"OvmfPkg/SmbiosPlatformDxe" is structured somewhat differently from the
drivers duplicated and trimmed thus far in this series. The final QEMU and
Xen versions will share a relatively significant amount of code, therefore
duplicating the whole driver is less useful, even temporarily. Instead,
duplicate the INF file, in preparation for customizing the entry point
function.
Because ArmVirtXen doesn't actually include OvmfPkg/SmbiosPlatformDxe [*],
there is only one platform that's supposed to consume the new driver:
OvmfXen. Switch OvmfXen to the new driver at once.
[*] See commit 164cf40383 ("OvmfPkg: SmbiosPlatformDxe: restrict current
Xen code to IA32/X64", 2015-07-26).
This patch is best viewed with "git show --find-copies-harder".
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Julien Grall <julien@xen.org>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=2122
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20210526201446.12554-42-lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Update gEfiMdePkgTokenSpaceGuid.PcdFSBClock so it can have the correct
value when SecPeiDxeTimerLibCpu start to use it for the APIC timer.
Currently, nothing appear to use the value in PcdFSBClock before
XenPlatformPei had a chance to set it even though TimerLib is included
in modules run before XenPlatformPei.
XenPlatformPei doesn't use any of the functions that would use that
value. No other modules in the PEI phase seems to use the TimerLib
before PcdFSBClock is set. There are currently two other modules in
the PEI phase that needs the TimerLib:
- S3Resume2Pei, but only because LocalApicLib needs it, but nothing is
using the value from PcdFSBClock.
- CpuMpPei, but I believe it only runs after XenPlatformPei
Before the PEI phase, there's the SEC phase, and SecMain needs
TimerLib because of LocalApicLib. And it initialise the APIC timers
for the debug agent. But I don't think any of the DebugLib that
OvmfXen could use are actually using the *Delay functions in TimerLib,
and so would not use the value from PcdFSBClock which would be
uninitialised.
A simple runtime test showed that TimerLib doesn't use PcdFSBClock
value before it is set.
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=2490
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20210412133003.146438-8-anthony.perard@citrix.com>
[lersek@redhat.com: cast Freq to UINT32 for PcdSet32S(), not for ASSERT()]
In NOOPT and DEBUG builds, if "PcdMaximumLinkedListLength" is nonzero,
then several LIST_ENTRY *node* APIs in BaseLib compare the *full* list
length against the PCD.
This turns the time complexity of node-level APIs from constant to linear,
and that of full-list manipulations from linear to quadratic.
As an example, consider the EFI_SHELL_FILE_INFO list, which is a data
structure that's widely used in the UEFI shell. I randomly extracted 5000
files from "/usr/include" on my laptop, spanning 1095 subdirectories out
of 1538, and then ran "DIR -R" in the UEFI shell on this tree. These are
the wall-clock times:
PcdMaximumLinkedListLength PcdMaximumLinkedListLength
=1,000,000 =0
-------------------------- ---------------------------
FAT 4 min 31 s 18 s
virtio-fs 5 min 13 s 1 min 33 s
Checking list lengths against an arbitrary maximum (default: 1,000,000)
seems useless even in NOOPT and DEBUG builds, while the cost is
significant; so set the PCD to 0.
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: Ard Biesheuvel <ard.biesheuvel@arm.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Julien Grall <julien@xen.org>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Peter Grehan <grehan@freebsd.org>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: Rebecca Cran <rebecca@bsdio.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=3152
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Ard Biesheuvel <ard.biesheuvel@arm.com>
Message-Id: <20210113085453.10168-10-lersek@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
We faced a problem with passing through a PCI device with 64GB BAR to UEFI
guest. The BAR is expectedly programmed into 64-bit PCI aperture at 64G
address which pushes physical address space to 37 bits. That is above
36-bit width that OVMF exposes currently to a guest without tweaking
PcdPciMmio64Size knob.
The reverse calculation using this knob was inhereted from QEMU-KVM
platform code where it serves the purpose of finding max accessible
physical address without necessary trusting emulated CPUID physbits value
(that could be different from host physbits). On Xen we expect to use
CPUID policy to level the data correctly to prevent situations with guest
physbits > host physbits e.g. across migrations.
The next aspect raising concern - resource consumption for DXE IPL page
tables and time required to map the whole address space in case of using
CPUID bits directly. That could be mitigated by enabling support for 1G
pages in DXE IPL configuration. 1G pages are available on most CPUs
produced in the last 10 years and those without don't have many phys bits.
Remove all the redundant code now (including PcdPciMmio64.. handling
that's not used on Xen anyway) and grab physbits directly from CPUID that
should be what baremetal UEFI systems do.
Signed-off-by: Igor Druzhinin <igor.druzhinin@citrix.com>
Message-Id: <1610509335-23314-1-git-send-email-igor.druzhinin@citrix.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Julien Grall <julien@xen.org>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
[lersek@redhat.com: fix up authorship from groups.io-mangled From line]
[lersek@redhat.com: wrap commit message at 74 characters]
Commit 55ee36b0c4
("EmbeddedPkg/RealTimeClockRuntimeDxe: Use helper functions from TimeBaseLib")
added a TimeBaseLib dependency to RealTimeClockRuntimeDxe, which now breaks
build of OvmfXen.dsc.
Add a resolution for EmbeddedPkg/Library/TimeBaseLib/TimeBaseLib.inf.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@arm.com>
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: Julien Grall <julien@xen.org>
Signed-off-by: Leif Lindholm <leif@nuviainc.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
OVMF booting stops with the assert if built with Xcode on macOS:
Loading driver at 0x0001FAB8000 EntryPoint=0x0001FABF249 LinuxInitrdDynamicShellCommand.efi
InstallProtocolInterface: BC62157E-3E33-4FEC-9920-2D3B36D750DF 1F218398
ProtectUefiImageCommon - 0x1F218140
- 0x000000001FAB8000 - 0x0000000000008A60
ASSERT_EFI_ERROR (Status = Unsupported)
ASSERT LinuxInitrdDynamicShellCommand.c(378): !EFI_ERROR (Status)
The assert comes from InitializeHiiPackage() after an attempt to
retrieve HII package list from ImageHandle.
Xcode still doesn't support HII resource section and
LinuxInitrdDynamicShellCommand depends on it. Likewise 277a3958d9
("OvmfPkg: Don't include TftpDynamicCommand in XCODE5 tool chain"),
disable initrd command if built with Xcode toolchain
Fixes: ec41733cfd ("OvmfPkg: add the 'initrd' dynamic shell command")
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Andrew Fish <afish@apple.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Message-Id: <20200514134820.62047-1-r.bolshakov@yadro.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Introduce DEBUG_ON_HYPERVISOR_CONSOLE build flag to enable logging
debug output to the Xen console.
This will work with both Xen HVM guest and Xen PVH guest whereas the
default PlatformDebugLibIoPort works only in HVM when QEMU is present.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20200423095358.2518197-6-anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Remove support for DEBUG_ON_SERIAL_PORT because OvmfXen can't be build
with it due to a circular dependency:
DebugLib : BaseDebugLibSerialPort ->
SerialPortLib : XenConsoleSerialPortLib ->
XenHypercallLib : XenHypercallLib ->
DebugLib
Also, if that dependency is fixed, I think it would be harder to find
which console the debug is sent to when running an HVM guest. The xen
console isn't the serial console used by default. Furthermore,
XenHypercallLib isn't initialised early enough, so we would loose
debug output from the SEC phase and early PEI phase.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20200423095358.2518197-2-anthony.perard@citrix.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
The BaseResetSystemLib instance is not suitable for OS runtime, because
its ResetShutdown() implementation calls PciRead16 (OVMF_HOSTBRIDGE_DID).
On q35, this boils down to a memory-mapped config space access -- but we
never ask the OS to map MMCONFIG for runtime.
There are at least three alternatives to approach this:
(1) Investigate "MdePkg/Library/DxeRuntimePciExpressLib", which offers
some kind of runtime mapping for MMCONFIG.
(2) Consume PciCf8Lib directly, rather than PciLib, in ResetSystemLib.
Then we'll read OVMF_HOSTBRIDGE_DID from the config space with IO port
accesses on q35 too, not just on i440fx. IO ports don't depend on page
tables.
(3) In the lib constructor, cache "mAcpiPmBaseAddress" based on
"PcdOvmfHostBridgePciDevId" (which is set by PlatformPei). Then the
host bridge type will be known at runtime without PCI config space
accesses.
This patch follows approach (3), in order to mirror AcpiTimerLib.
Notes:
* This patch is best viewed with "git show --find-copies-harder -C43".
* PCDs are not usable in the DXE_CORE, as the PCD PPI is gone, and the PCD
protocol is not available yet. (The DXE_CORE does consume ResetSystemLib
in practice, when OVMF is built with -D SOURCE_DEBUG_ENABLE.)
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: Ard Biesheuvel <ard.biesheuvel@arm.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Julien Grall <julien@xen.org>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: Rebecca Cran <rebecca@bsdio.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=2675
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20200417153751.7110-7-lersek@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@arm.com>
Reviewed-by: Rebecca Cran <rebecca@bsdio.com>
[lersek@redhat.com: move third Note (with repro info) to BZ comment]
Commit 859b55443a ("OvmfPkg/PlatformBootManagerLib: switch to
QemuLoadImageLib") replaced a dependency on LoadLinuxLib with one on
QemuLoadImageLib in the PlatformBootManagerLib implementation that is
shared between all OVMF builds, without taking into account that even
the Xen targeted builds incorporate this code, which is only used to
load kernels passed via the QEMU command line.
Since this is dead code on Xen, we can satisfy the dependency using
the generic version of QemuLoadImageLib, which does not rely on
LoadLinuxLib, which we can therefore drop from OvmfXen.dsc.
Fixes: 859b55443a
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Currently some tests check the value of SOURCE_DEBUG_ENABLE, and some
tests check if it's defined or not. Additionally, in UefiPayloadPkg as
well as some other trees, we define it as FALSE in the .dsc file.
This patch changes all of the Ovmf platforms to explicitly define it as
FALSE by default, and changes all of the checks to test if the value is
TRUE.
Signed-off-by: Peter Jones <pjones@redhat.com>
Message-Id: <20190920184507.909884-1-pjones@redhat.com>
[lersek@redhat.com: drop Contributed-under line, per TianoCore BZ#1373]
[lersek@redhat.com: replace "!= TRUE" with more idiomatic "== FALSE"]
Cc: Andrew Fish <afish@apple.com>
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Julien Grall <julien.grall@arm.com>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Cc: Peter Jones <pjones@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Philippe Mathieu-Daude <philmd@redhat.com>
On a Xen PVH guest, none of the existing serial or console interface
works, so we add a new one, based on XenConsoleSerialPortLib, and
implemented via SerialDxe.
That is a simple console implementation that can work on both PVH
guest and HVM guests, even if it is rarely going to be used on HVM.
Have PlatformBootManagerLib look for the new console, when running as a
Xen guest.
Since we use VENDOR_UART_DEVICE_PATH, fix its description and coding
style.
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1689
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20190813113119.14804-32-anthony.perard@citrix.com>
The ACPI Timer isn't present in a PVH guest, but local APIC works on
both PVH and HVM.
Note that the use of SecPeiDxeTimerLibCpu might be an issue with a
driver of type DXE_RUNTIME_DRIVER. I've attempted to find out which of
the DXE_RUNTIME_DRIVER uses the TimerLib at runtime. I've done that by
replacing the TimerLib evaluation in
[LibraryClasses.common.DXE_RUNTIME_DRIVER] by a different one and
checking every module that uses it (with the --report-file=report
build option).
ResetSystemRuntimeDxe is calling the TimerLib API at runtime to do the
operation "EfiResetCold", so this may never complete if the OS have
disabled the Local APIC Timer.
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1689
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20190813113119.14804-10-anthony.perard@citrix.com>
Introduce XenPlatformPei, a copy of OvmfPkg/PlatformPei without some
of QEMU specific initialization, Xen does not support QemuFwCfg.
This new module will be adjusted to accommodate Xen PVH.
fw_cfg dependents that have been removed, which are dynamically skipped
when running PlatformPei on Xen:
- GetFirstNonAddress(): controlling the 64-bit PCI MMIO aperture via the
(experimental) "opt/ovmf/X-PciMmio64Mb" file
- GetFirstNonAddress(): honoring the hotplug DIMM area
("etc/reserved-memory-end") in the placement of the 64-bit PCI MMIO
aperture
- NoexecDxeInitialization() is removed, so PcdPropertiesTableEnable and
PcdSetNxForStack are left constant FALSE (not set dynamically from
fw_cfg "opt/ovmf/PcdXxxx")
- MaxCpuCountInitialization(), PublishPeiMemory(): the max CPU count is
not taken from the QemuFwCfgItemSmpCpuCount fw_cfg key;
PcdCpuMaxLogicalProcessorNumber is used intact and
PcdCpuApInitTimeOutInMicroSeconds is never changed or used.
- InitializeXenPlatform(), S3Verification(): S3 is assumed disabled (not
consulting "etc/system-states" via QemuFwCfgS3Enabled()).
- InstallFeatureControlCallback(): the feature control MSR is not set
from "etc/msr_feature_control"
(also removed FeatureControl.c as there is nothing been executed)
Also removed:
- SMRAM/TSEG-related low mem size adjusting (PcdSmmSmramRequire is
assumed FALSE) in PublishPeiMemory(),
- QemuInitializeRam() entirely,
Xen related changes:
- Have removed the module variable mXen, as it should be always true.
- Have the platform PEI initialization fails if Xen has not been
detected.
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1689
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20190813113119.14804-5-anthony.perard@citrix.com>
Introduce XenResetVector, a copy of OvmfPkg/ResetVector, with one
changes:
- SEC_DEFAULT_CR0: enable cache (bit 30 or CD set to 0)
Xen copies the OVMF code to RAM, there is no need to disable cache.
This new module will later be modified to add a new entry point, more
detail in a following commit "OvmfPkg/XenResetVector: Add new entry point
for Xen PVH"
Value FILE_GUID of XenResetVector have not changed compare to ResetVector
because it is a special value (gEfiFirmwareVolumeTopFileGuid).
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1689
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20190813113119.14804-4-anthony.perard@citrix.com>
OvmfXen is a copy of OvmfX64, removing VirtIO and some SMM.
This new platform will be changed to make it works on two types of Xen
guest: HVM and PVH.
Compare to OvmfX64, this patch:
- changed: PLATFORM_GUID, OUTPUT_DIRECTORY, FLASH_DEFINITION
- removed: VirtioLib class resolution
- removed: all UEFI_DRIVER modules for virtio devices
- removed: DXE_SMM_DRIVER and SMM_CORE lib class resolutions
- removed: DXE_SMM_DRIVER and SMM_CORE FDF rules
- removed: Everything related to SMM_REQUIRE==true
- removed: Everything related to SECURE_BOOT_ENABLE==true
- removed: Everything related to TPM2_ENABLE==true
- changed: PcdPciDisableBusEnumeration dynamic default flipped to TRUE
- changed: default FD_SIZE_IN_KB to 2M.
- reverted d272449d9e, "OvmfPkg: raise DXEFV size to 11 MB"
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1689
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20190813113119.14804-3-anthony.perard@citrix.com>
