All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org,
	Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com>,
	Santosh Shilimkar <santosh.shilimkar@oracle.com>,
	Sasha Levin <alexander.levin@verizon.com>
Subject: [PATCH 3.18 02/35] RDS: RDMA: Fix the composite message user notification
Date: Fri,  6 Oct 2017 11:24:42 +0200	[thread overview]
Message-ID: <20171006092402.903168881@linuxfoundation.org> (raw)
In-Reply-To: <20171006092402.810400570@linuxfoundation.org>

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

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

From: Santosh Shilimkar <santosh.shilimkar@oracle.com>


[ Upstream commit 941f8d55f6d613a460a5e080d25a38509f45eb75 ]

When application sends an RDS RDMA composite message consist of
RDMA transfer to be followed up by non RDMA payload, it expect to
be notified *only* when the full message gets delivered. RDS RDMA
notification doesn't behave this way though.

Thanks to Venkat for debug and root casuing the issue
where only first part of the message(RDMA) was
successfully delivered but remainder payload delivery failed.
In that case, application should not be notified with
a false positive of message delivery success.

Fix this case by making sure the user gets notified only after
the full message delivery.

Reviewed-by: Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 net/rds/ib_send.c |   25 +++++++++++++++----------
 net/rds/rdma.c    |   10 ++++++++++
 net/rds/rds.h     |    1 +
 net/rds/send.c    |    4 +++-
 4 files changed, 29 insertions(+), 11 deletions(-)

--- a/net/rds/ib_send.c
+++ b/net/rds/ib_send.c
@@ -102,16 +102,6 @@ static void rds_ib_send_complete(struct
 	complete(rm, notify_status);
 }
 
-static void rds_ib_send_unmap_data(struct rds_ib_connection *ic,
-				   struct rm_data_op *op,
-				   int wc_status)
-{
-	if (op->op_nents)
-		ib_dma_unmap_sg(ic->i_cm_id->device,
-				op->op_sg, op->op_nents,
-				DMA_TO_DEVICE);
-}
-
 static void rds_ib_send_unmap_rdma(struct rds_ib_connection *ic,
 				   struct rm_rdma_op *op,
 				   int wc_status)
@@ -172,6 +162,21 @@ static void rds_ib_send_unmap_atomic(str
 		rds_ib_stats_inc(s_ib_atomic_fadd);
 }
 
+static void rds_ib_send_unmap_data(struct rds_ib_connection *ic,
+				   struct rm_data_op *op,
+				   int wc_status)
+{
+	struct rds_message *rm = container_of(op, struct rds_message, data);
+
+	if (op->op_nents)
+		ib_dma_unmap_sg(ic->i_cm_id->device,
+				op->op_sg, op->op_nents,
+				DMA_TO_DEVICE);
+
+	if (rm->rdma.op_active && rm->data.op_notify)
+		rds_ib_send_unmap_rdma(ic, &rm->rdma, wc_status);
+}
+
 /*
  * Unmap the resources associated with a struct send_work.
  *
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -625,6 +625,16 @@ int rds_cmsg_rdma_args(struct rds_sock *
 		}
 		op->op_notifier->n_user_token = args->user_token;
 		op->op_notifier->n_status = RDS_RDMA_SUCCESS;
+
+		/* Enable rmda notification on data operation for composite
+		 * rds messages and make sure notification is enabled only
+		 * for the data operation which follows it so that application
+		 * gets notified only after full message gets delivered.
+		 */
+		if (rm->data.op_sg) {
+			rm->rdma.op_notify = 0;
+			rm->data.op_notify = !!(args->flags & RDS_RDMA_NOTIFY_ME);
+		}
 	}
 
 	/* The cookie contains the R_Key of the remote memory region, and
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -360,6 +360,7 @@ struct rds_message {
 		} rdma;
 		struct rm_data_op {
 			unsigned int		op_active:1;
+			unsigned int		op_notify:1;
 			unsigned int		op_nents;
 			unsigned int		op_count;
 			struct scatterlist	*op_sg;
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -425,12 +425,14 @@ void rds_rdma_send_complete(struct rds_m
 	struct rm_rdma_op *ro;
 	struct rds_notifier *notifier;
 	unsigned long flags;
+	unsigned int notify = 0;
 
 	spin_lock_irqsave(&rm->m_rs_lock, flags);
 
+	notify =  rm->rdma.op_notify | rm->data.op_notify;
 	ro = &rm->rdma;
 	if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) &&
-	    ro->op_active && ro->op_notify && ro->op_notifier) {
+	    ro->op_active && notify && ro->op_notifier) {
 		notifier = ro->op_notifier;
 		rs = rm->m_rs;
 		sock_hold(rds_rs_to_sk(rs));

  parent reply	other threads:[~2017-10-06  9:32 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-06  9:24 [PATCH 3.18 00/35] 3.18.74-stable review Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 01/35] drm: bridge: add DT bindings for TI ths8135 Greg Kroah-Hartman
2017-10-06  9:24 ` Greg Kroah-Hartman [this message]
2017-10-06  9:24 ` [PATCH 3.18 03/35] MIPS: Ensure bss section ends on a long-aligned address Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 04/35] MIPS: kexec: Do not reserve invalid crashkernel memory on boot Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 06/35] hwmon: (gl520sm) Fix overflows and crash seen when writing into limit attributes Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 07/35] ARM: 8635/1: nommu: allow enabling REMAP_VECTORS_TO_RAM Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 08/35] tty: goldfish: Fix a parameter of a call to free_irq Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 09/35] IB/ipoib: Fix deadlock over vlan_mutex Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 10/35] IB/ipoib: rtnl_unlock can not come after free_netdev Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 11/35] IB/ipoib: Replace list_del of the neigh->list with list_del_init Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 12/35] USB: serial: mos7720: fix control-message error handling Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 13/35] USB: serial: mos7840: " Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 14/35] pinctrl: mvebu: Use seq_puts() in mvebu_pinconf_group_dbg_show() Greg Kroah-Hartman
2017-10-06  9:32   ` Joe Perches
2017-10-06  9:42     ` Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 15/35] partitions/efi: Fix integer overflow in GPT size calculation Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 16/35] audit: log 32-bit socketcalls Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 17/35] net: core: Prevent from dereferencing null pointer when releasing SKB Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 18/35] net/packet: check length in getsockopt() called with PACKET_HDRLEN Greg Kroah-Hartman
2017-10-06  9:24 ` [PATCH 3.18 19/35] team: fix memory leaks Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 21/35] mmc: sdio: fix alignment issue in struct sdio_func Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 22/35] netfilter: invoke synchronize_rcu after set the _hook_ to NULL Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 23/35] [media] exynos-gsc: Do not swap cb/cr for semi planar formats Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 24/35] netfilter: nfnl_cthelper: fix incorrect helper->expect_class_max Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 25/35] parisc: perf: Fix potential NULL pointer dereference Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 26/35] rds: ib: add error handle Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 27/35] md/raid10: submit bio directly to replacement disk Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 28/35] xfs: remove kmem_zalloc_greedy Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 29/35] libata: transport: Remove circular dependency at free time Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 30/35] IB/qib: fix false-postive maybe-uninitialized warning Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 31/35] ALSA: au88x0: avoid theoretical uninitialized access Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 32/35] [media] ttpci: address stringop overflow warning Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 33/35] staging: nvec: remove duplicated const Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 34/35] crypto: algif_skcipher - Load TX SG list after waiting Greg Kroah-Hartman
2017-10-06  9:25 ` [PATCH 3.18 35/35] mpi: Fix NULL ptr dereference in mpi_powm() [ver #3] Greg Kroah-Hartman
2017-10-06 14:03 ` [PATCH 3.18 00/35] 3.18.74-stable review Guenter Roeck
2017-10-07  9:38   ` Greg Kroah-Hartman
2017-10-06 17:30 ` Shuah Khan
2017-10-07  9:04   ` Greg Kroah-Hartman
2017-10-07  9:40 ` Greg Kroah-Hartman
2017-10-07 14:43   ` Guenter Roeck
2017-10-08  7:20     ` Greg Kroah-Hartman

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=20171006092402.903168881@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=alexander.levin@verizon.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=santosh.shilimkar@oracle.com \
    --cc=stable@vger.kernel.org \
    --cc=venkat.x.venkatsubra@oracle.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
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.