linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: bfields@fieldses.org
Cc: linux-nfs@vger.kernel.org, linux-rdma@vger.kernel.org
Subject: [PATCH v2 06/29] svcrdma: Fix backchannel return code
Date: Tue, 12 May 2020 17:22:36 -0400	[thread overview]
Message-ID: <20200512212236.5826.94193.stgit@klimt.1015granger.net> (raw)
In-Reply-To: <20200512211640.5826.77139.stgit@klimt.1015granger.net>

Way back when I was writing the RPC/RDMA server-side backchannel
code, I misread the TCP backchannel reply handler logic. When
svc_tcp_recvfrom() successfully receives a backchannel reply, it
does not return -EAGAIN. It sets XPT_DATA and returns zero.

Update svc_rdma_recvfrom() to return zero. Here, XPT_DATA doesn't
need to be set again: it is set whenever a new message is received,
behind a spin lock in a single threaded context.

Also, if handling the cb reply is not successful, the message is
simply dropped. There's no special message framing to deal with as
there is in the TCP case.

Now that the handle_bc_reply() return value is ignored, I've removed
the dprintk call sites in the error exit of handle_bc_reply() in
favor of trace points in other areas that already report the error
cases.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 include/linux/sunrpc/svc_rdma.h            |    5 +---
 net/sunrpc/xprtrdma/svc_rdma_backchannel.c |   38 +++++++---------------------
 net/sunrpc/xprtrdma/svc_rdma_recvfrom.c    |   11 ++++----
 3 files changed, 17 insertions(+), 37 deletions(-)

diff --git a/include/linux/sunrpc/svc_rdma.h b/include/linux/sunrpc/svc_rdma.h
index cbcfbd0521e3..8518c3f37e56 100644
--- a/include/linux/sunrpc/svc_rdma.h
+++ b/include/linux/sunrpc/svc_rdma.h
@@ -160,9 +160,8 @@ struct svc_rdma_send_ctxt {
 };
 
 /* svc_rdma_backchannel.c */
-extern int svc_rdma_handle_bc_reply(struct rpc_xprt *xprt,
-				    __be32 *rdma_resp,
-				    struct xdr_buf *rcvbuf);
+extern void svc_rdma_handle_bc_reply(struct svc_rqst *rqstp,
+				     struct svc_rdma_recv_ctxt *rctxt);
 
 /* svc_rdma_recvfrom.c */
 extern void svc_rdma_recv_ctxts_destroy(struct svcxprt_rdma *rdma);
diff --git a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
index d9aab4504f2c..b174f3b109a5 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
@@ -15,26 +15,25 @@
 #undef SVCRDMA_BACKCHANNEL_DEBUG
 
 /**
- * svc_rdma_handle_bc_reply - Process incoming backchannel reply
- * @xprt: controlling backchannel transport
- * @rdma_resp: pointer to incoming transport header
- * @rcvbuf: XDR buffer into which to decode the reply
+ * svc_rdma_handle_bc_reply - Process incoming backchannel Reply
+ * @rqstp: resources for handling the Reply
+ * @rctxt: Received message
  *
- * Returns:
- *	%0 if @rcvbuf is filled in, xprt_complete_rqst called,
- *	%-EAGAIN if server should call ->recvfrom again.
  */
-int svc_rdma_handle_bc_reply(struct rpc_xprt *xprt, __be32 *rdma_resp,
-			     struct xdr_buf *rcvbuf)
+void svc_rdma_handle_bc_reply(struct svc_rqst *rqstp,
+			      struct svc_rdma_recv_ctxt *rctxt)
 {
+	struct svc_xprt *sxprt = rqstp->rq_xprt;
+	struct rpc_xprt *xprt = sxprt->xpt_bc_xprt;
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
+	struct xdr_buf *rcvbuf = &rqstp->rq_arg;
 	struct kvec *dst, *src = &rcvbuf->head[0];
+	__be32 *rdma_resp = rctxt->rc_recv_buf;
 	struct rpc_rqst *req;
 	u32 credits;
 	size_t len;
 	__be32 xid;
 	__be32 *p;
-	int ret;
 
 	p = (__be32 *)src->iov_base;
 	len = src->iov_len;
@@ -49,14 +48,10 @@ int svc_rdma_handle_bc_reply(struct rpc_xprt *xprt, __be32 *rdma_resp,
 		__func__, (int)len, p);
 #endif
 
-	ret = -EAGAIN;
-	if (src->iov_len < 24)
-		goto out_shortreply;
-
 	spin_lock(&xprt->queue_lock);
 	req = xprt_lookup_rqst(xprt, xid);
 	if (!req)
-		goto out_notfound;
+		goto out_unlock;
 
 	dst = &req->rq_private_buf.head[0];
 	memcpy(&req->rq_private_buf, &req->rq_rcv_buf, sizeof(struct xdr_buf));
@@ -77,25 +72,12 @@ int svc_rdma_handle_bc_reply(struct rpc_xprt *xprt, __be32 *rdma_resp,
 	spin_unlock(&xprt->transport_lock);
 
 	spin_lock(&xprt->queue_lock);
-	ret = 0;
 	xprt_complete_rqst(req->rq_task, rcvbuf->len);
 	xprt_unpin_rqst(req);
 	rcvbuf->len = 0;
 
 out_unlock:
 	spin_unlock(&xprt->queue_lock);
-out:
-	return ret;
-
-out_shortreply:
-	dprintk("svcrdma: short bc reply: xprt=%p, len=%zu\n",
-		xprt, src->iov_len);
-	goto out;
-
-out_notfound:
-	dprintk("svcrdma: unrecognized bc reply: xprt=%p, xid=%08x\n",
-		xprt, be32_to_cpu(xid));
-	goto out_unlock;
 }
 
 /* Send a backwards direction RPC call.
diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
index efa5fcb5793f..eee7c6478b30 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
@@ -878,12 +878,9 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp)
 		goto out_drop;
 	rqstp->rq_xprt_hlen = ret;
 
-	if (svc_rdma_is_backchannel_reply(xprt, p)) {
-		ret = svc_rdma_handle_bc_reply(xprt->xpt_bc_xprt, p,
-					       &rqstp->rq_arg);
-		svc_rdma_recv_ctxt_put(rdma_xprt, ctxt);
-		return ret;
-	}
+	if (svc_rdma_is_backchannel_reply(xprt, p))
+		goto out_backchannel;
+
 	svc_rdma_get_inv_rkey(rdma_xprt, ctxt);
 
 	p += rpcrdma_fixed_maxsz;
@@ -913,6 +910,8 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp)
 	svc_rdma_recv_ctxt_put(rdma_xprt, ctxt);
 	return ret;
 
+out_backchannel:
+	svc_rdma_handle_bc_reply(rqstp, ctxt);
 out_drop:
 	svc_rdma_recv_ctxt_put(rdma_xprt, ctxt);
 	return 0;


  parent reply	other threads:[~2020-05-12 21:22 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-12 21:22 [PATCH v2 00/29] Possible NFSD patches for v5.8 Chuck Lever
2020-05-12 21:22 ` [PATCH v2 01/29] SUNRPC: Move xpt_mutex into socket xpo_sendto methods Chuck Lever
2020-05-12 21:22 ` [PATCH v2 02/29] svcrdma: Clean up the tracing for rw_ctx_init errors Chuck Lever
2020-05-12 21:22 ` [PATCH v2 03/29] svcrdma: Clean up handling of get_rw_ctx errors Chuck Lever
2020-05-12 21:22 ` [PATCH v2 04/29] svcrdma: Trace page overruns when constructing RDMA Reads Chuck Lever
2020-05-12 21:22 ` [PATCH v2 05/29] svcrdma: trace undersized Write chunks Chuck Lever
2020-05-12 21:22 ` Chuck Lever [this message]
2020-05-12 21:22 ` [PATCH v2 07/29] svcrdma: Remove backchannel dprintk call sites Chuck Lever
2020-05-12 21:22 ` [PATCH v2 08/29] svcrdma: Rename tracepoints that record header decoding errors Chuck Lever
2020-05-12 21:22 ` [PATCH v2 09/29] svcrdma: Remove the SVCRDMA_DEBUG macro Chuck Lever
2020-05-12 21:22 ` [PATCH v2 10/29] svcrdma: Displayed remote IP address should match stored address Chuck Lever
2020-05-12 21:23 ` [PATCH v2 11/29] svcrdma: Add tracepoints to report ->xpo_accept failures Chuck Lever
2020-05-12 21:23 ` [PATCH v2 12/29] SUNRPC: Remove kernel memory address from svc_xprt tracepoints Chuck Lever
2020-05-12 21:23 ` [PATCH v2 13/29] SUNRPC: Tracepoint to record errors in svc_xpo_create() Chuck Lever
2020-05-12 21:23 ` [PATCH v2 14/29] SUNRPC: Trace a few more generic svc_xprt events Chuck Lever
2020-05-12 21:23 ` [PATCH v2 15/29] SUNRPC: Remove "#include <trace/events/skb.h>" Chuck Lever
2020-05-12 21:23 ` [PATCH v2 16/29] SUNRPC: Add more svcsock tracepoints Chuck Lever
2020-05-12 21:23 ` [PATCH v2 17/29] SUNRPC: Replace dprintk call sites in TCP state change callouts Chuck Lever
2020-05-12 21:23 ` [PATCH v2 18/29] SUNRPC: Trace server-side rpcbind registration events Chuck Lever
2020-05-12 21:23 ` [PATCH v2 19/29] SUNRPC: Rename svc_sock::sk_reclen Chuck Lever
2020-05-12 21:23 ` [PATCH v2 20/29] SUNRPC: Restructure svc_tcp_recv_record() Chuck Lever
2020-05-12 21:23 ` [PATCH v2 21/29] SUNRPC: Refactor svc_recvfrom() Chuck Lever
2020-05-12 21:24 ` [PATCH v2 22/29] SUNRPC: Restructure svc_udp_recvfrom() Chuck Lever
2020-05-12 21:24 ` [PATCH v2 23/29] SUNRPC: svc_show_status() macro should have enum definitions Chuck Lever
2020-05-12 21:24 ` [PATCH v2 24/29] NFSD: Add tracepoints to NFSD's duplicate reply cache Chuck Lever
2020-05-12 21:24 ` [PATCH v2 25/29] NFSD: Add tracepoints to the NFSD state management code Chuck Lever
2020-05-12 21:24 ` [PATCH v2 26/29] NFSD: Add tracepoints for monitoring NFSD callbacks Chuck Lever
2020-05-12 21:24 ` [PATCH v2 27/29] SUNRPC: Clean up request deferral tracepoints Chuck Lever
2020-05-12 21:24 ` [PATCH v2 28/29] NFSD: Squash an annoying compiler warning Chuck Lever
2020-05-12 21:24 ` [PATCH v2 29/29] NFSD: Fix improperly-formatted Doxygen comments Chuck Lever
2020-05-19 16:11 ` [PATCH v2 00/29] Possible NFSD patches for v5.8 J. Bruce Fields
2020-05-19 16:14   ` Chuck Lever
2020-05-19 21:29     ` Bruce Fields
2020-05-19 22:25       ` Chuck Lever
2020-05-19 23:32         ` Chuck Lever
2020-05-20 16:46           ` Bruce Fields
2020-05-20 16:48             ` Chuck Lever
2020-05-20 17:01               ` Bruce Fields
2020-05-20 17:01             ` Calum Mackay
2020-05-20 17:14               ` Bruce Fields

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=20200512212236.5826.94193.stgit@klimt.1015granger.net \
    --to=chuck.lever@oracle.com \
    --cc=bfields@fieldses.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).