[PATCH 3.16.y-ckt 027/128] IB/srp: Fix possible send queue overflow

[Luis Henriques]

3.16.7-ckt23 -stable review patch.  If anyone has any objections, please let me know.

---8<------------------------------------------------------------

From: Sagi Grimberg <sagig at mellanox.com>

commit 09c0c0bea500a9ad362589990ee316c9b2482f44 upstream.

When using work request based memory registration (fast_reg)
we must reserve SQ entries for registration and invalidation
in addition to send operations. Each IO consumes 3 SQ entries
(registration, send, invalidation) so we need to allocate 3x
larger send-queue instead of 2x.

Signed-off-by: Sagi Grimberg <sagig at mellanox.com>
Reviewed-by: Christoph Hellwig <hch at lst.de>
Signed-off-by: Doug Ledford <dledford at redhat.com>
[ luis: backported to 3.16:
  - change function srp_create_target_ib() instead of srp_create_ch_ib()
  - adjusted context ]
Signed-off-by: Luis Henriques <luis.henriques at canonical.com>
---
 drivers/infiniband/ulp/srp/ib_srp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index 767000811cf9..36876960f97e 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -451,7 +451,7 @@ static int srp_create_target_ib(struct srp_target_port *target)
 	struct ib_qp *qp;
 	struct ib_fmr_pool *fmr_pool = NULL;
 	struct srp_fr_pool *fr_pool = NULL;
-	const int m = 1 + dev->use_fast_reg;
+	const int m = dev->use_fast_reg ? 3 : 1;
 	int ret;
 
 	init_attr = kzalloc(sizeof *init_attr, GFP_KERNEL);