diff mbox

[for-next,12/21] IB/iser: Use internal polling budget to avoid possible live-lock

Message ID 1412161337-25285-13-git-send-email-ogerlitz@mellanox.com (mailing list archive)
State Accepted, archived
Headers show

Commit Message

Or Gerlitz Oct. 1, 2014, 11:02 a.m. UTC 
From: Sagi Grimberg <sagig@mellanox.com>

We need a way to guarentee that we don't stay in soft-IRQ context
for too long. We might starve other pending CQ tasklets or worse
lock against application trying to issue IO on the running CPU.

Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Ariel Nahum <arieln@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/infiniband/ulp/iser/iser_verbs.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
diff mbox

Patch

diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index 94d1b46..e31ac57 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -42,6 +42,8 @@ 
 #define ISER_MAX_RX_CQ_LEN	(ISER_QP_MAX_RECV_DTOS * ISCSI_ISER_MAX_CONN)
 #define ISER_MAX_TX_CQ_LEN	(ISER_QP_MAX_REQ_DTOS  * ISCSI_ISER_MAX_CONN)
 
+static int iser_cq_poll_limit = 512;
+
 static void iser_cq_tasklet_fn(unsigned long data);
 static void iser_cq_callback(struct ib_cq *cq, void *cq_context);
 static int iser_drain_tx_cq(struct iser_comp *comp);
@@ -1248,6 +1250,8 @@  static void iser_cq_tasklet_fn(unsigned long data)
 		completed_rx++;
 		if (!(completed_rx & 63))
 			completed_tx += iser_drain_tx_cq(comp);
+		if (completed_rx >= iser_cq_poll_limit)
+			break;
 	}
 	/* #warning "it is assumed here that arming CQ only once its empty" *
 	 * " would not cause interrupts to be missed"                       */