From mboxrd@z Thu Jan  1 00:00:00 1970
From: minwoo.im.dev@gmail.com (Minwoo Im)
Date: Sat, 1 Jun 2019 11:37:54 +0900
Subject: [RFC PATCH 5/5] nvme-trace: Add tracing for req_comp in target
In-Reply-To: <b2e1e9f4-a96c-9f72-d787-1cd0a6248aa0@grimberg.me>
References: <20190527175346.29972-1-minwoo.im.dev@gmail.com>
 <20190527175346.29972-6-minwoo.im.dev@gmail.com>
 <b2e1e9f4-a96c-9f72-d787-1cd0a6248aa0@grimberg.me>
Message-ID: <20190601023752.GA3349@minwooim-desktop>

On 19-05-31 16:29:06, Sagi Grimberg wrote:
> 
> > @@ -207,20 +208,50 @@ TRACE_EVENT(nvme_complete_rq,
> >   		__field(u16, status)
> >   	    ),
> >   	    TP_fast_assign(
> > -		__entry->ctrl_id = nvme_req(req)->ctrl->instance;
> > -		__entry->qid = nvme_req_qid(req);
> > -		__entry->cid = req->tag;
> > -		__entry->result = le64_to_cpu(nvme_req(req)->result.u64);
> > -		__entry->retries = nvme_req(req)->retries;
> > -		__entry->flags = nvme_req(req)->flags;
> > -		__entry->status = nvme_req(req)->status;
> > -		__assign_disk_name(__entry->disk, req->rq_disk);
> > +		set_trace_type(__entry->type, req);
> > +		if (__entry->type != NVME_TRACE_TARGET) {
> > +			struct request *req = (struct request *) req;
> > +
> > +			__entry->ctrl_id = nvme_req(req)->ctrl->instance;
> > +			__entry->qid = nvme_req_qid(req);
> > +			__entry->cid = req->tag;
> > +			__entry->result =
> > +					le64_to_cpu(nvme_req(req)->result.u64);
> > +			__entry->retries = nvme_req(req)->retries;
> > +			__entry->flags = nvme_req(req)->flags;
> > +			__entry->status = nvme_req(req)->status;
> > +			__assign_disk_name(__entry->disk, req->rq_disk);
> > +		} else {
> > +			struct nvmet_ctrl *ctrl = nvmet_req_to_ctrl(req);
> > +			struct nvmet_cq *cq = ((struct nvmet_req *) req)->cq;
> > +			struct nvme_completion *cqe =
> > +					((struct nvmet_req *) req)->cqe;
> > +			struct nvmet_ns *ns = ((struct nvmet_req *) req)->ns;
> > +
> > +			__entry->ctrl_id = ctrl ? ctrl->cntlid : 0;
> > +			__entry->qid = cq->qid;
> > +			__entry->cid = cqe->command_id;
> > +			__entry->result = cqe->result.u64;
> > +			__entry->flags = 0;
> > +			__entry->status = cqe->status >> 1;
> 
> Why not keep the dnr bit?

Hi Sagi,

Thanks for your review on this.  It's just for phase tag which might
not need for us here.  Thanks to your review, I have realized that it
should be like:

	__entry->status = le16_to_cpu(cqe->status) >> 1;

I will prepare V5 patch series with this.

Thanks,