Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: [PATCH v4 8/8] NFSD: Make nfsd4_remove() wait before returning
 NFS4ERR_DELAY
From:   Chuck Lever <chuck.lever@oracle.com>
To:     linux-nfs@vger.kernel.org
Date:   Thu, 08 Sep 2022 18:14:25 -0400
Message-ID: <166267526571.1842.11502743715868609327.stgit@manet.1015granger.net>
In-Reply-To: <166267495153.1842.14474564029477470642.stgit@manet.1015granger.net>
References: <166267495153.1842.14474564029477470642.stgit@manet.1015granger.net>
User-Agent: StGit/1.5.dev2+g9ce680a5
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
Precedence: bulk

nfsd_unlink() can kick off a CB_RECALL (via
vfs_unlink() -> leases_conflict()) if a delegation is present.
Before returning NFS4ERR_DELAY, give the client holding that
delegation a chance to return it and then retry the nfsd_unlink()
again, once.

Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=354
Tested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 fs/nfsd/vfs.c |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index b597cb2af949..adb2624eca64 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -1773,9 +1773,18 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 
 	fh_fill_pre_attrs(fhp);
 	if (type != S_IFDIR) {
+		int retries;
+
 		if (rdentry->d_sb->s_export_op->flags & EXPORT_OP_CLOSE_BEFORE_UNLINK)
 			nfsd_close_cached_files(rdentry);
-		host_err = vfs_unlink(&init_user_ns, dirp, rdentry, NULL);
+
+		for (retries = 1;;) {
+			host_err = vfs_unlink(&init_user_ns, dirp, rdentry, NULL);
+			if (host_err != -EAGAIN || !retries--)
+				break;
+			if (!nfsd_wait_for_delegreturn(rqstp, rinode))
+				break;
+		}
 	} else {
 		host_err = vfs_rmdir(&init_user_ns, dirp, rdentry);
 	}