BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3223
In the current design, memory protection is not available till CpuDxe
is loaded. To resolve this, introduce CpuArchLib to move the
CPU Architectural initialization to DxeCore.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Vitaly Cheptsov <vit9696@protonmail.com>
Signed-off-by: Marvin Häuser <mhaeuser@posteo.de>
In order to provide an example and test out dynamic stack
cookies in edk2, dynamic stack cookies are added to ArmVirtPkg
platforms using ArmVirt.dsc.inc, per maintainer request.
This does require moving the ArmVirt.dsc.inc to be included
before MdeLibs.dsc.inc as the BaseTools does not expand all
the Defines sections before processing.
Signed-off-by: Oliver Smith-Denny <osde@linux.microsoft.com>
When running the verbose builds on arm for development and testing
it is quite convenient to use virtio serial to interact with the
system because console and logging are separated then. Except that
the virtio serial driver spams the debug log in that case. Set the
debug level for the driver to zero to avoid that.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Now that ArmPkg/Drivers/ArmGic no longer carries a combination of
libraries and DXE drivers, rename the directory to the more idiomatic
ArmPkg/Drivers/ArmGicDxe
Continuous-integration-options: PatchCheck.ignore-multi-package
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Before retiring ArmGicArchLib entirely, convert the existing
implementation in ArmVirtPkg into a NULL class library and inject it
into ArmGicDxe on all ARM virtual platforms. This ensures that the PCDs
describing the GIC are intialized correctly before ArmGicDxe is
dispatched.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
This issue showed up when addressing
https://bugzilla.tianocore.org/show_bug.cgi?id=4829
in https://github.com/tianocore/edk2/pull/6087 .
Various OvmfPkg and ArmVirtPkg platforms include some residual NetworkPkg
components when compiled with -D NETWORK_ENABLE=0, even though they use
NetworkPkg includes intended to allow all NetworkPkg components to be
disabled on this flag.
For the OvmfPkg Intel platforms only, commit
d933ec115bdf9be1d8dfe6a818414a14973cc0d3 started
the change of not including these residual NetworkPkg
components, and commit
7f17a155640a2a9e1f7b0f3522628ee2c6f62624 completed it.
This commit rolls these changes out to the ArmVirtPkg platforms where
they make sense in the same way.
Signed-off-by: Mike Beaton <mjsbeaton@gmail.com>
The first step is to add an option to disable PXE loading. The patch is
divided into 3 parts. This part adds the NETWORK_PXE_BOOT_ENABLE flag
to the ArmVirtPkg module. At the current stage the flag is not functional.
Signed-off-by: Alexander Gryanko <xpahos@gmail.com>
The implementation of ArmMonitorLib that selects the conduit (SMC or
HVC) based on the PSCI FDT node is suitable for other VMMs as well, so
rename it more appropriately.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Start using the include file in the ArmVirtPkg package to manage
dynamic network-related PCDs. This change removes the manual addition
of `PcdIPv4PXESupport` and `PcdIPv6PXESupport` from the DSC file,
relying instead on the centralized include file introduced in
NetworkPkg.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Aleksandr Goncharov <chat@joursoir.net>
ARCH as set by "build" command is not really useful for
conditionals, so move AArch64-specific Pcd to
[PcdsFixedAtBuild.AARCH64].
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
https://bugzilla.tianocore.org/show_bug.cgi?id=4829
7f17a15 (2024/02/22)
"OvmfPkg: Shell*.inc: allow building without network support"
breaks building OVMF with `-D NETWORK_ENABLE=0`.
Before this commit we could build OVMF e.g. with the following
command in the OvmfPkg directory:
./build.sh -D NETWORK_ENABLE=0
After the commit the same command fails early with:
/home/user/OpenSource/edk2/OvmfPkg/OvmfPkgX64.dsc(15):
error F001: Pcd (gEfiNetworkPkgTokenSpaceGuid.PcdAllowHttpConnections)
defined in DSC is not declared in DEC files referenced in INF files in
FDF. Arch: ['X64']
This problem also applies in the ArmVirtPkg
platforms which are modified here, but is currently
masked by another issue, namely that these platforms
incorrectly still include some network packages when
most are disabled. (A fix for this was previously applied,
for OvmfPkg Intel platforms only, by
d933ec1 followed by
7f17a15 .)
This commit was created at the same time as the
commits resolving this issue in NetworkPkg and
OvmfPkg. It makes conditional the Pcd references
in ArmVirtPkg platforms which will become references to
undefined Pcds as and when the other issue mentioned
above is fixed.
Signed-off-by: Mike Beaton <mjsbeaton@gmail.com>
Move all BaseCryptLib resolutions for 32-bit ARM to MbedTls, which does
not require a softfloat library, which can therefore be dropped from
EDK2 entirely going forward.
Note that this implies no TLS networking for 32-bit ARM, as this code
has a direct dependency on OpenSSL, so move the TlsLib resolution to a
AARCH64-only section to force the build to fail early when attempting to
build 32-bit ARM targets with NETWORK_TLS_ENABLE set.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Switch to the new SEC driver based on PrePeiCore, but with a sane name.
The old one will be retired once all users have migrated, including many
in edk2-platforms.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Retire the special ResetSystemLib implementation for the PEI phase on
virtual platforms, which has been superseded by the generic version
combined with a PEI-compatible implementation of ArmMonitorLib.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
PcdArmArchTimerFreqInHz lets platforms override the architectural timer
frequency, but this really isn't supported in hardware by anything lower
than EL3. Setting it to 0 skips the override - but that is also the
default. So drop the explicit setting in ArmVirtPkg platforms in
preparation for deleting the Pcd completely.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
This reverts commit 865229bcc8939c7a69d525f2b0627ef1532d5bc6, and
restores the old state where dynamic PCDs are only used when TPM support
is configured.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
This reverts commit c98f7f75508912c086158892fc56b0c3b85abcf1, which is
no longer needed: the SMCCC conduit will be converted back to being
hardcoded, as PrePi based ArmVirtPkg build cannot support dynamic PCDs,
and falling back to patchable PCDs does not work either.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Whether SMCCC calls use HVC or SMC generally depends on the exception
level that the firmware executes at, but also on whether or not EL2 is
implemented.
This is almost always known at build time, which is why the default
ArmMonitorLib used to model this as a feature PCD. However, on QEMU,
things are not that simple.
However, SMCCC specifies that the conduit is the same as the one used
for PSCI calls (which has been retrofitted into SMCCC when it was
defined). Given that QEMU provides this information via the device tree,
let's use it to select the conduit, using a special ArmMonitorLib
implementation.
This also removes the need to set the associated PCD at runtime, given
that its updated value will no longer be used.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Enable the non-hardcode version of FdtNorFlashQemuLib in ArmVirtQemu.dsc
and ArmVirtQemuKernel.dsc, and it can work rightly after enabling it.
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=4770
Build-tested (with "ArmVirtQemu.dsc" and "ArmVirtQemuKernel.dsc").
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Chao Li <lichao@loongson.cn>
Co-authored-by: Xianglai Li <lixianglai@loongson.cn>
Tested-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Add the RngDxe driver to the build, backed by either RNDR or TRNG, one
of which is expected to be available in most cases:
- RNDR is implemented by the 'max' CPU that QEMU implements in TCG mode
- TRNG is implemented by the KVM hypervisor, which backs QEMU's 'host'
CPU
Other TCG modes (e.g., the 'cortex-a*' CPUs) implement neither, which
should prevent the RngDxe driver from dispatching entirely, resulting
in the same situation as before.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.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>
MdeLibs.inc sets default library class resolutions which are much more
general than the ones that might be specified in ArmVirt.dsc.inc. So the
latter should be included *after* MdeLibs.inc to ensure that its
definitions take precedence.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.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>
On ARM systems, whether SMC or HVC instructions need to be used to issue
monitor calls is typically dependent on the exception level, but there
are also cases where EL1 might use SMC instructions, so there is no hard
and fast rule.
For ArmVirtQemu, this does depend strictly on the exception level, so
set the default to HVC (for EL1 execution) and override it to SMC when
booted at EL2.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.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>
Currently, only TPM2 builds enable the PCD PEIM, which is a prerequisite
for being able to use dynamic PCDs already at the PEI stage. This
facility will be used for other reasons too so move those pieces out of
code block that are conditional on TPM2_ENABLE
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.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>
This patch adds Hash2DxeCrypto to ArmVirtPkg. The Hash2DxeCrypto is
used to provide the hashing protocol services.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Doug Flick [MSFT] <doug.edk2@gmail.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The optimization that enabled entry with MMU and caches enabled at EL1
removed the strict alignment requirement for XIP code (roughly, any code
that might execute with the MMU and caches off, which means SEC and PEI
phase modules but also *all* BASE libraries), on the basis that QEMU can
only run guest payloads at EL2 in TCG emulation, which used to ignore
alignment violations, and execution at EL1 would always occur with the
MMU enabled.
This assumption no longer holds: not only does QEMU now enforce strict
alignment for memory accesses with device semantics, there are also
cases where this code might execute at EL2 under virtualization (i.e.,
under NV2 nested virtualization) where the strict alignment is required
too.
The latter case could be optimized too, by enabling VHE and pretending
execution is occurring at EL1, which would allow the existing logic for
entry with the MMU enabled to be reused. However, this would leave
non-VHE CPUs behind.
So in summary, strict alignment needs to be enforced for any code that
may execute with the MMU off, so drop the override that sets the XIP
flags to the empty string.
Cc: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Tested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Moved the PlatformBootManagerLib to OvmfPkg and renamed to
PlatformBootManagerLibLight for easy use by other ARCH.
Build-tested only (with "ArmVirtQemu.dsc and OvmfPkgX64.dsc").
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=4663
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Lazlo Ersek <lersek@redhat.com>
Signed-off-by: Chao Li <lichao@loongson.cn>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Move the PcdTerminalTypeGuidBuffer and PcdUninstallMemAttrProtocol into
OvmfPkg so other ARCH can easily use it.
Build-tested only (with "ArmVirtQemu.dsc and OvmfPkgX64.dsc").
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=4584
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Chao Li <lichao@loongson.cn>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Moved PcdDeviceTreeInitialBaseAddress and PcdDeviceTreeAllocationPadding
to OvmfPkg for easier use by other architectures.
Build-tested only (with "ArmVirtQemu.dsc").
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=4584
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Chao Li <lichao@loongson.cn>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
FEAT_VHE, introduced in ARMv8.1, adds a virtual EL2 timer.
However, this library verifies that exactly 3 or 4 12-byte timer
interrupts are provided in input DT, ASSERTing when the new timer
is added.
Change the assert to >= 36.
Extend the current logic, also initializing PcdArmArchTimerHypVirtIntrNum
if 5 interrupts are provided.
Signed-off-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Tested-by: Peter Maydell <peter.maydell@linaro.org>
Bhyve supports providing ACPI tables by FwCfg. Therefore,
InstallQemuFwCfgTables should be moved to AcpiPlatformLib to reuse the
code. As first step, move PciEncoding into AcpiPlatformLib.
Signed-off-by: Corvin Köhne <corvink@FreeBSD.org>
Acked-by: Peter Grehan <grehan@freebsd.org>
The PEI flavor of the ArmMmuLib will install a HOB that exposes its
implementation of the special helper routine that is used to update live
entries, so that other instantiations of ArmMmuLib can invoke it. This
is needed to ensure that splitting page tables using break-before-make
(BBM) does not unmap the code that is performing the split.
However, the BASE variety of ArmMmuLib discovers the HOB and sets a
global pointer to refer to it, which is not possible in PEIMs, and so
all PEIMs must use the PEI variety of this library if one does.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4076
PlatformHasAcpiDtDxe is required by other architectures also.
Hence, it is moved to OvmfPkg. So, update the consumers of this
module with the new location.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
Reviewed-by: Andrei Warkentin <andrei.warkentin@intel.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Commit 789a72328553 reclassified the NOR flash region as EFI_MEMORY_WC
in the OS visible EFI memory map, and dropped the explicit aligned
CopyMem() implementation, in the assumption that EFI_MEMORY_WC will be
honored by the OS, and that the region will be mapped in a way that
tolerates misaligned accesseses. However, Linux today uses device
attributes for all EFI MMIO regions, in spite of the memory type
attributes, and so using misaligned accesses is never safe.
So instead, switch to the generic CopyMem() implementation entirely,
just like we already did for VariableRuntimeDxe.
Fixes: 789a72328553 ("OvmfPkg/VirtNorFlashDxe: use EFI_MEMORY_WC and drop AlignedCopyMem()")
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
The early ID map used by ArmVirtQemu uses ASID scoped non-global
mappings, as this allows us to switch to the permanent ID map seamlessly
without the need for explicit TLB maintenance.
However, this triggers a known erratum on ThunderX, which does not
tolerate non-global mappings that are executable at EL1, as this appears
to result in I-cache corruption. (Linux disables the KPTI based Meltdown
mitigation on ThunderX for the same reason)
So work around this, by detecting the CPU implementor and part number,
and proceeding without the early ID map if a ThunderX CPU is detected.
Note that this requires the C code to be built with strict alignment
again, as we may end up executing it with the MMU and caches off.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: dann frazier <dann.frazier@canonical.com>
Since CcProbeLib is not used in AcpiPlatformDxe, CcProbeLib can be removed
from ArmVirtQemu.dsc.
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Leif Lindholm <quic_llindhol@quicinc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Min Xu <min.m.xu@intel.com>
Rely on CcProbe() to identify when running on TDX so that ACPI tables
can be retrieved differently for Cloud Hypervisor. Instead of relying on
the PVH structure to find the RSDP pointer, the tables are individually
passed through the HOB.
Signed-off-by: Jiaqi Gao <jiaqi.gao@intel.com>
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
Reviewed-by: Min Xu <min.m.xu@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Switch to the virt specific NorFlashDxe driver implementation that was
added recently.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Sunil V L <sunilvl@ventanamicro.com>
Clang does not support undoing the effects of -mstrict-align by passing
the -mno-strict-align counterpart, so appending the latter to the
compiler's XIPFLAGS does not work. Instead, clear the flags entirely.
This also removes -mgeneral-regs-only, but this is fine - we can
tolerate SIMD codegen in PEIMs or BASE libraries as they run with the
MMU and caches enabled.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The TPM discovery code relies on a dynamic PCD to communicate the TPM
base address to other components. But no other code relies on dynamic
PCDs in the PEI phase so let's drop the PCD PEIM when TPM support is not
enabled.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Due to the way we inherited the formerly fixed PCDs to describe the
system memory base and size from ArmPlatformPkg, we ended up with a
MemoryInit PEIM that relies on dynamic PCDs to communicate the size of
system memory between the constructor of one of its library dependencies
and the core module. This is unnecessary, and forces us to incorporate
the PCD PEIM as well, for no good reason. So instead, let's use a HOB.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
