All of lore.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 v4 06/21] xprtrdma: Don't invalidate FRMRs if registration fails
Date: Tue, 22 Jul 2014 10:51:51 -0400	[thread overview]
Message-ID: <20140722145151.6010.49734.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20140722144459.6010.99389.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>

If FRMR registration fails, it's likely to transition the QP to the
error state. Or, registration may have failed because the QP is
_already_ in ERROR.

Thus calling rpcrdma_deregister_external() in
rpcrdma_create_chunks() is useless in FRMR mode: the LOCAL_INVs just
get flushed.

It is safe to leave existing registrations: when FRMR registration
is tried again, rpcrdma_register_frmr_external() checks if each FRMR
is already/still VALID, and knocks it down first if it is.

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/rpc_rdma.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 54422f7..6166c98 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -271,9 +271,11 @@ rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target,
 	return (unsigned char *)iptr - (unsigned char *)headerp;
 
 out:
-	for (pos = 0; nchunks--;)
-		pos += rpcrdma_deregister_external(
-				&req->rl_segments[pos], r_xprt);
+	if (r_xprt->rx_ia.ri_memreg_strategy != RPCRDMA_FRMR) {
+		for (pos = 0; nchunks--;)
+			pos += rpcrdma_deregister_external(
+					&req->rl_segments[pos], r_xprt);
+	}
 	return n;
 }
 

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

WARNING: multiple messages have this Message-ID (diff)
From: Chuck Lever <chuck.lever@oracle.com>
To: Anna.Schumaker@netapp.com
Cc: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org
Subject: [PATCH v4 06/21] xprtrdma: Don't invalidate FRMRs if registration fails
Date: Tue, 22 Jul 2014 10:51:51 -0400	[thread overview]
Message-ID: <20140722145151.6010.49734.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20140722144459.6010.99389.stgit@manet.1015granger.net>

If FRMR registration fails, it's likely to transition the QP to the
error state. Or, registration may have failed because the QP is
_already_ in ERROR.

Thus calling rpcrdma_deregister_external() in
rpcrdma_create_chunks() is useless in FRMR mode: the LOCAL_INVs just
get flushed.

It is safe to leave existing registrations: when FRMR registration
is tried again, rpcrdma_register_frmr_external() checks if each FRMR
is already/still VALID, and knocks it down first if it is.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
Tested-by: Shirley Ma <shirley.ma@oracle.com>
Tested-by: Devesh Sharma <devesh.sharma@emulex.com>
---
 net/sunrpc/xprtrdma/rpc_rdma.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 54422f7..6166c98 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -271,9 +271,11 @@ rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target,
 	return (unsigned char *)iptr - (unsigned char *)headerp;
 
 out:
-	for (pos = 0; nchunks--;)
-		pos += rpcrdma_deregister_external(
-				&req->rl_segments[pos], r_xprt);
+	if (r_xprt->rx_ia.ri_memreg_strategy != RPCRDMA_FRMR) {
+		for (pos = 0; nchunks--;)
+			pos += rpcrdma_deregister_external(
+					&req->rl_segments[pos], r_xprt);
+	}
 	return n;
 }
 


  parent reply	other threads:[~2014-07-22 14:51 UTC|newest]

Thread overview: 64+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-22 14:50 [PATCH v4 00/21] NFS/RDMA client patches for 3.17 Chuck Lever
2014-07-22 14:50 ` Chuck Lever
     [not found] ` <20140722144459.6010.99389.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
2014-07-22 14:51   ` [PATCH v4 01/21] xprtrdma: Fix panic in rpcrdma_register_frmr_external() Chuck Lever
2014-07-22 14:51     ` Chuck Lever
2014-07-22 14:51   ` [PATCH v4 02/21] xprtrdma: Protect ia->ri_id when unmapping/invalidating MRs Chuck Lever
2014-07-22 14:51     ` Chuck Lever
2014-07-22 14:51   ` [PATCH v4 03/21] xprtrdma: Limit data payload size for ALLPHYSICAL Chuck Lever
2014-07-22 14:51     ` Chuck Lever
2014-07-22 14:51   ` [PATCH v4 04/21] xprtrdma: Update rkeys after transport reconnect Chuck Lever
2014-07-22 14:51     ` Chuck Lever
     [not found]     ` <20140722145133.6010.52757.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
2014-07-22 18:40       ` Anna Schumaker
2014-07-22 18:40         ` Anna Schumaker
     [not found]         ` <53CEB038.5090901-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org>
2014-07-24 20:03           ` Chuck Lever
2014-07-24 20:03             ` Chuck Lever
2014-07-22 14:51   ` [PATCH v4 05/21] xprtrdma: On disconnect, don't ignore pending CQEs Chuck Lever
2014-07-22 14:51     ` Chuck Lever
2014-07-22 14:51   ` Chuck Lever [this message]
2014-07-22 14:51     ` [PATCH v4 06/21] xprtrdma: Don't invalidate FRMRs if registration fails Chuck Lever
2014-07-22 14:51   ` [PATCH v4 07/21] xprtrdma: Unclutter struct rpcrdma_mr_seg Chuck Lever
2014-07-22 14:51     ` Chuck Lever
2014-07-22 14:52   ` [PATCH v4 08/21] xprtrdma: Back off rkey when FAST_REG_MR fails Chuck Lever
2014-07-22 14:52     ` Chuck Lever
     [not found]     ` <20140722145208.6010.76678.stgit-FYjufvaPoItvLzlybtyyYzGyq/o6K9yX@public.gmane.org>
2014-07-22 20:03       ` Anna Schumaker
2014-07-22 20:03         ` Anna Schumaker
     [not found]         ` <53CEC375.3010408-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org>
2014-07-22 21:18           ` Chuck Lever
2014-07-22 21:18             ` Chuck Lever
     [not found]             ` <A86981DE-2BD5-4763-81E7-C4A675245AA6-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
2014-07-23 13:44               ` Anna Schumaker
2014-07-23 13:44                 ` Anna Schumaker
2014-07-24 13:24       ` Anna Schumaker
2014-07-24 13:24         ` Anna Schumaker
2014-07-30 19:11       ` Anna Schumaker
2014-07-30 19:11         ` Anna Schumaker
2014-07-22 14:52   ` [PATCH v4 09/21] xprtrdma: Chain together all MWs in same buffer pool Chuck Lever
2014-07-22 14:52     ` Chuck Lever
2014-07-22 14:52   ` [PATCH v4 10/21] xprtrdma: Properly handle exhaustion of the rb_mws list Chuck Lever
2014-07-22 14:52     ` Chuck Lever
2014-07-22 14:52   ` [PATCH v4 11/21] xprtrdma: Reset FRMRs when FAST_REG_MR is flushed by a disconnect Chuck Lever
2014-07-22 14:52     ` Chuck Lever
2014-07-22 14:52   ` [PATCH v4 12/21] xprtrdma: Reset FRMRs after a flushed LOCAL_INV Work Request Chuck Lever
2014-07-22 14:52     ` Chuck Lever
2014-07-22 14:52   ` [PATCH v4 13/21] xprtrdma: Don't post a LOCAL_INV in rpcrdma_register_frmr_external() Chuck Lever
2014-07-22 14:52     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 14/21] xprtrdma: Disable completions for FAST_REG_MR Work Requests Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 15/21] xprtrdma: Disable completions for LOCAL_INV " Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 16/21] xprtrdma: Rename frmr_wr Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 17/21] xprtrdma: Allocate each struct rpcrdma_mw separately Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 18/21] xprtrdma: Schedule reply tasklet once per upcall Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 19/21] xprtrdma: Make rpcrdma_ep_disconnect() return void Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:53   ` [PATCH v4 20/21] xprtrdma: Remove RPCRDMA_PERSISTENT_REGISTRATION macro Chuck Lever
2014-07-22 14:53     ` Chuck Lever
2014-07-22 14:54   ` [PATCH v4 21/21] xprtrdma: Handle additional connection events Chuck Lever
2014-07-22 14:54     ` Chuck Lever
2014-07-23 18:53   ` [PATCH v4 00/21] NFS/RDMA client patches for 3.17 Anna Schumaker
2014-07-23 18:53     ` Anna Schumaker
     [not found]     ` <53D004A0.7030803-ZwjVKphTwtPQT0dZR+AlfA@public.gmane.org>
2014-07-23 21:23       ` Chuck Lever
2014-07-23 21:23         ` Chuck Lever
     [not found]         ` <37564E5B-5955-436B-88D2-888ADC3D7A00-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
2014-07-24 13:19           ` Anna Schumaker
2014-07-24 13:19             ` 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=20140722145151.6010.49734.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.