On 08/17/2009 10:00 AM, Hannes Reinecke wrote: > However, I still have the feeling there still is potential > for an I/O stall; this iscsi_check_cmdsn_window_closed() check > simply doesn't feel safe. >
Do you still have that test in iscsi_data_xmit or iscsi_xmit_task? If so there are the other problems I mentioned in the patch you modified iscsi_update_cmdsn's iscsi_conn_queue_work call. If you just mean the iscsi_check_cmdsn_window_closed call in iscsi_queuecommand, then I have been hitting problems with fcoe.ko when it returns SCSI_MLQUEUE_HOST_BUSY. The throughput drops from 1XX MB/s to a couple MB/s. I thought there might be bug in the SCSI_MLQUEUE_HOST_BUSY handling code. If you know the window size of your target you can remove the iscsi_check_cmdsn_window_closed call and set can_queue to it. Or you can also just set the can_queue to 1 command and see if you still hit the same problem. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "open-iscsi" group. To post to this group, send email to open-iscsi@googlegroups.com To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/open-iscsi -~----------~----~----~----~------~----~------~--~---