Capsule data can be loaded into the system either via UpdateCapsule
runtime service or files on a file system (of boot device).
The latter case is called "capsules on disk", and actual updates will
take place at the next boot time.
In this commit, we will support capsule on disk mechanism.
Please note that U-Boot itself has no notion of "boot device" and
all the capsule files to be executed will be detected only if they
are located in a specific directory, \EFI\UpdateCapsule, on a device
that is identified as a boot device by "BootXXXX" variables.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
In this commit, skeleton functions for capsule-related API's are
added under CONFIG_EFI_UPDATE_CAPSULE configuration.
Detailed implementation for a specific capsule type will be added
in the succeeding patches.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Since U-boot EFI implementation is getting richer it makes sense to
add support for EFI_TCG2_PROTOCOL taking advantage of any hardware TPM
available on the device.
This is the initial implementation of the protocol which only adds
support for GetCapability(). It's limited in the newer and safer
TPMv2 devices.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Acked-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When calling ExitBootServices during out unit tests we should not detach
devices as we need console output for runtime tests.
Fixes: 529441ca89 ("efi_loader: Disable devices before handing over control")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This function will be used from several places in UEFI subsystem
to generate some specific form of utf-16 variable name.
For example, L"Capsule0001"
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Move function to separate module.
Use char * as argument instead of u16 *.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Having an EFI_RNG_PROTOCOL without a backing RNG device leads to failure
to boot Linux 5.8.
Only install the EFI_RNG_PROTOCOL if we have a RNG device.
Reported-by: Scott K Logan <logans@cottsay.net>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We need to check that device paths provided via UEFI variables are not
malformed.
Provide function efi_dp_check_length() to check if a device path has an
end node within a given number of bytes.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Under the current implementation, all the signatures, if any, in
a signed image must be verified before loading it.
Meanwhile, UEFI specification v2.8b section 32.5.3.3 says,
Multiple signatures are allowed to exist in the binary’s certificate
table (as per PE/COFF Section “Attribute Certificate Table”). Only
one hash or signature is required to be present in db in order to pass
validation, so long as neither the SHA-256 hash of the binary nor any
present signature is reflected in dbx.
This patch makes the semantics of signature verification compliant with
the specification mentioned above.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In this commit, efi_signature_verify(with_sigdb) will be re-implemented
using pcks7_verify_one() in order to support certificates chain, where
the signer's certificate will be signed by an intermediate CA (certificate
authority) and the latter's certificate will also be signed by another CA
and so on.
What we need to do here is to search for certificates in a signature,
build up a chain of certificates and verify one by one. pkcs7_verify_one()
handles most of these steps except the last one.
pkcs7_verify_one() returns, if succeeded, the last certificate to verify,
which can be either a self-signed one or one that should be signed by one
of certificates in "db". Re-worked efi_signature_verify() will take care
of this step.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Up to now we used the value of the bootargs environment variable as load
options in the boot manager. This is not correct. The data has to be taken
from the Boot#### variable.
Let the boot manager copy the optional data of the EFI_LOAD_OPTION as load
options to the loaded image protocol.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The bootefi bootmgr command has to set the load options for a loaded image
from the value of BootXXXX variable. If the boot manager is not used, the
value is set from the environment variable bootargs (or efi_selftest).
Factor out a common function efi_set_load_options().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We need ConvertPointer() to adjust pointers when implementing runtime
services within U-Boot.
After ExitBootServices() gd is not available anymore. So we should not use
EFI_ENTRY() and EFI_EXIT().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In case that a type of certificate in "db" or "dbx" is
EFI_CERT_X509_SHA256_GUID, it is actually not a certificate which contains
a public key for RSA decryption, but a digest of image to be loaded.
If the value matches to a value calculated from a given binary image, it is
granted for loading.
With this patch, common digest check code, which used to be used for
unsigned image verification, will be extracted from
efi_signature_verify_with_sigdb() into efi_signature_lookup_digest(), and
extra step for digest check will be added to efi_image_authenticate().
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
A signed image may have multiple signatures in
- each WIN_CERTIFICATE in authenticode, and/or
- each SignerInfo in pkcs7 SignedData (of WIN_CERTIFICATE)
In the initial implementation of efi_image_authenticate(), the criteria
of verification check for multiple signatures case is a bit ambiguous
and it may cause inconsistent result.
With this patch, we will make sure that verification check in
efi_image_authenticate() should pass against all the signatures.
The only exception would be
- the case where a digest algorithm used in signature is not supported by
U-Boot, or
- the case where parsing some portion of authenticode has failed
In those cases, we don't know how the signature be handled and should
just ignore them.
Please note that, due to this change, efi_signature_verify_with_sigdb()'s
function prototype will be modified, taking "dbx" as well as "db"
instead of outputing a "certificate." If "dbx" is null, the behavior would
be the exact same as before.
The function's name will be changed to efi_signature_verify() once
current efi_signature_verify() has gone due to further improvement
in intermediate certificates support.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Export the UEFI sub-system initialization state. This will allow to treat
the setting of UEFI variables during and after initialization differently.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
For passing the optional data of the load option to the loaded imaged
protocol we need its size.
efi_deserialize_load_option() is changed to return the size of the optional
data.
As a by-product we get a partial validation of the load option.
Checking the length of the device path remains to be implemented.
Some Coverity defects identified the load options as user input because
get_unaligned_le32() and get_unaligned_le16() is called. But non of these
Coverity defects can be resolved without marking functions with Coverity
specific tags.
Reported-by: Coverity (CID 303760)
Reported-by: Coverity (CID 303768)
Reported-by: Coverity (CID 303776)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The UEFI specification does not define a device sub-type for virtio.
Let's use a vendor hardware node here.
This avoids creation of two handles with the same device path indicating
our root node.
Reported-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Virtually all callers of this function do the rounding on their own.
Some do it right, some don't. Instead of doing this in each caller,
do the rounding in efi_add_memory_map(). Change the size parameter
to bytes instead of pages and remove aligning and size calculation in
all callers.
There is no more need to make the original efi_add_memory_map() (which
takes pages as size) available outside the module. Thus rename it to
efi_add_memory_map_pg() and make it static to prevent further misuse
outside the module.
Signed-off-by: Michael Walle <michael@walle.cc>
Add missing comma in sunxi_display.c.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This function will check if a given handle to device is an EFI system
partition. It will be utilised in implementing capsule-on-disk feature.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Add function description. Return bool.
Reviewed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The guard doesn't make any difference, so remove it.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In subsequent patches UEFI variables shalled be stored on the EFI system
partition. Hence we need to identify the EFI system partition.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
With this commit, image validation can be enforced, as UEFI specification
section 32.5 describes, if CONFIG_EFI_SECURE_BOOT is enabled.
Currently we support
* authentication based on db and dbx,
so dbx-validated image will always be rejected.
* following signature types:
EFI_CERT_SHA256_GUID (SHA256 digest for unsigned images)
EFI_CERT_X509_GUID (x509 certificate for signed images)
Timestamp-based certificate revocation is not supported here.
Internally, authentication data is stored in one of certificates tables
of PE image (See efi_image_parse()) and will be verified by
efi_image_authenticate() before loading a given image.
It seems that UEFI specification defines the verification process
in a bit ambiguous way. I tried to implement it as closely to as
EDK2 does.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
With this commit, EFI_VARIABLE_TIME_BASED_AUTHENTICATED_WRITE_ACCESS
is supported for authenticated variables and the system secure state
will transfer between setup mode and user mode as UEFI specification
section 32.3 describes.
Internally, authentication data is stored as part of authenticated
variable's value. It is nothing but a pkcs7 message (but we need some
wrapper, see efi_variable_parse_signature()) and will be validated by
efi_variable_authenticate(), hence efi_signature_verify_with_db().
Associated time value will be encoded in "{...,time=...}" along with
other UEFI variable's attributes.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
efi_signature_parse_sigdb() is a helper function will be used to parse
signature database variable and instantiate a signature store structure
in later patches.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
In this commit, implemented are a couple of helper functions which will be
used to materialize variable authentication as well as image authentication
in later patches.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
'vendor' is both an input and an output parameter. So it cannot be
constant.
Fixes: 0bda81bfdc ("efi_loader: use const efi_guid_t * for variable services")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This is a preparatory patch.
Those functions will be used in an implementation of UEFI firmware
management protocol as part of my capsule update patch.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Following kernel's proposal for an arch-agnostic initrd loading
mechanism [1] let's implement the U-boot counterpart.
This new approach has a number of advantages compared to what we did up
to now. The file is loaded into memory only when requested limiting the
area of TOCTOU attacks. Users will be allowed to place the initramfs
file on any u-boot accessible partition instead of just the ESP one.
Finally this is an attempt of a generic interface across architectures
in the linux kernel so it makes sense to support that.
The file location is intentionally only supported as a config option
argument(CONFIG_EFI_INITRD_FILESPEC), in an effort to enhance security.
Although U-boot is not responsible for verifying the integrity of the
initramfs, we can enhance the offered security by only accepting a
built-in option, which will be naturally verified by UEFI Secure Boot.
This can easily change in the future if needed and configure that via ENV
or UEFI variable.
[1] https://lore.kernel.org/linux-efi/20200207202637.GA3464906@rani.riverdale.lan/T/#m4a25eb33112fab7a22faa0fd65d4d663209af32f
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Install the EFI_RNG_PROTOCOL implementation for it's subsequent use by
the kernel for features like kaslr.
Signed-off-by: Sughosh Ganu <sughosh.ganu@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Add guidcpy function to copy the source guid to the destination
guid. Use this function instead of memcpy for copying to the
destination guid.
Signed-off-by: Sughosh Ganu <sughosh.ganu@linaro.org>
Use void * instead of efi_guid_t * for arguments to allow copying unaligned
GUIDs. The GUIDs of configuration tables are __packed.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Use a pointer to addressable memory instead of a "physical" address in the
virtual address space of the sandbox to efi_install_fdt().
Export the efi_install_fdt() function.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
As part of moving the parsing of command line arguments to do_bootefi()
call efi_install_fdt() with the address of the device tree instead of a
string.
If the address is EFI_FDT_USE_INTERNAL (= 0), the internal device tree
is used.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
At present these two functions are defined in efi_loader.h but only if
CONFIG_EFI_LOADER is enabled. But these are functions that are useful to
other code, such as that which deals with Intel Handoff Blocks (HOBs).
Move these to the top of the function.
Possibly ascii2unicode() should not be an inline function, since this
might impact code size.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Sandbox's "host" devices are currently described as UCLASS_ROOT udevice
with DEV_IF_HOST block device. As the current implementation of
efi_device_path doesn't support such a type, any "host" device
on sandbox cannot be seen as a distinct object.
For example,
=> host bind 0 /foo/disk.img
=> efi devices
Scanning disk host0...
Found 1 disks
Device Device Path
================ ====================
0000000015c19970 /VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)
0000000015c19d70 /VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)
=> efi dh
Handle Protocols
================ ====================
0000000015c19970 Device Path, Device Path To Text, Device Path Utilities, Unicode Collation 2, HII String, HII Database, HII Config Routing
0000000015c19ba0 Driver Binding
0000000015c19c10 Simple Text Output
0000000015c19c80 Simple Text Input, Simple Text Input Ex
0000000015c19d70 Block IO, Device Path, Simple File System
As you can see here, efi_root (0x0000000015c19970) and host0 device
(0x0000000015c19d70) have the same representation of device path.
This is not only inconvenient, but also confusing since two different
efi objects are associated with the same device path and
efi_dp_find_obj() will possibly return a wrong result.
Solution:
Each "host" device should be given an additional device path node
of "vendor device path" to make it distinguishable.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
ascii2unicode() can only convert characters 0x00-0x7f from UTF-8 to UTF-16.
Use utf8_utf16_strcpy() instead.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Adding a conventional memory region to the memory map may require ram_top
limitation and it can be also commonly used. Extract adding a conventional
memory to the memory map in a separate routine for generic use.
Signed-off-by: Aiden Park <aiden.park@intel.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In packed structures GUIDs are not aligned. Avoid a build error with
GCC 9.1 by using const void * as argument for guidcmp().
Reported-by: Ramon Fried <rfried.dev@gmail.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We currently have some inconsistent use of efi_add_memory_map()
throughout the code. In particular the return value of efi_add_memory_map()
is not interpreted the same way by various users in the codebase.
This patch does the following:
- Changes efi_add_memory_map() to return efi_status_t.
- Adds a method description to efi_add_memory_map().
- Changes efi_add_memory_map() to return EFI_SUCCESS
- Returns non-zero for error in efi_add_memory_map()
- Updates efi_allocate_pages() to new efi_add_memory_map()
- Updates efi_free_pages() to new efi_add_memory_map()
- Updates efi_carve_out_dt_rsv() to new efi_add_memory_map()
- Updates efi_add_runtime_mmio() to new efi_add_memory_map()
Fixes: 5d00995c36 ("efi_loader: Implement memory allocation and map")
Fixes: 74c16acce3 ("efi_loader: Don't allocate from memory holes")
Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Alexander Graf <agraf@csgraf.de>
Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Linux can be called with a command line parameter efi=novamap, cf.
commit 4e46c2a95621 ("efi/arm/arm64: Allow SetVirtualAddressMap() to be
omitted"). In this case SetVirtualAddressMap() is not called after
ExitBootServices().
OpenBSD 32bit does not call SetVirtualAddressMap() either.
Runtime services must be set to an implementation supported at runtime
in ExitBootServices().
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Suggested-by: Alexander Graf <agraf@csgraf.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Our variable services are only provided at boottime. Therefore when
leaving boottime the variable function are replaced by dummy functions
returning EFI_UNSUPPORTED. Move this patching of the runtime table to the
variable services implementation. Executed it in ExitBootServices().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Let's keep similar things together.
Move efi_query_variable_info() to lib/efi_loader/efi_variable.c
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This variable is defined in UEFI specification 2.8, section 8.1.
Its value should be updated whenever we add any usable runtime services
function.
Currently we only support SetVirtualAddress() for all systems and
ResetSystem() for some.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Up to now we have only been using a flag queued for events. But this does
not satisfy the requirements of the UEFI spec. Events must be notified in
the sequence of decreasing TPL level and within a TPL level in the sequence
of signaling.
Implement a queue for signaled events.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
ExitBootServices() has to stop timer related activity before calling the
events of the EFI_EVENT_GROUP_EXIT_BOOT_SERVICES event group. But our
current implementation was stopping all other events.
All events have to observe the task priority level.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
