[PATCH 6.17 028/184] smb: client: queue post_recv_credits_work also if the peer raises the credit target

[PATCH 6.17 028/184] smb: client: queue post_recv_credits_work also if the peer raises the credit target

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stefan Metzmacher <metze@samba.org>

[ Upstream commit 02548c477a90481c1fd0d6e7c84b4504ec2fcc12 ]

This is already handled in the server, but currently it done
in a very complex way there. So we do it much simpler.

Note that put_receive_buffer() will take care of it
in case data_length is 0.

Cc: Steve French <smfrench@gmail.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Long Li <longli@microsoft.com>
Cc: linux-cifs@vger.kernel.org
Cc: samba-technical@lists.samba.org
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/smb/client/smbdirect.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c
index 6480945c24592..b3e04b410afe6 100644
--- a/fs/smb/client/smbdirect.c
+++ b/fs/smb/client/smbdirect.c
@@ -537,6 +537,7 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 	struct smbdirect_socket_parameters *sp = &sc->parameters;
 	struct smbd_connection *info =
 		container_of(sc, struct smbd_connection, socket);
+	int old_recv_credit_target;
 	u32 data_offset = 0;
 	u32 data_length = 0;
 	u32 remaining_data_length = 0;
@@ -599,6 +600,7 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 		}
 
 		atomic_dec(&info->receive_credits);
+		old_recv_credit_target = info->receive_credit_target;
 		info->receive_credit_target =
 			le16_to_cpu(data_transfer->credits_requested);
 		if (le16_to_cpu(data_transfer->credits_granted)) {
@@ -629,6 +631,9 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 		 * reassembly queue and wake up the reading thread
 		 */
 		if (data_length) {
+			if (info->receive_credit_target > old_recv_credit_target)
+				queue_work(info->workqueue, &info->post_send_credits_work);
+
 			enqueue_reassembly(info, response, data_length);
 			wake_up_interruptible(&sc->recv_io.reassembly.wait_queue);
 		} else
-- 
2.51.0

[PATCH 6.17 029/184] smb: client: limit the range of info->receive_credit_target

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stefan Metzmacher <metze@samba.org>

[ Upstream commit 9219f8cac296769324bbe8a28c289586114244c4 ]

This simplifies further changes...

Cc: Steve French <smfrench@gmail.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Long Li <longli@microsoft.com>
Cc: linux-cifs@vger.kernel.org
Cc: samba-technical@lists.samba.org
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/smb/client/smbdirect.c | 7 ++++++-
 fs/smb/client/smbdirect.h | 2 +-
 2 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c
index b3e04b410afe6..cbf1deff11065 100644
--- a/fs/smb/client/smbdirect.c
+++ b/fs/smb/client/smbdirect.c
@@ -429,6 +429,7 @@ static bool process_negotiation_response(
 		return false;
 	}
 	info->receive_credit_target = le16_to_cpu(packet->credits_requested);
+	info->receive_credit_target = min_t(u16, info->receive_credit_target, sp->recv_credit_max);
 
 	if (packet->credits_granted == 0) {
 		log_rdma_event(ERR, "error: credits_granted==0\n");
@@ -537,7 +538,7 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 	struct smbdirect_socket_parameters *sp = &sc->parameters;
 	struct smbd_connection *info =
 		container_of(sc, struct smbd_connection, socket);
-	int old_recv_credit_target;
+	u16 old_recv_credit_target;
 	u32 data_offset = 0;
 	u32 data_length = 0;
 	u32 remaining_data_length = 0;
@@ -603,6 +604,10 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 		old_recv_credit_target = info->receive_credit_target;
 		info->receive_credit_target =
 			le16_to_cpu(data_transfer->credits_requested);
+		info->receive_credit_target =
+			min_t(u16, info->receive_credit_target, sp->recv_credit_max);
+		info->receive_credit_target =
+			max_t(u16, info->receive_credit_target, 1);
 		if (le16_to_cpu(data_transfer->credits_granted)) {
 			atomic_add(le16_to_cpu(data_transfer->credits_granted),
 				&info->send_credits);
diff --git a/fs/smb/client/smbdirect.h b/fs/smb/client/smbdirect.h
index 4ca9b2b2c57f9..ed362267dd11d 100644
--- a/fs/smb/client/smbdirect.h
+++ b/fs/smb/client/smbdirect.h
@@ -63,7 +63,7 @@ struct smbd_connection {
 	int protocol;
 	atomic_t send_credits;
 	atomic_t receive_credits;
-	int receive_credit_target;
+	u16 receive_credit_target;
 
 	/* Memory registrations */
 	/* Maximum number of RDMA read/write outstanding on this connection */
-- 
2.51.0

[PATCH 6.17 030/184] smb: client: make use of ib_wc_status_msg() and skip IB_WC_WR_FLUSH_ERR logging

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stefan Metzmacher <metze@samba.org>

[ Upstream commit a8e970358b31a5abba8b5737a67ba7b8d26f4258 ]

There's no need to get log message for every IB_WC_WR_FLUSH_ERR
completion, but any other error should be logged at level ERR.

Cc: Steve French <smfrench@gmail.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Long Li <longli@microsoft.com>
Cc: linux-cifs@vger.kernel.org
Cc: samba-technical@lists.samba.org
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/smb/client/smbdirect.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c
index cbf1deff11065..99fad70356c57 100644
--- a/fs/smb/client/smbdirect.c
+++ b/fs/smb/client/smbdirect.c
@@ -362,8 +362,8 @@ static void send_done(struct ib_cq *cq, struct ib_wc *wc)
 	struct smbd_connection *info =
 		container_of(sc, struct smbd_connection, socket);
 
-	log_rdma_send(INFO, "smbdirect_send_io 0x%p completed wc->status=%d\n",
-		request, wc->status);
+	log_rdma_send(INFO, "smbdirect_send_io 0x%p completed wc->status=%s\n",
+		request, ib_wc_status_msg(wc->status));
 
 	for (i = 0; i < request->num_sge; i++)
 		ib_dma_unmap_single(sc->ib.dev,
@@ -372,8 +372,9 @@ static void send_done(struct ib_cq *cq, struct ib_wc *wc)
 			DMA_TO_DEVICE);
 
 	if (wc->status != IB_WC_SUCCESS || wc->opcode != IB_WC_SEND) {
-		log_rdma_send(ERR, "wc->status=%d wc->opcode=%d\n",
-			wc->status, wc->opcode);
+		if (wc->status != IB_WC_WR_FLUSH_ERR)
+			log_rdma_send(ERR, "wc->status=%s wc->opcode=%d\n",
+				ib_wc_status_msg(wc->status), wc->opcode);
 		mempool_free(request, sc->send_io.mem.pool);
 		smbd_disconnect_rdma_connection(info);
 		return;
@@ -543,13 +544,16 @@ static void recv_done(struct ib_cq *cq, struct ib_wc *wc)
 	u32 data_length = 0;
 	u32 remaining_data_length = 0;
 
-	log_rdma_recv(INFO, "response=0x%p type=%d wc status=%d wc opcode %d byte_len=%d pkey_index=%u\n",
-		      response, sc->recv_io.expected, wc->status, wc->opcode,
+	log_rdma_recv(INFO,
+		      "response=0x%p type=%d wc status=%s wc opcode %d byte_len=%d pkey_index=%u\n",
+		      response, sc->recv_io.expected,
+		      ib_wc_status_msg(wc->status), wc->opcode,
 		      wc->byte_len, wc->pkey_index);
 
 	if (wc->status != IB_WC_SUCCESS || wc->opcode != IB_WC_RECV) {
-		log_rdma_recv(INFO, "wc->status=%d opcode=%d\n",
-			wc->status, wc->opcode);
+		if (wc->status != IB_WC_WR_FLUSH_ERR)
+			log_rdma_recv(ERR, "wc->status=%s opcode=%d\n",
+				ib_wc_status_msg(wc->status), wc->opcode);
 		goto error;
 	}
 
-- 
2.51.0

[PATCH 6.17 031/184] smb: server: let smb_direct_flush_send_list() invalidate a remote key first

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stefan Metzmacher <metze@samba.org>

[ Upstream commit 1b53426334c3c942db47e0959a2527a4f815af50 ]

If we want to invalidate a remote key we should do that as soon as
possible, so do it in the first send work request.

Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Cc: Steve French <smfrench@gmail.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: linux-cifs@vger.kernel.org
Cc: samba-technical@lists.samba.org
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/smb/server/transport_rdma.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/smb/server/transport_rdma.c b/fs/smb/server/transport_rdma.c
index e1f659d3b4cf5..2363244ff5f75 100644
--- a/fs/smb/server/transport_rdma.c
+++ b/fs/smb/server/transport_rdma.c
@@ -939,12 +939,15 @@ static int smb_direct_flush_send_list(struct smb_direct_transport *t,
 			       struct smb_direct_sendmsg,
 			       list);
 
+	if (send_ctx->need_invalidate_rkey) {
+		first->wr.opcode = IB_WR_SEND_WITH_INV;
+		first->wr.ex.invalidate_rkey = send_ctx->remote_key;
+		send_ctx->need_invalidate_rkey = false;
+		send_ctx->remote_key = 0;
+	}
+
 	last->wr.send_flags = IB_SEND_SIGNALED;
 	last->wr.wr_cqe = &last->cqe;
-	if (is_last && send_ctx->need_invalidate_rkey) {
-		last->wr.opcode = IB_WR_SEND_WITH_INV;
-		last->wr.ex.invalidate_rkey = send_ctx->remote_key;
-	}
 
 	ret = smb_direct_post_send(t, &first->wr);
 	if (!ret) {
-- 
2.51.0

[PATCH 6.17 084/184] smb: client: get rid of d_drop() in cifs_do_rename()

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Paulo Alcantara <pc@manguebit.org>

commit 72ed55b4c335703c203b942972558173e1e5ddee upstream.

There is no need to force a lookup by unhashing the moved dentry after
successfully renaming the file on server.  The file metadata will be
re-fetched from server, if necessary, in the next call to
->d_revalidate() anyways.

Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: stable@vger.kernel.org
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 fs/smb/client/inode.c |    5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

--- a/fs/smb/client/inode.c
+++ b/fs/smb/client/inode.c
@@ -2484,11 +2484,8 @@ cifs_do_rename(const unsigned int xid, s
 	}
 #endif /* CONFIG_CIFS_ALLOW_INSECURE_LEGACY */
 do_rename_exit:
-	if (rc == 0) {
+	if (rc == 0)
 		d_move(from_dentry, to_dentry);
-		/* Force a new lookup */
-		d_drop(from_dentry);
-	}
 	cifs_put_tlink(tlink);
 	return rc;
 }

[PATCH 6.17 088/184] cifs: Fix TCP_Server_Info::credits to be signed

[Greg Kroah-Hartman]

6.17-stable review patch.  If anyone has any objections, please let me know.

------------------

From: David Howells <dhowells@redhat.com>

commit 5b2ff4873aeab972f919d5aea11c51393322bf58 upstream.

Fix TCP_Server_Info::credits to be signed, just as echo_credits and
oplock_credits are.  This also fixes what ought to get at least a
compilation warning if not an outright error in *get_credits_field() as a
pointer to the unsigned server->credits field is passed back as a pointer
to a signed int.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-cifs@vger.kernel.org
Cc: stable@vger.kernel.org
Acked-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Acked-by: Pavel Shilovskiy <pshilovskiy@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 fs/smb/client/cifsglob.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/fs/smb/client/cifsglob.h
+++ b/fs/smb/client/cifsglob.h
@@ -740,7 +740,7 @@ struct TCP_Server_Info {
 	bool nosharesock;
 	bool tcp_nodelay;
 	bool terminate;
-	unsigned int credits;  /* send no more requests at once */
+	int credits;  /* send no more requests at once */
 	unsigned int max_credits; /* can override large 32000 default at mnt */
 	unsigned int in_flight;  /* number of requests on the wire to server */
 	unsigned int max_in_flight; /* max number of requests that were on wire */