[PATCH 0/5 v2] qedr: batch of fixes for 4.10

[PATCH 0/5 v2] qedr: batch of fixes for 4.10

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

This series contains a batch of fixes for 4.10.

Based on origin/k.o/for-4.10, over
"[PATCH 0/5 v2] qedr: QP and CQ fixes for 4.10",
on top of 5cc8fab.

Changes from v1:
 - exactly the same, except that it is built on top of
   "[PATCH 0/5 v2] qedr: QP and CQ fixes for 4.10"

Ram Amrani (5):
  qedr: mark three functions as static
  qedr: fix formatting
  qedr: fix RDMA CM loopback
  qedr: fix memory leak in PD allocation and simplify it
  qedr: dispatch port active event from qedr_add

 drivers/infiniband/hw/qedr/main.c    | 23 +++++++++++++-------
 drivers/infiniband/hw/qedr/qedr.h    |  5 +++++
 drivers/infiniband/hw/qedr/qedr_cm.c |  4 ++--
 drivers/infiniband/hw/qedr/verbs.c   | 41 +++++++++++++++++++++++-------------
 4 files changed, 48 insertions(+), 25 deletions(-)

-- 
1.8.3.1

--
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 1/5 v2] qedr: mark three functions as static

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

mark qedr_get_state_from_ibqp(), __qedr_alloc_mr() and __qedr_post_send()
as static since they are only used in the same file.

Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
 drivers/infiniband/hw/qedr/verbs.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 189f9bd..2d55597 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -1600,7 +1600,7 @@ struct ib_qp *qedr_create_qp(struct ib_pd *ibpd,
 	return ERR_PTR(-EFAULT);
 }
 
-enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state)
+static enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state)
 {
 	switch (qp_state) {
 	case QED_ROCE_QP_STATE_RESET:
@@ -1621,7 +1621,8 @@ enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state)
 	return IB_QPS_ERR;
 }
 
-enum qed_roce_qp_state qedr_get_state_from_ibqp(enum ib_qp_state qp_state)
+static enum qed_roce_qp_state qedr_get_state_from_ibqp(
+					enum ib_qp_state qp_state)
 {
 	switch (qp_state) {
 	case IB_QPS_RESET:
@@ -2310,7 +2311,8 @@ int qedr_dereg_mr(struct ib_mr *ib_mr)
 	return rc;
 }
 
-struct qedr_mr *__qedr_alloc_mr(struct ib_pd *ibpd, int max_page_list_len)
+static struct qedr_mr *__qedr_alloc_mr(struct ib_pd *ibpd,
+				       int max_page_list_len)
 {
 	struct qedr_pd *pd = get_qedr_pd(ibpd);
 	struct qedr_dev *dev = get_qedr_dev(ibpd->device);
@@ -2712,7 +2714,7 @@ static int qedr_prepare_reg(struct qedr_qp *qp,
 	return 0;
 }
 
-enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode)
+static enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode)
 {
 	switch (opcode) {
 	case IB_WR_RDMA_WRITE:
@@ -2737,7 +2739,7 @@ enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode)
 	}
 }
 
-inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr)
+static inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr)
 {
 	int wq_is_full, err_wr, pbl_is_full;
 	struct qedr_dev *dev = qp->dev;
@@ -2774,7 +2776,7 @@ inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr)
 	return true;
 }
 
-int __qedr_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
+static int __qedr_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		     struct ib_send_wr **bad_wr)
 {
 	struct qedr_dev *dev = get_qedr_dev(ibqp->device);
-- 
1.8.3.1

--
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 2/5 v2] qedr: fix formatting

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

remove standalone ';'.
list function's parameters in a single line.

Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
 drivers/infiniband/hw/qedr/main.c  | 3 +--
 drivers/infiniband/hw/qedr/verbs.c | 1 -
 2 files changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c
index 7b74d09..eac0bfc 100644
--- a/drivers/infiniband/hw/qedr/main.c
+++ b/drivers/infiniband/hw/qedr/main.c
@@ -576,8 +576,7 @@ static int qedr_set_device_attr(struct qedr_dev *dev)
 	return 0;
 }
 
-void qedr_unaffiliated_event(void *context,
-			     u8 event_code)
+void qedr_unaffiliated_event(void *context, u8 event_code)
 {
 	pr_err("unaffiliated event not implemented yet\n");
 }
diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 2d55597..25fb75d 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -1874,7 +1874,6 @@ int qedr_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 			 qp_params.sgid.dwords[2], qp_params.sgid.dwords[3]);
 		DP_DEBUG(dev, QEDR_MSG_QP, "remote_mac=[%pM]\n",
 			 qp_params.remote_mac_addr);
-;
 
 		qp_params.mtu = qp->mtu;
 		qp_params.lb_indication = false;
-- 
1.8.3.1

--
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 2/5 v2] qedr: fix formatting

[Leon Romanovsky]

[-- Attachment #1: Type: text/plain, Size: 608 bytes --]

On Tue, Jan 24, 2017 at 01:51:40PM +0200, Ram Amrani wrote:
> From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>
>
> remove standalone ';'.
> list function's parameters in a single line.
>
> Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
> ---
>  drivers/infiniband/hw/qedr/main.c  | 3 +--
>  drivers/infiniband/hw/qedr/verbs.c | 1 -
>  2 files changed, 1 insertion(+), 3 deletions(-)
>

Thanks,
Reviewed-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[PATCH 3/5 v2] qedr: fix RDMA CM loopback

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

The loopback logic in RDMA CM packets compares Ethernet addresses and
was accidently inverse.

Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
 drivers/infiniband/hw/qedr/qedr_cm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/qedr_cm.c b/drivers/infiniband/hw/qedr/qedr_cm.c
index 00361f3..a9a8d87 100644
--- a/drivers/infiniband/hw/qedr/qedr_cm.c
+++ b/drivers/infiniband/hw/qedr/qedr_cm.c
@@ -398,9 +398,9 @@ static inline int qedr_gsi_build_packet(struct qedr_dev *dev,
 	}
 
 	if (ether_addr_equal(udh.eth.smac_h, udh.eth.dmac_h))
-		packet->tx_dest = QED_ROCE_LL2_TX_DEST_NW;
-	else
 		packet->tx_dest = QED_ROCE_LL2_TX_DEST_LB;
+	else
+		packet->tx_dest = QED_ROCE_LL2_TX_DEST_NW;
 
 	packet->roce_mode = roce_mode;
 	memcpy(packet->header.vaddr, ud_header_buffer, header_size);
-- 
1.8.3.1

--
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 4/5 v2] qedr: fix memory leak in PD allocation and simplify it

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

Free the PD if no internal resource were availbale.
Move userspace code to under the relevant 'if'.

Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
 drivers/infiniband/hw/qedr/verbs.c | 26 ++++++++++++++++++--------
 1 file changed, 18 insertions(+), 8 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 25fb75d..66bd297 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -471,8 +471,6 @@ struct ib_pd *qedr_alloc_pd(struct ib_device *ibdev,
 			    struct ib_ucontext *context, struct ib_udata *udata)
 {
 	struct qedr_dev *dev = get_qedr_dev(ibdev);
-	struct qedr_ucontext *uctx = NULL;
-	struct qedr_alloc_pd_uresp uresp;
 	struct qedr_pd *pd;
 	u16 pd_id;
 	int rc;
@@ -489,21 +487,33 @@ struct ib_pd *qedr_alloc_pd(struct ib_device *ibdev,
 	if (!pd)
 		return ERR_PTR(-ENOMEM);
 
-	dev->ops->rdma_alloc_pd(dev->rdma_ctx, &pd_id);
+	rc = dev->ops->rdma_alloc_pd(dev->rdma_ctx, &pd_id);
+	if (rc)
+		goto err;
 
-	uresp.pd_id = pd_id;
 	pd->pd_id = pd_id;
 
 	if (udata && context) {
+		struct qedr_alloc_pd_uresp uresp;
+
+		uresp.pd_id = pd_id;
+
 		rc = ib_copy_to_udata(udata, &uresp, sizeof(uresp));
-		if (rc)
+		if (rc) {
 			DP_ERR(dev, "copy error pd_id=0x%x.\n", pd_id);
-		uctx = get_qedr_ucontext(context);
-		uctx->pd = pd;
-		pd->uctx = uctx;
+			dev->ops->rdma_dealloc_pd(dev->rdma_ctx, pd_id);
+			goto err;
+		}
+
+		pd->uctx = get_qedr_ucontext(context);
+		pd->uctx->pd = pd;
 	}
 
 	return &pd->ibpd;
+
+err:
+	kfree(pd);
+	return ERR_PTR(rc);
 }
 
 int qedr_dealloc_pd(struct ib_pd *ibpd)
-- 
1.8.3.1

--
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 5/5 v2] qedr: dispatch port active event from qedr_add

[Ram Amrani]

From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>

Relying on qede to trigger qedr on startup is problematic. When probing
both if qedr loads slowly then qede can assume qedr is missing and not
trigger it. This patch adds a triggering from qedr and protects against
a race via an atomic bit.

Signed-off-by: Ram Amrani <Ram.Amrani-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
 drivers/infiniband/hw/qedr/main.c | 20 ++++++++++++++------
 drivers/infiniband/hw/qedr/qedr.h |  5 +++++
 2 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c
index eac0bfc..3ac8aa5 100644
--- a/drivers/infiniband/hw/qedr/main.c
+++ b/drivers/infiniband/hw/qedr/main.c
@@ -791,6 +791,9 @@ static struct qedr_dev *qedr_add(struct qed_dev *cdev, struct pci_dev *pdev,
 		if (device_create_file(&dev->ibdev.dev, qedr_attributes[i]))
 			goto sysfs_err;
 
+	if (!test_and_set_bit(QEDR_ENET_STATE_BIT, &dev->enet_state))
+		qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ACTIVE);
+
 	DP_DEBUG(dev, QEDR_MSG_INIT, "qedr driver loaded successfully\n");
 	return dev;
 
@@ -823,11 +826,10 @@ static void qedr_remove(struct qedr_dev *dev)
 	ib_dealloc_device(&dev->ibdev);
 }
 
-static int qedr_close(struct qedr_dev *dev)
+static void qedr_close(struct qedr_dev *dev)
 {
-	qedr_ib_dispatch_event(dev, 1, IB_EVENT_PORT_ERR);
-
-	return 0;
+	if (test_and_clear_bit(QEDR_ENET_STATE_BIT, &dev->enet_state))
+		qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ERR);
 }
 
 static void qedr_shutdown(struct qedr_dev *dev)
@@ -836,6 +838,12 @@ static void qedr_shutdown(struct qedr_dev *dev)
 	qedr_remove(dev);
 }
 
+static void qedr_open(struct qedr_dev *dev)
+{
+	if (!test_and_set_bit(QEDR_ENET_STATE_BIT, &dev->enet_state))
+		qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ACTIVE);
+}
+
 static void qedr_mac_address_change(struct qedr_dev *dev)
 {
 	union ib_gid *sgid = &dev->sgid_tbl[0];
@@ -862,7 +870,7 @@ static void qedr_mac_address_change(struct qedr_dev *dev)
 
 	ether_addr_copy(dev->gsi_ll2_mac_address, dev->ndev->dev_addr);
 
-	qedr_ib_dispatch_event(dev, 1, IB_EVENT_GID_CHANGE);
+	qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_GID_CHANGE);
 
 	if (rc)
 		DP_ERR(dev, "Error updating mac filter\n");
@@ -876,7 +884,7 @@ static void qedr_notify(struct qedr_dev *dev, enum qede_roce_event event)
 {
 	switch (event) {
 	case QEDE_UP:
-		qedr_ib_dispatch_event(dev, 1, IB_EVENT_PORT_ACTIVE);
+		qedr_open(dev);
 		break;
 	case QEDE_DOWN:
 		qedr_close(dev);
diff --git a/drivers/infiniband/hw/qedr/qedr.h b/drivers/infiniband/hw/qedr/qedr.h
index 94319ab..bb32e47 100644
--- a/drivers/infiniband/hw/qedr/qedr.h
+++ b/drivers/infiniband/hw/qedr/qedr.h
@@ -113,6 +113,8 @@ struct qedr_device_attr {
 	struct qed_rdma_events events;
 };
 
+#define QEDR_ENET_STATE_BIT	(0)
+
 struct qedr_dev {
 	struct ib_device	ibdev;
 	struct qed_dev		*cdev;
@@ -153,6 +155,8 @@ struct qedr_dev {
 	struct qedr_cq		*gsi_sqcq;
 	struct qedr_cq		*gsi_rqcq;
 	struct qedr_qp		*gsi_qp;
+
+	unsigned long enet_state;
 };
 
 #define QEDR_MAX_SQ_PBL			(0x8000)
@@ -188,6 +192,7 @@ struct qedr_dev {
 #define QEDR_ROCE_MAX_CNQ_SIZE		(0x4000)
 
 #define QEDR_MAX_PORT			(1)
+#define QEDR_PORT			(1)
 
 #define QEDR_UVERBS(CMD_NAME) (1ull << IB_USER_VERBS_CMD_##CMD_NAME)
 
-- 
1.8.3.1

--
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 0/5 v2] qedr: batch of fixes for 4.10

[Doug Ledford]

[-- Attachment #1: Type: text/plain, Size: 1477 bytes --]

On Tue, 2017-01-24 at 13:51 +0200, Ram Amrani wrote:
> From: Ram Amrani <Ram.Amrani-74tsMCuadCbQT0dZR+AlfA@public.gmane.org>
> 
> This series contains a batch of fixes for 4.10.
> 
> Based on origin/k.o/for-4.10, over
> "[PATCH 0/5 v2] qedr: QP and CQ fixes for 4.10",
> on top of 5cc8fab.
> 
> Changes from v1:
>  - exactly the same, except that it is built on top of
>    "[PATCH 0/5 v2] qedr: QP and CQ fixes for 4.10"
> 
> Ram Amrani (5):
>   qedr: mark three functions as static
>   qedr: fix formatting
>   qedr: fix RDMA CM loopback
>   qedr: fix memory leak in PD allocation and simplify it
>   qedr: dispatch port active event from qedr_add
> 
>  drivers/infiniband/hw/qedr/main.c    | 23 +++++++++++++-------
>  drivers/infiniband/hw/qedr/qedr.h    |  5 +++++
>  drivers/infiniband/hw/qedr/qedr_cm.c |  4 ++--
>  drivers/infiniband/hw/qedr/verbs.c   | 41 +++++++++++++++++++++++---
> ----------
>  4 files changed, 48 insertions(+), 25 deletions(-)

I took these, but I would like to point out a couple things.  First,
the commit messages could use some work.  I fixed a number of them up.
 Also, a fix to formatting isn't really an -rc fix.  I took it, but
next time let's be more selective on our submission please.

-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG KeyID: B826A3330E572FDD
   
Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

RE: [PATCH 0/5 v2] qedr: batch of fixes for 4.10

[Amrani, Ram]

> I took these, but I would like to point out a couple things.  First,
> the commit messages could use some work.  I fixed a number of them up.
>  Also, a fix to formatting isn't really an -rc fix.  I took it, but
> next time let's be more selective on our submission please.
> 

Thanks.
I have reviewed your changes for next time. I understood all but these:

I thought that introducing formatting fixes and making functions static are
simplifying and risk free operations that are acceptable. I now understand,
and correct me if I'm wrong, that only if it's important fix it should be introduced.

Sometimes I see IB/vendor and sometimes RDMA/vendor. When to use each?

Thanks,
Ram

Re: [PATCH 0/5 v2] qedr: batch of fixes for 4.10

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 1572 bytes --]

On 1/26/2017 8:44 AM, Amrani, Ram wrote:
>> I took these, but I would like to point out a couple things.  First,
>> the commit messages could use some work.  I fixed a number of them up.
>>  Also, a fix to formatting isn't really an -rc fix.  I took it, but
>> next time let's be more selective on our submission please.
>>
> 
> Thanks.
> I have reviewed your changes for next time. I understood all but these:
> 
> I thought that introducing formatting fixes and making functions static are
> simplifying and risk free operations that are acceptable.

Yes and no.  They are risk free, but they also provide 0 benefit to the
actual running of the kernel.  Therefore, to reduce churn during -rc
cycle kernels, you would normally not submit them.  You would hold on to
them until the next merge window.

> I now understand,
> and correct me if I'm wrong, that only if it's important fix it should be introduced.

Correct.

> Sometimes I see IB/vendor and sometimes RDMA/vendor. When to use each?

I'm trying to transition people over to using RDMA/* as much as possible
because the stack is much more than just IB these days.  But if someone
sends me patches that have IB/ as the prefix, I don't go in and modify
every subject line.  So, when I modified your subject lines, I used
RDMA, the IB subject lines you see in my current queue are from other
people.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]

RE: [PATCH 0/5 v2] qedr: batch of fixes for 4.10

[Amrani, Ram]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 1600 bytes --]

> On 1/26/2017 8:44 AM, Amrani, Ram wrote:
> >> I took these, but I would like to point out a couple things.  First,
> >> the commit messages could use some work.  I fixed a number of them up.
> >>  Also, a fix to formatting isn't really an -rc fix.  I took it, but
> >> next time let's be more selective on our submission please.
> >>
> >
> > Thanks.
> > I have reviewed your changes for next time. I understood all but these:
> >
> > I thought that introducing formatting fixes and making functions static are
> > simplifying and risk free operations that are acceptable.
> 
> Yes and no.  They are risk free, but they also provide 0 benefit to the
> actual running of the kernel.  Therefore, to reduce churn during -rc
> cycle kernels, you would normally not submit them.  You would hold on to
> them until the next merge window.
> 
> > I now understand,
> > and correct me if I'm wrong, that only if it's important fix it should be introduced.
> 
> Correct.
> 
> > Sometimes I see IB/vendor and sometimes RDMA/vendor. When to use each?
> 
> I'm trying to transition people over to using RDMA/* as much as possible
> because the stack is much more than just IB these days.  But if someone
> sends me patches that have IB/ as the prefix, I don't go in and modify
> every subject line.  So, when I modified your subject lines, I used
> RDMA, the IB subject lines you see in my current queue are from other
> people.
> 

OK

Ram

N‹§²æìr¸›yúèšØb²X¬¶Ç§vØ^–)Þº{.nÇ+‰·¥Š{±­ÙšŠ{ayº\x1dʇڙë,j\a­¢f£¢·hš‹»öì\x17/oSc¾™Ú³9˜uÀ¦æå‰È&jw¨®\x03(­éšŽŠÝ¢j"ú\x1a¶^[m§ÿïêäz¹Þ–Šàþf£¢·hšˆ§~ˆmš