1<!-- 2 - SPDX-License-Identifier: Apache-2.0 3 4 - Copyright (c) 2017-2020 Linaro LTD 5 - Copyright (c) 2017-2019 JUUL Labs 6 - Copyright (c) 2019-2024 Arm Limited 7 8 - Original license: 9 10 - Licensed to the Apache Software Foundation (ASF) under one 11 - or more contributor license agreements. See the NOTICE file 12 - distributed with this work for additional information 13 - regarding copyright ownership. The ASF licenses this file 14 - to you under the Apache License, Version 2.0 (the 15 - "License"); you may not use this file except in compliance 16 - with the License. You may obtain a copy of the License at 17 18 - http://www.apache.org/licenses/LICENSE-2.0 19 20 - Unless required by applicable law or agreed to in writing, 21 - software distributed under the License is distributed on an 22 - "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY 23 - KIND, either express or implied. See the License for the 24 - specific language governing permissions and limitations 25 - under the License. 26--> 27 28# Bootloader 29 30## [Summary](#summary) 31 32MCUboot comprises two packages: 33 34* The bootutil library (boot/bootutil) 35* The boot application (each port has its own at boot/<port>) 36 37The bootutil library performs most of the functions of a bootloader. In 38particular, the piece that is missing is the final step of actually jumping to 39the main image. This last step is instead implemented by the boot application. 40Bootloader functionality is separated in this manner to enable unit testing of 41the bootloader. A library can be unit tested, but an application can't. 42Therefore, functionality is delegated to the bootutil library when possible. 43 44## [Limitations](#limitations) 45 46The bootloader currently only supports images with the following 47characteristics: 48* Built to run from flash. 49* Built to run from a fixed location (i.e., not position-independent). 50 51## [Image format](#image-format) 52 53The following definitions describe the image format. 54 55``` c 56#define IMAGE_MAGIC 0x96f3b83d 57 58#define IMAGE_HEADER_SIZE 32 59 60struct image_version { 61 uint8_t iv_major; 62 uint8_t iv_minor; 63 uint16_t iv_revision; 64 uint32_t iv_build_num; 65}; 66 67/** Image header. All fields are in little endian byte order. */ 68struct image_header { 69 uint32_t ih_magic; 70 uint32_t ih_load_addr; 71 uint16_t ih_hdr_size; /* Size of image header (bytes). */ 72 uint16_t ih_protect_tlv_size; /* Size of protected TLV area (bytes). */ 73 uint32_t ih_img_size; /* Does not include header. */ 74 uint32_t ih_flags; /* IMAGE_F_[...]. */ 75 struct image_version ih_ver; 76 uint32_t _pad1; 77}; 78 79#define IMAGE_TLV_INFO_MAGIC 0x6907 80#define IMAGE_TLV_PROT_INFO_MAGIC 0x6908 81 82/** Image TLV header. All fields in little endian. */ 83struct image_tlv_info { 84 uint16_t it_magic; 85 uint16_t it_tlv_tot; /* size of TLV area (including tlv_info header) */ 86}; 87 88/** Image trailer TLV format. All fields in little endian. */ 89struct image_tlv { 90 uint8_t it_type; /* IMAGE_TLV_[...]. */ 91 uint8_t _pad; 92 uint16_t it_len; /* Data length (not including TLV header). */ 93}; 94 95/* 96 * Image header flags. 97 */ 98#define IMAGE_F_PIC 0x00000001 /* Not supported. */ 99#define IMAGE_F_ENCRYPTED_AES128 0x00000004 /* Encrypted using AES128. */ 100#define IMAGE_F_ENCRYPTED_AES256 0x00000008 /* Encrypted using AES256. */ 101#define IMAGE_F_NON_BOOTABLE 0x00000010 /* Split image app. */ 102#define IMAGE_F_RAM_LOAD 0x00000020 103 104/* 105 * Image trailer TLV types. 106 */ 107#define IMAGE_TLV_KEYHASH 0x01 /* hash of the public key */ 108#define IMAGE_TLV_SHA256 0x10 /* SHA256 of image hdr and body */ 109#define IMAGE_TLV_RSA2048_PSS 0x20 /* RSA2048 of hash output */ 110#define IMAGE_TLV_ECDSA224 0x21 /* ECDSA of hash output - Not supported anymore */ 111#define IMAGE_TLV_ECDSA_SIG 0x22 /* ECDSA of hash output */ 112#define IMAGE_TLV_RSA3072_PSS 0x23 /* RSA3072 of hash output */ 113#define IMAGE_TLV_ED25519 0x24 /* ED25519 of hash output */ 114#define IMAGE_TLV_ENC_RSA2048 0x30 /* Key encrypted with RSA-OAEP-2048 */ 115#define IMAGE_TLV_ENC_KW 0x31 /* Key encrypted with AES-KW-128 or 116 256 */ 117#define IMAGE_TLV_ENC_EC256 0x32 /* Key encrypted with ECIES-P256 */ 118#define IMAGE_TLV_ENC_X25519 0x33 /* Key encrypted with ECIES-X25519 */ 119#define IMAGE_TLV_DEPENDENCY 0x40 /* Image depends on other image */ 120#define IMAGE_TLV_SEC_CNT 0x50 /* security counter */ 121``` 122 123Optional type-length-value records (TLVs) containing image metadata are placed 124after the end of the image. 125 126The `ih_protect_tlv_size` field indicates the length of the protected TLV area. 127If protected TLVs are present then a TLV info header with magic equal to 128`IMAGE_TLV_PROT_INFO_MAGIC` must be present and the protected TLVs (plus the 129info header itself) have to be included in the hash calculation. Otherwise the 130hash is only calculated over the image header and the image itself. In this 131case the value of the `ih_protect_tlv_size` field is 0. 132 133The `ih_hdr_size` field indicates the length of the header, and therefore the 134offset of the image itself. This field provides for backwards compatibility in 135case of changes to the format of the image header. 136 137## [Flash map](#flash-map) 138 139A device's flash is partitioned according to its _flash map_. At a high 140level, the flash map maps numeric IDs to _flash areas_. A flash area is a 141region of disk with the following properties: 1421. An area can be fully erased without affecting any other areas. 1432. A write to one area does not restrict writes to other areas. 144 145The bootloader uses the following flash area IDs: 146```c 147/* Independent from multiple image boot */ 148#define FLASH_AREA_BOOTLOADER 0 149#define FLASH_AREA_IMAGE_SCRATCH 3 150``` 151```c 152/* If the bootloader is working with the first image */ 153#define FLASH_AREA_IMAGE_PRIMARY 1 154#define FLASH_AREA_IMAGE_SECONDARY 2 155``` 156```c 157/* If the bootloader is working with the second image */ 158#define FLASH_AREA_IMAGE_PRIMARY 5 159#define FLASH_AREA_IMAGE_SECONDARY 6 160``` 161 162The bootloader area contains the bootloader image itself. The other areas are 163described in subsequent sections. The flash could contain multiple executable 164images therefore the flash area IDs of primary and secondary areas are mapped 165based on the number of the active image (on which the bootloader is currently 166working). 167 168## [Image slots](#image-slots) 169 170A portion of the flash memory can be partitioned into multiple image areas, each 171contains two image slots: a primary slot and a secondary slot. 172Normally, the bootloader will only run an image from the primary slot, so 173images must be built such that they can run from that fixed location in flash 174(the exception to this is the [direct-xip](#direct-xip) and the 175[ram-load](#ram-load) upgrade mode). If the bootloader needs to run the 176image resident in the secondary slot, it must copy its contents into the primary 177slot before doing so, either by swapping the two images or by overwriting the 178contents of the primary slot. The bootloader supports either swap- or 179overwrite-based image upgrades, but must be configured at build time to choose 180one of these two strategies. 181 182### [Swap using scratch](#image-swap-using-scratch) 183 184When swap-using-scratch algorithm is used, in addition to the slots of 185image areas, the bootloader requires a scratch area to allow for reliable 186image swapping. The scratch area must have a size 187that is enough to store at least the largest sector that is going to be swapped. 188Many devices have small equally sized flash sectors, eg 4K, while others have 189variable sized sectors where the largest sectors might be 128K or 256K, so the 190scratch must be big enough to store that. The scratch is only ever used when 191swapping firmware, which means only when doing an upgrade. Given that, the main 192reason for using a larger size for the scratch is that flash wear will be more 193evenly distributed, because a single sector would be written twice the number of 194times than using two sectors, for example. To evaluate the ideal size of the 195scratch for your use case the following parameters are relevant: 196 197* the ratio of image size / scratch size 198* the number of erase cycles supported by the flash hardware 199 200The image size is used (instead of slot size) because only the slot's sectors 201that are actually used for storing the image are copied. The image/scratch ratio 202is the number of times the scratch will be erased on every upgrade. The number 203of erase cycles divided by the image/scratch ratio will give you the number of 204times an upgrade can be performed before the device goes out of spec. 205 206``` 207num_upgrades = number_of_erase_cycles / (image_size / scratch_size) 208``` 209 210Let's assume, for example, a device with 10000 erase cycles, an image size of 211150K and a scratch of 4K (usual minimum size of 4K sector devices). This would 212result in a total of: 213 214`10000 / (150 / 4) ~ 267` 215 216Increasing the scratch to 16K would give us: 217 218`10000 / (150 / 16) ~ 1067` 219 220There is no *best* ratio, as the right size is use-case dependent. Factors to 221consider include the number of times a device will be upgraded both in the field 222and during development, as well as any desired safety margin on the 223manufacturer's specified number of erase cycles. In general, using a ratio that 224allows hundreds to thousands of field upgrades in production is recommended. 225 226swap-using scratch algorithm assumes that the primary and the secondary image 227slot areas sizes are equal. 228The maximum image size available for the application 229will be: 230``` 231maximum-image-size = image-slot-size - image-trailer-size 232``` 233 234Where: 235 `image-slot-size` is the size of the image slot. 236 `image-trailer-size` is the size of the image trailer. 237 238### [Swap without using scratch](#image-swap-no-scratch) 239 240This algorithm is an alternative to the swap-using-scratch algorithm. 241It uses an additional sector in the primary slot to make swap possible. 242The algorithm works as follows: 243 244 1. Moves all sectors of the primary slot up by one sector. 245 Beginning from N=0: 246 2. Copies the N-th sector from the secondary slot to the N-th sector of the 247 primary slot. 248 3. Copies the (N+1)-th sector from the primary slot to the N-th sector of the 249 secondary slot. 250 4. Repeats steps 2. and 3. until all the slots' sectors are swapped. 251 252This algorithm is designed so that the higher sector of the primary slot is 253used only for allowing sectors to move up. Therefore the most 254memory-size-effective slot layout is when the primary slot is larger than 255the secondary slot by exactly one sector plus the size of the swap status area, 256rounded up to the total size of the sectors it occupies, 257although same-sized slots are allowed as well. 258The algorithm is limited to support sectors of the same 259sector layout. All slot's sectors should be of the same size. 260 261When using this algorithm the maximum image size available for the application 262will be: 263``` 264maximum-image-size = (N-1) * slot-sector-size - image-trailer-sectors-size 265``` 266 267Where: 268 `N` is the number of sectors in the primary slot. 269 `image-trailer-sectors-size` is the size of the image trailer rounded up to 270 the total size of sectors its occupied. For instance if the image-trailer-size 271 is equal to 1056 B and the sector size is equal to 1024 B, then 272 `image-trailer-sectors-size` will be equal to 2048 B. 273 274The algorithm does two erase cycles on the primary slot and one on the secondary 275slot during each swap. Assuming that receiving a new image by the DFU 276application requires 1 erase cycle on the secondary slot, this should result in 277leveling the flash wear between the slots. 278 279The algorithm is enabled using the `MCUBOOT_SWAP_USING_MOVE` option. 280 281### [Equal slots (direct-xip)](#direct-xip) 282 283When the direct-xip mode is enabled the active image flag is "moved" between the 284slots during image upgrade and in contrast to the above, the bootloader can 285run an image directly from either the primary or the secondary slot (without 286having to move/copy it into the primary slot). Therefore the image update 287client, which downloads the new images must be aware, which slot contains the 288active image and which acts as a staging area and it is responsible for loading 289the proper images into the proper slot. All this requires that the images be 290built to be executed from the corresponding slot. At boot time the bootloader 291first looks for images in the slots and then inspects the version numbers in the 292image headers. It selects the newest image (with the highest version number) and 293then checks its validity (integrity check, signature verification etc.). If the 294image is invalid MCUboot erases its memory slot and starts to validate the other 295image. After a successful validation of the selected image the bootloader 296chain-loads it. 297 298An additional "revert" mechanism is also supported. For more information, please 299read the [corresponding section](#direct-xip-revert). 300Handling the primary and secondary slots as equals has its drawbacks. Since the 301images are not moved between the slots, the on-the-fly image 302encryption/decryption can't be supported (it only applies to storing the image 303in an external flash on the device, the transport of encrypted image data is 304still feasible). 305 306The overwrite and the direct-xip upgrade strategies are substantially simpler to 307implement than the image swapping strategy, especially since the bootloader must 308work properly even when it is reset during the middle of an image swap. For this 309reason, the rest of the document describes its behavior when configured to swap 310images during an upgrade. 311 312### [RAM loading](#ram-load) 313 314In ram-load mode the slots are equal. Like the direct-xip mode, this mode 315also selects the newest image by reading the image version numbers in the image 316headers. But instead of executing it in place, the newest image is copied to the 317RAM for execution. The load address, the location in RAM where the image is 318copied to, is stored in the image header. The ram-load upgrade mode can be 319useful when there is no internal flash in the SoC, but there is a big enough 320internal RAM to hold the images. Usually in this case the images are stored 321in an external storage device. Execution from external storage has some 322drawbacks (lower execution speed, image is exposed to attacks) therefore the 323image is always copied to the internal RAM before the authentication and 324execution. Ram-load mode requires the image to be built to be executed from 325the RAM address range instead of the storage device address range. If 326ram-load is enabled then platform must define the following parameters: 327 328```c 329#define IMAGE_EXECUTABLE_RAM_START <area_base_addr> 330#define IMAGE_EXECUTABLE_RAM_SIZE <area_size_in_bytes> 331``` 332 333For multiple image load if multiple ram regions are used platform must define 334the `MULTIPLE_EXECUTABLE_RAM_REGIONS` flag instead and implement the following 335function: 336 337```c 338int boot_get_image_exec_ram_info(uint32_t image_id, 339 uint32_t *exec_ram_start, 340 uint32_t *exec_ram_size) 341``` 342 343When ram-load is enabled, the `--load-addr <addr>` option of the `imgtool` 344script must also be used when signing the images. This option set the `RAM_LOAD` 345flag in the image header which indicates that the image should be loaded to the 346RAM and also set the load address in the image header. 347 348When the encryption option is enabled (`MCUBOOT_ENC_IMAGES`) along with ram-load 349the image is checked for encryption. If the image is not encrypted, RAM loading 350happens as described above. If the image is encrypted, it is copied in RAM at 351the provided address and then decrypted. Finally, the decrypted image is 352authenticated in RAM and executed. 353 354## [Boot swap types](#boot-swap-types) 355 356When the device first boots under normal circumstances, there is an up-to-date 357firmware image in each primary slot, which MCUboot can validate and then 358chain-load. In this case, no image swaps are necessary. During device upgrades, 359however, new candidate image(s) is present in the secondary slot(s), which 360MCUboot must swap into the primary slot(s) before booting as discussed above. 361 362Upgrading an old image with a new one by swapping can be a two-step process. In 363this process, MCUboot performs a "test" swap of image data in flash and boots 364the new image or it will be executed during operation. The new image can then 365update the contents of flash at runtime to mark itself "OK", and MCUboot will 366then still choose to run it during the next boot. When this happens, the swap is 367made "permanent". If this doesn't happen, MCUboot will perform a "revert" swap 368during the next boot by swapping the image(s) back into its original location(s) 369, and attempting to boot the old image(s). 370 371Depending on the use case, the first swap can also be made permanent directly. 372In this case, MCUboot will never attempt to revert the images on the next reset. 373 374Test swaps are supported to provide a rollback mechanism to prevent devices 375from becoming "bricked" by bad firmware. If the device crashes immediately 376upon booting a new (bad) image, MCUboot will revert to the old (working) image 377at the next device reset, rather than booting the bad image again. This allows 378device firmware to make test swaps permanent only after performing a self-test 379routine. 380 381On startup, MCUboot inspects the contents of flash to decide for each images 382which of these "swap types" to perform; this decision determines how it 383proceeds. 384 385The possible swap types, and their meanings, are: 386 387- `BOOT_SWAP_TYPE_NONE`: The "usual" or "no upgrade" case; attempt to boot the 388 contents of the primary slot. 389 390- `BOOT_SWAP_TYPE_TEST`: Boot the contents of the secondary slot by swapping 391 images. Unless the swap is made permanent, revert back on the next boot. 392 393- `BOOT_SWAP_TYPE_PERM`: Permanently swap images, and boot the upgraded image 394 firmware. 395 396- `BOOT_SWAP_TYPE_REVERT`: A previous test swap was not made permanent; 397 swap back to the old image whose data are now in the secondary slot. If the 398 old image marks itself "OK" when it boots, the next boot will have swap type 399 `BOOT_SWAP_TYPE_NONE`. 400 401- `BOOT_SWAP_TYPE_FAIL`: Swap failed because image to be run is not valid. 402 403- `BOOT_SWAP_TYPE_PANIC`: Swapping encountered an unrecoverable error. 404 405The "swap type" is a high-level representation of the outcome of the 406boot. Subsequent sections describe how MCUboot determines the swap type from 407the bit-level contents of flash. 408 409### [Revert mechanism in direct-xip mode](#direct-xip-revert) 410 411The direct-xip mode also supports a "revert" mechanism which is the equivalent 412of the swap mode's "revert" swap. When the direct-xip mode is selected it can be 413enabled with the MCUBOOT_DIRECT_XIP_REVERT config option and an image trailer 414must also be added to the signed images (the "--pad" option of the `imgtool` 415script must be used). For more information on this please read the 416[Image Trailer](#image-trailer) section and the [imgtool](imgtool.md) 417documentation. Making the images permanent (marking them as confirmed in 418advance) is also supported just like in swap mode. The individual steps of the 419direct-xip mode's "revert" mechanism are the following: 420 4211. Select the slot which holds the newest potential image. 4222. Was the image previously selected to run (during a previous boot)? 423 + Yes: Did the image mark itself "OK" (was the self-test successful)? 424 + Yes. 425 - Proceed to step 3. 426 + No. 427 - Erase the image from the slot to prevent it from being selected 428 again during the next boot. 429 - Return to step 1 (the bootloader will attempt to select and 430 possibly boot the previous image if there is one). 431 + No. 432 - Mark the image as "selected" (set the copy_done flag in the trailer). 433 - Proceed to step 3. 4343. Proceed to image validation ... 435 436## [Image trailer](#image-trailer) 437 438For the bootloader to be able to determine the current state and what actions 439should be taken during the current boot operation, it uses metadata stored in 440the image flash areas. While swapping, some of this metadata is temporarily 441copied into and out of the scratch area. 442 443This metadata is located at the end of the image flash areas, and is called an 444image trailer. An image trailer has the following structure: 445 446``` 447 0 1 2 3 448 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 449 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 450 ~ ~ 451 ~ Swap status (BOOT_MAX_IMG_SECTORS * min-write-size * 3) ~ 452 ~ ~ 453 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 454 | Encryption key 0 (16 octets) [*] | 455 | | 456 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 457 | 0xff padding as needed | 458 | (BOOT_MAX_ALIGN minus 16 octets from Encryption key 0) [*] | 459 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 460 | Encryption key 1 (16 octets) [*] | 461 | | 462 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 463 | 0xff padding as needed | 464 | (BOOT_MAX_ALIGN minus 16 octets from Encryption key 1) [*] | 465 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 466 | Swap size (4 octets) | 467 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 468 | 0xff padding as needed | 469 | (BOOT_MAX_ALIGN minus 4 octets from Swap size) | 470 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 471 | Swap info | 0xff padding (BOOT_MAX_ALIGN minus 1 octet) | 472 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 473 | Copy done | 0xff padding (BOOT_MAX_ALIGN minus 1 octet) | 474 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 475 | Image OK | 0xff padding (BOOT_MAX_ALIGN minus 1 octet) | 476 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 477 | 0xff padding as needed | 478 | (BOOT_MAX_ALIGN minus 16 octets from MAGIC) | 479 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 480 | MAGIC (16 octets) | 481 | | 482 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 483``` 484 485[*]: Only present if the encryption option is enabled (`MCUBOOT_ENC_IMAGES`). 486 487The offset immediately following such a record represents the start of the next 488flash area. 489 490--- 491***Note*** 492 493*"min-write-size" is a property of the flash hardware. If the hardware* 494*allows individual bytes to be written at arbitrary addresses, then* 495*min-write-size is 1. If the hardware only allows writes at even addresses,* 496*then min-write-size is 2, and so on.* 497 498--- 499 500An image trailer contains the following fields: 501 5021. Swap status: A series of records which records the progress of an image 503 swap. To swap entire images, data are swapped between the two image areas 504 one or more sectors at a time, like this: 505 506 - sector data in the primary slot is copied into scratch, then erased 507 - sector data in the secondary slot is copied into the primary slot, 508 then erased 509 - sector data in scratch is copied into the secondary slot 510 511As it swaps images, the bootloader updates the swap status field in a way that 512allows it to compute how far this swap operation has progressed for each 513sector. The swap status field can thus used to resume a swap operation if the 514bootloader is halted while a swap operation is ongoing and later reset. The 515`BOOT_MAX_IMG_SECTORS` value is the configurable maximum number of sectors 516MCUboot supports for each image; its value defaults to 128, but allows for 517either decreasing this size, to limit RAM usage, or to increase it in devices 518that have massive amounts of Flash or very small sized sectors and thus require 519a bigger configuration to allow for the handling of all slot's sectors. 520The factor of min-write-size is due to the behavior of flash hardware. The factor 521of 3 is explained below. 522 5232. Encryption keys: key-encrypting keys (KEKs). These keys are needed for 524 image encryption and decryption. See the 525 [encrypted images](encrypted_images.md) document for more information. 526 5273. Swap size: When beginning a new swap operation, the total size that needs 528 to be swapped (based on the slot with largest image + TLVs) is written to 529 this location for easier recovery in case of a reset while performing the 530 swap. 531 5324. Swap info: A single byte which encodes the following information: 533 - Swap type: Stored in bits 0-3. Indicating the type of swap operation in 534 progress. When MCUboot resumes an interrupted swap, it uses this field to 535 determine the type of operation to perform. This field contains one of the 536 following values in the table below. 537 - Image number: Stored in bits 4-7. It has always 0 value at single image 538 boot. In case of multi image boot it indicates, which image was swapped when 539 interrupt happened. The same scratch area is used during in case of all 540 image swap operation. Therefore this field is used to determine which image 541 the trailer belongs to if boot status is found on scratch area when the swap 542 operation is resumed. 543 544| Name | Value | 545| ------------------------- | ----- | 546| `BOOT_SWAP_TYPE_TEST` | 2 | 547| `BOOT_SWAP_TYPE_PERM` | 3 | 548| `BOOT_SWAP_TYPE_REVERT` | 4 | 549 550 5515. Copy done: A single byte indicating whether the image in this slot is 552 complete (0x01=done; 0xff=not done). 553 5546. Image OK: A single byte indicating whether the image in this slot has been 555 confirmed as good by the user (0x01=confirmed; 0xff=not confirmed). 556 5577. MAGIC: A 16-byte field identifying the image trailer layout. It may assume 558 distinct values depending on the maximum supported write alignment 559 (`BOOT_MAX_ALIGN`) of the image, as defined by the following construct: 560 561``` c 562union boot_img_magic_t 563{ 564 struct { 565 uint16_t align; 566 uint8_t magic[14]; 567 }; 568 uint8_t val[16]; 569}; 570``` 571 If `BOOT_MAX_ALIGN` is **8 bytes**, then MAGIC contains the following 16 bytes: 572 573``` c 574const union boot_img_magic_t boot_img_magic = { 575 .val = { 576 0x77, 0xc2, 0x95, 0xf3, 577 0x60, 0xd2, 0xef, 0x7f, 578 0x35, 0x52, 0x50, 0x0f, 579 0x2c, 0xb6, 0x79, 0x80 580 } 581}; 582``` 583 584 In case `BOOT_MAX_ALIGN` is defined to any value different than **8**, then the maximum 585 supported write alignment value is encoded in the MAGIC field, followed by a fixed 586 14-byte pattern: 587 588``` c 589const union boot_img_magic_t boot_img_magic = { 590 .align = BOOT_MAX_ALIGN, 591 .magic = { 592 0x2d, 0xe1, 593 0x5d, 0x29, 0x41, 0x0b, 594 0x8d, 0x77, 0x67, 0x9c, 595 0x11, 0x0f, 0x1f, 0x8a 596 } 597}; 598``` 599 600--- 601***Note*** 602Be aware that the image trailers make the ending area of the image slot 603unavailable for carrying the image data. In particular, the swap status size 604could be huge. For example, for 128 slot sectors with a 4-byte alignment, 605it would become 1536 B. 606 607--- 608 609## [Image trailers](#image-trailers) 610 611At startup, the bootloader determines the boot swap type by inspecting the 612image trailers. When using the term "image trailers" what is meant is the 613aggregate information provided by both image slot's trailers. 614 615### [New swaps (non-resumes)](#new-swaps-non-resumes) 616 617For new swaps, MCUboot must inspect a collection of fields to determine which 618swap operation to perform. 619 620The image trailers records are structured around the limitations imposed by 621flash hardware. As a consequence, they do not have a very intuitive design, and 622it is difficult to get a sense of the state of the device just by looking at the 623image trailers. It is better to map all the possible trailer states to the swap 624types described above via a set of tables. These tables are reproduced below. 625 626--- 627***Note*** 628 629*An important caveat about the tables described below is that they must* 630*be evaluated in the order presented here. Lower state numbers must have a* 631*higher priority when testing the image trailers.* 632 633--- 634 635``` 636 State I 637 | primary slot | secondary slot | 638 -----------------+--------------+----------------| 639 magic | Any | Good | 640 image-ok | Any | Unset | 641 copy-done | Any | Any | 642 -----------------+--------------+----------------' 643 result: BOOT_SWAP_TYPE_TEST | 644 -------------------------------------------------' 645 646 647 State II 648 | primary slot | secondary slot | 649 -----------------+--------------+----------------| 650 magic | Any | Good | 651 image-ok | Any | 0x01 | 652 copy-done | Any | Any | 653 -----------------+--------------+----------------' 654 result: BOOT_SWAP_TYPE_PERM | 655 -------------------------------------------------' 656 657 658 State III 659 | primary slot | secondary slot | 660 -----------------+--------------+----------------| 661 magic | Good | Unset | 662 image-ok | 0xff | Any | 663 copy-done | 0x01 | Any | 664 -----------------+--------------+----------------' 665 result: BOOT_SWAP_TYPE_REVERT | 666 -------------------------------------------------' 667``` 668 669Any of the above three states results in MCUboot attempting to swap images. 670 671Otherwise, MCUboot does not attempt to swap images, resulting in one of the 672other three swap types, as illustrated by State IV. 673 674``` 675 State IV 676 | primary slot | secondary slot | 677 -----------------+--------------+----------------| 678 magic | Any | Any | 679 image-ok | Any | Any | 680 copy-done | Any | Any | 681 -----------------+--------------+----------------' 682 result: BOOT_SWAP_TYPE_NONE, | 683 BOOT_SWAP_TYPE_FAIL, or | 684 BOOT_SWAP_TYPE_PANIC | 685 -------------------------------------------------' 686``` 687 688In State IV, when no errors occur, MCUboot will attempt to boot the contents of 689the primary slot directly, and the result is `BOOT_SWAP_TYPE_NONE`. If the image 690in the primary slot is not valid, the result is `BOOT_SWAP_TYPE_FAIL`. If a 691fatal error occurs during boot, the result is `BOOT_SWAP_TYPE_PANIC`. If the 692result is either `BOOT_SWAP_TYPE_FAIL` or `BOOT_SWAP_TYPE_PANIC`, MCUboot hangs 693rather than booting an invalid or compromised image. 694 695--- 696***Note*** 697 698*An important caveat to the above is the result when a swap is requested* 699*and the image in the secondary slot fails to validate, due to a hashing or* 700*signing error. This state behaves as State IV with the extra action of* 701*marking the image in the primary slot as "OK", to prevent further attempts* 702*to swap.* 703 704--- 705 706### [Resumed swaps](#resumed-swaps) 707 708If MCUboot determines that it is resuming an interrupted swap (i.e., a reset 709occurred mid-swap), it fully determines the operation to resume by reading the 710`swap info` field from the active trailer and extracting the swap type from bits 7110-3. The set of tables in the previous section are not necessary in the resume 712case. 713 714## [High-level operation](#high-level-operation) 715 716With the terms defined, we can now explore the bootloader's operation. First, 717a high-level overview of the boot process is presented. Then, the following 718sections describe each step of the process in more detail. 719 720Procedure: 721 7221. Inspect swap status region; is an interrupted swap being resumed? 723 + Yes: Complete the partial swap operation; skip to step 3. 724 + No: Proceed to step 2. 725 7262. Inspect image trailers; is a swap requested? 727 + Yes: 728 1. Is the requested image valid (integrity and security check)? 729 + Yes. 730 a. Perform swap operation. 731 b. Persist completion of swap procedure to image trailers. 732 c. Proceed to step 3. 733 + No. 734 a. Erase invalid image. 735 b. Persist failure of swap procedure to image trailers. 736 c. Proceed to step 3. 737 738 + No: Proceed to step 3. 739 7403. Boot into image in primary slot. 741 742### [Multiple image boot](#multiple-image-boot) 743 744When the flash contains multiple executable images the bootloader's operation 745is a bit more complex but similar to the previously described procedure with 746one image. Every image can be updated independently therefore the flash is 747partitioned further to arrange two slots for each image. 748``` 749+--------------------+ 750| MCUboot | 751+--------------------+ 752 ~~~~~ <- memory might be not contiguous 753+--------------------+ 754| Image 0 | 755| primary slot | 756+--------------------+ 757| Image 0 | 758| secondary slot | 759+--------------------+ 760 ~~~~~ <- memory might be not contiguous 761+--------------------+ 762| Image N | 763| primary slot | 764+--------------------+ 765| Image N | 766| secondary slot | 767+--------------------+ 768| Scratch | 769+--------------------+ 770``` 771MCUboot is also capable of handling dependencies between images. For example 772if an image needs to be reverted it might be necessary to revert another one too 773(e.g. due to API incompatibilities) or simply to prevent from being updated 774because of an unsatisfied dependency. Therefore all aborted swaps have to be 775completed and all the swap types have to be determined for each image before 776the dependency checks. Dependency handling is described in more detail in a 777following section. The multiple image boot procedure is organized in loops which 778iterate over all the firmware images. The high-level overview of the boot 779process is presented below. 780 781+ Loop 1. Iterate over all images 782 1. Inspect swap status region of current image; is an interrupted swap being 783 resumed? 784 + Yes: 785 + Review the validity of previously determined swap types 786 of other images. 787 + Complete the partial swap operation. 788 + Mark the swap type as `None`. 789 + Skip to next image. 790 + No: Proceed to step 2. 791 792 2. Inspect image trailers in the primary and secondary slot; is an image 793 swap requested? 794 + Yes: Review the validity of previously determined swap types of other 795 images. Is the requested image valid (integrity and security 796 check)? 797 + Yes: 798 + Set the previously determined swap type for the current image. 799 + Skip to next image. 800 + No: 801 + Erase invalid image. 802 + Persist failure of swap procedure to image trailers. 803 + Mark the swap type as `Fail`. 804 + Skip to next image. 805 + No: 806 + Mark the swap type as `None`. 807 + Skip to next image. 808 809+ Loop 2. Iterate over all images 810 1. Does the current image depend on other image(s)? 811 + Yes: Are all the image dependencies satisfied? 812 + Yes: Skip to next image. 813 + No: 814 + Modify swap type depending on what the previous type was. 815 + Restart dependency check from the first image. 816 + No: Skip to next image. 817 818+ Loop 3. Iterate over all images 819 1. Is an image swap requested? 820 + Yes: 821 + Perform image update operation. 822 + Persist completion of swap procedure to image trailers. 823 + Skip to next image. 824 + No: Skip to next image. 825 826+ Loop 4. Iterate over all images 827 1. Validate image in the primary slot (integrity and security check) or 828 at least do a basic sanity check to avoid booting into an empty flash 829 area. 830 831+ Boot into image in the primary slot of the 0th image position\ 832 (other image in the boot chain is started by another image). 833 834### [Multiple image boot for RAM loading and direct-xip](#multiple-image-boot-for-ram-loading-and-direct-xip) 835 836The operation of the bootloader is different when the ram-load or the 837direct-xip strategy is chosen. The flash map is very similar to the swap 838strategy but there is no need for Scratch area. 839 840+ Loop 1. Until all images are loaded and all dependencies are satisfied 841 1. Subloop 1. Iterate over all images 842 + Does any of the slots contain an image? 843 + Yes: 844 + Choose the newer image. 845 + Copy it to RAM in case of ram-load strategy. 846 + Validate the image (integrity and security check). 847 + If validation fails delete the image from flash and try the other 848 slot. (Image must be deleted from RAM too in case of ram-load 849 strategy.) 850 + No: Return with failure. 851 852 2. Subloop 2. Iterate over all images 853 + Does the current image depend on other image(s)? 854 + Yes: Are all the image dependencies satisfied? 855 + Yes: Skip to next image. 856 + No: 857 + Delete the image from RAM in case of ram-load strategy, but 858 do not delete it from flash. 859 + Try to load the image from the other slot. 860 + Restart dependency check from the first image. 861 + No: Skip to next image. 862 863+ Loop 2. Iterate over all images 864 + Increase the security counter if needed. 865 + Do the measured boot and the data sharing if needed. 866 867+ Boot the loaded slot of image 0. 868 869## [Image swapping](#image-swapping) 870 871The bootloader swaps the contents of the two image slots for two reasons: 872 873 * User has issued a "set pending" operation; the image in the secondary slot 874 should be run once (state I) or repeatedly (state II), depending on 875 whether a permanent swap was specified. 876 * Test image rebooted without being confirmed; the bootloader should 877 revert to the original image currently in the secondary slot (state III). 878 879If the image trailers indicates that the image in the secondary slot should be 880run, the bootloader needs to copy it to the primary slot. The image currently 881in the primary slot also needs to be retained in flash so that it can be used 882later. Furthermore, both images need to be recoverable if the bootloader 883resets in the middle of the swap operation. The two images are swapped 884according to the following procedure: 885 8861. Determine if both slots are compatible enough to have their images swapped. 887 To be compatible, both have to have only sectors that can fit into the 888 scratch area and if one of them has larger sectors than the other, it must 889 be able to entirely fit some rounded number of sectors from the other slot. 890 In the next steps we'll use the terminology "region" for the total amount of 891 data copied/erased because this can be any amount of sectors depending on 892 how many the scratch is able to fit for some swap operation. 8932. Iterate the list of region indices in descending order (i.e., starting 894 with the greatest index); only regions that are predetermined to be part of 895 the image are copied; current element = "index". 896 + a. Erase scratch area. 897 + b. Copy secondary_slot[index] to scratch area. 898 - If this is the last region in the slot, scratch area has a temporary 899 status area initialized to store the initial state, because the 900 primary slot's last region will have to be erased. In this case, 901 only the data that was calculated to amount to the image is copied. 902 - Else if this is the first swapped region but not the last region in 903 the slot, initialize the status area in primary slot and copy the 904 full region contents. 905 - Else, copy entire region contents. 906 + c. Write updated swap status (i). 907 + d. Erase secondary_slot[index] 908 + e. Copy primary_slot[index] to secondary_slot[index] according to amount 909 previosly copied at step b. 910 - If this is not the last region in the slot, erase the trailer in the 911 secondary slot, to always use the one in the primary slot. 912 + f. Write updated swap status (ii). 913 + g. Erase primary_slot[index]. 914 + h. Copy scratch area to primary_slot[index] according to amount 915 previously copied at step b. 916 - If this is the last region in the slot, the status is read from 917 scratch (where it was stored temporarily) and written anew in the 918 primary slot. 919 + i. Write updated swap status (iii). 9203. Persist completion of swap procedure to the primary slot image trailer. 921 922The additional caveats in step 2f are necessary so that the secondary slot image 923trailer can be written by the user at a later time. With the image trailer 924unwritten, the user can test the image in the secondary slot 925(i.e., transition to state I). 926 927--- 928***Note*** 929 930*If the region being copied contains the last sector, then swap status is* 931*temporarily maintained on scratch for the duration of this operation, always* 932*using the primary slot's area otherwise.* 933 934--- 935***Note*** 936 937*The bootloader tries to copy only used sectors (based on largest image* 938*installed on any of the slots), minimizing the amount of sectors copied and* 939*reducing the amount of time required for a swap operation.* 940 941--- 942 943The particulars of step 3 vary depending on whether an image is being tested, 944permanently used, reverted or a validation failure of the secondary slot 945happened when a swap was requested: 946 947 * test: 948 o Write primary_slot.copy_done = 1 949 (swap caused the following values to be written: 950 primary_slot.magic = BOOT_MAGIC 951 secondary_slot.magic = UNSET 952 primary_slot.image_ok = Unset) 953 954 * permanent: 955 o Write primary_slot.copy_done = 1 956 (swap caused the following values to be written: 957 primary_slot.magic = BOOT_MAGIC 958 secondary_slot.magic = UNSET 959 primary_slot.image_ok = 0x01) 960 961 * revert: 962 o Write primary_slot.copy_done = 1 963 o Write primary_slot.image_ok = 1 964 (swap caused the following values to be written: 965 primary_slot.magic = BOOT_MAGIC) 966 967 * failure to validate the secondary slot: 968 o Write primary_slot.image_ok = 1 969 970After completing the operations as described above the image in the primary slot 971should be booted. 972 973## [Swap status](#swap-status) 974 975The swap status region allows the bootloader to recover in case it restarts in 976the middle of an image swap operation. The swap status region consists of a 977series of single-byte records. These records are written independently, and 978therefore must be padded according to the minimum write size imposed by the 979flash hardware. In the below figure, a min-write-size of 1 is assumed for 980simplicity. The structure of the swap status region is illustrated below. In 981this figure, a min-write-size of 1 is assumed for simplicity. 982 983``` 984 0 1 2 3 985 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 986 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 987 |sec127,state 0 |sec127,state 1 |sec127,state 2 |sec126,state 0 | 988 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 989 |sec126,state 1 |sec126,state 2 |sec125,state 0 |sec125,state 1 | 990 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 991 |sec125,state 2 | | 992 +-+-+-+-+-+-+-+-+ + 993 ~ ~ 994 ~ [Records for indices 124 through 1 ~ 995 ~ ~ 996 ~ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 997 ~ |sec000,state 0 |sec000,state 1 |sec000,state 2 | 998 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 999``` 1000 1001The above is probably not helpful at all; here is a description in English. 1002 1003Each image slot is partitioned into a sequence of flash sectors. If we were to 1004enumerate the sectors in a single slot, starting at 0, we would have a list of 1005sector indices. Since there are two image slots, each sector index would 1006correspond to a pair of sectors. For example, sector index 0 corresponds to 1007the first sector in the primary slot and the first sector in the secondary slot. 1008Finally, reverse the list of indices such that the list starts with index 1009`BOOT_MAX_IMG_SECTORS - 1` and ends with 0. The swap status region is a 1010representation of this reversed list. 1011 1012During a swap operation, each sector index transitions through four separate 1013states: 1014``` 10150. primary slot: image 0, secondary slot: image 1, scratch: N/A 10161. primary slot: image 0, secondary slot: N/A, scratch: image 1 (1->s, erase 1) 10172. primary slot: N/A, secondary slot: image 0, scratch: image 1 (0->1, erase 0) 10183. primary slot: image 1, secondary slot: image 0, scratch: N/A (s->0) 1019``` 1020 1021Each time a sector index transitions to a new state, the bootloader writes a 1022record to the swap status region. Logically, the bootloader only needs one 1023record per sector index to keep track of the current swap state. However, due 1024to limitations imposed by flash hardware, a record cannot be overwritten when 1025an index's state changes. To solve this problem, the bootloader uses three 1026records per sector index rather than just one. 1027 1028Each sector-state pair is represented as a set of three records. The record 1029values map to the above four states as follows 1030 1031``` 1032 | rec0 | rec1 | rec2 1033 --------+------+------+------ 1034 state 0 | 0xff | 0xff | 0xff 1035 state 1 | 0x01 | 0xff | 0xff 1036 state 2 | 0x01 | 0x02 | 0xff 1037 state 3 | 0x01 | 0x02 | 0x03 1038``` 1039 1040The swap status region can accommodate `BOOT_MAX_IMG_SECTORS` sector indices. 1041Hence, the size of the region, in bytes, is 1042`BOOT_MAX_IMG_SECTORS * min-write-size * 3`. The only requirement for the index 1043count is that it is great enough to account for a maximum-sized image 1044(i.e., at least as great as the total sector count in an image slot). If a 1045device's image slots have been configured with `BOOT_MAX_IMG_SECTORS: 128` and 1046use less than 128 sectors, the first record that gets written will be somewhere 1047in the middle of the region. For example, if a slot uses 64 sectors, the first 1048sector index that gets swapped is 63, which corresponds to the exact halfway 1049point within the region. 1050 1051--- 1052***Note*** 1053 1054*Since the scratch area only ever needs to record swapping of the last* 1055*sector, it uses at most min-write-size * 3 bytes for its own status area.* 1056 1057--- 1058 1059## [Reset recovery](#reset-recovery) 1060 1061If the bootloader resets in the middle of a swap operation, the two images may 1062be discontiguous in flash. Bootutil recovers from this condition by using the 1063image trailers to determine how the image parts are distributed in flash. 1064 1065The first step is determine where the relevant swap status region is located. 1066Because this region is embedded within the image slots, its location in flash 1067changes during a swap operation. The below set of tables map image trailers 1068contents to swap status location. In these tables, the "source" field 1069indicates where the swap status region is located. In case of multi image boot 1070the images primary area and the single scratch area is always examined in pairs. 1071If swap status found on scratch area then it might not belong to the current 1072image. The swap_info field of swap status stores the corresponding image number. 1073If it does not match then "source: none" is returned. 1074 1075``` 1076 | primary slot | scratch | 1077 ----------+--------------+--------------| 1078 magic | Good | Any | 1079 copy-done | 0x01 | N/A | 1080 ----------+--------------+--------------' 1081 source: none | 1082 ----------------------------------------' 1083 1084 | primary slot | scratch | 1085 ----------+--------------+--------------| 1086 magic | Good | Any | 1087 copy-done | 0xff | N/A | 1088 ----------+--------------+--------------' 1089 source: primary slot | 1090 ----------------------------------------' 1091 1092 | primary slot | scratch | 1093 ----------+--------------+--------------| 1094 magic | Any | Good | 1095 copy-done | Any | N/A | 1096 ----------+--------------+--------------' 1097 source: scratch | 1098 ----------------------------------------' 1099 1100 | primary slot | scratch | 1101 ----------+--------------+--------------| 1102 magic | Unset | Any | 1103 copy-done | 0xff | N/A | 1104 ----------+--------------+--------------| 1105 source: primary slot | 1106 ----------------------------------------+------------------------------+ 1107 This represents one of two cases: | 1108 o No swaps ever (no status to read, so no harm in checking). | 1109 o Mid-revert; status in the primary slot. | 1110 For this reason we assume the primary slot as source, to trigger a | 1111 check of the status area and find out if there was swapping under way. | 1112 -----------------------------------------------------------------------' 1113``` 1114 1115If the swap status region indicates that the images are not contiguous, MCUboot 1116determines the type of swap operation that was interrupted by reading the `swap 1117info` field in the active image trailer and extracting the swap type from bits 11180-3 then resumes the operation. In other words, it applies the procedure defined 1119in the previous section, moving image 1 into the primary slot and image 0 into 1120the secondary slot. If the boot status indicates that an image part is present 1121in the scratch area, this part is copied into the correct location by starting 1122at step e or step h in the area-swap procedure, depending on whether the part 1123belongs to image 0 or image 1. 1124 1125After the swap operation has been completed, the bootloader proceeds as though 1126it had just been started. 1127 1128## [Integrity check](#integrity-check) 1129 1130An image is checked for integrity immediately before it gets copied into the 1131primary slot. If the bootloader doesn't perform an image swap, then it can 1132perform an optional integrity check of the image in the primary slot if 1133`MCUBOOT_VALIDATE_PRIMARY_SLOT` is set, otherwise it doesn't perform an 1134integrity check. 1135 1136During the integrity check, the bootloader verifies the following aspects of 1137an image: 1138 1139 * 32-bit magic number must be correct (`IMAGE_MAGIC`). 1140 * Image must contain an `image_tlv_info` struct, identified by its magic 1141 (`IMAGE_TLV_PROT_INFO_MAGIC` or `IMAGE_TLV_INFO_MAGIC`) exactly following 1142 the firmware (`hdr_size` + `img_size`). If `IMAGE_TLV_PROT_INFO_MAGIC` is 1143 found then after `ih_protect_tlv_size` bytes, another `image_tlv_info` 1144 with magic equal to `IMAGE_TLV_INFO_MAGIC` must be present. 1145 * Image must contain a SHA256 TLV. 1146 * Calculated SHA256 must match SHA256 TLV contents. 1147 * Image *may* contain a signature TLV. If it does, it must also have a 1148 KEYHASH TLV with the hash of the key that was used to sign. The list of 1149 keys will then be iterated over looking for the matching key, which then 1150 will then be used to verify the image contents. 1151 1152For low performance MCU's where the validation is a heavy process at boot 1153(~1-2 seconds on a arm-cortex-M0), the `MCUBOOT_VALIDATE_PRIMARY_SLOT_ONCE` 1154could be used. This option will cache the validation result as described above 1155into the magic area of the primary slot. The next boot, the validation will be 1156skipped if the previous validation was succesfull. This option is reducing the 1157security level since if an attacker could modify the contents of the flash after 1158a good image has been validated, the attacker could run his own image without 1159running validation again. Enabling this option should be done with care. 1160 1161## [Security](#security) 1162 1163As indicated above, the final step of the integrity check is signature 1164verification. The bootloader can have one or more public keys embedded in it 1165at build time. During signature verification, the bootloader verifies that an 1166image was signed with a private key that corresponds to the embedded KEYHASH 1167TLV. 1168 1169For information on embedding public keys in the bootloader, as well as 1170producing signed images, see: [signed_images](signed_images.md). 1171 1172If you want to enable and use encrypted images, see: 1173[encrypted_images](encrypted_images.md). 1174 1175--- 1176***Note*** 1177 1178*Image encryption is not supported when the direct-xip upgrade strategy* 1179*is selected.* 1180 1181--- 1182 1183### [Using hardware keys for verification](#hw-key-support) 1184 1185By default, the whole public key is embedded in the bootloader code and its 1186hash is added to the image manifest as a KEYHASH TLV entry. As an alternative 1187the bootloader can be made independent of the keys (avoiding the incorporation 1188of the public key into the code) by using one of the following options: 1189`MCUBOOT_HW_KEY` or `MCUBOOT_BUILTIN_KEY`. 1190 1191Using any of these options makes MCUboot independent from the public key(s). 1192The key(s) can be provisioned any time and by different parties. 1193 1194Hardware KEYs support options details: 1195- `MCUBOOT_HW_KEY`: In this case the hash of the public key must be 1196provisioned to the target device and MCUboot must be able to retrieve the 1197key-hash from there. For this reason the target must provide a definition 1198for the `boot_retrieve_public_key_hash()` function which is declared in 1199`boot/bootutil/include/bootutil/sign_key.h`. It is also required to use 1200the `full` option for the `--public-key-format` imgtool argument in order to 1201add the whole public key (PUBKEY TLV) to the image manifest instead of its 1202hash (KEYHASH TLV). During boot the public key is validated before using it for 1203signature verification, MCUboot calculates the hash of the public key from the 1204TLV area and compares it with the key-hash that was retrieved from the device. 1205- `MCUBOOT_BUILTIN_KEY`: With this option the whole public key(s) used for 1206signature verification must be provisioned to the target device and the used 1207[cryptographic library](PORTING.md) must support the usage of builtin keys based 1208on key IDs. In this case, neither the code nor the image metadata needs to 1209contain any public key data. During image validation only a key ID is passed to 1210the verifier function. The key handling is entirely the responsibility of the 1211crypto library and the details of the key handling mechanism are abstracted away 1212from the boot code.\ 1213***Note:*** *At the moment the usage of builtin keys is only available with the* 1214*PSA Crypto API based crypto backend (`MCUBOOT_USE_PSA_CRYPTO`) for ECDSA* 1215*signatures.* 1216 1217## [Protected TLVs](#protected-tlvs) 1218 1219If the TLV area contains protected TLV entries, by beginning with a `struct 1220image_tlv_info` with a magic value of `IMAGE_TLV_PROT_INFO_MAGIC` then the 1221data of those TLVs must also be integrity and authenticity protected. Beyond 1222the full size of the protected TLVs being stored in the `image_tlv_info`, 1223the size of the protected TLVs together with the size of the `image_tlv_info` 1224struct itself are also saved in the `ih_protected_size` field inside the 1225header. 1226 1227Whenever an image has protected TLVs the SHA256 has to be calculated over 1228not just the image header and the image but also the TLV info header and the 1229protected TLVs. 1230 1231``` 1232A +---------------------+ 1233 | Header | <- struct image_header 1234 +---------------------+ 1235 | Payload | 1236 +---------------------+ 1237 | TLV area | 1238 | +-----------------+ | struct image_tlv_info with 1239 | | TLV area header | | <- IMAGE_TLV_PROT_INFO_MAGIC (optional) 1240 | +-----------------+ | 1241 | | Protected TLVs | | <- Protected TLVs (struct image_tlv) 1242B | +-----------------+ | 1243 | | TLV area header | | <- struct image_tlv_info with IMAGE_TLV_INFO_MAGIC 1244C | +-----------------+ | 1245 | | SHA256 hash | | <- hash from A - B (struct image_tlv) 1246D | +-----------------+ | 1247 | | Keyhash | | <- indicates which pub. key for sig (struct image_tlv) 1248 | +-----------------+ | 1249 | | Signature | | <- signature from C - D (struct image_tlv), only hash 1250 | +-----------------+ | 1251 +---------------------+ 1252``` 1253 1254## [Dependency check](#dependency-check) 1255 1256MCUboot can handle multiple firmware images. It is possible to update them 1257independently but in many cases it can be desired to be able to describe 1258dependencies between the images (e.g. to ensure API compliance and avoid 1259interoperability issues). 1260 1261The dependencies between images can be described with additional TLV entries in 1262the protected TLV area after the end of an image. There can be more than one 1263dependency entry, but in practice if the platform only supports two individual 1264images then there can be maximum one entry which reflects to the other image. 1265 1266At the phase of dependency check all aborted swaps are finalized if there were 1267any. During the dependency check the bootloader verifies whether the image 1268dependencies are all satisfied. If at least one of the dependencies of an image 1269is not fulfilled then the swap type of that image has to be modified 1270accordingly and the dependency check needs to be restarted. This way the number 1271of unsatisfied dependencies will decrease or remain the same. There is always at 1272least 1 valid configuration. In worst case, the system returns to the initial 1273state after dependency check. 1274 1275For more information on adding dependency entries to an image, 1276see: [imgtool](imgtool.md). 1277 1278## [Downgrade prevention](#downgrade-prevention) 1279 1280Downgrade prevention is a feature which enforces that the new image must have a 1281higher version/security counter number than the image it is replacing, thus 1282preventing the malicious downgrading of the device to an older and possibly 1283vulnerable version of its firmware. 1284 1285### [Software-based downgrade prevention](#sw-downgrade-prevention) 1286 1287During the software based downgrade prevention the image version numbers are 1288compared. This feature is enabled with the `MCUBOOT_DOWNGRADE_PREVENTION` 1289option. In this case downgrade prevention is only available when the 1290overwrite-based image update strategy is used (i.e. `MCUBOOT_OVERWRITE_ONLY` 1291is set). 1292 1293### [Hardware-based downgrade prevention](#hw-downgrade-prevention) 1294 1295Each signed image can contain a security counter in its protected TLV area, which 1296can be added to the image using the `-s` option of the [imgtool](imgtool.md) script. 1297During the hardware based downgrade prevention (alias rollback protection) the 1298new image's security counter will be compared with the currently active security 1299counter value which must be stored in a non-volatile and trusted component of 1300the device. It is beneficial to handle this counter independently from image 1301version number: 1302 1303 * It does not need to increase with each software release, 1304 * It makes it possible to do software downgrade to some extent: if the 1305 security counter has the same value in the older image then it is accepted. 1306 1307It is an optional step of the image validation process and can be enabled with 1308the `MCUBOOT_HW_ROLLBACK_PROT` config option. When enabled, the target must 1309provide an implementation of the security counter interface defined in 1310`boot/bootutil/include/security_cnt.h`. 1311 1312## [Measured boot and data sharing](#boot-data-sharing) 1313 1314MCUboot defines a mechanism for sharing boot status information (also known as 1315measured boot) and an interface for sharing application specific information 1316with the runtime software. If any of these are enabled the target must provide 1317a shared data area between the bootloader and runtime firmware and define the 1318following parameters: 1319 1320```c 1321#define MCUBOOT_SHARED_DATA_BASE <area_base_addr> 1322#define MCUBOOT_SHARED_DATA_SIZE <area_size_in_bytes> 1323``` 1324 1325In the shared memory area all data entries are stored in a type-length-value 1326(TLV) format. Before adding the first data entry, the whole area is overwritten 1327with zeros and a TLV header is added at the beginning of the area during an 1328initialization phase. This TLV header contains a `tlv_magic` field with a value 1329of `SHARED_DATA_TLV_INFO_MAGIC` and a `tlv_tot_len` field which is indicating 1330the total length of shared TLV area including this header. The header is 1331followed by the the data TLV entries which are composed from a 1332`shared_data_tlv_entry` header and the data itself. In the data header there is 1333a `tlv_type` field which identifies the consumer of the entry (in the runtime 1334software) and specifies the subtype of that data item. More information about 1335the `tlv_type` field and data types can be found in the 1336`boot/bootutil/include/bootutil/boot_status.h` file. The type is followed by a 1337`tlv_len` field which indicates the size of the data entry in bytes, not 1338including the entry header. After this header structure comes the actual data. 1339 1340```c 1341/** Shared data TLV header. All fields in little endian. */ 1342struct shared_data_tlv_header { 1343 uint16_t tlv_magic; 1344 uint16_t tlv_tot_len; /* size of whole TLV area (including this header) */ 1345}; 1346 1347/** Shared data TLV entry header format. All fields in little endian. */ 1348struct shared_data_tlv_entry { 1349 uint16_t tlv_type; 1350 uint16_t tlv_len; /* TLV data length (not including this header). */ 1351}; 1352``` 1353 1354The measured boot can be enabled with the `MCUBOOT_MEASURED_BOOT` config option. 1355When enabled, the `--boot_record` argument of the imgtool script must also be 1356used during the image signing process to add a BOOT_RECORD TLV to the image 1357manifest. This TLV contains the following attributes/measurements of the 1358image in CBOR encoded format: 1359 1360 * Software type (role of the software component) 1361 * Software version 1362 * Signer ID (identifies the signing authority) 1363 * Measurement value (hash of the image) 1364 * Measurement type (algorithm used to calculate the measurement value) 1365 1366The `sw_type` string that is passed as the `--boot_record` option's parameter 1367will be the value of the "Software type" attribute in the generated BOOT_RECORD 1368TLV. The target must also define the `MAX_BOOT_RECORD_SZ` macro which indicates 1369the maximum size of the CBOR encoded boot record in bytes. 1370During boot, MCUboot will look for these TLVs (in case of multiple images) in 1371the manifests of the active images (the latest and validated) and copy the CBOR 1372encoded binary data to the shared data area. Preserving all these image 1373attributes from the boot stage for use by later runtime services (such as an 1374attestation service) is known as a measured boot. 1375 1376Setting the `MCUBOOT_DATA_SHARING` option enables the sharing of application 1377specific data using the same shared data area as for the measured boot. For 1378this, the target must provide a definition for the `boot_save_shared_data()` 1379function which is declared in `boot/bootutil/include/bootutil/boot_record.h`. 1380The `boot_add_data_to_shared_area()` function can be used for adding new TLV 1381entries to the shared data area. Alternatively, setting the 1382`MCUBOOT_DATA_SHARING_BOOTINFO` option will provide a default function for 1383this which saves information such as the maximum application size, bootloader 1384version (if available), running slot number, if recovery is part of MCUboot 1385and the signature type. Details of the TLVs for this information can be found 1386in `boot/bootutil/include/bootutil/boot_status.h` with `BLINFO_` prefixes. 1387 1388## [Testing in CI](#testing-in-ci) 1389 1390### [Testing Fault Injection Hardening (FIH)](#testing-fih) 1391 1392The CI currently tests the Fault Injection Hardening feature of MCUboot by 1393executing instruction skip during execution, and looking at whether a corrupted 1394image was booted by the bootloader or not. 1395 1396The main idea is that instruction skipping can be automated by scripting a 1397debugger to automatically execute the following steps: 1398 1399- Set breakpoint at specified address. 1400- Continue execution. 1401- On breakpoint hit increase the Program Counter. 1402- Continue execution. 1403- Detach from target after a timeout reached. 1404 1405Whether or not the corrupted image was booted or not can be decided by looking 1406for certain entries in the log. 1407 1408As MCUboot is deployed on a microcontroller, testing FI would not make much 1409sense in the simulator environment running on a host machine with different 1410architecture than the MCU's, as the degree of hardening depends on compiler 1411behavior. For example, (a bit counterintuitively) the code produced by gcc 1412with `-O0` optimisation is more resilient against FI attacks than the code 1413generated with `-O3` or `-Os` optimizations. 1414 1415To run on a desired architecture in the CI, the tests need to be executed on an 1416emulator (as real devices are not available in the CI environment). For this 1417implementation QEMU is selected. 1418 1419For the tests MCUboot needs a set of drivers and an implementation of a main 1420function. For the purpose of this test Trusted-Firmware-M has been selected as 1421it supports Armv8-M platforms that are also emulated by QEMU. 1422 1423The tests run in a docker container inside the CI VMs, to make it more easy to 1424deploy build and test environment (QEMU, compilers, interpreters). The CI VMs 1425seems to be using quite old Ubuntu (16.04). 1426 1427The sequence of the testing is the following (pseudo code): 1428 1429```sh 1430fn main() 1431 # Implemented in ci/fih-tests_install.sh 1432 generate_docker_image(Dockerfile) 1433 1434 # See details below. Implemented in ci/fih-tests_run.sh. 1435 # Calling the function with different parameters is done by Travis CI based on 1436 # the values provided in the .travis.yaml 1437 start_docker_image(skip_sizes, build_type, damage_type, fih_level) 1438 1439fn start_docker_image(skip_sizes, build_type, damage_type, fih_level) 1440 # implemented in ci/fih_test_docker/execute_test.sh 1441 compile_mcuboot(build_type) 1442 1443 # implemented in ci/fih_test_docker/damage_image.py 1444 damage_image(damage_type) 1445 1446 # implemented in ci/fih_test_docker/run_fi_test.sh 1447 ranges = generate_address_ranges() 1448 for s in skip_sizes 1449 for r in ranges 1450 do_skip_in_qemu(s, r) # See details below 1451 evaluate_logs() 1452 1453fn do_skip_in_qemu(size, range) 1454 for a in r 1455 run_qemu(a, size) # See details below 1456 1457# this part is implemented in ci/fih_test_docker/fi_tester_gdb.sh 1458fn run_qemu(a, size) 1459 script = create_debugger_script(a, size) 1460 start_qemu_in_bacground() # logs serial out to a file 1461 gdb_attach_to_qemu(script) 1462 kill_qemu() 1463 1464 # This checks the debugger and the quemu logs, and decides whether the tets 1465 # was executed successfully, and whether the image is booted or not. Then 1466 # emits a yaml fragment on the standard out to be processed by the caller 1467 # script 1468 evaluate_run(qemu_log_file) 1469``` 1470 1471Further notes: 1472 1473- The image is corrupted by changing its signature. 1474- MCUBOOT_FIH_PROFILE_MAX is not tested as it requires TRNG, and the AN521 1475platform has no support for it. However this profile adds the random 1476execution delay to the code, so should not affect the instruction skip results 1477too much, because break point is placed at exact address. But in practice this 1478makes harder the accurate timing of the attack. 1479- The test cases defined in .travis.yml always return `passed`, if they were 1480executed successfully. A yaml file is created during test execution with the 1481details of the test execution results. A summary of the collected results is 1482printed in the log at the end of the test. 1483 1484An advantage of having the tests running in a docker image is that it is 1485possible to run the tests on a local machine that has git and docker, without 1486installing any additional software. 1487 1488So, running the test on the host looks like the following (The commands below 1489are issued from the MCUboot source directory): 1490 1491```sh 1492$ mkdir docker 1493$ ./ci/fih-tests_install.sh 1494$ FIH_LEVEL=MEDIUM BUILD_TYPE=RELEASE SKIP_SIZE=2 DAMAGE_TYPE=SIGNATURE \ 1495 ./ci/fih-tests_run.sh 1496``` 1497On the travis CI the environment variables in the last command are set based on 1498the configs provided in the `.travis.yaml` 1499 1500This starts the tests, however the shell that it is running in is not 1501interactive, it is not possible to examine the results of the test run. To have 1502an interactive shell where the results can be examined, the following can be 1503done: 1504 1505- The docker image needs to be built with `ci/fih-tests_install.sh` as described 1506 above. 1507- Start the docker image with the following command: 1508 `docker run -i -t mcuboot/fih-test`. 1509- Execute the test with a command similar to the following: 1510 `/root/execute_test.sh 8 RELEASE SIGNATURE MEDIUM`. After the test finishes, 1511 the shell returns, and it is possible to investigate the results. It is also 1512 possible to stop the test with _Ctrl+c_. The parameters to the 1513 `execute_test.sh` are `SKIP_SIZE`, `BUILD_TYPE`, `DAMAGE_TYPE`, `FIH_LEVEL` in 1514 order. 1515
