The trusted boot framework on Marvell Armada 38x
================================================

Contents:

1. Overview of the trusted boot
2. Terminology
3. Boot image layout
4. The secured header
5. The secured boot flow
6. Usage example
7. Work to be done
8. Bibliography

1. Overview of the trusted boot
-------------------------------

The Armada's trusted boot framework enables the SoC to cryptographically verify
a specially prepared boot image. This can be used to establish a chain of trust
from the boot firmware all the way to the OS.

To achieve this, the Armada SoC requires a specially prepared boot image, which
contains the relevant cryptographic data, as well as other information
pertaining to the boot process. Furthermore, a eFuse structure (a
one-time-writeable memory) need to be configured in the correct way.

Roughly, the secure boot process works as follows:

* Load the header block of the boot image, extract a special "root" public RSA
  key from it, and verify its SHA-256 hash against a SHA-256 stored in a eFuse
  field.
* Load an array of code signing public RSA keys from the header block, and
  verify its RSA signature (contained in the header block as well) using the
  "root" RSA key.
* Choose a code signing key, and use it to verify the header block (excluding
  the key array).
* Verify the binary image's signature (contained in the header block) using the
  code signing key.
* If all checks pass successfully, boot the image.

The chain of trust is thus as follows:

* The SHA-256 value in the eFuse field verifies the "root" public key.
* The "root" public key verifies the code signing key array.
* The selected code signing key verifies the header block and the binary image.

In the special case of building a boot image containing U-Boot as the binary
image, which employs this trusted boot framework, the following tasks need to
be addressed:

1. Creation of the needed cryptographic key material.
2. Creation of a conforming boot image containing the U-Boot image as binary
   image.
3. Burning the necessary eFuse values.

(1) will be addressed later, (2) will be taken care of by U-Boot's build
system (some user configuration is required, though), and for (3) the necessary
data (essentially a series of U-Boot commands to be entered at the U-Boot
command prompt) will be created by the build system as well.

The documentation of the trusted boot mode is contained in part 1, chapter
7.2.5 in the functional specification [1], and in application note [2].

2. Terminology
--------------

                   CSK - Code Signing Key(s): An array of RSA key pairs, which
                         are used to sign and verify the secured header and the
                         boot loader image.
                   KAK - Key Authentication Key: A RSA key pair, which is used
                         to sign and verify the array of CSKs.
          Header block - The first part of the boot image, which contains the
                         image's headers (also known as "headers block", "boot
                         header", and "image header")
                 eFuse - A one-time-writeable memory.
               BootROM - The Armada's built-in boot firmware, which is
                         responsible for verifying and starting secure images.
            Boot image - The complete image the SoC's boot firmware loads
                         (contains the header block and the binary image)
           Main header - The header in the header block containing information
                         and data pertaining to the boot process (used for both
                         the regular and secured boot processes)
          Binary image - The binary code payload of the boot image; in this
                         case the U-Boot's code (also known as "source image",
                         or just "image")
        Secured header - The specialized header in the header block that
                         contains information and data pertaining to the
                         trusted boot (also known as "security header")
     Secured boot mode - A special boot mode of the Armada SoC in which secured
                         images are verified (non-secure images won't boot);
                         the mode is activated by setting a eFuse field.
    Trusted debug mode - A special mode for the trusted boot that allows
                         debugging of devices employing the trusted boot
                         framework in a secure manner (untested in the current
                         implementation).
Trusted boot framework - The ARMADA SoC's implementation of a secure verified
                         boot process.

3. Boot image layout
--------------------

+-- Boot image --------------------------------------------+
|                                                          |
| +-- Header block --------------------------------------+ |
| | Main header                                          | |
| +------------------------------------------------------+ |
| | Secured header                                       | |
| +------------------------------------------------------+ |
| | BIN header(s)                                        | |
| +------------------------------------------------------+ |
| | REG header(s)                                        | |
| +------------------------------------------------------+ |
| | Padding                                              | |
| +------------------------------------------------------+ |
|                                                          |
| +------------------------------------------------------+ |
| | Binary image + checksum                              | |
| +------------------------------------------------------+ |
+----------------------------------------------------------+

4. The secured header
---------------------

For the trusted boot framework, a additional header is added to the boot image.
The following data are relevant for the secure boot:

                   KAK: The KAK is contained in the secured header in the form
                        of a RSA-2048 public key in DER format with a length of
                        524 bytes.
Header block signature: The RSA signature of the header block (excluding the
                        CSK array), created using the selected CSK.
Binary image signature: The RSA signature of the binary image, created using
                        the selected CSK.
             CSK array: The array of the 16 CSKs as RSA-2048 public keys in DER
                        format with a length of 8384 = 16 * 524 bytes.
   CSK block signature: The RSA signature of the CSK array, created using the
                        KAK.

NOTE: The JTAG delay, Box ID, and Flash ID header fields do play a role in the
trusted boot process to enable and configure secure debugging, but they were
not tested in the current implementation of the trusted boot in U-Boot.

5. The secured boot flow
------------------------

The steps in the boot flow that are relevant for the trusted boot framework
proceed as follows:

1) Check if trusted boot is enabled, and perform regular boot if it is not.
2) Load the secured header, and verify its checksum.
3) Select the lowest valid CSK from CSK0 to CSK15.
4) Verify the SHA-256 hash of the KAK embedded in the secured header.
5) Verify the RSA signature of the CSK block from the secured header with the
   KAK.
6) Verify the header block signature (which excludes the CSK block) from the
   secured header with the selected CSK.
7) Load the binary image to the main memory and verify its checksum.
8) Verify the binary image's RSA signature from the secured header with the
   selected CSK.
9) Continue the boot process as in the case of the regular boot.

NOTE: All RSA signatures are verified according to the PKCS #1 v2.1 standard
described in [3].

NOTE: The Box ID and Flash ID are checked after step 6, and the trusted debug
mode may be entered there, but since this mode is untested in the current
implementation, it is not described further.

6. Usage example
----------------

### Create key material

To employ the trusted boot framework, cryptographic key material needs to be
created. In the current implementation, two keys are needed to build a valid
secured boot image: The KAK private key and a CSK private key (both have to be
2048 bit RSA keys in PEM format). Note that the usage of more than one CSK is
currently not supported.

NOTE: Since the public key can be generated from the private key, it is
sufficient to store the private key for each key pair.

OpenSSL can be used to generate the needed files kwb_kak.key and kwb_csk.key
(the names of these files have to be configured, see the next section on
kwbimage.cfg settings):

openssl genrsa -out kwb_kak.key 2048
openssl genrsa -out kwb_csk.key 2048

The generated files have to be placed in the U-Boot root directory.

Alternatively, instead of copying the files, symlinks to the private keys can
be placed in the U-Boot root directory.

WARNING: Knowledge of the KAK or CSK private key would enable an attacker to
generate secured boot images containing arbitrary code. Hence, the private keys
should be carefully guarded.

### Create/Modifiy kwbimage.cfg

The Kirkwook architecture in U-Boot employs a special board-specific
configuration file (kwbimage.cfg), which controls various boot image settings
that are interpreted by the BootROM, such as the boot medium. The support the
trusted boot framework, several new options were added to faciliate
configuration of the secured boot.

The configuration file's layout has been retained, only the following new
options were added:

                KAK - The name of the KAK RSA private key file in the U-Boot
                      root directory, without the trailing extension of ".key".
                CSK - The name of the (active) CSK RSA private key file in the
                      U-Boot root directory, without the trailing extension of
                      ".key".
             BOX_ID - The BoxID to be used for trusted debugging (a integer
                      value).
           FLASH_ID - The FlashID to be used for trusted debugging (a integer
                      value).
         JTAG_DELAY - The JTAG delay to be used for trusted debugging (a
                      integer value).
          CSK_INDEX - The index of the active CSK (a integer value).
SEC_SPECIALIZED_IMG - Flag to indicate whether to include the BoxID and FlashID
                      in the image (that is, whether to use the trusted debug
                      mode or not); no parameters.
       SEC_BOOT_DEV - The boot device from which the trusted boot is allowed to
                      proceed, identified via a numeric ID. The tested values
                      are 0x34 = NOR flash, 0x31 = SDIO/MMC card; for
                      additional ID values, consult the documentation in [1].
      SEC_FUSE_DUMP - Dump the "fuse prog" commands necessary for writing the
                      correct eFuse values to a text file in the U-Boot root
                      directory. The parameter is the architecture for which to
                      dump the commands (currently only "a38x" is supported).

The parameter values may be hardcoded into the file, but it is also possible to
employ a dynamic approach of creating a Autoconf-like kwbimage.cfg.in, then
reading configuration values from Kconfig options or from the board config
file, and generating the actual kwbimage.cfg from this template using Makefile
mechanisms (see board/gdsys/a38x/Makefile as an example for this approach).

### Set config options

To enable the generation of trusted boot images, the corresponding support
needs to be activated, and a index for the active CSK needs to be selected as
well.

Furthermore, eFuse writing support has to be activated in order to burn the
eFuse structure's values (this option is just needed for programming the eFuse
structure; production boot images may disable it).

ARM architecture
 -> [*] Build image for trusted boot
    (0)   Index of active CSK
 -> [*] Enable eFuse support
    [ ]   Fake eFuse access (dry run)

### Build and test boot image

The creation of the boot image is done via the usual invocation of make (with a
suitably set CROSS_COMPILE environment variable, of course). The resulting boot
image u-boot-spl.kwb can then be tested, if so desired. The hdrparser from [5]
can be used for this purpose. To build the tool, invoke make in the
'tools/marvell/doimage_mv' directory of [5], which builds a stand-alone
hdrparser executable. A test can be conducted by calling hdrparser with the
produced boot image and the following (mandatory) parameters:

./hdrparser -k 0 -t u-boot-spl.kwb

Here we assume that the CSK index is 0 and the boot image file resides in the
same directory (adapt accordingly if needed). The tool should report that all
checksums are valid ("GOOD"), that all signature verifications succeed
("PASSED"), and, finally, that the overall test was successful
("T E S T   S U C C E E D E D" in the last line of output).

### Burn eFuse structure

+----------------------------------------------------------+
| WARNING: Burning the eFuse structure is a irreversible   |
| operation! Should wrong or corrupted values be used, the |
| board won't boot anymore, and recovery is likely         |
| impossible!                                              |
+----------------------------------------------------------+

After the build process has finished, and the SEC_FUSE_DUMP option was set in
the kwbimage.cfg was set, a text file kwb_fuses_a38x.txt should be present in
the U-Boot top-level directory. It contains all the necessary commands to set
the eFuse structure to the values needed for the used KAK digest, as well as
the CSK index, Flash ID and Box ID that were selected in kwbimage.cfg.

Sequentially executing the commands in this file at the U-Boot command prompt
will write these values to the eFuse structure.

If the SEC_FUSE_DUMP option was not set, the commands needed to burn the fuses
have to be crafted by hand. The needed fuse lines can be looked up in [1]; a
rough overview of the process is:

* Burn the KAK public key hash. The hash itself can be found in the file
  pub_kak_hash.txt in the U-Boot top-level directory; be careful to account for
  the endianness!
* Burn the CSK selection, BoxID, and FlashID
* Enable trusted boot by burning the corresponding fuse (WARNING: this must be
  the last fuse line written!)
* Lock the unused fuse lines

The command to employ is the "fuse prog" command previously enabled by setting
the corresponding configuration option.

For the trusted boot, the fuse prog command has a special syntax, since the
ARMADA SoC demands that whole fuse lines (64 bit values) have to be written as
a whole. The fuse prog command itself allows lists of 32 bit words to be
written at a time, but this is translated to a series of single 32 bit write
operations to the fuse line, where the individual 32 bit words are identified
by a "word" counter that is increased for each write.

To work around this restriction, we interpret each line to have three "words"
(0-2): The first and second words are the values to be written to the fuse
line, and the third is a lock flag, which is supposed to lock the fuse line
when set to 1. Writes to the first and second words are memoized between
function calls, and the fuse line is only really written and locked (on writing
the third word) if both words were previously set, so that "incomplete" writes
are prevented. An exception to this is a single write to the third word (index
2) without previously writing neither the first nor the second word, which
locks the fuse line without setting any value; this is needed to lock the
unused fuse lines.

As an example, to write the value 0011223344556677 to fuse line 10, we would
use the following command:

fuse prog -y 10 0 00112233 44556677 1

Here 10 is the fuse line number, 0 is the index of the first word to be
written, 00112233 and 44556677 are the values to be written to the fuse line
(first and second word) and the trailing 1 is the value for the third word
responsible for locking the line.

A "lock-only" command would look like this:

fuse prog -y 11 2 1

Here 11 is the fuse number, 2 is the index of the first word to be written
(notice that we only write to word 2 here; the third word for fuse line
locking), and the 1 is the value for the word we are writing to.

WARNING: According to application note [4], the VHV pin of the SoC must be
connected to a 1.8V source during eFuse programming, but *must* be disconnected
for normal operation. The AN [4] describes a software-controlled circuit (based
on a N-channel or P-channel FET and a free GPIO pin of the SoC) to achieve
this, but a jumper-based circuit should suffice as well. Regardless of the
chosen circuit, the issue needs to be addressed accordingly!

7. Work to be done
------------------

* Add the ability to populate more than one CSK
* Test secure debug
* Test on Armada XP

8. Bibliography
---------------

[1] ARMADA(R) 38x Family High-Performance Single/Dual CPU System on Chip
    Functional Specification; MV-S109094-00, Rev. C; August 2, 2015,
    Preliminary
[2] AN-383: ARMADA(R) 38x Families Secure Boot Mode Support; MV-S302501-00
    Rev.  A; March 11, 2015, Preliminary
[3] Public-Key Cryptography Standards (PKCS) #1: RSA Cryptography
    Specifications Version 2.1; February 2003;
    https://www.ietf.org/rfc/rfc3447.txt
[4] AN-389: ARMADA(R) VHV Power; MV-S302545-00 Rev. B; January 28, 2016,
    Released
[5] Marvell Armada 38x U-Boot support; November 25, 2015;
    https://github.com/MarvellEmbeddedProcessors/u-boot-marvell

2017-01-05, Mario Six <mario.six@gdsys.cc>