RDMA/hns: Modify the print level of CQE error
authorChengchang Tang <tangchengchang@huawei.com>
Fri, 12 Apr 2024 09:16:16 +0000 (17:16 +0800)
committerLeon Romanovsky <leon@kernel.org>
Tue, 16 Apr 2024 12:06:47 +0000 (15:06 +0300)
Too much print may lead to a panic in kernel. Change ibdev_err() to
ibdev_err_ratelimited(), and change the printing level of cqe dump
to debug level.

Fixes: 7c044adca272 ("RDMA/hns: Simplify the cqe code of poll cq")
Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com>
Link: https://lore.kernel.org/r/20240412091616.370789-11-huangjunxian6@hisilicon.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
drivers/infiniband/hw/hns/hns_roce_hw_v2.c

index 5d526b5c4b818b4733ca62f84c70456a76e95778..4287818a737f9780e15efc77bdb522bbed18c571 100644 (file)
@@ -3713,8 +3713,9 @@ static void get_cqe_status(struct hns_roce_dev *hr_dev, struct hns_roce_qp *qp,
                   wc->status == IB_WC_WR_FLUSH_ERR))
                return;
 
-       ibdev_err(&hr_dev->ib_dev, "error cqe status 0x%x:\n", cqe_status);
-       print_hex_dump(KERN_ERR, "", DUMP_PREFIX_NONE, 16, 4, cqe,
+       ibdev_err_ratelimited(&hr_dev->ib_dev, "error cqe status 0x%x:\n",
+                             cqe_status);
+       print_hex_dump(KERN_DEBUG, "", DUMP_PREFIX_NONE, 16, 4, cqe,
                       cq->cqe_size, false);
        wc->vendor_err = hr_reg_read(cqe, CQE_SUB_STATUS);