Initial implementation of an NTFS filesystem crate, with access to NTFS files, resident attributes, StandardInformation and FileName structures, string and time parsing, with real filesystem tests for all of that.

23 files changed, 2095 insertions, 0 deletions

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..96ef6c0
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+/target
+Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
new file mode 100644
index 0000000..8e4da25
--- /dev/null
+++ b/Cargo.toml
@@ -0,0 +1,19 @@
+[package]
+name = "ntfs"
+version = "0.1.0"
+authors = ["Colin Finck <colin@reactos.org>"]
+edition = "2018"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+binread = { path = "../binread/binread", features = ["const_generics"], default-features = false }
+chrono = { version = "0.4.19", optional = true }
+bitflags = "1.2.1"
+displaydoc = { version = "0.1.7", default-features = false }
+enumn = "0.1.3"
+memoffset = "0.6.1"
+
+[features]
+default = ["std"]
+std = ["binread/std"]
diff --git a/LICENSE b/LICENSE
new file mode 100644
index 0000000..d159169
--- /dev/null
+++ b/LICENSE
@@ -0,0 +1,339 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                            NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.
diff --git a/src/attribute.rs b/src/attribute.rs
new file mode 100644
index 0000000..254fa58
--- /dev/null
+++ b/src/attribute.rs
@@ -0,0 +1,410 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::attribute_value::{NtfsAttributeResidentValue, NtfsAttributeValue};
+use crate::error::{NtfsError, Result};
+use crate::ntfs_file::NtfsFile;
+use crate::string::NtfsString;
+use crate::structured_values::{NtfsFileName, NtfsStandardInformation, NtfsStructuredValue};
+use binread::io::{Read, Seek, SeekFrom};
+use binread::{BinRead, BinReaderExt};
+use bitflags::bitflags;
+use core::iter::FusedIterator;
+use core::mem;
+use core::ops::Range;
+use enumn::N;
+
+/// On-disk structure of the generic header of an NTFS attribute.
+#[allow(unused)]
+#[derive(BinRead)]
+struct NtfsAttributeHeader {
+    /// Type of the attribute, known types are in [`NtfsAttributeType`].
+    ty: u32,
+    /// Length of the resident part of this attribute, in bytes.
+    length: u32,
+    /// 0 if this attribute has a resident value, 1 if this attribute has a non-resident value.
+    is_non_resident: u8,
+    /// Length of the name, in UTF-16 code points (every code point is 2 bytes).
+    name_length: u8,
+    /// Offset to the beginning of the name, in bytes from the beginning of this header.
+    name_offset: u16,
+    /// Flags of the attribute, known flags are in [`NtfsAttributeFlags`].
+    flags: u16,
+    /// Identifier of this attribute that is unique within the [`NtfsFile`].
+    instance: u16,
+}
+
+impl NtfsAttributeHeader {
+    fn is_resident(&self) -> bool {
+        self.is_non_resident == 0
+    }
+}
+
+bitflags! {
+    pub struct NtfsAttributeFlags: u16 {
+        /// The attribute value is compressed.
+        const COMPRESSED = 0x0001;
+        /// The attribute value is encrypted.
+        const ENCRYPTED = 0x4000;
+        /// The attribute value is stored sparsely.
+        const SPARSE = 0x8000;
+    }
+}
+
+/// On-disk structure of the extra header of an NTFS attribute that has a resident value.
+#[allow(unused)]
+#[derive(BinRead)]
+struct NtfsAttributeResidentHeader {
+    /// Length of the value, in bytes.
+    value_length: u32,
+    /// Offset to the beginning of the value, in bytes from the beginning of the [`NtfsAttributeHeader`].
+    value_offset: u16,
+    /// 1 if this attribute (with resident value) is referenced in an index.
+    indexed_flag: u8,
+}
+
+/// On-disk structure of the extra header of an NTFS attribute that has a non-resident value.
+#[allow(unused)]
+#[derive(BinRead)]
+struct NtfsAttributeNonResidentHeader {
+    /// Lower boundary of Virtual Cluster Numbers (VCNs) referenced by this attribute.
+    /// This becomes relevant when file data is split over multiple attributes.
+    /// Otherwise, it's zero.
+    lowest_vcn: i64,
+    /// Upper boundary of Virtual Cluster Numbers (VCNs) referenced by this attribute.
+    /// This becomes relevant when file data is split over multiple attributes.
+    /// Otherwise, it's zero (or even -1 for zero-length files according to NTFS-3G).
+    highest_vcn: i64,
+    /// Offset to the beginning of the value data runs.
+    data_runs_offset: u16,
+    /// Binary exponent denoting the number of clusters in a compression unit.
+    /// A typical value is 4, meaning that 2^4 = 16 clusters are part of a compression unit.
+    /// A value of zero means no compression (but that should better be determined via
+    /// [`NtfsAttributeFlags`]).
+    compression_unit_exponent: u8,
+    reserved: [u8; 5],
+    /// Allocated space for the attribute value, in bytes. This is always a multiple of the cluster size.
+    /// For compressed files, this is always a multiple of the compression unit size.
+    allocated_size: i64,
+    /// Size of the attribute value, in bytes.
+    /// This can be larger than `allocated_size` if the value is compressed or stored sparsely.
+    data_size: i64,
+    /// Size of the initialized part of the attribute value, in bytes.
+    /// This is usually the same as `data_size`.
+    initialized_size: i64,
+}
+
+#[derive(Clone, Copy, Debug, Eq, N, PartialEq)]
+#[repr(u32)]
+pub enum NtfsAttributeType {
+    StandardInformation = 0x10,
+    AttributeList = 0x20,
+    FileName = 0x30,
+    ObjectId = 0x40,
+    SecurityDescriptor = 0x50,
+    VolumeName = 0x60,
+    VolumeInformation = 0x70,
+    Data = 0x80,
+    IndexRoot = 0x90,
+    IndexAllocation = 0xA0,
+    Bitmap = 0xB0,
+    ReparsePoint = 0xC0,
+    EAInformation = 0xD0,
+    EA = 0xE0,
+    PropertySet = 0xF0,
+    LoggedUtilityStream = 0x100,
+    End = 0xFFFF_FFFF,
+}
+
+enum NtfsAttributeExtraHeader {
+    Resident(NtfsAttributeResidentHeader),
+    NonResident(NtfsAttributeNonResidentHeader),
+}
+
+impl NtfsAttributeExtraHeader {
+    fn new<T>(fs: &mut T, header: &NtfsAttributeHeader) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        if header.is_resident() {
+            // Read the resident header.
+            let resident_header = fs.read_le::<NtfsAttributeResidentHeader>()?;
+            Ok(Self::Resident(resident_header))
+        } else {
+            // Read the non-resident header.
+            let non_resident_header = fs.read_le::<NtfsAttributeNonResidentHeader>()?;
+            Ok(Self::NonResident(non_resident_header))
+        }
+    }
+}
+
+pub struct NtfsAttribute {
+    position: u64,
+    header: NtfsAttributeHeader,
+    extra_header: NtfsAttributeExtraHeader,
+}
+
+impl NtfsAttribute {
+    fn new<T>(fs: &mut T, position: u64) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        // Read the common header for resident and non-resident attributes.
+        fs.seek(SeekFrom::Start(position))?;
+        let header = fs.read_le::<NtfsAttributeHeader>()?;
+
+        // This must be a real attribute and not an end marker!
+        // The caller must have already checked for potential end markers.
+        debug_assert!(header.ty != NtfsAttributeType::End as u32);
+
+        // Read the extra header specific to the attribute type.
+        let extra_header = NtfsAttributeExtraHeader::new(fs, &header)?;
+
+        let attribute = Self {
+            position,
+            header,
+            extra_header,
+        };
+        Ok(attribute)
+    }
+
+    /// Returns the length of this NTFS attribute, in bytes.
+    ///
+    /// This denotes the length of the attribute structure on disk.
+    /// Apart from various headers, this structure also includes the name and,
+    /// for resident attributes, the actual value.
+    pub fn attribute_length(&self) -> u32 {
+        self.header.length
+    }
+
+    /// Returns flags set for this attribute as specified by [`NtfsAttributeFlags`].
+    pub fn flags(&self) -> NtfsAttributeFlags {
+        NtfsAttributeFlags::from_bits_truncate(self.header.flags)
+    }
+
+    /// Returns `true` if this is a resident attribute, i.e. one where its value
+    /// is part of the attribute structure.
+    pub fn is_resident(&self) -> bool {
+        self.header.is_resident()
+    }
+
+    /// Returns the length of the name of this NTFS attribute, in bytes.
+    ///
+    /// An attribute name has a maximum length of 255 UTF-16 code points (510 bytes).
+    /// It is always part of the attribute itself and hence also of the length
+    /// returned by [`NtfsAttribute::attribute_length`].
+    pub fn name_length(&self) -> usize {
+        self.header.name_length as usize * mem::size_of::<u16>()
+    }
+
+    /// Returns the absolute position of this NTFS attribute within the filesystem, in bytes.
+    pub fn position(&self) -> u64 {
+        self.position
+    }
+
+    /// Reads the name of this NTFS attribute into the given buffer, and returns an
+    /// [`NtfsString`] wrapping that buffer.
+    ///
+    /// Note that most NTFS attributes have no name and are distinguished by their types.
+    /// Use [`NtfsAttribute::ty`] to get the attribute type.
+    pub fn read_name<'a, T>(&self, fs: &mut T, buf: &'a mut [u8]) -> Result<NtfsString<'a>>
+    where
+        T: Read + Seek,
+    {
+        let name_length = self.name_length();
+        if buf.len() < name_length {
+            return Err(NtfsError::BufferTooSmall {
+                expected: name_length,
+                actual: buf.len(),
+            });
+        }
+
+        let name_position = self.position + self.header.name_offset as u64;
+        fs.seek(SeekFrom::Start(name_position))?;
+        fs.read_exact(&mut buf[..name_length])?;
+
+        Ok(NtfsString(&buf[..name_length]))
+    }
+
+    pub fn read_structured_value<T>(&self, fs: &mut T) -> Result<NtfsStructuredValue>
+    where
+        T: Read + Seek,
+    {
+        let attached_value = self.value().attach(fs);
+
+        match self.ty()? {
+            NtfsAttributeType::StandardInformation => {
+                let inner = NtfsStandardInformation::new(
+                    self.position,
+                    attached_value,
+                    self.value_length(),
+                )?;
+                Ok(NtfsStructuredValue::StandardInformation(inner))
+            }
+            NtfsAttributeType::AttributeList => panic!("TODO"),
+            NtfsAttributeType::FileName => {
+                let inner = NtfsFileName::new(self.position, attached_value, self.value_length())?;
+                Ok(NtfsStructuredValue::FileName(inner))
+            }
+            NtfsAttributeType::ObjectId => panic!("TODO"),
+            NtfsAttributeType::SecurityDescriptor => panic!("TODO"),
+            NtfsAttributeType::VolumeName => panic!("TODO"),
+            NtfsAttributeType::VolumeInformation => panic!("TODO"),
+            ty => Err(NtfsError::UnsupportedStructuredValue {
+                position: self.position,
+                ty,
+            }),
+        }
+    }
+
+    /// Returns the type of this NTFS attribute, or [`NtfsError::UnsupportedNtfsAttributeType`]
+    /// if it's an unknown type.
+    pub fn ty(&self) -> Result<NtfsAttributeType> {
+        NtfsAttributeType::n(self.header.ty).ok_or(NtfsError::UnsupportedNtfsAttributeType {
+            position: self.position,
+            actual: self.header.ty,
+        })
+    }
+
+    /// Returns an [`NtfsAttributeValue`] structure to read the value of this NTFS attribute.
+    pub fn value(&self) -> NtfsAttributeValue {
+        match &self.extra_header {
+            NtfsAttributeExtraHeader::Resident(resident_header) => {
+                let value_position = self.position + resident_header.value_offset as u64;
+                let value_length = resident_header.value_length;
+                let value = NtfsAttributeResidentValue::new(value_position, value_length);
+                NtfsAttributeValue::Resident(value)
+            }
+            NtfsAttributeExtraHeader::NonResident(_non_resident_header) => {
+                panic!("TODO")
+            }
+        }
+    }
+
+    /// Returns the length of the value of this NTFS attribute, in bytes.
+    pub fn value_length(&self) -> u64 {
+        match &self.extra_header {
+            NtfsAttributeExtraHeader::Resident(resident_header) => {
+                resident_header.value_length as u64
+            }
+            NtfsAttributeExtraHeader::NonResident(non_resident_header) => {
+                non_resident_header.data_size as u64
+            }
+        }
+    }
+}
+
+pub struct NtfsAttributes<'a, T: Read + Seek> {
+    fs: &'a mut T,
+    items_range: Range<u64>,
+}
+
+impl<'a, T> NtfsAttributes<'a, T>
+where
+    T: Read + Seek,
+{
+    pub(crate) fn new(fs: &'a mut T, file: &NtfsFile) -> Self {
+        let start = file.position + file.header.first_attribute_offset as u64;
+        let end = file.position + file.header.used_size as u64;
+        let items_range = start..end;
+
+        Self { fs, items_range }
+    }
+}
+
+impl<'a, T> Iterator for NtfsAttributes<'a, T>
+where
+    T: Read + Seek,
+{
+    type Item = Result<NtfsAttribute>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        if self.items_range.is_empty() {
+            return None;
+        }
+
+        // This may be an entire attribute or just the 4-byte end marker.
+        // Check if this marks the end of the attribute list.
+        let offset = self.items_range.start;
+        iter_try!(self.fs.seek(SeekFrom::Start(offset)));
+        let ty = iter_try!(self.fs.read_le::<u32>());
+        if ty == NtfsAttributeType::End as u32 {
+            return None;
+        }
+
+        // It's a real attribute.
+        let attribute = iter_try!(NtfsAttribute::new(&mut self.fs, offset));
+        self.items_range.start += attribute.attribute_length() as u64;
+
+        Some(Ok(attribute))
+    }
+}
+
+impl<'a, T> FusedIterator for NtfsAttributes<'a, T> where T: Read + Seek {}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::ntfs::Ntfs;
+    use crate::ntfs_file::KnownNtfsFile;
+    use crate::time::tests::NT_TIMESTAMP_2021_01_01;
+
+    #[test]
+    fn test_data() {
+        let mut testfs1 = crate::helpers::tests::testfs1();
+        let ntfs = Ntfs::new(&mut testfs1).unwrap();
+        let mft = ntfs
+            .ntfs_file(&mut testfs1, KnownNtfsFile::MFT as u64)
+            .unwrap();
+        let mut mft_attributes = mft.attributes(&mut testfs1);
+
+        // Check the StandardInformation attribute.
+        let attribute = mft_attributes.next().unwrap().unwrap();
+        assert_eq!(
+            attribute.ty().unwrap(),
+            NtfsAttributeType::StandardInformation,
+        );
+        assert_eq!(attribute.attribute_length(), 96);
+        assert!(attribute.is_resident());
+        assert_eq!(attribute.name_length(), 0);
+        assert_eq!(attribute.value_length(), 72);
+
+        // Check the FileName attribute.
+        let attribute = mft_attributes.next().unwrap().unwrap();
+        assert_eq!(attribute.ty().unwrap(), NtfsAttributeType::FileName);
+        assert_eq!(attribute.attribute_length(), 104);
+        assert!(attribute.is_resident());
+        assert_eq!(attribute.name_length(), 0);
+        assert_eq!(attribute.value_length(), 74);
+
+        // Check the actual "file name" of the MFT.
+        let value = attribute.read_structured_value(&mut testfs1).unwrap();
+        let file_name = match value {
+            NtfsStructuredValue::FileName(file_name) => file_name,
+            v => panic!("Unexpected NtfsStructuredValue: {:?}", v),
+        };
+
+        let creation_time = file_name.creation_time();
+        assert!(*creation_time > NT_TIMESTAMP_2021_01_01);
+        assert_eq!(creation_time, file_name.modification_time());
+        assert_eq!(creation_time, file_name.mft_record_modification_time());
+        assert_eq!(creation_time, file_name.access_time());
+
+        let allocated_size = file_name.allocated_size();
+        assert!(allocated_size > 0);
+        assert_eq!(allocated_size, file_name.data_size());
+
+        assert_eq!(file_name.name_length(), 8);
+
+        let mut buf = [0u8; 8];
+        let file_name_string = file_name.read_name(&mut testfs1, &mut buf).unwrap();
+
+        // Test various ways to compare the same string.
+        assert_eq!(file_name_string, "$MFT");
+        assert_eq!(file_name_string.to_string_lossy(), String::from("$MFT"));
+        assert_eq!(
+            file_name_string,
+            NtfsString(&[b'$', 0, b'M', 0, b'F', 0, b'T', 0])
+        );
+    }
+}
diff --git a/src/attribute_value.rs b/src/attribute_value.rs
new file mode 100644
index 0000000..20593a7
--- /dev/null
+++ b/src/attribute_value.rs
@@ -0,0 +1,215 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use binread::io;
+use binread::io::{Error, ErrorKind, Read, Seek, SeekFrom};
+use core::cmp;
+
+pub trait NtfsAttributeRead<T>
+where
+    T: Read + Seek,
+{
+    fn read(&mut self, fs: &mut T, buf: &mut [u8]) -> io::Result<usize>;
+
+    fn read_exact(&mut self, fs: &mut T, mut buf: &mut [u8]) -> io::Result<()> {
+        // This implementation is taken from https://github.com/rust-lang/rust/blob/5662d9343f0696efcc38a1264656737c9f22d427/library/std/src/io/mod.rs
+        // It handles all corner cases properly and outputs the known `io` error messages.
+        while !buf.is_empty() {
+            match self.read(fs, buf) {
+                Ok(0) => break,
+                Ok(n) => {
+                    buf = &mut buf[n..];
+                }
+                Err(ref e) if e.kind() == ErrorKind::Interrupted => {}
+                Err(e) => return Err(e),
+            }
+        }
+
+        if !buf.is_empty() {
+            Err(Error::new(
+                ErrorKind::UnexpectedEof,
+                "failed to fill whole buffer",
+            ))
+        } else {
+            Ok(())
+        }
+    }
+}
+
+pub enum NtfsAttributeValue {
+    Resident(NtfsAttributeResidentValue),
+    NonResident(NtfsAttributeNonResidentValue),
+}
+
+impl NtfsAttributeValue {
+    pub fn attach<'a, T>(self, fs: &'a mut T) -> NtfsAttributeValueAttached<'a, T>
+    where
+        T: Read + Seek,
+    {
+        NtfsAttributeValueAttached { fs, value: self }
+    }
+
+    pub(crate) fn position(&self) -> u64 {
+        match self {
+            Self::Resident(inner) => inner.position(),
+            Self::NonResident(inner) => inner.position(),
+        }
+    }
+}
+
+impl<T> NtfsAttributeRead<T> for NtfsAttributeValue
+where
+    T: Read + Seek,
+{
+    fn read(&mut self, fs: &mut T, buf: &mut [u8]) -> io::Result<usize> {
+        match self {
+            Self::Resident(inner) => inner.read(fs, buf),
+            Self::NonResident(inner) => inner.read(fs, buf),
+        }
+    }
+}
+
+impl Seek for NtfsAttributeValue {
+    fn seek(&mut self, pos: SeekFrom) -> io::Result<u64> {
+        match self {
+            Self::Resident(inner) => inner.seek(pos),
+            Self::NonResident(inner) => inner.seek(pos),
+        }
+    }
+}
+
+pub struct NtfsAttributeValueAttached<'a, T: Read + Seek> {
+    fs: &'a mut T,
+    value: NtfsAttributeValue,
+}
+
+impl<'a, T> NtfsAttributeValueAttached<'a, T>
+where
+    T: Read + Seek,
+{
+    pub fn detach(self) -> NtfsAttributeValue {
+        self.value
+    }
+
+    pub fn position(&self) -> u64 {
+        self.value.position()
+    }
+}
+
+impl<'a, T> Read for NtfsAttributeValueAttached<'a, T>
+where
+    T: Read + Seek,
+{
+    fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
+        self.value.read(self.fs, buf)
+    }
+}
+
+impl<'a, T> Seek for NtfsAttributeValueAttached<'a, T>
+where
+    T: Read + Seek,
+{
+    fn seek(&mut self, pos: SeekFrom) -> io::Result<u64> {
+        self.value.seek(pos)
+    }
+}
+
+pub struct NtfsAttributeResidentValue {
+    /// Absolute position of the attribute's value within the filesystem, in bytes.
+    position: u64,
+    /// Total length of the attribute's value, in bytes.
+    length: u32,
+    /// Current relative seek position within the value, in bytes.
+    seek_position: u64,
+}
+
+impl NtfsAttributeResidentValue {
+    pub(crate) fn new(position: u64, length: u32) -> Self {
+        Self {
+            position,
+            length,
+            seek_position: 0,
+        }
+    }
+
+    pub fn position(&self) -> u64 {
+        self.position
+    }
+}
+
+impl<T> NtfsAttributeRead<T> for NtfsAttributeResidentValue
+where
+    T: Read + Seek,
+{
+    fn read(&mut self, fs: &mut T, buf: &mut [u8]) -> io::Result<usize> {
+        let bytes_left = (self.length as u64).saturating_sub(self.seek_position);
+        if bytes_left == 0 {
+            return Ok(0);
+        }
+
+        let bytes_to_read = cmp::min(buf.len(), bytes_left as usize);
+
+        fs.seek(SeekFrom::Start(self.position + self.seek_position))?;
+        fs.read(&mut buf[..bytes_to_read])?;
+
+        self.seek_position += bytes_to_read as u64;
+        Ok(bytes_to_read)
+    }
+}
+
+impl Seek for NtfsAttributeResidentValue {
+    fn seek(&mut self, pos: SeekFrom) -> io::Result<u64> {
+        // This implementation is taken from https://github.com/rust-lang/rust/blob/18c524fbae3ab1bf6ed9196168d8c68fc6aec61a/library/std/src/io/cursor.rs
+        // It handles all signed/unsigned arithmetics properly and outputs the known `io` error messages.
+        let (base_pos, offset) = match pos {
+            SeekFrom::Start(n) => {
+                self.seek_position = n;
+                return Ok(n);
+            }
+            SeekFrom::End(n) => (self.length as u64, n),
+            SeekFrom::Current(n) => (self.seek_position, n),
+        };
+
+        let new_pos = if offset >= 0 {
+            base_pos.checked_add(offset as u64)
+        } else {
+            base_pos.checked_sub(offset.wrapping_neg() as u64)
+        };
+
+        match new_pos {
+            Some(n) => {
+                self.seek_position = n;
+                Ok(self.seek_position)
+            }
+            None => Err(Error::new(
+                ErrorKind::InvalidInput,
+                "invalid seek to a negative or overflowing position",
+            )),
+        }
+    }
+}
+
+pub struct NtfsAttributeNonResidentValue {
+    // TODO
+}
+
+impl NtfsAttributeNonResidentValue {
+    pub fn position(&self) -> u64 {
+        panic!("TODO")
+    }
+}
+
+impl<T> NtfsAttributeRead<T> for NtfsAttributeNonResidentValue
+where
+    T: Read + Seek,
+{
+    fn read(&mut self, _fs: &mut T, _buf: &mut [u8]) -> io::Result<usize> {
+        panic!("TODO")
+    }
+}
+
+impl Seek for NtfsAttributeNonResidentValue {
+    fn seek(&mut self, _pos: SeekFrom) -> io::Result<u64> {
+        panic!("TODO")
+    }
+}
diff --git a/src/boot_sector.rs b/src/boot_sector.rs
new file mode 100644
index 0000000..0ab2f64
--- /dev/null
+++ b/src/boot_sector.rs
@@ -0,0 +1,91 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::error::{NtfsError, Result};
+use binread::BinRead;
+use memoffset::offset_of;
+
+/// The usual exponent of `BiosParameterBlock::file_record_size_info` is 10 (2^10 = 1024 bytes).
+/// Exponents > 10 would come as a surprise, but our code should still be able to handle those.
+/// Exponents > 32 (2^32 = 4 GiB) would make no sense, exceed a u32, and must be outright denied.
+const MAXIMUM_SIZE_INFO_EXPONENT: u32 = 32;
+
+// Sources:
+// - https://en.wikipedia.org/wiki/NTFS#Partition_Boot_Sector_(VBR)
+// - https://en.wikipedia.org/wiki/BIOS_parameter_block#NTFS
+// - https://wiki.osdev.org/NTFS
+// - The iBored tool from https://apps.tempel.org/iBored/
+#[allow(unused)]
+#[derive(BinRead)]
+pub(crate) struct BiosParameterBlock {
+    pub(crate) bytes_per_sector: u16,
+    pub(crate) sectors_per_cluster: u8,
+    zeros_1: [u8; 7],
+    media: u8,
+    zeros_2: [u8; 2],
+    dummy_sectors_per_track: u16,
+    dummy_heads: u16,
+    hidden_sectors: u32,
+    zeros_3: u32,
+    physical_drive_number: u8,
+    flags: u8,
+    extended_boot_signature: u8,
+    reserved: u8,
+    pub(crate) total_sectors: u64,
+    /// Logical Cluster Number (LCN) to the beginning of the Master File Table (MFT).
+    pub(crate) mft_lcn: u64,
+    mft_mirror_lcn: u64,
+    pub(crate) file_record_size_info: i8,
+    zeros_4: [u8; 3],
+    index_record_size_info: i8,
+    zeros_5: [u8; 3],
+    pub(crate) serial_number: u64,
+    checksum: u32,
+}
+
+impl BiosParameterBlock {
+    /// Source: https://en.wikipedia.org/wiki/NTFS#Partition_Boot_Sector_(VBR)
+    pub(crate) fn record_size(size_info_value: i8, bytes_per_cluster: u32) -> Result<u32> {
+        if size_info_value > 0 {
+            // The size field denotes a cluster count.
+            Ok(size_info_value as u32 * bytes_per_cluster as u32)
+        } else {
+            // The size field denotes a binary exponent after negation.
+            let exponent = (-size_info_value) as u32;
+            if exponent > MAXIMUM_SIZE_INFO_EXPONENT {
+                return Err(NtfsError::InvalidRecordSizeExponent {
+                    expected: MAXIMUM_SIZE_INFO_EXPONENT,
+                    actual: exponent,
+                });
+            }
+
+            Ok(1 << exponent)
+        }
+    }
+}
+
+#[allow(unused)]
+#[derive(BinRead)]
+pub(crate) struct BootSector {
+    bootjmp: [u8; 3],
+    oem_name: [u8; 8],
+    pub(crate) bpb: BiosParameterBlock,
+    boot_code: [u8; 426],
+    signature: [u8; 2],
+}
+
+impl BootSector {
+    pub(crate) fn validate(&self) -> Result<()> {
+        // Validate the infamous [0x55, 0xAA] signature at the end of the boot sector.
+        let expected_signature = &[0x55, 0xAA];
+        if &self.signature != expected_signature {
+            return Err(NtfsError::InvalidTwoByteSignature {
+                position: offset_of!(BootSector, signature) as u64,
+                expected: expected_signature,
+                actual: self.signature,
+            });
+        }
+
+        Ok(())
+    }
+}
diff --git a/src/error.rs b/src/error.rs
new file mode 100644
index 0000000..ef85bb9
--- /dev/null
+++ b/src/error.rs
@@ -0,0 +1,73 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::attribute::NtfsAttributeType;
+use displaydoc::Display;
+
+/// Central result type of ntfs.
+pub type Result<T, E = NtfsError> = core::result::Result<T, E>;
+
+/// Central error type of ntfs.
+#[derive(Debug, Display)]
+pub enum NtfsError {
+    /// The given buffer should have at least {expected} bytes, but it only has {actual} bytes
+    BufferTooSmall { expected: usize, actual: usize },
+    /// The NTFS attribute at byte position {position:#010x} of type {ty:?} should have at least {expected} bytes, but it only has {actual} bytes
+    InvalidAttributeSize {
+        position: u64,
+        ty: NtfsAttributeType,
+        expected: u64,
+        actual: u64,
+    },
+    /// The requested NTFS file {n} is invalid
+    InvalidNtfsFile { n: u64 },
+    /// The NTFS file at byte position {position:#010x} should have signature {expected:?}, but it has signature {actual:?}
+    InvalidNtfsFileSignature {
+        position: u64,
+        expected: &'static [u8],
+        actual: [u8; 4],
+    },
+    /// The given time can't be represented as an NtfsTime
+    InvalidNtfsTime,
+    /// A record size field in the BIOS Parameter Block denotes the exponent {actual}, but the maximum valid one is {expected}
+    InvalidRecordSizeExponent { expected: u32, actual: u32 },
+    /// The 2-byte signature field at byte position {position:#010x} should contain {expected:?}, but it contains {actual:?}
+    InvalidTwoByteSignature {
+        position: u64,
+        expected: &'static [u8],
+        actual: [u8; 2],
+    },
+    /// I/O error: {0:?}
+    Io(binread::io::Error),
+    /// The cluster size is {actual} bytes, but the maximum supported one is {expected}
+    UnsupportedClusterSize { expected: u32, actual: u32 },
+    /// The type of the NTFS attribute at byte position {position:#010x} is {actual:#010x}, which is not supported
+    UnsupportedNtfsAttributeType { position: u64, actual: u32 },
+    /// The namespace of the NTFS file name starting at byte position {position:#010x} is {actual}, which is not supported
+    UnsupportedNtfsFileNamespace { position: u64, actual: u8 },
+    /// The NTFS attribute at byte position {position:#010x} has type {ty:?}, which cannot be read as a structured value
+    UnsupportedStructuredValue {
+        position: u64,
+        ty: NtfsAttributeType,
+    },
+}
+
+impl From<binread::error::Error> for NtfsError {
+    fn from(error: binread::error::Error) -> Self {
+        if let binread::error::Error::Io(io_error) = error {
+            Self::Io(io_error)
+        } else {
+            // We don't use any binread attributes that result in other errors.
+            unreachable!("Got a binread error of unexpected type: {:?}", error);
+        }
+    }
+}
+
+impl From<binread::io::Error> for NtfsError {
+    fn from(error: binread::io::Error) -> Self {
+        Self::Io(error)
+    }
+}
+
+#[cfg(feature = "std")]
+impl std::error::Error for NtfsError {}
diff --git a/src/helpers.rs b/src/helpers.rs
new file mode 100644
index 0000000..02abdcf
--- /dev/null
+++ b/src/helpers.rs
@@ -0,0 +1,26 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+macro_rules! iter_try {
+    ($e:expr) => {
+        match $e {
+            Ok(x) => x,
+            Err(e) => return Some(Err(e.into())),
+        }
+    };
+}
+
+#[cfg(test)]
+pub mod tests {
+    use std::fs::File;
+    use std::io::{Cursor, Read};
+
+    pub fn testfs1() -> Cursor<Vec<u8>> {
+        let mut buffer = Vec::new();
+        File::open("testdata/testfs1")
+            .unwrap()
+            .read_to_end(&mut buffer)
+            .unwrap();
+        Cursor::new(buffer)
+    }
+}
diff --git a/src/lib.rs b/src/lib.rs
new file mode 100644
index 0000000..673092b
--- /dev/null
+++ b/src/lib.rs
@@ -0,0 +1,28 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#![cfg_attr(not(feature = "std"), no_std)]
+#![forbid(unsafe_code)]
+
+extern crate alloc;
+
+#[macro_use]
+mod helpers;
+
+mod attribute;
+mod attribute_value;
+mod boot_sector;
+mod error;
+mod ntfs;
+mod ntfs_file;
+mod string;
+pub mod structured_values;
+mod time;
+
+pub use crate::attribute::*;
+pub use crate::attribute_value::*;
+pub use crate::error::*;
+pub use crate::ntfs::*;
+pub use crate::ntfs_file::*;
+pub use crate::string::*;
+pub use crate::time::*;
diff --git a/src/ntfs.rs b/src/ntfs.rs
new file mode 100644
index 0000000..b8998f3
--- /dev/null
+++ b/src/ntfs.rs
@@ -0,0 +1,129 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::boot_sector::{BiosParameterBlock, BootSector};
+//use crate::dir::Dir;
+use crate::error::{NtfsError, Result};
+use crate::ntfs_file::NtfsFile;
+use binread::io::{Read, Seek, SeekFrom};
+use binread::BinReaderExt;
+
+/// The maximum cluster size supported by Windows.
+/// Source: https://support.microsoft.com/en-us/topic/default-cluster-size-for-ntfs-fat-and-exfat-9772e6f1-e31a-00d7-e18f-73169155af95
+const MAXIMUM_CLUSTER_SIZE: u32 = 65536;
+
+pub struct Ntfs {
+    /// How many bytes a sector occupies. This is usually 512.
+    bytes_per_sector: u16,
+    /// How many sectors a cluster occupies. This is usually 8.
+    sectors_per_cluster: u8,
+    /// Size of the filesystem, in bytes.
+    size: u64,
+    /// Absolute position of the Master File Table (MFT), in bytes.
+    mft_position: u64,
+    /// Size of a single file record, in bytes.
+    pub(crate) file_record_size: u32,
+    /// Serial number of the NTFS volume.
+    serial_number: u64,
+}
+
+impl Ntfs {
+    pub fn new<T>(fs: &mut T) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        // Read and validate the boot sector.
+        fs.seek(SeekFrom::Start(0))?;
+        let boot_sector = fs.read_le::<BootSector>()?;
+        boot_sector.validate()?;
+
+        let bytes_per_sector = boot_sector.bpb.bytes_per_sector;
+        let sectors_per_cluster = boot_sector.bpb.sectors_per_cluster;
+        let bytes_per_cluster = sectors_per_cluster as u32 * bytes_per_sector as u32;
+        if bytes_per_cluster > MAXIMUM_CLUSTER_SIZE {
+            return Err(NtfsError::UnsupportedClusterSize {
+                expected: MAXIMUM_CLUSTER_SIZE,
+                actual: bytes_per_cluster,
+            });
+        }
+
+        let size = boot_sector.bpb.total_sectors * bytes_per_sector as u64;
+        let mft_position = boot_sector.bpb.mft_lcn * bytes_per_cluster as u64;
+        let file_record_size = BiosParameterBlock::record_size(
+            boot_sector.bpb.file_record_size_info,
+            bytes_per_cluster,
+        )?;
+        let serial_number = boot_sector.bpb.serial_number;
+
+        Ok(Self {
+            bytes_per_sector,
+            sectors_per_cluster,
+            size,
+            mft_position,
+            file_record_size,
+            serial_number,
+        })
+    }
+
+    /// Returns the size of a single cluster, in bytes.
+    pub fn cluster_size(&self) -> u16 {
+        self.bytes_per_sector * self.sectors_per_cluster as u16
+    }
+
+    /// Returns the [`NtfsFile`] for the `n`-th NTFS file record.
+    ///
+    /// The first few NTFS files have fixed indexes and contain filesystem
+    /// management information (see the [`KnownNtfsFile`] enum).
+    ///
+    /// TODO:
+    /// - Check if `n` can be u32 instead of u64.
+    /// - Check if `n` should be in a newtype, with easier conversion from
+    ///   KnownNtfsFile.
+    pub fn ntfs_file<T>(&self, fs: &mut T, n: u64) -> Result<NtfsFile>
+    where
+        T: Read + Seek,
+    {
+        let offset = n
+            .checked_mul(self.file_record_size as u64)
+            .ok_or(NtfsError::InvalidNtfsFile { n })?;
+        let position = self
+            .mft_position
+            .checked_add(offset)
+            .ok_or(NtfsError::InvalidNtfsFile { n })?;
+        NtfsFile::new(fs, position)
+    }
+
+    /// Returns the root [`Dir`] of this NTFS volume.
+    pub fn root_dir(&self) -> ! {
+        panic!("TODO")
+    }
+
+    /// Returns the size of a single sector in bytes.
+    pub fn sector_size(&self) -> u16 {
+        self.bytes_per_sector
+    }
+
+    /// Returns the 64-bit serial number of this NTFS volume.
+    pub fn serial_number(&self) -> u64 {
+        self.serial_number
+    }
+
+    /// Returns the partition size in bytes.
+    pub fn size(&self) -> u64 {
+        self.size
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_ntfs() {
+        let mut testfs1 = crate::helpers::tests::testfs1();
+        let ntfs = Ntfs::new(&mut testfs1).unwrap();
+        assert_eq!(ntfs.cluster_size(), 512);
+        assert_eq!(ntfs.sector_size(), 512);
+        assert_eq!(ntfs.size(), 1049088);
+    }
+}
diff --git a/src/ntfs_file.rs b/src/ntfs_file.rs
new file mode 100644
index 0000000..e663848
--- /dev/null
+++ b/src/ntfs_file.rs
@@ -0,0 +1,98 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::attribute::NtfsAttributes;
+use crate::error::{NtfsError, Result};
+use binread::io::{Read, Seek, SeekFrom};
+use binread::{BinRead, BinReaderExt};
+use bitflags::bitflags;
+
+#[repr(u64)]
+pub enum KnownNtfsFile {
+    MFT = 0,
+    MFTMirr = 1,
+    LogFile = 2,
+    Volume = 3,
+    AttrDef = 4,
+    RootDirectory = 5,
+    Bitmap = 6,
+    Boot = 7,
+    BadClus = 8,
+    Secure = 9,
+    UpCase = 10,
+    Extend = 11,
+}
+
+#[allow(unused)]
+#[derive(BinRead)]
+struct NtfsRecordHeader {
+    signature: [u8; 4],
+    update_sequence_array_offset: u16,
+    update_sequence_array_size: u16,
+    logfile_sequence_number: u64,
+}
+
+#[allow(unused)]
+#[derive(BinRead)]
+pub(crate) struct NtfsFileRecordHeader {
+    record_header: NtfsRecordHeader,
+    sequence_number: u16,
+    hard_link_count: u16,
+    pub(crate) first_attribute_offset: u16,
+    flags: u16,
+    pub(crate) used_size: u32,
+    allocated_size: u32,
+    base_file_record: u64,
+    next_attribute_number: u16,
+}
+
+bitflags! {
+    struct NtfsFileRecordFlags: u16 {
+        /// Record is in use.
+        const IN_USE = 0x0001;
+        /// Record is a directory.
+        const IS_DIRECTORY = 0x0002;
+    }
+}
+
+pub struct NtfsFile {
+    pub(crate) header: NtfsFileRecordHeader,
+    pub(crate) position: u64,
+}
+
+impl NtfsFile {
+    pub(crate) fn new<T>(fs: &mut T, position: u64) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        fs.seek(SeekFrom::Start(position))?;
+        let header = fs.read_le::<NtfsFileRecordHeader>()?;
+
+        let file = Self { header, position };
+        file.validate_signature()?;
+
+        Ok(file)
+    }
+
+    pub fn attributes<'a, T>(&self, fs: &'a mut T) -> NtfsAttributes<'a, T>
+    where
+        T: Read + Seek,
+    {
+        NtfsAttributes::new(fs, &self)
+    }
+
+    fn validate_signature(&self) -> Result<()> {
+        let signature = &self.header.record_header.signature;
+        let expected = b"FILE";
+
+        if signature == expected {
+            Ok(())
+        } else {
+            Err(NtfsError::InvalidNtfsFileSignature {
+                position: self.position,
+                expected,
+                actual: *signature,
+            })
+        }
+    }
+}
diff --git a/src/string.rs b/src/string.rs
new file mode 100644
index 0000000..18acc4d
--- /dev/null
+++ b/src/string.rs
@@ -0,0 +1,163 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use alloc::string::String;
+use core::char;
+use core::cmp::Ordering;
+use core::convert::TryInto;
+use core::fmt;
+
+/// Zero-copy representation of a string stored in an NTFS filesystem structure.
+#[derive(Clone, Debug, Eq)]
+pub struct NtfsString<'a>(pub &'a [u8]);
+
+impl<'a> NtfsString<'a> {
+    fn cmp_iter<TI, OI>(mut this_iter: TI, mut other_iter: OI) -> Ordering
+    where
+        TI: Iterator<Item = u16>,
+        OI: Iterator<Item = u16>,
+    {
+        loop {
+            match (this_iter.next(), other_iter.next()) {
+                (Some(this_code_unit), Some(other_code_unit)) => {
+                    // We have two UTF-16 code units to compare.
+                    if this_code_unit != other_code_unit {
+                        return this_code_unit.cmp(&other_code_unit);
+                    }
+                }
+                (Some(_), None) => {
+                    // `this_iter` is longer than `other_iter` but otherwise equal.
+                    return Ordering::Greater;
+                }
+                (None, Some(_)) => {
+                    // `other_iter` is longer than `this_iter` but otherwise equal.
+                    return Ordering::Less;
+                }
+                (None, None) => {
+                    // We made it to the end of both strings, so they must be equal.
+                    return Ordering::Equal;
+                }
+            }
+        }
+    }
+
+    fn cmp_str(&self, other: &str) -> Ordering {
+        let other_iter = other.encode_utf16();
+        Self::cmp_iter(self.utf16le_iter(), other_iter)
+    }
+
+    fn utf16le_iter(&'a self) -> impl Iterator<Item = u16> + 'a {
+        self.0
+            .chunks_exact(2)
+            .map(|two_bytes| u16::from_le_bytes(two_bytes.try_into().unwrap()))
+    }
+
+    /// Returns `true` if `self` has a length of zero bytes.
+    pub const fn is_empty(&self) -> bool {
+        self.len() == 0
+    }
+
+    /// Returns the length of `self`.
+    ///
+    /// This length is in bytes, not characters! In other words,
+    /// it may not be what a human considers the length of the string.
+    pub const fn len(&self) -> usize {
+        self.0.len()
+    }
+
+    /// Attempts to convert `self` to an owned `String`.
+    /// Returns `Some(String)` if all characters could be converted successfully or `None` if a decoding error occurred.
+    pub fn to_string_checked(&self) -> Option<String> {
+        char::decode_utf16(self.utf16le_iter())
+            .map(|x| x.ok())
+            .collect::<Option<String>>()
+    }
+
+    /// Converts `self` to an owned `String`, replacing invalid data with the replacement character (U+FFFD).
+    pub fn to_string_lossy(&self) -> String {
+        char::decode_utf16(self.utf16le_iter())
+            .map(|x| x.unwrap_or(char::REPLACEMENT_CHARACTER))
+            .collect()
+    }
+}
+
+impl<'a> fmt::Display for NtfsString<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let utf16_iter = char::decode_utf16(self.utf16le_iter())
+            .map(|x| x.unwrap_or(char::REPLACEMENT_CHARACTER));
+
+        for single_char in utf16_iter {
+            single_char.fmt(f)?;
+        }
+
+        Ok(())
+    }
+}
+
+impl<'a> Ord for NtfsString<'a> {
+    fn cmp(&self, other: &Self) -> Ordering {
+        Self::cmp_iter(self.utf16le_iter(), other.utf16le_iter())
+    }
+}
+
+impl<'a> PartialEq for NtfsString<'a> {
+    /// Checks that two strings are a (case-sensitive!) match.
+    fn eq(&self, other: &Self) -> bool {
+        let ordering = self.cmp(other);
+        ordering == Ordering::Equal
+    }
+}
+
+impl<'a> PartialEq<str> for NtfsString<'a> {
+    fn eq(&self, other: &str) -> bool {
+        self.cmp_str(other) == Ordering::Equal
+    }
+}
+
+impl<'a> PartialEq<NtfsString<'a>> for str {
+    fn eq(&self, other: &NtfsString<'a>) -> bool {
+        other.cmp_str(self) == Ordering::Equal
+    }
+}
+
+impl<'a> PartialEq<&str> for NtfsString<'a> {
+    fn eq(&self, other: &&str) -> bool {
+        self.cmp_str(other) == Ordering::Equal
+    }
+}
+
+impl<'a> PartialEq<NtfsString<'a>> for &str {
+    fn eq(&self, other: &NtfsString<'a>) -> bool {
+        other.cmp_str(self) == Ordering::Equal
+    }
+}
+
+impl<'a> PartialOrd for NtfsString<'a> {
+    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
+        Some(self.cmp(other))
+    }
+}
+
+impl<'a> PartialOrd<str> for NtfsString<'a> {
+    fn partial_cmp(&self, other: &str) -> Option<Ordering> {
+        Some(self.cmp_str(other))
+    }
+}
+
+impl<'a> PartialOrd<NtfsString<'a>> for str {
+    fn partial_cmp(&self, other: &NtfsString<'a>) -> Option<Ordering> {
+        Some(other.cmp_str(self))
+    }
+}
+
+impl<'a> PartialOrd<&str> for NtfsString<'a> {
+    fn partial_cmp(&self, other: &&str) -> Option<Ordering> {
+        Some(self.cmp_str(other))
+    }
+}
+
+impl<'a> PartialOrd<NtfsString<'a>> for &str {
+    fn partial_cmp(&self, other: &NtfsString<'a>) -> Option<Ordering> {
+        Some(other.cmp_str(self))
+    }
+}
diff --git a/src/structured_values/attribute_list.rs b/src/structured_values/attribute_list.rs
new file mode 100644
index 0000000..53769a7
--- /dev/null
+++ b/src/structured_values/attribute_list.rs
@@ -0,0 +1,2 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
diff --git a/src/structured_values/file_name.rs b/src/structured_values/file_name.rs
new file mode 100644
index 0000000..19f50cc
--- /dev/null
+++ b/src/structured_values/file_name.rs
@@ -0,0 +1,142 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::attribute::NtfsAttributeType;
+use crate::attribute_value::NtfsAttributeValueAttached;
+use crate::error::{NtfsError, Result};
+use crate::string::NtfsString;
+use crate::structured_values::NtfsFileAttributeFlags;
+use crate::time::NtfsTime;
+use binread::io::{Read, Seek, SeekFrom};
+use binread::{BinRead, BinReaderExt};
+use core::mem;
+use enumn::N;
+
+/// Size of all [`FileNameHeader`] fields.
+const FILE_NAME_HEADER_SIZE: u64 = 66;
+
+/// The smallest FileName attribute has a name containing just a single character.
+const FILE_NAME_MIN_SIZE: u64 = FILE_NAME_HEADER_SIZE + mem::size_of::<u16>() as u64;
+
+#[allow(unused)]
+#[derive(BinRead, Clone, Debug)]
+struct FileNameHeader {
+    parent_directory_ref: u64,
+    creation_time: NtfsTime,
+    modification_time: NtfsTime,
+    mft_record_modification_time: NtfsTime,
+    access_time: NtfsTime,
+    allocated_size: u64,
+    data_size: u64,
+    file_attributes: u32,
+    reparse_point_tag: u32,
+    name_length: u8,
+    namespace: u8,
+}
+
+#[derive(Clone, Copy, Debug, Eq, N, PartialEq)]
+#[repr(u8)]
+pub enum NtfsFileNamespace {
+    Posix = 0,
+    Win32 = 1,
+    Dos = 2,
+    Win32AndDos = 3,
+}
+
+#[derive(Clone, Debug)]
+pub struct NtfsFileName {
+    header: FileNameHeader,
+    name_position: u64,
+}
+
+impl NtfsFileName {
+    pub(crate) fn new<T>(
+        attribute_position: u64,
+        mut value_attached: NtfsAttributeValueAttached<'_, T>,
+        value_length: u64,
+    ) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        if value_length < FILE_NAME_MIN_SIZE {
+            return Err(NtfsError::InvalidAttributeSize {
+                position: attribute_position,
+                ty: NtfsAttributeType::FileName,
+                expected: FILE_NAME_MIN_SIZE,
+                actual: value_length,
+            });
+        }
+
+        let header = value_attached.read_le::<FileNameHeader>()?;
+        let name_position = value_attached.position() + FILE_NAME_HEADER_SIZE;
+
+        Ok(Self {
+            header,
+            name_position,
+        })
+    }
+
+    pub fn access_time(&self) -> NtfsTime {
+        self.header.access_time
+    }
+
+    pub fn allocated_size(&self) -> u64 {
+        self.header.allocated_size
+    }
+
+    pub fn creation_time(&self) -> NtfsTime {
+        self.header.creation_time
+    }
+
+    pub fn data_size(&self) -> u64 {
+        self.header.data_size
+    }
+
+    pub fn file_attributes(&self) -> NtfsFileAttributeFlags {
+        NtfsFileAttributeFlags::from_bits_truncate(self.header.file_attributes)
+    }
+
+    pub fn mft_record_modification_time(&self) -> NtfsTime {
+        self.header.mft_record_modification_time
+    }
+
+    pub fn modification_time(&self) -> NtfsTime {
+        self.header.modification_time
+    }
+
+    /// Returns the file name length, in bytes.
+    ///
+    /// A file name has a maximum length of 255 UTF-16 code points (510 bytes).
+    pub fn name_length(&self) -> usize {
+        self.header.name_length as usize * mem::size_of::<u16>()
+    }
+
+    /// Returns the namespace this name belongs to, or [`NtfsError::UnsupportedNtfsFileNamespace`]
+    /// if it's an unknown namespace.
+    pub fn namespace(&self) -> Result<NtfsFileNamespace> {
+        NtfsFileNamespace::n(self.header.namespace).ok_or(NtfsError::UnsupportedNtfsFileNamespace {
+            position: self.name_position,
+            actual: self.header.namespace,
+        })
+    }
+
+    /// Reads the file name into the given buffer, and returns an
+    /// [`NtfsString`] wrapping that buffer.
+    pub fn read_name<'a, T>(&self, fs: &mut T, buf: &'a mut [u8]) -> Result<NtfsString<'a>>
+    where
+        T: Read + Seek,
+    {
+        let name_length = self.name_length();
+        if buf.len() < name_length {
+            return Err(NtfsError::BufferTooSmall {
+                expected: name_length,
+                actual: buf.len(),
+            });
+        }
+
+        fs.seek(SeekFrom::Start(self.name_position))?;
+        fs.read_exact(&mut buf[..name_length])?;
+
+        Ok(NtfsString(&buf[..name_length]))
+    }
+}
diff --git a/src/structured_values/mod.rs b/src/structured_values/mod.rs
new file mode 100644
index 0000000..df4aa90
--- /dev/null
+++ b/src/structured_values/mod.rs
@@ -0,0 +1,44 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+mod attribute_list;
+mod file_name;
+mod object_id;
+mod security_descriptor;
+mod standard_information;
+mod volume_information;
+mod volume_name;
+
+pub use attribute_list::*;
+pub use file_name::*;
+pub use object_id::*;
+pub use security_descriptor::*;
+pub use standard_information::*;
+pub use volume_information::*;
+pub use volume_name::*;
+
+use bitflags::bitflags;
+
+bitflags! {
+    pub struct NtfsFileAttributeFlags: u32 {
+        const READ_ONLY = 0x0001;
+        const HIDDEN = 0x0002;
+        const SYSTEM = 0x0004;
+        const ARCHIVE = 0x0020;
+        const DEVICE = 0x0040;
+        const NORMAL = 0x0080;
+        const TEMPORARY = 0x0100;
+        const SPARSE_FILE = 0x0200;
+        const REPARSE_POINT = 0x0400;
+        const COMPRESSED = 0x0800;
+        const OFFLINE = 0x1000;
+        const NOT_CONTENT_INDEXED = 0x2000;
+        const ENCRYPTED = 0x4000;
+    }
+}
+
+#[derive(Clone, Debug)]
+pub enum NtfsStructuredValue {
+    StandardInformation(NtfsStandardInformation),
+    FileName(NtfsFileName),
+}
diff --git a/src/structured_values/object_id.rs b/src/structured_values/object_id.rs
new file mode 100644
index 0000000..53769a7
--- /dev/null
+++ b/src/structured_values/object_id.rs
@@ -0,0 +1,2 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
diff --git a/src/structured_values/security_descriptor.rs b/src/structured_values/security_descriptor.rs
new file mode 100644
index 0000000..53769a7
--- /dev/null
+++ b/src/structured_values/security_descriptor.rs
@@ -0,0 +1,2 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
diff --git a/src/structured_values/standard_information.rs b/src/structured_values/standard_information.rs
new file mode 100644
index 0000000..f566278
--- /dev/null
+++ b/src/structured_values/standard_information.rs
@@ -0,0 +1,119 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use crate::attribute::NtfsAttributeType;
+use crate::attribute_value::NtfsAttributeValueAttached;
+use crate::error::{NtfsError, Result};
+use crate::structured_values::NtfsFileAttributeFlags;
+use crate::time::NtfsTime;
+use binread::io::{Read, Seek};
+use binread::{BinRead, BinReaderExt};
+
+/// Size of all [`StandardInformationData`] fields plus some reserved bytes.
+const STANDARD_INFORMATION_SIZE_NTFS1: u64 = 48;
+
+/// Size of all [`StandardInformationData`] plus [`StandardInformationDataNtfs3`] fields.
+const STANDARD_INFORMATION_SIZE_NTFS3: u64 = 72;
+
+#[derive(BinRead, Clone, Debug)]
+struct StandardInformationData {
+    creation_time: NtfsTime,
+    modification_time: NtfsTime,
+    mft_record_modification_time: NtfsTime,
+    access_time: NtfsTime,
+    file_attributes: u32,
+}
+
+#[derive(BinRead, Clone, Debug)]
+struct StandardInformationDataNtfs3 {
+    maximum_versions: u32,
+    version: u32,
+    class_id: u32,
+    owner_id: u32,
+    security_id: u32,
+    quota_charged: u64,
+    usn: u64,
+}
+
+#[derive(Clone, Debug)]
+pub struct NtfsStandardInformation {
+    data: StandardInformationData,
+    ntfs3_data: Option<StandardInformationDataNtfs3>,
+}
+
+impl NtfsStandardInformation {
+    pub(crate) fn new<T>(
+        attribute_position: u64,
+        mut value_attached: NtfsAttributeValueAttached<'_, T>,
+        value_length: u64,
+    ) -> Result<Self>
+    where
+        T: Read + Seek,
+    {
+        if value_length < STANDARD_INFORMATION_SIZE_NTFS1 {
+            return Err(NtfsError::InvalidAttributeSize {
+                position: attribute_position,
+                ty: NtfsAttributeType::StandardInformation,
+                expected: STANDARD_INFORMATION_SIZE_NTFS1,
+                actual: value_length,
+            });
+        }
+
+        let data = value_attached.read_le::<StandardInformationData>()?;
+        let ntfs3_data = if value_length >= STANDARD_INFORMATION_SIZE_NTFS3 as u64 {
+            Some(value_attached.read_le::<StandardInformationDataNtfs3>()?)
+        } else {
+            None
+        };
+
+        Ok(Self { data, ntfs3_data })
+    }
+
+    pub fn access_time(&self) -> NtfsTime {
+        self.data.access_time
+    }
+
+    pub fn class_id(&self) -> Option<u32> {
+        self.ntfs3_data.as_ref().map(|x| x.class_id)
+    }
+
+    pub fn creation_time(&self) -> NtfsTime {
+        self.data.creation_time
+    }
+
+    pub fn file_attributes(&self) -> NtfsFileAttributeFlags {
+        NtfsFileAttributeFlags::from_bits_truncate(self.data.file_attributes)
+    }
+
+    pub fn maximum_versions(&self) -> Option<u32> {
+        self.ntfs3_data.as_ref().map(|x| x.maximum_versions)
+    }
+
+    pub fn mft_record_modification_time(&self) -> NtfsTime {
+        self.data.mft_record_modification_time
+    }
+
+    pub fn modification_time(&self) -> NtfsTime {
+        self.data.modification_time
+    }
+
+    pub fn owner_id(&self) -> Option<u32> {
+        self.ntfs3_data.as_ref().map(|x| x.owner_id)
+    }
+
+    pub fn quota_charged(&self) -> Option<u64> {
+        self.ntfs3_data.as_ref().map(|x| x.quota_charged)
+    }
+
+    pub fn security_id(&self) -> Option<u32> {
+        self.ntfs3_data.as_ref().map(|x| x.security_id)
+    }
+
+    pub fn usn(&self) -> Option<u64> {
+        self.ntfs3_data.as_ref().map(|x| x.usn)
+    }
+
+    pub fn version(&self) -> Option<u32> {
+        self.ntfs3_data.as_ref().map(|x| x.version)
+    }
+}
diff --git a/src/structured_values/volume_information.rs b/src/structured_values/volume_information.rs
new file mode 100644
index 0000000..53769a7
--- /dev/null
+++ b/src/structured_values/volume_information.rs
@@ -0,0 +1,2 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
diff --git a/src/structured_values/volume_name.rs b/src/structured_values/volume_name.rs
new file mode 100644
index 0000000..53769a7
--- /dev/null
+++ b/src/structured_values/volume_name.rs
@@ -0,0 +1,2 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
diff --git a/src/time.rs b/src/time.rs
new file mode 100644
index 0000000..ea351da
--- /dev/null
+++ b/src/time.rs
@@ -0,0 +1,162 @@
+// Copyright 2021 Colin Finck <colin@reactos.org>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+use binread::BinRead;
+use core::ops::Deref;
+
+#[cfg(any(feature = "chrono", feature = "std"))]
+use core::convert::TryFrom;
+
+#[cfg(feature = "chrono")]
+use {
+    crate::error::NtfsError,
+    chrono::{DateTime, Datelike, NaiveDate, Timelike, Utc},
+};
+
+#[cfg(feature = "std")]
+use std::time::{SystemTime, SystemTimeError};
+
+/// How many days we have between 0001-01-01 and 1601-01-01.
+#[cfg(feature = "chrono")]
+const DAYS_FROM_0001_TO_1601: i32 = 584389;
+
+/// Difference in 100-nanosecond intervals between the Windows/NTFS epoch (1601-01-01) and the Unix epoch (1970-01-01).
+#[cfg(feature = "std")]
+const EPOCH_DIFFERENCE_IN_INTERVALS: i64 = 116_444_736_000_000_000;
+
+/// How many 100-nanosecond intervals we have in a second.
+#[cfg(any(feature = "chrono", feature = "std"))]
+const INTERVALS_PER_SECOND: u64 = 10_000_000;
+
+/// How many 100-nanosecond intervals we have in a day.
+#[cfg(feature = "chrono")]
+const INTERVALS_PER_DAY: u64 = 24 * 60 * 60 * INTERVALS_PER_SECOND;
+
+#[derive(BinRead, Clone, Copy, Debug, Eq, Ord, PartialEq, PartialOrd)]
+pub struct NtfsTime(u64);
+
+impl Deref for NtfsTime {
+    type Target = u64;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+#[cfg(feature = "chrono")]
+impl TryFrom<DateTime<Utc>> for NtfsTime {
+    type Error = NtfsError;
+
+    fn try_from(dt: DateTime<Utc>) -> Result<Self, Self::Error> {
+        // First do the time calculations, which safely fit into a u64.
+        let mut intervals = dt.hour() as u64;
+
+        intervals *= 60;
+        intervals += dt.minute() as u64;
+
+        intervals *= 60;
+        intervals += dt.second() as u64;
+
+        intervals *= INTERVALS_PER_SECOND;
+        intervals += dt.nanosecond() as u64 / 100;
+
+        // Now do checked arithmetics for the day calculations, which may
+        // exceed the lower bounds (years before 1601) or upper bounds
+        // (dates after approximately 28 May 60056).
+        let num_days_from_ce = dt.num_days_from_ce();
+        let num_days_from_1601 = num_days_from_ce
+            .checked_sub(DAYS_FROM_0001_TO_1601)
+            .ok_or(NtfsError::InvalidNtfsTime)?;
+        let intervals_days = INTERVALS_PER_DAY
+            .checked_mul(num_days_from_1601 as u64)
+            .ok_or(NtfsError::InvalidNtfsTime)?;
+        intervals = intervals
+            .checked_add(intervals_days)
+            .ok_or(NtfsError::InvalidNtfsTime)?;
+
+        Ok(Self(intervals))
+    }
+}
+
+#[cfg(feature = "chrono")]
+impl From<NtfsTime> for DateTime<Utc> {
+    fn from(nt: NtfsTime) -> DateTime<Utc> {
+        let mut remainder = *nt;
+
+        let nano = (remainder % INTERVALS_PER_SECOND) as u32 * 100;
+        remainder /= INTERVALS_PER_SECOND;
+
+        let sec = (remainder % 60) as u32;
+        remainder /= 60;
+
+        let min = (remainder % 60) as u32;
+        remainder /= 60;
+
+        let hour = (remainder % 24) as u32;
+        remainder /= 24;
+
+        let num_days_from_1601 = remainder as i32;
+        let num_days_from_ce = num_days_from_1601 + DAYS_FROM_0001_TO_1601;
+
+        let ndt =
+            NaiveDate::from_num_days_from_ce(num_days_from_ce).and_hms_nano(hour, min, sec, nano);
+        DateTime::<Utc>::from_utc(ndt, Utc)
+    }
+}
+
+#[cfg(feature = "std")]
+impl TryFrom<SystemTime> for NtfsTime {
+    type Error = SystemTimeError;
+
+    fn try_from(st: SystemTime) -> Result<Self, Self::Error> {
+        let duration_since_unix_epoch = st.duration_since(SystemTime::UNIX_EPOCH)?;
+        let intervals_since_unix_epoch = duration_since_unix_epoch.as_secs() as u64
+            * INTERVALS_PER_SECOND
+            + duration_since_unix_epoch.subsec_nanos() as u64 / 100;
+        let intervals_since_windows_epoch =
+            intervals_since_unix_epoch + EPOCH_DIFFERENCE_IN_INTERVALS as u64;
+        Ok(Self(intervals_since_windows_epoch))
+    }
+}
+
+#[cfg(test)]
+pub(crate) mod tests {
+    use super::*;
+
+    #[cfg(feature = "chrono")]
+    use chrono::TimeZone;
+
+    pub(crate) const NT_TIMESTAMP_2021_01_01: u64 = 132539328000000000u64;
+
+    #[cfg(feature = "chrono")]
+    #[test]
+    fn test_chrono() {
+        let dt = Utc.ymd(2013, 1, 5).and_hms(18, 15, 00);
+        let nt = NtfsTime::try_from(dt).unwrap();
+        assert_eq!(*nt, 130018833000000000u64);
+
+        let dt2 = DateTime::<Utc>::from(nt);
+        assert_eq!(dt, dt2);
+
+        let dt = Utc.ymd(1601, 1, 1).and_hms(0, 0, 0);
+        let nt = NtfsTime::try_from(dt).unwrap();
+        assert_eq!(*nt, 0u64);
+
+        let dt = Utc.ymd(1600, 12, 31).and_hms(23, 59, 59);
+        assert!(NtfsTime::try_from(dt).is_err());
+
+        let dt = Utc.ymd(60056, 5, 28).and_hms(0, 0, 0);
+        assert!(NtfsTime::try_from(dt).is_ok());
+
+        let dt = Utc.ymd(60056, 5, 29).and_hms(0, 0, 0);
+        assert!(NtfsTime::try_from(dt).is_err());
+    }
+
+    #[cfg(feature = "std")]
+    #[test]
+    fn test_systemtime() {
+        let st = SystemTime::now();
+        let nt = NtfsTime::try_from(st).unwrap();
+        assert!(*nt > NT_TIMESTAMP_2021_01_01);
+    }
+}
diff --git a/testdata/create-testfs1.sh b/testdata/create-testfs1.sh
new file mode 100644
index 0000000..65dac63
--- /dev/null
+++ b/testdata/create-testfs1.sh
@@ -0,0 +1,25 @@
+#!/bin/bash
+set -eu
+
+if [ "`whoami`" != "root" ]; then
+    echo Needs to be run as root!
+    exit 1
+fi
+
+dd if=/dev/zero of=testfs1 bs=1k count=1025
+mkntfs -c 512 -L mylabel -F testfs1
+
+mkdir mnt
+mount -t ntfs-3g -o loop testfs1 mnt
+cd mnt
+
+touch -m -t 202101011337 empty-file
+dd if=/dev/zero of=file-with-5-zeros bs=1 count=5
+dd if=/dev/zero of=big-sparse-file skip=5M bs=1 count=1
+
+mkdir -p subdir/subsubdir
+echo abcdef > subdir/subsubdir/file-with-6-letters
+
+cd ..
+umount mnt
+rmdir mnt
diff --git a/testdata/testfs1 b/testdata/testfs1
new file mode 100644
index 0000000..4afb44a
--- /dev/null
+++ b/testdata/testfs1