linux-kernel - [PATCH v2 2/2] nvmet-rdma: add unbound

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20240717091451.111158-3-jacky_gam_2001@163.com>
Date: Wed, 17 Jul 2024 17:14:51 +0800
From: Ping Gan <jacky_gam_2001@....com>
To: hare@...e.de,
	sagi@...mberg.me,
	hch@....de,
	kch@...dia.com,
	linux-nvme@...ts.infradead.org,
	linux-kernel@...r.kernel.org
Cc: ping.gan@...l.com,
	Ping Gan <jacky_gam_2001@....com>
Subject: [PATCH v2 2/2] nvmet-rdma:  add unbound_wq support for nvmet-rdma

To define a module parameter use_unbound_wq to enable unbound
workqueue to handle RDMA's IO of CQ.

Signed-off-by: Ping Gan <jacky_gam_2001@....com>
Reviewed-by: Hannes Reinecke <hare@...e.de>
---
 drivers/nvme/target/rdma.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c
index 1eff8ca6a5f1..bfd7106316bc 100644
--- a/drivers/nvme/target/rdma.c
+++ b/drivers/nvme/target/rdma.c
@@ -155,6 +155,10 @@ static int nvmet_rdma_srq_size = 1024;
 module_param_cb(srq_size, &srq_size_ops, &nvmet_rdma_srq_size, 0644);
 MODULE_PARM_DESC(srq_size, "set Shared Receive Queue (SRQ) size, should >= 256 (default: 1024)");
 
+static bool use_unbound_wq;
+module_param(use_unbound_wq, bool, 0444);
+MODULE_PARM_DESC(use_unbound_wq, "use unbound workqueue to handle IO request: Default false");
+
 static DEFINE_IDA(nvmet_rdma_queue_ida);
 static LIST_HEAD(nvmet_rdma_queue_list);
 static DEFINE_MUTEX(nvmet_rdma_queue_mutex);
@@ -1259,7 +1263,11 @@ static int nvmet_rdma_create_queue_ib(struct nvmet_rdma_queue *queue)
 	 */
 	nr_cqe = queue->recv_queue_size + 2 * queue->send_queue_size;
 
-	queue->cq = ib_cq_pool_get(ndev->device, nr_cqe + 1,
+	if (use_unbound_wq)
+		queue->cq = ib_cq_pool_get(ndev->device, nr_cqe + 1,
+				   queue->comp_vector, IB_POLL_UNBOUND_WORKQUEUE);
+	else
+		queue->cq = ib_cq_pool_get(ndev->device, nr_cqe + 1,
 				   queue->comp_vector, IB_POLL_WORKQUEUE);
 	if (IS_ERR(queue->cq)) {
 		ret = PTR_ERR(queue->cq);
-- 
2.26.2