Linux RDMA and InfiniBand development
 help / color / mirror / Atom feed
From: Leon Romanovsky <leonro@nvidia.com>
To: "yangx.jy@fujitsu.com" <yangx.jy@fujitsu.com>
Cc: "yanjun.zhu@linux.dev" <yanjun.zhu@linux.dev>,
	"rpearsonhpe@gmail.com" <rpearsonhpe@gmail.com>,
	"linux-rdma@vger.kernel.org" <linux-rdma@vger.kernel.org>,
	"jgg@nvidia.com" <jgg@nvidia.com>
Subject: Re: [PATCH v2] RDMA/rxe: Generate a completion on error after getting a wqe
Date: Thu, 31 Mar 2022 11:51:03 +0300	[thread overview]
Message-ID: <YkVrdzqZaDKPOEpC@unreal> (raw)
In-Reply-To: <e54bd3f3-92be-8574-894b-c5fb7aa59e58@fujitsu.com>

On Thu, Mar 31, 2022 at 05:55:59AM +0000, yangx.jy@fujitsu.com wrote:
> On 2022/3/30 15:44, Leon Romanovsky wrote:
> > You will need to open IBTA spec and try to understand from it.
> > But realistically speaking, I think that it will be too hard to do it
> > and not sure if it is really important.
> >
> > Please resubmit your patch with updated diff and commit message.
> 
> Hi Leon,
> 
> It seems that your change is based on old code.
> 
> After looking into the latest code, this change seem not simpler and 
> clearer. do you think so?

No, something like that will do the trick.

diff --git a/drivers/infiniband/sw/rxe/rxe_av.c b/drivers/infiniband/sw/rxe/rxe_av.c
index 3b05314ca739..5938cba936d9 100644
--- a/drivers/infiniband/sw/rxe/rxe_av.c
+++ b/drivers/infiniband/sw/rxe/rxe_av.c
@@ -104,9 +104,6 @@ struct rxe_av *rxe_get_av(struct rxe_pkt_info *pkt, struct rxe_ah **ahp)
        struct rxe_ah *ah;
        u32 ah_num;

-       if (ahp)
-               *ahp = NULL;
-
        if (!pkt || !pkt->qp)
                return NULL;

diff --git a/drivers/infiniband/sw/rxe/rxe_req.c b/drivers/infiniband/sw/rxe/rxe_req.c
index ae5fbc79dd5c..9afc8cf50863 100644
--- a/drivers/infiniband/sw/rxe/rxe_req.c
+++ b/drivers/infiniband/sw/rxe/rxe_req.c
@@ -648,26 +648,30 @@ int rxe_requester(void *arg)
                psn_compare(qp->req.psn, (qp->comp.psn +
                                RXE_MAX_UNACKED_PSNS)) > 0)) {
                qp->req.wait_psn = 1;
-               goto exit;
+               wqe->status = IB_WC_LOC_QP_OP_ERR;
+               goto err;
        }

        /* Limit the number of inflight SKBs per QP */
        if (unlikely(atomic_read(&qp->skb_out) >
                     RXE_INFLIGHT_SKBS_PER_QP_HIGH)) {
                qp->need_req_skb = 1;
-               goto exit;
+               wqe->status = IB_WC_LOC_QP_OP_ERR;
+               goto err;
        }

        opcode = next_opcode(qp, wqe, wqe->wr.opcode);
        if (unlikely(opcode < 0)) {
                wqe->status = IB_WC_LOC_QP_OP_ERR;
-               goto exit;
+               goto err;
        }
 
        mask = rxe_opcode[opcode].mask;
        if (unlikely(mask & RXE_READ_OR_ATOMIC_MASK)) {
-               if (check_init_depth(qp, wqe))
-                       goto exit;
+               if (check_init_depth(qp, wqe)) {
+                       wqe->status = IB_WC_LOC_QP_OP_ERR;
+                       goto err;
+               }
        }
 
        mtu = get_mtu(qp);
@@ -704,32 +708,27 @@ int rxe_requester(void *arg)
        pkt.wqe = wqe;
 
        av = rxe_get_av(&pkt, &ah);
-       if (unlikely(!av)) {
-               pr_err("qp#%d Failed no address vector\n", qp_num(qp));
-               wqe->status = IB_WC_LOC_QP_OP_ERR;
-               goto err_drop_ah;
-       }
+       if (unlikely(!av))
+               goto qp_err;
 
        skb = init_req_packet(qp, av, wqe, opcode, payload, &pkt);
        if (unlikely(!skb)) {
-               pr_err("qp#%d Failed allocating skb\n", qp_num(qp));
-               wqe->status = IB_WC_LOC_QP_OP_ERR;
-               goto err_drop_ah;
+               rxe_put(ah);
+               goto qp_err;
        }
 
        ret = finish_packet(qp, av, wqe, &pkt, skb, payload);
        if (unlikely(ret)) {
-               pr_debug("qp#%d Error during finish packet\n", qp_num(qp));
                if (ret == -EFAULT)
                        wqe->status = IB_WC_LOC_PROT_ERR;
                else
                        wqe->status = IB_WC_LOC_QP_OP_ERR;
                kfree_skb(skb);
-               goto err_drop_ah;
+               rxe_put(ah);
+               goto err;
        }
 
-       if (ah)
-               rxe_put(ah);
+       rxe_put(ah);
 
        /*
         * To prevent a race on wqe access between requester and completer,
@@ -751,17 +750,15 @@ int rxe_requester(void *arg)
                        goto exit;
                }
 
-               wqe->status = IB_WC_LOC_QP_OP_ERR;
-               goto err;
+               goto qp_err;
        }
 
        update_state(qp, wqe, &pkt);
 
        goto next_wqe;
 
-err_drop_ah:
-       if (ah)
-               rxe_put(ah);
+qp_err:
+       wqe->status = IB_WC_LOC_QP_OP_ERR;
 err:
        wqe->state = wqe_state_error;
        __rxe_do_task(&qp->comp.task);


> 
> Best Regards,
> 
> Xiao Yang
> 
> >
> > Thanks

      reply	other threads:[~2022-03-31  9:05 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-28 15:17 [PATCH v2] RDMA/rxe: Generate a completion on error after getting a wqe Xiao Yang
2022-03-28 18:47 ` Leon Romanovsky
2022-03-29  3:55   ` yangx.jy
2022-03-30  7:44     ` Leon Romanovsky
2022-03-31  5:55       ` yangx.jy
2022-03-31  8:51         ` Leon Romanovsky [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YkVrdzqZaDKPOEpC@unreal \
    --to=leonro@nvidia.com \
    --cc=jgg@nvidia.com \
    --cc=linux-rdma@vger.kernel.org \
    --cc=rpearsonhpe@gmail.com \
    --cc=yangx.jy@fujitsu.com \
    --cc=yanjun.zhu@linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox