mbed TLS provides the debug API `mbedtls_memory_buffer_alloc_status()`
to analyse the RAM usage of the library.
When RSA is selected as algorithm, the maximum heap usage in FVP and
Juno has been determined empirically to be approximately 5.5 KiB.
However, The default heap size used when RSA is selected is 8 KiB.
This patch reduces the buffer from 8 KiB to 6 KiB so that the BSS
sections of both BL1 and BL2 are 2 KiB smaller when the firmware is
compiled with TBB support.
Change-Id: I43878a4e7af50c97be9c8d027c728c8483f24fbf
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
If UFS device is initialized, we could just make it out of
hibernation by UFS_FLAGS_SKIPINIT. And vendor's dirver is always
focus on PHY setting. We could use UFS driver directly if it
exits from hibernation.
There're eight LUNs in UFS device. The UFS driver only provides
the read/write API with LUN. User could define his own read/write
since user may want to access different LUNs.
Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
The Trusted Firmware uses a subset of the APIs provided by mbed TLS.
This subset internally uses `snprintf`, but the only format specifier
used is '%d', which is supported by `tf_snprintf`.
This patch makes mbed TLS use `tf_snprintf` instead of `snprintf`,
saving 3 KB in both debug and release builds when TBBR is enabled.
Change-Id: I7f992a21015930d7c0f4660e7a28ceefd60b9597
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Platforms aligned with TBBR are supposed to use their own OIDs, but
defining the same macros with different OIDs does not provide any
value (at least technically).
For easier use of TBBR, this commit allows platforms to reuse the OIDs
obtained by ARM Ltd. This will be useful for non-ARM vendors that
do not need their own extension fields in their certificate files.
The OIDs of ARM Ltd. have been moved to include/tools_share/tbbr_oid.h
Platforms can include <tbbr_oid.h> instead of <platform_oid.h> by
defining USE_TBBR_DEFS as 1. USE_TBBR_DEFS is 0 by default to keep the
backward compatibility.
For clarification, I inserted a blank line between headers from the
include/ directory (#include <...>) and ones from a local directory
(#include "..." ).
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
This patch renames MBEDTLS_KEY_ALG to TF_MBEDTLS_KEY_ALG. This
completes the migration of TF specific macros so that they do not
have the MBEDTLS_ suffix (see arm-trusted-firmware#874).
Change-Id: Iad7632477e220b0af987c4db3cf52229fb127d00
Signed-off-by: David Cunado <david.cunado@arm.com>
An earlier patch (arm-trusted-firmware#874) migrated MBEDTLS_ suffixed
macros to have a TBBR_ suffix to avoid any potential clash with future
mbedtls macros.
But on reflection the TBBR_ suffix could be confusing as the macros
are used to drive TF-specific configuration of mbedtls. As such
this patch migrates these macros from TBBR_suffix to TF_MBEDTLS_
suffix which more accurately conveys their use.
Change-Id: Ic87642b653ceeaa03d62f724976abd5e12e867d4
Signed-off-by: David Cunado <david.cunado@arm.com>
To make software license auditing simpler, use SPDX[0] license
identifiers instead of duplicating the license text in every file.
NOTE: Files that have been imported by FreeBSD have not been modified.
[0]: https://spdx.org/
Change-Id: I80a00e1f641b8cc075ca5a95b10607ed9ed8761a
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
This patch removes the code that touched UART_FCR, from
console_core_putc(). The check for whether transmit FIFO is
full is sufficient before writing to UART TX FIFO. In fact
setting UARTFCR_TXCLR immediately after a byte is written to
FIFO might even result in loss of that byte, if UART hasn't
sent that byte out yet.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
Many asserts depend on code that is conditionally compiled based on the
DEBUG define. This patch modifies the conditional inclusion of such code
so that it is based on the ENABLE_ASSERTIONS build option.
Change-Id: I6406674788aa7e1ad7c23d86ce94482ad3c382bd
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
C files shouldn't be included into others. This file only contains some
macros and functions that can be made `static inline`, so it is ok to
convert it into a header file.
This is the only occurrence of a C file being included in another one in
the codebase instead of using a header, other occurrences are a way of
achieving backwards-compatibility.
Functions therein have been qualified as `inline`.
Change-Id: I88fe300f6d85a7f0740ef14c9cb8fa54849218e6
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
The amount of console output is controlled by the LOG_LEVEL build
option. Using tf_printf without any #ifdef depending on the LOG_LEVEL
doesn't give the user that flexibility.
This patch replaces all occurrences of tf_printf that prints error, but
aren't dependent on LOG_LEVEL, with the ERROR macro.
Change-Id: Ib5147f14fc1579398a11f19ddd0e840ff6692831
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
It is needed to add placeholders for this function because, as this is
not a `plat_xxx()` function, there aren't weak definitions of it in any
file.
If `console_flush()` is used and there isn't an implementation of
`console_core_flush()` in any file, the compilation will fail.
Change-Id: I50eb56d085c4c9fbc85d40c343e86af6412f3020
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
This function ensures that console output is flushed, for example
before shutting down or use by another component
In line with other console APIs, console_flush() wraps
console_core_flush().
Also implement console_core_flush() for PL011.
Change-Id: I3db365065e4de04a454a5c2ce21be335a23a01e4
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
These macros are not part of mbed TLS so they should not be prefixed
with `MBEDTLS_` to avoid potential collision in the future. Use the
`TBBR_` suffix to highlight that they only used in TF.
`MBEDTLS_KEY_ALG` was not modified because that is documented and used
by platforms to select the key algorithm.
Change-Id: Ief224681715c481691c80810501830ce16e210b0
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
Due to incorrect conditional compilation checks, bakery locks were
excluded from the CCN driver and the power controller driver for FVP
when BL32 was built as the EL3 Runtime Software in AArch32 mode.
This patch corrects the same.
Change-Id: Ib1f163d9167a5c38e4d622232c4835cad9c235aa
Signed-off-by: Soby Mathew <soby.mathew@arm.com>
The GIC driver data is initialized by the primary CPU with caches
enabled. When the secondary CPU boots up, it initializes the
GICC/GICR interface with the caches disabled and there is a chance that
the driver data is not yet written back to the memory. This patch fixes
this problem by flushing the driver data after they have been
initialized.
Change-Id: Ie9477029683846209593ff005d2bac559bb8f5e6
Signed-off-by: Soby Mathew <soby.mathew@arm.com>
Static checks flag an assert added in commit 1f786b0 that compares
unsigned value to 0, which will never fail.
Change-Id: I4b02031c2cfbd9a25255d12156919dda7d4805a0
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>
This reverts commit b621fb503c.
Because of the Trusted Firmware design, timing-safe functions are not
needed. Using them may be misleading as it could be interpreted as being
a protection against private data leakage, which isn't the case here.
For each image, the SHA-256 hash is calculated. Some padding is appended
and the result is encrypted with a private key using RSA-2048. This is
the signature of the image. The public key is stored along with BL1 in
read-only memory and the encrypted hash is stored in the FIP.
When authenticating an image, the TF decrypts the hash stored in the FIP
and recalculates the hash of the image. If they don't match, the boot
sequence won't continue.
A constant-time comparison does not provide additional security as all
the data involved in this process is already known to any attacker.
There is no private data that can leaked through a timing attack when
authenticating an image.
`timingsafe_bcmp()` is kept in the codebase because it could be useful
in the future.
Change-Id: I44bdcd58faa586a050cc89447e38c142508c9888
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
cppcheck highlighted variables that were initialized but then later
reassigned.
Change-Id: Ie12742c01fd3bf48b2d6c05a3b448da91d57a2e4
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
The memmap IO driver doesn't perform bounds check when reading, writing,
or seeking. The onus to vet parameters is on the caller, and this patch
asserts that:
- non-negative size is specified for for backing memory;
- valid parameters are passed into the driver for read, write and seek
operations.
Change-Id: I6518c4065817e640e9e7e39a8a4577655f2680f7
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>
Replace all use of memset by zeromem when zeroing moderately-sized
structure by applying the following transformation:
memset(x, 0, sizeof(x)) => zeromem(x, sizeof(x))
As the Trusted Firmware is compiled with -ffreestanding, it forbids the
compiler from using __builtin_memset and forces it to generate calls to
the slow memset implementation. Zeromem is a near drop in replacement
for this use case, with a more efficient implementation on both AArch32
and AArch64.
Change-Id: Ia7f3a90e888b96d056881be09f0b4d65b41aa79e
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
To avoid timing side-channel attacks, it is needed to use a constant
time memory comparison function when comparing hashes. The affected
code only cheks for equality so it isn't needed to use any variant of
memcmp(), bcmp() is enough.
Also, timingsafe_bcmp() is as fast as memcmp() when the two compared
regions are equal, so this change incurrs no performance hit in said
case. In case they are unequal, the boot sequence wouldn't continue as
normal, so performance is not an issue.
Change-Id: I1c7c70ddfa4438e6031c8814411fef79fd3bb4df
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
One nasty part of ATF is some of boolean macros are always defined
as 1 or 0, and the rest of them are only defined under certain
conditions.
For the former group, "#if FOO" or "#if !FOO" must be used because
"#ifdef FOO" is always true. (Options passed by $(call add_define,)
are the cases.)
For the latter, "#ifdef FOO" or "#ifndef FOO" should be used because
checking the value of an undefined macro is strange.
Here, IMAGE_BL* is handled by make_helpers/build_macro.mk like
follows:
$(eval IMAGE := IMAGE_BL$(call uppercase,$(3)))
$(OBJ): $(2)
@echo " CC $$<"
$$(Q)$$(CC) $$(TF_CFLAGS) $$(CFLAGS) -D$(IMAGE) -c $$< -o $$@
This means, IMAGE_BL* is defined when building the corresponding
image, but *undefined* for the other images.
So, IMAGE_BL* belongs to the latter group where we should use #ifdef
or #ifndef.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
In mbedtls_x509_parser.c there are some static arrays that are filled
during the integrity check and then read whenever an authentication
parameter is requested. However, they aren't cleared in case of an
integrity check failure, which can be problematic from a security
point of view. This patch clears these arrays in the case of failure.
Change-Id: I9d48f5bc71fa13e5a75d6c45b5e34796ef13aaa2
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
We are duplicating this macro define, and it is useful enough
to be placed in the common place.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
tbz check for RDR status is to check for a bit being zero.
Unfortunately, we are using a mask rather than the bit position.
Further as per http://www.ti.com/lit/ds/symlink/pc16550d.pdf (page 17),
LSR register bit 0 is Data ready status (RDR), not bit position 2.
Update the same to match the specification.
Reported-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
This comment block says the default algorithm is ESDSA, while the
code obviously sets the default to RSA:
ifeq (${MBEDTLS_KEY_ALG},)
MBEDTLS_KEY_ALG := rsa
endif
The git log of commit 7d37aa1711 ("TBB: add mbedTLS authentication
related libraries") states available options are:
* 'rsa' (for RSA-2048) (default option)
* 'ecdsa' (for ECDSA-SECP256R1)
So, my best guess is the comment block is wrong.
The mismatch between the code and the comment is confusing. Fix it.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The previous code required that a certificate be signed with the ROT
key before the platform's NV counter could be updated with the value
in the certificate. This implies that the Non-Trusted NV counter was
not being updated for Non-Trusted content certificates, as they cannot
be signed with the ROT key in the TBBR CoT scheme.
The code is reworked to only allow updating the platform's Trusted NV
counter when a certificate protected by the Trusted NV counter is
signed with the ROT key.
Content certificates protected by the Non-Trusted NV counter are
allowed to update the platform's Non-Trusted NV counter, assuming
that the certificate value is higher than the platform's value.
A new optional platform API has been introduced, named
plat_set_nv_ctr2(). Platforms may choose to implement it and perform
additional checks based on the authentication image descriptor before
modifying the NV counters. A default weak implementation is available
that just calls into plat_set_nv_ctr().
FixesARM-software/tf-issues#426
Change-Id: I4fc978fd28a3007bc0cef972ff1f69ad0413b79c
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
Some GICv3 implementations have provision for power management
operations at Redistributor level. This patch introduces and provides
place-holders for Redistributor power management. The default
implementations are empty stubs, but are weakly bound so as to enable
implementation-specific drivers to override them.
Change-Id: I4fec1358693d3603ca5dce242a2f7f0e730516d8
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>
This patch makes following miscellaneous fixes:
* pl011_console.S: Fixed the bit mask used to check if the
transmit FIFO is full or empty.
* smcc_macros.S: Added `_fsxc` suffix while updating the SPSR.
By default the assembler assumes `_fc` suffix which does not
update all the fields in SPSR. By adding `_fsxc` suffix all
the fields gets updated.
* platform_helpers.S: Removed the weak definition for
`plat_my_core_pos()` as this is a mandatory function which
needs to be defined by all platforms.
Change-Id: I8302292533c943686fff8d7c749a07132c052a3b
Signed-off-by: Yatharth Kochar <yatharth.kochar@arm.com>
Currently the GICv3 driver mandates that platform populate
both G1S and G0 interrupts. However, it is possible that a
given platform is not interested in both the groups and
just needs to specify either one of them.
This patch modifies the `gicv3_rdistif_init()` & `gicv3_distif_init()`
functions to allow either G1S or G0 interrupts to be configured.
FixesARM-software/tf-issues#400
Change-Id: I43572b0e08ae30bed5af9334f25d35bf439b0d2b
Support CMD23. When CMD23 is used, CMD12 could be avoided.
Two scenarios:
1. CMD17 for single block, CMD18 + CMD12 for multiple blocks.
2. CMD23 + CMD18 for both single block and multiple blocks.
The emmc_init() should initialize whether CMD23 is supported
or not.
Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
As per the GICv3 specification, to power down a processor using GICv3
and allow automatic power-on if an interrupt must be sent to a processor,
software must set Enable to zero for all interrupt groups(by writing to
GICC_CTLR or ICC_IGRPEN{0,1}_EL1/3 as appropriate.
Also, NonSecure EL1 software may not be aware of the CPU power state
details and fail to choose right states that require quiescing the CPU
interface. So it's preferred that the PSCI implementation handles it as
it is fully aware of the CPU power states.
This patch adds disabling of Group1 NonSecure interrupts during processor
power down along with Group0 and Group1 Secure interrupts so that all the
interrupt groups are handled at once as per specification.
Change-Id: Ib564d773c9c4c41f2ca9471451c030e3de75e641
This patch modifies GICv3 and TZC drivers to add AArch32 support.
No modifications are required for the GICv2 driver for AArch32 support.
The TZC driver assumes that the secure world is running in Little-Endian
mode to do 64 bit manipulations. Assertions are present to validate the
assumption.
Note: The legacy GICv3 driver is not supported for AArch32.
Change-Id: Id1bc75a9f5dafb9715c9500ca77b4606eb1e2458
This patch moves the various assembly console drivers
into `aarch64` architecture specific folder. Stub files,
which include files from new location, are retained at the
original location for platform compatibility reasons.
Change-Id: I0069b6c1c0489ca47f5204d4e26e3bc3def533a8
If buffer address parameter isn't aligned, it may cause
DMA issue in block device driver, as eMMC. Now check the buffer
address. If it's not aligned, use temporary buffer in io block
driver instead.
Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
This patch fixes the offset of GICD_IROUTER register defined in gicv3.h.
Although the GICv3 documention mentions that the offset for this register
is 0x6100-0x7FD8, the offset calculation for an interrupt id `n` is :
0x6000 + 8n, where n >= 32
This requires the offset for GICD_IROUTER to be defined as 0x6000.
FixesARM-software/tf-issues#410
Change-Id: If9e91e30d946afe7f1f60fea4f065c7567093fa8
This patch reworks type usage in generic code, drivers and ARM platform files
to make it more portable. The major changes done with respect to
type usage are as listed below:
* Use uintptr_t for storing address instead of uint64_t or unsigned long.
* Review usage of unsigned long as it can no longer be assumed to be 64 bit.
* Use u_register_t for register values whose width varies depending on
whether AArch64 or AArch32.
* Use generic C types where-ever possible.
In addition to the above changes, this patch also modifies format specifiers
in print invocations so that they are AArch64/AArch32 agnostic. Only files
related to upcoming feature development have been reworked.
Change-Id: I9f8c78347c5a52ba7027ff389791f1dad63ee5f8
The legacy GIC driver assumes that the SGIs and PPIs are Group0 during
initialization. This is true if the driver is the first one to initialize
the GIC hardware after reset. But in some cases, earlier BL stages could
have already initialized the GIC hardware which means that SGI and PPI
configuration are not the expected reset values causing assertion failure
in `gicd_set_ipriorityr()`. This patch explicitly resets the SGI and PPI
to Group0 prior to their initialization in the driver. The same patch is
not done in the GICv2-only driver because unlike in the legacy driver,
`gicd_set_ipriorityr()` of GICv2 driver doesn't enforce this policy and
the appropriate group is set irrespective of the initial value.
FixesARM-software/tf-issues#396
Change-Id: I521d35caa37470ce542c796c2ba99716e4763105
A production ROM with TBB enabled must have the ability to boot test software
before a real ROTPK is deployed (e.g. manufacturing mode). Previously the
function plat_get_rotpk_info() must return a valid ROTPK for TBB to succeed.
This patch adds an additional bit `ROTPK_NOT_DEPLOYED` in the output `flags`
parameter from plat_get_rotpk_info(). If this bit is set, then the ROTPK
in certificate is used without verifying against the platform value.
FixesARM-software/tf-issues#381
Change-Id: Icbbffab6bff8ed76b72431ee21337f550d8fdbbb
On some platform gpio can set/get pull status when input, add these
function so we can set/get gpio pull status when need it. And they are
optional function.
This patch adds the API `ccn_get_part0_id` to query the PART0 ID from the
PERIPHERAL_ID 0 register in the CCN driver. This ID allows to distinguish
the variant of CCN present on the system and possibly enable dynamic
configuration of the IP based on the variant. Also added an assert in
`ccn_master_to_rn_id_map()` to ensure that the master map bitfield provided
by the platform is within the expected interface id.
Change-Id: I92d2db7bd93a9be8a7fbe72a522cbcba0aba2d0e
Add delay timer implementation based on the system generic counter.
This either uses the platform's implementation of
`plat_get_syscnt_freq()` or explicit clock multiplier/divider values
provided by the platform.
The current implementation of udelay has been modified to avoid
unnecessary calculations while waiting on the loop and to make it
easier to check for overflows.
Change-Id: I9062e1d506dc2f68367fd9289250b93444721732
Allow to handle cases where some images are pre-loaded (by debugger for
instance) without introducing many switches in files calling load_* functions.
Fixes: arm-software/tf-issues#398
Signed-off-by: Gerald Lejeune <gerald.lejeune@st.com>
In a lot of embedded platforms, eMMC device is the only one storage
device. So loading content from eMMC device is required in ATF.
Create the emmc stack that could co-work with IO block driver.
Support to read/write/erase eMMC blocks on both rpmb and normal
user area. Support to change the IO speed and bus width.
Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
FIP is accessed as memory-mapped type. eMMC is block device type.
In order to support FIP based on eMMC, add the new io_block layer.
io_block always access eMMC device as block size. And it'll only
copy the required data into buffer in io_block driver. So preparing
an temporary buffer is required.
When use io_block device, MAX_IO_BLOCK_DEVICES should be declared
in platform_def.h. It's used to support multiple block devices.
Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
Antonio Nino Diaz
Haojian Zhuang
Masahiro Yamada
David Cunado
dp-arm
Varun Wadekar
Soby Mathew
Jeenu Viswambharan
Douglas Raillard
Nishanth Menon
Yatharth Kochar
Sudeep Holla
Caesar Wang
Gerald Lejeune