blob: f18a38a8a0951d54ec53c299d2692f1fd7723fca [file] [log] [blame]
Jooyung Han5d94bfc2021-08-06 14:07:49 +09001/*
2 * Copyright (C) 2021 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17//! Utilities for zip handling
18
19use anyhow::{bail, Result};
Jooyung Hand8397852021-08-10 16:29:36 +090020use bytes::{Buf, BufMut};
Jooyung Han5d94bfc2021-08-06 14:07:49 +090021use std::io::{Read, Seek, SeekFrom};
22use zip::ZipArchive;
23
24const EOCD_MIN_SIZE: usize = 22;
Jooyung Hancee6de62021-08-11 15:52:07 +090025const EOCD_CENTRAL_DIRECTORY_SIZE_FIELD_OFFSET: usize = 12;
Jooyung Han5d94bfc2021-08-06 14:07:49 +090026const EOCD_CENTRAL_DIRECTORY_OFFSET_FIELD_OFFSET: usize = 16;
27const EOCD_MAGIC: u32 = 0x06054b50;
Jooyung Hancee6de62021-08-11 15:52:07 +090028const ZIP64_MARK: u32 = 0xffffffff;
Jooyung Han5d94bfc2021-08-06 14:07:49 +090029
30#[derive(Debug, PartialEq)]
31pub struct ZipSections {
32 pub central_directory_offset: u32,
33 pub central_directory_size: u32,
34 pub eocd_offset: u32,
35 pub eocd_size: u32,
36}
37
38/// Discover the layout of a zip file.
39pub fn zip_sections<R: Read + Seek>(mut reader: R) -> Result<(R, ZipSections)> {
40 // open a zip to parse EOCD
41 let archive = ZipArchive::new(reader)?;
42 let eocd_size = archive.comment().len() + EOCD_MIN_SIZE;
43 if archive.offset() != 0 {
44 bail!("Invalid ZIP: offset should be 0, but {}.", archive.offset());
45 }
46 // retrieve reader back
47 reader = archive.into_inner();
48 // the current position should point EOCD offset
Jooyung Hancee6de62021-08-11 15:52:07 +090049 let eocd_offset = reader.seek(SeekFrom::Current(0))? as u32;
Jooyung Han5d94bfc2021-08-06 14:07:49 +090050 let mut eocd = vec![0u8; eocd_size as usize];
51 reader.read_exact(&mut eocd)?;
52 if (&eocd[0..]).get_u32_le() != EOCD_MAGIC {
53 bail!("Invalid ZIP: ZipArchive::new() should point EOCD after reading.");
54 }
Jooyung Hancee6de62021-08-11 15:52:07 +090055 let (central_directory_size, central_directory_offset) = get_central_directory(&eocd)?;
56 if central_directory_offset == ZIP64_MARK || central_directory_size == ZIP64_MARK {
57 bail!("Unsupported ZIP: ZIP64 is not supported.");
58 }
59 if central_directory_offset + central_directory_size != eocd_offset {
60 bail!("Invalid ZIP: EOCD should follow CD with no extra data or overlap.");
61 }
62
Jooyung Han5d94bfc2021-08-06 14:07:49 +090063 Ok((
64 reader,
65 ZipSections {
66 central_directory_offset,
67 central_directory_size,
Jooyung Hancee6de62021-08-11 15:52:07 +090068 eocd_offset,
Jooyung Han5d94bfc2021-08-06 14:07:49 +090069 eocd_size: eocd_size as u32,
70 },
71 ))
72}
73
Jooyung Hancee6de62021-08-11 15:52:07 +090074fn get_central_directory(buf: &[u8]) -> Result<(u32, u32)> {
Jooyung Han5d94bfc2021-08-06 14:07:49 +090075 if buf.len() < EOCD_MIN_SIZE {
76 bail!("Invalid EOCD size: {}", buf.len());
77 }
Jooyung Hancee6de62021-08-11 15:52:07 +090078 let mut buf = &buf[EOCD_CENTRAL_DIRECTORY_SIZE_FIELD_OFFSET..];
79 let size = buf.get_u32_le();
80 let offset = buf.get_u32_le();
81 Ok((size, offset))
Jooyung Han5d94bfc2021-08-06 14:07:49 +090082}
Jooyung Hand8397852021-08-10 16:29:36 +090083
84/// Update EOCD's central_directory_offset field.
85pub fn set_central_directory_offset(buf: &mut [u8], value: u32) -> Result<()> {
86 if buf.len() < EOCD_MIN_SIZE {
87 bail!("Invalid EOCD size: {}", buf.len());
88 }
89 (&mut buf[EOCD_CENTRAL_DIRECTORY_OFFSET_FIELD_OFFSET..]).put_u32_le(value);
90 Ok(())
91}
Jooyung Hancee6de62021-08-11 15:52:07 +090092
93#[cfg(test)]
94mod tests {
95 use super::*;
Andrew Walbran117cd5e2021-08-13 11:42:13 +000096 use crate::testing::assert_contains;
Jooyung Hancee6de62021-08-11 15:52:07 +090097 use std::io::{Cursor, Write};
98 use zip::{write::FileOptions, ZipWriter};
99
100 fn create_test_zip() -> Cursor<Vec<u8>> {
101 let mut writer = ZipWriter::new(Cursor::new(Vec::new()));
102 writer.start_file("testfile", FileOptions::default()).unwrap();
103 writer.write_all(b"testcontent").unwrap();
104 writer.finish().unwrap()
105 }
106
107 #[test]
108 fn test_zip_sections() {
109 let (cursor, sections) = zip_sections(create_test_zip()).unwrap();
110 assert_eq!(sections.eocd_offset, (cursor.get_ref().len() - EOCD_MIN_SIZE) as u32);
111 }
112
113 #[test]
114 fn test_reject_if_extra_data_between_cd_and_eocd() {
115 // prepare normal zip
116 let buf = create_test_zip().into_inner();
117
118 // insert garbage between CD and EOCD.
119 // by the way, to mock zip-rs, use CD as garbage. This is implementation detail of zip-rs,
120 // which reads CD at (eocd_offset - cd_size) instead of at cd_offset from EOCD.
121 let (pre_eocd, eocd) = buf.split_at(buf.len() - EOCD_MIN_SIZE);
122 let (_, cd_offset) = get_central_directory(eocd).unwrap();
123 let cd = &pre_eocd[cd_offset as usize..];
124
125 // ZipArchive::new() succeeds, but we should reject
126 let res = zip_sections(Cursor::new([pre_eocd, cd, eocd].concat()));
127 assert!(res.is_err());
Andrew Walbran117cd5e2021-08-13 11:42:13 +0000128 assert_contains(&res.err().unwrap().to_string(), "Invalid ZIP: offset should be 0");
Jooyung Hancee6de62021-08-11 15:52:07 +0900129 }
130}