Signature verification: verify the kernel against lists of hashes that are
either in dbx or db list. If it is not in the dbx list then the trusted keys
from the db list are used to verify the signature.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If secure boot is enabled with static key management mode, the trusted
certificates will be extracted from the GRUB ELF Note and added to db list.
If secure boot is enabled with dynamic key management mode, the trusted
certificates and certificate/binary hash will be extracted from the PKS
and added to db list. The distrusted certificates, certificate/binary hash
are read from the PKS and added to dbx list. Both dbx and db lists usage is
added by a subsequent patch.
Note:
- If db does not exist in the PKS storage, then read the static keys as a db
default keys from the GRUB ELF Note and add them into the db list.
- If the certificate or the certificate hash exists in the dbx list, then do not
add that certificate/certificate hash to the db list.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Introducing the appended signature key management environment variable. It is
automatically set to either "static" or "dynamic" based on the Platform KeyStore.
"static": Enforce static key management signature verification. This is the
default. When the GRUB is locked down, user cannot change the value
by setting the appendedsig_key_mgmt variable back to "dynamic".
"dynamic": Enforce dynamic key management signature verification. When the GRUB
is locked down, user cannot change the value by setting the
appendedsig_key_mgmt variable back to "static".
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Enhancing the infrastructure to enable the Platform Keystore (PKS) feature,
which provides access to the SB_VERSION, db, and dbx secure boot variables
from PKS.
If PKS is enabled, it will read secure boot variables such as db and dbx
from PKS and extract EFI Signature List (ESL) from it. The ESLs would be
saved in the Platform Keystore buffer, and the appendedsig module would
read it later to extract the certificate's details from ESL.
In the following scenarios, static key management mode will be activated:
1. When Secure Boot is enabled with static key management mode
2. When SB_VERSION is unavailable but Secure Boot is enabled
3. When PKS support is unavailable but Secure Boot is enabled
Note:
SB_VERSION: Key Management Mode
1 - Enable dynamic key management mode. Read the db and dbx variables from PKS,
and use them for signature verification.
0 - Enable static key management mode. Read keys from the GRUB ELF Note and
use it for signature verification.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Building on the parsers and the ability to embed X.509 certificates, as well
as the existing gcrypt functionality, add a module for verifying appended
signatures.
This includes a signature verifier that requires that the Linux kernel and
GRUB modules have appended signatures for verification.
Signature verification must be enabled by setting check_appended_signatures.
If secure boot is enabled with enforce mode when the appendedsig module is
loaded, signature verification will be enabled, and trusted keys will be
extracted from the GRUB ELF Note and stored in the db and locked automatically.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Read secure boot mode from 'ibm,secure-boot' property and if the secure boot
mode is set to 2 (enforce), enter lockdown. Else it is considered as disabled.
There are three secure boot modes. They are
0 - disabled
No signature verification is performed. This is the default.
1 - audit
Signature verification is performed and if signature verification fails,
display the errors and allow the boot to continue.
2 - enforce
Lockdown the GRUB. Signature verification is performed and if signature
verification fails, display the errors and stop the boot.
Now, only support disabled and enforce.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This code allows us to parse:
- X.509 certificates: at least enough to verify the signatures on the PKCS#7
messages. We expect that the certificates embedded in GRUB will be leaf
certificates, not CA certificates. The parser enforces this.
- X.509 certificates support the Extended Key Usage extension and handle it by
verifying that the certificate has a Code Signing usage.
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com> # EKU support
Reported-by: Michal Suchanek <msuchanek@suse.com> # key usage issue
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This code allows us to parse:
- PKCS#7 signed data messages. Only a single signer info is supported, which
is all that the Linux sign-file utility supports creating out-of-the-box.
Only RSA, SHA-256 and SHA-512 are supported. Any certificate embedded in
the PKCS#7 message will be ignored.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This code allows us to parse ASN1 node and allocating memory to store it.
It will work for anything where the size libtasn1 returns is right:
- Integers
- Octet strings
- DER encoding of other structures
It will _not_ work for things where libtasn1 size requires adjustment:
- Strings that require an extra NULL byte at the end
- Bit strings because libtasn1 returns the length in bits, not bytes.
If the function returns a non-NULL value, the caller must free it.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In order to parse PKCS#7 messages and X.509 certificates with libtasn1, we need
some information about how they are encoded. We get these from GNUTLS, which has
the benefit that they support the features we need and are well tested.
The GNUTLS files are from:
- https://github.com/gnutls/gnutls/blob/master/lib/gnutls.asn
- https://github.com/gnutls/gnutls/blob/master/lib/pkix.asn
The GNUTLS license is LGPLv2.1+, which is GPLv3 compatible, allowing us to import
it without issue.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Prior to the addition of the X.509 public key support for appended signature,
current PGP signature relied on the GPG public key. Changing the enum name
from "OBJ_TYPE_PUBKEY" to "OBJ_TYPE_GPG_PUBKEY" to differentiate between x509
certificate based appended signature and GPG certificate based PGP signature.
Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The way gcry_rsa and friends (the asymmetric ciphers) are loaded for the
pgp module is a bit quirky.
include/grub/crypto.h contains:
extern struct gcry_pk_spec *grub_crypto_pk_rsa;
commands/pgp.c contains the actual storage:
struct gcry_pk_spec *grub_crypto_pk_rsa;
And the module itself saves to the storage in pgp.c:
GRUB_MOD_INIT(gcry_rsa)
{
grub_crypto_pk_rsa = &_gcry_pubkey_spec_rsa;
}
This is annoying: gcry_rsa now has a dependency on pgp!
We want to be able to bring in gcry_rsa without bringing in PGP, so move the
storage to crypto.c.
Previously, gcry_rsa depended on pgp and mpi. Now it depends on crypto and mpi.
As pgp depends on crypto, this doesn't add any new module dependencies using
the PGP verfier.
[FWIW, the story is different for the symmetric ciphers. cryptodisk and friends
(zfs encryption etc) use grub_crypto_lookup_cipher_by_name() to get a cipher
handle. That depends on grub_ciphers being populated by people calling
grub_cipher_register. import_gcry.py ensures that the symmetric ciphers call it.]
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On some targets, size_t and grub_size_t may not be the same type
(unsigned long / unsigned int). This breaks the compilation because the
definition of _gpgrt_b64dec_proc() differs from gpgrt_b64dec_proc()
declaration. Fix it by using grub_size_t in the _gpgrt_b64dec_proc()
definition.
Signed-off-by: Anaëlle Cazuc <acazuc@acazuc.fr>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Some Qualcomm-based UEFI platforms only provide volume up, volume down,
and power keys. The volume keys are already mapped to SCAN_UP and SCAN_DOWN,
while the power key is mapped to SCAN_SUSPEND (key.scan_code 0x0102).
On such devices, the power key is commonly used as the Enter (confirm)
button, since no dedicated Enter key exists. This patch treats key.scan_code
0x0102 as Enter to improve usability on these platforms.
Signed-off-by: Kancy Joe <kancy2333@outlook.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A Unified Kernel Image (UKI) is a single UEFI PE file that combines
a UEFI boot stub, a Linux kernel image, an initrd, and further resources.
The uki command will locate where the UKI file is and create a GRUB menu
entry to load it.
The Unified Kernel Image Specification: https://uapi-group.org/specifications/specs/unified_kernel_image/
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Irritatingly, BLS defines paths relative to the mountpoint of the
filesystem which contains its snippets, not / or any other fixed
location. So grub-emu needs to know whether /boot is a separate
filesystem from / and conditionally prepend a path.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The BootLoaderSpec (BLS) defines a scheme where different bootloaders can
share a format for boot items and a configuration directory that accepts
these common configurations as drop-in files.
The BLS Specification: https://uapi-group.org/specifications/specs/boot_loader_specification/
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Will Thompson <wjt@endlessm.com>
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add the functions grub_strtok() and grub_strtok_r() to help parse strings into
tokens separated by characters in the "delim" parameter. These functions are
present in gnulib but calling them directly from the gnulib code is quite
challenging since the call "#include <string.h>" would include the header file
grub-core/lib/posix_wrap/string.h instead of grub-core/lib/gnulib/string.h,
where strtok() and strtok_r() are declared. Since this overlap is quite
problematic, the simpler solution was to implement the code in the GRUB based
on gnulib's implementation. For more information on these functions, visit the
Linux Programmer's Manual, man strtok.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Xen traditionally allows customizing guest behavior by passing arguments
to the VM kernel via the kernel command line. This is no longer possible
when using GRUB with Xen, as the kernel command line is decided by the
GRUB configuration file within the guest, not data passed to the guest
by Xen.
To work around this limitation, enable GRUB to parse a command line
passed to it by Xen, and expose data from the command line to the GRUB
configuration as environment variables. These variables can be used in
the GRUB configuration for any desired purpose, such as extending the
kernel command line passed to the guest. The command line format is
inspired by the Linux kernel's command line format.
To reduce the risk of misuse, abuse, or accidents in production, the
command line will only be parsed if it consists entirely of 7-bit ASCII
characters, only alphabetical characters and underscores are permitted
in variable names, and all variable names must start with the string
"xen_grub_env_". This also allows room for expanding the command line
arguments accepted by GRUB in the future, should other arguments end up
becoming desirable in the future.
Signed-off-by: Aaron Rainbolt <arraybolt3@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We need to avoid clobbering existing table between starting of chunk movers
and the moment we install target page table. Generate temporary table for
this rather than hoping that we don't clobber existing one.
Fixes 64-bit GhostBSD on 64-bit EFI.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The cmdline_size already account for NUL terminator, you can see
this in xen_boot_binary_load(). The same property is set correctly
for Xen command line.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
From UEFI specifications 2.10, section 13.2.2, EFI_LOAD_FILE2_PROTOCOL.LoadFile
(see https://uefi.org/specs/UEFI/2.10/13_Protocols_Media_Access.html), for BufferSize:
On input the size of Buffer in bytes. On output with a return code
of EFI_SUCCESS, the amount of data transferred to Buffer. On output
with a return code of EFI_BUFFER_TOO_SMALL, the size of Buffer
required to retrieve the requested file.
So, set *buffer_size correctly.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It turns out checking from userspace is not 100% reliable to figure out
whether the firmware had TPM2 support enabled or not. For example with
EDK2 arm64, the default upstream build config bundles TPM2 support with
SecureBoot support, so if the latter is disabled, TPM2 is also unavailable.
But still, the ACPI TPM2 table is created just as if it was enabled. So,
/sys/firmware/acpi/tables/TPM2 exists and looks correct but there are no
measurements, neither the firmware nor the loader/stub can do them, and
/sys/kernel/security/tpm0/binary_bios_measurements does not exist.
So, userspace cannot really tell what was going on in UEFI mode.
The loader can use the apposite UEFI protocol to check, which is a more
definitive answer. Export the bitmask with the list of active banks as-is.
If it's not 0, then in userspace we can be sure a working TPM2 was available
in UEFI mode.
systemd-boot and systemd-stub v258 (current main) set this variable and
userspace portion consumes it to be able to tell what was available in
the firmware context.
Signed-off-by: Luca Boccassi <luca.boccassi@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
libgcrypt itself is compiled with -Wno-sign-compare. Do the same for consistency.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Libgcrypt code assumes that on x64 all SSE registers are fair game.
While it's true that CPUs in question support it, we disable it in
our compilation options. Disable the offending optimization.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows us to test purely the integration of the implementation of
DSA and RSA from libgcrypt without concerning with additional code.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patches modifies the GRUB-libgcrypt API to match new libgcrypt 1.11.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We currently use an old version of libgcrypt which results in us having
fewer ciphers and missing on many other improvements.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Not reusing these handles will result in image measurements showing up
twice in the event log.
On the occasion add missing grub_free() call.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Not reusing these handles will result in image measurements showing up
twice in the event log.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Not reusing these handles will result in image measurements showing up
twice in the event log.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use loader protocol for image verification where available, otherwise
fall back to the old shim lock protocol.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These can be used to register a different implementation later,
for example, when shim provides a protocol with those functions.
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Xen image is loaded with an alignment, not always at "start".
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
More readable, could consume less space.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We need to use FOR_LIST_ELEMENTS_SAFE() instead of FOR_LIST_ELEMENTS()
as single_binary_unload(), called during the loop, is changing the list
using grub_list_remove(). Given the environment probably the old code
simply removed only the first module on the list not freeing all the others.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
In grub_cmd_initrd(), memory is allocated for variable initrd_ctx
before calling grub_relocator_alloc_chunk_align_safe(). When the
function call fails, initrd_ctx should be freed before exiting
grub_cmd_initrd().
Fixes: CID 473852
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Unload previous Linux kernel/initrd before updating the global variable
kernel_size. Otherwise the previous Linux kernel gets deallocated with
the kernel_size of the Linux kernel that is being currently loaded.
Signed-off-by: Adriano Cordova <adriano.cordova@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If a simple string for arguments are passed it should be NUL terminated.
This is true for other code but not for "linux" command.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is more consistent with the above code using sizeof(grub_efi_char16_t).
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Although the length should not exceed 2^31 grub_size_t is more
suitable for that variable. len is used to compute the size
of buffers which in C is a size_t, not a int. It is used
for GRUB_EFI_BYTES_TO_PAGES which expects unsigned values.
It is assigned to load_options_size which is unsigned, not signed.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The size passed to grub_utf8_to_utf16() for the source string is
used as a limit for the string if NUL character is not encountered.
However, len, which is "strlen(src) * 2 + 2" is surely greater than
strlen(src). Pass the exact correct length.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the following allocation fails this would leave load_options NULL
while load_options_size not valid. If the allocation succeed
load_options_size is overwritten.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
