NFS client bugfixes for Linux 4.7

Stable bugfixes:
 - Fix _cancel_empty_pagelist
 - Fix a double page unlock
 - Make nfs_atomic_open() call d_drop() on all ->open_context() errors.
 - Fix another OPEN_DOWNGRADE bug
 
 Other bugfixes:
 - Ensure we handle delegation errors in nfs4_proc_layoutget()
 - Layout stateids start out as being invalid
 - Add sparse lock annotations for pnfs_find_alloc_layout
 - Handle bad delegation stateids in nfs4_layoutget_handle_exception
 - Fix up O_DIRECT results
 - Fix potential use after free of state in nfs4_do_reclaim.
 - Mark the layout stateid invalid when all segments are removed
 - Don't let readdirplus revalidate an inode that was marked as stale
 - Fix potential race in nfs_fhget()
 - Fix an unused variable warning
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2
 
 iQIcBAABCAAGBQJXdDWkAAoJENfLVL+wpUDrpsUP/1F2zu12r/Zkv3ZEFhShcpQc
 2N1TRD9X7Lruod2pUD95qqjjdw+/vu3LjcyJljrasRaJENijvZ2GQhKkB7xPODlu
 qxZcmnQsH+WmpmJKcqAByAW1czcNGMoMHnt4tV0gG21NH+XUb92fgn+aGeIJDVrK
 Hcd9d8TfnFWO70ZgTUXW/hv0CXwu4MEJhN2JfF4lolbxUkmjLHHLoxSDDm0AdXGC
 EE8f0V9/7xurvOeLe5bQOQXfZPedBydsLNXa1ZacMGKgmBUoRNxJ5yCpPUtcTVBx
 HwbiY+WDFQ7MdKTzUQqqbnrIqKw8Hu4SugIV/vHRqR+Lhc6u29YGOqdU4d2G8IKW
 Nh8MBqS+dDefCkL3TJoE7MpjhP3EOO6HXnv5FjMZLOuu2X2o+Sz3+DkhYCq6pj/g
 fFh480vZfZYaTsfDf1ttvVN8kIvQ+1Uk3LK6aC2EVwgPrv+0OIRu36F0JQQimxOp
 EbYDlhVk7mzH/ZQ31GmPPbSIk+3sm2V58lqXnUMovoqPFPiN3xZDuBcnlyFrrzaI
 NjOvsdVxkOdHWbYZyBQzj16Vo651EYbAAUEwsud70N8C3aCgkTxCZ30Q0v+KqqxU
 pP5kz3zYUdkXQeHxE6T0iXG9fGcv/nGS21hTfT01YJCK7v67K8TYRNMrOEVURVgk
 LSD/CZJXJHVJn1Vr4F7o
 =IGNO
 -----END PGP SIGNATURE-----

Merge tag 'nfs-for-4.7-2' of git://git.linux-nfs.org/projects/anna/linux-nfs

Pull NFS client bugfixes from Anna Schumaker:
 "Stable bugfixes:
   - Fix _cancel_empty_pagelist
   - Fix a double page unlock
   - Make nfs_atomic_open() call d_drop() on all ->open_context() errors.
   - Fix another OPEN_DOWNGRADE bug

  Other bugfixes:
   - Ensure we handle delegation errors in nfs4_proc_layoutget()
   - Layout stateids start out as being invalid
   - Add sparse lock annotations for pnfs_find_alloc_layout
   - Handle bad delegation stateids in nfs4_layoutget_handle_exception
   - Fix up O_DIRECT results
   - Fix potential use after free of state in nfs4_do_reclaim.
   - Mark the layout stateid invalid when all segments are removed
   - Don't let readdirplus revalidate an inode that was marked as stale
   - Fix potential race in nfs_fhget()
   - Fix an unused variable warning"

* tag 'nfs-for-4.7-2' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  NFS: Fix another OPEN_DOWNGRADE bug
  make nfs_atomic_open() call d_drop() on all ->open_context() errors.
  NFS: Fix an unused variable warning
  NFS: Fix potential race in nfs_fhget()
  NFS: Don't let readdirplus revalidate an inode that was marked as stale
  NFSv4.1/pnfs: Mark the layout stateid invalid when all segments are removed
  NFS: Fix a double page unlock
  pnfs_nfs: fix _cancel_empty_pagelist
  nfs4: Fix potential use after free of state in nfs4_do_reclaim.
  NFS: Fix up O_DIRECT results
  NFS/pnfs: handle bad delegation stateids in nfs4_layoutget_handle_exception
  NFSv4.1/pnfs: Add sparse lock annotations for pnfs_find_alloc_layout
  NFSv4.1/pnfs: Layout stateids start out as being invalid
  NFSv4.1/pnfs: Ensure we handle delegation errors in nfs4_proc_layoutget()

fs/nfs/dir.c

@@ -424,12 +424,17 @@ static int xdr_decode(nfs_readdir_descriptor_t *desc,
 static
 int nfs_same_file(struct dentry *dentry, struct nfs_entry *entry)
 {
+	struct inode *inode;
 	struct nfs_inode *nfsi;
 
 	if (d_really_is_negative(dentry))
 		return 0;
 
-	nfsi = NFS_I(d_inode(dentry));
+	inode = d_inode(dentry);
+	if (is_bad_inode(inode) || NFS_STALE(inode))
+		return 0;
+
+	nfsi = NFS_I(inode);
 	if (entry->fattr->fileid == nfsi->fileid)
 		return 1;
 	if (nfs_compare_fh(entry->fh, &nfsi->fh) == 0)
@@ -1363,7 +1368,6 @@ EXPORT_SYMBOL_GPL(nfs_dentry_operations);
 struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned int flags)
 {
 	struct dentry *res;
-	struct dentry *parent;
 	struct inode *inode = NULL;
 	struct nfs_fh *fhandle = NULL;
 	struct nfs_fattr *fattr = NULL;
@@ -1393,7 +1397,6 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in
 	if (IS_ERR(label))
 		goto out;
 
-	parent = dentry->d_parent;
 	/* Protect against concurrent sillydeletes */
 	trace_nfs_lookup_enter(dir, dentry, flags);
 	error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr, label);
@@ -1536,9 +1539,9 @@ int nfs_atomic_open(struct inode *dir, struct dentry *dentry,
 		err = PTR_ERR(inode);
 		trace_nfs_atomic_open_exit(dir, ctx, open_flags, err);
 		put_nfs_open_context(ctx);
+		d_drop(dentry);
 		switch (err) {
 		case -ENOENT:
-			d_drop(dentry);
 			d_add(dentry, NULL);
 			nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 			break;

fs/nfs/direct.c

@@ -353,10 +353,12 @@ static ssize_t nfs_direct_wait(struct nfs_direct_req *dreq)
 
 	result = wait_for_completion_killable(&dreq->completion);
 
+	if (!result) {
+		result = dreq->count;
+		WARN_ON_ONCE(dreq->count < 0);
+	}
 	if (!result)
 		result = dreq->error;
-	if (!result)
-		result = dreq->count;
 
 out:
 	return (ssize_t) result;
@@ -386,8 +388,10 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq, bool write)
 
 	if (dreq->iocb) {
 		long res = (long) dreq->error;
-		if (!res)
+		if (dreq->count != 0) {
 			res = (long) dreq->count;
+			WARN_ON_ONCE(dreq->count < 0);
+		}
 		dreq->iocb->ki_complete(dreq->iocb, res, 0);
 	}
 

fs/nfs/inode.c

@@ -282,6 +282,7 @@ nfs_init_locked(struct inode *inode, void *opaque)
 	struct nfs_fattr	*fattr = desc->fattr;
 
 	set_nfs_fileid(inode, fattr->fileid);
+	inode->i_mode = fattr->mode;
 	nfs_copy_fh(NFS_FH(inode), desc->fh);
 	return 0;
 }

fs/nfs/nfs4proc.c

@@ -2882,12 +2882,11 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data)
 			call_close |= is_wronly;
 		else if (is_wronly)
 			calldata->arg.fmode |= FMODE_WRITE;
+		if (calldata->arg.fmode != (FMODE_READ|FMODE_WRITE))
+			call_close |= is_rdwr;
 	} else if (is_rdwr)
 		calldata->arg.fmode |= FMODE_READ|FMODE_WRITE;
 
-	if (calldata->arg.fmode == 0)
-		call_close |= is_rdwr;
-
 	if (!nfs4_valid_open_stateid(state))
 		call_close = 0;
 	spin_unlock(&state->owner->so_lock);
@@ -7924,8 +7923,8 @@ nfs4_layoutget_handle_exception(struct rpc_task *task,
 			break;
 		}
 		lo = NFS_I(inode)->layout;
-		if (lo && nfs4_stateid_match(&lgp->args.stateid,
+		if (lo && !test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) &&
-					&lo->plh_stateid)) {
+		    nfs4_stateid_match_other(&lgp->args.stateid, &lo->plh_stateid)) {
 			LIST_HEAD(head);
 
 			/*
@@ -7936,10 +7935,10 @@ nfs4_layoutget_handle_exception(struct rpc_task *task,
 			pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0);
 			spin_unlock(&inode->i_lock);
 			pnfs_free_lseg_list(&head);
+			status = -EAGAIN;
+			goto out;
 		} else
 			spin_unlock(&inode->i_lock);
-		status = -EAGAIN;
-		goto out;
 	}
 
 	status = nfs4_handle_exception(server, status, exception);
@@ -8036,7 +8035,10 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout, gfp_t gfp_flags)
 		.flags = RPC_TASK_ASYNC,
 	};
 	struct pnfs_layout_segment *lseg = NULL;
-	struct nfs4_exception exception = { .timeout = *timeout };
+	struct nfs4_exception exception = {
+		.inode = inode,
+		.timeout = *timeout,
+	};
 	int status = 0;
 
 	dprintk("--> %s\n", __func__);

fs/nfs/nfs4state.c

@@ -1488,9 +1488,9 @@ restart:
 					}
 					spin_unlock(&state->state_lock);
 				}
-				nfs4_put_open_state(state);
 				clear_bit(NFS_STATE_RECLAIM_NOGRACE,
 					&state->flags);
+				nfs4_put_open_state(state);
 				spin_lock(&sp->so_lock);
 				goto restart;
 			}

fs/nfs/pnfs.c

@@ -361,8 +361,10 @@ pnfs_layout_remove_lseg(struct pnfs_layout_hdr *lo,
 	list_del_init(&lseg->pls_list);
 	/* Matched by pnfs_get_layout_hdr in pnfs_layout_insert_lseg */
 	atomic_dec(&lo->plh_refcount);
-	if (list_empty(&lo->plh_segs))
+	if (list_empty(&lo->plh_segs)) {
+		set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
 		clear_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
+	}
 	rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq);
 }
 
@@ -1290,6 +1292,7 @@ alloc_init_layout_hdr(struct inode *ino,
 	INIT_LIST_HEAD(&lo->plh_bulk_destroy);
 	lo->plh_inode = ino;
 	lo->plh_lc_cred = get_rpccred(ctx->cred);
+	lo->plh_flags |= 1 << NFS_LAYOUT_INVALID_STID;
 	return lo;
 }
 
@@ -1297,6 +1300,8 @@ static struct pnfs_layout_hdr *
 pnfs_find_alloc_layout(struct inode *ino,
 		       struct nfs_open_context *ctx,
 		       gfp_t gfp_flags)
+	__releases(&ino->i_lock)
+	__acquires(&ino->i_lock)
 {
 	struct nfs_inode *nfsi = NFS_I(ino);
 	struct pnfs_layout_hdr *new = NULL;
@@ -1565,8 +1570,7 @@ lookup_again:
 	 * stateid, or it has been invalidated, then we must use the open
 	 * stateid.
 	 */
-	if (lo->plh_stateid.seqid == 0 ||
+	if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags)) {
-	    test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags)) {
 
 		/*
 		 * The first layoutget for the file. Need to serialize per

fs/nfs/pnfs_nfs.c

@@ -247,7 +247,11 @@ void pnfs_fetch_commit_bucket_list(struct list_head *pages,
 }
 
 /* Helper function for pnfs_generic_commit_pagelist to catch an empty
- * page list. This can happen when two commits race. */
+ * page list. This can happen when two commits race.
+ *
+ * This must be called instead of nfs_init_commit - call one or the other, but
+ * not both!
+ */
 static bool
 pnfs_generic_commit_cancel_empty_pagelist(struct list_head *pages,
 					  struct nfs_commit_data *data,
@@ -256,7 +260,11 @@ pnfs_generic_commit_cancel_empty_pagelist(struct list_head *pages,
 	if (list_empty(pages)) {
 		if (atomic_dec_and_test(&cinfo->mds->rpcs_out))
 			wake_up_atomic_t(&cinfo->mds->rpcs_out);
-		nfs_commitdata_release(data);
+		/* don't call nfs_commitdata_release - it tries to put
+		 * the open_context which is not acquired until nfs_init_commit
+		 * which has not been called on @data */
+		WARN_ON_ONCE(data->context);
+		nfs_commit_free(data);
 		return true;
 	}
 

fs/nfs/read.c

@@ -367,13 +367,13 @@ readpage_async_filler(void *data, struct page *page)
 		nfs_list_remove_request(new);
 		nfs_readpage_release(new);
 		error = desc->pgio->pg_error;
-		goto out_unlock;
+		goto out;
 	}
 	return 0;
 out_error:
 	error = PTR_ERR(new);
-out_unlock:
 	unlock_page(page);
+out:
 	return error;
 }