IB/core: Add an unbound WQ type to the new CQ API

The upstream kernel commit cited below modified the workqueue in the
new CQ API to be bound to a specific CPU (instead of being unbound).
This caused ALL users of the new CQ API to use the same bound WQ.

Specifically, MAD handling was severely delayed when the CPU bound
to the WQ was busy handling (higher priority) interrupts.

This caused a delay in the MAD "heartbeat" response handling,
which resulted in ports being incorrectly classified as "down".

To fix this, add a new "unbound" WQ type to the new CQ API, so that users
have the option to choose either a bound WQ or an unbound WQ.

For MADs, choose the new "unbound" WQ.

Fixes: b7363e67b23e ("IB/device: Convert ib-comp-wq to be CPU-bound")
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.m>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>

1 files changed, 14 insertions, 1 deletions

diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index db3b6271f09d..6d8ac51a39cc 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -61,6 +61,7 @@ struct ib_client_data {
 };
 
 struct workqueue_struct *ib_comp_wq;
+struct workqueue_struct *ib_comp_unbound_wq;
 struct workqueue_struct *ib_wq;
 EXPORT_SYMBOL_GPL(ib_wq);
 
@@ -1166,10 +1167,19 @@ static int __init ib_core_init(void)
 		goto err;
 	}
 
+	ib_comp_unbound_wq =
+		alloc_workqueue("ib-comp-unb-wq",
+				WQ_UNBOUND | WQ_HIGHPRI | WQ_MEM_RECLAIM |
+				WQ_SYSFS, WQ_UNBOUND_MAX_ACTIVE);
+	if (!ib_comp_unbound_wq) {
+		ret = -ENOMEM;
+		goto err_comp;
+	}
+
 	ret = class_register(&ib_class);
 	if (ret) {
 		pr_warn("Couldn't create InfiniBand device class\n");
-		goto err_comp;
+		goto err_comp_unbound;
 	}
 
 	ret = rdma_nl_init();
@@ -1218,6 +1228,8 @@ err_ibnl:
 	rdma_nl_exit();
 err_sysfs:
 	class_unregister(&ib_class);
+err_comp_unbound:
+	destroy_workqueue(ib_comp_unbound_wq);
 err_comp:
 	destroy_workqueue(ib_comp_wq);
 err:
@@ -1236,6 +1248,7 @@ static void __exit ib_core_cleanup(void)
 	addr_cleanup();
 	rdma_nl_exit();
 	class_unregister(&ib_class);
+	destroy_workqueue(ib_comp_unbound_wq);
 	destroy_workqueue(ib_comp_wq);
 	/* Make sure that any pending umem accounting work is done. */
 	destroy_workqueue(ib_wq);