linux-kernel - [PATCH AUTOSEL 4.19 217/671] NFS/pnfs: Bulk destroy of layouts needs to be safe w.r.t. umount

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20200116165940.10720-100-sashal@kernel.org>
Date:   Thu, 16 Jan 2020 11:52:06 -0500
From:   Sasha Levin <sashal@...nel.org>
To:     linux-kernel@...r.kernel.org, stable@...r.kernel.org
Cc:     Trond Myklebust <trond.myklebust@...merspace.com>,
        Sasha Levin <sashal@...nel.org>, linux-nfs@...r.kernel.org
Subject: [PATCH AUTOSEL 4.19 217/671] NFS/pnfs: Bulk destroy of layouts needs to be safe w.r.t. umount

From: Trond Myklebust <trond.myklebust@...merspace.com>

[ Upstream commit 5085607d209102b37b169bc94d0aa39566a9842a ]

If a bulk layout recall or a metadata server reboot coincides with a
umount, then holding a reference to an inode is unsafe unless we
also hold a reference to the super block.

Fixes: fd9a8d7160937 ("NFSv4.1: Fix bulk recall and destroy of layouts")
Signed-off-by: Trond Myklebust <trond.myklebust@...merspace.com>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 fs/nfs/pnfs.c | 33 +++++++++++++++++++++++----------
 fs/nfs/pnfs.h |  1 +
 2 files changed, 24 insertions(+), 10 deletions(-)

diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index c818f9886f61..66f699e18755 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -758,22 +758,35 @@ static int
 pnfs_layout_bulk_destroy_byserver_locked(struct nfs_client *clp,
 		struct nfs_server *server,
 		struct list_head *layout_list)
+	__must_hold(&clp->cl_lock)
+	__must_hold(RCU)
 {
 	struct pnfs_layout_hdr *lo, *next;
 	struct inode *inode;
 
 	list_for_each_entry_safe(lo, next, &server->layouts, plh_layouts) {
-		if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags))
+		if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) ||
+		    test_bit(NFS_LAYOUT_INODE_FREEING, &lo->plh_flags) ||
+		    !list_empty(&lo->plh_bulk_destroy))
 			continue;
+		/* If the sb is being destroyed, just bail */
+		if (!nfs_sb_active(server->super))
+			break;
 		inode = igrab(lo->plh_inode);
-		if (inode == NULL)
-			continue;
-		list_del_init(&lo->plh_layouts);
-		if (pnfs_layout_add_bulk_destroy_list(inode, layout_list))
-			continue;
-		rcu_read_unlock();
-		spin_unlock(&clp->cl_lock);
-		iput(inode);
+		if (inode != NULL) {
+			list_del_init(&lo->plh_layouts);
+			if (pnfs_layout_add_bulk_destroy_list(inode,
+						layout_list))
+				continue;
+			rcu_read_unlock();
+			spin_unlock(&clp->cl_lock);
+			iput(inode);
+		} else {
+			rcu_read_unlock();
+			spin_unlock(&clp->cl_lock);
+			set_bit(NFS_LAYOUT_INODE_FREEING, &lo->plh_flags);
+		}
+		nfs_sb_deactive(server->super);
 		spin_lock(&clp->cl_lock);
 		rcu_read_lock();
 		return -EAGAIN;
@@ -811,7 +824,7 @@ pnfs_layout_free_bulk_destroy_list(struct list_head *layout_list,
 		/* Free all lsegs that are attached to commit buckets */
 		nfs_commit_inode(inode, 0);
 		pnfs_put_layout_hdr(lo);
-		iput(inode);
+		nfs_iput_and_deactive(inode);
 	}
 	return ret;
 }
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index ece367ebde69..3ba44819a88a 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -104,6 +104,7 @@ enum {
 	NFS_LAYOUT_RETURN_REQUESTED,	/* Return this layout ASAP */
 	NFS_LAYOUT_INVALID_STID,	/* layout stateid id is invalid */
 	NFS_LAYOUT_FIRST_LAYOUTGET,	/* Serialize first layoutget */
+	NFS_LAYOUT_INODE_FREEING,	/* The inode is being freed */
 };
 
 enum layoutdriver_policy_flags {
-- 
2.20.1