[PATCH for-next 00/11] IB/hfi1: Driver fixes for 11/6/2017

[PATCH for-next 00/11] IB/hfi1: Driver fixes for 11/6/2017

[Dennis Dalessandro]

Hi Doug,

More fixes for our drivers. Nothing scary or out of the ordinary here. There
is a core change, but it is for extended lids. This does include a patch
suggested by Dan and is based off of his patch and the ensuing discussion on the
list.

As always patches available for browsing at:
https://github.com/ddalessa/kernel/tree/for-4.15

---

Dennis Dalessandro (1):
      IB/hfi1: Handle initial value of 0 for CCTI setting

Don Hiatt (2):
      IB/core: Convert OPA AH to IB for Extended LIDs only
      IB/hfi1: Mask upper 16Bits of Extended LID prior to rvt_cq_entry

Grzegorz Morys (1):
      IB/hfi1: Prohibit invalid Init to Armed state transition

Jakub Byczkowski (1):
      IB/hfi1: Reduce 8051 command timeout

Jan Sokolowski (3):
      IB/hfi1: Allow MgmtAllowed on B2B setups
      IB/hfi1: Remove unnecessary if check
      IB/hfi1: Send 'reboot' as planned down remote reason

Kamenee Arumugam (1):
      IB/hfi1: Remove wrapper function in mmu_rb

Mike Marciniszyn (1):
      IB/hfi1: Fix a wrapping test to insure the correct timeout

Niranjana Vishwanathapura (1):
      IB/hfi1: Do not allocate PIO send contexts for VNIC


 drivers/infiniband/core/uverbs_marshall.c |    3 +-
 drivers/infiniband/hw/hfi1/chip.c         |   55 +++++------------------------
 drivers/infiniband/hw/hfi1/chip.h         |    2 +
 drivers/infiniband/hw/hfi1/driver.c       |    5 +--
 drivers/infiniband/hw/hfi1/hfi.h          |    3 ++
 drivers/infiniband/hw/hfi1/init.c         |    3 +-
 drivers/infiniband/hw/hfi1/intr.c         |   43 +++++++++++++++++++++++
 drivers/infiniband/hw/hfi1/mad.c          |   44 ++++++++++++++---------
 drivers/infiniband/hw/hfi1/mad.h          |    2 +
 drivers/infiniband/hw/hfi1/mmu_rb.c       |   24 ++++---------
 drivers/infiniband/hw/hfi1/pio.c          |   17 ---------
 drivers/infiniband/hw/hfi1/pio.h          |    6 ---
 drivers/infiniband/hw/hfi1/rc.c           |    4 +-
 drivers/infiniband/hw/hfi1/ruc.c          |    2 +
 drivers/infiniband/hw/hfi1/uc.c           |    2 +
 drivers/infiniband/hw/hfi1/ud.c           |    6 ++-
 drivers/infiniband/hw/hfi1/user_sdma.c    |   30 ++++++++--------
 drivers/infiniband/hw/hfi1/vnic_main.c    |   37 +-------------------
 18 files changed, 118 insertions(+), 170 deletions(-)

--
-Denny
--
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-next 01/11] IB/hfi1: Allow MgmtAllowed on B2B setups

[Dennis Dalessandro]

From: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

HFI's are hard-wired to send Device Info frames with
MgmtAllowed bit set to 0. This means in B2B setups,
MgmtAllowed would never be allowed, which prevents
remote opa management tools from working properly.

Assume MgmtAllowed if a neighbor is also an HFI.

Fixes: 98b9ee2002a8 ("IB/hfi1: Cache neighbor secure data after link up")
Reviewed-by: Sebastian Sanchez <sebastian.sanchez-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/chip.c |   39 +---------------------------------
 drivers/infiniband/hw/hfi1/hfi.h  |    3 +++
 drivers/infiniband/hw/hfi1/intr.c |   43 +++++++++++++++++++++++++++++++++++++
 3 files changed, 47 insertions(+), 38 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index 05e03a0..2523b62 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -1036,7 +1036,6 @@ static void read_vc_local_link_width(struct hfi1_devdata *dd, u8 *misc_bits,
 				     u8 *flag_bits, u16 *link_widths);
 static void read_remote_device_id(struct hfi1_devdata *dd, u16 *device_id,
 				  u8 *device_rev);
-static void read_mgmt_allowed(struct hfi1_devdata *dd, u8 *mgmt_allowed);
 static void read_local_lni(struct hfi1_devdata *dd, u8 *enable_lane_rx);
 static int read_tx_settings(struct hfi1_devdata *dd, u8 *enable_lane_tx,
 			    u8 *tx_polarity_inversion,
@@ -7197,27 +7196,6 @@ static int lcb_to_port_ltp(int lcb_crc)
 	return port_ltp;
 }
 
-/*
- * Our neighbor has indicated that we are allowed to act as a fabric
- * manager, so place the full management partition key in the second
- * (0-based) pkey array position (see OPAv1, section 20.2.2.6.8). Note
- * that we should already have the limited management partition key in
- * array element 1, and also that the port is not yet up when
- * add_full_mgmt_pkey() is invoked.
- */
-static void add_full_mgmt_pkey(struct hfi1_pportdata *ppd)
-{
-	struct hfi1_devdata *dd = ppd->dd;
-
-	/* Sanity check - ppd->pkeys[2] should be 0, or already initialized */
-	if (!((ppd->pkeys[2] == 0) || (ppd->pkeys[2] == FULL_MGMT_P_KEY)))
-		dd_dev_warn(dd, "%s pkey[2] already set to 0x%x, resetting it to 0x%x\n",
-			    __func__, ppd->pkeys[2], FULL_MGMT_P_KEY);
-	ppd->pkeys[2] = FULL_MGMT_P_KEY;
-	(void)hfi1_set_ib_cfg(ppd, HFI1_IB_CFG_PKEYS, 0);
-	hfi1_event_pkey_change(ppd->dd, ppd->port);
-}
-
 static void clear_full_mgmt_pkey(struct hfi1_pportdata *ppd)
 {
 	if (ppd->pkeys[2] != 0) {
@@ -7414,11 +7392,7 @@ void handle_verify_cap(struct work_struct *work)
 			      &partner_supported_crc);
 	read_vc_remote_link_width(dd, &remote_tx_rate, &link_widths);
 	read_remote_device_id(dd, &device_id, &device_rev);
-	/*
-	 * And the 'MgmtAllowed' information, which is exchanged during
-	 * LNI, is also be available at this point.
-	 */
-	read_mgmt_allowed(dd, &ppd->mgmt_allowed);
+
 	/* print the active widths */
 	get_link_widths(dd, &active_tx, &active_rx);
 	dd_dev_info(dd,
@@ -7546,9 +7520,6 @@ void handle_verify_cap(struct work_struct *work)
 	write_csr(dd, DC_LCB_ERR_EN, 0); /* mask LCB errors */
 	set_8051_lcb_access(dd);
 
-	if (ppd->mgmt_allowed)
-		add_full_mgmt_pkey(ppd);
-
 	/* tell the 8051 to go to LinkUp */
 	set_link_state(ppd, HLS_GOING_UP);
 }
@@ -8960,14 +8931,6 @@ static void read_local_lni(struct hfi1_devdata *dd, u8 *enable_lane_rx)
 	*enable_lane_rx = (frame >> ENABLE_LANE_RX_SHIFT) & ENABLE_LANE_RX_MASK;
 }
 
-static void read_mgmt_allowed(struct hfi1_devdata *dd, u8 *mgmt_allowed)
-{
-	u32 frame;
-
-	read_8051_config(dd, REMOTE_LNI_INFO, GENERAL_CONFIG, &frame);
-	*mgmt_allowed = (frame >> MGMT_ALLOWED_SHIFT) & MGMT_ALLOWED_MASK;
-}
-
 static void read_last_local_state(struct hfi1_devdata *dd, u32 *lls)
 {
 	read_8051_config(dd, LAST_LOCAL_STATE_COMPLETE, GENERAL_CONFIG, lls);
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 08d394f..4a9b4d7 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -95,6 +95,9 @@
 #define DROP_PACKET_OFF		0
 #define DROP_PACKET_ON		1
 
+#define NEIGHBOR_TYPE_HFI		0
+#define NEIGHBOR_TYPE_SWITCH	1
+
 extern unsigned long hfi1_cap_mask;
 #define HFI1_CAP_KGET_MASK(mask, cap) ((mask) & HFI1_CAP_##cap)
 #define HFI1_CAP_UGET_MASK(mask, cap) \
diff --git a/drivers/infiniband/hw/hfi1/intr.c b/drivers/infiniband/hw/hfi1/intr.c
index 3e3184d..387305b 100644
--- a/drivers/infiniband/hw/hfi1/intr.c
+++ b/drivers/infiniband/hw/hfi1/intr.c
@@ -55,6 +55,40 @@
 
 #define LINK_UP_DELAY  500  /* in microseconds */
 
+static void set_mgmt_allowed(struct hfi1_pportdata *ppd)
+{
+	u32 frame;
+	struct hfi1_devdata *dd = ppd->dd;
+
+	if (ppd->neighbor_type == NEIGHBOR_TYPE_HFI) {
+		ppd->mgmt_allowed = 1;
+	} else {
+		read_8051_config(dd, REMOTE_LNI_INFO, GENERAL_CONFIG, &frame);
+		ppd->mgmt_allowed = (frame >> MGMT_ALLOWED_SHIFT)
+		& MGMT_ALLOWED_MASK;
+	}
+}
+
+/*
+ * Our neighbor has indicated that we are allowed to act as a fabric
+ * manager, so place the full management partition key in the second
+ * (0-based) pkey array position. Note that we should already have
+ * the limited management partition key in array element 1, and also
+ * that the port is not yet up when add_full_mgmt_pkey() is invoked.
+ */
+static void add_full_mgmt_pkey(struct hfi1_pportdata *ppd)
+{
+	struct hfi1_devdata *dd = ppd->dd;
+
+	/* Sanity check - ppd->pkeys[2] should be 0, or already initialized */
+	if (!((ppd->pkeys[2] == 0) || (ppd->pkeys[2] == FULL_MGMT_P_KEY)))
+		dd_dev_warn(dd, "%s pkey[2] already set to 0x%x, resetting it to 0x%x\n",
+			    __func__, ppd->pkeys[2], FULL_MGMT_P_KEY);
+	ppd->pkeys[2] = FULL_MGMT_P_KEY;
+	(void)hfi1_set_ib_cfg(ppd, HFI1_IB_CFG_PKEYS, 0);
+	hfi1_event_pkey_change(ppd->dd, ppd->port);
+}
+
 /**
  * format_hwmsg - format a single hwerror message
  * @msg message buffer
@@ -163,6 +197,15 @@ void handle_linkup_change(struct hfi1_devdata *dd, u32 linkup)
 		/* HW needs LINK_UP_DELAY to settle, give it that chance */
 		udelay(LINK_UP_DELAY);
 
+		/*
+		 * 'MgmtAllowed' information, which is exchanged during
+		 * LNI, is available at this point.
+		 */
+		set_mgmt_allowed(ppd);
+
+		if (ppd->mgmt_allowed)
+			add_full_mgmt_pkey(ppd);
+
 		/* physical link went up */
 		ppd->linkup = 1;
 		ppd->offline_disabled_reason =

--
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-next 02/11] IB/hfi1: Reduce 8051 command timeout

[Dennis Dalessandro]

From: Jakub Byczkowski <jakub.byczkowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Timeout of 20 seconds is too long for active wait performed
for 8051 command completion. It was required for scenarios
when transition to polling was requested before offline.quiet
state was reached. Currently wait for offline.quiet is
properly implemented and timeout can be reduced to 1 second.

Reviewed-by: Dean Luick <dean.luick-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Duane McCrory <duane.mccrory-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Jakub Byczkowski <jakub.byczkowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/chip.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/chip.h b/drivers/infiniband/hw/hfi1/chip.h
index fc122c1..133e313 100644
--- a/drivers/infiniband/hw/hfi1/chip.h
+++ b/drivers/infiniband/hw/hfi1/chip.h
@@ -560,7 +560,7 @@ enum {
 /* timeouts */
 #define LINK_RESTART_DELAY 1000		/* link restart delay, in ms */
 #define TIMEOUT_8051_START 5000         /* 8051 start timeout, in ms */
-#define DC8051_COMMAND_TIMEOUT 20000	/* DC8051 command timeout, in ms */
+#define DC8051_COMMAND_TIMEOUT 1000	/* DC8051 command timeout, in ms */
 #define FREEZE_STATUS_TIMEOUT 20	/* wait for freeze indicators, in ms */
 #define VL_STATUS_CLEAR_TIMEOUT 5000	/* per-VL status clear, in ms */
 #define CCE_STATUS_TIMEOUT 10		/* time to clear CCE Status, in ms */

--
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-next 03/11] IB/hfi1: Remove wrapper function in mmu_rb

[Dennis Dalessandro]

From: Kamenee Arumugam <kamenee.arumugam-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Wrapper functions were used to call the same function
mmu_notifier_mem_invalidate for 2 callbacks in
mmu_notifier.
The commit 7def96f0a973 ("IB/hfi1: update to new mmu_notifier semantic")
removed the invalidate_page callback.
Therefore, the wrapper function is no longer needed.

Reviewed-by: Michael J. Ruhl <michael.j.ruhl-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Alex Estrin <alex.estrin-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Kamenee Arumugam <kamenee.arumugam-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/mmu_rb.c |   24 +++++++-----------------
 1 files changed, 7 insertions(+), 17 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/mmu_rb.c b/drivers/infiniband/hw/hfi1/mmu_rb.c
index 175002c..e7b3ce1 100644
--- a/drivers/infiniband/hw/hfi1/mmu_rb.c
+++ b/drivers/infiniband/hw/hfi1/mmu_rb.c
@@ -67,12 +67,9 @@ struct mmu_rb_handler {
 
 static unsigned long mmu_node_start(struct mmu_rb_node *);
 static unsigned long mmu_node_last(struct mmu_rb_node *);
-static inline void mmu_notifier_range_start(struct mmu_notifier *,
-					    struct mm_struct *,
-					    unsigned long, unsigned long);
-static void mmu_notifier_mem_invalidate(struct mmu_notifier *,
-					struct mm_struct *,
-					unsigned long, unsigned long);
+static void mmu_notifier_range_start(struct mmu_notifier *,
+				     struct mm_struct *,
+				     unsigned long, unsigned long);
 static struct mmu_rb_node *__mmu_rb_search(struct mmu_rb_handler *,
 					   unsigned long, unsigned long);
 static void do_remove(struct mmu_rb_handler *handler,
@@ -286,17 +283,10 @@ void hfi1_mmu_rb_remove(struct mmu_rb_handler *handler,
 	handler->ops->remove(handler->ops_arg, node);
 }
 
-static inline void mmu_notifier_range_start(struct mmu_notifier *mn,
-					    struct mm_struct *mm,
-					    unsigned long start,
-					    unsigned long end)
-{
-	mmu_notifier_mem_invalidate(mn, mm, start, end);
-}
-
-static void mmu_notifier_mem_invalidate(struct mmu_notifier *mn,
-					struct mm_struct *mm,
-					unsigned long start, unsigned long end)
+static void mmu_notifier_range_start(struct mmu_notifier *mn,
+				     struct mm_struct *mm,
+				     unsigned long start,
+				     unsigned long end)
 {
 	struct mmu_rb_handler *handler =
 		container_of(mn, struct mmu_rb_handler, mn);

--
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-next 04/11] IB/hfi1: Fix a wrapping test to insure the correct timeout

[Dennis Dalessandro]

From: Mike Marciniszyn <mike.marciniszyn-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

The "2 * UINT_MAX"  statement:
	if ((u64)(ts - cce->timestamp) > 2 * UINT_MAX) {

is equivalent to:
	if ((u64)(ts - cce->timestamp) > UINT_MAX - 1) {

This results in a premature timeout of the cong log entry.

Fix by using unsigned 64 bit integers, removing casts, and using
an algebraic equivalent test to avoid the "2 * UINT_MAX" issue.

Also make use of kernel API to get nanoseconds instead of
open coding.

Reported-by: Dan Carpenter <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/mad.c |    6 +++---
 drivers/infiniband/hw/hfi1/mad.h |    2 +-
 drivers/infiniband/hw/hfi1/rc.c  |    2 +-
 3 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index e648f58..0a867e5 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -3760,7 +3760,7 @@ static int __subn_get_opa_hfi1_cong_log(struct opa_smp *smp, u32 am,
 	struct hfi1_ibport *ibp = to_iport(ibdev, port);
 	struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);
 	struct opa_hfi1_cong_log *cong_log = (struct opa_hfi1_cong_log *)data;
-	s64 ts;
+	u64 ts;
 	int i;
 
 	if (am || smp_length_check(sizeof(*cong_log), max_len)) {
@@ -3778,7 +3778,7 @@ static int __subn_get_opa_hfi1_cong_log(struct opa_smp *smp, u32 am,
 	       ppd->threshold_cong_event_map,
 	       sizeof(cong_log->threshold_cong_event_map));
 	/* keep timestamp in units of 1.024 usec */
-	ts = ktime_to_ns(ktime_get()) / 1024;
+	ts = ktime_get_ns() / 1024;
 	cong_log->current_time_stamp = cpu_to_be32(ts);
 	for (i = 0; i < OPA_CONG_LOG_ELEMS; i++) {
 		struct opa_hfi1_cong_log_event_internal *cce =
@@ -3790,7 +3790,7 @@ static int __subn_get_opa_hfi1_cong_log(struct opa_smp *smp, u32 am,
 		 * required to wrap the counter are supposed to
 		 * be zeroed (CA10-49 IBTA, release 1.2.1, V1).
 		 */
-		if ((u64)(ts - cce->timestamp) > (2 * UINT_MAX))
+		if ((ts - cce->timestamp) / 2 > U32_MAX)
 			continue;
 		memcpy(cong_log->events[i].local_qp_cn_entry, &cce->lqpn, 3);
 		memcpy(cong_log->events[i].remote_qp_number_cn_entry,
diff --git a/drivers/infiniband/hw/hfi1/mad.h b/drivers/infiniband/hw/hfi1/mad.h
index a4f8ac1..c4938f3 100644
--- a/drivers/infiniband/hw/hfi1/mad.h
+++ b/drivers/infiniband/hw/hfi1/mad.h
@@ -239,7 +239,7 @@ struct opa_hfi1_cong_log_event_internal {
 	u8 sl;
 	u8 svc_type;
 	u32 rlid;
-	s64 timestamp; /* wider than 32 bits to detect 32 bit rollover */
+	u64 timestamp; /* wider than 32 bits to detect 32 bit rollover */
 };
 
 struct opa_hfi1_cong_log_event {
diff --git a/drivers/infiniband/hw/hfi1/rc.c b/drivers/infiniband/hw/hfi1/rc.c
index 32d7bbe..3f21b05 100644
--- a/drivers/infiniband/hw/hfi1/rc.c
+++ b/drivers/infiniband/hw/hfi1/rc.c
@@ -1965,7 +1965,7 @@ static void log_cca_event(struct hfi1_pportdata *ppd, u8 sl, u32 rlid,
 	cc_event->svc_type = svc_type;
 	cc_event->rlid = rlid;
 	/* keep timestamp in units of 1.024 usec */
-	cc_event->timestamp = ktime_to_ns(ktime_get()) / 1024;
+	cc_event->timestamp = ktime_get_ns() / 1024;
 
 	spin_unlock_irqrestore(&ppd->cc_log_lock, flags);
 }

--
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-next 05/11] IB/hfi1: Remove unnecessary if check

[Dennis Dalessandro]

From: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

A for loop condition of data_iovs in user_sdma_free_request
is unnecessarily repeated before the loop as an if check.

Remove the if enveloping the loop.

Reviewed-by: Jakub Byczkowski <jakub.byczkowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/user_sdma.c |   30 +++++++++++++++---------------
 1 files changed, 15 insertions(+), 15 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c
index 01b9a9c..833d0b7 100644
--- a/drivers/infiniband/hw/hfi1/user_sdma.c
+++ b/drivers/infiniband/hw/hfi1/user_sdma.c
@@ -1428,6 +1428,8 @@ static inline void pq_update(struct hfi1_user_sdma_pkt_q *pq)
 
 static void user_sdma_free_request(struct user_sdma_request *req, bool unpin)
 {
+	int i;
+
 	if (!list_empty(&req->txps)) {
 		struct sdma_txreq *t, *p;
 
@@ -1439,22 +1441,20 @@ static void user_sdma_free_request(struct user_sdma_request *req, bool unpin)
 			kmem_cache_free(req->pq->txreq_cache, tx);
 		}
 	}
-	if (req->data_iovs) {
-		struct sdma_mmu_node *node;
-		int i;
-
-		for (i = 0; i < req->data_iovs; i++) {
-			node = req->iovs[i].node;
-			if (!node)
-				continue;
-
-			if (unpin)
-				hfi1_mmu_rb_remove(req->pq->handler,
-						   &node->rb);
-			else
-				atomic_dec(&node->refcount);
-		}
+
+	for (i = 0; i < req->data_iovs; i++) {
+		struct sdma_mmu_node *node = req->iovs[i].node;
+
+		if (!node)
+			continue;
+
+		if (unpin)
+			hfi1_mmu_rb_remove(req->pq->handler,
+					   &node->rb);
+		else
+			atomic_dec(&node->refcount);
 	}
+
 	kfree(req->tids);
 	clear_bit(req->info.comp_idx, req->pq->req_in_use);
 }

--
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-next 06/11] IB/hfi1: Do not allocate PIO send contexts for VNIC

[Dennis Dalessandro]

From: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

OPA VNIC does not use PIO contexts and instead only uses SDMA
engines. Do not allocate PIO contexts for VNIC ports.

Reviewed-by: Michael J. Ruhl <michael.j.ruhl-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/chip.c      |   10 ++++-----
 drivers/infiniband/hw/hfi1/driver.c    |    5 ++--
 drivers/infiniband/hw/hfi1/init.c      |    3 +--
 drivers/infiniband/hw/hfi1/pio.c       |   17 ---------------
 drivers/infiniband/hw/hfi1/pio.h       |    6 -----
 drivers/infiniband/hw/hfi1/vnic_main.c |   37 ++------------------------------
 6 files changed, 10 insertions(+), 68 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index 2523b62..bb3b65a 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -6816,7 +6816,8 @@ static void rxe_kernel_unfreeze(struct hfi1_devdata *dd)
 		rcd = hfi1_rcd_get_by_index(dd, i);
 
 		/* Ensure all non-user contexts(including vnic) are enabled */
-		if (!rcd || !rcd->sc || (rcd->sc->type == SC_USER)) {
+		if (!rcd ||
+		    (i >= dd->first_dyn_alloc_ctxt && !rcd->is_vnic)) {
 			hfi1_rcd_put(rcd);
 			continue;
 		}
@@ -8093,8 +8094,7 @@ static void is_rcv_avail_int(struct hfi1_devdata *dd, unsigned int source)
 		rcd = hfi1_rcd_get_by_index(dd, source);
 		if (rcd) {
 			/* Check for non-user contexts, including vnic */
-			if ((source < dd->first_dyn_alloc_ctxt) ||
-			    (rcd->sc && (rcd->sc->type == SC_KERNEL)))
+			if (source < dd->first_dyn_alloc_ctxt || rcd->is_vnic)
 				rcd->do_interrupt(rcd, 0);
 			else
 				handle_user_interrupt(rcd);
@@ -8124,8 +8124,8 @@ static void is_rcv_urgent_int(struct hfi1_devdata *dd, unsigned int source)
 		rcd = hfi1_rcd_get_by_index(dd, source);
 		if (rcd) {
 			/* only pay attention to user urgent interrupts */
-			if ((source >= dd->first_dyn_alloc_ctxt) &&
-			    (!rcd->sc || (rcd->sc->type == SC_USER)))
+			if (source >= dd->first_dyn_alloc_ctxt &&
+			    !rcd->is_vnic)
 				handle_user_interrupt(rcd);
 
 			hfi1_rcd_put(rcd);
diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c
index 07a32a1..4f65ac6 100644
--- a/drivers/infiniband/hw/hfi1/driver.c
+++ b/drivers/infiniband/hw/hfi1/driver.c
@@ -929,10 +929,9 @@ void set_all_slowpath(struct hfi1_devdata *dd)
 		rcd = hfi1_rcd_get_by_index(dd, i);
 		if (!rcd)
 			continue;
-		if ((i < dd->first_dyn_alloc_ctxt) ||
-		    (rcd->sc && (rcd->sc->type == SC_KERNEL))) {
+		if (i < dd->first_dyn_alloc_ctxt || rcd->is_vnic)
 			rcd->do_interrupt = &handle_receive_interrupt;
-		}
+
 		hfi1_rcd_put(rcd);
 	}
 }
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index f29e7a3..8e3b3e7 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -1807,8 +1807,7 @@ int hfi1_create_rcvhdrq(struct hfi1_devdata *dd, struct hfi1_ctxtdata *rcd)
 		amt = PAGE_ALIGN(rcd->rcvhdrq_cnt * rcd->rcvhdrqentsize *
 				 sizeof(u32));
 
-		if ((rcd->ctxt < dd->first_dyn_alloc_ctxt) ||
-		    (rcd->sc && (rcd->sc->type == SC_KERNEL)))
+		if (rcd->ctxt < dd->first_dyn_alloc_ctxt || rcd->is_vnic)
 			gfp_flags = GFP_KERNEL;
 		else
 			gfp_flags = GFP_USER;
diff --git a/drivers/infiniband/hw/hfi1/pio.c b/drivers/infiniband/hw/hfi1/pio.c
index 7108a4b..c929f5b 100644
--- a/drivers/infiniband/hw/hfi1/pio.c
+++ b/drivers/infiniband/hw/hfi1/pio.c
@@ -703,7 +703,6 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
 {
 	struct send_context_info *sci;
 	struct send_context *sc = NULL;
-	int req_type = type;
 	dma_addr_t dma;
 	unsigned long flags;
 	u64 reg;
@@ -730,13 +729,6 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
 		return NULL;
 	}
 
-	/*
-	 * VNIC contexts are dynamically allocated.
-	 * Hence, pick a user context for VNIC.
-	 */
-	if (type == SC_VNIC)
-		type = SC_USER;
-
 	spin_lock_irqsave(&dd->sc_lock, flags);
 	ret = sc_hw_alloc(dd, type, &sw_index, &hw_context);
 	if (ret) {
@@ -746,15 +738,6 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
 		return NULL;
 	}
 
-	/*
-	 * VNIC contexts are used by kernel driver.
-	 * Hence, mark them as kernel contexts.
-	 */
-	if (req_type == SC_VNIC) {
-		dd->send_contexts[sw_index].type = SC_KERNEL;
-		type = SC_KERNEL;
-	}
-
 	sci = &dd->send_contexts[sw_index];
 	sci->sc = sc;
 
diff --git a/drivers/infiniband/hw/hfi1/pio.h b/drivers/infiniband/hw/hfi1/pio.h
index 99ca5ed..058b08f 100644
--- a/drivers/infiniband/hw/hfi1/pio.h
+++ b/drivers/infiniband/hw/hfi1/pio.h
@@ -54,12 +54,6 @@
 #define SC_USER   3	/* must be the last one: it may take all left */
 #define SC_MAX    4	/* count of send context types */
 
-/*
- * SC_VNIC types are allocated (dynamically) from the user context pool,
- * (SC_USER) and used by kernel driver as kernel contexts (SC_KERNEL).
- */
-#define SC_VNIC   SC_MAX
-
 /* invalid send context index */
 #define INVALID_SCI 0xff
 
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
index 1a17708..5d65582 100644
--- a/drivers/infiniband/hw/hfi1/vnic_main.c
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -67,8 +67,6 @@ static int setup_vnic_ctxt(struct hfi1_devdata *dd, struct hfi1_ctxtdata *uctxt)
 	unsigned int rcvctrl_ops = 0;
 	int ret;
 
-	hfi1_init_ctxt(uctxt->sc);
-
 	uctxt->do_interrupt = &handle_receive_interrupt;
 
 	/* Now allocate the RcvHdr queue and eager buffers. */
@@ -96,8 +94,6 @@ static int setup_vnic_ctxt(struct hfi1_devdata *dd, struct hfi1_ctxtdata *uctxt)
 		rcvctrl_ops |= HFI1_RCVCTRL_TAILUPD_ENB;
 
 	hfi1_rcvctrl(uctxt->dd, rcvctrl_ops, uctxt);
-
-	uctxt->is_vnic = true;
 done:
 	return ret;
 }
@@ -122,20 +118,7 @@ static int allocate_vnic_ctxt(struct hfi1_devdata *dd,
 			HFI1_CAP_KGET(NODROP_EGR_FULL) |
 			HFI1_CAP_KGET(DMA_RTAIL);
 	uctxt->seq_cnt = 1;
-
-	/* Allocate and enable a PIO send context */
-	uctxt->sc = sc_alloc(dd, SC_VNIC, uctxt->rcvhdrqentsize,
-			     uctxt->numa_id);
-
-	ret = uctxt->sc ? 0 : -ENOMEM;
-	if (ret)
-		goto bail;
-
-	dd_dev_dbg(dd, "allocated vnic send context %u(%u)\n",
-		   uctxt->sc->sw_index, uctxt->sc->hw_context);
-	ret = sc_enable(uctxt->sc);
-	if (ret)
-		goto bail;
+	uctxt->is_vnic = true;
 
 	if (dd->num_msix_entries)
 		hfi1_set_vnic_msix_info(uctxt);
@@ -144,11 +127,7 @@ static int allocate_vnic_ctxt(struct hfi1_devdata *dd,
 	dd_dev_dbg(dd, "created vnic context %d\n", uctxt->ctxt);
 	*vnic_ctxt = uctxt;
 
-	return ret;
-bail:
-	hfi1_free_ctxt(uctxt);
-	dd_dev_dbg(dd, "vnic allocation failed. rc %d\n", ret);
-	return ret;
+	return 0;
 }
 
 static void deallocate_vnic_ctxt(struct hfi1_devdata *dd,
@@ -170,18 +149,6 @@ static void deallocate_vnic_ctxt(struct hfi1_devdata *dd,
 		     HFI1_RCVCTRL_ONE_PKT_EGR_DIS |
 		     HFI1_RCVCTRL_NO_RHQ_DROP_DIS |
 		     HFI1_RCVCTRL_NO_EGR_DROP_DIS, uctxt);
-	/*
-	 * VNIC contexts are allocated from user context pool.
-	 * Release them back to user context pool.
-	 *
-	 * Reset context integrity checks to default.
-	 * (writes to CSRs probably belong in chip.c)
-	 */
-	write_kctxt_csr(dd, uctxt->sc->hw_context, SEND_CTXT_CHECK_ENABLE,
-			hfi1_pkt_default_send_ctxt_mask(dd, SC_USER));
-	sc_disable(uctxt->sc);
-
-	dd->send_contexts[uctxt->sc->sw_index].type = SC_USER;
 
 	uctxt->event_flags = 0;
 

--
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-next 07/11] IB/hfi1: Prohibit invalid Init to Armed state transition

[Dennis Dalessandro]

From: Grzegorz Morys <grzegorz.morys-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

It is invalid to change Link state from Init to Armed if
IsSmConfigurationStarted bit is not set in Attribute modifier
for Set subnet management method in case of PortInfo
and PortStateInfo attribute.
Set response MAD status field bits accordingly to react correctly
in such situations and avoid changing Link state.

Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Grzegorz Morys <grzegorz.morys-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/mad.c |   32 ++++++++++++++++++--------------
 1 files changed, 18 insertions(+), 14 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index 0a867e5..1277e1c 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -1238,8 +1238,7 @@ static int port_states_transition_allowed(struct hfi1_pportdata *ppd,
 }
 
 static int set_port_states(struct hfi1_pportdata *ppd, struct opa_smp *smp,
-			   u32 logical_state, u32 phys_state,
-			   int suppress_idle_sma)
+			   u32 logical_state, u32 phys_state)
 {
 	struct hfi1_devdata *dd = ppd->dd;
 	u32 link_state;
@@ -1320,7 +1319,7 @@ static int set_port_states(struct hfi1_pportdata *ppd, struct opa_smp *smp,
 		break;
 	case IB_PORT_ARMED:
 		ret = set_link_state(ppd, HLS_UP_ARMED);
-		if ((ret == 0) && (suppress_idle_sma == 0))
+		if (!ret)
 			send_idle_sma(dd, SMA_IDLE_ARM);
 		break;
 	case IB_PORT_ACTIVE:
@@ -1614,8 +1613,10 @@ static int __subn_set_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
 			if (ls_new == ls_old || (ls_new == IB_PORT_ARMED))
 				ppd->is_sm_config_started = 1;
 		} else if (ls_new == IB_PORT_ARMED) {
-			if (ppd->is_sm_config_started == 0)
+			if (ppd->is_sm_config_started == 0) {
 				invalid = 1;
+				smp->status |= IB_SMP_INVALID_FIELD;
+			}
 		}
 	}
 
@@ -1632,9 +1633,11 @@ static int __subn_set_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
 	 * is down or is being set to down.
 	 */
 
-	ret = set_port_states(ppd, smp, ls_new, ps_new, invalid);
-	if (ret)
-		return ret;
+	if (!invalid) {
+		ret = set_port_states(ppd, smp, ls_new, ps_new);
+		if (ret)
+			return ret;
+	}
 
 	ret = __subn_get_opa_portinfo(smp, am, data, ibdev, port, resp_len,
 				      max_len);
@@ -2111,17 +2114,18 @@ static int __subn_set_opa_psi(struct opa_smp *smp, u32 am, u8 *data,
 			if (ls_new == ls_old || (ls_new == IB_PORT_ARMED))
 				ppd->is_sm_config_started = 1;
 		} else if (ls_new == IB_PORT_ARMED) {
-			if (ppd->is_sm_config_started == 0)
+			if (ppd->is_sm_config_started == 0) {
 				invalid = 1;
+				smp->status |= IB_SMP_INVALID_FIELD;
+			}
 		}
 	}
 
-	ret = set_port_states(ppd, smp, ls_new, ps_new, invalid);
-	if (ret)
-		return ret;
-
-	if (invalid)
-		smp->status |= IB_SMP_INVALID_FIELD;
+	if (!invalid) {
+		ret = set_port_states(ppd, smp, ls_new, ps_new);
+		if (ret)
+			return ret;
+	}
 
 	return __subn_get_opa_psi(smp, am, data, ibdev, port, resp_len,
 				  max_len);

--
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-next 08/11] IB/hfi1: Send 'reboot' as planned down remote reason

[Dennis Dalessandro]

From: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

On host shutdown, driver sends 'SMA_Disabled' as a reason
for link down. This is incorrect.

Send 'reboot' as a linkdown reason.

Signed-off-by: Jan Sokolowski <jan.sokolowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Jakub Byczkowski <jakub.byczkowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/chip.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index bb3b65a..4f057e8 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -9786,9 +9786,9 @@ void hfi1_quiet_serdes(struct hfi1_pportdata *ppd)
 	cancel_delayed_work_sync(&ppd->start_link_work);
 
 	ppd->offline_disabled_reason =
-			HFI1_ODR_MASK(OPA_LINKDOWN_REASON_SMA_DISABLED);
-	set_link_down_reason(ppd, OPA_LINKDOWN_REASON_SMA_DISABLED, 0,
-			     OPA_LINKDOWN_REASON_SMA_DISABLED);
+			HFI1_ODR_MASK(OPA_LINKDOWN_REASON_REBOOT);
+	set_link_down_reason(ppd, OPA_LINKDOWN_REASON_REBOOT, 0,
+			     OPA_LINKDOWN_REASON_REBOOT);
 	set_link_state(ppd, HLS_DN_OFFLINE);
 
 	/* disable the port */

--
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-next 09/11] IB/core: Convert OPA AH to IB for Extended LIDs only

[Dennis Dalessandro]

From: Don Hiatt <don.hiatt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

When deciding to convert an OPA AH to IB we were incorrectly
including the IB multicast range. At this layer, all Extended
LIDs will be larger than IB_LID_PERMISSIVE. Change comparison
accordingly.

Fixes: d541e45500bd ("IB/core: Convert ah_attr from OPA to IB when copying to user")
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Don Hiatt <don.hiatt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/core/uverbs_marshall.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c
index aeb2824..bb372b4 100644
--- a/drivers/infiniband/core/uverbs_marshall.c
+++ b/drivers/infiniband/core/uverbs_marshall.c
@@ -69,8 +69,7 @@ void ib_copy_ah_attr_to_user(struct ib_device *device,
 	memset(&dst->grh.reserved, 0, sizeof(dst->grh.reserved));
 
 	if ((ah_attr->type == RDMA_AH_ATTR_TYPE_OPA) &&
-	    (rdma_ah_get_dlid(ah_attr) >=
-	     be16_to_cpu(IB_MULTICAST_LID_BASE)) &&
+	    (rdma_ah_get_dlid(ah_attr) > be16_to_cpu(IB_LID_PERMISSIVE)) &&
 	    (!rdma_ah_conv_opa_to_ib(device, &conv_ah, ah_attr)))
 		src = &conv_ah;
 

--
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-next 10/11] IB/hfi1: Mask upper 16Bits of Extended LID prior to rvt_cq_entry

[Dennis Dalessandro]

From: Don Hiatt <don.hiatt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Pass only the lower 16Bits of an Extended LIDs to rvt_cq_entry
to avoid triggering a WARN_ON_ONCE during conversion there.
These upper 16Bits are okay to drop as they are obtained elsewhere.

Fixes: 62ede7779904 ("Add OPA extended LID support")
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Don Hiatt <don.hiatt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/rc.c  |    2 +-
 drivers/infiniband/hw/hfi1/ruc.c |    2 +-
 drivers/infiniband/hw/hfi1/uc.c  |    2 +-
 drivers/infiniband/hw/hfi1/ud.c  |    6 +++---
 4 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/rc.c b/drivers/infiniband/hw/hfi1/rc.c
index 3f21b05..fd01a76 100644
--- a/drivers/infiniband/hw/hfi1/rc.c
+++ b/drivers/infiniband/hw/hfi1/rc.c
@@ -2219,7 +2219,7 @@ void hfi1_rc_rcv(struct hfi1_packet *packet)
 			wc.opcode = IB_WC_RECV;
 		wc.qp = &qp->ibqp;
 		wc.src_qp = qp->remote_qpn;
-		wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr);
+		wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr) & U16_MAX;
 		/*
 		 * It seems that IB mandates the presence of an SL in a
 		 * work completion only for the UD transport (see section
diff --git a/drivers/infiniband/hw/hfi1/ruc.c b/drivers/infiniband/hw/hfi1/ruc.c
index d450d4d..1a3ed6e 100644
--- a/drivers/infiniband/hw/hfi1/ruc.c
+++ b/drivers/infiniband/hw/hfi1/ruc.c
@@ -560,7 +560,7 @@ static void ruc_loopback(struct rvt_qp *sqp)
 	wc.byte_len = wqe->length;
 	wc.qp = &qp->ibqp;
 	wc.src_qp = qp->remote_qpn;
-	wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr);
+	wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr) & U16_MAX;
 	wc.sl = rdma_ah_get_sl(&qp->remote_ah_attr);
 	wc.port_num = 1;
 	/* Signal completion event if the solicited bit is set. */
diff --git a/drivers/infiniband/hw/hfi1/uc.c b/drivers/infiniband/hw/hfi1/uc.c
index fcd8186..1ee80a5 100644
--- a/drivers/infiniband/hw/hfi1/uc.c
+++ b/drivers/infiniband/hw/hfi1/uc.c
@@ -462,7 +462,7 @@ void hfi1_uc_rcv(struct hfi1_packet *packet)
 		wc.status = IB_WC_SUCCESS;
 		wc.qp = &qp->ibqp;
 		wc.src_qp = qp->remote_qpn;
-		wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr);
+		wc.slid = rdma_ah_get_dlid(&qp->remote_ah_attr) & U16_MAX;
 		/*
 		 * It seems that IB mandates the presence of an SL in a
 		 * work completion only for the UD transport (see section
diff --git a/drivers/infiniband/hw/hfi1/ud.c b/drivers/infiniband/hw/hfi1/ud.c
index 84c7dbe..5260b27 100644
--- a/drivers/infiniband/hw/hfi1/ud.c
+++ b/drivers/infiniband/hw/hfi1/ud.c
@@ -265,8 +265,8 @@ static void ud_loopback(struct rvt_qp *sqp, struct rvt_swqe *swqe)
 	} else {
 		wc.pkey_index = 0;
 	}
-	wc.slid = ppd->lid | (rdma_ah_get_path_bits(ah_attr) &
-				   ((1 << ppd->lmc) - 1));
+	wc.slid = (ppd->lid | (rdma_ah_get_path_bits(ah_attr) &
+				   ((1 << ppd->lmc) - 1))) & U16_MAX;
 	/* Check for loopback when the port lid is not set */
 	if (wc.slid == 0 && sqp->ibqp.qp_type == IB_QPT_GSI)
 		wc.slid = be16_to_cpu(IB_LID_PERMISSIVE);
@@ -1037,7 +1037,7 @@ void hfi1_ud_rcv(struct hfi1_packet *packet)
 	}
 	if (slid_is_permissive)
 		slid = be32_to_cpu(OPA_LID_PERMISSIVE);
-	wc.slid = slid;
+	wc.slid = slid & U16_MAX;
 	wc.sl = sl_from_sc;
 
 	/*

--
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-next 11/11] IB/hfi1: Handle initial value of 0 for CCTI setting

[Dennis Dalessandro]

When the driver is loaded it sets the default CCTI value to be 0. When the FM
starts and CCA is disabled the driver sets the max value to 65535 due the driver
subtracting 1 from 0 and the fact that the CCTI value is a u16.

Special case the subtraction to find the index for a 0 value.

Reviewed-by: Mike Marciniszyn <mike.marciniszyn-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/infiniband/hw/hfi1/mad.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index 1277e1c..cf8dba3 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -3701,7 +3701,11 @@ static void apply_cc_state(struct hfi1_pportdata *ppd)
 
 	*new_cc_state = *old_cc_state;
 
-	new_cc_state->cct.ccti_limit = ppd->total_cct_entry - 1;
+	if (ppd->total_cct_entry)
+		new_cc_state->cct.ccti_limit = ppd->total_cct_entry - 1;
+	else
+		new_cc_state->cct.ccti_limit = 0;
+
 	memcpy(new_cc_state->cct.entries, ppd->ccti_entries,
 	       ppd->total_cct_entry * sizeof(struct ib_cc_table_entry));
 

--
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-next 00/11] IB/hfi1: Driver fixes for 11/6/2017

[Doug Ledford]

[-- Attachment #1: Type: text/plain, Size: 1525 bytes --]

On Mon, 2017-11-06 at 06:38 -0800, Dennis Dalessandro wrote:
> Hi Doug,
> 
> More fixes for our drivers. Nothing scary or out of the ordinary here. There
> is a core change, but it is for extended lids. This does include a patch
> suggested by Dan and is based off of his patch and the ensuing discussion on the
> list.
> 
> As always patches available for browsing at:
> https://github.com/ddalessa/kernel/tree/for-4.15
> 
> ---
> 
> Dennis Dalessandro (1):
>       IB/hfi1: Handle initial value of 0 for CCTI setting
> 
> Don Hiatt (2):
>       IB/core: Convert OPA AH to IB for Extended LIDs only
>       IB/hfi1: Mask upper 16Bits of Extended LID prior to rvt_cq_entry
> 
> Grzegorz Morys (1):
>       IB/hfi1: Prohibit invalid Init to Armed state transition
> 
> Jakub Byczkowski (1):
>       IB/hfi1: Reduce 8051 command timeout
> 
> Jan Sokolowski (3):
>       IB/hfi1: Allow MgmtAllowed on B2B setups
>       IB/hfi1: Remove unnecessary if check
>       IB/hfi1: Send 'reboot' as planned down remote reason
> 
> Kamenee Arumugam (1):
>       IB/hfi1: Remove wrapper function in mmu_rb
> 
> Mike Marciniszyn (1):
>       IB/hfi1: Fix a wrapping test to insure the correct timeout
> 
> Niranjana Vishwanathapura (1):
>       IB/hfi1: Do not allocate PIO send contexts for VNIC

Thanks Denny, applied.

-- 
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: 833 bytes --]