ArmGicArchLib is no longer use so remove all remaining references and
implementations.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
StandaloneMmCpu driver is only used for Arm architecture and
StandaloneMmCoreEntryPointLib for Arm has specific implementation with
StandaloneMmCpu driver.
Move StandaloneMmCpu Driver and StandaloneMmCoreEntryPointLib for Arm
to ArmPkg.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
The StandaloneMm implementation for Arm sets up the stack in
the early startup code using the data section reserved in the
assembly code.
When TF-A loads the StandaloneMM binary in the DRAM it maps
the entire StandaloneMM memory region as Read Only.
Therefore, the initial startup assembly code updates the mem
permissions of the stack region to Read Write.
However, when the StandaloneMmCore is loaded the function
UpdateMmFoundationPeCoffPermissions() starts applying the
memory permissions based on the PE COFF sections as below:
A. If the section is not executable, it first removes the
executable permission of the section by calling TextUpdate().
TextUpdate() is the StandaloneMM MMU library function
ArmSetMemoryRegionNoExec().
B. It then checks if the section is writable, and if it is
it calls ReadWriteUpdater(), which invokes the StandaloneMM
MMU library function ArmClearMemoryRegionReadOnly() to make
the section writable.
However, this results in the stack being made read-only
between A and B. To understand this please see the following
flow.
1. TF-A sets the entire StandaloneMM region as Read Only.
2. The stack is reserved in the data section by the early
assembly entry point code.
+--------------------+ <--- Start of Data Section
| |
| Data Section |
| |
| +----------------+ | <--- Stack region
| | Stack | |
| +----------------+ |
| |
+--------------------+
3. The StanaloneMM early entry point code updates the
attributes of the stack to Read Write.
4. When UpdateMmFoundationPeCoffPermissions() sets the
permission of the data section to remove the Execute
attribute, it calls ArmSetMemoryRegionNoExec().
5. The ArmSetMemoryRegionNoExec() implementation gets the
attributes of the first granule which is at the start
of the data section, then clears the execute permission
and applies the attribute for the entire data section.
6. Since TF-A has mapped the entire section as read only
the first granule of the data section is read only and
therefore the stack region attributes are changed to
Read Only no execute.
7. Since the stack is read only after point A any updates
to the stack result in an exception.
To resolve this issue with update the library with FF-A v1.2,
get/set memory permission per page unit.
Links: https://developer.arm.com/documentation/den0140/latest/ [0]
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Add ArmFfaLib used in StandaloneMmCore/StandaloneMm Driver.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
PcdFfaEnabled is no more used because ArmFfaLib could find whether FF-A
is supported dynamically.
This patch removes usage of PcdFfaEnabled.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
Introduce a new helper function SpmMmStatusToEfiStatus() to convert
the SPM_MM status values to EFI_STATUS values.
Signed-off-by: Levi yun <yeoreum.yun@arm.com>
Change naming convention in ArmMmSvc.h with
MM to SPM_MM
This would make it clear to discern ABI protocol used to communicate
with secure partition.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
ArmTransferList is used to pass boot information according to
firmware handoff protocol specification [0].
When initializing StandaloneMm, it gets boot information from
the PHIT HOB in the TransferList.
[0] https://github.com/FirmwareHandoff/firmware_handoff
Signed-off-by: Levi Yun <yeoreum.yun@arm.com>
The current DescribeExceptionSyndrome is taking ESR as 32bit value.
However, ESR should be a 64 bit value. This change updates the function
to intake a 64bit value.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Signed-off-by: Kun Qin <kun.qin@microsoft.com>
The current VectorBase is taking value from a 64bit PCD into a UINTN
value, which could have truncated value for 32bit system.
In addition, the comparison between UINTN and INTN could lead to
undesired comparison outcome or compiler complaints.
This change updates all of them to be UINT64 based operation.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Signed-off-by: Kun Qin <kun.qin@microsoft.com>
Update function implementation to match interface definition. The return
should be bound to 0xffff0000, which is guaranteed to be a UINT32.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Signed-off-by: Kun Qin <kun.qin@microsoft.com>
The existing operation in ArmArchTimerLib is operating on UINT32 or
UINT64 based on the target system. This casting game originates from the
fact that timer frequency is UINTN type.
This change will simply promote all operations to UINT64 based, which
will remove the casting and conditional #if in the code for better
portability and readability.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Signed-off-by: Kun Qin <kun.qin@microsoft.com>
With VHE enabled, EL2 uses the EL2&0 translation regime, which is
compatible with the EL1&0 translation regime when it comes to the TCR
configuration register and the page table descriptor.
Given that some CPUs may have VHE force enabled when executing at EL2,
the MMU code needs to be able to deal with this even if it doesn't
enable VHE itself.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Remove the code path for execution at EL3, which just dumps an error.
None of the other code is remotely suitable for execution at EL3, and so
just ASSERT()'ing here is sufficient, and simplifies future changes
related to VHE.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
When VHE is enabled, some pre-existing timer system register specifiers
are redirected to the HYP timer. To access the conventional timer,
special aliases have to be used that end in _EL02.
These aliases are not understood by Clang's internal assembler, so use
the generic mnemonics instead.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
When building OvmfPkg in RELEASE mode in the XCODE5 toolchain, the
ASSERT_EFI_ERROR change prevents this error:
.../MdePkg/Library/UefiMemoryAllocationLib/MemoryAllocationLib.c:141:15:
error: variable 'Status' set but not used [-Werror,-Wunused-but-set-variable]
EFI_STATUS Status;
^
which is currently stopping the build.
When building in RELEASE mode in the CLANGPDB toolchain,the DEBUG macro
change prevents this error:
.../edk2/OvmfPkg/VirtioSerialDxe/VirtioSerial.c:28:22: error:
variable 'EventNames' is not needed and will not be
emitted [-Werror,-Wunneeded-internal-declaration]
STATIC CONST CHAR8 *EventNames[] = {
^
which is currently stopping the build.
CLANGDWARF produces the same error as CLANGPDB above, if
-Wno-unneeded-internal-declaration is removed from its build flags.
With the null DEBUG macro change, this warning suppression
can be removed from CLANGDWARF, which is considered a benefit
as it has the potential to catch real coding errors. This is
done in a subsequent commit.
This commit has the desirable side effect that we no longer require
(and cannot use) explicit `#ifndef MDEPKG_NDEBUG` around items only
used in DEBUG macros. This requires the ArmPkg change made here to
be in the same commit as the MdePkg changes.
Note: In common with existing macros in EDK II, including the pre-existing
and unchanged DEBUG/NOOPT versions of the macros which are modified here,
we use the standard approach of adding `do { ... } while (FALSE)` wrapping
to ensure that the macros behave correctly with surrounding code
(e.g. require a following ';' and do not combine in unexpected ways with
nearby conditionals).
Continuous-integration-options: PatchCheck.ignore-multi-package
Co-authored-by: Mikhail Krichanov <krichanov@ispras.ru>
Signed-off-by: Mike Beaton <mjsbeaton@gmail.com>
Cache maintenance operations by set/way are not broadcast, and operate
on individual architected caches, making them suitable only for
en/disabling cache levels, which is the job of secure firmware, to be
carried out while the CPU in question is not taking part in the
cache coherency protocol.
Managing the clean/dirty state of a memory range can only be done using
cache maintenance by virtual address.
So drop the set/way handling from ArmLib for ARM and AARCH64, as there
is no context where it can be used correctly from EDK2.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
As per the emailed RFC in
https://edk2.groups.io/g/devel/topic/rfc_move/107675828,
this patch moves CompilerIntrinsicsLib from ArmPkg to
MdePkg as this library provides compiler intrinsics, which
are industry standard.
This aligns with the goal of integrating ArmPkg into existing
packages: https://bugzilla.tianocore.org/show_bug.cgi?id=4121.
The newly placed CompilerIntrinsicsLib is added to MdeLibs.dsc.inc
as every DSC that builds ARM/AARCH64 needs this library added. The
old location is removed from every DSC in edk2 in this commit also
to not break bisectability with minimal hoop jumping.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
AsmMacroIoLib.h and AsmMacroIoLibV8.h are used by the
CompilerIntrinsicsLib, which is moving to MdePkg. These
functions provide standard definitions for ARM/AARCH64
assembly code, respectively, and so are moved to the arch
directories in MdePkg to avoid MdePkg having a
dependency on ArmPkg.
Now that the files are in Arm/ and AArch64/ directories,
the filenames are changed to AsmMacroLib.h as we can
distinguish the architecture from the path.
AsmMacroIoLib.inc is unused and so is removed.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
AArch64/ashlti3.S was using AsmMacroIoLib.h which is the ARM version
of these definitions. AsmMacroIoLibV8.h is the AARCH64 version of
these defintions. This patch moves that file to use the proper arch
file.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
The timer counter register can wrap around and when this happens,
we'll get misbehavior for any MicroSecondDelay() calls. This adds
handling for that.
Signed-off-by: Carsten Haitzler <carsten.haitzler@foss.arm.com>
Drop the pointless constructor in ArmArchTimerLib, which does nothing
useful, especially because AArch64 mandates the presence of the generic
timer, and 32-bit ARM is mostly obsolete these days.
To preserve the existing behavior in DEBUG builds when actually using
the timer, move the ASSERT () on a non-zero frequency to the associated
accessor helper function.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
This library is no longer used, and has been superseded by
ArmPsciResetSystemLib based on ArmMonitorLib. So delete it.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The SMCCC protocol stipulates the following:
- on AARCH64, 18 arguments can be passed, and 18 values can be returned,
via registers X0-x17;
- on ARM, 8 arguments can be passed, and 8 values can be returned.
This makes ArmSmcLib and ArmHvcLib as implemented currently unsuitable
for use with SMCCC services in general, although for PSCI in particular,
they work fine.
The dependency on both ArmSmcLib and ArmHvcLib is also impractical
because it requires every platform that consumes ArmMonitorLib to
provide resolutions for each, even though most platforms will only ever
need one of these (and the choice is made at compile time)
So let's drop these dependencies, and re-implement the asm helpers from
scratch. Note that the only difference is the actual instruction used
-HVC vs SMC- and so all other code can be shared.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Feature PCDs and fixed-at-build PCDs are identical in concept, but the
latter are accessible from assembler, whereas the former are not. So
convert the SMCCC conduit selection PCD to fixed-at-build so we can make
use of this in a subsequent patch.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The ARM SMCCC specification stipulates that the conduit for ARM monitor
calls can be determined by taking the conduit for PSCI calls. The
converse is also true, and so we can implement ResetSystemLib for PSCI
using ArmMonitorLib to choose between SMC and HVC instructions.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Fix up a handful of uncrustify infractions and drop the AuditOnly
override for uncrustify so we get the same uncrustify CI coverage as
other packages.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
ArmPsciResetSystemLib has been deprecated since commit
b2c55e7328 in 2017. The
lib itself has not been meaningfully updated in 10 years.
This commit removes the library to complete the deprecation
process and remove confusion over which library to use for
resetting an ARM platform.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
This patch moves GnuNoteBti.bin from ArmPkg to BaseTools as it
is used during the build by GCC. This removes an unnecessary
dependency on ArmPkg from BaseTools and keeps build related
files in BaseTools.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
This moves the GccLto files from ArmPkg to BaseTools as they
are files that are only used in the build. This removes an
artificial dependency on ArmPkg from BaseTools and keeps build
related files in BaseTools.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
This adds logging in failure cases of SetMemoryAttributes. This
is useful generally as if an attribute update fails, code will
typically break, but is added in particular to make debugging
incorrect bootloader usage of the Memory Attribute Protocol. This
patch updates the AARCH64 UpdateRegionMapping.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
This adds logging in failure cases of SetMemoryAttributes. This
is useful generally as if an attribute update fails, code will
typically break, but is added in particular to make debugging
incorrect bootloader usage of the Memory Attribute Protocol. This
patch updates the ARM32 SetMemoryAttributes.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
PcdArmArchTimerFreqInHz lets platforms override the architectural timer
frequency, but this really isn't supported in hardware by anything lower
than EL3. And we no longer support running edk2 at EL3.
So drop this Pcd and update ArmArchTimerLib to unconditionally use
ArmGenericTimerGetTimerFreq () instead.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
There is a bit of legacy code in ArmArchTimerLib that attempts to program
the generic timer based on a Pcd. This was only ever done for (32-bit)
ARM in case the cpu lacked the Security Extensions (because if they were
there, Secure firmware would be expected to have configured this before
switching to Non-secure state).
We don't support operation without Secure firmware anymore, but also the
code has been incorrectly invoked only on platforms that *do* implement
the Security Extensions since December 2020 and commit 0dd0d42ab5
("ArmPkg: use helper to check for Security extensions in ArmArchTimerLib")
As a result, edk2 has been overriding any programming done by Secure
Firmware on ARM platforms. Which is wrong.
Drop this programming in preparation for deleting the Pcd completely.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
Drop logic from the ARM architectural support libraries that can only
execute in EL3 on AArch64 or Monitor mode on 32-bit ARM. While early
32-bit ports (and even some early 64-bit code) included some monitor
logic in EDK2, UEFI per the spec runs in non-secure execution contexts
only, and secure monitor and other secure world duties are usually
delegated to TF-A (Trusted Firmware for the A profile).
Since there are no longer users of this code in EDK2 or the
edk2-platforms tree, let's remove it from the core support libraries.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The compiler will use this function if it is
left shifting a 128 bit value.
Seen when updating OpenSSL.
Signed-off-by: Jeff Brasen <jbrasen@nvidia.com>
Following the discussion at [1] and as the ArmLib relies on them,
move ArmPkg/Chipset/Aarch64[|Mmu].h files to the MdePkg.
Update the path to correctly include the moved files.
[1] https://edk2.groups.io/g/devel/message/111566
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Following the discussion at [1] and as the ArmLib relies on them,
move ArmPkg/Chipset/ArmV7[|Mmu].h files to the MdePkg.
Update the path to correctly include the moved files.
[1] https://edk2.groups.io/g/devel/message/111566
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
This reverts commit 32460bb5b1, which is
no longer needed as ArmVirtQemu now has its own special implementation
for ArmMonitorLib.
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
ArmVirtQemu may execute at EL2, in which case monitor calls are
generally made using SMC instructions instead of HVC instructions.
Whether or not this is the case can only be decided at runtime, and so
the associated PCD needs to be settable at runtime, if the platform
definition chooses so. This implies a boolean PCD, given that a feature
PCD is build-time configurable only.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Committed-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Doug Flick [MSFT] <doug.edk2@gmail.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The DebugPeCoffExtraActionLib implemention in ArmPkg contains some cruft
that dates back to the original RVCT based ARM port, and support for
RVCT was dropped a while ago.
Also drop the handling of Cygwin specific paths, which is highly
unlikely to be still depended upon by anyone.
Tweak the logic so that only two versions of the DEBUG() invocations
remain: one for __GNUC__ when PdbPointer is set, and the fallback that
just prints the image address and the address of the entrypoint.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Sami Mujawar <sami.mujawar@arm.com>
mReplaceLiveEntryFunc is a function pointer but assigned as a VOID*
pointer:
mReplaceLiveEntryFunc = *(VOID **)GET_GUID_HOB_DATA (Hob);
This leads to the Visual Studio warning:
nonstandard extension, function/data pointer conversion in
expression
This change updates the assignment to avoid using a data pointer and
defines a type for the function pointer to succinctly and accurately
refer to the type when it is used in the library code.
Signed-off-by: Michael Kubacki <michael.kubacki@microsoft.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Create a helper function to query whether ID_AA64MFR1_EL1 indicates
presence of the Embedded Trace Extension (ETE). This feature is only
visible in AARCH64 state.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Pierre Gondois <pierre.gondois@arm.com>
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Pierre Gondois <pierre.gondois@arm.com>
Create a helper function to query whether ID_AA64MFR1_EL1 indicates
presence of the Trace Buffer Extension (TRBE). This feature is only
visible in AARCH64 state.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Pierre Gondois <pierre.gondois@arm.com>
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Pierre Gondois <pierre.gondois@arm.com>
Create a helper function to query whether ID_AA64MFR1_EL1 indicates
presence of the Virtualization Host Extensions. This feature is only
visible in AARCH64 state.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Sami Mujawar <sami.mujawar@arm.com>
Tested-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Jake reports that the IS_ARM_MEMORY_REGION_ATTRIBUTES_SECURE() macro is
no longer accurate since commit 852227a9d5 ("ArmPkg/Mmu: Remove
handling of NONSECURE memory regions").
Fortunately, it only affects the NS bit in level 1 short descriptors,
which is ignored when executing in non-secure mode. And given that
running UEFI in the secure world is not a use case we aim to support,
let's just drop this logic altogether.
Reported-by: Jake Garver <jake@nvidia.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
If gST->ConOut is available when Arm's DefaultExceptionHandler is
running, AsciiPrint will get called to attempt to print to ConOut, in
addition to the serial output.
AsciiPrint calls AsciiInternalPrint in UefiLibPrint.c which in turn
calls AllocatePool to allocate a buffer to convert the Ascii input
string to a Unicode string to pass to ConOut->OutputString.
Per the comment on DefaultExceptionHandler, we should not be allocating
memory in the exception handler, as this can cause the exception handler
to fail if we had a memory exception or the system state is such that we
cannot allocate memory.
It has been observed on ArmVirtQemu that exceptions generated in the
memory handling code will fail to output the stack dump and CPU state
that is critical to debugging because the AllocatePool will fail.
This patch fixes the Arm and AARCH64 DefaultExceptionHandlers to not
allocate memory when ConOut is available and instead use stack memory to
convert the Ascii string needed for SerialPortWrite to the Unicode
string needed for ConOut->OutputString. Correspondingly, ArmVirtQemu can
now output the stack dump and CPU state when hitting an exception in
memory code.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
Ard Biesheuvel
Levi Yun
Kun Qin
Mike Beaton
Oliver Smith-Denny
Carsten Haitzler
Leif Lindholm
Jeff Brasen
Pierre Gondois
Michael Kubacki
Sami Mujawar