With the SerialDxe change in commit 4cf3f37c87 ("MdeModulePkg SerialDxe:
Process timeout consistently in SerialRead", 2017-07-18), setting
EFI_SERIAL_INPUT_BUFFER_EMPTY in the "Control" output parameter, in the
GetControl() SerialPortLib function, is no longer a "small optimization".
Namely, due to the SerialDxe change, the GetOneKeyFromSerial() call in
TerminalDxe's TerminalConInTimerHandler() can take very long if the input
queue is empty, even if GetOneKeyFromSerial()'s return value causes the
loop to be exited right after, in the first iteration.
This issue causes a boot hang in ArmVirtQemu: with the input queue empty,
TerminalConInTimerHandler() takes so long to return that, by the time it
returns, there's another execution queued already (due to the associated
timer event being signaled meanwhile). The boot process is stuck in the
timer event handler.
Therefore even the first GetOneKeyFromSerial() iteration must be prevented
in TerminalConInTimerHandler() if the input queue is empty, and that
requires implementing GetControl() for real.
Implement the SetAttributes(), SetControl() and GetControl() APIs (of
SerialPortExtLib origin) in FdtPL011SerialPortLib with calls to matching
PL011UartLib functions. This follows the example of
"ArmPlatformPkg/Library/PL011SerialPortLib" and also matches Star's
original idea under [1].
The patch can be considered a continuation of commit ad7f6bc2e1
("ArmVirtPkg: Use SerialDxe in MdeModulePkg instead of EmbeddedPkg",
2015-11-26), based on the mailing list threads [1] [2] [3].
[1] http://mid.mail-archive.com/1447752930-32880-12-git-send-email-star.zeng@intel.com
[2] http://mid.mail-archive.com/1448243067-1880-12-git-send-email-star.zeng@intel.com
[3] http://mid.mail-archive.com/b748580c-cb51-32c9-acf9-780841ef15da@redhat.com
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Heyi Guo <guoheyi@huawei.com>
Cc: Star Zeng <star.zeng@intel.com>
Originally-suggested-by: Star Zeng <star.zeng@intel.com>
Reported-by: Brijesh Singh <brijesh.singh@amd.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Brijesh Singh <brijesh.singh@amd.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
https://bugzilla.tianocore.org/show_bug.cgi?id=642
Add top level License.txt file with the BSD 2-Clause
License that is used by the majority of the EKD II open
source project content. Merge copyright statements
from the BSD 2-Clause License files in each package
directory and remove the duplication License.txt
file from package directories.
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Andrew Fish <afish@apple.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
https://bugzilla.tianocore.org/show_bug.cgi?id=629
Move Contributions.txt that contains the TianoCore
Contribution Agreement 1.0 to the root of the edk2
repository and remove the duplicate Contributions.txt
files from all packages.
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Andrew Fish <afish@apple.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Commit 7b1dc6c569 'ArmVirtPkg: switch to generic ResetSystemRuntimeDxe'
replaced all references in ArmVirtPkg to the deprecated ResetRuntimeDxe
from EmbeddedPkg with the well maintained generic alternative that lives
in MdeModulePkg.
However, as it turns out, the generic driver has a dependency on the
library class ReportStatusCodeLib, whose default resolution is an
implementation that is not safe for use at runtime, resulting in crashes
when trying to invoke it from the OS.
Since we have no use for status codes in any of the ArmVirtPkg
platforms, let's replace all resolutions with a common one to the NULL
implementation.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
For obscure reasons, ARM platforms use a different implementation of
the ResetSystem() runtime service call than other platforms. So let's
switch all ArmVirtPkg platforms to the generic version instead.
Given that all platforms use an implementation of EfiResetSystemLib [as
consumed by the ResetRuntimeDxe in EmbeddedPkg that we are replacing]
which is unlikely to be depended upon by out of tree platforms, let's
simply modify this library into an implementation of ResetSystemLib
instead [which is what the generic driver in MdeModulePkg consumes]
This does mean we need to update all clients at the same time, which
is why all changes are part of the same patch.
As before, warm reset and platform specific reset are mapped onto
cold reset (which is the only thing PSCI implements, at least the
version we depend on). The new library function EnterS3WithImmediateWake()
is left unimplemented, as permitted by the ResetSystemLib library class.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Remove the functions now provided by TimeBaseLib from
PL031RealTimeClockLib. Add TimeBaseLib resolution to ArmVirtPkg
in same commit to prevent breakage.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Include XenPlatformHasAcpiDtDxe and PlatformHasAcpiDtDxe in the 32-bit
builds too.
Please see https://bugzilla.tianocore.org/show_bug.cgi?id=524
why it is needed. With this patch my arm uefi VM boots.
Fixes: 3a2c1548fe
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Nerijus Baliūnas <nerijus@users.sourceforge.net>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
[lersek@redhat.com: move long subj to commit msg body, add short subj]
[lersek@redhat.com: add Fixes reference]
[lersek@redhat.com: keep ACPI DXE modules grouped in QEMU DSCs]
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Remove the library class resolution for ARM's BdsLib: no included
module actually depends on it, and it will be removed shortly.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Given the agreement on the edk2-devel regarding the fact that the
notion whether or not a 'platform has ACPI' is a universal one, move
the PlatformHasAcpi GUID to MdeModulePkg.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: "Zeng, Star" <star.zeng@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
The relocatable build of ArmVirtQemuKernel is designed to be executed
from RAM, and contains some scratch memory at the start of the image
to use as a stack very early on, and to preserve the DTB image received
from QEMU while it discovers and initializes memory.
It turns out that 8 KB is a bit on the small side here, especially when
executing with secure world emulation enabled, in which case there are
additional nodes present.
So increase the slack space to 32 KB.
While at it, remove a stale Xen reference that was copy/pasted when this
file was created.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Laszlo Ersek <lersek@redhat.com>
In some cases, (e.g., when running QEMU with TrustZone emulation), the
DT may contain DT nodes whose status is set to 'secure'. Similarly, the
status may be set to 'disabled' if the consumer of the DT image is
expected to treat it as if it weren't there.
So check whether a 'status' property is present, and if so, ignore the
node if the status is not 'okay'.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
In some cases, (e.g., when running QEMU with TrustZone emulation), the
DT may contain DT nodes whose status is set to 'secure'. Similarly, the
status may be set to 'disabled' if the consumer of the DT image is
expected to treat it as if it weren't there.
So check whether a 'status' property is present, and if so, ignore the
node if the status is not 'okay'.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
In some cases, (e.g., when running QEMU with TrustZone emulation), the
DT may contain memory nodes whose status is set to 'secure'. Similarly,
the status may be set to 'disabled' if the consumer of the DT image is
expected to treat it as if it weren't there.
So check whether a 'status' property is present, and if so, ignore the
node if the status is not 'okay'.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
In general, we should not present two separate (and inevitably different)
hardware descriptions to the OS, in the form of ACPI tables and a device
tree blob. For this reason, we recently added the logic to ArmVirtQemu to
only expose the ACPI 2.0 entry point if no DT binary is being passed, and
vice versa.
However, this is arguably a regression for those who relied on DT
descriptions being available, even if the former behavior can be
restored by passing the -no-acpi switch to QEMU.
So allow a secret handshake with the UEFI Shell, to set a variable that
will result in ACPI to be disabled on subsequent boots even if -no-acpi
was not passed on the QEMU command line.
setvar -nv -bs -guid 50bea1e5-a2c5-46e9-9b3a-59596516b00a ForceNoAcpi =01
To delete the variable and revert to the old situation, simply omit the
value after the =
setvar -nv -bs -guid 50bea1e5-a2c5-46e9-9b3a-59596516b00a ForceNoAcpi =
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Marc Zyngier <marc.zyngier@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
ArmCpuLib is never used anywhere, and is about to be removed. So remove
any references from our .DSC files.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
The build flag and the FeaturePCD have no effect any longer, remove them.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
This will let QEMU's "-no-acpi" option exclusively expose DT vs. ACPI to
the guest. Showing both is never needed (it is actually detrimental to the
adoption of standards, such as SBSA / SBBR).
* Without "-no-acpi", the firmware logs (from PlatformHasAcpiDtDxe)
> Found FwCfg @ 0x9020008/0x9020000
> Found FwCfg DMA @ 0x9020010
> InstallProtocolInterface: [EdkiiPlatformHasAcpi] 0
plus the usual messages. Later the guest kernel logs
> [ 0.000000] efi: SMBIOS 3.0=0x13bdb0000 ACPI 2.0=0x138440000
> MEMATTR=0x13a675018
before it lists the ACPI tables one by one.
In addition, in the guest, the "/sys/firmware/devicetree/*" shell pattern
matches no files, while the "/sys/firmware/acpi/tables/*" pattern matches
the ACPI tables.
* With "-no-acpi", the firmware logs:
> PlatformHasAcpiDtDxe | Found FwCfg @ 0x9020008/0x9020000
> PlatformHasAcpiDtDxe | Found FwCfg DMA @ 0x9020010
> PlatformHasAcpiDtDxe | InstallProtocolInterface:
> PlatformHasAcpiDtDxe | [EdkiiPlatformHasDeviceTree] 0
> FdtClientDxe | OnPlatformHasDeviceTree: exposing DTB @
> FdtClientDxe | 0x13FFBF000 to OS
> ...
> DXE_CORE | Driver [AcpiTableDxe] was discovered but not
> DXE_CORE | loaded!!
> DXE_CORE | Driver [QemuFwCfgAcpiPlatform] was discovered but
> DXE_CORE | not loaded!!
> ...
> RamDiskDxe | RamDiskAcpiCheck: Cannot locate the EFI ACPI
> RamDiskDxe | Table Protocol, unable to publish RAM disks to
> RamDiskDxe | NFIT.
(BootGraphicsResourceTableDxe's ReadyToBoot callback --
InstallBootGraphicsResourceTable() -- handles the lack of
EFI_ACPI_TABLE_PROTOCOL silently.) Later the guest kernel logs
> [ 0.000000] efi: SMBIOS 3.0=0x13bdb0000 MEMATTR=0x138caa018
In addition, in the guest, the "/sys/firmware/devicetree/*" shell pattern
matches the directory "/sys/firmware/devicetree/base", which contains a
large number of DT nodes, while the "/sys/firmware/acpi/tables/*" pattern
matches no files.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Ref: https://bugzilla.redhat.com/show_bug.cgi?id=1430262
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Replace the dependency on PcdPureAcpiBoot with a Platform Has Device Tree
notification callback. Move the sysconfig table installation from the
entry point function to the callback.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
In this patch, the ACPI protocol / driver chain is enabled dynamically,
when appropriate. This is being done in one larger patch, because
ArmVirt.dsc.inc, where AcpiTableDxe is built, is used by all the platform
DSCs.
No change in behavior should be observable after this patch on any
ArmVirtPkg platform.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
This driver produces the EDKII Platform Has ACPI and Platform Has Device
Tree protocols, exactly matching the current ACPI / DT exposure on Xen,
according to ARM vs. AARCH64. At this point it differs from the QEMU
driver PlatformHasAcpiDtDxe in that this one always installs the DT.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
This driver produces the EDKII Platform Has ACPI and Platform Has Device
Tree protocols, exactly matching the current ACPI / DT exposure on QEMU,
according to ARM vs. AARCH64, and (in the latter case) to PcdPureAcpiBoot.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
This reverts commit 18f6d4df9e.
We realized that DXE drivers that are independent of AcpiPlatformDxe (that
is, independent of QEMU's ACPI generation), such as RamDiskDxe and
BootGraphicsResourceTableDxe, may produce and/or manipulate ACPI tables,
at driver dispatch or even at Ready To Boot.
This makes it unsafe for us to check for ACPI presence in the UEFI system
config table in a Ready To Boot callback, in order to decide about
exposing the DT.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
This reverts commit 78c41ff519.
We realized that DXE drivers that are independent of AcpiPlatformDxe (that
is, independent of QEMU's ACPI generation), such as RamDiskDxe and
BootGraphicsResourceTableDxe, may produce and/or manipulate ACPI tables,
at driver dispatch or even at Ready To Boot.
This makes it unsafe for us to check for ACPI presence in the UEFI system
config table in a Ready To Boot callback, in order to decide about
exposing the DT.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Currently, the file GUID reference of the UEFI Shell app is indirected
via the PCD gEfiIntelFrameworkModulePkgTokenSpaceGuid.PcdShellFile,
which is set to a fixed value for our platforms.
So instead, use the new symbolic GUID added for this purpose, and drop
the reference to this PCD, and to the IntelFrameworkModulePkg package
entirely.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Instead of looking at the PCD gArmTokenSpaceGuid.PcdSystemMemoryBase
to decide which DT node covers the memory we are already using, query
the GCD memory space map, which is the authoritative source for this
kind of information
This fixes a problem observed by Michael on platforms where this PCD
is of the 'Patchable' type, which means updates to its value do not
propagate to other modules.
Reported-by: Michael Zimmermann <sigmaepsilon92@gmail.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Instead of invoking gDS->SetMemorySpaceAttributes to set the EFI_MEMORY_XP
attribute on newly added regions, which is guaranteed to fail if the same
attribute was not declared as a capability of the region when it as added,
invoke the CPU arch protocol directly to set the EFI_MEMORY_XP attribute
if our memory protection policy demands it.
Reported-by: Michael Zimmermann <sigmaepsilon92@gmail.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
At this point we're ready to retire QemuFwCfgS3Enabled() from the
QemuFwCfgLib class, together with its implementations in:
- ArmVirtPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c
- OvmfPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c
Extend all modules that call the function with a new QemuFwCfgS3Lib class
dependency. Thanks to the previously added library class, instances, and
class resolutions, we can do this switch now as tightly as possible.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=394
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
QemuFwCfgS3Enabled() in "ArmVirtPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c"
returns constant FALSE.
The same implementation is now available factored-out in
"OvmfPkg/Library/QemuFwCfgS3Lib/QemuFwCfgS3Base.c".
Resolve QemuFwCfgS3Lib to BaseQemuFwCfgS3LibNull.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=394
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Instead of having a build time switch to prevent the FDT configuration
table from being installed, make this behavior dependent on whether we
are passing ACPI tables to the OS. This is done by looking for the
ACPI 2.0 configuration table, and only installing the FDT one if the
ACPI one cannot be found.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Defer FDT configuration table installation until ReadyToBoot is signaled.
This allows any driver to make modifications in the mean time, and will
also allow us to defer the decision of whether to install it in the first
place to later on in the boot.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Disable the PL031 RTC DT node unconditionally rather than only when
the DT will be exposed to the OS. This allows us to defer the decision
whether to expose it to the OS to a later time without creating an
additional dependency on the FDT client code by the RTC driver.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Include DXE_CORE in the BuildOptions that are set to force 4 KB section
alignment for PE/COFF images in order to allow them to be mapped with
strict memory permissions.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Now that ARM has grown support for managing memory permissions in
ArmMmuLib, we can enable the non-executable DXE stack for all virt
platforms. Note that this includes the AARCH64 Xen platform as well.
Note that this is not [entirely] redundant: the non-executable stack
is configured before DxeCore is invoked. The image and memory protection
features configured during DXE only take affect when the CPU arch
protocol implementation is registered.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Like for AARCH64, enable PE/COFF image and NX memory protection for all
32-bit ARM virt platforms.
Note that this does not [yet] protect EfiLoaderData regions, due to
compatibility issues with GRUB.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
This sets the recently introduced PCD PcdDxeNxMemoryProtectionPolicy to
a value that protects all memory regions except code regions against
inadvertent execution.
Note that this does not [yet] protect EfiLoaderData regions, due to
compatibility issues with shim and GRUB.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Recent changes to ShellPkg require a resolution for UefiBootManagerLib
for all platforms in ArmVirtPkg. So move the resolution to the shared
include ArmVirt.dsc.inc.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Using DxeServices::SetMemorySpaceAttributes to set cacheability
attributes has the side effect of stripping permission attributes,
given that those are bits in the same bitfield, and so setting the
Attributes argument to EFI_MEMORY_WB implies not setting EFI_MEMORY_XP
or EFI_MEMORY_RO attributes.
In fact, the situation is even worse, given that the descriptor returned
by DxeServices::GetMemorySpaceDescriptor does not reflect the permission
attributes that may have been set by the preceding call to
DxeServices::AddMemorySpace if PcdDxeNxMemoryProtectionPolicy has been
configured to map EfiConventionalMemory with non-executable permissions.
Note that this applies equally to the non-executable stack and to PE/COFF
sections that may have been mapped with R-X or RW- permissions. This is
due to the ambiguity in the meaning of the EFI_MEMORY_RO/EFI_MEMORY_XP
attributes when used in the GCD memory map, i.e., between signifying
that an underlying RAM region has the controls to be configured as
read-only or non-executable, and signifying that the contents of a
certain UEFI memory region allow them to be mapped with certain
restricted permissions.
So let's check the policy in PcdDxeNxMemoryProtectionPolicy directly,
and set the EFI_MEMORY_XP attribute if appropriate for
EfiConventionalMemory regions.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
This removes the PCD PcdArmUncachedMemoryMask from ArmPkg, along with
any remaining references to it in various platform .DSC files. It is
no longer used now that we removed the virtual uncached pages protocol
and the associated DebugUncachedMemoryAllocationLib library instance.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
The only observeable effect of having PcdPerformanceLibraryPropertyMask
set to 1 is that a EfiReservedMemory region of 4 pages is allocated right
below the 4 GB mark. This region is out of bounds for the OS, which means
it is not even allowed to map it, to avoid speculative loads from it.
On Linux, this may prevent the kernel from using a 1 GB block mapping for
this region, and instead it has to carve up the block as follows:
0xffffffff80000000-0xffffffffbe000000 992M PMD CON BLK
0xffffffffbe000000-0xffffffffbfe00000 30M PMD BLK
0xffffffffbfe00000-0xffffffffbfff0000 1984K PTE CON
0xffffffffbfff0000-0xffffffffbfffc000 48K PTE
where it would otherwise use a single 1 GB mapping (*), i.e.,
0xffffffff80000000-0xffffffffc0000000 1G PGD
To clarify, the latter is a single 8 byte entry in the top level page
table, whereas in the former case, we have two additional levels of
paging, requiring two extra 4 KB pages (on a 4 KB pagesize kernel).
The real cost, however, is the TLB footprint, which goes up from a
single entry to a number between 90 and 1020, depending on whether
contiguous hints are honoured by the hardware.
So let's remove PcdPerformanceLibraryPropertyMask until we find a reason
why we need it.
(*) provided that no other allocations were deliberately located right
below the 4 GB mark, and that we are running with more than 3 GB of
memory, in which case most allocations will be over 4 GB, given EDK2's
default top-down allocation policy.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
The OpensslLibCrypto library instance (which does not contain libssl
functions) is sufficient for the Secure Boot feature. It would not be
sufficient for HTTPS booting (which requires TLS), but in ArmVirtPkg, we
don't even enable plaintext HTTP booting for the time being.
Ease security analysis by excluding libssl functionality from the
OpensslLib instance we use.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Tomas Hoger <thoger@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Enable the new DXE image protection for all image, i.e., FV images but
also external images that originate from disk or the network, such as
OS loaders.
This complements work that is underway on the arm64/Linux kernel side,
to emit the OS loader with 4 KB section alignment, and a suitable split
between code and data.
http://marc.info/?l=linux-arm-kernel&m=148655557227819
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
The debug implementation of the UncachedMemoryAllocationLib library
class relies on the creation of an uncached alias of a memory range,
while keeping the original cached mapping, but with read-only attributes
to trap inadvertent write accesses.
This is not a terribly good idea, given that the ARM architecture does
not allow mismatched attributes, and so creating them deliberately is
not something we should encourage by doing it in reference code.
So remove the library, and replace all references to it with a reference
to the non-debug version (unless the platform does not require a resolution
for it in the first place, in which case all UncachedMemoryAllocationLib
references can be removed altogether).
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
We are now sufficiently equipped to implement the new QemuFwCfgSkipBytes()
API.
The previous patch and this one enable ArmVirtPkg/QemuFwCfgLib to
overwrite part of a writeable fw_cfg file, which will be particularly
useful for the upcoming QEMU_LOADER_WRITE_POINTER command in
OvmfPkg/AcpiPlatformDxe.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=359
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
We use the "InternalQemuFwCfgReadBytes" static function pointer to
dispatch the reading of fw_cfg bytes between MMIO and DMA. This pointer is
initialized to MMIO, and we set it to DMA in the library constructor if
DMA is available.
Unlike the above, we write fw_cfg bytes only with MMIO at the moment.
Extend the write functionality so that it follows the read pattern:
- introduce the new function typedef WRITE_BYTES_FUNCTION,
- extract the current (MMIO-only) write internals from
QemuFwCfgWriteBytes() to MmioWriteBytes(),
- provide a DMA-based implementation in DmaWriteBytes() -- a thin wrapper
around DmaTransferBytes(),
- set the new static function pointer "InternalQemuFwCfgWriteBytes"
according to the DMA feature provided by QEMU,
- In QemuFwCfgWriteBytes(), call the best available method through
"InternalQemuFwCfgWriteBytes".
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=359
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
The DmaReadBytes() function that we currently use only for reading --
through the InternalQemuFwCfgReadBytes function pointer, in case the DMA
interface is available -- is suitable with minimal changes for two more
operations provided by the DMA interface, WRITE and SKIP. Expose the
Control parameter in the function prototype, rename the function to
DmaTransferBytes(), and rebase DmaReadBytes() to it.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=359
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
While debugging OS for ACPI BGRT support (especially on VMs),
it is very useful to have the EFI firmware to export the
ACPI BGRT table.
This patch tries to add this support in ArmVirtPkg.
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Bhupesh Sharma <bhsharma@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Laszlo Ersek
Michael D Kinney
Ard Biesheuvel
Leif Lindholm
Nerijus Baliūnas
Bhupesh Sharma