This patch fixes up the AArch64 assembly code to use
adrp/adr instructions instead of ldr instruction for
reference to symbols. This allows these assembly
sequences to be Position Independant. Note that the
the reference to sizes have been replaced with
calculation of size at runtime. This is because size
is a constant value and does not depend on execution
address and using PC relative instructions for loading
them makes them relative to execution address. Also
we cannot use `ldr` instruction to load size as it
generates a dynamic relocation entry which must *not*
be fixed up and it is difficult for a dynamic loader
to differentiate which entries need to be skipped.
Change-Id: I8bf4ed5c58a9703629e5498a27624500ef40a836
Signed-off-by: Soby Mathew <soby.mathew@arm.com>
The `finish_console_register` macro is used by the multi console
framework to register the `console_t` driver callbacks. It relied
on weak references to the `ldr` instruction to populate 0 to the
callback in case the driver has not defined the appropriate
function. Use of `ldr` instruction to load absolute address to a
reference makes the binary position dependant. These instructions
should be replaced with adrp/adr instruction for position independant
executable(PIE). But adrp/adr instructions don't work well with weak
references as described in GNU ld bugzilla issue 22589.
This patch defines a new version of `finish_console_register` macro
which can spcify which driver callbacks are valid and deprecates the
old one. If any of the argument is not specified, then the macro
populates 0 for that callback. Hence the functionality of the previous
deprecated macro is preserved. The USE_FINISH_CONSOLE_REG_2 define
is used to select the new variant of the macro and will be removed
once the deprecated variant is removed.
All the upstream console drivers have been migrated to use the new
macro in this patch.
NOTE: Platforms be aware that the new variant of the
`finish_console_register` should be used and the old variant is
deprecated.
Change-Id: Ia6a67aaf2aa3ba93932992d683587bbd0ad25259
Signed-off-by: Soby Mathew <soby.mathew@arm.com>
Pointer authentication is an Armv8.3 feature that introduces
instructions that can be used to authenticate and verify pointers.
Pointer authentication instructions are allowed to be accessed from all
ELs but only when EL3 explicitly allows for it; otherwise, their usage
will trap to EL3. Since EL3 doesn't have trap handling in place, this
patch unconditionally disables all related traps to EL3 to avoid
potential misconfiguration leading to an unhandled EL3 exception.
FixesARM-software/tf-issues#629
Change-Id: I9bd2efe0dc714196f503713b721ffbf05672c14d
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>
Check_vector_size checks if the size of the vector fits
in the size reserved for it. This check creates problems in
the Clang assembler. A new macro, end_vector_entry, is added
and check_vector_size is deprecated.
This new macro fills the current exception vector until the next
exception vector. If the size of the current vector is bigger
than 32 instructions then it gives an error.
Change-Id: Ie8545cf1003a1e31656a1018dd6b4c28a4eaf671
Signed-off-by: Roberto Vargas <roberto.vargas@arm.com>
These directives are only used when stabs debugging information
is used, but we use ELF which uses DWARF debugging information.
Clang assembler doesn't support these directives, and removing
them makes the code more compatible with clang.
Change-Id: I2803f22ebd24c0fe248e04ef1b17de9cec5f89c4
Signed-off-by: Roberto Vargas <roberto.vargas@arm.com>
At present, the function provided by the translation library to enable
MMU constructs appropriate values for translation library, and programs
them to the right registers. The construction of initial values,
however, is only required once as both the primary and secondaries
program the same values.
Additionally, the MMU-enabling function is written in C, which means
there's an active stack at the time of enabling MMU. On some systems,
like Arm DynamIQ, having active stack while enabling MMU during warm
boot might lead to coherency problems.
This patch addresses both the above problems by:
- Splitting the MMU-enabling function into two: one that sets up
values to be programmed into the registers, and another one that
takes the pre-computed values and writes to the appropriate
registers. With this, the primary effectively calls both functions
to have the MMU enabled, but secondaries only need to call the
latter.
- Rewriting the function that enables MMU in assembly so that it
doesn't use stack.
This patch fixes a bunch of MISRA issues on the way.
Change-Id: I0faca97263a970ffe765f0e731a1417e43fbfc45
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>
For the adr instruction, it require the label's offset from the
address of this instruction must be in the range +/-1MB. If the
option "BL2_IN_XIP_MEM" is set to '1', in some cases, BL2's RW
memory will not in the range of +/-1MB from BL2's RO memory region.
so we need to use ldr instruction to cover this case.
Signed-off-by: Jiafei Pan <Jiafei.Pan@nxp.com>
In some use-cases BL2 will be stored in eXecute In Place (XIP) memory,
like BL1. In these use-cases, it is necessary to initialize the RW sections
in RAM, while leaving the RO sections in place. This patch enable this
use-case with a new build option, BL2_IN_XIP_MEM. For now, this option
is only supported when BL2_AT_EL3 is 1.
Signed-off-by: Jiafei Pan <Jiafei.Pan@nxp.com>
This patch enables BL2 to execute at the highest exception level
without any dependancy on TF BL1. This enables platforms which already
have a non-TF Boot ROM to directly load and execute BL2 and subsequent BL
stages without need for BL1. This is not currently possible because
BL2 executes at S-EL1 and cannot jump straight to EL3.
Change-Id: Ief1efca4598560b1b8c8e61fbe26d1f44e929d69
Signed-off-by: Roberto Vargas <roberto.vargas@arm.com>
Invalidate the Branch Target Buffer (BTB) on entry to EL3 by disabling
and enabling the MMU. To achieve this without performing any branch
instruction, a per-cpu vbar is installed which executes the workaround
and then branches off to the corresponding vector entry in the main
vector table. A side effect of this change is that the main vbar is
configured before any reset handling. This is to allow the per-cpu
reset function to override the vbar setting.
This workaround is enabled by default on the affected CPUs.
Change-Id: I97788d38463a5840a410e3cea85ed297a1678265
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>
This patch overhauls the console API to allow for multiple console
instances of different drivers that are active at the same time. Instead
of binding to well-known function names (like console_core_init),
consoles now provide a register function (e.g. console_16550_register())
that will hook them into the list of active consoles. All console
operations will be dispatched to all consoles currently in the list.
The new API will be selected by the build-time option MULTI_CONSOLE_API,
which defaults to ${ERROR_DEPRECATED} for now. The old console API code
will be retained to stay backwards-compatible to older platforms, but
should no longer be used for any newly added platforms and can hopefully
be removed at some point in the future.
The new console API is intended to be used for both normal (bootup) and
crash use cases, freeing platforms of the need to set up the crash
console separately. Consoles can be individually configured to be active
active at boot (until first handoff to EL2), at runtime (after first
handoff to EL2), and/or after a crash. Console drivers should set a sane
default upon registration that can be overridden with the
console_set_scope() call. Code to hook up the crash reporting mechanism
to this framework will be added with a later patch.
This patch only affects AArch64, but the new API could easily be ported
to AArch32 as well if desired.
Change-Id: I35c5aa2cb3f719cfddd15565eb13c7cde4162549
Signed-off-by: Julius Werner <jwerner@chromium.org>
This patch adds a new build option, ENABLE_SVE_FOR_NS, which when set
to one EL3 will check to see if the Scalable Vector Extension (SVE) is
implemented when entering and exiting the Non-secure world.
If SVE is implemented, EL3 will do the following:
- Entry to Non-secure world: SIMD, FP and SVE functionality is enabled.
- Exit from Non-secure world: SIMD, FP and SVE functionality is
disabled. As SIMD and FP registers are part of the SVE Z-registers
then any use of SIMD / FP functionality would corrupt the SVE
registers.
The build option default is 1. The SVE functionality is only supported
on AArch64 and so the build option is set to zero when the target
archiecture is AArch32.
This build option is not compatible with the CTX_INCLUDE_FPREGS - an
assert will be raised on platforms where SVE is implemented and both
ENABLE_SVE_FOR_NS and CTX_INCLUDE_FPREGS are set to 1.
Also note this change prevents secure world use of FP&SIMD registers on
SVE-enabled platforms. Existing Secure-EL1 Payloads will not work on
such platforms unless ENABLE_SVE_FOR_NS is set to 0.
Additionally, on the first entry into the Non-secure world the SVE
functionality is enabled and the SVE Z-register length is set to the
maximum size allowed by the architecture. This includes the use case
where EL2 is implemented but not used.
Change-Id: Ie2d733ddaba0b9bef1d7c9765503155188fe7dae
Signed-off-by: David Cunado <david.cunado@arm.com>
Factor out SPE operations in a separate file. Use the publish
subscribe framework to drain the SPE buffers before entering secure
world. Additionally, enable SPE before entering normal world.
A side effect of this change is that the profiling buffers are now
only drained when a transition from normal world to secure world
happens. Previously they were drained also on return from secure
world, which is unnecessary as SPE is not supported in S-EL1.
Change-Id: I17582c689b4b525770dbb6db098b3a0b5777b70a
Signed-off-by: Dimitris Papastamos <dimitris.papastamos@arm.com>
A Secure Partition is a software execution environment instantiated in
S-EL0 that can be used to implement simple management and security
services. Since S-EL0 is an unprivileged exception level, a Secure
Partition relies on privileged firmware e.g. ARM Trusted Firmware to be
granted access to system and processor resources. Essentially, it is a
software sandbox that runs under the control of privileged software in
the Secure World and accesses the following system resources:
- Memory and device regions in the system address map.
- PE system registers.
- A range of asynchronous exceptions e.g. interrupts.
- A range of synchronous exceptions e.g. SMC function identifiers.
A Secure Partition enables privileged firmware to implement only the
absolutely essential secure services in EL3 and instantiate the rest in
a partition. Since the partition executes in S-EL0, its implementation
cannot be overly complex.
The component in ARM Trusted Firmware responsible for managing a Secure
Partition is called the Secure Partition Manager (SPM). The SPM is
responsible for the following:
- Validating and allocating resources requested by a Secure Partition.
- Implementing a well defined interface that is used for initialising a
Secure Partition.
- Implementing a well defined interface that is used by the normal world
and other secure services for accessing the services exported by a
Secure Partition.
- Implementing a well defined interface that is used by a Secure
Partition to fulfil service requests.
- Instantiating the software execution environment required by a Secure
Partition to fulfil a service request.
Change-Id: I6f7862d6bba8732db5b73f54e789d717a35e802f
Co-authored-by: Douglas Raillard <douglas.raillard@arm.com>
Co-authored-by: Sandrine Bailleux <sandrine.bailleux@arm.com>
Co-authored-by: Achin Gupta <achin.gupta@arm.com>
Co-authored-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Add Call Frame Information assembler directives to vector entries so
that debuggers display the backtrace of functions that triggered a
synchronous exception. For example, a function triggering a data abort
will be easier to debug if the backtrace can be displayed from a
breakpoint at the beginning of the synchronous exception vector.
DS-5 needs CFI otherwise it will not attempt to display the backtrace.
Other debuggers might have other needs. These debug information are
stored in the ELF file but not in the final binary.
Change-Id: I32dc4e4b7af02546c93c1a45c71a1f6d710d36b1
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
SPE is only supported in non-secure state. Accesses to SPE specific
registers from SEL1 will trap to EL3. During a world switch, before
`TTBR` is modified the SPE profiling buffers are drained. This is to
avoid a potential invalid memory access in SEL1.
SPE is architecturally specified only for AArch64.
Change-Id: I04a96427d9f9d586c331913d815fdc726855f6b0
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
This patch updates the el3_arch_init_common macro so that it fully
initialises essential control registers rather then relying on hardware
to set the reset values.
The context management functions are also updated to fully initialise
the appropriate control registers when initialising the non-secure and
secure context structures and when preparing to leave EL3 for a lower
EL.
This gives better alignement with the ARM ARM which states that software
must initialise RES0 and RES1 fields with 0 / 1.
This patch also corrects the following typos:
"NASCR definitions" -> "NSACR definitions"
Change-Id: Ia8940b8351dc27bc09e2138b011e249655041cfc
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>
Introduce new build option ENABLE_STACK_PROTECTOR. It enables
compilation of all BL images with one of the GCC -fstack-protector-*
options.
A new platform function plat_get_stack_protector_canary() is introduced.
It returns a value that is used to initialize the canary for stack
corruption detection. Returning a random value will prevent an attacker
from predicting the value and greatly increase the effectiveness of the
protection.
A message is printed at the ERROR level when a stack corruption is
detected.
To be effective, the global data must be stored at an address
lower than the base of the stacks. Failure to do so would allow an
attacker to overwrite the canary as part of an attack which would void
the protection.
FVP implementation of plat_get_stack_protector_canary is weak as
there is no real source of entropy on the FVP. It therefore relies on a
timer's value, which could be predictable.
Change-Id: Icaaee96392733b721fa7c86a81d03660d3c1bc06
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
Trusted Firmware currently has no support for secure self-hosted
debug. To avoid unexpected exceptions, disable software debug
exceptions, other than software breakpoint instruction exceptions,
from all exception levels in secure state. This applies to both
AArch32 and AArch64 EL3 initialization.
Change-Id: Id097e54a6bbcd0ca6a2be930df5d860d8d09e777
Signed-off-by: dp-arm <dimitris.papastamos@arm.com>
Introduce zeromem_dczva function on AArch64 that can handle unaligned
addresses and make use of DC ZVA instruction to zero a whole block at a
time. This zeroing takes place directly in the cache to speed it up
without doing external memory access.
Remove the zeromem16 function on AArch64 and replace it with an alias to
zeromem. This zeromem16 function is now deprecated.
Remove the 16-bytes alignment constraint on __BSS_START__ in
firmware-design.md as it is now not mandatory anymore (it used to comply
with zeromem16 requirements).
Change the 16-bytes alignment constraints in SP min's linker script to a
8-bytes alignment constraint as the AArch32 zeromem implementation is now
more efficient on 8-bytes aligned addresses.
Introduce zero_normalmem and zeromem helpers in platform agnostic header
that are implemented this way:
* AArch32:
* zero_normalmem: zero using usual data access
* zeromem: alias for zero_normalmem
* AArch64:
* zero_normalmem: zero normal memory using DC ZVA instruction
(needs MMU enabled)
* zeromem: zero using usual data access
Usage guidelines: in most cases, zero_normalmem should be preferred.
There are 2 scenarios where zeromem (or memset) must be used instead:
* Code that must run with MMU disabled (which means all memory is
considered device memory for data accesses).
* Code that fills device memory with null bytes.
Optionally, the following rule can be applied if performance is
important:
* Code zeroing small areas (few bytes) that are not secrets should use
memset to take advantage of compiler optimizations.
Note: Code zeroing security-related critical information should use
zero_normalmem/zeromem instead of memset to avoid removal by
compilers' optimizations in some cases or misbehaving versions of GCC.
FixesARM-software/tf-issues#408
Change-Id: Iafd9663fc1070413c3e1904e54091cf60effaa82
Signed-off-by: Douglas Raillard <douglas.raillard@arm.com>
At present, spin locks can only defined from C files. Add some macros
such that they can be defined from assembly files too.
Change-Id: I64f0c214062f5c15b3c8b412c7f25c908e87d970
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@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>
There are many instances in ARM Trusted Firmware where control is
transferred to functions from which return isn't expected. Such jumps
are made using 'bl' instruction to provide the callee with the location
from which it was jumped to. Additionally, debuggers infer the caller by
examining where 'lr' register points to. If a 'bl' of the nature
described above falls at the end of an assembly function, 'lr' will be
left pointing to a location outside of the function range. This misleads
the debugger back trace.
This patch defines a 'no_ret' macro to be used when jumping to functions
from which return isn't expected. The macro ensures to use 'bl'
instruction for the jump, and also, for debug builds, places a 'nop'
instruction immediately thereafter (unless instructed otherwise) so as
to leave 'lr' pointing within the function range.
Change-Id: Ib34c69fc09197cfd57bc06e147cc8252910e01b0
Co-authored-by: Douglas Raillard <douglas.raillard@arm.com>
Signed-off-by: Jeenu Viswambharan <jeenu.viswambharan@arm.com>
In order to avoid unexpected traps into EL3/MON mode, this patch
resets the debug registers, MDCR_EL3 and MDCR_EL2 for AArch64,
and SDCR and HDCR for AArch32.
MDCR_EL3/SDCR is zero'ed when EL3/MON mode is entered, at the
start of BL1 and BL31/SMP_MIN.
For MDCR_EL2/HDCR, this patch zero's the bits that are
architecturally UNKNOWN values on reset. This is done when
exiting from EL3/MON mode but only on platforms that support
EL2/HYP mode but choose to exit to EL1/SVC mode.
FixesARM-software/tf-issues#430
Change-Id: Idb992232163c072faa08892251b5626ae4c3a5b6
Signed-off-by: David Cunado <david.cunado@arm.com>
This patch moves assembler macros which are not architecture specific
to a new file `asm_macros_common.S` and moves the `el3_common_macros.S`
into `aarch64` specific folder.
Change-Id: I444a1ee3346597bf26a8b827480cd9640b38c826
This patch moves the PSCI services and BL31 frameworks like context
management and per-cpu data into new library components `PSCI` and
`el3_runtime` respectively. This enables PSCI to be built independently from
BL31. A new `psci_lib.mk` makefile is introduced which adds the relevant
PSCI library sources and gets included by `bl31.mk`. Other changes which
are done as part of this patch are:
* The runtime services framework is now moved to the `common/` folder to
enable reuse.
* The `asm_macros.S` and `assert_macros.S` helpers are moved to architecture
specific folder.
* The `plat_psci_common.c` is moved from the `plat/common/aarch64/` folder
to `plat/common` folder. The original file location now has a stub which
just includes the file from new location to maintain platform compatibility.
Most of the changes wouldn't affect platform builds as they just involve
changes to the generic bl1.mk and bl31.mk makefiles.
NOTE: THE `plat_psci_common.c` FILE HAS MOVED LOCATION AND THE STUB FILE AT
THE ORIGINAL LOCATION IS NOW DEPRECATED. PLATFORMS SHOULD MODIFY THEIR
MAKEFILES TO INCLUDE THE FILE FROM THE NEW LOCATION.
Change-Id: I6bd87d5b59424995c6a65ef8076d4fda91ad5e86
Soby Mathew
Antonio Niño Díaz
Jeenu Viswambharan
Roberto Vargas
Qixiang Xu
Jiafei Pan
davidcunado-arm
Dimitris Papastamos
Julius Werner
Douglas Raillard
danh-arm
Masahiro Yamada