tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
1============
2dm-integrity
3============
4
5The dm-integrity target emulates a block device that has additional
6per-sector tags that can be used for storing integrity information.
7
8A general problem with storing integrity tags with every sector is that
9writing the sector and the integrity tag must be atomic - i.e. in case of
10crash, either both sector and integrity tag or none of them is written.
11
12To guarantee write atomicity, the dm-integrity target uses journal, it
13writes sector data and integrity tags into a journal, commits the journal
14and then copies the data and integrity tags to their respective location.
15
16The dm-integrity target can be used with the dm-crypt target - in this
17situation the dm-crypt target creates the integrity data and passes them
18to the dm-integrity target via bio_integrity_payload attached to the bio.
19In this mode, the dm-crypt and dm-integrity targets provide authenticated
20disk encryption - if the attacker modifies the encrypted device, an I/O
21error is returned instead of random data.
22
23The dm-integrity target can also be used as a standalone target, in this
24mode it calculates and verifies the integrity tag internally. In this
25mode, the dm-integrity target can be used to detect silent data
26corruption on the disk or in the I/O path.
27
28There's an alternate mode of operation where dm-integrity uses bitmap
29instead of a journal. If a bit in the bitmap is 1, the corresponding
30region's data and integrity tags are not synchronized - if the machine
31crashes, the unsynchronized regions will be recalculated. The bitmap mode
32is faster than the journal mode, because we don't have to write the data
33twice, but it is also less reliable, because if data corruption happens
34when the machine crashes, it may not be detected.
35
36When loading the target for the first time, the kernel driver will format
37the device. But it will only format the device if the superblock contains
38zeroes. If the superblock is neither valid nor zeroed, the dm-integrity
39target can't be loaded.
40
41To use the target for the first time:
42
431. overwrite the superblock with zeroes
442. load the dm-integrity target with one-sector size, the kernel driver
45 will format the device
463. unload the dm-integrity target
474. read the "provided_data_sectors" value from the superblock
485. load the dm-integrity target with the the target size
49 "provided_data_sectors"
506. if you want to use dm-integrity with dm-crypt, load the dm-crypt target
51 with the size "provided_data_sectors"
52
53
54Target arguments:
55
561. the underlying block device
57
582. the number of reserved sector at the beginning of the device - the
59 dm-integrity won't read of write these sectors
60
613. the size of the integrity tag (if "-" is used, the size is taken from
62 the internal-hash algorithm)
63
644. mode:
65
66 D - direct writes (without journal)
67 in this mode, journaling is
68 not used and data sectors and integrity tags are written
69 separately. In case of crash, it is possible that the data
70 and integrity tag doesn't match.
71 J - journaled writes
72 data and integrity tags are written to the
73 journal and atomicity is guaranteed. In case of crash,
74 either both data and tag or none of them are written. The
75 journaled mode degrades write throughput twice because the
76 data have to be written twice.
77 B - bitmap mode - data and metadata are written without any
78 synchronization, the driver maintains a bitmap of dirty
79 regions where data and metadata don't match. This mode can
80 only be used with internal hash.
81 R - recovery mode - in this mode, journal is not replayed,
82 checksums are not checked and writes to the device are not
83 allowed. This mode is useful for data recovery if the
84 device cannot be activated in any of the other standard
85 modes.
86
875. the number of additional arguments
88
89Additional arguments:
90
91journal_sectors:number
92 The size of journal, this argument is used only if formatting the
93 device. If the device is already formatted, the value from the
94 superblock is used.
95
96interleave_sectors:number
97 The number of interleaved sectors. This values is rounded down to
98 a power of two. If the device is already formatted, the value from
99 the superblock is used.
100
101meta_device:device
102 Don't interleave the data and metadata on on device. Use a
103 separate device for metadata.
104
105buffer_sectors:number
106 The number of sectors in one buffer. The value is rounded down to
107 a power of two.
108
109 The tag area is accessed using buffers, the buffer size is
110 configurable. The large buffer size means that the I/O size will
111 be larger, but there could be less I/Os issued.
112
113journal_watermark:number
114 The journal watermark in percents. When the size of the journal
115 exceeds this watermark, the thread that flushes the journal will
116 be started.
117
118commit_time:number
119 Commit time in milliseconds. When this time passes, the journal is
120 written. The journal is also written immediatelly if the FLUSH
121 request is received.
122
123internal_hash:algorithm(:key) (the key is optional)
124 Use internal hash or crc.
125 When this argument is used, the dm-integrity target won't accept
126 integrity tags from the upper target, but it will automatically
127 generate and verify the integrity tags.
128
129 You can use a crc algorithm (such as crc32), then integrity target
130 will protect the data against accidental corruption.
131 You can also use a hmac algorithm (for example
132 "hmac(sha256):0123456789abcdef"), in this mode it will provide
133 cryptographic authentication of the data without encryption.
134
135 When this argument is not used, the integrity tags are accepted
136 from an upper layer target, such as dm-crypt. The upper layer
137 target should check the validity of the integrity tags.
138
139recalculate
140 Recalculate the integrity tags automatically. It is only valid
141 when using internal hash.
142
143journal_crypt:algorithm(:key) (the key is optional)
144 Encrypt the journal using given algorithm to make sure that the
145 attacker can't read the journal. You can use a block cipher here
146 (such as "cbc(aes)") or a stream cipher (for example "chacha20",
147 "salsa20" or "ctr(aes)").
148
149 The journal contains history of last writes to the block device,
150 an attacker reading the journal could see the last sector nubmers
151 that were written. From the sector numbers, the attacker can infer
152 the size of files that were written. To protect against this
153 situation, you can encrypt the journal.
154
155journal_mac:algorithm(:key) (the key is optional)
156 Protect sector numbers in the journal from accidental or malicious
157 modification. To protect against accidental modification, use a
158 crc algorithm, to protect against malicious modification, use a
159 hmac algorithm with a key.
160
161 This option is not needed when using internal-hash because in this
162 mode, the integrity of journal entries is checked when replaying
163 the journal. Thus, modified sector number would be detected at
164 this stage.
165
166block_size:number
167 The size of a data block in bytes. The larger the block size the
168 less overhead there is for per-block integrity metadata.
169 Supported values are 512, 1024, 2048 and 4096 bytes. If not
170 specified the default block size is 512 bytes.
171
172sectors_per_bit:number
173 In the bitmap mode, this parameter specifies the number of
174 512-byte sectors that corresponds to one bitmap bit.
175
176bitmap_flush_interval:number
177 The bitmap flush interval in milliseconds. The metadata buffers
178 are synchronized when this interval expires.
179
180fix_padding
181 Use a smaller padding of the tag area that is more
182 space-efficient. If this option is not present, large padding is
183 used - that is for compatibility with older kernels.
184
185allow_discards
186 Allow block discard requests (a.k.a. TRIM) for the integrity device.
187 Discards are only allowed to devices using internal hash.
188
189The journal mode (D/J), buffer_sectors, journal_watermark, commit_time and
190allow_discards can be changed when reloading the target (load an inactive
191table and swap the tables with suspend and resume). The other arguments
192should not be changed when reloading the target because the layout of disk
193data depend on them and the reloaded target would be non-functional.
194
195
196The layout of the formatted block device:
197
198* reserved sectors
199 (they are not used by this target, they can be used for
200 storing LUKS metadata or for other purpose), the size of the reserved
201 area is specified in the target arguments
202
203* superblock (4kiB)
204 * magic string - identifies that the device was formatted
205 * version
206 * log2(interleave sectors)
207 * integrity tag size
208 * the number of journal sections
209 * provided data sectors - the number of sectors that this target
210 provides (i.e. the size of the device minus the size of all
211 metadata and padding). The user of this target should not send
212 bios that access data beyond the "provided data sectors" limit.
213 * flags
214 SB_FLAG_HAVE_JOURNAL_MAC
215 - a flag is set if journal_mac is used
216 SB_FLAG_RECALCULATING
217 - recalculating is in progress
218 SB_FLAG_DIRTY_BITMAP
219 - journal area contains the bitmap of dirty
220 blocks
221 * log2(sectors per block)
222 * a position where recalculating finished
223* journal
224 The journal is divided into sections, each section contains:
225
226 * metadata area (4kiB), it contains journal entries
227
228 - every journal entry contains:
229
230 * logical sector (specifies where the data and tag should
231 be written)
232 * last 8 bytes of data
233 * integrity tag (the size is specified in the superblock)
234
235 - every metadata sector ends with
236
237 * mac (8-bytes), all the macs in 8 metadata sectors form a
238 64-byte value. It is used to store hmac of sector
239 numbers in the journal section, to protect against a
240 possibility that the attacker tampers with sector
241 numbers in the journal.
242 * commit id
243
244 * data area (the size is variable; it depends on how many journal
245 entries fit into the metadata area)
246
247 - every sector in the data area contains:
248
249 * data (504 bytes of data, the last 8 bytes are stored in
250 the journal entry)
251 * commit id
252
253 To test if the whole journal section was written correctly, every
254 512-byte sector of the journal ends with 8-byte commit id. If the
255 commit id matches on all sectors in a journal section, then it is
256 assumed that the section was written correctly. If the commit id
257 doesn't match, the section was written partially and it should not
258 be replayed.
259
260* one or more runs of interleaved tags and data.
261 Each run contains:
262
263 * tag area - it contains integrity tags. There is one tag for each
264 sector in the data area
265 * data area - it contains data sectors. The number of data sectors
266 in one run must be a power of two. log2 of this value is stored
267 in the superblock.