xprt_iter_init_listoffline(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt_switch *xps);
+
 extern void xprt_iter_destroy(struct rpc_xprt_iter *xpi);
 
 extern struct rpc_xprt_switch *xprt_iter_xchg_switch(
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 2b079c4d8af1..68021b70340d 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -786,7 +786,8 @@ out_revert:
 EXPORT_SYMBOL_GPL(rpc_switch_client_transport);
 
 static
-int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
+int _rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi,
+			     void func(struct rpc_xprt_iter *xpi, struct rpc_xprt_switch *xps))
 {
 	struct rpc_xprt_switch *xps;
 
@@ -795,11 +796,17 @@ int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
 	rcu_read_unlock();
 	if (xps == NULL)
 		return -EAGAIN;
-	xprt_iter_init_listall(xpi, xps);
+	func(xpi, xps);
 	xprt_switch_put(xps);
 	return 0;
 }
 
+static
+int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
+{
+	return _rpc_clnt_xprt_iter_init(clnt, xpi, xprt_iter_init_listall);
+}
+
 /**
  * rpc_clnt_iterate_for_each_xprt - Apply a function to all transports
  * @clnt: pointer to client
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 1693f81aae37..8def8423fc0a 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -27,6 +27,7 @@ typedef struct rpc_xprt *(*xprt_switch_find_xprt_t)(struct rpc_xprt_switch *xps,
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_singular;
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_roundrobin;
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_listall;
+static const struct rpc_xprt_iter_ops rpc_xprt_iter_listoffline;
 
 static void xprt_switch_add_xprt_locked(struct rpc_xprt_switch *xps,
 		struct rpc_xprt *xprt)
@@ -248,6 +249,18 @@ struct rpc_xprt *xprt_switch_find_first_entry(struct list_head *head)
 	return NULL;
 }
 
+static
+struct rpc_xprt *xprt_switch_find_first_entry_offline(struct list_head *head)
+{
+	struct rpc_xprt *pos;
+
+	list_for_each_entry_rcu(pos, head, xprt_switch) {
+		if (!xprt_is_active(pos))
+			return pos;
+	}
+	return NULL;
+}
+
 static
 struct rpc_xprt *xprt_iter_first_entry(struct rpc_xprt_iter *xpi)
 {
@@ -259,8 +272,9 @@ struct rpc_xprt *xprt_iter_first_entry(struct rpc_xprt_iter *xpi)
 }
 
 static
-struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
-		const struct rpc_xprt *cur)
+struct rpc_xprt *_xprt_switch_find_current_entry(struct list_head *head,
+						 const struct rpc_xprt *cur,
+						 bool find_active)
 {
 	struct rpc_xprt *pos;
 	bool found = false;
@@ -268,14 +282,25 @@ struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
 	list_for_each_entry_rcu(pos, head, xprt_switch) {
 		if (cur == pos)
 			found = true;
-		if (found && xprt_is_active(pos))
+		if (found && ((find_active && xprt_is_active(pos)) ||
+			      (!find_active && xprt_is_active(pos))))
 			return pos;
 	}
 	return NULL;
 }
 
 static
-struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
+struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
+						const struct rpc_xprt *cur)
+{
+	return _xprt_switch_find_current_entry(head, cur, true);
+}
+
+static
+struct rpc_xprt * _xprt_iter_current_entry(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt *first_entry(struct list_head *head),
+		struct rpc_xprt *current_entry(struct list_head *head,
+					       const struct rpc_xprt *cur))
 {
 	struct rpc_xprt_switch *xps = rcu_dereference(xpi->xpi_xpswitch);
 	struct list_head *head;
@@ -284,8 +309,30 @@ struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
 		return NULL;
 	head = &xps->xps_xprt_list;
 	if (xpi->xpi_cursor == NULL || xps->xps_nxprts < 2)
-		return xprt_switch_find_first_entry(head);
-	return xprt_switch_find_current_entry(head, xpi->xpi_cursor);
+		return first_entry(head);
+	return current_entry(head, xpi->xpi_cursor);
+}
+
+static
+struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
+{
+	return _xprt_iter_current_entry(xpi, xprt_switch_find_first_entry,
+			xprt_switch_find_current_entry);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_current_entry_offline(struct list_head *head,
+		const struct rpc_xprt *cur)
+{
+	return _xprt_switch_find_current_entry(head, cur, false);
+}
+
+static
+struct rpc_xprt *xprt_iter_current_entry_offline(struct rpc_xprt_iter *xpi)
+{
+	return _xprt_iter_current_entry(xpi,
+			xprt_switch_find_first_entry_offline,
+			xprt_switch_find_current_entry_offline);
 }
 
 bool rpc_xprt_switch_has_addr(struct rpc_xprt_switch *xps,
@@ -310,7 +357,7 @@ bool rpc_xprt_switch_has_addr(struct rpc_xprt_switch *xps,
 
 static
 struct rpc_xprt *xprt_switch_find_next_entry(struct list_head *head,
-		const struct rpc_xprt *cur)
+		const struct rpc_xprt *cur, bool check_active)
 {
 	struct rpc_xprt *pos, *prev = NULL;
 	bool found = false;
@@ -318,7 +365,12 @@ struct rpc_xprt *xprt_switch_find_next_entry(struct list_head *head,
 	list_for_each_entry_rcu(pos, head, xprt_switch) {
 		if (cur == prev)
 			found = true;
-		if (found && xprt_is_active(pos))
+		/* for request to return active transports return only
+		 * active, for request to return offline transports
+		 * return only offline
+		 */
+		if (found && ((check_active && xprt_is_active(pos)) ||
+			      (!check_active && !xprt_is_active(pos))))
 			return pos;
 		prev = pos;
 	}
@@ -355,7 +407,7 @@ struct rpc_xprt *__xprt_switch_find_next_entry_roundrobin(struct list_head *head
 {
 	struct rpc_xprt *ret;
 
-	ret = xprt_switch_find_next_entry(head, cur);
+	ret = xprt_switch_find_next_entry(head, cur, true);
 	if (ret != NULL)
 		return ret;
 	return xprt_switch_find_first_entry(head);
@@ -397,7 +449,14 @@ static
 struct rpc_xprt *xprt_switch_find_next_entry_all(struct rpc_xprt_switch *xps,
 		const struct rpc_xprt *cur)
 {
-	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur);
+	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur, true);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_next_entry_offline(struct rpc_xprt_switch *xps,
+		const struct rpc_xprt *cur)
+{
+	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur, false);
 }
 
 static
@@ -407,6 +466,13 @@ struct rpc_xprt *xprt_iter_next_entry_all(struct rpc_xprt_iter *xpi)
 			xprt_switch_find_next_entry_all);
 }
 
+static
+struct rpc_xprt *xprt_iter_next_entry_offline(struct rpc_xprt_iter *xpi)
+{
+	return xprt_iter_next_entry_multiple(xpi,
+			xprt_switch_find_next_entry_offline);
+}
+
 /*
  * xprt_iter_rewind - Resets the xprt iterator
  * @xpi: pointer to rpc_xprt_iter
@@ -460,6 +526,12 @@ void xprt_iter_init_listall(struct rpc_xprt_iter *xpi,
 	__xprt_iter_init(xpi, xps, &rpc_xprt_iter_listall);
 }
 
+void xprt_iter_init_listoffline(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt_switch *xps)
+{
+	__xprt_iter_init(xpi, xps, &rpc_xprt_iter_listoffline);
+}
+
 /**
  * xprt_iter_xchg_switch - Atomically swap out the rpc_xprt_switch
  * @xpi: pointer to rpc_xprt_iter
@@ -574,3 +646,10 @@ const struct rpc_xprt_iter_ops rpc_xprt_iter_listall = {
 	.xpi_xprt = xprt_iter_current_entry,
 	.xpi_next = xprt_iter_next_entry_all,
 };
+
+static
+const struct rpc_xprt_iter_ops rpc_xprt_iter_listoffline = {
+	.xpi_rewind = xprt_iter_default_rewind,
+	.xpi_xprt = xprt_iter_current_entry_offline,
+	.xpi_next = xprt_iter_next_entry_offline,
+};
-- 
cgit v1.2.3


From 9368fd6c75053630e95a6dbd17c9522e82101276 Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <kolga@netapp.com>
Date: Mon, 25 Jul 2022 09:32:25 -0400
Subject: SUNRPC enable back offline transports in trunking discovery

When we are adding a transport to a xprt_switch that's already on
the list but has been marked OFFLINE, then make the state ONLINE
since it's been tested now.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/clnt.h |  1 +
 net/sunrpc/clnt.c           | 14 ++++++++++++++
 2 files changed, 15 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index d14333f4947a..71a3a1dd7e81 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -242,6 +242,7 @@ void rpc_clnt_xprt_switch_put(struct rpc_clnt *);
 void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *, struct rpc_xprt *);
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 			const struct sockaddr *sap);
+void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt);
 void rpc_cleanup_clids(void);
 
 static inline int rpc_reply_expected(struct rpc_task *task)
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 68021b70340d..9dbce3b0d3a2 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -3095,8 +3095,22 @@ void rpc_clnt_xprt_switch_put(struct rpc_clnt *clnt)
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_put);
 
+void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
+{
+	struct rpc_xprt_switch *xps;
+
+	rcu_read_lock();
+	xps = rcu_dereference(clnt->cl_xpi.xpi_xpswitch);
+	rcu_read_unlock();
+	xprt_set_online_locked(xprt, xps);
+}
+
 void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
 {
+	if (rpc_clnt_xprt_switch_has_addr(clnt,
+		(const struct sockaddr *)&xprt->addr)) {
+		return rpc_clnt_xprt_set_online(clnt, xprt);
+	}
 	rcu_read_lock();
 	rpc_xprt_switch_add_xprt(rcu_dereference(clnt->cl_xpi.xpi_xpswitch),
 				 xprt);
-- 
cgit v1.2.3


From 497e6464d6adcee64f071b18fc826e63cfd2f0a5 Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:26 -0400
Subject: SUNRPC create an rpc function that allows xprt removal from rpc_clnt

Expose a function that allows a removal of xprt from the rpc_clnt.

When called from NFS that's running a trunked transport then don't
decrement the active transport counter.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/clnt.h          |  1 +
 include/linux/sunrpc/xprtmultipath.h |  2 +-
 net/sunrpc/clnt.c                    | 16 +++++++++++++++-
 net/sunrpc/xprt.c                    |  2 +-
 net/sunrpc/xprtmultipath.c           | 11 ++++++-----
 5 files changed, 24 insertions(+), 8 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 71a3a1dd7e81..7a43fd514398 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -240,6 +240,7 @@ const char *rpc_proc_name(const struct rpc_task *task);
 
 void rpc_clnt_xprt_switch_put(struct rpc_clnt *);
 void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *, struct rpc_xprt *);
+void rpc_clnt_xprt_switch_remove_xprt(struct rpc_clnt *, struct rpc_xprt *);
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 			const struct sockaddr *sap);
 void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt);
diff --git a/include/linux/sunrpc/xprtmultipath.h b/include/linux/sunrpc/xprtmultipath.h
index 688ca7eb1d01..9fff0768d942 100644
--- a/include/linux/sunrpc/xprtmultipath.h
+++ b/include/linux/sunrpc/xprtmultipath.h
@@ -55,7 +55,7 @@ extern void rpc_xprt_switch_set_roundrobin(struct rpc_xprt_switch *xps);
 extern void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
 		struct rpc_xprt *xprt);
 extern void rpc_xprt_switch_remove_xprt(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt);
+		struct rpc_xprt *xprt, bool offline);
 
 extern void xprt_iter_init(struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *xps);
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 9dbce3b0d3a2..26f3102500bb 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2144,7 +2144,8 @@ call_connect_status(struct rpc_task *task)
 				xprt_release(task);
 				value = atomic_long_dec_return(&xprt->queuelen);
 				if (value == 0)
-					rpc_xprt_switch_remove_xprt(xps, saved);
+					rpc_xprt_switch_remove_xprt(xps, saved,
+								    true);
 				xprt_put(saved);
 				task->tk_xprt = NULL;
 				task->tk_action = call_start;
@@ -3118,6 +3119,19 @@ void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_add_xprt);
 
+void rpc_clnt_xprt_switch_remove_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
+{
+	struct rpc_xprt_switch *xps;
+
+	rcu_read_lock();
+	xps = rcu_dereference(clnt->cl_xpi.xpi_xpswitch);
+	rpc_xprt_switch_remove_xprt(rcu_dereference(clnt->cl_xpi.xpi_xpswitch),
+				    xprt, 0);
+	xps->xps_nunique_destaddr_xprts--;
+	rcu_read_unlock();
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_remove_xprt);
+
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 				   const struct sockaddr *sap)
 {
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 8f8e3c952f24..44348c9f4b00 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -2182,5 +2182,5 @@ void xprt_delete_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
 
 	if (!xprt->sending.qlen && !xprt->pending.qlen &&
 	    !xprt->backlog.qlen && !atomic_long_read(&xprt->queuelen))
-		rpc_xprt_switch_remove_xprt(xps, xprt);
+		rpc_xprt_switch_remove_xprt(xps, xprt, true);
 }
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 8def8423fc0a..55da01730311 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -62,11 +62,11 @@ void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
 }
 
 static void xprt_switch_remove_xprt_locked(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt)
+		struct rpc_xprt *xprt, bool offline)
 {
 	if (unlikely(xprt == NULL))
 		return;
-	if (!test_bit(XPRT_OFFLINE, &xprt->state))
+	if (!test_bit(XPRT_OFFLINE, &xprt->state) && offline)
 		xps->xps_nactive--;
 	xps->xps_nxprts--;
 	if (xps->xps_nxprts == 0)
@@ -79,14 +79,15 @@ static void xprt_switch_remove_xprt_locked(struct rpc_xprt_switch *xps,
  * rpc_xprt_switch_remove_xprt - Removes an rpc_xprt from a rpc_xprt_switch
  * @xps: pointer to struct rpc_xprt_switch
  * @xprt: pointer to struct rpc_xprt
+ * @offline: indicates if the xprt that's being removed is in an offline state
  *
  * Removes xprt from the list of struct rpc_xprt in xps.
  */
 void rpc_xprt_switch_remove_xprt(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt)
+		struct rpc_xprt *xprt, bool offline)
 {
 	spin_lock(&xps->xps_lock);
-	xprt_switch_remove_xprt_locked(xps, xprt);
+	xprt_switch_remove_xprt_locked(xps, xprt, offline);
 	spin_unlock(&xps->xps_lock);
 	xprt_put(xprt);
 }
@@ -155,7 +156,7 @@ static void xprt_switch_free_entries(struct rpc_xprt_switch *xps)
 
 		xprt = list_first_entry(&xps->xps_xprt_list,
 				struct rpc_xprt, xprt_switch);
-		xprt_switch_remove_xprt_locked(xps, xprt);
+		xprt_switch_remove_xprt_locked(xps, xprt, true);
 		spin_unlock(&xps->xps_lock);
 		xprt_put(xprt);
 		spin_lock(&xps->xps_lock);
-- 
cgit v1.2.3


From 273d6aed9e5a1859dda15256f45561315c3d237a Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:29 -0400
Subject: SUNRPC export xprt_iter_rewind function

Make xprt_iter_rewind callable outside of xprtmultipath.c

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xprtmultipath.h | 2 ++
 net/sunrpc/xprtmultipath.c           | 1 -
 2 files changed, 2 insertions(+), 1 deletion(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xprtmultipath.h b/include/linux/sunrpc/xprtmultipath.h
index 9fff0768d942..c0514c684b2c 100644
--- a/include/linux/sunrpc/xprtmultipath.h
+++ b/include/linux/sunrpc/xprtmultipath.h
@@ -68,6 +68,8 @@ extern void xprt_iter_init_listoffline(struct rpc_xprt_iter *xpi,
 
 extern void xprt_iter_destroy(struct rpc_xprt_iter *xpi);
 
+extern void xprt_iter_rewind(struct rpc_xprt_iter *xpi);
+
 extern struct rpc_xprt_switch *xprt_iter_xchg_switch(
 		struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *newswitch);
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 55da01730311..685db598acbe 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -481,7 +481,6 @@ struct rpc_xprt *xprt_iter_next_entry_offline(struct rpc_xprt_iter *xpi)
  * Resets xpi to ensure that it points to the first entry in the list
  * of transports.
  */
-static
 void xprt_iter_rewind(struct rpc_xprt_iter *xpi)
 {
 	rcu_read_lock();
-- 
cgit v1.2.3


From 92cc04f60ab4ae199eee507e5cd4d5aa6c722e9c Mon Sep 17 00:00:00 2001
From: Olga Kornievskaia <olga.kornievskaia@gmail.com>
Date: Mon, 25 Jul 2022 09:32:30 -0400
Subject: SUNRPC create a function that probes only offline transports

For only offline transports, attempt to check connectivity via
a NULL call and, if that succeeds, call a provided session trunking
detection function.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/clnt.h |  2 ++
 net/sunrpc/clnt.c           | 65 +++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 67 insertions(+)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 7a43fd514398..75eea5ebb179 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -235,6 +235,8 @@ int		rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *,
 			struct rpc_xprt *,
 			void *);
 void		rpc_clnt_manage_trunked_xprts(struct rpc_clnt *);
+void		rpc_clnt_probe_trunked_xprts(struct rpc_clnt *,
+			struct rpc_add_xprt_test *);
 
 const char *rpc_proc_name(const struct rpc_task *task);
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 9c9712274ca8..bbfc47f03480 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -807,6 +807,13 @@ int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
 	return _rpc_clnt_xprt_iter_init(clnt, xpi, xprt_iter_init_listall);
 }
 
+static
+int rpc_clnt_xprt_iter_offline_init(struct rpc_clnt *clnt,
+				    struct rpc_xprt_iter *xpi)
+{
+	return _rpc_clnt_xprt_iter_init(clnt, xpi, xprt_iter_init_listoffline);
+}
+
 /**
  * rpc_clnt_iterate_for_each_xprt - Apply a function to all transports
  * @clnt: pointer to client
@@ -3018,6 +3025,64 @@ out_put_switch:
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_add_xprt);
 
+static int rpc_xprt_probe_trunked(struct rpc_clnt *clnt,
+				  struct rpc_xprt *xprt,
+				  struct rpc_add_xprt_test *data)
+{
+	struct rpc_xprt_switch *xps;
+	struct rpc_xprt *main_xprt;
+	int status = 0;
+
+	xprt_get(xprt);
+
+	rcu_read_lock();
+	main_xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
+	xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+	status = rpc_cmp_addr_port((struct sockaddr *)&xprt->addr,
+				   (struct sockaddr *)&main_xprt->addr);
+	rcu_read_unlock();
+	xprt_put(main_xprt);
+	if (status || !test_bit(XPRT_OFFLINE, &xprt->state))
+		goto out;
+
+	status = rpc_clnt_add_xprt_helper(clnt, xprt, data);
+out:
+	xprt_put(xprt);
+	xprt_switch_put(xps);
+	return status;
+}
+
+/* rpc_clnt_probe_trunked_xprt -- probe offlined transport for session trunking
+ * @clnt rpc_clnt structure
+ *
+ * For each offlined transport found in the rpc_clnt structure call
+ * the function rpc_xprt_probe_trunked() which will determine if this
+ * transport still belongs to the trunking group.
+ */
+void rpc_clnt_probe_trunked_xprts(struct rpc_clnt *clnt,
+				  struct rpc_add_xprt_test *data)
+{
+	struct rpc_xprt_iter xpi;
+	int ret;
+
+	ret = rpc_clnt_xprt_iter_offline_init(clnt, &xpi);
+	if (ret)
+		return;
+	for (;;) {
+		struct rpc_xprt *xprt = xprt_iter_get_next(&xpi);
+
+		if (!xprt)
+			break;
+		ret = rpc_xprt_probe_trunked(clnt, xprt, data);
+		xprt_put(xprt);
+		if (ret < 0)
+			break;
+		xprt_iter_rewind(&xpi);
+	}
+	xprt_iter_destroy(&xpi);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_probe_trunked_xprts);
+
 static int rpc_xprt_offline(struct rpc_clnt *clnt,
 			    struct rpc_xprt *xprt,
 			    void *data)
-- 
cgit v1.2.3


From 72691a269f0baad6d5f4aa7af97c29081b86d70f Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Wed, 27 Jul 2022 13:02:27 -0400
Subject: SUNRPC: Don't reuse bvec on retransmission of the request

If a request is re-encoded and then retransmitted, we need to make sure
that we also re-encode the bvec, in case the page lists have changed.

Fixes: ff053dbbaffe ("SUNRPC: Move the call to xprt_send_pagedata() out of xprt_sock_sendmsg()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 include/linux/sunrpc/xprt.h |  3 ++-
 net/sunrpc/clnt.c           |  1 -
 net/sunrpc/xprt.c           | 27 ++++++++++++++++++---------
 net/sunrpc/xprtsock.c       | 12 ++----------
 4 files changed, 22 insertions(+), 21 deletions(-)

(limited to 'include/linux')

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index 0d51b9f9ea37..b9f59aabee53 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -144,7 +144,8 @@ struct rpc_xprt_ops {
 	unsigned short	(*get_srcport)(struct rpc_xprt *xprt);
 	int		(*buf_alloc)(struct rpc_task *task);
 	void		(*buf_free)(struct rpc_task *task);
-	int		(*prepare_request)(struct rpc_rqst *req);
+	int		(*prepare_request)(struct rpc_rqst *req,
+					   struct xdr_buf *buf);
 	int		(*send_request)(struct rpc_rqst *req);
 	void		(*wait_for_reply_request)(struct rpc_task *task);
 	void		(*timer)(struct rpc_xprt *xprt, struct rpc_task *task);
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index bbfc47f03480..b098e707ad41 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1870,7 +1870,6 @@ rpc_xdr_encode(struct rpc_task *task)
 	req->rq_snd_buf.head[0].iov_len = 0;
 	xdr_init_encode(&xdr, &req->rq_snd_buf,
 			req->rq_snd_buf.head[0].iov_base, req);
-	xdr_free_bvec(&req->rq_snd_buf);
 	if (rpc_encode_header(task, &xdr))
 		return;
 
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 44348c9f4b00..d71eec494826 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -73,7 +73,7 @@ static void	xprt_init(struct rpc_xprt *xprt, struct net *net);
 static __be32	xprt_alloc_xid(struct rpc_xprt *xprt);
 static void	xprt_destroy(struct rpc_xprt *xprt);
 static void	xprt_request_init(struct rpc_task *task);
-static int	xprt_request_prepare(struct rpc_rqst *req);
+static int	xprt_request_prepare(struct rpc_rqst *req, struct xdr_buf *buf);
 
 static DEFINE_SPINLOCK(xprt_list_lock);
 static LIST_HEAD(xprt_list);
@@ -1149,7 +1149,7 @@ xprt_request_enqueue_receive(struct rpc_task *task)
 	if (!xprt_request_need_enqueue_receive(task, req))
 		return 0;
 
-	ret = xprt_request_prepare(task->tk_rqstp);
+	ret = xprt_request_prepare(task->tk_rqstp, &req->rq_rcv_buf);
 	if (ret)
 		return ret;
 	spin_lock(&xprt->queue_lock);
@@ -1179,8 +1179,11 @@ xprt_request_dequeue_receive_locked(struct rpc_task *task)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
 
-	if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
+	if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate)) {
 		xprt_request_rb_remove(req->rq_xprt, req);
+		xdr_free_bvec(&req->rq_rcv_buf);
+		req->rq_private_buf.bvec = NULL;
+	}
 }
 
 /**
@@ -1336,8 +1339,14 @@ xprt_request_enqueue_transmit(struct rpc_task *task)
 {
 	struct rpc_rqst *pos, *req = task->tk_rqstp;
 	struct rpc_xprt *xprt = req->rq_xprt;
+	int ret;
 
 	if (xprt_request_need_enqueue_transmit(task, req)) {
+		ret = xprt_request_prepare(task->tk_rqstp, &req->rq_snd_buf);
+		if (ret) {
+			task->tk_status = ret;
+			return;
+		}
 		req->rq_bytes_sent = 0;
 		spin_lock(&xprt->queue_lock);
 		/*
@@ -1397,6 +1406,7 @@ xprt_request_dequeue_transmit_locked(struct rpc_task *task)
 	} else
 		list_del(&req->rq_xmit2);
 	atomic_long_dec(&req->rq_xprt->xmit_queuelen);
+	xdr_free_bvec(&req->rq_snd_buf);
 }
 
 /**
@@ -1433,8 +1443,6 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 	    test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) ||
 	    xprt_is_pinned_rqst(req)) {
 		spin_lock(&xprt->queue_lock);
-		xprt_request_dequeue_transmit_locked(task);
-		xprt_request_dequeue_receive_locked(task);
 		while (xprt_is_pinned_rqst(req)) {
 			set_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
 			spin_unlock(&xprt->queue_lock);
@@ -1442,6 +1450,8 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 			spin_lock(&xprt->queue_lock);
 			clear_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
 		}
+		xprt_request_dequeue_transmit_locked(task);
+		xprt_request_dequeue_receive_locked(task);
 		spin_unlock(&xprt->queue_lock);
 	}
 }
@@ -1449,18 +1459,19 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 /**
  * xprt_request_prepare - prepare an encoded request for transport
  * @req: pointer to rpc_rqst
+ * @buf: pointer to send/rcv xdr_buf
  *
  * Calls into the transport layer to do whatever is needed to prepare
  * the request for transmission or receive.
  * Returns error, or zero.
  */
 static int
-xprt_request_prepare(struct rpc_rqst *req)
+xprt_request_prepare(struct rpc_rqst *req, struct xdr_buf *buf)
 {
 	struct rpc_xprt *xprt = req->rq_xprt;
 
 	if (xprt->ops->prepare_request)
-		return xprt->ops->prepare_request(req);
+		return xprt->ops->prepare_request(req, buf);
 	return 0;
 }
 
@@ -1961,8 +1972,6 @@ void xprt_release(struct rpc_task *task)
 	spin_unlock(&xprt->transport_lock);
 	if (req->rq_buffer)
 		xprt->ops->buf_free(task);
-	xdr_free_bvec(&req->rq_rcv_buf);
-	xdr_free_bvec(&req->rq_snd_buf);
 	if (req->rq_cred != NULL)
 		put_rpccred(req->rq_cred);
 	if (req->rq_release_snd_buf)
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index eba1be9984f8..e976007f4fd0 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -822,17 +822,9 @@ static int xs_stream_nospace(struct rpc_rqst *req, bool vm_wait)
 	return ret;
 }
 
-static int
-xs_stream_prepare_request(struct rpc_rqst *req)
+static int xs_stream_prepare_request(struct rpc_rqst *req, struct xdr_buf *buf)
 {
-	gfp_t gfp = rpc_task_gfp_mask();
-	int ret;
-
-	ret = xdr_alloc_bvec(&req->rq_snd_buf, gfp);
-	if (ret < 0)
-		return ret;
-	xdr_free_bvec(&req->rq_rcv_buf);
-	return xdr_alloc_bvec(&req->rq_rcv_buf, gfp);
+	return xdr_alloc_bvec(buf, rpc_task_gfp_mask());
 }
 
 /*
-- 
cgit v1.2.3


From 3c59366c207e4c6c6569524af606baf017a55c61 Mon Sep 17 00:00:00 2001
From: NeilBrown <neilb@suse.de>
Date: Mon, 1 Aug 2022 10:33:34 +1000
Subject: NFS: don't unhash dentry during unlink/rename

NFS unlink() (and rename over existing target) must determine if the
file is open, and must perform a "silly rename" instead of an unlink (or
before rename) if it is.  Otherwise the client might hold a file open
which has been removed on the server.

Consequently if it determines that the file isn't open, it must block
any subsequent opens until the unlink/rename has been completed on the
server.

This is currently achieved by unhashing the dentry.  This forces any
open attempt to the slow-path for lookup which will block on i_rwsem on
the directory until the unlink/rename completes.  A future patch will
change the VFS to only get a shared lock on i_rwsem for unlink, so this
will no longer work.

Instead we introduce an explicit interlock.  A special value is stored
in dentry->d_fsdata while the unlink/rename is running and
->d_revalidate blocks while that value is present.  When ->d_revalidate
unblocks, the dentry will be invalid.  This closes the race
without requiring exclusion on i_rwsem.

d_fsdata is already used in two different ways.
1/ an IS_ROOT directory dentry might have a "devname" stored in
   d_fsdata.  Such a dentry doesn't have a name and so cannot be the
   target of unlink or rename.  For safety we check if an old devname
   is still stored, and remove it if it is.
2/ a dentry with DCACHE_NFSFS_RENAMED set will have a 'struct
   nfs_unlinkdata' stored in d_fsdata.  While this is set maydelete()
   will fail, so an unlink or rename will never proceed on such
   a dentry.

Neither of these can be in effect when a dentry is the target of unlink
or rename.  So we can expect d_fsdata to be NULL, and store a special
value ((void*)1) which is given the name NFS_FSDATA_BLOCKED to indicate
that any lookup will be blocked.

The d_count() is incremented under d_lock() when a lookup finds the
dentry, so we check d_count() is low, and set NFS_FSDATA_BLOCKED under
the same lock to avoid any races.

Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/dir.c           | 72 +++++++++++++++++++++++++++++++++++++-------------
 include/linux/nfs_fs.h |  9 +++++++
 2 files changed, 63 insertions(+), 18 deletions(-)

(limited to 'include/linux')

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 7b2230297f6b..dbab3caa15ed 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1782,6 +1782,8 @@ __nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags,
 	int ret;
 
 	if (flags & LOOKUP_RCU) {
+		if (dentry->d_fsdata == NFS_FSDATA_BLOCKED)
+			return -ECHILD;
 		parent = READ_ONCE(dentry->d_parent);
 		dir = d_inode_rcu(parent);
 		if (!dir)
@@ -1790,6 +1792,9 @@ __nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags,
 		if (parent != READ_ONCE(dentry->d_parent))
 			return -ECHILD;
 	} else {
+		/* Wait for unlink to complete */
+		wait_var_event(&dentry->d_fsdata,
+			       dentry->d_fsdata != NFS_FSDATA_BLOCKED);
 		parent = dget_parent(dentry);
 		ret = reval(d_inode(parent), dentry, flags);
 		dput(parent);
@@ -2458,7 +2463,6 @@ out:
 int nfs_unlink(struct inode *dir, struct dentry *dentry)
 {
 	int error;
-	int need_rehash = 0;
 
 	dfprintk(VFS, "NFS: unlink(%s/%lu, %pd)\n", dir->i_sb->s_id,
 		dir->i_ino, dentry);
@@ -2473,15 +2477,25 @@ int nfs_unlink(struct inode *dir, struct dentry *dentry)
 		error = nfs_sillyrename(dir, dentry);
 		goto out;
 	}
-	if (!d_unhashed(dentry)) {
-		__d_drop(dentry);
-		need_rehash = 1;
-	}
+	/* We must prevent any concurrent open until the unlink
+	 * completes.  ->d_revalidate will wait for ->d_fsdata
+	 * to clear.  We set it here to ensure no lookup succeeds until
+	 * the unlink is complete on the server.
+	 */
+	error = -ETXTBSY;
+	if (WARN_ON(dentry->d_flags & DCACHE_NFSFS_RENAMED) ||
+	    WARN_ON(dentry->d_fsdata == NFS_FSDATA_BLOCKED))
+		goto out;
+	if (dentry->d_fsdata)
+		/* old devname */
+		kfree(dentry->d_fsdata);
+	dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+
 	spin_unlock(&dentry->d_lock);
 	error = nfs_safe_remove(dentry);
 	nfs_dentry_remove_handle_error(dir, dentry, error);
-	if (need_rehash)
-		d_rehash(dentry);
+	dentry->d_fsdata = NULL;
+	wake_up_var(&dentry->d_fsdata);
 out:
 	trace_nfs_unlink_exit(dir, dentry, error);
 	return error;
@@ -2588,6 +2602,15 @@ nfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
 }
 EXPORT_SYMBOL_GPL(nfs_link);
 
+static void
+nfs_unblock_rename(struct rpc_task *task, struct nfs_renamedata *data)
+{
+	struct dentry *new_dentry = data->new_dentry;
+
+	new_dentry->d_fsdata = NULL;
+	wake_up_var(&new_dentry->d_fsdata);
+}
+
 /*
  * RENAME
  * FIXME: Some nfsds, like the Linux user space nfsd, may generate a
@@ -2618,8 +2641,9 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 {
 	struct inode *old_inode = d_inode(old_dentry);
 	struct inode *new_inode = d_inode(new_dentry);
-	struct dentry *dentry = NULL, *rehash = NULL;
+	struct dentry *dentry = NULL;
 	struct rpc_task *task;
+	bool must_unblock = false;
 	int error = -EBUSY;
 
 	if (flags)
@@ -2637,18 +2661,27 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 	 * the new target.
 	 */
 	if (new_inode && !S_ISDIR(new_inode->i_mode)) {
-		/*
-		 * To prevent any new references to the target during the
-		 * rename, we unhash the dentry in advance.
+		/* We must prevent any concurrent open until the unlink
+		 * completes.  ->d_revalidate will wait for ->d_fsdata
+		 * to clear.  We set it here to ensure no lookup succeeds until
+		 * the unlink is complete on the server.
 		 */
-		if (!d_unhashed(new_dentry)) {
-			d_drop(new_dentry);
-			rehash = new_dentry;
+		error = -ETXTBSY;
+		if (WARN_ON(new_dentry->d_flags & DCACHE_NFSFS_RENAMED) ||
+		    WARN_ON(new_dentry->d_fsdata == NFS_FSDATA_BLOCKED))
+			goto out;
+		if (new_dentry->d_fsdata) {
+			/* old devname */
+			kfree(new_dentry->d_fsdata);
+			new_dentry->d_fsdata = NULL;
 		}
 
+		spin_lock(&new_dentry->d_lock);
 		if (d_count(new_dentry) > 2) {
 			int err;
 
+			spin_unlock(&new_dentry->d_lock);
+
 			/* copy the target dentry's name */
 			dentry = d_alloc(new_dentry->d_parent,
 					 &new_dentry->d_name);
@@ -2661,14 +2694,19 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 				goto out;
 
 			new_dentry = dentry;
-			rehash = NULL;
 			new_inode = NULL;
+		} else {
+			new_dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+			must_unblock = true;
+			spin_unlock(&new_dentry->d_lock);
 		}
+
 	}
 
 	if (S_ISREG(old_inode->i_mode))
 		nfs_sync_inode(old_inode);
-	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry, NULL);
+	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry,
+				must_unblock ? nfs_unblock_rename : NULL);
 	if (IS_ERR(task)) {
 		error = PTR_ERR(task);
 		goto out;
@@ -2692,8 +2730,6 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 		spin_unlock(&old_inode->i_lock);
 	}
 out:
-	if (rehash)
-		d_rehash(rehash);
 	trace_nfs_rename_exit(old_dir, old_dentry,
 			new_dir, new_dentry, error);
 	if (!error) {
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index a17c337dbdf1..b32ed68e7dc4 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -617,6 +617,15 @@ nfs_fileid_to_ino_t(u64 fileid)
 
 #define NFS_JUKEBOX_RETRY_TIME (5 * HZ)
 
+/* We need to block new opens while a file is being unlinked.
+ * If it is opened *before* we decide to unlink, we will silly-rename
+ * instead. If it is opened *after*, then we need to create or will fail.
+ * If we allow the two to race, we could end up with a file that is open
+ * but deleted on the server resulting in ESTALE.
+ * So use ->d_fsdata to record when the unlink is happening
+ * and block dentry revalidation while it is set.
+ */
+#define NFS_FSDATA_BLOCKED ((void*)1)
 
 # undef ifdebug
 # ifdef NFS_DEBUG
-- 
cgit v1.2.3


From af887e437bb298752b2edc5834048b8151b8aea0 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Tue, 9 Aug 2022 12:50:28 -0400
Subject: NFS: Improve write error tracing

Don't leak request pointers, but use the "device:inode" labelling that
is used by all the other trace points. Furthermore, replace use of page
indexes with an offset, again in order to align behaviour with other
NFS trace points.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 fs/nfs/nfstrace.h        | 36 +++++++++++++++++++++---------------
 fs/nfs/write.c           |  8 +++++---
 include/linux/nfs_page.h |  3 +--
 3 files changed, 27 insertions(+), 20 deletions(-)

(limited to 'include/linux')

diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h
index 8bd0c13a7c4b..731eecfdf49a 100644
--- a/fs/nfs/nfstrace.h
+++ b/fs/nfs/nfstrace.h
@@ -1447,44 +1447,50 @@ TRACE_EVENT(nfs_writeback_done,
 
 DECLARE_EVENT_CLASS(nfs_page_error_class,
 		TP_PROTO(
+			const struct inode *inode,
 			const struct nfs_page *req,
 			int error
 		),
 
-		TP_ARGS(req, error),
+		TP_ARGS(inode, req, error),
 
 		TP_STRUCT__entry(
-			__field(const void *, req)
-			__field(pgoff_t, index)
-			__field(unsigned int, offset)
-			__field(unsigned int, pgbase)
-			__field(unsigned int, bytes)
+			__field(dev_t, dev)
+			__field(u32, fhandle)
+			__field(u64, fileid)
+			__field(loff_t, offset)
+			__field(unsigned int, count)
 			__field(int, error)
 		),
 
 		TP_fast_assign(
-			__entry->req = req;
-			__entry->index = req->wb_index;
-			__entry->offset = req->wb_offset;
-			__entry->pgbase = req->wb_pgbase;
-			__entry->bytes = req->wb_bytes;
+			const struct nfs_inode *nfsi = NFS_I(inode);
+			__entry->dev = inode->i_sb->s_dev;
+			__entry->fileid = nfsi->fileid;
+			__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
+			__entry->offset = req_offset(req);
+			__entry->count = req->wb_bytes;
 			__entry->error = error;
 		),
 
 		TP_printk(
-			"req=%p index=%lu offset=%u pgbase=%u bytes=%u error=%d",
-			__entry->req, __entry->index, __entry->offset,
-			__entry->pgbase, __entry->bytes, __entry->error
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"offset=%lld count=%u", __entry->error,
+			MAJOR(__entry->dev), MINOR(__entry->dev),
+			(unsigned long long)__entry->fileid,
+			__entry->fhandle, __entry->offset,
+			__entry->count
 		)
 );
 
 #define DEFINE_NFS_PAGEERR_EVENT(name) \
 	DEFINE_EVENT(nfs_page_error_class, name, \
 			TP_PROTO( \
+				const struct inode *inode, \
 				const struct nfs_page *req, \
 				int error \
 			), \
-			TP_ARGS(req, error))
+			TP_ARGS(inode, req, error))
 
 DEFINE_NFS_PAGEERR_EVENT(nfs_write_error);
 DEFINE_NFS_PAGEERR_EVENT(nfs_comp_error);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 4adf2b488da1..4a3796811b4b 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -592,7 +592,8 @@ nfs_lock_and_join_requests(struct page *page)
 
 static void nfs_write_error(struct nfs_page *req, int error)
 {
-	trace_nfs_write_error(req, error);
+	trace_nfs_write_error(page_file_mapping(req->wb_page)->host, req,
+			      error);
 	nfs_mapping_set_error(req->wb_page, error);
 	nfs_inode_remove_request(req);
 	nfs_end_page_writeback(req);
@@ -1000,7 +1001,7 @@ static void nfs_write_completion(struct nfs_pgio_header *hdr)
 		nfs_list_remove_request(req);
 		if (test_bit(NFS_IOHDR_ERROR, &hdr->flags) &&
 		    (hdr->good_bytes < bytes)) {
-			trace_nfs_comp_error(req, hdr->error);
+			trace_nfs_comp_error(hdr->inode, req, hdr->error);
 			nfs_mapping_set_error(req->wb_page, hdr->error);
 			goto remove_req;
 		}
@@ -1882,7 +1883,8 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data)
 			(long long)req_offset(req));
 		if (status < 0) {
 			if (req->wb_page) {
-				trace_nfs_commit_error(req, status);
+				trace_nfs_commit_error(data->inode, req,
+						       status);
 				nfs_mapping_set_error(req->wb_page, status);
 				nfs_inode_remove_request(req);
 			}
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index f0373a6cb5fb..ba7e2e4b0926 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -202,8 +202,7 @@ nfs_list_entry(struct list_head *head)
 	return list_entry(head, struct nfs_page, wb_list);
 }
 
-static inline
-loff_t req_offset(struct nfs_page *req)
+static inline loff_t req_offset(const struct nfs_page *req)
 {
 	return (((loff_t)req->wb_index) << PAGE_SHIFT) + req->wb_offset;
 }
-- 
cgit v1.2.3