304 lines · plain
1============2dm-integrity3============4 5The dm-integrity target emulates a block device that has additional6per-sector tags that can be used for storing integrity information.7 8A general problem with storing integrity tags with every sector is that9writing the sector and the integrity tag must be atomic - i.e. in case of10crash, either both sector and integrity tag or none of them is written.11 12To guarantee write atomicity, the dm-integrity target uses journal, it13writes sector data and integrity tags into a journal, commits the journal14and 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 this17situation the dm-crypt target creates the integrity data and passes them18to the dm-integrity target via bio_integrity_payload attached to the bio.19In this mode, the dm-crypt and dm-integrity targets provide authenticated20disk encryption - if the attacker modifies the encrypted device, an I/O21error is returned instead of random data.22 23The dm-integrity target can also be used as a standalone target, in this24mode it calculates and verifies the integrity tag internally. In this25mode, the dm-integrity target can be used to detect silent data26corruption on the disk or in the I/O path.27 28There's an alternate mode of operation where dm-integrity uses a bitmap29instead of a journal. If a bit in the bitmap is 1, the corresponding30region's data and integrity tags are not synchronized - if the machine31crashes, the unsynchronized regions will be recalculated. The bitmap mode32is faster than the journal mode, because we don't have to write the data33twice, but it is also less reliable, because if data corruption happens34when the machine crashes, it may not be detected.35 36When loading the target for the first time, the kernel driver will format37the device. But it will only format the device if the superblock contains38zeroes. If the superblock is neither valid nor zeroed, the dm-integrity39target can't be loaded.40 41Accesses to the on-disk metadata area containing checksums (aka tags) are42buffered using dm-bufio. When an access to any given metadata area43occurs, each unique metadata area gets its own buffer(s). The buffer size44is capped at the size of the metadata area, but may be smaller, thereby45requiring multiple buffers to represent the full metadata area. A smaller46buffer size will produce a smaller resulting read/write operation to the47metadata area for small reads/writes. The metadata is still read even in48a full write to the data covered by a single buffer.49 50To use the target for the first time:51 521. overwrite the superblock with zeroes532. load the dm-integrity target with one-sector size, the kernel driver54 will format the device553. unload the dm-integrity target564. read the "provided_data_sectors" value from the superblock575. load the dm-integrity target with the target size58 "provided_data_sectors"596. if you want to use dm-integrity with dm-crypt, load the dm-crypt target60 with the size "provided_data_sectors"61 62 63Target arguments:64 651. the underlying block device66 672. the number of reserved sector at the beginning of the device - the68 dm-integrity won't read of write these sectors69 703. the size of the integrity tag (if "-" is used, the size is taken from71 the internal-hash algorithm)72 734. mode:74 75 D - direct writes (without journal)76 in this mode, journaling is77 not used and data sectors and integrity tags are written78 separately. In case of crash, it is possible that the data79 and integrity tag doesn't match.80 J - journaled writes81 data and integrity tags are written to the82 journal and atomicity is guaranteed. In case of crash,83 either both data and tag or none of them are written. The84 journaled mode degrades write throughput twice because the85 data have to be written twice.86 B - bitmap mode - data and metadata are written without any87 synchronization, the driver maintains a bitmap of dirty88 regions where data and metadata don't match. This mode can89 only be used with internal hash.90 R - recovery mode - in this mode, journal is not replayed,91 checksums are not checked and writes to the device are not92 allowed. This mode is useful for data recovery if the93 device cannot be activated in any of the other standard94 modes.95 965. the number of additional arguments97 98Additional arguments:99 100journal_sectors:number101 The size of journal, this argument is used only if formatting the102 device. If the device is already formatted, the value from the103 superblock is used.104 105interleave_sectors:number (default 32768)106 The number of interleaved sectors. This values is rounded down to107 a power of two. If the device is already formatted, the value from108 the superblock is used.109 110meta_device:device111 Don't interleave the data and metadata on the device. Use a112 separate device for metadata.113 114buffer_sectors:number (default 128)115 The number of sectors in one metadata buffer. The value is rounded116 down to a power of two.117 118journal_watermark:number (default 50)119 The journal watermark in percents. When the size of the journal120 exceeds this watermark, the thread that flushes the journal will121 be started.122 123commit_time:number (default 10000)124 Commit time in milliseconds. When this time passes, the journal is125 written. The journal is also written immediately if the FLUSH126 request is received.127 128internal_hash:algorithm(:key) (the key is optional)129 Use internal hash or crc.130 When this argument is used, the dm-integrity target won't accept131 integrity tags from the upper target, but it will automatically132 generate and verify the integrity tags.133 134 You can use a crc algorithm (such as crc32), then integrity target135 will protect the data against accidental corruption.136 You can also use a hmac algorithm (for example137 "hmac(sha256):0123456789abcdef"), in this mode it will provide138 cryptographic authentication of the data without encryption.139 140 When this argument is not used, the integrity tags are accepted141 from an upper layer target, such as dm-crypt. The upper layer142 target should check the validity of the integrity tags.143 144recalculate145 Recalculate the integrity tags automatically. It is only valid146 when using internal hash.147 148journal_crypt:algorithm(:key) (the key is optional)149 Encrypt the journal using given algorithm to make sure that the150 attacker can't read the journal. You can use a block cipher here151 (such as "cbc(aes)") or a stream cipher (for example "chacha20"152 or "ctr(aes)").153 154 The journal contains history of last writes to the block device,155 an attacker reading the journal could see the last sector numbers156 that were written. From the sector numbers, the attacker can infer157 the size of files that were written. To protect against this158 situation, you can encrypt the journal.159 160journal_mac:algorithm(:key) (the key is optional)161 Protect sector numbers in the journal from accidental or malicious162 modification. To protect against accidental modification, use a163 crc algorithm, to protect against malicious modification, use a164 hmac algorithm with a key.165 166 This option is not needed when using internal-hash because in this167 mode, the integrity of journal entries is checked when replaying168 the journal. Thus, modified sector number would be detected at169 this stage.170 171block_size:number (default 512)172 The size of a data block in bytes. The larger the block size the173 less overhead there is for per-block integrity metadata.174 Supported values are 512, 1024, 2048 and 4096 bytes.175 176sectors_per_bit:number177 In the bitmap mode, this parameter specifies the number of178 512-byte sectors that corresponds to one bitmap bit.179 180bitmap_flush_interval:number181 The bitmap flush interval in milliseconds. The metadata buffers182 are synchronized when this interval expires.183 184allow_discards185 Allow block discard requests (a.k.a. TRIM) for the integrity device.186 Discards are only allowed to devices using internal hash.187 188fix_padding189 Use a smaller padding of the tag area that is more190 space-efficient. If this option is not present, large padding is191 used - that is for compatibility with older kernels.192 193fix_hmac194 Improve security of internal_hash and journal_mac:195 196 - the section number is mixed to the mac, so that an attacker can't197 copy sectors from one journal section to another journal section198 - the superblock is protected by journal_mac199 - a 16-byte salt stored in the superblock is mixed to the mac, so200 that the attacker can't detect that two disks have the same hmac201 key and also to disallow the attacker to move sectors from one202 disk to another203 204legacy_recalculate205 Allow recalculating of volumes with HMAC keys. This is disabled by206 default for security reasons - an attacker could modify the volume,207 set recalc_sector to zero, and the kernel would not detect the208 modification.209 210The journal mode (D/J), buffer_sectors, journal_watermark, commit_time and211allow_discards can be changed when reloading the target (load an inactive212table and swap the tables with suspend and resume). The other arguments213should not be changed when reloading the target because the layout of disk214data depend on them and the reloaded target would be non-functional.215 216For example, on a device using the default interleave_sectors of 32768, a217block_size of 512, and an internal_hash of crc32c with a tag size of 4218bytes, it will take 128 KiB of tags to track a full data area, requiring219256 sectors of metadata per data area. With the default buffer_sectors of220128, that means there will be 2 buffers per metadata area, or 2 buffers221per 16 MiB of data.222 223Status line:224 2251. the number of integrity mismatches2262. provided data sectors - that is the number of sectors that the user227 could use2283. the current recalculating position (or '-' if we didn't recalculate)229 230 231The layout of the formatted block device:232 233* reserved sectors234 (they are not used by this target, they can be used for235 storing LUKS metadata or for other purpose), the size of the reserved236 area is specified in the target arguments237 238* superblock (4kiB)239 * magic string - identifies that the device was formatted240 * version241 * log2(interleave sectors)242 * integrity tag size243 * the number of journal sections244 * provided data sectors - the number of sectors that this target245 provides (i.e. the size of the device minus the size of all246 metadata and padding). The user of this target should not send247 bios that access data beyond the "provided data sectors" limit.248 * flags249 SB_FLAG_HAVE_JOURNAL_MAC250 - a flag is set if journal_mac is used251 SB_FLAG_RECALCULATING252 - recalculating is in progress253 SB_FLAG_DIRTY_BITMAP254 - journal area contains the bitmap of dirty255 blocks256 * log2(sectors per block)257 * a position where recalculating finished258* journal259 The journal is divided into sections, each section contains:260 261 * metadata area (4kiB), it contains journal entries262 263 - every journal entry contains:264 265 * logical sector (specifies where the data and tag should266 be written)267 * last 8 bytes of data268 * integrity tag (the size is specified in the superblock)269 270 - every metadata sector ends with271 272 * mac (8-bytes), all the macs in 8 metadata sectors form a273 64-byte value. It is used to store hmac of sector274 numbers in the journal section, to protect against a275 possibility that the attacker tampers with sector276 numbers in the journal.277 * commit id278 279 * data area (the size is variable; it depends on how many journal280 entries fit into the metadata area)281 282 - every sector in the data area contains:283 284 * data (504 bytes of data, the last 8 bytes are stored in285 the journal entry)286 * commit id287 288 To test if the whole journal section was written correctly, every289 512-byte sector of the journal ends with 8-byte commit id. If the290 commit id matches on all sectors in a journal section, then it is291 assumed that the section was written correctly. If the commit id292 doesn't match, the section was written partially and it should not293 be replayed.294 295* one or more runs of interleaved tags and data.296 Each run contains:297 298 * tag area - it contains integrity tags. There is one tag for each299 sector in the data area. The size of this area is always 4KiB or300 greater.301 * data area - it contains data sectors. The number of data sectors302 in one run must be a power of two. log2 of this value is stored303 in the superblock.304