From: Chengchang Tang <tangchengchang@xxxxxxxxxx>
Flush CQE handler has not been called if QP state gets into errored
mode in DWQE path. So, the new added outstanding WQEs will never be
flushed.
It leads to a hung task timeout when using NFS over RDMA:
__switch_to+0x7c/0xd0
__schedule+0x350/0x750
schedule+0x50/0xf0
schedule_timeout+0x2c8/0x340
wait_for_common+0xf4/0x2b0
wait_for_completion+0x20/0x40
__ib_drain_sq+0x140/0x1d0 [ib_core]
ib_drain_sq+0x98/0xb0 [ib_core]
rpcrdma_xprt_disconnect+0x68/0x270 [rpcrdma]
xprt_rdma_close+0x20/0x60 [rpcrdma]
xprt_autoclose+0x64/0x1cc [sunrpc]
process_one_work+0x1d8/0x4e0
worker_thread+0x154/0x420
kthread+0x108/0x150
ret_from_fork+0x10/0x18
Fixes: 01584a5edcc4 ("RDMA/hns: Add support of direct wqe")
Signed-off-by: Chengchang Tang <tangchengchang@xxxxxxxxxx>
Signed-off-by: Junxian Huang <huangjunxian6@xxxxxxxxxxxxx>
---
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index d0469d27c63c..0144e7210d05 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -670,6 +670,10 @@ static void write_dwqe(struct hns_roce_dev *hr_dev, struct hns_roce_qp *qp,
#define HNS_ROCE_SL_SHIFT 2
struct hns_roce_v2_rc_send_wqe *rc_sq_wqe = wqe;
+ if (unlikely(qp->state == IB_QPS_ERR)) {
+ flush_cqe(hr_dev, qp);
+ return;
+ }
/* All kinds of DirectWQE have the same header field layout */
hr_reg_enable(rc_sq_wqe, RC_SEND_WQE_FLAG);
hr_reg_write(rc_sq_wqe, RC_SEND_WQE_DB_SL_L, qp->sl);
--
2.33.0
