public inbox for linux-rdma@vger.kernel.org
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
To: Anna.Schumaker-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: [PATCH v5 11/21] xprtrdma: Reset FRMRs when FAST_REG_MR is flushed by a disconnect
Date: Tue, 29 Jul 2014 17:24:45 -0400	[thread overview]
Message-ID: <20140729212445.2812.69481.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20140729211534.2812.3128.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>

FAST_REG_MR Work Requests update a Memory Region's rkey. Rkey's are
used to block unwanted access to the memory controlled by an MR. The
rkey is passed to the receiver (the NFS server, in our case), and is
also used by xprtrdma to invalidate the MR when the RPC is complete.

When a FAST_REG_MR Work Request is flushed after a transport
disconnect, xprtrdma cannot tell whether the WR actually hit the
adapter or not. So it is indeterminant at that point whether the
existing rkey is still valid.

After the transport connection is re-established, the next
FAST_REG_MR or LOCAL_INV Work Request against that MR can sometimes
fail because the rkey value does not match what xprtrdma expects.

The only reliable way to recover in this case is to deregister and
register the MR before it is used again. These operations can be
done only in a process context, so handle it in the transport
connect worker.

Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Tested-by: Steve Wise <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>
Tested-by: Shirley Ma <shirley.ma-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Tested-by: Devesh Sharma <devesh.sharma-laKkSmNT4hbQT0dZR+AlfA@public.gmane.org>
---
 net/sunrpc/xprtrdma/verbs.c     |   67 ++++++++++++++++++++++++++++++++++++++-
 net/sunrpc/xprtrdma/xprt_rdma.h |    1 +
 2 files changed, 66 insertions(+), 2 deletions(-)

diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index addda7f..4bb1209 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -61,6 +61,8 @@
 # define RPCDBG_FACILITY	RPCDBG_TRANS
 #endif
 
+static void rpcrdma_reset_frmrs(struct rpcrdma_ia *);
+
 /*
  * internal functions
  */
@@ -152,8 +154,10 @@ rpcrdma_sendcq_process_wc(struct ib_wc *wc)
 
 	if (wc->wr_id == 0ULL)
 		return;
-	if (wc->status != IB_WC_SUCCESS)
+	if (wc->status != IB_WC_SUCCESS) {
+		frmr->r.frmr.fr_state = FRMR_IS_STALE;
 		return;
+	}
 
 	if (wc->opcode == IB_WC_FAST_REG_MR)
 		frmr->r.frmr.fr_state = FRMR_IS_VALID;
@@ -881,6 +885,9 @@ retry:
 				" status %i\n", __func__, rc);
 		rpcrdma_flush_cqs(ep);
 
+		if (ia->ri_memreg_strategy == RPCRDMA_FRMR)
+			rpcrdma_reset_frmrs(ia);
+
 		xprt = container_of(ia, struct rpcrdma_xprt, rx_ia);
 		id = rpcrdma_create_id(xprt, ia,
 				(struct sockaddr *)&xprt->rx_data.addr);
@@ -1256,6 +1263,62 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf)
 	kfree(buf->rb_pool);
 }
 
+/* After a disconnect, a flushed FAST_REG_MR can leave an FRMR in
+ * an unusable state. Find FRMRs in this state and dereg / reg
+ * each.  FRMRs that are VALID and attached to an rpcrdma_req are
+ * also torn down.
+ *
+ * This gives all in-use FRMRs a fresh rkey and leaves them INVALID.
+ *
+ * This is invoked only in the transport connect worker in order
+ * to serialize with rpcrdma_register_frmr_external().
+ */
+static void
+rpcrdma_reset_frmrs(struct rpcrdma_ia *ia)
+{
+	struct rpcrdma_xprt *r_xprt =
+				container_of(ia, struct rpcrdma_xprt, rx_ia);
+	struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
+	struct list_head *pos;
+	struct rpcrdma_mw *r;
+	int rc;
+
+	list_for_each(pos, &buf->rb_all) {
+		r = list_entry(pos, struct rpcrdma_mw, mw_all);
+
+		if (r->r.frmr.fr_state == FRMR_IS_INVALID)
+			continue;
+
+		rc = ib_dereg_mr(r->r.frmr.fr_mr);
+		if (rc)
+			dprintk("RPC:       %s: ib_dereg_mr failed %i\n",
+				__func__, rc);
+		ib_free_fast_reg_page_list(r->r.frmr.fr_pgl);
+
+		r->r.frmr.fr_mr = ib_alloc_fast_reg_mr(ia->ri_pd,
+					ia->ri_max_frmr_depth);
+		if (IS_ERR(r->r.frmr.fr_mr)) {
+			rc = PTR_ERR(r->r.frmr.fr_mr);
+			dprintk("RPC:       %s: ib_alloc_fast_reg_mr"
+				" failed %i\n", __func__, rc);
+			continue;
+		}
+		r->r.frmr.fr_pgl = ib_alloc_fast_reg_page_list(
+					ia->ri_id->device,
+					ia->ri_max_frmr_depth);
+		if (IS_ERR(r->r.frmr.fr_pgl)) {
+			rc = PTR_ERR(r->r.frmr.fr_pgl);
+			dprintk("RPC:       %s: "
+				"ib_alloc_fast_reg_page_list "
+				"failed %i\n", __func__, rc);
+
+			ib_dereg_mr(r->r.frmr.fr_mr);
+			continue;
+		}
+		r->r.frmr.fr_state = FRMR_IS_INVALID;
+	}
+}
+
 /* "*mw" can be NULL when rpcrdma_buffer_get_mrs() fails, leaving
  * some req segments uninitialized.
  */
@@ -1572,7 +1635,7 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg,
 	dprintk("RPC:       %s: Using frmr %p to map %d segments\n",
 		__func__, mw, i);
 
-	if (unlikely(frmr->fr_state == FRMR_IS_VALID)) {
+	if (unlikely(frmr->fr_state != FRMR_IS_INVALID)) {
 		dprintk("RPC:       %s: frmr %x left valid, posting invalidate.\n",
 			__func__, mr->rkey);
 		/* Invalidate before using. */
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index c1d8652..1ee6db3 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -161,6 +161,7 @@ struct rpcrdma_rep {
 enum rpcrdma_frmr_state {
 	FRMR_IS_INVALID,	/* ready to be used */
 	FRMR_IS_VALID,		/* in use */
+	FRMR_IS_STALE,		/* failed completion */
 };
 
 struct rpcrdma_frmr {

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

  parent reply	other threads:[~2014-07-29 21:24 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-29 21:23 [PATCH v5 00/21] NFS/RDMA client patches for 3.17 Chuck Lever
     [not found] ` <20140729211534.2812.3128.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
2014-07-29 21:23   ` [PATCH v5 01/21] xprtrdma: Fix panic in rpcrdma_register_frmr_external() Chuck Lever
2014-07-29 21:23   ` [PATCH v5 02/21] xprtrdma: Protect ia->ri_id when unmapping/invalidating MRs Chuck Lever
2014-07-29 21:23   ` [PATCH v5 03/21] xprtrdma: Limit data payload size for ALLPHYSICAL Chuck Lever
2014-07-29 21:23   ` [PATCH v5 04/21] xprtrdma: Update rkeys after transport reconnect Chuck Lever
2014-07-29 21:23   ` [PATCH v5 05/21] xprtrdma: On disconnect, don't ignore pending CQEs Chuck Lever
2014-07-29 21:24   ` [PATCH v5 06/21] xprtrdma: Don't invalidate FRMRs if registration fails Chuck Lever
2014-07-29 21:24   ` [PATCH v5 07/21] xprtrdma: Unclutter struct rpcrdma_mr_seg Chuck Lever
2014-07-29 21:24   ` [PATCH v5 08/21] xprtrdma: Back off rkey when FAST_REG_MR fails Chuck Lever
2014-07-29 21:24   ` [PATCH v5 09/21] xprtrdma: Chain together all MWs in same buffer pool Chuck Lever
2014-07-29 21:24   ` [PATCH v5 10/21] xprtrdma: Properly handle exhaustion of the rb_mws list Chuck Lever
2014-07-29 21:24   ` Chuck Lever [this message]
2014-07-29 21:24   ` [PATCH v5 12/21] xprtrdma: Reset FRMRs after a flushed LOCAL_INV Work Request Chuck Lever
2014-07-29 21:25   ` [PATCH v5 13/21] xprtrdma: Don't post a LOCAL_INV in rpcrdma_register_frmr_external() Chuck Lever
2014-07-29 21:25   ` [PATCH v5 14/21] xprtrdma: Disable completions for FAST_REG_MR Work Requests Chuck Lever
2014-07-29 21:25   ` [PATCH v5 15/21] xprtrdma: Disable completions for LOCAL_INV " Chuck Lever
2014-07-29 21:25   ` [PATCH v5 16/21] xprtrdma: Rename frmr_wr Chuck Lever
2014-07-29 21:25   ` [PATCH v5 17/21] xprtrdma: Allocate each struct rpcrdma_mw separately Chuck Lever
2014-07-29 21:25   ` [PATCH v5 18/21] xprtrdma: Schedule reply tasklet once per upcall Chuck Lever
2014-07-29 21:25   ` [PATCH v5 19/21] xprtrdma: Make rpcrdma_ep_disconnect() return void Chuck Lever
2014-07-29 21:26   ` [PATCH v5 20/21] xprtrdma: Remove RPCRDMA_PERSISTENT_REGISTRATION macro Chuck Lever
2014-07-29 21:26   ` [PATCH v5 21/21] xprtrdma: Handle additional connection events Chuck Lever
2014-07-31 20:44   ` [PATCH v5 00/21] NFS/RDMA client patches for 3.17 Anna Schumaker

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=20140729212445.2812.69481.stgit@manet.1015granger.net \
    --to=chuck.lever-qhclzuegtsvqt0dzr+alfa@public.gmane.org \
    --cc=Anna.Schumaker-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org \
    --cc=linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.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