I am moving this command outside of EDK2.
This reverts commit 2c2cb23528.
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
The top two bits of the Extended BIOS ROM Size field indicates the unit
used for the remaining 14 bits. If the size is greater than 16GB, the
unit is gigabytes.
The test for this uses the local BiosPhysicalSize variable, which is a
UINTN, meaning that when building for ARM/CLANGDWARF we have a
tautological constant comparison, which the toolchain flags now we've
stopped disabling that warning.
So switch the BiosPhysicalSize variable to UINT64.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Rebecca Cran <rebecca@bsdio.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
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>
Much of the MMU logic was written without function headers. This patch
adds function headers where absent and updates function headers which
do not match the EDK2 standard.
Signed-off-by: Taylor Beebe <t@taylorbeebe.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
There are ASSERTs present in the MMU logic to ensure various
functions return successfully, but these ASSERTs may be ignored
on release builds causing unsafe behavior. This patch updates
the logic to handle unexpected return values and branch safely.
Signed-off-by: Taylor Beebe <t@taylorbeebe.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
This patch updates the GetMemoryRegion() function to handle the case
where there is no mapping for the requested address.
The original logic for the ARM would hit an ASSERT after
GetMemoryRegionPage() returned EFI_SUCCESS but did not update The
RegionLength parameter.
The original logic for the AARCH64 would never initialize the
RegionLength parameter to zero and return EFI_SUCCESS after
traversing an unknown number of pages.
To fix this, update the logic for both architecture to return
EFI_NO_MAPPING if the BaseAddress being checked is unmapped.
Signed-off-by: Taylor Beebe <t@taylorbeebe.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
This patch applies Uncrustify to the following files:
ArmPkg/Drivers/MmCommunicationPei/MmCommunicationPei.c
ArmPkg/Include/IndustryStandard/ArmStdSmc.h
Signed-off-by: Taylor Beebe <t@taylorbeebe.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4464
This change introduced the MM communicate support in PEI phase for ARM
based platforms. Similar to the DXE counterpart, `PcdMmBufferBase` is
used as communicate buffer and SMC will be invoked to communicate to
TrustZone when MMI is requested.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Co-authored-by: Ronny Hansen <hansen.ronny@microsoft.com>
Co-authored-by: Shriram Masanamuthu Chinnathurai <shriramma@microsoft.com>
Co-authored-by: Preshit Harlikar <pharlikar@microsoft.com>
Signed-off-by: Kun Qin <kuqin@microsoft.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The OP-TEE secure OS exposes a non-secure memory region for
communication between the secure OS itself and any clients in the
non-secure firmware. This memory is writable by non-secure and is not
used for code only data, and so it should be mapped non-executable.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Now that we have a sane API to set and clear memory permissions that
works the same on ARM and AArch64, we no longer have a need for the
individual set/clear no-access/read-only/no-exec helpers so let's drop
them.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Now that ArmSetMemoryAttributes() permits a mask to be provided, we can
simplify the implementation the UEFI memory attribute protocol
substantially, and just pass on the requested mask to be set or cleared
directly.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Oliver Smith-Denny <osde@linux.microsoft.com>
Reviewed-by: Michael Kubacki <michael.kubacki@microsoft.com>
Implement the newly defined PPI that permits the PEI core and DXE IPL to
manage memory permissions on ranges of DRAM, for doing things like
mapping the stack non-executable, or granting executable permissions to
shadowed PEIMs.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Oliver Smith-Denny <osde@linux.microsoft.com>
Reviewed-by: Michael Kubacki <michael.kubacki@microsoft.com>
Currently, ArmSetMemoryAttributes () takes a combination of
EFI_MEMORY_xx constants describing the memory type and permission
attributes that should be set on a region of memory. In cases where the
memory type is omitted, we assume that the memory permissions being set
are final, and that existing memory permissions can be discarded.
This is problematic, because we aim to map memory non-executable
(EFI_MEMORY_XP) by default, and only relax this requirement for code
regions that are mapped read-only (EFI_MEMORY_RO). Currently, setting
one permission clears the other, and so code managing these permissions
has to be aware of the existing permissions in order to be able to
preserve them, and this is not always tractable (e.g., the UEFI memory
attribute protocol implements an abstraction that promises to preserve
memory permissions that it is not operating on explicitly).
So let's add an AttributeMask parameter to ArmSetMemoryAttributes(),
which is permitted to be non-zero if no memory type is being provided,
in which case only memory permission attributes covered in the mask will
be affected by the update.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Oliver Smith-Denny <osde@linux.microsoft.com>
Reviewed-by: Michael Kubacki <michael.kubacki@microsoft.com>
They are useful for those platforms where SMC SiP calls exist.
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Uncrustify checks are too rigid, making them counter-productive:
- it leads to code that is arguably harder to parse visually (e.g.,
the changes to ArmPkg/Include/Chipset/AArch64Mmu.h in commit
429309e0c6)
- it forces indentation-only changes to code in the vicinity of actual
changes, making the code history more bloated than necessary (see
commit 7f198321ee for an example)
- finding out from the web UI what exactly Uncrustify objected to is not
straight-forward.
So let's enable AuditMode for ArmPkg, so that interested parties can see
the uncrustify recommendations if desired, but without preventing the
changes from being merged. This leaves it at the discretion of the
ArmPkg maintainers to decide which level of conformance is required.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: "Kinney, Michael D" <michael.d.kinney@intel.com>
Cc: Michael Kubacki <mikuback@linux.microsoft.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
The ArmGicAcknowledgeInterrupt () returns the value returned by the
Interrupt Acknowledge Register and the InterruptID separately in an out
parameter.
The function documents the following: 'InterruptId is returned
separately from the register value because in the GICv2 the register
value contains the CpuId and InterruptId while in the GICv3 the register
value is only the InterruptId.'
This function skips setting the InterruptId in the out parameter for
GICv3. Although the return value from the function is the InterruptId
for GICv3, this breaks the function usage model as the caller expects
the InterruptId in the out parameter for the function. e.g. The caller
may end up using the InterruptID which could potentially be an
uninitialised variable value.
Therefore, set the InterruptID in the function out parameter for GICv3
as well.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
According to the GIC architecture version 3 and 4 specification, the
maximum number of INTID bits supported in the CPU interface is 24.
Considering this the RegShift variable is not required to be more than 8
bits. Therefore, make the RegShift variable type to UINT8. Also add
necessary typecasts when calculating the RegOffset and RegShift values.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
GICD_SGIR is a 32-bit register, of which INTID is bits [3:0] and Bits
[14:4] is RES0. Since SgiId parameter in the function ArmGicSendSgiTo ()
is UINT8, mask unused bits of SgiId before writing to the GICD_SGIR
register to prevent accidental setting of the RES0 bits.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The IrqInterruptHandler () and ExitBootServicesEvent () function
declarations were unused. Therefore, remove these declarations.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
The EIOR register of the Gic CPU interface is a 32 bit register.
However, the HARDWARE_INTERRUPT_SOURCE used to represent the interrupt
source (Interrupt ID) is typedefed as UINTN, see
EmbeddedPkg\Include\Protocol\HardwareInterrupt.h
Therfore, typecast the interrupt ID (Source) value to UINT32 before
setting the EOIR register. Also, add an assert to check that the value
does not exceed 32 bits.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Although the maximum interrupt ID on GicV2 is 10bit and for GicV3/4 is
24bit, and that the IAR and EOIR registers of the Gic CPU interface are
32 bit; the typedef HARDWARE_INTERRUPT_SOURCE is defined as UINTN in
EmbeddedPkg\Include\Protocol\HardwareInterrupt.h
Therefore, use UINTN for Gic Interrupt variables and use appropriate
typecasts wherever needed.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
The CPU Interface Identification Register (GICC_IIDR) is a 32-bit
register. Since ArmGicGetInterfaceIdentification () returns the value
read from the GICC_IIDR register, update the return type for this
function to UINT32.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
According to edk2 coding standard specification, Non-Boolean comparisons
must use a compare operator (==, !=, >, < >=, <=). See Section 5.7.2.1
at https://edk2-docs.gitbook.io/
edk-ii-c-coding-standards-specification/5_source_files/ 57_c_programming
Therefore, fix the comparison in ArmGicEnableDistributor()
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
The Software Generated Interrupt Register (GICD_SGIR) is a 32 bit
register with the following bit assignment:
TargetListFilter, bits [25:24]
CPUTargetList, bits [23:16]
NSATT, bit [15]
SGIINTID, bits [3:0]
Therefore, modify the TargetListFilter, CPUTargetList, SGI Interrupt ID
parameters of the ArmGicSendSgiTo () to use UINT8 instead of INTN.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
The data type used by variables representing the
GicInterruptInterfaceBase has been inconsistently used in the ArmGic
driver and the library. The PCD defined for the GIC Interrupt interface
base address is UINT64. However, the data types for the variables used
is UINTN, INTN, and at some places UINT32.
Therefore, update the data types to use UINTN and add necessary
typecasts when reading values from the PCD. This should then be
consistent across AArch32 and AArch64 builds.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The data type used by variables representing the GicDistributorBase has
been inconsistently used in the ArmGic driver and the library. The PCD
defined for the GIC Distributor base address is UINT64. However, the
data types for the variables used is UINTN, INTN, and at some places
UINT32.
Therefore, update the data types to use UINTN and add necessary
typecasts when reading values from the PCD. This should then be
consistent across AArch32 and AArch64 builds.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
The GIC v2 base addresses can be 64bit, don't limit to 32 on 64bit
machines.
Signed-off-by: Neil Jones <neil.jones@blaize.com>
Reviewed-by: Pedro Falcato <pedro.falcato@gmail.com>
Reviewed-by: Sami Mujawar <sami.mujawar@arm.com>
The DXE core implementation of PcdDxeNxMemoryProtectionPolicy already
contains an assertion that EfiConventionalMemory and EfiBootServicesData
are subjected to the same policy when it comes to the use of NX
permissions. The reason for this is that we may otherwise end up with
unbounded recursion in the page table code, given that allocating a page
table would then involve a permission attribute change, and this could
result in the need for a block entry to be split, which would trigger
the allocation of a page table recursively.
For the same reason, a shortcut exists in ApplyMemoryProtectionPolicy()
where, instead of setting the memory attributes unconditionally, we
compare the NX policies and avoid touching the page tables if they are
the same for the old and the new memory types. Without this shortcut, we
may end up in a situation where, as the CPU arch protocol DXE driver is
ramping up, the same unbounded recursion is triggered, due to the fact
that the NX policy for EfiConventionalMemory has not been applied yet.
To break this cycle, let's remap all EfiConventionalMemory regions
according to the NX policy for EfiBootServicesData before exposing the
CPU arch protocol to the DXE core and other drivers. This ensures that
creating EfiBootServicesData allocations does not result in memory
attribute changes, and therefore no recursion.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
The PCD gEfiMdeModulePkgTokenSpaceGuid.PcdEmuVariableNvModeEnable
indicates if a variable driver will emulate the variable NV mode.
This PCD is defined as [PcdsFixedAtBuild, PcdsPatchableInModule,
PcdsDynamic, PcdsDynamicEx].
Some firmware builds may define this PCD as a dynamic PCD and
initialise the value at runtime. Therefore, move the PCD declaration
from the [FixedPcd] section to the [Pcd] section in the platform
boot manager library file PlatformBootManagerLib.inf. Without this
change the build would not succeed.
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4463
When the AARCH64 CpuDxe attempts to SyncCacheConfig() with the GCD, it
collects the page attributes as:
EntryAttribute = Entry & TT_ATTR_INDX_MASK
However, TT_ATTR_INDX_MASK only masks the cacheability attributes and
drops the memory protections attributes. Importantly, it also drops the
TT_AF (access flag) which is now wired up in EDK2 to represent
EFI_MEMORY_RP, so by default all SystemMem pages will report as
EFI_MEMORY_RP to the GCD. The GCD currently drops that silently, because
the Capabilities field in the GCD does not support EFI_MEMORY_RP by
default.
However, some ranges may support EFI_MEMORY_RP and incorrectly mark
those ranges as read protected. In conjunction with another change on
the mailing list (see: https://edk2.groups.io/g/devel/topic/98505340),
this causes an access flag fault incorrectly. See the linked BZ below
for full details.
This patch exposes all memory protections attributes to the GCD layer so
it can correctly set pages as EFI_MEMORY[RP|XP|RO] when it initially
syncs.
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Taylor Beebe <t@taylorbeebe.com>
Cc: Sean Brogan <sean.brogan@microsoft.com>
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Michael Kubacki <michael.kubacki@microsoft.com>
The helper that updates live page table entries writes a zero entry,
invalidates the covered address range from the TLBs, and finally writes
the actual entry. This ensures that no TLB conflicts can occur.
Writing the final entry needs to complete before any translations can be
performed, as otherwise, the zero entry, which describes an invalid
translation, may be observed by the page table walker, resulting in a
translation fault. For this reason, the final write is followed by a DSB
barrier instruction.
However, this barrier will not stall the pipeline, and instruction
fetches may still hit this invalid translation, as has been observed and
reported by Oliver. To ensure that the new translation is fully active
before returning from this helper, we have to insert an ISB barrier as
well.
Reported-by: Oliver Steffen <osteffen@redhat.com>
Tested-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Acked-by: Michael D Kinney <michael.d.kinney@intel.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The UEFI Shell is a non-active boot option, at the opposite of UiApp.
If no valid boot option is found, UiApp is selected. UiApp requires a
human interaction. When installing a new EDKII image in CIs or when
scripting is required, this is problematic.
If no valid boot option is discovered, add a path to directly go to
the UEFI Shell where the startup.nsh script is automatically executed.
The UEFI Shell is launched after connecting possible devices, but
before the reset that is meant to automatically make them visible.
The new PcdUefiShellDefaultBootEnable must be set to TRUE to enable
this behaviour. The Pcd is set to false by default.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Tested-by: Patrik Berglund <patrik.berglund@arm.com>
App goes through ID_AA64*_EL1 system registers and decode their values.
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
ArmCpuInfo uses those to read system registers and other parts of EDK2
may find them useful.
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
ArmCpuInfo needs to be able to read ID_AA64ISAR2_EL1 system register.
Older toolchains do not know it.
Same solution as one for QEMU:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg929586.html
Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
As the ASM_FUNC() macro performs a section switch, the preceding
.balign directive applies the alignment constraint to the current
location in the previous section. As the linker may not merge the
sections in-order, ArmReplaceLiveTranslationEntry() may be left
unaligned.
Replace the explicit invocation of .balign with the ASM_FUNC_ALIGN()
macro, which guarantees the alignment constraint is applied correctly.
To make sure related issues are reliably caught in the future, align the
end of the function before checking the total occupied size. This
ensures crossing a 0x200 boundary will cause a compilation error.
Signed-off-by: Marvin Häuser <mhaeuser@posteo.de>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
With the current ASM_FUNC() macro, there is no good way to declare an
alignment constraint for a function. As ASM_FUNC() switches sections,
declaring the constraint before the macro invocation applies it to the
current location in the previous section. Declaring the constraint after
the macro invocation lets the function label point to the location prior
to alignment. Depending on toolchain behaviour, this may cause the label
to point to alignment padding preceding the actual function definition.
To address these issues, introduce the ASM_FUNC_ALIGN() macro, which
declares the alignment constraint right before the function label.
Signed-off-by: Marvin Häuser <mhaeuser@posteo.de>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
__FUNCTION__ is a pre-standard extension that gcc and Visual C++ among
others support, while __func__ was standardized in C99.
Since it's more standard, replace __FUNCTION__ with __func__ throughout
ArmPkg.
Signed-off-by: Rebecca Cran <rebecca@bsdio.com>
Reviewed-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The ELF based toolchains use objcopy to create HII object files, which
contain only a single .hii section. This means no GNU note is inserted
that describes the object as compatible with BTI, even though the lack
of executable code in such an object makes the distinction irrelevant.
However, the linker will not add the note globally to the resulting ELF
executable, and this breaks BTI compatibility.
So let's insert a GNU BTI-compatible ELF note by hand when generating
such object files.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Oliver Smith-Denny <osd@smith-denny.com>
The GccLto helper library does not contain any code, as its only purpose
is to pull in other libraries that implement intrinsics to which the
linker's codegen pass may emit calls.
So mark it as BTI compatible, so that the linker does not complain about
unannotated objects.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Oliver Smith-Denny <osd@smith-denny.com>
When building with -mbranch-protection=bti, which affects the compiler
codegen only, ensure that the assembler based codegen is aligned with
this, by emitting the BTI C opcode at the start of each exported
function. While most exported functions are not in fact ever called
indirectly, whether or not this is the case is a property of the caller
so annotating every exported function is a reasonable default.
While at it, fix two occurrences in ArmPkg of exported functions that
did not use the ASM_FUNC() macro.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Oliver Smith-Denny <osd@smith-denny.com>
The BIOS Firmware Version in the SMBIOS Type 0 can be fetched from
the fixed PcdFirmwareVersionString or platform specific OemMiscLib.
In fact, the support from OemMiscLib comes into play when the firmware
version may be modified at boot time for extended information.
Therefore, the priority of getting the version from OemMiscLib should
be higher.
In case there is no modification in the OemMiscLib, we have to keep
HII string STR_MISC_BIOS_VERSION empty or 'Not Specified'
to indicate that the firmware version should be fetched from
the PcdFirmwareVersionString.
Signed-off-by: Tinh Nguyen <tinhnguyen@os.amperecomputing.com>
Reviewed-by: Rebecca Cran <rebecca@bsdio.com>
Reviewed-by: Oliver Smith-Denny <osd@smith-denny.com>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
To prepare for the enablement of booting EFI with the SCTLR.WXN control
enabled, which makes all writeable memory regions non-executable by
default, introduce a memory type that we will use to describe the flash
region that carries the SEC and PEIM modules that execute in place. Even
if these are implicitly read-only due to the ROM nature, they need to be
mapped with read-only attributes in the page tables to be able to
execute from them.
Also add the XP counterpart which will be used for all normal DRAM right
at the outset.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Non-secure memory is a distinction that only matters when executing code
in the secure world that reasons about the secure vs non-secure address
spaces. EDK2 was not designed for that, and the AArch64 version of the
MMU handling library already treats them as identical, so let's just
drop the ARM memory region types that mark memory as 'non-secure'
explicitly.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Expose the protocol introduced in v2.10 that permits the caller to
manage mapping permissions in the page tables.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
In preparation for introducing an implementation of the EFI memory
attributes protocol that is shared between ARM and AArch64, unify the
existing code that converts a page table descriptor into a
EFI_MEMORY_xx bitfield, so it can be called from the generic code.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Currently, the ARM MMU page table logic will break down any block entry
that overlaps with the region being mapped, even if the block entry in
question is using the same attributes as the new region.
This means that creating a non-executable mapping inside a region that
is already mapped non-executable at a coarser granularity may trigger a
call to AllocatePages (), which may recurse back into the page table
code to update the attributes on the newly allocated page tables.
Let's avoid this, by preserving the block entry if it already covers the
region being mapped with the correct attributes.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Implement support for read-protected memory by wiring it up to the
access flag in the page table descriptor. The resulting mapping is
implicitly non-writable and non-executable as well, but this is good
enough for implementing this attribute, as we never rely on write or
execute permissions without read permissions.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Currently, the MMU code that is supposed to clear the RO or XP
attributes from a region just clears both unconditionally. This
approximates the desired behavior to some extent, but it does mean that
setting the RO bit first on a code region, and then clearing the XP bit
results both RO and XP being cleared, and we end up with writable code,
and avoiding that is the point of all these protections.
Once we introduce RP support, this will only get worse, so let's fix
this up, by reshuffling the attribute update code to take the entry mask
from the caller, and use the mask to preserve other attributes when
clearing RO or XP.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Split the ARM permission fields in the short descriptors into an access
flag and AP[2:1] as per the recommendation in the ARM ARM. This makes
the access flag available separately, which allows us to implement
EFI_MEMORY_RP memory analogous to how it will be implemented for
AArch64.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Marcin Juszkiewicz
Leif Lindholm
Oliver Smith-Denny
Taylor Beebe
Kun Qin
Ard Biesheuvel
Sami Mujawar
Neil Jones
Pierre Gondois
Marvin Häuser
Rebecca Cran
Tinh Nguyen