A dm-verity protected filesystem image is not hashed by U-Boot when it
is loaded; its integrity is delegated to the kernel, which validates the
filesystem on the fly against the roothash taken from the FIT dm-verity
subnode. The roothash is therefore the sole integrity anchor for the
filesystem, yet fit_config_add_hash() only adds the image node, its
hash subnodes and its cipher subnode to the signed region, leaving the
dm-verity subnode (roothash, salt and block parameters) unsigned.
An attacker able to rewrite the boot medium could then replace both the
filesystem and the roothash, recompute a matching dm-verity tree and
keep the configuration signature valid, defeating verified boot for the
root filesystem.
Add the dm-verity subnode to the list of nodes covered by the
configuration signature, both when signing (tools/image-host.c) and when
verifying (boot/image-fit-sig.c), so the roothash and salt are
authenticated together with the rest of the configuration.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Both the boot-side and host-side fit_config_add_hash() repeat the same
sequence to append a node's path to the hashed-node list three times:
for the image node, for each hash subnode and for the cipher subnode.
Extract it into a helper, fit_config_add_node(), in each file, with no
functional change.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When mkimage encounters a dm-verity subnode inside a component image
node it now automatically invokes veritysetup(8) with --no-superblock
to generate the Merkle hash tree, screen-scrapes the Root hash and Salt
from the tool output, and writes the computed properties back into the
FIT blob.
The user only needs to specify algorithm, data-block-size, and
hash-block-size in the ITS; mkimage fills in digest, salt,
num-data-blocks, and hash-start-block. Because --no-superblock is
used, hash-start-block equals num-data-blocks with no off-by-one.
The image data property is replaced with the expanded content (original
data followed directly by the hash tree) so that subsequent hash and
signature subnodes operate on the complete image.
fit_image_add_verification_data() is restructured into two passes:
dm-verity first (may grow data), then hashes and signatures.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Commit 488445cefa ("doc: Move FIT into its own directory") moved the
documentation in doc/uImage.FIT to doc/usage/fit, subsequently all
documents and example sources have been converted to reStructuredText.
Fix (almost) all of the remaining occurrences of the old path and
filenames across the tree.
The exception is doc/uImage.FIT/command_syntax_extensions.txt which
apparently has been removed entirely, or at least I was unable to
locate where that document is now.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
If realloc fails, list->strings was set to NULL and
it would create a leak. This commit ensures that if we cannot
add a string to the list, the list stays in a good state.
Signed-off-by: Francois Berder <fberder@outlook.fr>
Currently, when one wants to use an OpenSSL engine to sign a FIT image,
one needs to pass a keydir (via -k) to mkimage which will then be
prepended to the value of the key-name-hint before being passed as
key_id argument to the OpenSSL Engine API, or pass a keyfile (via -G) to
mkimage.
My OpenSSL engine only has "slots" which are not mapped like
directories, so using keydir is not proper, though I could simply have
-k '' I guess but this won't work currently with binman anyway.
Additionally, passing a keyfile (-G) when using an engine doesn't make
sense as the key is stored in the engine.
Let simply allow FIT images be signed if both keydir and keyfile are
missing but an engine is to be used.
The keyname member is already filled by looking at key-name-hint
property in the FIT and passed to the engine, which is exactly what is
needed here.
Reviewed-by: Wolfgang Wallner <wolfgang.wallner@br-automation.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
The symbol TOOLS_IMAGE_PRE_LOAD doesn't depend on TOOLS_LIBCRYPTO.
If we choose to build tools without openssl, rsa_verify_openssl()
will attempt to call the unavailable openssl library functions.
Fixes: 942c8c8e66 ("rsa: Add rsa_verify_openssl() to use openssl for host builds")
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
rsa_verify_openssl() is used in lib/rsa/rsa-verify.c to authenticate data
when building host tools.
Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
This reverts commit c8750efe02, reversing
changes made to 8c6cf8aeea.
Unfortunately these changes do not build on macOS hosts.
Signed-off-by: Tom Rini <trini@konsulko.com>
rsa_verify_openssl() is used in lib/rsa/rsa-verify.c to authenticate data
when building host tools.
Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
The PRNG implementing the random() function only has 2^31 states and
therefore is unsafe to use for cryptography. Use arc4random() instead.
Fixes: cc34f04efd ("tools: image-host.c: use random instead of rand")
Addresses-Coverity-ID: 312953 Calling risky function
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Report of the static analyzer:
1. NULL_AFTER_DEREF Pointer 'str', which is dereferenced at
image-host.c:688 by calling function 'strdup', is compared to a NULL
value at image-host.c:691.
2. NULL_AFTER_DEREF Pointer 'list', which is dereferenced at
image-host.c:689, is compared to a NULL value at image-host.c:691.
Corrections explained:
1. Checking for NULL before using pointers: The if (!list || !str) check
is now performed before calling strdup and realloc, which prevents
null pointer dereferences.
2. Checking the result of strdup: strdup can return NULL if memory
allocation fails. This also needs to be checked.
3. Checking the result of realloc: If realloc returns NULL, then memory
has not been allocated and dup must be freed to avoid memory leaks.
Triggers found by static analyzer Svace.
Signed-off-by: Anton Moryakov <ant.v.moryakov@gmail.com>
This function can only be used with FITs that use embedded data. Rename
it so this is clear.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When the initialisation vector is randomly generated, its value shall be
stored in the FIT together with the encrypted data. The changes allow to
store the IV in the FIT also in the case where the key is not stored in
the DTB but retrieved somewhere else at runtime.
Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Fix error message like this:
Can't add verification data for node 'fdt-1' (<unknown error>)
We get unknown error because we decode error as fdt error
but actually it is system error.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
U-Boot configured for verified boot with the "required" option set to
"conf" also checks scripts put in FIT images for a valid signature, and
refuses to source and run such a script if the signature for the
configuration is bad or missing. Such a script could not be packaged
before, because mkimage failed like this:
% tools/mkimage -T script -C none -d tmp/my.scr -f auto-conf -k tmp -g dev -o sha256,rsa4096 my.uimg
Failed to find any images for configuration 'conf-1/signature'
tools/mkimage Can't add hashes to FIT blob: -1
Error: Bad parameters for FIT image type
This is especially unfortunate if LEGACY_IMAGE_FORMAT is disabled as
recommended.
Listing the script configuration in a "sign-images" subnode instead,
would have added even more complexity to the already complex auto fit
generation code.
Signed-off-by: Alexander Dahl <ada@thorsis.com>
A recent refactoring in image-host.c messed up the return values of
the function that reads the encryptiong keys. This patch fixes this
and also makes sure that error output goes to stderr instead of to
stdout.
Signed-off-by: Hugo Cornelis <hugo.cornelis@essensium.com>
This patch adds a function fit_image_read_key_iv_data that checks the
return value of snprintf and allows to generate a sensible error
message when generating binary images using filenames that are too
long for the OS to handle.
This is especially relevant for automated builds such as Buildroot and
Yocto builds.
Signed-off-by: Hugo Cornelis <hugo.cornelis@essensium.com>
Commit cb9faa6f98 ("tools: Use a single target-independent config to
enable OpenSSL") introduced a target-independent configuration to build
crypto features in host tools.
But since commit 2c21256b27 ("hash: Use Kconfig to enable hashing in
host tools and SPL") the build without OpenSSL is broken, due to FIT
signature/encryption features. Add missing conditional compilation
tokens to fix this.
Signed-off-by: Paul-Erwan Rio <paulerwan.rio@gmail.com>
Tested-by: Alexander Dahl <ada@thorsis.com>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The make by default cuts off the stdout output from external tools,
so all error messages from the image-host are not shown in a make
output. Besides that, it is a common approach to use stderr stream
for error messages.
Use stderr for all error messages in image-host.
Signed-off-by: Oleksandr Suvorov <oleksandr.suvorov@foundries.io>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add messages to make it clearer which part of the FIT creation is failing.
This can happen when an invalid 'algo' property is provided in the .its
file.
Signed-off-by: Simon Glass <sjg@chromium.org>
Just like we exclude data-size, data-position, and data-offset from
fit_config_check_sig, we must exclude them while signing as well.
While we're at it, use the FIT_DATA_* defines for fit_config_check_sig
as welll.
Fixes: 8edecd3110 ("fit: Fix verification of images with external data")
Fixes: c522949a29 ("rsa: sig: fix config signature check for fit with padding")
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Putting these definitions in a header will allow signatures to be
validated independently of bootm.
Signed-off-by: Steven Lawrance <steven.lawrance@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This commit enhances mkimage to update the node
/image/pre-load/sig with the public key.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
At present mkimage displays the node information but it is not clear what
signing action was taken. Add a message that shows it. For now it only
supports showing a single signing action, since that is the common case.
Sample:
Signature written to 'sha1-basic/test.fit',
node '/configurations/conf-1/signature'
Public key written to 'sha1-basic/sandbox-u-boot.dtb',
node '/signature/key-dev'
Signed-off-by: Simon Glass <sjg@chromium.org>
Update the two functions that call add_verify_data() so that the caller
can see the node that was written to.
Signed-off-by: Simon Glass <sjg@chromium.org>
This permits to prepare FIT image description that do not hard-code the
final choice of the signature algorithm, possibly requiring the user to
patch the sources.
When -o <algo> is specified, this information is used in favor of the
'algo' property in the signature node. Furthermore, that property is set
accordingly when writing the image.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
This value is either 0 for success or -1 for error. Coverity reports that
"ret" is passed to a parameter that cannot be negative, pointing to the
condition 'if (ret < 0)'.
Adjust it to just check for non-zero and avoid showing -1 in the error
message, which is pointless. Perhaps these changes will molify Coverity.
Reported-by: Coverity (CID: 312956)
Signed-off-by: Simon Glass <sjg@chromium.org>
The return value '-ENOSPC' of fit_set_timestamp function does not match
the caller fit_image_write_sig's expection which is '-FDT_ERR_NOSPACE'.
Fix it by not calling fit_set_timestamp, but call fdt_setprop instead.
This fixes a following mkimage error:
| Can't write signature for 'signature@1' signature node in
| 'conf@imx6ull-colibri-wifi-eval-v3.dtb' conf node: <unknown error>
| mkimage Can't add hashes to FIT blob: -1
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Tested-by: Igor Opaniuk <igor.opaniuk@foundries.io>
It's not always desirable to use 'keydir' and some ad-hoc heuristics
to get the filename of the signing key. More often, just passing the
filename is the simpler, easier, and logical thing to do.
Since mkimage doesn't use long options, we're slowly running out of
letters. I've chosen '-G' because it was available.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The propoerty sign-images points to images in the configuration
node. But thoses images may references severals "sub-images" (for
example for images loadable). This commit adds the support of
severals sub-images.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
This commit creates a function fit_config_add_hash that will be
used in the next commit to support several 'sub-images'.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
In the function get_random_data, strerrno is called with
the variable ret (which is the return of the function
clock_gettime). It should be called with errnor. This
commit fixes this mistake.
Reported-by: Coverity (CID: 312956)
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
According to the manpage of rand, it is recommended
to use random instead of rand. This commit updates
the function get_random_data to use random.
Reported-by: Coverity (CID: 312953)
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When building on a 32bit host the following warning occurs:
tools/image-host.c: In function ‘fit_image_read_data’:
tools/image-host.c:296:56: warning: format ‘%ld’ expects argument of
type ‘long int’, but argument 3 has type ‘__off64_t’
{aka ‘long long int’} [-Wformat=]
printf("File %s don't have the expected size (size=%ld, expected=%d)\n",
~~^
%lld
filename, sbuf.st_size, expected_size);
~~~~~~~~~~~~
tools/image-host.c:311:62: warning: format ‘%ld’ expects argument of
type ‘long int’, but argument 4 has type ‘__off64_t’
{aka ‘long long int’} [-Wformat=]
printf("Can't read all file %s (read %zd bytes, expexted %ld)\n",
~~^
%lld
filename, n, sbuf.st_size);
~~~~~~~~~~~~
Fix the format strings.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Binaries may be encrypted in a FIT image with AES. This
algo needs a key and an IV (Initialization Vector). The
IV is provided in a file (pointer by iv-name-hint in the
ITS file) when building the ITB file.
This commits adds provide an alternative way to manage
the IV. If the property iv-name-hint is not provided in
the ITS file, the tool mkimage will generate an random
IV and store it in the FIT image.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
This patch addresses issue #2 for signed configurations.
-----8<-----
Including the image cipher properties in the configuration signature
prevents an attacker from modifying cipher, key or iv properties.
Signed-off-by: Patrick Oppenlander <patrick.oppenlander@gmail.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Previously, mkimage -F could be run multiple times causing already
ciphered image data to be ciphered again.
Signed-off-by: Patrick Oppenlander <patrick.oppenlander@gmail.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Also replace fdt_delprop/fdt_setprop with fdt_setprop as fdt_setprop can
replace an existing property value.
Signed-off-by: Patrick Oppenlander <patrick.oppenlander@gmail.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Daniel GolleandTom Rini
Francois BerderandTom Rini
Quentin SchulzandTom Rini
Shiji YangandTom Rini
Paul HENRYSandTom Rini
Anton MoryakovandTom Rini
Simon GlassandTom Rini
Alexander KochetkovandSimon Glass
Alexander DahlandTom Rini
Hugo CornelisandTom Rini
Paul-Erwan RioandTom Rini
Oleksandr SuvorovandTom Rini
Sean AndersonandTom Rini
Steven LawranceandTom Rini
Mark KettenisandTom Rini
Philippe ReynesandTom Rini
Jan KiszkaandTom Rini
Vagrant CascadianandTom Rini
Ming LiuandTom Rini
Alexandru GagniucandTom Rini
Heinrich SchuchardtandTom Rini
Patrick OppenlanderandTom Rini