GitLab

ROMLIB extract functions code from BL images to put them inside ROM.
This has for effect to reduce the size of the BL images.

This patch take this size reduction into consideration to optimize the
memory layout of BL2.
A new "PLAT_ARM_BL2_ROMLIB_OPTIMIZATION" macro is defined and used to
reduce "PLAT_ARM_MAX_BL2_SIZE". This allows to remove the gap between
BL1 and BL2 when ROMLIB is used and provides more room for BL31.

The current memory gain is 0x6000 for fvp and 0x8000 for juno.

Change-Id: I71c2c2c63b57bce5b22a125efaefc486ff3e87be
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

N1SDP supports multichip configuration wherein n1sdp boards are
connected over high speed coherent CCIX link, for now only dual-chip
is supported.

Whether or not multiple chips are present is dynamically probed by
SCP firmware and passed on to TF-A, routing table will be set up
only if multiple chips are present.

Initialize GIC-600 multichip operation by overriding the default GICR
frames with array of GICR frames and setting the chip 0 as routing
table owner.

Change-Id: Ida35672be4bbf4c517469a5b330548d75e593ff2
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

Earlier PIE support was enabled for all arm platforms when
RESET_TO_BL31=1, but later on it was restricted only to FVP with patch
SHA d4580d17

 because of n1sdp platform.

Now it has been verified that PIE does work for n1sdp platform also, so
enabling it again for all arm platforms.

Change-Id: I05ad4f1775ef72e7cb578ec9245cde3fbce971a5
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

ARM platform can have a non-contiguous GICR frames. For instance, a
multi socket platform can have two or more GIC Redistributor frames
which are 4TB apart. Hence it is necessary for the `gicv3_rdistif_probe`
function to probe all the GICR frames available in the platform.

Introduce `plat_arm_override_gicr_frames` function which platforms can
use to override the default gicr_frames which holds the GICR base
address of the primary cpu.

Change-Id: I1f537b0d871a679cb256092944737f2e55ab866e
Signed-off-by: Vijayenthiran Subramaniam <vijayenthiran.subramaniam@arm.com>

N1SDP supports multichip configuration wherein n1sdp boards are
connected over high speed coherent CCIX link  for now only dual-chip is
supported.

A single instance of TF-A runs on master chip which should be aware of
slave chip's CPU and memory topology.

This patch updates platform macros to include remote chip's information
and also ensures that a single version of firmware works for both single
and dual-chip setup.

Change-Id: I75799fd46dc10527aa99585226099d836c21da70
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

Platform information structure holds information about platform's DDR
size(local/remote) which will be used to zero out the memory before
enabling the ECC capability as well as information about multichip
setup. Multichip and remote DDR information can only be probed in SCP,
SDS region will be used by TF-A to get this information at boot up.

This patch introduces a new SDS to store platform information, which is
populated dynamically by SCP Firmware.previously used mem_info SDS is
also made part of this structure itself.

The platform information is also passed to BL33 by copying it to Non-
Secure SRAM.

Change-Id: I4781dc6a7232c3c0a3219b164d943ce9e3e469ee
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

There are some platforms which uses MPIDR Affinity level 3 for storing
extra affinity information e.g. N1SDP uses it for keeping chip id in a
multichip setup, for such platforms MPIDR validation should not fail.

This patch adds Aff3 bits also as part of mpidr validation mask, for
platforms which does not uses Aff3 will not have any impact as these
bits will be all zeros.

Change-Id: Ia8273972fa7948fdb11708308d0239d2dc4dfa85
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

Change-Id: I1ea2bf088f1e001cdbd377cbfb7c6a2866af0422
Signed-off-by: Imre Kis <imre.kis@arm.com>

a5ds only has always-on power domain and there is no power control
present. However, without the pwr_domain_off handler, the kernel
panics when the user will try to switch off secondary cores. The
a5ds_pwr_domain_off handler will prevent kernel from crashing,
i.e. the kernel will attempt but fail to shut down the secondary CPUs
if the user tries to switch them offline.

Change-Id: I3c2239a1b6f035113ddbdda063c8495000cbe30c
Signed-off-by: Usama Arif <usama.arif@arm.com>

Change-Id: I645442d52a295706948e2cac88c36c1a3cb0bc47
Signed-off-by: Imre Kis <imre.kis@arm.com>

Not tested on FVP Model.

Change-Id: Iedebc5c1fbc7ea577e94142b7feafa5546f1f4f9
Signed-off-by: Artsem Artsemenka <artsem.artsemenka@arm.com>

This patch invokes the new function gicv3_rdistif_probe() in the
ARM platform specific gicv3 driver. Since this API modifies the
shared GIC related data structure, it must be invoked coherently
by using the platform specific pwr_domain_on_finish_late hook.

Change-Id: I6efb17d5da61545a1c5a6641b8f58472b31e62a8
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>

The Fast Models provide a non-volatile counter component, which is used
in the Trusted Board Boot implementation to protect against rollback
attacks.

This component comes in 2 versions (see [1]).

- Version 0 is the default and models a locked non-volatile counter,
  whose value is fixed.

- Version 1 of the counter may be incremented in a monotonic fashion.

plat_set_nv_ctr() must cope with both versions. This is achieved by:
1) Attempting to write the new value in the counter.
2) Reading the value back.
3) If there is a mismatch, we know the counter upgrade failed.

When using version 0 of the counter, no upgrade is possible so the
function is expected to fail all the time. However, the code is
missing a compiler barrier between the write operation and the next
read. Thus, the compiler may optimize and remove the read operation on
the basis that the counter value has not changed. With the default
optimization level used in TF-A (-Os), this is what's happening.

The fix introduced in this patch marks the write and subsequent read
accesses to the counter as volatile, such that the compiler makes no
assumption about the value of the counter.

Note that the comment above plat_set_nv_ctr() was clearly stating
that when using the read-only version of the non-volatile counter,
"we expect the values in the certificates to always match the RO
values so that this function is never called". However, the fact that
the counter value was read back seems to contradict this comment, as
it is implementing a counter-measure against misuse of the
function. The comment has been reworded to avoid any confusion.

Without this patch, this bug may be demonstrated on the Base AEM FVP:
- Using version 0 of the non-volatile counter (default version).
- With certificates embedding a revision number value of 32
  (compiling TF-A with TFW_NVCTR_VAL=32).

In this configuration, the non-volatile counter is tied to value 31 by
default. When BL1 loads the Trusted Boot Firmware certificate, it
notices that the two values do not match and tries to upgrade the
non-volatile counter. This write operation is expected to fail
(because the counter is locked) and the function is expected to return
an error but it succeeds instead.

As a result, the trusted boot does not abort as soon as it should and
incorrectly boots BL2. The boot is finally aborted when BL2 verifies
the BL31 image and figures out that the version of the SoC Firmware
Key Certificate does not match. On Arm platforms, only certificates
signed with the Root-of-Trust Key may trigger an upgrade of the
non-volatile Trusted counter.

[1] https://developer.arm.com/docs/100964/1160/fast-models-components/peripheral-components/nonvolatilecounter



Change-Id: I9979f29c23b47b338b9b484013d1fb86c59db92f
Signed-off-by: Sandrine Bailleux <sandrine.bailleux@arm.com>

Enable cores 1-3 using psci. On receiving the smc call from kernel,
core 0 will bring the secondary cores out pen and signal an event for
the cores. Currently on switching the cores is enabled i.e. it is not
possible to suspend, switch cores off, etc.

Change-Id: I6087e1d2ec650e1d587fd543efc1b08cbb50ae5f
Signed-off-by: Usama Arif <usama.arif@arm.com>

For the secondary CPUs, hold the cpu in wfe rather then panic.
This will be needed when multicore support is added to a5ds as
the smc call will write to the hold base and signal an event to
power on the secondary CPUs.

Change-Id: I0ffc2059e9ef894c21375ca5c94def859bfa6599
Signed-off-by: Usama Arif <usama.arif@arm.com>

This patch provides the following features and makes modifications
listed below:
- Individual APIAKey key generation for each CPU.
- New key generation on every BL31 warm boot and TSP CPU On event.
- Per-CPU storage of APIAKey added in percpu_data[]
  of cpu_data structure.
- `plat_init_apiakey()` function replaced with `plat_init_apkey()`
  which returns 128-bit value and uses Generic timer physical counter
  value to increase the randomness of the generated key.
  The new function can be used for generation of all ARMv8.3-PAuth keys
- ARMv8.3-PAuth specific code placed in `lib\extensions\pauth`.
- New `pauth_init_enable_el1()` and `pauth_init_enable_el3()` functions
  generate, program and enable APIAKey_EL1 for EL1 and EL3 respectively;
  pauth_disable_el1()` and `pauth_disable_el3()` functions disable
  PAuth for EL1 and EL3 respectively;
  `pauth_load_bl31_apiakey()` loads saved per-CPU APIAKey_EL1 from
  cpu-data structure.
- Combined `save_gp_pauth_registers()` function replaces calls to
  `save_gp_registers()` and `pauth_context_save()`;
  `restore_gp_pauth_registers()` replaces `pauth_context_restore()`
  and `restore_gp_registers()` calls.
- `restore_gp_registers_eret()` function removed with corresponding
  code placed in `el3_exit()`.
- Fixed the issue when `pauth_t pauth_ctx` structure allocated space
  for 12 uint64_t PAuth registers instead of 10 by removal of macro
  CTX_PACGAKEY_END from `include/lib/el3_runtime/aarch64/context.h`
  and assigning its value to CTX_PAUTH_REGS_END.
- Use of MODE_SP_ELX and MODE_SP_EL0 macro definitions
  in `msr	spsel`  instruction instead of hard-coded values.
- Changes in documentation related to ARMv8.3-PAuth and ARMv8.5-BTI.

Change-Id: Id18b81cc46f52a783a7e6a09b9f149b6ce803211
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>

Some cores support only AArch64 from EL1 and above, e.g. A76, N1 etc. If
TF-A is compiled with CTX_INCLUDE_AARCH32_REGS=0 so as to properly
handle those cores, only the AArch64 cores' assembly is included in the
TF-A binary. In other words, for FVP, TF-A assumes that AArch64 only
cores will never exist in the same cluster with cores that also support
AArch32.

However, A55 and A75 can be used as AArch64 only cores, despite
supporting AArch32, too. This patch enables A55 and A75 to exist in
clusters together with AArch64 cores.

Change-Id: I58750ad6c3d76ce77eb354784c2a42f2c179031d
Signed-off-by: John Tsichritzis <john.tsichritzis@arm.com>

This patch adds support for Corstone-700 foundation IP, which integrates
both Cortex-M0+ and Cortex-A(Host) processors in one handy, flexible
subsystem.
This is an example implementation of Corstone-700 IP host firmware.

Cortex-M0+ will take care of boot stages 1 and 2(BL1/BL2) as well as
bringing Host out RESET. Host will start execution directly from BL32 and
then will jump to Linux.

It is an initial port and additional features are expected to be added
later.

Change-Id: I7b5c0278243d574284b777b2408375d007a7736e
Signed-off-by: Manish Pandey <manish.pandey2@arm.com>

SMMUv3 driver functions which are called from BL1 and BL31
currently use counter-based poll method for testing status
bits. Adding Delay Timer driver to BL1 and BL31 is required
for timeout-based implementation using timer delay functions
for SMMU and other drivers.
This patch adds new function `fvp_timer_init()` which
initialises either System level generic or SP804 timer based on
FVP_USE_SP804_TIMER build flag.
In BL2U `bl2u_early_platform_setup()` function the call to
`arm_bl2u_early_platform_setup()` (which calls
`generic_delay_timer_init()` ignoring FVP_USE_SP804_TIMER flag),
is replaced with `arm_console_boot_init()` and `fvp_timer_init()`.

Change-Id: Ifd8dcebf4019e877b9bc5641551deef77a44c0d1
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>

Change-Id: Ia1ecad58ebf9de3f3a44b17ad1de57424b431125
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

Add the disable_auth dynamic parameter, that allows to disable the
authentication when TBBR is enabled. This parameter is for development
only.

Change-Id: Ic24ad16738517f7e07c4f506dcf69a1ae8df7d2d
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

NOTE: AARCH32/AARCH64 macros are now deprecated in favor of __aarch64__.

All common C compilers pre-define the same macros to signal which
architecture the code is being compiled for: __arm__ for AArch32 (or
earlier versions) and __aarch64__ for AArch64. There's no need for TF-A
to define its own custom macros for this. In order to unify code with
the export headers (which use __aarch64__ to avoid another dependency),
let's deprecate the AARCH32 and AARCH64 macros and switch the code base
over to the pre-defined standard macro. (Since it is somewhat
unintuitive that __arm__ only means AArch32, let's standardize on only
using __aarch64__.)

Change-Id: Ic77de4b052297d77f38fc95f95f65a8ee70cf200
Signed-off-by: Julius Werner <jwerner@chromium.org>

NOTE: __ASSEMBLY__ macro is now deprecated in favor of __ASSEMBLER__.

All common C compilers predefine a macro called __ASSEMBLER__ when
preprocessing a .S file. There is no reason for TF-A to define it's own
__ASSEMBLY__ macro for this purpose instead. To unify code with the
export headers (which use __ASSEMBLER__ to avoid one extra dependency),
let's deprecate __ASSEMBLY__ and switch the code base over to the
predefined standard.

Change-Id: Id7d0ec8cf330195da80499c68562b65cb5ab7417
Signed-off-by: Julius Werner <jwerner@chromium.org>

Fixed the path to a source file specified in tsp makefile
Created a platform specific tsp makefile

Change-Id: I89565127c67eff510e48e21fd450af4c3088c2d4
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>

Code using Cryptocell specific APIs was used as part of the
arm common board ROT support, instead of being abstracted
in Cryptocell specific driver code, creating two problems:
- Any none arm board that uses Cryptocell wuld need to
  copy and paste the same code.
- Inability to cleanly support multiple versions of Cryptocell
  API and products.

Move over Cryptocell specific API calls into the Cryptocell
driver, creating abstraction API where needed.
Signed-off-by: Gilad Ben-Yossef <gilad.benyossef@arm.com>
Change-Id: I9e03ddce90fcc47cfdc747098bece86dbd11c58e

The DMC-620 memory controllers in N1SDP platform has to be put
into CONFIG state before writing to ERR0CTLR0 register to enable
ECC.

This patch fixes the sequence so that DMCs are set to CONFIG
state before writing to ERR0CTLR0 register and moved back to
READY state after writing.

Change-Id: I1252f3ae0991603bb29234029cddb5fbf869c1b2
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

The watchdog is configured with a default value of 256 seconds in order
to implement the Trusted Board Boot Requirements.

For the FVP and Juno platforms, the FWU process relies on a watchdog
reset. In order to automate the test of FWU, the length of this process
needs to be as short as possible. Instead of waiting for those 4 minutes
to have a reset by the watchdog, tell it to reset immediately.

There are no side effects as the value of the watchdog's load register
resets to 0xFFFFFFFF.

Tested on Juno.

Change-Id: Ib1aea80ceddc18ff1e0813a5b98dd141ba8a3ff2
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

Change-Id: Iab767e9937f5c6c8150953fcdc3b37e8ee83fa63
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

This patch adds support for Cortex-A5 FVP for the
DesignStart program. DesignStart aims at providing
low cost and fast access to Arm IP.

Currently with this patch only the primary CPU is booted
and the rest of them wait for an interrupt.
Signed-off-by: Usama Arif <usama.arif@arm.com>
Change-Id: I3a2281ce6de2402dda4610a89939ed53aa045fab

Change-Id: I755e4c42242d9a052570fd1132ca3d937acadb13
Signed-off-by: Balint Dobszay <balint.dobszay@arm.com>

The new API becomes the default one.

Change-Id: Ic1d602da3dff4f4ebbcc158b885295c902a24fec
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>

N1SDP platform supports RDIMMs with ECC capability. To use the ECC
capability, the entire DDR memory space has to be zeroed out before
enabling the ECC bits in DMC620. Zeroing out several gigabytes of
memory from SCP is quite time consuming so functions are added that
zeros out the DDR memory from application processor which is
much faster compared to SCP. BL33 binary cannot be copied to DDR memory
before enabling ECC so this is also done by TF-A from IOFPGA-DDR3
memory to main DDR4 memory after ECC is enabled.

Original PLAT_PHY_ADDR_SPACE_SIZE was limited to 36-bits with which
the entire DDR space cannot be accessed as DRAM2 starts in base
0x8080000000. So these macros are redefined for all ARM platforms.

Change-Id: If09524fb65b421b7a368b1b9fc52c49f2ddb7846
Signed-off-by: Manoj Kumar <manoj.kumar3@arm.com>

BL2U should not build when RESET_TO_SP_MIN flag is set, like BL1 and
BL2.

Change-Id: Iac516121f98611ca1f58d2b5efdec6525b06ce4e
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

During the secondary cores' cold boot path, the cores initialise the GIC
CPU interface. However this is a redundant action since 1) the cores are
powered down immediately after that, 2) the GIC CPU interface is
initialised from scratch when the secondary cores are powered up again
later.

Moreover, this part of code was introducing a bug. In a GICv3 system,
the GIC's CPU interface system registers must not be written without the
core being marked as "awake" in the redistributor. However, this
sequence was performing such accesses and this would cause those cores
to hang. The hang was caused by the DSB instruction that would never
complete because of the GIC not recognising those writes.

For the two aforementioned reasons, the entire part of the GIC CPU
interface initialisation is removed.

Change-Id: I6c33a1edda69dd5b6add16a27390a70731b5532a
Signed-off-by: John Tsichritzis <john.tsichritzis@arm.com>

Some cores support only AArch64 mode. In those cores, only a limited
subset of the AArch32 system registers are implemented. Hence, if TF-A
is supposed to run on AArch64-only cores, it must be compiled with
CTX_INCLUDE_AARCH32_REGS=0.

Currently, the default settings for compiling TF-A are with the AArch32
system registers included. So, if we compile TF-A the default way and
attempt to run it on an AArch64-only core, we only get a runtime panic.

Now a compile-time check has been added to ensure that this flag has the
appropriate value when AArch64-only cores are included in the build.

Change-Id: I298ec550037fafc9347baafb056926d149197d4c
Signed-off-by: John Tsichritzis <john.tsichritzis@arm.com>

N1SDP exhibits the behavior similar to Juno wherein CNTBaseN.CNTFRQ
can be written but does not reflect the value of the CNTFRQ register
in CNTCTLBase frame. This doesn't follow ARM ARM in that the value
updated in CNTCTLBase.CNTFRQ is not reflected in CNTBaseN.CNTFRQ.

Hence enable the workaround (applied to Juno) for N1SDP that updates
the CNTFRQ register in the Non Secure CNTBaseN frame.

Change-Id: Id89ee1bca0f25c9d62f8f794f2c4f4e618cdf092
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>

The default DRAM2 start address for Arm platforms
is 0x880000000. However, for N1SDP platform this is
0x8080000000.

Fix the DRAM2 start address by initialising
PLAT_ARM_DRAM2_BASE.

Without this fix there is a mismatch of the System
memory region view as seen by the BL31 runtime
firmware (PSCI) versus the view of the OS (which
is based on the description provided by UEFI. In
this case UEFI is correctly describing the DRAM2
start address).

This implicates in secondary cores failing to start
on some Operating Systems if the OS decides to place
the secondary start address in the mismatched region.

Change-Id: I57220e753219353dda429868b4c5e1a69944cc64
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>

The default DRAM2 base address for Arm platforms
is 0x880000000. However, on some platforms the
firmware may want to move the start address to
a different value.

To support this introduce PLAT_ARM_DRAM2_BASE that
defaults to 0x880000000; but can be overridden by
a platform (e.g. in platform_def.h).

Change-Id: I0d81195e06070bc98f376444b48ada2db1666e28
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>

For security DMA should be blocked at the SMMU by default
unless explicitly enabled for a device. SMMU is disabled
after reset with all streams bypassing the SMMU, and
abortion of all incoming transactions implements a default
deny policy on reset.
This patch also moves "bl1_platform_setup()" function from
arm_bl1_setup.c to FVP platforms' fvp_bl1_setup.c and
fvp_ve_bl1_setup.c files.

Change-Id: Ie0ffedc10219b1b884eb8af625bd4b6753749b1a
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>

This patch fixes this issue:
https://github.com/ARM-software/tf-issues/issues/660



The introduced changes are the following:

1) Some cores implement cache coherency maintenance operation on the
hardware level. For those cores, such as - but not only - the DynamIQ
cores, it is mandatory that TF-A is compiled with the
HW_ASSISTED_COHERENCY flag. If not, the core behaviour at runtime is
unpredictable. To prevent this, compile time checks have been added and
compilation errors are generated, if needed.

2) To enable this change for FVP, a logical separation has been done for
the core libraries. A system cannot contain cores of both groups, i.e.
cores that manage coherency on hardware and cores that don't do it. As
such, depending on the HW_ASSISTED_COHERENCY flag, FVP includes the
libraries only of the relevant cores.

3) The neoverse_e1.S file has been added to the FVP sources.

Change-Id: I787d15819b2add4ec0d238249e04bf0497dc12f3
Signed-off-by: John Tsichritzis <john.tsichritzis@arm.com>