[PATCH AUTOSEL 5.10 09/31] NFSv4: Fix a pNFS layout related use-after-free race when freeing the inode

[PATCH AUTOSEL 5.10 09/31] NFSv4: Fix a pNFS layout related use-after-free race when freeing the inode

[Sasha Levin]

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

[ Upstream commit b6d49ecd1081740b6e632366428b960461f8158b ]

When returning the layout in nfs4_evict_inode(), we need to ensure that
the layout is actually done being freed before we can proceed to free the
inode itself.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/nfs4super.c |  2 +-
 fs/nfs/pnfs.c      | 33 +++++++++++++++++++++++++++++++--
 fs/nfs/pnfs.h      |  5 +++++
 3 files changed, 37 insertions(+), 3 deletions(-)

diff --git a/fs/nfs/nfs4super.c b/fs/nfs/nfs4super.c
index 93f5c1678ec29..984cc42ee54d8 100644
--- a/fs/nfs/nfs4super.c
+++ b/fs/nfs/nfs4super.c
@@ -67,7 +67,7 @@ static void nfs4_evict_inode(struct inode *inode)
 	nfs_inode_evict_delegation(inode);
 	/* Note that above delegreturn would trigger pnfs return-on-close */
 	pnfs_return_layout(inode);
-	pnfs_destroy_layout(NFS_I(inode));
+	pnfs_destroy_layout_final(NFS_I(inode));
 	/* First call standard NFS clear_inode() code */
 	nfs_clear_inode(inode);
 	nfs4_xattr_cache_zap(inode);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 0e50b9d45c320..07f59dc8cb2e7 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -294,6 +294,7 @@ void
 pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 {
 	struct inode *inode;
+	unsigned long i_state;
 
 	if (!lo)
 		return;
@@ -304,8 +305,12 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 		if (!list_empty(&lo->plh_segs))
 			WARN_ONCE(1, "NFS: BUG unfreed layout segments.\n");
 		pnfs_detach_layout_hdr(lo);
+		i_state = inode->i_state;
 		spin_unlock(&inode->i_lock);
 		pnfs_free_layout_hdr(lo);
+		/* Notify pnfs_destroy_layout_final() that we're done */
+		if (i_state & (I_FREEING | I_CLEAR))
+			wake_up_var(lo);
 	}
 }
 
@@ -734,8 +739,7 @@ pnfs_free_lseg_list(struct list_head *free_me)
 	}
 }
 
-void
-pnfs_destroy_layout(struct nfs_inode *nfsi)
+static struct pnfs_layout_hdr *__pnfs_destroy_layout(struct nfs_inode *nfsi)
 {
 	struct pnfs_layout_hdr *lo;
 	LIST_HEAD(tmp_list);
@@ -753,9 +757,34 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 		pnfs_put_layout_hdr(lo);
 	} else
 		spin_unlock(&nfsi->vfs_inode.i_lock);
+	return lo;
+}
+
+void pnfs_destroy_layout(struct nfs_inode *nfsi)
+{
+	__pnfs_destroy_layout(nfsi);
 }
 EXPORT_SYMBOL_GPL(pnfs_destroy_layout);
 
+static bool pnfs_layout_removed(struct nfs_inode *nfsi,
+				struct pnfs_layout_hdr *lo)
+{
+	bool ret;
+
+	spin_lock(&nfsi->vfs_inode.i_lock);
+	ret = nfsi->layout != lo;
+	spin_unlock(&nfsi->vfs_inode.i_lock);
+	return ret;
+}
+
+void pnfs_destroy_layout_final(struct nfs_inode *nfsi)
+{
+	struct pnfs_layout_hdr *lo = __pnfs_destroy_layout(nfsi);
+
+	if (lo)
+		wait_var_event(lo, pnfs_layout_removed(nfsi, lo));
+}
+
 static bool
 pnfs_layout_add_bulk_destroy_list(struct inode *inode,
 		struct list_head *layout_list)
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 2661c44c62db4..78c3893918486 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -266,6 +266,7 @@ struct pnfs_layout_segment *pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_layoutget_free(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
+void pnfs_destroy_layout_final(struct nfs_inode *);
 void pnfs_destroy_all_layouts(struct nfs_client *);
 int pnfs_destroy_layouts_byfsid(struct nfs_client *clp,
 		struct nfs_fsid *fsid,
@@ -710,6 +711,10 @@ static inline void pnfs_destroy_layout(struct nfs_inode *nfsi)
 {
 }
 
+static inline void pnfs_destroy_layout_final(struct nfs_inode *nfsi)
+{
+}
+
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
2.27.0

[PATCH AUTOSEL 5.10 21/31] NFSv4.2: Don't error when exiting early on a READ_PLUS buffer overflow

[Sasha Levin]

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

[ Upstream commit 503b934a752f7e789a5f33217520e0a79f3096ac ]

Expanding the READ_PLUS extents can cause the read buffer to overflow.
If it does, then don't error, but just exit early.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/nfs42xdr.c | 36 +++++++++++++++++-------------------
 1 file changed, 17 insertions(+), 19 deletions(-)

diff --git a/fs/nfs/nfs42xdr.c b/fs/nfs/nfs42xdr.c
index 8432bd6b95f08..c078f88552695 100644
--- a/fs/nfs/nfs42xdr.c
+++ b/fs/nfs/nfs42xdr.c
@@ -1019,29 +1019,24 @@ static int decode_deallocate(struct xdr_stream *xdr, struct nfs42_falloc_res *re
 	return decode_op_hdr(xdr, OP_DEALLOCATE);
 }
 
-static int decode_read_plus_data(struct xdr_stream *xdr, struct nfs_pgio_res *res,
-				 uint32_t *eof)
+static int decode_read_plus_data(struct xdr_stream *xdr,
+				 struct nfs_pgio_res *res)
 {
 	uint32_t count, recvd;
 	uint64_t offset;
 	__be32 *p;
 
 	p = xdr_inline_decode(xdr, 8 + 4);
-	if (unlikely(!p))
-		return -EIO;
+	if (!p)
+		return 1;
 
 	p = xdr_decode_hyper(p, &offset);
 	count = be32_to_cpup(p);
 	recvd = xdr_align_data(xdr, res->count, count);
 	res->count += recvd;
 
-	if (count > recvd) {
-		dprintk("NFS: server cheating in read reply: "
-				"count %u > recvd %u\n", count, recvd);
-		*eof = 0;
+	if (count > recvd)
 		return 1;
-	}
-
 	return 0;
 }
 
@@ -1052,18 +1047,16 @@ static int decode_read_plus_hole(struct xdr_stream *xdr, struct nfs_pgio_res *re
 	__be32 *p;
 
 	p = xdr_inline_decode(xdr, 8 + 8);
-	if (unlikely(!p))
-		return -EIO;
+	if (!p)
+		return 1;
 
 	p = xdr_decode_hyper(p, &offset);
 	p = xdr_decode_hyper(p, &length);
 	recvd = xdr_expand_hole(xdr, res->count, length);
 	res->count += recvd;
 
-	if (recvd < length) {
-		*eof = 0;
+	if (recvd < length)
 		return 1;
-	}
 	return 0;
 }
 
@@ -1088,12 +1081,12 @@ static int decode_read_plus(struct xdr_stream *xdr, struct nfs_pgio_res *res)
 
 	for (i = 0; i < segments; i++) {
 		p = xdr_inline_decode(xdr, 4);
-		if (unlikely(!p))
-			return -EIO;
+		if (!p)
+			goto early_out;
 
 		type = be32_to_cpup(p++);
 		if (type == NFS4_CONTENT_DATA)
-			status = decode_read_plus_data(xdr, res, &eof);
+			status = decode_read_plus_data(xdr, res);
 		else if (type == NFS4_CONTENT_HOLE)
 			status = decode_read_plus_hole(xdr, res, &eof);
 		else
@@ -1102,12 +1095,17 @@ static int decode_read_plus(struct xdr_stream *xdr, struct nfs_pgio_res *res)
 		if (status < 0)
 			return status;
 		if (status > 0)
-			break;
+			goto early_out;
 	}
 
 out:
 	res->eof = eof;
 	return 0;
+early_out:
+	if (unlikely(!i))
+		return -EIO;
+	res->eof = 0;
+	return 0;
 }
 
 static int decode_seek(struct xdr_stream *xdr, struct nfs42_seek_res *res)
-- 
2.27.0