On Nov 11, 2008, at 1:41 AM, FUJITA Tomonori wrote:
On Fri, 31 Oct 2008 08:34:47 -0700
Seokmann Ju <seokmann.ju@xxxxxxxxxx> wrote:
Attachment is also available at the bottom.
---
From 4c6fc04a25afd6dd495975434e9732ede4a1bae1 Mon Sep 17 00:00:00
2001
From: root <root@xxxxxxxxxxxxxxxxxxxxxxx>
Date: Fri, 31 Oct 2008 08:15:56 -0700
Subject: [PATCH] qla2xxx: FC pass through support
This patch will add FC passs through support.
With this patch, the driver support,
- ELS (Extended Link Services)
- FC-CT (Generic Services)
Signed-off-by: Seokmann Ju <seokmann.ju@xxxxxxxxxx>
---
drivers/scsi/qla2xxx/qla_attr.c | 253 ++++++++++++++++++++++++++++
++
+++++++++
drivers/scsi/qla2xxx/qla_def.h | 5 +
drivers/scsi/qla2xxx/qla_fw.h | 35 ++++++
drivers/scsi/qla2xxx/qla_gbl.h | 4 +
drivers/scsi/qla2xxx/qla_gs.c | 2 +-
drivers/scsi/qla2xxx/qla_init.c | 11 ++
drivers/scsi/qla2xxx/qla_iocb.c | 9 +-
drivers/scsi/qla2xxx/qla_isr.c | 68 +++++++++++
8 files changed, 380 insertions(+), 7 deletions(-)
diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/
qla_attr.c
index b223842..945c840 100644
--- a/drivers/scsi/qla2xxx/qla_attr.c
+++ b/drivers/scsi/qla2xxx/qla_attr.c
@@ -1257,6 +1257,253 @@ qla24xx_vport_disable(struct fc_vport
*fc_vport, bool disable)
return 0;
}
+static int
+qla2x00_execute_fc_service(struct fc_service *service)
+{
+ struct fc_rport *rport = service->rport;
+ fc_port_t *fcport = *(fc_port_t **) rport->dd_data;
+ struct Scsi_Host *host = rport_to_shost(rport);
+ scsi_qla_host_t *vha = shost_priv(host);
+ struct qla_hw_data *ha = vha->hw;
+ scsi_qla_host_t *base_vha = pci_get_drvdata(ha->pdev);
+ struct req_que *req = ha->req;
+ struct els_entry_24xx *els_iocb;
+ struct ct_entry_24xx *ct_iocb;
+ request_t *iocb;
+ srb_t *sp;
+ unsigned long flags;
+ uint32_t handle;
+ int index;
+ int req_sg_cnt, rsp_sg_cnt;
+ uint8_t els_pkt[20];
+ size_t copy_size;
+ uint16_t avail_dsds;
+ uint32_t *cur_dsd;
+ struct scatterlist *sg;
+
+ /* At this time, pass throug supported by 4Gb or higher */
+ if (!IS_FWI2_CAPABLE(ha))
+ goto pt_error0;
+
+ if (rport->port_state != FC_PORTSTATE_ONLINE || !fcport)
+ goto pt_error0;
+
+ if ((service->srv_request->request_type != FC_FRAME_TYPE_ELS) &&
+ (service->srv_request->request_type != FC_FRAME_TYPE_FC_CT))
+ goto pt_error0;
+
+ /* ELS doesn't support multiple SG */
+ if ((service->srv_request->request_type == FC_FRAME_TYPE_ELS) &&
+ (service->req_sg_cnt > 1 || service->rsp_sg_cnt > 1)) {
+ printk(KERN_WARNING "ERROR: multiple SG of request/response "
+ "[%u/%u] are not supported for ELS services\n",
+ service->req_sg_cnt, service->rsp_sg_cnt);
+ service->srv_reply.status = FC_SERVICE_UNSUPPORTED;
+ goto pt_error0;
+ }
+
+ sp = mempool_alloc(ha->srb_mempool, GFP_ATOMIC);
+ if (!sp)
+ goto pt_error0;
+
+ sp->vha = vha;
+ sp->fcport = fcport;
+ sp->cmd = (struct scsi_cmnd *)service;
+ sp->flags = SRB_ELS_CT;
+
+ spin_lock_irqsave(&ha->hardware_lock, flags);
+ iocb = qla2x00_req_pkt(base_vha);
+ if (iocb == NULL) {
+ qla_printk(KERN_WARNING, ha,
+ "Passthru request failed to get request packet\n");
+ goto pt_error1;
+ }
+
+ req_sg_cnt = dma_map_sg(&ha->pdev->dev, service->sg_req,
+ service->req_sg_cnt, DMA_TO_DEVICE);
+ if (!req_sg_cnt)
+ goto pt_error1;
+ rsp_sg_cnt = dma_map_sg(&ha->pdev->dev, service->sg_rsp,
+ service->rsp_sg_cnt, DMA_FROM_DEVICE);
+ if (!rsp_sg_cnt)
+ goto pt_error2;
+
+ copy_size = sg_copy_to_buffer(service->sg_req, req_sg_cnt,
+ els_pkt, 20);
You need to use service->sg_rsp instead of req_sg_cnt here.
sg_copy_to_buffer takes the number entry of scatter gather. dma_map_sg
returns the number of physical segments mapped. They could be
different (when IOMMU merges the entries).
Yes, I've got your point. I will make change it.
+ if (!copy_size)
+ goto pt_error3;
+
+ /* Check for room in outstanding command list. */
+ handle = req->current_outstanding_cmd;
+ for (index = 1; index < MAX_OUTSTANDING_COMMANDS; index++) {
+ handle++;
+ if (handle == MAX_OUTSTANDING_COMMANDS)
+ handle = 1;
+ if (!req->outstanding_cmds[handle])
+ break;
+ }
Hmm, needs to check the index overflow here?
I think I've missed your point, here.
Could you please elaborate a bit further?
Thank you,
Seokmann
--
To unsubscribe from this list: 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
