- Added the NOR Flash region to the Runtime UEFI Memory Mapped IO
- Caught the gEfiEventVirtualAddressChangeGuid event to fixup the NOR Flash pointers
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15438 6f19259b-4bc3-4df7-8a09-765794883524
- No allocation during Runtime mode (post ExitBootServices())
- Allocate all the persistent data into runtime space
- Do not access BootServices API during Runtime mode
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15437 6f19259b-4bc3-4df7-8a09-765794883524
- Removed PCD base address from the macro definition. The base address needs to be fixup when the driver runs in UEFI Runtime mode
- Added the PL031 controller memory region to the Runtime UEFI Memory Mapped IO
- Caught the gEfiEventVirtualAddressChangeGuid event to fixup the PL031 Base address
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15435 6f19259b-4bc3-4df7-8a09-765794883524
OVMF's SecMain is unique in the sense that it links against the following
two libraries *in combination*:
- IntelFrameworkModulePkg/Library/LzmaCustomDecompressLib/
LzmaCustomDecompressLib.inf
- MdePkg/Library/BaseExtractGuidedSectionLib/
BaseExtractGuidedSectionLib.inf
The ExtractGuidedSectionLib library class allows decompressor modules to
register themselves (keyed by GUID) with it, and it allows clients to
decompress file sections with a registered decompressor module that
matches the section's GUID.
BaseExtractGuidedSectionLib is a library instance (of type BASE) for this
library class. It has no constructor function.
LzmaCustomDecompressLib is a compatible decompressor module (of type
BASE). Its section type GUID is
gLzmaCustomDecompressGuid == EE4E5898-3914-4259-9D6E-DC7BD79403CF
When OVMF's SecMain module starts, the LzmaCustomDecompressLib constructor
function is executed, which registers its LZMA decompressor with the above
GUID, by calling into BaseExtractGuidedSectionLib:
LzmaDecompressLibConstructor() [GuidedSectionExtraction.c]
ExtractGuidedSectionRegisterHandlers() [BaseExtractGuidedSectionLib.c]
GetExtractGuidedSectionHandlerInfo()
PcdGet64 (PcdGuidedExtractHandlerTableAddress) -- NOTE THIS
Later, during a normal (non-S3) boot, SecMain utilizes this decompressor
to get information about, and to decompress, sections of the OVMF firmware
image:
SecCoreStartupWithStack() [OvmfPkg/Sec/SecMain.c]
SecStartupPhase2()
FindAndReportEntryPoints()
FindPeiCoreImageBase()
DecompressMemFvs()
ExtractGuidedSectionGetInfo() [BaseExtractGuidedSectionLib.c]
ExtractGuidedSectionDecode() [BaseExtractGuidedSectionLib.c]
Notably, only the extraction depends on full-config-boot; the registration
of LzmaCustomDecompressLib occurs unconditionally in the SecMain EFI
binary, triggered by the library constructor function.
This is where the bug happens. BaseExtractGuidedSectionLib maintains the
table of GUIDed decompressors (section handlers) at a fixed memory
location; selected by PcdGuidedExtractHandlerTableAddress (declared in
MdePkg.dec). The default value of this PCD is 0x1000000 (16 MB).
This causes SecMain to corrupt guest OS memory during S3, leading to
random crashes. Compare the following two memory dumps, the first taken
right before suspending, the second taken right after resuming a RHEL-7
guest:
crash> rd -8 -p 1000000 0x50
1000000: c0 00 08 00 02 00 00 00 00 00 00 00 00 00 00 00 ................
1000010: d0 33 0c 00 00 c9 ff ff c0 10 00 01 00 88 ff ff .3..............
1000020: 0a 6d 57 32 0f 00 00 00 38 00 00 01 00 88 ff ff .mW2....8.......
1000030: 00 00 00 00 00 00 00 00 73 69 67 6e 61 6c 6d 6f ........signalmo
1000040: 64 75 6c 65 2e 73 6f 00 00 00 00 00 00 00 00 00 dule.so.........
vs.
crash> rd -8 -p 1000000 0x50
1000000: 45 47 53 49 01 00 00 00 20 00 00 01 00 00 00 00 EGSI.... .......
1000010: 20 01 00 01 00 00 00 00 a0 01 00 01 00 00 00 00 ...............
1000020: 98 58 4e ee 14 39 59 42 9d 6e dc 7b d7 94 03 cf .XN..9YB.n.{....
1000030: 00 00 00 00 00 00 00 00 73 69 67 6e 61 6c 6d 6f ........signalmo
1000040: 64 75 6c 65 2e 73 6f 00 00 00 00 00 00 00 00 00 dule.so.........
The "EGSI" signature corresponds to EXTRACT_HANDLER_INFO_SIGNATURE
declared in
MdePkg/Library/BaseExtractGuidedSectionLib/BaseExtractGuidedSectionLib.c.
Additionally, the gLzmaCustomDecompressGuid (quoted above) is visible at
guest-phys offset 0x1000020.
Fix the problem as follows:
- Carve out 4KB from the 36KB gap that we currently have between
PcdOvmfLockBoxStorageBase + PcdOvmfLockBoxStorageSize == 8220 KB
and
PcdOvmfSecPeiTempRamBase == 8256 KB.
- Point PcdGuidedExtractHandlerTableAddress to 8220 KB (0x00807000).
- Cover the area with an EfiACPIMemoryNVS type memalloc HOB, if S3 is
supported and we're not currently resuming.
The 4KB size that we pick is an upper estimate for
BaseExtractGuidedSectionLib's internal storage size. The latter is
calculated as follows (see GetExtractGuidedSectionHandlerInfo()):
sizeof(EXTRACT_GUIDED_SECTION_HANDLER_INFO) + // 32
PcdMaximumGuidedExtractHandler * (
sizeof(GUID) + // 16
sizeof(EXTRACT_GUIDED_SECTION_DECODE_HANDLER) + // 8
sizeof(EXTRACT_GUIDED_SECTION_GET_INFO_HANDLER) // 8
)
OVMF sets PcdMaximumGuidedExtractHandler to 16 decimal (which is the
MdePkg default too), yielding 32 + 16 * (16 + 8 + 8) == 544 bytes.
Regarding the lifecycle of the new area:
(a) when and how it is initialized after first boot of the VM
The library linked into SecMain finds that the area lacks the signature.
It initializes the signature, plus the rest of the structure. This is
independent of S3 support.
Consumption of the area is also limited to SEC (but consumption does
depend on full-config-boot).
(b) how it is protected from memory allocations during DXE
It is not, in the general case; and we don't need to. Nothing else links
against BaseExtractGuidedSectionLib; it's OK if DXE overwrites the area.
(c) how it is protected from the OS
When S3 is enabled, we cover it with AcpiNVS in InitializeRamRegions().
When S3 is not supported, the range is not protected.
(d) how it is accessed on the S3 resume path
Examined by the library linked into SecMain. Registrations update the
table in-place (based on GUID matches).
(e) how it is accessed on the warm reset path
If S3 is enabled, then the OS won't damage the table (due to (c)), hence
see (d).
If S3 is unsupported, then the OS may or may not overwrite the
signature. (It likely will.) This is identical to the pre-patch status.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15433 6f19259b-4bc3-4df7-8a09-765794883524
The stack size in the entrypoint (ie: $ARCH/ModuleEntryPoint.S) is calculated such as
StackSize = PrimaryCoreStack + (core_count - 1) * SecondaryCoreStack
While we were declaring the stacksize into the stack hob as:
StackSize = PrimaryCoreStack + (cluster * 8) * SecondaryCoreStack
If the number of cluster (ie: PcdClusterCount) were not defined correctly then
the stack size declaration were not correct.
It could cause stack corruption if the allocator allocates memory in this range.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15431 6f19259b-4bc3-4df7-8a09-765794883524
The function was allocating a buffer for the read value from the UEFI Variable.
But it was returning the pointer of the default value when the variable was
not present.
It could cause error when the default value and the returned value were free
when these addresses were the same (double FreePool on the same address).
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15427 6f19259b-4bc3-4df7-8a09-765794883524
In many cases, the second node in /pci@i0cf8/XYZ@DD,FF node is enough
to match a UEFI device path; a typical cases is a NIC that is assigned
from the host to the guest. Add a catch-all case for PCI devices, and
reuse it for NICs since it works well for those too.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15422 6f19259b-4bc3-4df7-8a09-765794883524
VideoDxe is a UEFI_DRIVER, so it has by default a null instance
of PcdLib. It accesses two PCDs that are now dynamic
(gEfiMdeModulePkgTokenSpaceGuid.PcdVideoHorizontalResolution
and gEfiMdeModulePkgTokenSpaceGuid.PcdVideoVerticalResolution).
Similar to r15362 (OvmfPkg: non-null PcdLib instance for
GraphicsConsoleDxe, 2014-03-22), we need to specify a non-null
instance of PcdLib.
This patch unbreaks the CSM VideoDxe module for OvmfPkg.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15421 6f19259b-4bc3-4df7-8a09-765794883524
Recent qemu versions compose all ACPI tables on the host side, according
to the target hardware configuration, and make the tables available to any
guest firmware over fw_cfg.
See version compatibility information below.
The feature moves the burden of keeping ACPI tables up-to-date from boot
firmware to qemu (which is the source of hardware configuration anyway).
This patch adds client code for this feature. Benefits of the
qemu-provided ACPI tables include PCI hotplug for example.
Qemu provides the following three fw_cfg files:
- etc/acpi/rsdp
- etc/acpi/tables
- etc/table-loader
"etc/acpi/rsdp" and "etc/acpi/tables" are similar, they are only kept
separate because they have different allocation requirements in SeaBIOS.
Both of these fw_cfg files contain preformatted ACPI payload.
"etc/acpi/rsdp" contains only the RSDP table, while "etc/acpi/tables"
contains all other tables, concatenated.
The tables in these two fw_cfg files are filled in by qemu, but two kinds
of fields are left incomplete in each table: pointers to other tables, and
checksums (which depend on the pointers).
Qemu initializes each pointer with a relative offset into the fw_cfg file
that contains the pointed-to ACPI table. The final pointer values depend
on where the fw_cfg files, holding the pointed-to ACPI tables, will be
placed in memory by the guest. That is, the pointer fields need to be
"relocated" (incremented) by the base addresses of where "/etc/acpi/rsdp"
and "/etc/acpi/tables" will be placed in guest memory.
This is where the third file, "/etc/table-loader" comes in the picture. It
is a linker/loader script that has several command types:
One command type instructs the guest to download the other two files.
Another command type instructs the guest to increment ("absolutize") a
pointer field (having a relative initial value) in the pointing ACPI
table, present in some fw_cfg file, with the dynamic base address of the
same (or another) fw_cfg file, holding the pointed-to ACPI table.
The third command type instructs the guest to compute checksums over
ranges and to store them.
In edk2, EFI_ACPI_TABLE_PROTOCOL knows about table relationships -- it
handles linkage automatically when a table is installed. The protocol
takes care of checksumming too. RSDP is installed automatically. Hence we
only need to care about the "etc/acpi/tables" fw_cfg file, determining the
boundaries of each ACPI table inside it, and installing those tables.
Qemu compatibility information:
--------------+---------------------+-------------------------------------
qemu version | qemu machine type | effects of the patch
--------------+---------------------+-------------------------------------
up to 1.6.x | any pc-i440fx | None. OVMF's built-in ACPI tables
| | are used.
--------------+---------------------+-------------------------------------
any | up to pc-i440fx-1.6 | None. OVMF's built-in ACPI tables
| | are used.
--------------+---------------------+-------------------------------------
1.7.0 | pc-i440fx-1.7 | Potential guest OS crash, dependent
| (default for 1.7.0) | on guest RAM size.
| |
| | DO NOT RUN OVMF on the (1.7.0,
| | pc-i440fx-1.7) qemu / machine type
| | combination.
--------------+---------------------+-------------------------------------
1.7.1 | pc-i440fx-1.7 | OVMF downloads valid ACPI tables
| (default for 1.7.1) | from qemu and passes them to the
| | guest OS.
--------------+---------------------+-------------------------------------
2.0.0-rc0 | pc-i440fx-1.7 or | OVMF downloads valid ACPI tables
| later | from qemu and passes them to the
| | guest OS.
-------------+---------------------+-------------------------------------
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15420 6f19259b-4bc3-4df7-8a09-765794883524
The previous patch ensures that the LockBox is protected during DXE (but
the OS can still drop it) if S3 is unsupported or disabled. However, S3
related drivers not only save data in the lockbox, they allocate objects
with Reserved and AcpiNVS memory types too, which the OS can't (must not)
release. This is a waste when S3 is unsupported or disabled.
In OVMF a good "choke point" for these drivers is the entry point of
AcpiS3SaveDxe. The messages of the following commits are relevant to the
data and control flow:
- SVN r15290 (git commit 8f5ca05b)
- SVN r15305 (git commit 5a217a06)
- SVN r15306 (git commit d4ba06df)
Prevent AcpiS3SaveDxe from loading when S3 is unsupported or disabled.
This should keep away (most of the) dependent drivers too.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Matt Fleming <matt.fleming@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15419 6f19259b-4bc3-4df7-8a09-765794883524
If (mBootMode == BOOT_ON_S3_RESUME) -- that is, we are resuming --, then
the patch has no observable effect.
If (mBootMode != BOOT_ON_S3_RESUME && mS3Supported) -- that is, we are
booting or rebooting, and S3 is supported), then the patch has no
observable effect either.
If (mBootMode != BOOT_ON_S3_RESUME && !mS3Supported) -- that is, we are
booting or rebooting, and S3 is unsupported), then the patch effects the
following two fixes:
- The LockBox storage is reserved from DXE (but not the OS). Drivers in
DXE may save data in the LockBox regardless of S3 support, potentially
corrupting any overlapping allocations. Make sure there's no overlap.
- The LockBox storage is cleared. A LockBox inherited across a non-resume
reboot, populated with well-known GUIDs, breaks drivers that want to
save entries with those GUIDs.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Matt Fleming <matt.fleming@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15418 6f19259b-4bc3-4df7-8a09-765794883524
There are several instances of asserts which do not also handle
the error condition in Release builds.
Because these functions are called in different location of the
code and their parameters might change during the execution, it
is safer to handle the error.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15399 6f19259b-4bc3-4df7-8a09-765794883524
On CortexA5x the SMP bit is BIT6 of CPUECTLR_EL1 register.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15398 6f19259b-4bc3-4df7-8a09-765794883524
This register is A5x specific. It is the reason why the code moved from ArmLib
to ArmCpuLib/ArmCortexA5xLib.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15397 6f19259b-4bc3-4df7-8a09-765794883524
PartNum is the field of MIDR that returns the CPU name.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15395 6f19259b-4bc3-4df7-8a09-765794883524
The arguments passed to the semi-hosing backend should
be of 'native' size to match register widths.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Harry Liebel <Harry.Liebel@arm.com>
Reviewed-by: Olivier Martin <olivier.martin@arm.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@15389 6f19259b-4bc3-4df7-8a09-765794883524
Olivier Martin
Eric Dong
Laszlo Ersek
Chris Phillips
Dong Guo
Jaben Carsey
Paolo Bonzini
Ruiyu Ni
Star Zeng
Elvin Li
Jeff Fan
Gao, Liming
Harry Liebel