[PATCH] RDMA/addr: change addr_wq back to unordered workqueue

Surabhi Gogte posted 1 patch 2 months, 1 week ago
drivers/infiniband/core/addr.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
[PATCH] RDMA/addr: change addr_wq back to unordered workqueue
Posted by Surabhi Gogte 2 months, 1 week ago
Commit 5fff41e1f89d ("IB/core: Fix race condition in resolving IP to MAC")
changed the workqueue "addr_wq" to a single-threaded wq.
Commit e19c0d237873 ("RDMA/rdma_cm: Remove process_req and timer sorting")
eliminated global work and started using per-req work.
Now we no longer have the race, change "addr_wq" back to multi-threaded
workqueue to speed up multiple addr resolutions.

Fixes: e19c0d237873 ("RDMA/rdma_cm: Remove process_req and timer sorting")
Signed-off-by: Surabhi Gogte <sgogte@purestorage.com>
---
 drivers/infiniband/core/addr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c
index 866746695712..7ce012f389bd 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -848,7 +848,7 @@ static struct notifier_block nb = {
 
 int addr_init(void)
 {
-	addr_wq = alloc_ordered_workqueue("ib_addr", 0);
+	addr_wq = alloc_workqueue("ib_addr", WQ_UNBOUND, 0);
 	if (!addr_wq)
 		return -ENOMEM;
 
-- 
2.53.0
Re: [PATCH] RDMA/addr: change addr_wq back to unordered workqueue
Posted by Leon Romanovsky 1 month, 1 week ago
On Thu, 09 Apr 2026 18:15:49 -0600, Surabhi Gogte wrote:
> Commit 5fff41e1f89d ("IB/core: Fix race condition in resolving IP to MAC")
> changed the workqueue "addr_wq" to a single-threaded wq.
> Commit e19c0d237873 ("RDMA/rdma_cm: Remove process_req and timer sorting")
> eliminated global work and started using per-req work.
> Now we no longer have the race, change "addr_wq" back to multi-threaded
> workqueue to speed up multiple addr resolutions.
> 
> [...]

Applied, thanks!

[1/1] RDMA/addr: change addr_wq back to unordered workqueue
      https://git.kernel.org/rdma/rdma/c/dd2403fe1499bd

Best regards,
-- 
Leon Romanovsky <leon@kernel.org>