GitLab

The Cortex-A76 implements SMCCC_ARCH_WORKAROUND_2 as defined in
"Firmware interfaces for mitigating cache speculation vulnerabilities
System Software on Arm Systems"[0].

Dynamic mitigation for CVE-2018-3639 is enabled/disabled by
setting/clearning bit 16 (Disable load pass store) of `CPUACTLR2_EL1`.

NOTE: The generic code that implements dynamic mitigation does not
currently implement the expected semantics when dispatching an SDEI
event to a lower EL.  This will be fixed in a separate patch.

[0] https://developer.arm.com/cache-speculation-vulnerability-firmware-specification



Change-Id: I8fb2862b9ab24d55a0e9693e48e8be4df32afb5a
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

The workaround uses the instruction patching feature of the Ares cpu.

Change-Id: I868fce0dc0e8e41853dcce311f01ee3867aabb59
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

Change-Id: Ia170c12d3929a616ba80eb7645c301066641f5cc
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

Both Cortex-Ares and Cortex-A76 CPUs use the ARM DynamIQ Shared Unit
(DSU).  The power-down and power-up sequences are therefore mostly
managed in hardware, and required software operations are simple.

Change-Id: I3a9447b5bdbdbc5ed845b20f6564d086516fa161
Signed-off-by: Isla Mitchell <isla.mitchell@arm.com>

The patch changes the layout of BL images in memory to enable
more efficient use of available space. Previously BL31 was loaded
with the expectation that BL2 memory would be reclaimed by BL32
loaded in SRAM. But with increasing memory requirements in the
firmware, we can no longer fit BL32 in SRAM anymore which means the
BL2 memory is not reclaimed by any runtime image. Positioning BL2
below BL1-RW and above BL31 means that the BL31 NOBITS can be
overlaid on BL2 and BL1-RW.

This patch also propogates the same memory layout to BL32 for AArch32
mode. The reset addresses for the following configurations are also
changed :
   * When RESET_TO_SP_MIN=1 for BL32 in AArch32 mode
   * When BL2_AT_EL3=1 for BL2

The restriction on BL31 to be only in DRAM when SPM is enabled
is now removed with this change. The update to the firmware design
guide for the BL memory layout is done in the following patch.

Change-Id: Icca438e257abe3e4f5a8215f945b9c3f9fbf29c9
Signed-off-by: Soby Mathew <soby.mathew@arm.com>

BL31 is running out of space, and the use-case of SPM doesn't require it
to be in SRAM. To prevent BL31 from running out of space in the future,
move BL31 to DRAM if SPM is enabled.

Secure Partition Manager design document updated to reflect the changes.

Increased the size of the stack of BL31 for builds with SPM.

The translation tables used by SPM in Arm platforms have been moved back
to the 'xlat_tables' region instead of 'arm_el3_tzc_dram'. Everything is
in DRAM now, so it doesn't make sense to treat them in a different way.

Change-Id: Ia6136c8e108b8da9edd90e9d72763dada5e5e5dc
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Move all information related to a Secure Partition to the struct
secure_partition_context_t.

This requires an in-depth refactor because most of the previous code of
SPM relied on global information.

Change-Id: I0a23e93817dcc191ce1d7506b8bc671d376123c4
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

This function can be currently accessed through the wrappers
cm_init_context_by_index() and cm_init_my_context(). However, they only
work on contexts that are associated to a CPU.

By making this function public, it is possible to set up a context that
isn't associated to any CPU. For consistency, it has been renamed to
cm_setup_context().

Change-Id: Ib2146105abc8137bab08745a8adb30ca2c4cedf4
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Some CPUS may benefit from using a dynamic mitigation approach for
CVE-2018-3639.  A new SMC interface is defined to allow software
executing in lower ELs to enable or disable the mitigation for their
execution context.

It should be noted that regardless of the state of the mitigation for
lower ELs, code executing in EL3 is always mitigated against
CVE-2018-3639.

NOTE: This change is a compatibility break for any platform using
the declare_cpu_ops_workaround_cve_2017_5715 macro.  Migrate to
the declare_cpu_ops_wa macro instead.

Change-Id: I3509a9337ad217bbd96de9f380c4ff8bf7917013
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

Implement static mitigation for CVE-2018-3639 on
Cortex A57 and A72.

Change-Id: I83409a16238729b84142b19e258c23737cc1ddc3
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

For affected CPUs, this approach enables the mitigation during EL3
initialization, following every PE reset. No mechanism is provided to
disable the mitigation at runtime.

This approach permanently mitigates the entire software stack and no
additional mitigation code is required in other software components.

TF-A implements this approach for the following affected CPUs:

*   Cortex-A57 and Cortex-A72, by setting bit 55 (Disable load pass store) of
    `CPUACTLR_EL1` (`S3_1_C15_C2_0`).

*   Cortex-A73, by setting bit 3 of `S3_0_C15_C0_0` (not documented in the
    Technical Reference Manual (TRM)).

*   Cortex-A75, by setting bit 35 (reserved in TRM) of `CPUACTLR_EL1`
    (`S3_0_C15_C1_0`).

Additionally, a new SMC interface is implemented to allow software
executing in lower ELs to discover whether the system is mitigated
against CVE-2018-3639.

Refer to "Firmware interfaces for mitigating cache speculation
vulnerabilities System Software on Arm Systems"[0] for more
information.

[0] https://developer.arm.com/cache-speculation-vulnerability-firmware-specification



Change-Id: I084aa7c3bc7c26bf2df2248301270f77bed22ceb
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

This patch renames symbols and files relating to CVE-2017-5715 to make
it easier to introduce new symbols and files for new CVE mitigations.

Change-Id: I24c23822862ca73648c772885f1690bed043dbc7
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>

File: include/common/aarch64/el3_common_macros.S

Change-Id: I619401e961a3f627ad8864781b5f90bc747c3ddb
Signed-off-by: Qixiang Xu <qixiang.xu@arm.com>

This patch adds soc_fw_config, tos_fw_config and nt_fw_config to the FVP.
The config files are placeholders and do not have any useful bindings
defined. The tos_fw_config is packaged in FIP and loaded by BL2 only
if SPD=tspd. The load address of these configs are specified in tb_fw_config
via new bindings defined for these configs. Currently, in FVP, the
soc_fw_config and tos_fw_config is loaded in the page between BL2_BASE
and ARM_SHARED_RAM. This memory was typically used for BL32 when
ARM_TSP_RAM_LOCATION=tsram but since we cannot fit BL32 in that
space anymore, it should be safe to use this memory for these configs.
There is also a runtime check in arm_bl2_dyn_cfg_init() which ensures
that this overlap doesn't happen.

The previous arm_dyn_get_hwconfig_info() is modified to accept configs
other than hw_config and hence renamed to arm_dyn_get_config_load_info().
The patch also corrects the definition of ARM_TB_FW_CONFIG_LIMIT to be
BL2_BASE.

Change-Id: I03a137d9fa1f92c862c254be808b8330cfd17a5a
Signed-off-by: Soby Mathew <soby.mathew@arm.com>

This patch implements support for adding dynamic configurations for
BL31 (soc_fw_config), BL32 (tos_fw_config) and BL33 (nt_fw_config). The
necessary cert tool support and changes to default chain of trust are made
for these configs.

Change-Id: I25f266277b5b5501a196d2f2f79639d838794518
Signed-off-by: Soby Mathew <soby.mathew@arm.com>

This patch adds capability to FVP to disable authentication dynamically
via the `disable_auth` property in TB_FW_CONFIG. Both BL1 and BL2 parses
the TB_FW_CONFIG for the `disable_auth` property and invokes the
`load_dyn_disable_auth()` API to disable authentication if the
property is set to 1. The DYN_DISABLE_AUTH is enabled by default for
FVP as it is a development platform. Note that the TB_FW_CONFIG has to
be authenticated by BL1 irrespective of these settings.

The arm_bl2_dyn_cfg_init() is now earlier in bl2_plat_preload_setup()
rather than in bl2_platform_setup() as we need to get the value of
`disable_auth` property prior to authentication of any image by BL2.

Change-Id: I734acd59572849793e5020ec44c6ac51f654a4d1
Signed-off-by: Soby Mathew <soby.mathew@arm.com>

This patch allows platforms to dynamically disable authentication of
images during cold boot. This capability is controlled via the
DYN_DISABLE_AUTH build flag and is only meant for development
purposes.

Change-Id: Ia3df8f898824319bb76d5cc855b5ad6c3d227260
Signed-off-by: Soby Mathew <soby.mathew@arm.com>

In 'console_set_scope' and when registering a console, field 'flags' of
'console_t' is assigned a 32-bit value. However, when it is actually
used, the functions perform 64-bit reads to access its value. This patch
changes all 64-bit reads to 32-bit reads.

Change-Id: I181349371409e60065335f078857946fa3c32dc1
Signed-off-by: Daniel Boulby <daniel.boulby@arm.com>

This reverts commit 2f18aa1f

.

It is causing some tests to fail. Until the cause is found and fixed, it
is needed to remove this commit from master.

Change-Id: Ic5ff7a841903a15613e00379e87cbbd8a0e85152
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Change-Id: I7e73c0ab134da11c49f990b739245110c59eac2b
Signed-off-by: Chris Kay <chris.kay@arm.com>

SGI-575's NSRAM is neither in the same place nor the same size as Juno's.

Change-Id: Id6d692e9c7e9c1360014bb525eda966ebe29c823
Signed-off-by: Chris Kay <chris.kay@arm.com>

The bounds check in ARM_CASSERT_MMAP does not take into account the
array sentinel in plat_arm_mmap. This commit fixes this, and adds an
additional check to ensure the number of entries in the array is
within the bounds of PLAT_ARM_MMAP_ENTRIES.

Change-Id: Ie6df10c0aa0890d62826bc3224ad7b3e36fd53e2
Signed-off-by: Chris Kay <chris.kay@arm.com>

There are three calls to mmap_add_region() that always occur in
arm_setup_page_tables(), and two further calls based on whether coherent
memory is enabled, and whether SPM is enabled in BL31.

This commit adapts the ARM_BL_REGIONS definition to match the number of
calls made inside arm_setup_page_tables() so that the MAX_MMAP_REGIONS
is realigned with what is actually occurring.

Change-Id: I7adc05951abccf2cbd5c86280eb874911e6a1566
Signed-off-by: Chris Kay <chris.kay@arm.com>

The old API is deprecated and will eventually be removed.

Arm platforms now use the multi console driver for boot and runtime
consoles. However, the crash console uses the direct console API because
it doesn't need any memory access to work. This makes it more robust
during crashes.

The AArch32 port of the Trusted Firmware doesn't support this new API
yet, so it is only enabled in AArch64 builds. Because of this, the
common code must maintain compatibility with both systems. SP_MIN
doesn't have to be updated because it's only used in AArch32 builds.
The TSP is only used in AArch64, so it only needs to support the new
API without keeping support for the old one.

Special care must be taken because of PSCI_SYSTEM_SUSPEND. In Juno, this
causes the UARTs to reset (except for the one used by the TSP). This
means that they must be unregistered when suspending and re-registered
when resuming. This wasn't a problem with the old driver because it just
restarted the UART, and there were no problems associated with
registering and unregistering consoles.

The size of BL31 has been increased in builds with SPM.

Change-Id: Icefd117dd1eb9c498921181a21318c2d2435c441
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

In the multi console driver, allowing to register the same console more
than once may result in an infinte loop when putc is called.

If, for example, a boot message is trying to be printed, but the
consoles in the loop in the linked list are runtime consoles, putc will
iterate forever looking for a console that can print boot messages (or
a NULL pointer that will never come).

This loop in the linked list can occur after restoring the system from a
system suspend. The boot console is registered during the cold boot in
BL31, but the runtime console is registered even in the warm boot path.
Consoles are always added to the start of the linked list when they are
registered, so this it what should happen if they were actually
different structures:

   console_list -> NULL
   console_list -> BOOT -> NULL
   console_list -> RUNTIME -> BOOT -> NULL
   console_list -> RUNTIME -> RUNTIME -> BOOT -> NULL

In practice, the two runtime consoles are the same one, so they create
this loop:

   console_list -> RUNTIME -.    X -> BOOT -> NULL
                       ^    |
                       `----'

This patch adds an assertion to detect this problem. The assertion will
fail whenever the same structure tries to be registered while being on
the list.

In order to assert this, console_is_registered() has been implemented.
It returns 1 if the specified console is registered, 0 if not.

Change-Id: I922485e743775ca9bd1af9cbd491ddd360526a6d
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

When TF is compiled for aarch32 MAX_VIRT_ADDR_SPACE_SIZE is 2^32 in some cases,
which makes the test (size) <= MAX_VIRT_ADDR_SPACE_SIZE a tautology because
uintptr_t is a 32 bit value. The cast remove the warning for clang.

Change-Id: I1345f3400f8fbbe4ffd3caa990a90e7ba593dba5
Signed-off-by: Roberto Vargas <roberto.vargas@arm.com>

  - Assign 0x10 for RAS exceptions on ARM platforms, and install
    EHF priority descriptor.

  - Call the common RAS initialisation from ARM BL31 setup.

  - Add empty definitions for platform error records and RAS interrupts.

Change-Id: I0675f299b7840be4c83a9c7a81073a95c605dc90
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

The ARMv8.4 RAS extensions introduce architectural support for software
to inject faults into the system in order to test fault-handling
software. This patch introduces the build option FAULT_HANDLING_SUPPORT
to allow for lower ELs to use registers in the Standard Error Record to
inject fault. The build option RAS_EXTENSIONS must also be enabled along
with fault injection.

This feature is intended for testing purposes only, and is advisable to
keep disabled for production images.

Change-Id: I6f7a4454b15aec098f9505a10eb188c2f928f7ea
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

EHF currently allows for registering interrupt handlers for a defined
priority ranges. This is primarily targeted at various EL3 dispatchers
to own ranges of secure interrupt priorities in order to delegate
execution to lower ELs.

The RAS support added by earlier patches necessitates registering
handlers based on interrupt number so that error handling agents shall
receive and handle specific Error Recovery or Fault Handling interrupts
at EL3.

This patch introduces a macro, RAS_INTERRUPTS() to declare an array of
interrupt numbers and handlers. Error handling agents can use this macro
to register handlers for individual RAS interrupts. The array is
expected to be sorted in the increasing order of interrupt numbers.

As part of RAS initialisation, the list of all RAS interrupts are sorted
based on their ID so that, given an interrupt, its handler can be looked
up with a simple binary search.

For an error handling agent that wants to handle a RAS interrupt,
platform must:

  - Define PLAT_RAS_PRI to be the priority of all RAS exceptions.

  - Enumerate interrupts to have the GIC driver program individual EL3
    interrupts to the required priority range. This is required by EHF
    even before this patch.

Documentation to follow.

Change-Id: I9471e4887ff541f8a7a63309e9cd8f771f76aeda
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

Previous patches added frameworks for handling RAS errors. This patch
introduces features that the platform can use to enumerate and iterate
RAS nodes:

  - The REGISTER_RAS_NODES() can be used to expose an array of
    ras_node_info_t structures. Each ras_node_info_t describes a RAS
    node, along with handlers for probing the node for error, and if
    did record an error, another handler to handle it.

  - The macro for_each_ras_node() can be used to iterate over the
    registered RAS nodes, probe for, and handle any errors.

The common platform EA handler has been amended using error handling
primitives introduced by both this and previous patches.

Change-Id: I2e13f65a88357bc48cd97d608db6c541fad73853
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

The ARMv8 RAS Extensions introduced Standard Error Records which are a
set of standard registers through which:

  - Platform can configure RAS node policy; e.g., notification
    mechanism;

  - RAS nodes can record and expose error information for error handling
    agents.

Standard Error Records can either be accessed via. memory-mapped
or System registers. This patch adds helper functions to access
registers and fields within an error record.

Change-Id: I6594ba799f4a1789d7b1e45b3e17fd40e7e0ba5c
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

RAS extensions are mandatory for ARMv8.2 CPUs, but are also optional
extensions to base ARMv8.0 architecture.

This patch adds build system support to enable RAS features in ARM
Trusted Firmware. A boolean build option RAS_EXTENSION is introduced for
this.

With RAS_EXTENSION, an Exception Synchronization Barrier (ESB) is
inserted at all EL3 vector entry and exit. ESBs will synchronize pending
external aborts before entering EL3, and therefore will contain and
attribute errors to lower EL execution. Any errors thus synchronized are
detected via. DISR_EL1 register.

When RAS_EXTENSION is set to 1, HANDLE_EL3_EA_FIRST must also be set to 1.

Change-Id: I38a19d84014d4d8af688bd81d61ba582c039383a
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

At present, any External Abort routed to EL3 is reported as an unhandled
exception and cause a panic. This patch enables ARM Trusted Firmware to
handle External Aborts routed to EL3.

With this patch, when an External Abort is received at EL3, its handling
is delegated to plat_ea_handler() function. Platforms can provide their
own implementation of this function. This patch adds a weak definition
of the said function that prints out a message and just panics.

In order to support handling External Aborts at EL3, the build option
HANDLE_EA_EL3_FIRST must be set to 1.

Before this patch, HANDLE_EA_EL3_FIRST wasn't passed down to
compilation; this patch fixes that too.

Change-Id: I4d07b7e65eb191ff72d63b909ae9512478cd01a1
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>

Change-Id: Ice141dcc17f504025f922acace94d98f84acba9e
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Previously mem_protect used to be only supported from BL2. This is not
helpful in the case when ARM TF-A BL2 is not used. This patch demonstrates
mem_protect from el3_runtime firmware on ARM Platforms specifically
when RESET_TO_BL31 or RESET_TO_SP_MIN flag is set as BL2 may be absent
in these cases. The Non secure DRAM is dynamically mapped into EL3 mmap
tables temporarily and then the protected regions are then cleared. This
avoids the need to map the non secure DRAM permanently to BL31/sp_min.

The stack size is also increased, because DYNAMIC_XLAT_TABLES require
a bigger stack.

Change-Id: Ia44c594192ed5c5adc596c0cff2c7cc18c001fde
Signed-off-by: Roberto Vargas <roberto.vargas@arm.com>

Since commit 031dbb12

 ("AArch32: Add essential Arch helpers"),
it is difficult to use consistent format strings for printf() family
between aarch32 and aarch64.

For example, uint64_t is defined as 'unsigned long long' for aarch32
and as 'unsigned long' for aarch64.  Likewise, uintptr_t is defined
as 'unsigned int' for aarch32, and as 'unsigned long' for aarch64.

A problem typically arises when you use printf() in common code.

One solution could be, to cast the arguments to a type long enough
for both architectures.  For example, if 'val' is uint64_t type,
like this:

  printf("val = %llx\n", (unsigned long long)val);

Or, somebody may suggest to use a macro provided by <inttypes.h>,
like this:

  printf("val = %" PRIx64 "\n", val);

But, both would make the code ugly.

The solution adopted in Linux kernel is to use the same typedefs for
all architectures.  The fixed integer types in the kernel-space have
been unified into int-ll64, like follows:

    typedef signed char           int8_t;
    typedef unsigned char         uint8_t;

    typedef signed short          int16_t;
    typedef unsigned short        uint16_t;

    typedef signed int            int32_t;
    typedef unsigned int          uint32_t;

    typedef signed long long      int64_t;
    typedef unsigned long long    uint64_t;

[ Linux commit: 0c79a8e29b5fcbcbfd611daf9d500cfad8370fcf ]

This gets along with the codebase shared between 32 bit and 64 bit,
with the data model called ILP32, LP64, respectively.

The width for primitive types is defined as follows:

                   ILP32           LP64
    int            32              32
    long           32              64
    long long      64              64
    pointer        32              64

'long long' is 64 bit for both, so it is used for defining uint64_t.
'long' has the same width as pointer, so for uintptr_t.

We still need an ifdef conditional for (s)size_t.

All 64 bit architectures use "unsigned long" size_t, and most 32 bit
architectures use "unsigned int" size_t.  H8/300, S/390 are known as
exceptions; they use "unsigned long" size_t despite their architecture
is 32 bit.

One idea for simplification might be to define size_t as 'unsigned long'
across architectures, then forbid the use of "%z" string format.
However, this would cause a distortion between size_t and sizeof()
operator.  We have unknowledge about the native type of sizeof(), so
we need a guess of it anyway.  I want the following formula to always
return 1:

  __builtin_types_compatible_p(size_t, typeof(sizeof(int)))

Fortunately, ARM is probably a majority case.  As far as I know, all
32 bit ARM compilers use "unsigned int" size_t.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

u_register_t is preferred rather than uint64_t.  This is more
consistent with the aarch32 implementation.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

According to the ARMv8 ARM issue C.a:

    AP[1] is valid only for stage 1 of a translation regime that can
    support two VA ranges. It is RES 1 when stage 1 translations can
    support only one VA range.

This means that, even though this bit is ignored, it should be set to 1
in the EL3 and EL2 translation regimes.

For translation regimes consisting on EL0 and a higher regime this bit
selects between control at EL0 or at the higher Exception level. The
regimes that support two VA ranges are EL1&0 and EL2&0 (the later one
is only available since ARMv8.1).

This fix has to be applied to both versions of the translation tables
library.

Change-Id: If19aaf588551bac7aeb6e9a686cf0c2068e7c181
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Due to differences in the bitfields of the SMC IDs, it is not possible
to support SMCCC 1.X and 2.0 at the same time.

The behaviour of `SMCCC_MAJOR_VERSION` has changed. Now, it is a build
option that specifies the major version of the SMCCC that the Trusted
Firmware supports. The only two allowed values are 1 and 2, and it
defaults to 1. The value of `SMCCC_MINOR_VERSION` is derived from it.

Note: Support for SMCCC v2.0 is an experimental feature to enable
prototyping of secure partition specifications. Support for this
convention is disabled by default and could be removed without notice.

Change-Id: I88abf9ccf08e9c66a13ce55c890edea54d9f16a7
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>

Change-Id: I989c1f4aef8e3cb20d5d19e6347575e6449bb60b
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>