Commit a35a3c16 authored by Chengchang Tang's avatar Chengchang Tang Committed by li-huisong
Browse files

RDMA/hns: Fix missing flush CQE for DWQE 

maillist inclusion
category: bugfix
bugzilla: https://gitee.com/openeuler/kernel/issues/IBDW6Z
CVE: NA

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git/commit/?id=e3debdd48423d3d75b9d366399228d7225d902cd



----------------------------------------------------------------------

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: 01584a5e ("RDMA/hns: Add support of direct wqe")
Signed-off-by: default avatarChengchang Tang <tangchengchang@huawei.com>
Signed-off-by: default avatarJunxian Huang <huangjunxian6@hisilicon.com>
Link: https://patch.msgid.link/20241220055249.146943-5-huangjunxian6@hisilicon.com


Signed-off-by: default avatarLeon Romanovsky <leon@kernel.org>
Signed-off-by: default avatarXinghai Cen <cenxinghai@h-partners.com>
parent d3729802

drivers/infiniband/hw/hns/hns_roce_hw_v2.c

+4 −0
Original line number Diff line number Diff line
@@ -771,6 +771,10 @@ 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);