staging/hfi1: check for ARMED->ACTIVE change in recv int

The link state will transition from ARMED to ACTIVE when a non-SC15
packet arrives, but the driver might not notice the change.  With this
fix, if the slowpath receive interrupt handler sees a non-SC15 packet
while in the ARMED state, we queue work to call linkstate_active_work
from process context to promote it to ACTIVE.

Reviewed-by: Dean Luick <dean.luick@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Jim Snow <jim.m.snow@intel.com>
Signed-off-by: Brendan Cunningham <brendan.cunningham@intel.com>
Signed-off-by: Jubin John <jubin.john@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>

5 files changed, 89 insertions, 2 deletions

diff --git a/drivers/staging/rdma/hfi1/chip.c b/drivers/staging/rdma/hfi1/chip.c
index f7bf90202ad1..63d5d71e8c92 100644
--- a/drivers/staging/rdma/hfi1/chip.c
+++ b/drivers/staging/rdma/hfi1/chip.c
@@ -7878,7 +7878,7 @@ static inline void clear_recv_intr(struct hfi1_ctxtdata *rcd)
 }
 
 /* force the receive interrupt */
-static inline void force_recv_intr(struct hfi1_ctxtdata *rcd)
+void force_recv_intr(struct hfi1_ctxtdata *rcd)
 {
 	write_csr(rcd->dd, CCE_INT_FORCE + (8 * rcd->ireg), rcd->imask);
 }
@@ -7977,7 +7977,7 @@ u32 read_physical_state(struct hfi1_devdata *dd)
 				& DC_DC8051_STS_CUR_STATE_PORT_MASK;
 }
 
-static u32 read_logical_state(struct hfi1_devdata *dd)
+u32 read_logical_state(struct hfi1_devdata *dd)
 {
 	u64 reg;
 
@@ -9952,6 +9952,7 @@ int set_link_state(struct hfi1_pportdata *ppd, u32 state)
 				ppd->link_enabled = 1;
 		}
 
+		set_all_slowpath(ppd->dd);
 		ret = set_local_link_attributes(ppd);
 		if (ret)
 			break;
diff --git a/drivers/staging/rdma/hfi1/chip.h b/drivers/staging/rdma/hfi1/chip.h
index b46ef6675d45..78ba42567f2b 100644
--- a/drivers/staging/rdma/hfi1/chip.h
+++ b/drivers/staging/rdma/hfi1/chip.h
@@ -690,6 +690,8 @@ u64 read_dev_cntr(struct hfi1_devdata *dd, int index, int vl);
 u64 write_dev_cntr(struct hfi1_devdata *dd, int index, int vl, u64 data);
 u64 read_port_cntr(struct hfi1_pportdata *ppd, int index, int vl);
 u64 write_port_cntr(struct hfi1_pportdata *ppd, int index, int vl, u64 data);
+u32 read_logical_state(struct hfi1_devdata *dd);
+void force_recv_intr(struct hfi1_ctxtdata *rcd);
 
 /* Per VL indexes */
 enum {
diff --git a/drivers/staging/rdma/hfi1/driver.c b/drivers/staging/rdma/hfi1/driver.c
index 321852060213..d096f11c0baa 100644
--- a/drivers/staging/rdma/hfi1/driver.c
+++ b/drivers/staging/rdma/hfi1/driver.c
@@ -862,6 +862,37 @@ static inline void set_all_dma_rtail(struct hfi1_devdata *dd)
 			&handle_receive_interrupt_dma_rtail;
 }
 
+void set_all_slowpath(struct hfi1_devdata *dd)
+{
+	int i;
+
+	/* HFI1_CTRL_CTXT must always use the slow path interrupt handler */
+	for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
+		dd->rcd[i]->do_interrupt = &handle_receive_interrupt;
+}
+
+static inline int set_armed_to_active(struct hfi1_ctxtdata *rcd,
+				      struct hfi1_packet packet,
+				      struct hfi1_devdata *dd)
+{
+	struct work_struct *lsaw = &rcd->ppd->linkstate_active_work;
+	struct hfi1_message_header *hdr = hfi1_get_msgheader(packet.rcd->dd,
+							     packet.rhf_addr);
+
+	if (hdr2sc(hdr, packet.rhf) != 0xf) {
+		int hwstate = read_logical_state(dd);
+
+		if (hwstate != LSTATE_ACTIVE) {
+			dd_dev_info(dd, "Unexpected link state %d\n", hwstate);
+			return 0;
+		}
+
+		queue_work(rcd->ppd->hfi1_wq, lsaw);
+		return 1;
+	}
+	return 0;
+}
+
 /*
  * handle_receive_interrupt - receive a packet
  * @rcd: the context
@@ -929,6 +960,11 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread)
 			last = skip_rcv_packet(&packet, thread);
 			skip_pkt = 0;
 		} else {
+			/* Auto activate link on non-SC15 packet receive */
+			if (unlikely(rcd->ppd->host_link_state ==
+				     HLS_UP_ARMED) &&
+			    set_armed_to_active(rcd, packet, dd))
+				goto bail;
 			last = process_rcv_packet(&packet, thread);
 		}
 
@@ -984,6 +1020,42 @@ bail:
 }
 
 /*
+ * We may discover in the interrupt that the hardware link state has
+ * changed from ARMED to ACTIVE (due to the arrival of a non-SC15 packet),
+ * and we need to update the driver's notion of the link state.  We cannot
+ * run set_link_state from interrupt context, so we queue this function on
+ * a workqueue.
+ *
+ * We delay the regular interrupt processing until after the state changes
+ * so that the link will be in the correct state by the time any application
+ * we wake up attempts to send a reply to any message it received.
+ * (Subsequent receive interrupts may possibly force the wakeup before we
+ * update the link state.)
+ *
+ * The rcd is freed in hfi1_free_ctxtdata after hfi1_postinit_cleanup invokes
+ * dd->f_cleanup(dd) to disable the interrupt handler and flush workqueues,
+ * so we're safe from use-after-free of the rcd.
+ */
+void receive_interrupt_work(struct work_struct *work)
+{
+	struct hfi1_pportdata *ppd = container_of(work, struct hfi1_pportdata,
+						  linkstate_active_work);
+	struct hfi1_devdata *dd = ppd->dd;
+	int i;
+
+	/* Received non-SC15 packet implies neighbor_normal */
+	ppd->neighbor_normal = 1;
+	set_link_state(ppd, HLS_UP_ACTIVE);
+
+	/*
+	 * Interrupt all kernel contexts that could have had an
+	 * interrupt during auto activation.
+	 */
+	for (i = HFI1_CTRL_CTXT; i < dd->first_user_ctxt; i++)
+		force_recv_intr(dd->rcd[i]);
+}
+
+/*
  * Convert a given MTU size to the on-wire MAD packet enumeration.
  * Return -1 if the size is invalid.
  */
diff --git a/drivers/staging/rdma/hfi1/hfi.h b/drivers/staging/rdma/hfi1/hfi.h
index 5bc385a574c5..23d7e0249f32 100644
--- a/drivers/staging/rdma/hfi1/hfi.h
+++ b/drivers/staging/rdma/hfi1/hfi.h
@@ -729,6 +729,7 @@ struct hfi1_pportdata {
 	u8 remote_link_down_reason;
 	/* Error events that will cause a port bounce. */
 	u32 port_error_action;
+	struct work_struct linkstate_active_work;
 };
 
 typedef int (*rhf_rcv_function_ptr)(struct hfi1_packet *packet);
@@ -1177,6 +1178,7 @@ void hfi1_free_ctxtdata(struct hfi1_devdata *, struct hfi1_ctxtdata *);
 int handle_receive_interrupt(struct hfi1_ctxtdata *, int);
 int handle_receive_interrupt_nodma_rtail(struct hfi1_ctxtdata *, int);
 int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *, int);
+void set_all_slowpath(struct hfi1_devdata *dd);
 
 /* receive packet handler dispositions */
 #define RCV_PKT_OK      0x0 /* keep going */
@@ -1197,6 +1199,15 @@ static inline u32 driver_lstate(struct hfi1_pportdata *ppd)
 	return ppd->lstate; /* use the cached value */
 }
 
+void receive_interrupt_work(struct work_struct *work);
+
+/* extract service channel from header and rhf */
+static inline int hdr2sc(struct hfi1_message_header *hdr, u64 rhf)
+{
+	return ((be16_to_cpu(hdr->lrh[0]) >> 12) & 0xf) |
+	       ((!!(rhf & RHF_DC_INFO_MASK)) << 4);
+}
+
 static inline u16 generate_jkey(kuid_t uid)
 {
 	return from_kuid(current_user_ns(), uid) & 0xffff;
diff --git a/drivers/staging/rdma/hfi1/init.c b/drivers/staging/rdma/hfi1/init.c
index 48269a2ab756..27b31fc88592 100644
--- a/drivers/staging/rdma/hfi1/init.c
+++ b/drivers/staging/rdma/hfi1/init.c
@@ -498,6 +498,7 @@ void hfi1_init_pportdata(struct pci_dev *pdev, struct hfi1_pportdata *ppd,
 	INIT_WORK(&ppd->link_downgrade_work, handle_link_downgrade);
 	INIT_WORK(&ppd->sma_message_work, handle_sma_message);
 	INIT_WORK(&ppd->link_bounce_work, handle_link_bounce);
+	INIT_WORK(&ppd->linkstate_active_work, receive_interrupt_work);
 	mutex_init(&ppd->hls_lock);
 	spin_lock_init(&ppd->sdma_alllock);
 	spin_lock_init(&ppd->qsfp_info.qsfp_lock);