Regarding to discussion on LKML: http://marc.theaimsgroup.com/?t=112865351000001&r=1&w=2 It turns out that qla2x00 driver unnecessarily activated generic blk tag queue without actually use any of the tag maintained by the block layer. I suggest we turn that off until there is a real consumer of the tag queuing in the qla driver. Please apply. Signed-off-by: Ken Chen <kenneth.w.chen@xxxxxxxxx> --- ./drivers/scsi/qla2xxx/qla_os.c.orig 2005-10-06 23:24:32.547238409 -0700 +++ ./drivers/scsi/qla2xxx/qla_os.c 2005-10-06 23:30:44.004265109 -0700 @@ -1101,11 +1101,7 @@ qla2xxx_slave_configure(struct scsi_devi scsi_qla_host_t *ha = to_qla_host(sdev->host); struct fc_rport *rport = starget_to_rport(sdev->sdev_target); - if (sdev->tagged_supported) - scsi_activate_tcq(sdev, 32); - else - scsi_deactivate_tcq(sdev, 32); - + scsi_adjust_queue_depth(sdev, scsi_get_tag_type(sdev), 32); rport->dev_loss_tmo = ha->port_down_retry_count + 5; return 0; @@ -1127,13 +1123,9 @@ qla2x00_change_queue_depth(struct scsi_d static int qla2x00_change_queue_type(struct scsi_device *sdev, int tag_type) { - if (sdev->tagged_supported) { + if (sdev->tagged_supported) scsi_set_tag_type(sdev, tag_type); - if (tag_type) - scsi_activate_tcq(sdev, sdev->queue_depth); - else - scsi_deactivate_tcq(sdev, sdev->queue_depth); - } else + else tag_type = 0; return tag_type; - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
