Merge tag 'ntfs3_for_6.6' of https://github.com/Paragon-Software-Grou…

…p/linux-ntfs3

Pull ntfs3 fixes from Konstantin Komarov:

 - memory leak

 - some logic errors, NULL dereferences

 - some code was refactored

 - more sanity checks

* tag 'ntfs3_for_6.6' of https://github.com/Paragon-Software-Group/linux-ntfs3:
  fs/ntfs3: Avoid possible memory leak
  fs/ntfs3: Fix directory element type detection
  fs/ntfs3: Fix possible null-pointer dereference in hdr_find_e()
  fs/ntfs3: Fix OOB read in ntfs_init_from_boot
  fs/ntfs3: fix panic about slab-out-of-bounds caused by ntfs_list_ea()
  fs/ntfs3: Fix NULL pointer dereference on error in attr_allocate_frame()
  fs/ntfs3: Fix possible NULL-ptr-deref in ni_readpage_cmpr()
  fs/ntfs3: Do not allow to change label if volume is read-only
  fs/ntfs3: Add more info into /proc/fs/ntfs3/<dev>/volinfo
  fs/ntfs3: Refactoring and comments
  fs/ntfs3: Fix alternative boot searching
  fs/ntfs3: Allow repeated call to ntfs3_put_sbi
  fs/ntfs3: Use inode_set_ctime_to_ts instead of inode_set_ctime
  fs/ntfs3: Fix shift-out-of-bounds in ntfs_fill_super
  fs/ntfs3: fix deadlock in mark_as_free_ex
  fs/ntfs3: Add more attributes checks in mi_enum_attr()
  fs/ntfs3: Use kvmalloc instead of kmalloc(... __GFP_NOWARN)
  fs/ntfs3: Write immediately updated ntfs state
  fs/ntfs3: Add ckeck in ni_update_parent()

fs/ntfs3/attrib.c

@@ -1106,10 +1106,10 @@ int attr_data_get_block(struct ntfs_inode *ni, CLST vcn, CLST clen, CLST *lcn,
 		}
 	}
 
-	/* 
+	/*
 	 * The code below may require additional cluster (to extend attribute list)
-	 * and / or one MFT record 
-	 * It is too complex to undo operations if -ENOSPC occurs deep inside 
+	 * and / or one MFT record
+	 * It is too complex to undo operations if -ENOSPC occurs deep inside
 	 * in 'ni_insert_nonresident'.
 	 * Return in advance -ENOSPC here if there are no free cluster and no free MFT.
 	 */
@@ -1736,10 +1736,8 @@ int attr_allocate_frame(struct ntfs_inode *ni, CLST frame, size_t compr_size,
 			le_b = NULL;
 			attr_b = ni_find_attr(ni, NULL, &le_b, ATTR_DATA, NULL,
 					      0, NULL, &mi_b);
-			if (!attr_b) {
-				err = -ENOENT;
-				goto out;
-			}
+			if (!attr_b)
+				return -ENOENT;
 
 			attr = attr_b;
 			le = le_b;

fs/ntfs3/attrlist.c

@@ -52,7 +52,8 @@ int ntfs_load_attr_list(struct ntfs_inode *ni, struct ATTRIB *attr)
 
 	if (!attr->non_res) {
 		lsize = le32_to_cpu(attr->res.data_size);
-		le = kmalloc(al_aligned(lsize), GFP_NOFS | __GFP_NOWARN);
+		/* attr is resident: lsize < record_size (1K or 4K) */
+		le = kvmalloc(al_aligned(lsize), GFP_KERNEL);
 		if (!le) {
 			err = -ENOMEM;
 			goto out;
@@ -80,7 +81,17 @@ int ntfs_load_attr_list(struct ntfs_inode *ni, struct ATTRIB *attr)
 		if (err < 0)
 			goto out;
 
-		le = kmalloc(al_aligned(lsize), GFP_NOFS | __GFP_NOWARN);
+		/* attr is nonresident.
+		 * The worst case:
+		 * 1T (2^40) extremely fragmented file.
+		 * cluster = 4K (2^12) => 2^28 fragments
+		 * 2^9 fragments per one record => 2^19 records
+		 * 2^5 bytes of ATTR_LIST_ENTRY per one record => 2^24 bytes.
+		 *
+		 * the result is 16M bytes per attribute list.
+		 * Use kvmalloc to allocate in range [several Kbytes - dozen Mbytes]
+		 */
+		le = kvmalloc(al_aligned(lsize), GFP_KERNEL);
 		if (!le) {
 			err = -ENOMEM;
 			goto out;

fs/ntfs3/bitmap.c

@@ -125,6 +125,7 @@ void wnd_close(struct wnd_bitmap *wnd)
 	struct rb_node *node, *next;
 
 	kfree(wnd->free_bits);
+	wnd->free_bits = NULL;
 	run_close(&wnd->run);
 
 	node = rb_first(&wnd->start_tree);
@@ -659,7 +660,8 @@ int wnd_init(struct wnd_bitmap *wnd, struct super_block *sb, size_t nbits)
 		wnd->bits_last = wbits;
 
 	wnd->free_bits =
-		kcalloc(wnd->nwnd, sizeof(u16), GFP_NOFS | __GFP_NOWARN);
+		kvmalloc_array(wnd->nwnd, sizeof(u16), GFP_KERNEL | __GFP_ZERO);
+
 	if (!wnd->free_bits)
 		return -ENOMEM;
 

fs/ntfs3/dir.c

@@ -309,7 +309,11 @@ static inline int ntfs_filldir(struct ntfs_sb_info *sbi, struct ntfs_inode *ni,
 		return 0;
 	}
 
-	dt_type = (fname->dup.fa & FILE_ATTRIBUTE_DIRECTORY) ? DT_DIR : DT_REG;
+	/* NTFS: symlinks are "dir + reparse" or "file + reparse" */
+	if (fname->dup.fa & FILE_ATTRIBUTE_REPARSE_POINT)
+		dt_type = DT_LNK;
+	else
+		dt_type = (fname->dup.fa & FILE_ATTRIBUTE_DIRECTORY) ? DT_DIR : DT_REG;
 
 	return !dir_emit(ctx, (s8 *)name, name_len, ino, dt_type);
 }

fs/ntfs3/file.c

@@ -745,8 +745,8 @@ static ssize_t ntfs_file_read_iter(struct kiocb *iocb, struct iov_iter *iter)
 }
 
 static ssize_t ntfs_file_splice_read(struct file *in, loff_t *ppos,
-				     struct pipe_inode_info *pipe,
-				     size_t len, unsigned int flags)
+				     struct pipe_inode_info *pipe, size_t len,
+				     unsigned int flags)
 {
 	struct inode *inode = in->f_mapping->host;
 	struct ntfs_inode *ni = ntfs_i(inode);

fs/ntfs3/frecord.c

@@ -2148,7 +2148,7 @@ int ni_readpage_cmpr(struct ntfs_inode *ni, struct page *page)
 
 	for (i = 0; i < pages_per_frame; i++) {
 		pg = pages[i];
-		if (i == idx)
+		if (i == idx || !pg)
 			continue;
 		unlock_page(pg);
 		put_page(pg);
@@ -3208,6 +3208,12 @@ static bool ni_update_parent(struct ntfs_inode *ni, struct NTFS_DUP_INFO *dup,
 		if (!fname || !memcmp(&fname->dup, dup, sizeof(fname->dup)))
 			continue;
 
+		/* Check simple case when parent inode equals current inode. */
+		if (ino_get(&fname->home) == ni->vfs_inode.i_ino) {
+			ntfs_set_state(sbi, NTFS_DIRTY_ERROR);
+			continue;
+		}
+
 		/* ntfs_iget5 may sleep. */
 		dir = ntfs_iget5(sb, &fname->home, NULL);
 		if (IS_ERR(dir)) {

fs/ntfs3/fslog.c

@@ -2168,8 +2168,10 @@ static int last_log_lsn(struct ntfs_log *log)
 
 			if (!page) {
 				page = kmalloc(log->page_size, GFP_NOFS);
-				if (!page)
-					return -ENOMEM;
+				if (!page) {
+					err = -ENOMEM;
+					goto out;
+				}
 			}
 
 			/*

fs/ntfs3/fsntfs.c

@@ -983,18 +983,11 @@ int ntfs_set_state(struct ntfs_sb_info *sbi, enum NTFS_DIRTY_FLAGS dirty)
 	if (err)
 		return err;
 
-	mark_inode_dirty(&ni->vfs_inode);
+	mark_inode_dirty_sync(&ni->vfs_inode);
 	/* verify(!ntfs_update_mftmirr()); */
 
-	/*
-	 * If we used wait=1, sync_inode_metadata waits for the io for the
-	 * inode to finish. It hangs when media is removed.
-	 * So wait=0 is sent down to sync_inode_metadata
-	 * and filemap_fdatawrite is used for the data blocks.
-	 */
-	err = sync_inode_metadata(&ni->vfs_inode, 0);
-	if (!err)
-		err = filemap_fdatawrite(ni->vfs_inode.i_mapping);
+	/* write mft record on disk. */
+	err = _ni_write_inode(&ni->vfs_inode, 1);
 
 	return err;
 }
@@ -2461,10 +2454,12 @@ void mark_as_free_ex(struct ntfs_sb_info *sbi, CLST lcn, CLST len, bool trim)
 {
 	CLST end, i, zone_len, zlen;
 	struct wnd_bitmap *wnd = &sbi->used.bitmap;
+	bool dirty = false;
 
 	down_write_nested(&wnd->rw_lock, BITMAP_MUTEX_CLUSTERS);
 	if (!wnd_is_used(wnd, lcn, len)) {
-		ntfs_set_state(sbi, NTFS_DIRTY_ERROR);
+		/* mark volume as dirty out of wnd->rw_lock */
+		dirty = true;
 
 		end = lcn + len;
 		len = 0;
@@ -2518,6 +2513,8 @@ void mark_as_free_ex(struct ntfs_sb_info *sbi, CLST lcn, CLST len, bool trim)
 
 out:
 	up_write(&wnd->rw_lock);
+	if (dirty)
+		ntfs_set_state(sbi, NTFS_DIRTY_ERROR);
 }
 
 /*

fs/ntfs3/index.c

@@ -729,6 +729,9 @@ static struct NTFS_DE *hdr_find_e(const struct ntfs_index *indx,
 	u32 total = le32_to_cpu(hdr->total);
 	u16 offs[128];
 
+	if (unlikely(!cmp))
+		return NULL;
+
 fill_table:
 	if (end > total)
 		return NULL;

fs/ntfs3/inode.c

@@ -170,8 +170,8 @@ static struct inode *ntfs_read_mft(struct inode *inode,
 		nt2kernel(std5->cr_time, &ni->i_crtime);
 #endif
 		nt2kernel(std5->a_time, &inode->i_atime);
-		ctime = inode_get_ctime(inode);
 		nt2kernel(std5->c_time, &ctime);
+		inode_set_ctime_to_ts(inode, ctime);
 		nt2kernel(std5->m_time, &inode->i_mtime);
 
 		ni->std_fa = std5->fa;
@@ -1660,7 +1660,8 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 	d_instantiate(dentry, inode);
 
 	/* Set original time. inode times (i_ctime) may be changed in ntfs_init_acl. */
-	inode->i_atime = inode->i_mtime = inode_set_ctime_to_ts(inode, ni->i_crtime);
+	inode->i_atime = inode->i_mtime =
+		inode_set_ctime_to_ts(inode, ni->i_crtime);
 	dir->i_mtime = inode_set_ctime_to_ts(dir, ni->i_crtime);
 
 	mark_inode_dirty(dir);

fs/ntfs3/namei.c

@@ -156,8 +156,8 @@ static int ntfs_link(struct dentry *ode, struct inode *dir, struct dentry *de)
 	err = ntfs_link_inode(inode, de);
 
 	if (!err) {
-		dir->i_mtime = inode_set_ctime_to_ts(inode,
-						     inode_set_ctime_current(dir));
+		dir->i_mtime = inode_set_ctime_to_ts(
+			inode, inode_set_ctime_current(dir));
 		mark_inode_dirty(inode);
 		mark_inode_dirty(dir);
 		d_instantiate(de, inode);
@@ -373,7 +373,7 @@ static int ntfs_atomic_open(struct inode *dir, struct dentry *dentry,
 
 #ifdef CONFIG_NTFS3_FS_POSIX_ACL
 	if (IS_POSIXACL(dir)) {
-		/* 
+		/*
 		 * Load in cache current acl to avoid ni_lock(dir):
 		 * ntfs_create_inode -> ntfs_init_acl -> posix_acl_create ->
 		 * ntfs_get_acl -> ntfs_get_acl_ex -> ni_lock

fs/ntfs3/ntfs.h

@@ -847,7 +847,7 @@ struct OBJECT_ID {
 	// Birth Volume Id is the Object Id of the Volume on.
 	// which the Object Id was allocated. It never changes.
 	struct GUID BirthVolumeId; //0x10:
-	
+
 	// Birth Object Id is the first Object Id that was
 	// ever assigned to this MFT Record. I.e. If the Object Id
 	// is changed for some reason, this field will reflect the

fs/ntfs3/ntfs_fs.h

@@ -42,9 +42,11 @@ enum utf16_endian;
 #define MINUS_ONE_T			((size_t)(-1))
 /* Biggest MFT / smallest cluster */
 #define MAXIMUM_BYTES_PER_MFT		4096
+#define MAXIMUM_SHIFT_BYTES_PER_MFT	12
 #define NTFS_BLOCKS_PER_MFT_RECORD	(MAXIMUM_BYTES_PER_MFT / 512)
 
 #define MAXIMUM_BYTES_PER_INDEX		4096
+#define MAXIMUM_SHIFT_BYTES_PER_INDEX	12
 #define NTFS_BLOCKS_PER_INODE		(MAXIMUM_BYTES_PER_INDEX / 512)
 
 /* NTFS specific error code when fixup failed. */
@@ -495,8 +497,6 @@ int ntfs_getattr(struct mnt_idmap *idmap, const struct path *path,
 		 struct kstat *stat, u32 request_mask, u32 flags);
 int ntfs3_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 		  struct iattr *attr);
-void ntfs_sparse_cluster(struct inode *inode, struct page *page0, CLST vcn,
-			 CLST len);
 int ntfs_file_open(struct inode *inode, struct file *file);
 int ntfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		__u64 start, __u64 len);

fs/ntfs3/record.c

@@ -189,12 +189,19 @@ int mi_read(struct mft_inode *mi, bool is_mft)
 	return err;
 }
 
+/*
+ * mi_enum_attr - start/continue attributes enumeration in record.
+ *
+ * NOTE: mi->mrec - memory of size sbi->record_size
+ * here we sure that mi->mrec->total == sbi->record_size (see mi_read)
+ */
 struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 {
 	const struct MFT_REC *rec = mi->mrec;
 	u32 used = le32_to_cpu(rec->used);
-	u32 t32, off, asize;
+	u32 t32, off, asize, prev_type;
 	u16 t16;
+	u64 data_size, alloc_size, tot_size;
 
 	if (!attr) {
 		u32 total = le32_to_cpu(rec->total);
@@ -213,6 +220,7 @@ struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 		if (!is_rec_inuse(rec))
 			return NULL;
 
+		prev_type = 0;
 		attr = Add2Ptr(rec, off);
 	} else {
 		/* Check if input attr inside record. */
@@ -226,11 +234,11 @@ struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 			return NULL;
 		}
 
-		if (off + asize < off) {
-			/* Overflow check. */
+		/* Overflow check. */
+		if (off + asize < off)
 			return NULL;
-		}
 
+		prev_type = le32_to_cpu(attr->type);
 		attr = Add2Ptr(attr, asize);
 		off += asize;
 	}
@@ -250,7 +258,11 @@ struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 
 	/* 0x100 is last known attribute for now. */
 	t32 = le32_to_cpu(attr->type);
-	if ((t32 & 0xf) || (t32 > 0x100))
+	if (!t32 || (t32 & 0xf) || (t32 > 0x100))
+		return NULL;
+
+	/* attributes in record must be ordered by type */
+	if (t32 < prev_type)
 		return NULL;
 
 	/* Check overflow and boundary. */
@@ -259,16 +271,15 @@ struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 
 	/* Check size of attribute. */
 	if (!attr->non_res) {
+		/* Check resident fields. */
 		if (asize < SIZEOF_RESIDENT)
 			return NULL;
 
 		t16 = le16_to_cpu(attr->res.data_off);
-
 		if (t16 > asize)
 			return NULL;
 
-		t32 = le32_to_cpu(attr->res.data_size);
-		if (t16 + t32 > asize)
+		if (t16 + le32_to_cpu(attr->res.data_size) > asize)
 			return NULL;
 
 		t32 = sizeof(short) * attr->name_len;
@@ -278,21 +289,52 @@ struct ATTRIB *mi_enum_attr(struct mft_inode *mi, struct ATTRIB *attr)
 		return attr;
 	}
 
-	/* Check some nonresident fields. */
-	if (attr->name_len &&
-	    le16_to_cpu(attr->name_off) + sizeof(short) * attr->name_len >
-		    le16_to_cpu(attr->nres.run_off)) {
+	/* Check nonresident fields. */
+	if (attr->non_res != 1)
+		return NULL;
+
+	t16 = le16_to_cpu(attr->nres.run_off);
+	if (t16 > asize)
+		return NULL;
+
+	t32 = sizeof(short) * attr->name_len;
+	if (t32 && le16_to_cpu(attr->name_off) + t32 > t16)
+		return NULL;
+
+	/* Check start/end vcn. */
+	if (le64_to_cpu(attr->nres.svcn) > le64_to_cpu(attr->nres.evcn) + 1)
 		return NULL;
-	}
 
-	if (attr->nres.svcn || !is_attr_ext(attr)) {
+	data_size = le64_to_cpu(attr->nres.data_size);
+	if (le64_to_cpu(attr->nres.valid_size) > data_size)
+		return NULL;
+
+	alloc_size = le64_to_cpu(attr->nres.alloc_size);
+	if (data_size > alloc_size)
+		return NULL;
+
+	t32 = mi->sbi->cluster_mask;
+	if (alloc_size & t32)
+		return NULL;
+
+	if (!attr->nres.svcn && is_attr_ext(attr)) {
+		/* First segment of sparse/compressed attribute */
+		if (asize + 8 < SIZEOF_NONRESIDENT_EX)
+			return NULL;
+
+		tot_size = le64_to_cpu(attr->nres.total_size);
+		if (tot_size & t32)
+			return NULL;
+
+		if (tot_size > alloc_size)
+			return NULL;
+	} else {
 		if (asize + 8 < SIZEOF_NONRESIDENT)
 			return NULL;
 
 		if (attr->nres.c_unit)
 			return NULL;
-	} else if (asize + 8 < SIZEOF_NONRESIDENT_EX)
-		return NULL;
+	}
 
 	return attr;
 }