[PATCH for-3.10 0/4] iSER initiator updates for 3.10

[PATCH for-3.10 0/4] iSER initiator updates for 3.10

[Or Gerlitz]

Hi Roland,

Here are the iser initiator updates for the 3.10 merge window, the 
"Return error to upper layers on EAGAIN registration failures" patch 
is marked for -stable fixing a bug that went it in 3.9

Or.

Or Gerlitz (2):
  IB/iser: Return error to upper layers on EAGAIN registration failures
  IB/iser: Add support for iser CM REQ additional info

Roi Dayan (2):
  IB/iser: Add module version
  IB/iser: Move informational messages from error to info level

 drivers/infiniband/ulp/iser/iscsi_iser.c  |   24 +++++++++---------
 drivers/infiniband/ulp/iser/iscsi_iser.h  |   24 ++++++++++++++++---
 drivers/infiniband/ulp/iser/iser_memory.c |    3 +-
 drivers/infiniband/ulp/iser/iser_verbs.c  |   36 +++++++++++++++++-----------
 4 files changed, 56 insertions(+), 31 deletions(-)

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-3.10 1/4] IB/iser: Add module version

[Or Gerlitz]

From: Roi Dayan <roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Add displaying module version, update the version to 1.1,
and remove the DRV_DATE define.

Signed-off-by: Roi Dayan <roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Or Gerlitz <ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/ulp/iser/iscsi_iser.c |    6 +++---
 drivers/infiniband/ulp/iser/iscsi_iser.h |    3 +--
 2 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index 0ab8c9c..69e0d7c 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -82,10 +82,10 @@ module_param_named(max_lun, iscsi_max_lun, uint, S_IRUGO);
 
 int iser_debug_level = 0;
 
-MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover "
-		   "v" DRV_VER " (" DRV_DATE ")");
+MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover");
 MODULE_LICENSE("Dual BSD/GPL");
 MODULE_AUTHOR("Alex Nezhinsky, Dan Bar Dov, Or Gerlitz");
+MODULE_VERSION(DRV_VER);
 
 module_param_named(debug_level, iser_debug_level, int, 0644);
 MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0 (default:disabled)");
@@ -682,7 +682,7 @@ static umode_t iser_attr_is_visible(int param_type, int param)
 
 static struct scsi_host_template iscsi_iser_sht = {
 	.module                 = THIS_MODULE,
-	.name                   = "iSCSI Initiator over iSER, v." DRV_VER,
+	.name                   = "iSCSI Initiator over iSER",
 	.queuecommand           = iscsi_queuecommand,
 	.change_queue_depth	= iscsi_change_queue_depth,
 	.sg_tablesize           = ISCSI_ISER_SG_TABLESIZE,
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 5babdb3..0da04bf 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -65,8 +65,7 @@
 
 #define DRV_NAME	"iser"
 #define PFX		DRV_NAME ": "
-#define DRV_VER		"0.1"
-#define DRV_DATE	"May 7th, 2006"
+#define DRV_VER		"1.1"
 
 #define iser_dbg(fmt, arg...)				\
 	do {						\
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-3.10 2/4] IB/iser: Move informational messages from error to info level

[Or Gerlitz]

From: Roi Dayan <roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Introduce iser_info and move informational messages that were treated
as errors with respect to printing them to use that macro. Also,
cleanup printk leftovers to use the existing macros.

Signed-off-by: Roi Dayan <roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Or Gerlitz <ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/ulp/iser/iscsi_iser.c |   18 +++++++++---------
 drivers/infiniband/ulp/iser/iscsi_iser.h |   12 ++++++++++--
 drivers/infiniband/ulp/iser/iser_verbs.c |   29 +++++++++++++++--------------
 3 files changed, 34 insertions(+), 25 deletions(-)

diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index 69e0d7c..f19b099 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -370,8 +370,8 @@ iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
 	/* binds the iSER connection retrieved from the previously
 	 * connected ep_handle to the iSCSI layer connection. exchanges
 	 * connection pointers */
-	iser_err("binding iscsi/iser conn %p %p to ib_conn %p\n",
-					conn, conn->dd_data, ib_conn);
+	iser_info("binding iscsi/iser conn %p %p to ib_conn %p\n",
+		  conn, conn->dd_data, ib_conn);
 	iser_conn = conn->dd_data;
 	ib_conn->iser_conn = iser_conn;
 	iser_conn->ib_conn  = ib_conn;
@@ -475,28 +475,28 @@ iscsi_iser_set_param(struct iscsi_cls_conn *cls_conn,
 	case ISCSI_PARAM_HDRDGST_EN:
 		sscanf(buf, "%d", &value);
 		if (value) {
-			printk(KERN_ERR "DataDigest wasn't negotiated to None");
+			iser_err("DataDigest wasn't negotiated to None");
 			return -EPROTO;
 		}
 		break;
 	case ISCSI_PARAM_DATADGST_EN:
 		sscanf(buf, "%d", &value);
 		if (value) {
-			printk(KERN_ERR "DataDigest wasn't negotiated to None");
+			iser_err("DataDigest wasn't negotiated to None");
 			return -EPROTO;
 		}
 		break;
 	case ISCSI_PARAM_IFMARKER_EN:
 		sscanf(buf, "%d", &value);
 		if (value) {
-			printk(KERN_ERR "IFMarker wasn't negotiated to No");
+			iser_err("IFMarker wasn't negotiated to No");
 			return -EPROTO;
 		}
 		break;
 	case ISCSI_PARAM_OFMARKER_EN:
 		sscanf(buf, "%d", &value);
 		if (value) {
-			printk(KERN_ERR "OFMarker wasn't negotiated to No");
+			iser_err("OFMarker wasn't negotiated to No");
 			return -EPROTO;
 		}
 		break;
@@ -596,7 +596,7 @@ iscsi_iser_ep_poll(struct iscsi_endpoint *ep, int timeout_ms)
 	     ib_conn->state == ISER_CONN_DOWN))
 		rc = -1;
 
-	iser_err("ib conn %p rc = %d\n", ib_conn, rc);
+	iser_info("ib conn %p rc = %d\n", ib_conn, rc);
 
 	if (rc > 0)
 		return 1; /* success, this is the equivalent of POLLOUT */
@@ -623,7 +623,7 @@ iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
 		iscsi_suspend_tx(ib_conn->iser_conn->iscsi_conn);
 
 
-	iser_err("ib conn %p state %d\n",ib_conn, ib_conn->state);
+	iser_info("ib conn %p state %d\n", ib_conn, ib_conn->state);
 	iser_conn_terminate(ib_conn);
 }
 
@@ -740,7 +740,7 @@ static int __init iser_init(void)
 	iser_dbg("Starting iSER datamover...\n");
 
 	if (iscsi_max_lun < 1) {
-		printk(KERN_ERR "Invalid max_lun value of %u\n", iscsi_max_lun);
+		iser_err("Invalid max_lun value of %u\n", iscsi_max_lun);
 		return -EINVAL;
 	}
 
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 0da04bf..f6b8ff0 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -42,6 +42,7 @@
 
 #include <linux/types.h>
 #include <linux/net.h>
+#include <linux/printk.h>
 #include <scsi/libiscsi.h>
 #include <scsi/scsi_transport_iscsi.h>
 
@@ -69,15 +70,22 @@
 
 #define iser_dbg(fmt, arg...)				\
 	do {						\
-		if (iser_debug_level > 1)		\
+		if (iser_debug_level > 2)		\
 			printk(KERN_DEBUG PFX "%s:" fmt,\
 				__func__ , ## arg);	\
 	} while (0)
 
 #define iser_warn(fmt, arg...)				\
 	do {						\
+		if (iser_debug_level > 1)		\
+			printk(KERN_WARNING PFX "%s:" fmt,\
+				__func__ , ## arg);	\
+	} while (0)
+
+#define iser_info(fmt, arg...)				\
+	do {						\
 		if (iser_debug_level > 0)		\
-			printk(KERN_DEBUG PFX "%s:" fmt,\
+			pr_info(PFX "%s:" fmt,          \
 				__func__ , ## arg);	\
 	} while (0)
 
diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index 4debadc..13838ad 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -74,8 +74,9 @@ static int iser_create_device_ib_res(struct iser_device *device)
 	struct iser_cq_desc *cq_desc;
 
 	device->cqs_used = min(ISER_MAX_CQ, device->ib_device->num_comp_vectors);
-	iser_err("using %d CQs, device %s supports %d vectors\n", device->cqs_used,
-		 device->ib_device->name, device->ib_device->num_comp_vectors);
+	iser_info("using %d CQs, device %s supports %d vectors\n",
+		  device->cqs_used, device->ib_device->name,
+		  device->ib_device->num_comp_vectors);
 
 	device->cq_desc = kmalloc(sizeof(struct iser_cq_desc) * device->cqs_used,
 				  GFP_KERNEL);
@@ -262,7 +263,7 @@ static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
 			min_index = index;
 	device->cq_active_qps[min_index]++;
 	mutex_unlock(&ig.connlist_mutex);
-	iser_err("cq index %d used for ib_conn %p\n", min_index, ib_conn);
+	iser_info("cq index %d used for ib_conn %p\n", min_index, ib_conn);
 
 	init_attr.event_handler = iser_qp_event_callback;
 	init_attr.qp_context	= (void *)ib_conn;
@@ -280,9 +281,9 @@ static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
 		goto out_err;
 
 	ib_conn->qp = ib_conn->cma_id->qp;
-	iser_err("setting conn %p cma_id %p: fmr_pool %p qp %p\n",
-		 ib_conn, ib_conn->cma_id,
-		 ib_conn->fmr_pool, ib_conn->cma_id->qp);
+	iser_info("setting conn %p cma_id %p: fmr_pool %p qp %p\n",
+		  ib_conn, ib_conn->cma_id,
+		  ib_conn->fmr_pool, ib_conn->cma_id->qp);
 	return ret;
 
 out_err:
@@ -299,9 +300,9 @@ static int iser_free_ib_conn_res(struct iser_conn *ib_conn, int can_destroy_id)
 	int cq_index;
 	BUG_ON(ib_conn == NULL);
 
-	iser_err("freeing conn %p cma_id %p fmr pool %p qp %p\n",
-		 ib_conn, ib_conn->cma_id,
-		 ib_conn->fmr_pool, ib_conn->qp);
+	iser_info("freeing conn %p cma_id %p fmr pool %p qp %p\n",
+		  ib_conn, ib_conn->cma_id,
+		  ib_conn->fmr_pool, ib_conn->qp);
 
 	/* qp is created only once both addr & route are resolved */
 	if (ib_conn->fmr_pool != NULL)
@@ -379,7 +380,7 @@ static void iser_device_try_release(struct iser_device *device)
 {
 	mutex_lock(&ig.device_list_mutex);
 	device->refcount--;
-	iser_err("device %p refcount %d\n",device,device->refcount);
+	iser_info("device %p refcount %d\n", device, device->refcount);
 	if (!device->refcount) {
 		iser_free_device_ib_res(device);
 		list_del(&device->ig_list);
@@ -558,8 +559,8 @@ static int iser_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *eve
 {
 	int ret = 0;
 
-	iser_err("event %d status %d conn %p id %p\n",
-		event->event, event->status, cma_id->context, cma_id);
+	iser_info("event %d status %d conn %p id %p\n",
+		  event->event, event->status, cma_id->context, cma_id);
 
 	switch (event->event) {
 	case RDMA_CM_EVENT_ADDR_RESOLVED:
@@ -619,8 +620,8 @@ int iser_connect(struct iser_conn   *ib_conn,
 	/* the device is known only --after-- address resolution */
 	ib_conn->device = NULL;
 
-	iser_err("connecting to: %pI4, port 0x%x\n",
-		 &dst_addr->sin_addr, dst_addr->sin_port);
+	iser_info("connecting to: %pI4, port 0x%x\n",
+		  &dst_addr->sin_addr, dst_addr->sin_port);
 
 	ib_conn->state = ISER_CONN_PENDING;
 
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-3.10 3/4] IB/iser: Return error to upper layers on EAGAIN registration failures

[Or Gerlitz]

Commit 819a087 "IB/iser: Avoid error prints on EAGAIN registration
failures" not only eliminated the error print on that case, but
rather also modified the code such that it doesn't return any error
to upper layers. As a result a wrong mapping was used. Fix this
such that we do return error on that case.

Signed-off-by: Or Gerlitz <ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/ulp/iser/iser_memory.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/infiniband/ulp/iser/iser_memory.c b/drivers/infiniband/ulp/iser/iser_memory.c
index be1edb0..68ebb7f 100644
--- a/drivers/infiniband/ulp/iser/iser_memory.c
+++ b/drivers/infiniband/ulp/iser/iser_memory.c
@@ -416,8 +416,9 @@ int iser_reg_rdma_mem(struct iscsi_iser_task *iser_task,
 			for (i=0 ; i<ib_conn->page_vec->length ; i++)
 				iser_err("page_vec[%d] = 0x%llx\n", i,
 					 (unsigned long long) ib_conn->page_vec->pages[i]);
-			return err;
 		}
+		if (err)
+			return err;
 	}
 	return 0;
 }
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-3.10 4/4] IB/iser: Add support for iser CM REQ additional info

[Or Gerlitz]

Annex A12 of the IBTA spec defines additional information that needs
to be provided through the CM exchange and relate to usage of ZBVA
(Zero Based VAs) and Send With Invalidate over this iser connection.

Currently, the initiator set both to not supported, but does provide
the header such that existing iser targets can be patched too, so
they start looking on the private data carried by the CM.

This is preparation step to enable iser with HW drivers for which
FMRs are not supported, such as mlx4 VF instances or new HW devices
which might support only FRWR (Fast Registration Work-Requests) along
the details of the IB_DEVICE_MEM_MGT_EXTENSIONS device capability.

Signed-off-by: Or Gerlitz <ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/ulp/iser/iscsi_iser.h |    9 +++++++++
 drivers/infiniband/ulp/iser/iser_verbs.c |    7 +++++++
 2 files changed, 16 insertions(+), 0 deletions(-)

diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index f6b8ff0..cae6084 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -140,6 +140,15 @@ struct iser_hdr {
 	__be64  read_va;
 } __attribute__((packed));
 
+
+#define ISER_ZBVA_NOT_SUPPORTED		0x80
+#define ISER_SEND_W_INV_NOT_SUPPORTED	0x40
+
+struct iser_cm_hdr {
+	u8      flags;
+	u8      rsvd[3];
+} __packed;
+
 /* Constant PDU lengths calculations */
 #define ISER_HEADERS_LEN  (sizeof(struct iser_hdr) + sizeof(struct iscsi_hdr))
 
diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index 13838ad..5278916 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -499,6 +499,7 @@ static int iser_route_handler(struct rdma_cm_id *cma_id)
 {
 	struct rdma_conn_param conn_param;
 	int    ret;
+	struct iser_cm_hdr req_hdr;
 
 	ret = iser_create_ib_conn_res((struct iser_conn *)cma_id->context);
 	if (ret)
@@ -510,6 +511,12 @@ static int iser_route_handler(struct rdma_cm_id *cma_id)
 	conn_param.retry_count	       = 7;
 	conn_param.rnr_retry_count     = 6;
 
+	memset(&req_hdr, 0, sizeof(req_hdr));
+	req_hdr.flags = (ISER_ZBVA_NOT_SUPPORTED |
+			ISER_SEND_W_INV_NOT_SUPPORTED);
+	conn_param.private_data		= (void *)&req_hdr;
+	conn_param.private_data_len	= sizeof(struct iser_cm_hdr);
+
 	ret = rdma_connect(cma_id, &conn_param);
 	if (ret) {
 		iser_err("failure connecting: %d\n", ret);
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH for-3.10 0/4] iSER initiator updates for 3.10

[Roland Dreier]

Thanks, applied all 4.

I just moved and the box with some of my private keys etc. still needs
to be set up, so it will be a day or two until I get this pushed out
to kernel.org.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH for-3.10 0/4] iSER initiator updates for 3.10

[Or Gerlitz]

On Thu, May 2, 2013 at 3:37 AM, Roland Dreier <roland-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
> Thanks, applied all 4.

Good. Anything else you need for the flow steering kernel patches so
they can be accepted? Steve and Christoph acknowledged them and there
wasn't any objection.

Or.


> I just moved and the box with some of my private keys etc. still needs
> to be set up, so it will be a day or two until I get this pushed out
> to kernel.org.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html