Sorry about the fix. This is a false positive which should be ignored.
The work will be posted on ha->dpc_thread. In qla4xxx_remove_adapter,
it will call destroy_workqueue(ha->dpc_thread);
Regards,
Zheng
Zheng Wang <zyytlz.wz@xxxxxxx> 于2023年3月10日周五 02:38写道：
>
> In qla4xxx_probe_adapter, &ha->dpc_work is bound with
> qla4xxx_do_dpc. qla4xxx_post_aen_work may be called
> to start the work.
>
> If we remove the module which will call qla4xxx_remove_adapter
> to make cleanup, there may be a unfinished work. The possible
> sequence is as follows:
>
> Fix it by canceling the work before cleanup in qla4xxx_remove_adapter
>
> CPU0 CPUc1
>
> |qla4xxx_do_dpc
> qla4xxx_remove_adapter|
> scsi_remove_host |
> kfree(ha->host) |
> |
> | iscsi_host_for_each_session
> | //use ha->host
> Fixes: afaf5a2d341d ("[SCSI] Initial Commit of qla4xxx")
> Signed-off-by: Zheng Wang <zyytlz.wz@xxxxxxx>
> ---
> drivers/scsi/qla4xxx/ql4_os.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/scsi/qla4xxx/ql4_os.c b/drivers/scsi/qla4xxx/ql4_os.c
> index 005502125b27..d530cc853f43 100644
> --- a/drivers/scsi/qla4xxx/ql4_os.c
> +++ b/drivers/scsi/qla4xxx/ql4_os.c
> @@ -9001,6 +9001,7 @@ static void qla4xxx_remove_adapter(struct pci_dev *pdev)
> if (!pci_is_enabled(pdev))
> return;
>
> + cancel_work_sync(&ha->dpc_work);
> ha = pci_get_drvdata(pdev);
>
> if (is_qla40XX(ha))
> --
> 2.25.1
>
