mirror_edk2.git
4 years agoUefiCpuPkg: Add GUID gEdkiiCpuFeaturesSetDoneGuid
Jeff Fan [Tue, 7 Mar 2017 07:59:22 +0000 (15:59 +0800)]
UefiCpuPkg: Add GUID gEdkiiCpuFeaturesSetDoneGuid

GUID gEdkiiCpuFeaturesSetDoneGuid is used to indicate if CPU feature related
setting are set finished. For example, PCD PcdCpuFeaturesUserConfiguration.

On PEI phase, one gEdkiiCpuFeaturesSetDoneGuid PPI will be installed after
platform set CPU feature setting.
On DXE phase, one gEdkiiCpuFeaturesSetDoneGuid Protocol will be installed after
platform set CPU feature setting.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoUefiCpuPkg/UefiCpuPkg.dec: Add a set of CPU features PCDs
Jeff Fan [Tue, 7 Mar 2017 07:54:23 +0000 (15:54 +0800)]
UefiCpuPkg/UefiCpuPkg.dec: Add a set of CPU features PCDs

PcdCpuFeaturesSupport supports PcdsFixedAtBuild/PcdsPatchableInModule types and
used to add/remove CPU feature from firmware during build time.

PcdCpuFeaturesUserConfiguration supports all PCD types and used to configurate
CPU features by platforms.

PcdCpuFeaturesCapability supports PcdsDynamic PCD and used to indicate the CPU
features capability on processors.

PcdCpuFeaturesSetting supports PcdsDynamic PCD and used to indicate the current
CPU features setting on processors.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoUefiCpuPkg/Msr: Add CPUID signature check MACROs
Jeff Fan [Tue, 7 Mar 2017 03:14:35 +0000 (11:14 +0800)]
UefiCpuPkg/Msr: Add CPUID signature check MACROs

All model-specific MSRs are related to processor signatures that are defined in
each section in Chapter 35 Model-Specific-Registers (MSR), Intel(R) 64 and
IA-32 Architectures Software Developer's Manual, Volume 3, September 2016.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: Skip if AllocatedSize is 0
Jeff Fan [Tue, 7 Mar 2017 12:01:51 +0000 (20:01 +0800)]
UefiCpuPkg/PiSmmCpuDxeSmm: Skip if AllocatedSize is 0

Needn't to copy register table if AllocatedSize is 0.

v4:
  Fix potential uninitialized variable issue.

v5:
  Set DestinationRegisterTableList[Index].RegisterTableEntry before
  RegisterTableEntry is updated.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoUefiCpuPkg/CpuS3DataDxe: Consume the existing PcdCpuS3DataAddress
Jeff Fan [Tue, 7 Mar 2017 11:39:27 +0000 (19:39 +0800)]
UefiCpuPkg/CpuS3DataDxe: Consume the existing PcdCpuS3DataAddress

If PCD PcdCpuS3DataAddress is set before, CpuS3DataDxe should get RegisterTable
and PreSmmRegisterTable from existing PCD pointed buffer and needn't to allocate
new buffer for RegisterTable and PreSmmRegisterTable.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoUefiCpuPkg/AcpiCpuData: Update RegisterTableEntry type
Jeff Fan [Tue, 7 Mar 2017 06:32:28 +0000 (14:32 +0800)]
UefiCpuPkg/AcpiCpuData: Update RegisterTableEntry type

Current RegisterTableEntry filed in CPU_REGISTER_TABLE is one pointer to
CPU_REGISTER_TABLE_ENTRY. If CPU register table wants to be passed from 32bit
PEI to x64 DXE/SMM, x64 DXE/SMM cannot get the correct RegisterTableEntry.

This update is to update RegisterTableEntry type to EFI_PHYSICAL_ADDRESS and
make RegisterTableEntry is fixed length.

Cc: Feng Tian <feng.tian@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoMdeModulePkg/MemoryProtection: split protect and unprotect paths
Ard Biesheuvel [Tue, 21 Mar 2017 13:49:08 +0000 (13:49 +0000)]
MdeModulePkg/MemoryProtection: split protect and unprotect paths

Currently, the PE/COFF image memory protection code uses the same code
paths for protecting and unprotecting an image. This is strange, since
unprotecting an image involves a single call into the CPU arch protocol
to clear the permission attributes of the entire range, and there is no
need to parse the PE/COFF headers again.

So let's store the ImageRecord entries in a linked list, so we can find
it again at unprotect time, and simply clear the permissions.

Note that this fixes a DEBUG hang on an ASSERT() that occurs when the
PE/COFF image fails to load, which causes UnprotectUefiImage() to be
invoked before the image is fully loaded.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoArmVirtPkg/HighMemDxe: check new regions against GCD memory space map
Ard Biesheuvel [Tue, 21 Mar 2017 09:12:56 +0000 (09:12 +0000)]
ArmVirtPkg/HighMemDxe: check new regions against GCD memory space map

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>
4 years agoArmVirtPkg/HighMemDxe: use CPU arch protocol to apply memprotect policy
Ard Biesheuvel [Tue, 21 Mar 2017 08:47:23 +0000 (08:47 +0000)]
ArmVirtPkg/HighMemDxe: use CPU arch protocol to apply memprotect policy

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>
4 years agoMdeModulePkg/BootGraphicsResourceTableDxe: don't allocate below 4 GB
Ard Biesheuvel [Sun, 19 Mar 2017 17:19:48 +0000 (17:19 +0000)]
MdeModulePkg/BootGraphicsResourceTableDxe: don't allocate below 4 GB

The BGRT table has an 8 byte field for the memory address of the image
data, and yet the driver explicitly allocates below 4 GB. This results
in an ASSERT() on systems that do not have any memory below 4 GB to begin
with.

Since neither the PI, the UEFI or the ACPI spec contain any mention of
why this data should reside below 4 GB, replace the allocation call
with an ordinary AllocatePages() call.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/AcpiTableDxe: consider version mask when removing tables
Ard Biesheuvel [Mon, 20 Mar 2017 14:51:36 +0000 (14:51 +0000)]
MdeModulePkg/AcpiTableDxe: consider version mask when removing tables

Invocations of EFI_ACPI_TABLE_PROTOCOL::UninstallAcpiTable() may
result in a crash when the value of PcdAcpiExposedTableVersions does
not include EFI_ACPI_TABLE_VERSION_1_0B.

The reason is that EFI_ACPI_TABLE_PROTOCOL::InstallAcpiTable() will
only populate the Rsdt1/Rsdt3 pointers when EFI_ACPI_TABLE_VERSION_1_0B
is set, whereas EFI_ACPI_TABLE_PROTOCOL::UninstallAcpiTable() will
invoke PublishTables with EFI_ACPI_TABLE_VERSION_1_0B alawys set,
resulting in a NULL pointer dereference of the Rsdt1/Rsdt3 pointers.

So take PcdAcpiExposedTableVersions into account for UninstallAcpiTable
as well.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg CapsuleApp: Add -NR (no-reset) option support
Star Zeng [Fri, 17 Mar 2017 03:27:37 +0000 (11:27 +0800)]
MdeModulePkg CapsuleApp: Add -NR (no-reset) option support

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=388

Add -NR (no-reset) option support, once the option is specified,
no reset will be trigger for the capsule with flag
CAPSULE_FLAGS_PERSIST_ACROSS_RESET and no CAPSULE_FLAGS_INITIATE_RESET.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Xiaofeng Wang <winggundum82@163.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg/NvmExpressDxe: Memory leak fix in async code flow
Suman Prakash [Mon, 20 Mar 2017 08:34:55 +0000 (16:34 +0800)]
MdeModulePkg/NvmExpressDxe: Memory leak fix in async code flow

For async commands, the buffer allocated for Prp list is
not getting freed, which will cause memory leak for async
read write command. For example testing async command flow
with custom application to send multiple read write commands
were resulting in decrease of available memory page in memmap,
which eventually resulted in system hang. Hence freeing
AsyncRequest->MapData, AsyncRequest->MapMeta, AsyncRequest->MapPrpList and
AsyncRequest->PrpListHost when async command is completed.

Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Suman Prakash <suman.p@samsung.com.com>
Reviewed-by: Hao Wu <hao.a.wu@intel.com>
4 years agoMdeModulePkg/Core/Dxe: downgrade "CodeSegmentCount is 0" msg to DEBUG_WARN
Laszlo Ersek [Mon, 20 Mar 2017 10:24:23 +0000 (11:24 +0100)]
MdeModulePkg/Core/Dxe: downgrade "CodeSegmentCount is 0" msg to DEBUG_WARN

UEFI executables that consist of a single read+write+exec PE/COFF section
trigger this message, but such a binary layout isn't actually an error.
The image can be launched alright, only image protection cannot be applied
to it fully.

One example that elicits the message is (some) Linux kernels (with the EFI
stub of course).

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg/RamDiskDxe: fix C string literal catenation in info messages
Laszlo Ersek [Fri, 17 Mar 2017 18:57:09 +0000 (19:57 +0100)]
MdeModulePkg/RamDiskDxe: fix C string literal catenation in info messages

RamDiskDxe installs the RamDiskAcpiCheck() Ready To Boot callback
function. If EFI_ACPI_TABLE_PROTOCOL and/or EFI_ACPI_SDT_PROTOCOL are not
found, then informational messages are logged, and the RAM disks are not
published to the (nonexistent) NFIT table.

The logic is fine, but the info messages are not concatenated correctly
from multiple string literals -- the second parts are passed as (unused)
arguments to DEBUG(). Fix the typos.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Hao Wu <hao.a.wu@intel.com>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
Reviewed-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoEmulatorPkg: Add SortLib/UefiBootManagerLib
Dandan Bi [Mon, 20 Mar 2017 06:03:55 +0000 (14:03 +0800)]
EmulatorPkg: Add SortLib/UefiBootManagerLib

ConPlatformDxe uses the UefiBootManagerLib,
and UefiBootManagerLib consumes SortLib.

Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoDuetPkg: Add SortLib/UefiBootManagerLib
Dandan Bi [Mon, 20 Mar 2017 06:02:28 +0000 (14:02 +0800)]
DuetPkg: Add SortLib/UefiBootManagerLib

ConPlatformDxe uses the UefiBootManagerLib,
and UefiBootManagerLib consumes SortLib.

Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoMdeModulePkg/DxeCapsuleLibFmp: Fix build failure
Dandan Bi [Mon, 20 Mar 2017 01:26:44 +0000 (09:26 +0800)]
MdeModulePkg/DxeCapsuleLibFmp: Fix build failure

Move the definition of variable "mEsrtTable" and
"mIsVirtualAddrConverted" to DxeCapsuleLib.c.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: jiewen.yao@intel.com
4 years agoBaseTools: GenFds get the Size info for FV image in the FD region
Yonghong Zhu [Mon, 13 Mar 2017 07:11:38 +0000 (15:11 +0800)]
BaseTools: GenFds get the Size info for FV image in the FD region

When the FV size is specify in the FD region, Tool generate the FV file
may not use the correct size.

Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=387
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoBaseTools: Update some tool with shell=True
Yonghong Zhu [Mon, 13 Mar 2017 07:27:17 +0000 (15:27 +0800)]
BaseTools: Update some tool with shell=True

Pkcs7Sign, Rsa2048Sha256Sign and Rsa2048Sha256GenerateKeys doesn't work
on Linux. It needs to be changed with shell=True.

Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=423
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/DxeCore: deal with allocations spanning several memmap entries
Ard Biesheuvel [Thu, 16 Mar 2017 11:35:28 +0000 (11:35 +0000)]
MdeModulePkg/DxeCore: deal with allocations spanning several memmap entries

When attempting to perform page allocations using AllocateAddress, we
fail to check whether the entire region is free before splitting the
region. This may lead to memory being leaked further into the routine,
when it turns out that one of the memory map entries intersected by the
region is already occupied. In this case, prior conversions are not rolled
back.

For instance, starting from this situation

0x000040000000-0x00004007ffff [ConventionalMemory ]
0x000040080000-0x00004009ffff [Boot Data          ]
0x0000400a0000-0x000047ffffff [ConventionalMemory ]

a failed EfiLoaderData allocation @ 0x40000000 that covers the BootData
region will fail, but leave the first part of the allocation converted,
so we end up with

0x000040000000-0x00004007ffff [Loader Data        ]
0x000040080000-0x00004009ffff [Boot Data          ]
0x0000400a0000-0x000047ffffff [ConventionalMemory ]

even though the AllocatePages() call returned an error.

So let's check beforehand that AllocateAddress allocations are covered
by a single memory map entry, so that it either succeeds or fails
completely, rather than leaking allocations.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoUefiCpuPkg/CpuDxe: Remove MSR_IA32_MISC_ENABLE check
Jeff Fan [Wed, 15 Mar 2017 01:47:52 +0000 (09:47 +0800)]
UefiCpuPkg/CpuDxe: Remove MSR_IA32_MISC_ENABLE check

The architectural MSR MSR_IA32_MISC_ENABLE is not supported by AMD processors.
Because reading CPUID.80000001H:EDK[20] is enough to check if XD feature is
supported or not, we just remove checking MSR_IA32_MISC_ENABLE(0x1A0).

Cc: Anthony PERARD <anthony.perard@citrix.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Tested-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoMdeModulePkg/AcpiTableDxe: Not make FADT.{DSDT,X_DSDT} mutual exclusion
Star Zeng [Thu, 16 Mar 2017 07:26:03 +0000 (15:26 +0800)]
MdeModulePkg/AcpiTableDxe: Not make FADT.{DSDT,X_DSDT} mutual exclusion

198a46d768fb90d2f9b16e26451b4814e7469eaf improved the DSDT and X_DSDT
fields mutual exclusion by checking FADT revision, but that breaks
some OS that has assumption to only consume X_DSDT field even the
DSDT address is < 4G.

To have better compatibility, this patch is to update the code to not
make FADT.{DSDT,X_DSDT} mutual exclusion, but always set both DSDT and
X_DSDT fields in the FADT when the DSDT address is < 4G.

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Jeff Fan <jeff.fan@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Jeff Fan <jeff.fan@intel.com>
4 years agoMdeModulePkg/LogoDxe: Add missing dependency gEfiHiiImageExProtocolGuid
Ruiyu Ni [Thu, 16 Mar 2017 06:05:04 +0000 (14:05 +0800)]
MdeModulePkg/LogoDxe: Add missing dependency gEfiHiiImageExProtocolGuid

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
4 years agoNt32Pkg: Use Shell source code directly
Ruiyu Ni [Thu, 16 Mar 2017 05:57:11 +0000 (13:57 +0800)]
Nt32Pkg: Use Shell source code directly

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoShellPkg/HandleParsingLib: Correct format specifier for LoadedImage
Jeff Westfahl [Tue, 14 Mar 2017 21:02:01 +0000 (05:02 +0800)]
ShellPkg/HandleParsingLib: Correct format specifier for LoadedImage

The format specifier for the LoadOptions field of the LoadedImage protocol
is "%s". However, the data in LoadOptions is often generic binary data. A
format specifier of "%x" is more appropriate for this field.

Using "dh -v" with format specifier "%s" on BIOS images based on EDK II
source before commit 891d844 can cause a crash.

Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Jaben Carsey <jaben.carsey@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Westfahl <jeff.westfahl@ni.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoMdeModulePkg/ConPlatform: Support GOP created as PCI's grandson
Ruiyu Ni [Thu, 16 Mar 2017 05:14:25 +0000 (13:14 +0800)]
MdeModulePkg/ConPlatform: Support GOP created as PCI's grandson

The original logic assumes GOP hande is son of PCI handle but it
is not always true.
Below wordings are from UEFI Spec:
If a graphics device supports multiple frame buffers, then
handles for the frame buffers must be created first, and then the
handles for the video output devices can be created as children of
the frame buffer handles.

So the GOP handle could be grandson of the PCI handle.

EfiBootManagerGetGopDevicePath(VideoController) is used to fix
this bug.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
4 years agoNetworkPkg: Fix service binding issue in TCP dxe.
Zhang Lubo [Thu, 16 Mar 2017 09:52:51 +0000 (17:52 +0800)]
NetworkPkg: Fix service binding issue in TCP dxe.

v2: Handle error case in SockCreateChild and fix typo issue

when we destroy the socket Sock and its associated
protocol control block, we need to first close the
parent protocol, then remove the protocol from childHandle
and last to free any data structures that allocated in
CreateChild. But currently, we free the socket data (Socket ConfigureState)
before removing the protocol form  the childhandle. So if the up layer
perform the driverbing stop to abort tcp session and send the tcp reset
packet, it will failed.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Zhang Lubo <lubo.zhang@intel.com>
Cc: Wu Jiaxin <jiaxin.wu@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
4 years agoMdeModulePkg: Fix service binding issue in TCP4 and Ip4 dxe.
Zhang Lubo [Thu, 16 Mar 2017 10:03:36 +0000 (18:03 +0800)]
MdeModulePkg: Fix service binding issue in TCP4 and Ip4 dxe.

v2: Handle error case in SockCreateChild and fix typo issue

when we destroy the socket Sock and its associated
protocol control block, we need to first close the
parent protocol, then remove the protocol from childHandle
and last to free any data structures that allocated in
CreateChild. But currently, we free the socket data
(Socket ConfigureState) before removing the protocol
form  the childhandle. So if the up layer want to send the
tcp reset packet in it's driver binding stop function, it will failed.

The IpInstance destroy state is redundant and may cause
ip transmit failed if up layer want to send ip packet.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Zhang Lubo <lubo.zhang@intel.com>
Cc: Wu Jiaxin <jiaxin.wu@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
4 years agoMdeModulePkg: Fix bug in DxeHttplib when converting port number.
Zhang Lubo [Thu, 16 Mar 2017 07:06:41 +0000 (15:06 +0800)]
MdeModulePkg: Fix bug in DxeHttplib when converting port number.

Http boot on X64 platform is faild, this is caused by the incorrect
type conversion when getting port number from Url.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Zhang Lubo <lubo.zhang@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Cc: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
4 years agoMdeModulePkg/Ip4Dxe: Add Ip/Netmask pair check for Ip4Config2
Jiaxin Wu [Thu, 16 Mar 2017 01:39:34 +0000 (09:39 +0800)]
MdeModulePkg/Ip4Dxe: Add Ip/Netmask pair check for Ip4Config2

v2:
* Add the check in Ip4Config2SetDefaultIf to avoid the DHCP configuration
case.

Ip4config2 doesn't check the validity of Ip/Netmask pair, which
leads to the invalid combination of Ip and Netmask setting.
This patch is to resolve the issue.

Cc: Hegde Nagaraj P <nagaraj-p.hegde@hpe.com>
Cc: Subramanian Sriram <sriram-s@hpe.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Sriram Subramanian <sriram-s@hpe.com>
Reviewed-by: Hegde Nagaraj P <nagaraj-p.hegde@hpe.com>
4 years agoMdeModulePkg/DxeCapsuleLibFmp: Fix build failure issues
Dandan Bi [Fri, 17 Mar 2017 01:28:11 +0000 (09:28 +0800)]
MdeModulePkg/DxeCapsuleLibFmp: Fix build failure issues

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoOvmfPkg/build.sh: Parse qemu version in more cases
Jordan Justen [Fri, 3 Feb 2017 21:10:39 +0000 (13:10 -0800)]
OvmfPkg/build.sh: Parse qemu version in more cases

Since qemu v2.7.0, the pkgversion appears to have a bug:

$ ./configure --target-list=x86_64-softmmu --with-pkgversion=foo

Results in this output:

$ x86_64-softmmu/qemu-system-x86_64 -version
QEMU emulator version 2.8.90(foo)
Copyright (c) 2003-2017 Fabrice Bellard and the QEMU Project developers

This appears to have been introduced in:

67a1de0d19 Makefile: Derive "PKGVERSION" from "git describe" by default

The previous commit (077de81a4c) produces this output:

$ x86_64-softmmu/qemu-system-x86_64 -version
QEMU emulator version 2.6.50 (foo), Copyright (c) 2003-2008 Fabrice Bellard

Now the OvmfPkg/build.sh script uses grep with '-o' to return only the
matched text.

grep -E is also used with a simple regex to extract only the digits of
the version.

qemu-bug: https://bugs.launchpad.net/bugs/1673373

Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoIntelFsp2Pkg: Raise exception for invalid BSF option
Thomaiyar, Richard Marian [Wed, 15 Mar 2017 09:24:23 +0000 (17:24 +0800)]
IntelFsp2Pkg: Raise exception for invalid BSF option

Raise exception for invalid BSF option in GenCfgOpt.py

Cc: Maurice Ma <maurice.ma@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Richard Thomaiyar <richard.marian.thomaiyar@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg/CapsuleLib: Fix runtime issue
Jiewen Yao [Wed, 15 Mar 2017 05:12:33 +0000 (13:12 +0800)]
MdeModulePkg/CapsuleLib: Fix runtime issue

The previous patch has problem on handling capsule
request at runtime.

This patch fixed the issue.

Cc: Chao Zhang <chao.b.zhang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Chao Zhang <chao.b.zhang@intel.com>
4 years agoOvmfPkg/QemuFwCfgS3Lib: Fix VS tool chain build failure
Dandan Bi [Thu, 16 Mar 2017 01:25:52 +0000 (09:25 +0800)]
OvmfPkg/QemuFwCfgS3Lib: Fix VS tool chain build failure

Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoMdeModulePkg BmPerf: Handle "/" separator in debug path for GCC build
Star Zeng [Wed, 15 Mar 2017 05:39:27 +0000 (13:39 +0800)]
MdeModulePkg BmPerf: Handle "/" separator in debug path for GCC build

Cc: Liming Gao <liming.gao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoShellPkg UefiDpLib: Handle "/" separator in debug path for GCC build
Star Zeng [Wed, 15 Mar 2017 05:37:45 +0000 (13:37 +0800)]
ShellPkg UefiDpLib: Handle "/" separator in debug path for GCC build

Cc: Liming Gao <liming.gao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Jaben Carsey <jaben.carsey@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoPerformancePkg Dp_App: Handle "/" separator in debug path for GCC build
Star Zeng [Wed, 15 Mar 2017 05:36:14 +0000 (13:36 +0800)]
PerformancePkg Dp_App: Handle "/" separator in debug path for GCC build

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg PCD: Allow SkuId to be changed only once
Star Zeng [Tue, 28 Feb 2017 04:48:39 +0000 (12:48 +0800)]
MdeModulePkg PCD: Allow SkuId to be changed only once

Current PI spec has no clear description about whether the
SkuId could be changed multiple times or not during one boot.

If the SkuId could be changed multiple times during one boot,
different modules may get inconsistent PCD values.
And DynamicHii PCD maps to UEFI variable, once one DynamicHii
PCD(UEFI variable) is set for one SkuId, then the PCD value
will be always from UEFI variable but not PCD database, even
the SkuId is set to other value.

This patch is to update PCD drivers to allow SkuId to be
changed only once during one boot.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoArmPkg/UncachedMemoryAllocationLib: set XP bit via CPU arch protocol
Ard Biesheuvel [Tue, 14 Mar 2017 19:52:11 +0000 (19:52 +0000)]
ArmPkg/UncachedMemoryAllocationLib: set XP bit via CPU arch protocol

Commit e7b24ec9785d ("ArmPkg/UncachedMemoryAllocationLib: map uncached
allocations non-executable") adds code that manipulates the GCD memory
space attributes of a newly allocated uncached region without checking
whether this region expose these attributes in its capabilities mask.

Given that the intent is to remove executable permissions from the region,
this is a fairly pointless exercise to begin with, regardless of whether
it is correct or not. The reason is that RO/XP memory attributes in the
GCD memory space map or the UEFI memory map are completely disconnected
from the actual mapping permissions used in the page tables.

So instead, invoke the CPU arch protocol directly, and add the non-exec
attributes in the page tables directly.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Tested-by: Ryan Harkin <ryan.harkin@linaro.org>
4 years agoCryptoPkg:SmmCryptLib: Add real Pkcs5Pbkdf2.c.
Jiewen Yao [Sat, 7 Jan 2017 01:35:04 +0000 (17:35 -0800)]
CryptoPkg:SmmCryptLib: Add real Pkcs5Pbkdf2.c.

Cc: Qin Long <qin.long@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
4 years agoMdeModulePkg DxeCore: Remove unreferenced symbol for memory profile
Star Zeng [Tue, 14 Mar 2017 09:05:36 +0000 (17:05 +0800)]
MdeModulePkg DxeCore: Remove unreferenced symbol for memory profile

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=370

Use GLOBAL_REMOVE_IF_UNREFERENCED for some memory profile global variables,
then their symbols could be removed when memory profile is disabled.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg PiSmmCore: Remove unreferenced symbol for SMRAM profile
Star Zeng [Tue, 14 Mar 2017 08:51:31 +0000 (16:51 +0800)]
MdeModulePkg PiSmmCore: Remove unreferenced symbol for SMRAM profile

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=370

Use GLOBAL_REMOVE_IF_UNREFERENCED for some SMRAM profile global variables,
then their symbols could be removed when SMRAM profile is disabled.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoNetworkPkg: Fix potential bug if the iSCSI use dns protocol.
Zhang, Lubo [Thu, 9 Mar 2017 08:17:56 +0000 (16:17 +0800)]
NetworkPkg: Fix potential bug if the iSCSI use dns protocol.

Since we use the Attempt and index as the attempt variable name instead of
the MAC address plus index, we need to update this to check the whether
the Controller handle is configured to use DNS protocol

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Zhang Lubo <lubo.zhang@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Cc: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
4 years agoMdePkg/UefiDevicePathLib: Fix the wrong MAC address length
Jiaxin Wu [Fri, 10 Mar 2017 06:45:12 +0000 (14:45 +0800)]
MdePkg/UefiDevicePathLib: Fix the wrong MAC address length

Network interface type should be checked before the conversion between
text device path node and MAC device path. Otherwise, the MAC text string
can't be converted to the representation of a device node, which leads to
the series failure of network HII configuration(e.g. IP, VLAN, HTTP Boot
configuration in Network Device List).

Cc: Liming Gao <liming.gao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
4 years agoShellPkg/UefiShellBcfgCommandLib: Fix VS2012 build failure
Dandan Bi [Tue, 14 Mar 2017 01:34:59 +0000 (09:34 +0800)]
ShellPkg/UefiShellBcfgCommandLib: Fix VS2012 build failure

Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Jaben Carsey <jaben.carsey@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoMdePkg/HiiImageEx: Fix incorrect comments of functions
Dandan Bi [Fri, 10 Mar 2017 01:25:10 +0000 (09:25 +0800)]
MdePkg/HiiImageEx: Fix incorrect comments of functions

https://bugzilla.tianocore.org/show_bug.cgi?id=383

Cc: Liming Gao <liming.gao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoOvmfPkg/AcpiPlatformDxe: save fw_cfg boot script with QemuFwCfgS3Lib
Laszlo Ersek [Wed, 22 Feb 2017 23:22:17 +0000 (00:22 +0100)]
OvmfPkg/AcpiPlatformDxe: save fw_cfg boot script with QemuFwCfgS3Lib

Drop the explicit S3SaveState protocol and opcode management; instead,
create ACPI S3 Boot Script opcodes for the WRITE_POINTER commands with
QemuFwCfgS3Lib functions.

In this case, we have a dynamically allocated Context structure, hence the
patch demonstrates how the FW_CFG_BOOT_SCRIPT_CALLBACK_FUNCTION takes
ownership of Context.

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>
4 years agoOvmfPkg/SmmControl2Dxe: save fw_cfg boot script with QemuFwCfgS3Lib
Laszlo Ersek [Wed, 22 Feb 2017 21:48:00 +0000 (22:48 +0100)]
OvmfPkg/SmmControl2Dxe: save fw_cfg boot script with QemuFwCfgS3Lib

We cannot entirely eliminate the manual boot script building in this
driver, as it also programs lower-level chipset registers (SMI_EN,
GEN_PMCON_1) at S3 resume, not just registers exposed via fw_cfg.

We can nonetheless replace the manually built opcodes for the latter class
of registers with QemuFwCfgS3Lib function calls. We preserve the ordering
between the two sets of registers (low-level chipset first, fw_cfg
second).

This patch demonstrates that manual handling of S3SaveState protocol
installation can be combined with QemuFwCfgS3Lib, even without upsetting
the original order between boot script fragments. An S3SaveState notify
function running at TPL_CALLBACK can safely queue another S3SaveState
notify function at TPL_CALLBACK with QemuFwCfgS3CallWhenBootScriptReady().

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for DXE fw_cfg instance
Laszlo Ersek [Wed, 22 Feb 2017 16:46:43 +0000 (17:46 +0100)]
OvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for DXE fw_cfg instance

In the DXE fw_cfg instance:

- QemuFwCfgS3Enabled() queries S3 enablement via fw_cfg. This behavior is
  shared with the PEI fw_cfg instance, and the DXE fw_cfg instance already
  pulls in the function from "QemuFwCfgS3PeiDxe.c".

- If QemuFwCfgS3Enabled() returns TRUE, the client module is permitted to
  call QemuFwCfgS3CallWhenBootScriptReady().

  We provide a fully functional implementation for
  QemuFwCfgS3CallWhenBootScriptReady(). A protocol notify is installed at
  TPL_CALLBACK for EFI_S3_SAVE_STATE_PROTOCOL. If / once the protocol is
  available, the client module's Callback() function is called, which is
  expected to produce ACPI S3 Boot Script opcodes using the helper
  functions listed below. In QemuFwCfgS3CallWhenBootScriptReady(), we also
  allocate a reserved memory buffer, sized & typed by the client module,
  for the opcodes and (internally) the fw_cfg DMA operations to work upon,
  during S3 resume.

  This behavior is unique to the DXE fw_cfg instance. Thus, add the
  function to "QemuFwCfgS3Dxe.c".

- The QemuFwCfgS3ScriptWriteBytes(), QemuFwCfgS3ScriptReadBytes(),
  QemuFwCfgS3ScriptSkipBytes(), and QemuFwCfgS3ScriptCheckValue()
  functions are also implemented usefully, since the client module's
  Callback() function is expected to invoke them.

  Each of the first three functions produces MEM_WRITE, IO_WRITE, and
  MEM_POLL opcodes, to set up the DMA command in reserved memory, to start
  the DMA transfer, and to check the DMA result, respectively.

  The QemuFwCfgS3ScriptCheckValue() function produces a MEM_POLL opcode to
  validate an unsigned integer field in data that was read via
  QemuFwCfgS3ScriptReadBytes().

  This behavior is again unique to the DXE fw_cfg instance, so add the
  functions to "QemuFwCfgS3Dxe.c".

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for PEI fw_cfg instance
Laszlo Ersek [Wed, 22 Feb 2017 09:34:48 +0000 (10:34 +0100)]
OvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for PEI fw_cfg instance

In the PEI fw_cfg instance:

- QemuFwCfgS3Enabled() queries S3 enablement via fw_cfg. This behavior is
  shared with the DXE fw_cfg instance, and the PEI fw_cfg instance already
  pulls in the function from "QemuFwCfgS3PeiDxe.c".

- If QemuFwCfgS3Enabled() returns TRUE, the client module is permitted to
  call QemuFwCfgS3CallWhenBootScriptReady(). However, in the PEI phase we
  have no support for capturing ACPI S3 Boot Script opcodes, hence we
  return RETURN_UNSUPPORTED unconditionally. This behavior is unique to
  the PEI fw_cfg instance, so add the function to "QemuFwCfgS3Pei.c".

- Consequently, the QemuFwCfgS3ScriptWriteBytes(),
  QemuFwCfgS3ScriptReadBytes(), QemuFwCfgS3ScriptSkipBytes(), and
  QemuFwCfgS3ScriptCheckValue() functions must never be called. (They
  could only be called from the client module's callback, but
  QemuFwCfgS3CallWhenBootScriptReady() will never install such callback in
  the PEI fw_cfg instance -- see above.)

  This behavior is not unique to the PEI fw_cfg instance (it is shared
  with the Base Null instance), so pull in these functions from
  "QemuFwCfgS3BasePei.c".

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for Base Null instance
Laszlo Ersek [Wed, 22 Feb 2017 09:34:48 +0000 (10:34 +0100)]
OvmfPkg/QemuFwCfgS3Lib: implement opcode APIs for Base Null instance

In the Base Null instance:

- QemuFwCfgS3Enabled() returns constant FALSE. This is unique to the Base
  Null instance, and the function is already present in
  "QemuFwCfgS3Base.c".

- The QemuFwCfgS3CallWhenBootScriptReady() function must never be called
  (according to the documentation, given the above). This is also unique
  to the Base Null instance, so implement the function in
  "QemuFwCfgS3Base.c".

- Consequently, the QemuFwCfgS3ScriptWriteBytes(),
  QemuFwCfgS3ScriptReadBytes(), QemuFwCfgS3ScriptSkipBytes(), and
  QemuFwCfgS3ScriptCheckValue() functions must never be called either.
  This behavior is not unique to the Base Null instance (it will be shared
  with the PEI fw_cfg instance), so add these functions to
  "QemuFwCfgS3BasePei.c".

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: add boot script opcode generation APIs to libclass
Laszlo Ersek [Wed, 22 Feb 2017 05:58:55 +0000 (06:58 +0100)]
OvmfPkg/QemuFwCfgS3Lib: add boot script opcode generation APIs to libclass

Introduce the following APIs:

- QemuFwCfgS3CallWhenBootScriptReady(): central function that registers a
  callback function, with a context parameter, for when ACPI S3 Boot
  Script opcodes can be produced. This function also allocates reserved
  memory for the opcodes to operate upon.

  The client module is supposed to produce the boot script fragment in the
  callback function.

- QemuFwCfgS3ScriptWriteBytes(), QemuFwCfgS3ScriptReadBytes(),
  QemuFwCfgS3ScriptSkipBytes(), QemuFwCfgS3ScriptCheckValue(): helper
  functions, available only to the above callback function, for composing
  the boot script fragment. QemuFwCfgS3ScriptSkipBytes() can double as a
  plain "select" whenever necessary.

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>
4 years agoArmVirtPkg, OvmfPkg: retire QemuFwCfgS3Enabled() from QemuFwCfgLib
Laszlo Ersek [Wed, 22 Feb 2017 02:47:18 +0000 (03:47 +0100)]
ArmVirtPkg, OvmfPkg: retire QemuFwCfgS3Enabled() from QemuFwCfgLib

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>
4 years agoOvmfPkg: resolve QemuFwCfgS3Lib
Laszlo Ersek [Wed, 22 Feb 2017 02:11:07 +0000 (03:11 +0100)]
OvmfPkg: resolve QemuFwCfgS3Lib

QemuFwCfgS3Enabled() in "OvmfPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c"
queries the "etc/system-states" fw_cfg file.

The same implementation is now available factored-out in
"OvmfPkg/Library/QemuFwCfgS3Lib/QemuFwCfgS3PeiDxe.c". It is available to
PEIMs through the PeiQemuFwCfgS3LibFwCfg instance, and to DXE_DRIVER and
DXE_RUNTIME_DRIVER modules through the DxeQemuFwCfgS3LibFwCfg instance.

Resolve QemuFwCfgS3Lib accordingly.

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>
4 years agoArmVirtPkg: resolve QemuFwCfgS3Lib
Laszlo Ersek [Wed, 22 Feb 2017 01:17:00 +0000 (02:17 +0100)]
ArmVirtPkg: resolve QemuFwCfgS3Lib

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: add initial PEI and DXE fw_cfg library instances
Laszlo Ersek [Wed, 22 Feb 2017 01:40:32 +0000 (02:40 +0100)]
OvmfPkg/QemuFwCfgS3Lib: add initial PEI and DXE fw_cfg library instances

This patch introduces PeiQemuFwCfgS3LibFwCfg, a limited functionality
QemuFwCfgS3Lib instance, for PEI phase modules.

The patch also introduces DxeQemuFwCfgS3LibFwCfg, a full functionality
QemuFwCfgS3Lib instance, for DXE_DRIVER and DXE_RUNTIME_DRIVER modules.

These library instances share the QemuFwCfgS3Enabled() function. The
function actually uses fw_cfg; the implementation is copied from
"OvmfPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c".

The library instances will diverge in the following patches.

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>
4 years agoOvmfPkg/QemuFwCfgS3Lib: add initial Base Null library instance
Laszlo Ersek [Wed, 22 Feb 2017 01:09:47 +0000 (02:09 +0100)]
OvmfPkg/QemuFwCfgS3Lib: add initial Base Null library instance

This library instance returns constant FALSE from QemuFwCfgS3Enabled(),
and all other library functions trigger assertion failures. It is suitable
for QEMU targets and machine types that never enable S3.

The QemuFwCfgS3Enabled() implementation is copied from
"ArmVirtPkg/Library/QemuFwCfgLib/QemuFwCfgLib.c". Stubs for further
QemuFwCfgS3Lib APIs (with assertion failures, see above) will be added
later.

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>
4 years agoOvmfPkg: introduce QemuFwCfgS3Lib class
Laszlo Ersek [Wed, 22 Feb 2017 00:59:41 +0000 (01:59 +0100)]
OvmfPkg: introduce QemuFwCfgS3Lib class

This library class will enable driver modules (a) to query whether S3
support was enabled on the QEMU command line, (b) to produce fw_cfg DMA
operations that are to be replayed at S3 resume time.

Declare the library class in OvmfPkg/OvmfPkg.dec, and add the library
class header under OvmfPkg/Include/Library/. At the moment, the only API
we expose is QemuFwCfgS3Enabled(), which we'll first migrate from
QemuFwCfgLib. Further interfaces will be added in later patches.

Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=394
Suggested-by: Jordan Justen <jordan.l.justen@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoEmbeddedPkg/PrePiLib: allocate code pages for DxeCore
Ard Biesheuvel [Tue, 14 Mar 2017 08:01:04 +0000 (08:01 +0000)]
EmbeddedPkg/PrePiLib: allocate code pages for DxeCore

The recently introduced memory protection features inadvertently broke
the boot on all PrePi platforms, because the changes to explicitly use
EfiBootServicesCode for loading the DxeCore PE/COFF image need to be
applied in a different way for PrePi. So add a simple helper function
that sets the type of an allocation to EfiBootServicesCode, and invoke
it to allocate the space for DxeCore.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Michael Zimmermann <sigmaepsilon92@gmail.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoBaseTools/UPT: Fix an issue in subst command
Hess Chen [Fri, 10 Mar 2017 00:25:21 +0000 (08:25 +0800)]
BaseTools/UPT: Fix an issue in subst command

UPT used to use "B:" as the temp directory which may cause conflict,
now it will choose a valid volume. Also UPT now accepts empty sections.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hess Chen <hesheng.chen@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoArmPkg: Fix modsi3.S compilation across toolchains.
Marvin Häuser [Sat, 11 Mar 2017 22:05:26 +0000 (22:05 +0000)]
ArmPkg: Fix modsi3.S compilation across toolchains.

modsi3.S references the symbol '__divsi3' by '___divsi3' which assumes
the prefix is always required and supported.  Use ASM_PFX() instead
to support all compilers.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Marvin Haeuser <Marvin.Haeuser@outlook.com>
4 years agoMdeModulePkg/SdMmc: Add break to avoid dead loop when polling OCR Reg
Feng Tian [Mon, 13 Mar 2017 03:20:41 +0000 (11:20 +0800)]
MdeModulePkg/SdMmc: Add break to avoid dead loop when polling OCR Reg

At worst case, OCR register may always not set BIT31. It will cause
original code enter to dead loop. Adding a break for such case.

Cc: Hao Wu <hao.a.wu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Feng Tian <feng.tian@intel.com>
Reviewed-by: Hao Wu <hao.a.wu@intel.com>
4 years agoMdeModulePkg/SmmCore: Add Context in SmiHandlerProfileUnregister.
Jiewen Yao [Fri, 10 Mar 2017 03:45:32 +0000 (11:45 +0800)]
MdeModulePkg/SmmCore: Add Context in SmiHandlerProfileUnregister.

The reason is that we observe that a platform may use same Handler
for different context.

In order to support Unregister such handler, we have to input
context information as well.

Cc: Jeff Fan <jeff.fan@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Bret Barkelew <Bret.Barkelew@microsoft.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoMdeModulePkg/SmiHandlerProfile: Add Context support in Unregister
Jiewen Yao [Fri, 10 Mar 2017 03:44:38 +0000 (11:44 +0800)]
MdeModulePkg/SmiHandlerProfile: Add Context support in Unregister

The reason is that we observe that a platform may use same Handler
for different context.

In order to support Unregister such handler, we have to input
context information as well.

Cc: Jeff Fan <jeff.fan@intel.com>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Bret Barkelew <Bret.Barkelew@microsoft.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoMdePkg/SmiHandlerProfile: Add Context support in Unregister
Jiewen Yao [Fri, 10 Mar 2017 03:42:53 +0000 (11:42 +0800)]
MdePkg/SmiHandlerProfile: Add Context support in Unregister

The reason is that we observe that a platform may use same Handler
for different context.

In order to support Unregister such handler, we have to input
context information as well.

Cc: Jeff Fan <jeff.fan@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Bret Barkelew <Bret.Barkelew@microsoft.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
4 years agoMdeModulePkg/Print2: Refine the comment for ASCII_VALUE_TO_STRING_S
Hao Wu [Tue, 7 Mar 2017 08:04:04 +0000 (16:04 +0800)]
MdeModulePkg/Print2: Refine the comment for ASCII_VALUE_TO_STRING_S

The commit refines the comment description for EFI_PRINT2S_PROTOCOL
service ASCII_VALUE_TO_STRING_S.

This service will not ASSERT when the input/output parameter 'Buffer' is
not aligned on a 16-bit boundary.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/PrintLib: Refine the comment for AsciiValueToStringS API
Hao Wu [Tue, 7 Mar 2017 08:02:27 +0000 (16:02 +0800)]
MdeModulePkg/PrintLib: Refine the comment for AsciiValueToStringS API

The commit refines the comment description for PrintLib API
AsciiValueToStringS.

This API will not ASSERT when the input/output parameter 'Buffer' is not
aligned on a 16-bit boundary.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdePkg/BasePrintLib: Refine the comment for AsciiValueToStringS API
Hao Wu [Tue, 7 Mar 2017 07:57:52 +0000 (15:57 +0800)]
MdePkg/BasePrintLib: Refine the comment for AsciiValueToStringS API

The commit refines the comment description for PrintLib API
AsciiValueToStringS.

This API will not ASSERT when the input/output parameter 'Buffer' is not
aligned on a 16-bit boundary.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/UefiBootManagerLib: Generate boot description for NVME
Ruiyu Ni [Fri, 10 Mar 2017 07:41:21 +0000 (15:41 +0800)]
MdeModulePkg/UefiBootManagerLib: Generate boot description for NVME

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
Reviewed-by: Sunny Wang <sunnywang@hpe.com>
4 years agoMdeModulePkg/CapsuleLib: Free the buffer returned by GetVariable2 API
Hao Wu [Fri, 10 Mar 2017 02:22:30 +0000 (10:22 +0800)]
MdeModulePkg/CapsuleLib: Free the buffer returned by GetVariable2 API

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg/CapsuleLib: Add lib destructors to handle unclosed events
Hao Wu [Fri, 10 Mar 2017 02:13:32 +0000 (10:13 +0800)]
MdeModulePkg/CapsuleLib: Add lib destructors to handle unclosed events

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdeModulePkg/PlatVarCleanupLib: Add lib destructor for unclosed event
Hao Wu [Fri, 10 Mar 2017 05:57:39 +0000 (13:57 +0800)]
MdeModulePkg/PlatVarCleanupLib: Add lib destructor for unclosed event

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoIntelFsp2WrapperPkg: Add Readme.md
Jiewen Yao [Thu, 9 Mar 2017 02:32:15 +0000 (10:32 +0800)]
IntelFsp2WrapperPkg: Add Readme.md

Cc: Maurice Ma <maurice.ma@intel.com>
Cc: Richard Thomaiyar <richard.marian.thomaiyar@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoIntelFsp2Pkg: Add Readme.md
Jiewen Yao [Thu, 9 Mar 2017 02:31:57 +0000 (10:31 +0800)]
IntelFsp2Pkg: Add Readme.md

Cc: Maurice Ma <maurice.ma@intel.com>
Cc: Richard Thomaiyar <richard.marian.thomaiyar@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoQuarkPlatformPkg: Add UefiBootManagerLib to fix Shell build failure
Ruiyu Ni [Fri, 10 Mar 2017 09:36:39 +0000 (17:36 +0800)]
QuarkPlatformPkg: Add UefiBootManagerLib to fix Shell build failure

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoMdePkg/Pci22.h: Remove deprecated macros
Ruiyu Ni [Mon, 6 Mar 2017 06:38:59 +0000 (14:38 +0800)]
MdePkg/Pci22.h: Remove deprecated macros

The following deprecated macros are removed.The removal
doesn't cause any build failure to existing packages.

#define DEVICE_ID_NOCARE    0xFFFF
#define PCI_BAR_OLD_ALIGN   0xFFFFFFFFFFFFFFFFULL
#define PCI_BAR_EVEN_ALIGN  0xFFFFFFFFFFFFFFFEULL
#define PCI_BAR_SQUAD_ALIGN 0xFFFFFFFFFFFFFFFDULL
#define PCI_BAR_DQUAD_ALIGN 0xFFFFFFFFFFFFFFFCULL
#define PCI_BAR_ALL         0xFF
#define PCI_ACPI_UNUSED     0
#define PCI_BAR_NOCHANGE    0

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoSignedCapsulePkg: Add Readme.md
Jiewen Yao [Thu, 9 Mar 2017 02:30:14 +0000 (10:30 +0800)]
SignedCapsulePkg: Add Readme.md

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoAdd SignedCapsule content.
Jiewen Yao [Thu, 9 Mar 2017 02:13:43 +0000 (10:13 +0800)]
Add SignedCapsule content.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoArmVirtPkg/FdtClientDxe: make DT table installation !ACPI dependent
Ard Biesheuvel [Thu, 9 Mar 2017 15:59:34 +0000 (16:59 +0100)]
ArmVirtPkg/FdtClientDxe: make DT table installation !ACPI dependent

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>
4 years agoArmVirtPkg/FdtClientDxe: install DT configuration table at ReadyToBoot
Ard Biesheuvel [Thu, 9 Mar 2017 15:45:07 +0000 (16:45 +0100)]
ArmVirtPkg/FdtClientDxe: install DT configuration table at ReadyToBoot

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>
4 years agoArmVirtPkg/ArmVirtPL031FdtClientLib: unconditionally disable DT node
Ard Biesheuvel [Thu, 9 Mar 2017 15:38:53 +0000 (16:38 +0100)]
ArmVirtPkg/ArmVirtPL031FdtClientLib: unconditionally disable DT node

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>
4 years agoArmVirtPkg/FdtClientDxe: supplement missing EFIAPI calling conv specifiers
Laszlo Ersek [Wed, 8 Mar 2017 19:05:06 +0000 (20:05 +0100)]
ArmVirtPkg/FdtClientDxe: supplement missing EFIAPI calling conv specifiers

Add missing EFIAPI calling convention specifiers to STATIC function
that are exposed via the FdtClientProtocol interface.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoMdeModulePkg/AcpiTableDxe: improve FADT.{DSDT,X_DSDT} mutual exclusion
Laszlo Ersek [Wed, 1 Feb 2017 14:33:57 +0000 (15:33 +0100)]
MdeModulePkg/AcpiTableDxe: improve FADT.{DSDT,X_DSDT} mutual exclusion

The ACPI specification, up to and including revision 5.1 Errata A, allows
the DSDT and X_DSDT fields to be both set in the FADT. (Obviously, this
only makes sense if the DSDT address is representable in 4 bytes.)

Starting with 5.1 Errata B, specifically for Mantis 1393
<https://mantis.uefi.org/mantis/view.php?id=1393>, the spec requires at
most one of DSDT and X_DSDT to be set to a nonzero value.

MdeModulePkg/AcpiTableDxe handles this mutual exclusion somewhat
inconsistently.

- If the caller of EFI_ACPI_TABLE_PROTOCOL.InstallAcpiTable() installs the
  tables in "DSDT, FADT" order, then we enforce the exclusion between the
  DSDT and X_DSDT fields:

  DSDT under 4GB  FADT.DSDT  FADT.X_DSDT    [VARIANT B]
  --------------  ---------  -----------
  yes             set        clear
  no              clear      set

  This behavior conforms to 5.1 Errata B. (And it's not required by
  earlier versions of the spec.)

- If the caller passes in the tables in "FADT, DSDT" relative order, then
  we do not enforce the exclusion:

  DSDT under 4GB  FADT.DSDT  FADT.X_DSDT    [VARIANT A]
  --------------  ---------  -----------
  yes             set        set
  no              clear      set

  This satisfies 5.1 Errata A and earlier, but breaks 5.1 Errata B and
  later.

Unify the handling of both relative orders. In particular, check the major
and minor version numbers in the FADT. If the FADT version is strictly
before 5.1, then implement [VARIANT A]. If the FADT version is equal to or
larger than 5.1, then implement [VARIANT B].

We make three observations:

- We can't check the FADT table version precisely against "5.1 Errata B";
  erratum levels are not captured in the table. We err in the safe
  direction, namely we enforce the exclusion for "5.1" and "5.1 Errata A".

- The same applies to "6.0" versus "6.0 Errata A". Because we cannot
  distinguish these two, we consider "6.0" to be "equal to or larger than
  5.1", and apply [VARIANT B], enforcing the exclusion.

- While a blanket [VARIANT B] would be simpler, there is a significant
  benefit to [VARIANT A], under the spec versions that permit it:
  compatibility with a wider range of OSPMs (typically, older ones).

  For example, Igor reported about a "DELL R430 system with rev4 FADT
  where DSDT and X_DSDT are pointing to the same address". Michael also
  reported about several systems that exhibit the same.

Regression tested with the following KVM guests (QEMU built at
ata0def594286d, "Merge remote-tracking branch
'remotes/bonzini/tags/for-upstream' into staging", 2017-01-30):

- OVMF: boot and S3 suspend/resume
  - Ia32, Q35, SMM
    - Fedlet 20141209
  - Ia32X64, Q35, SMM
    - Fedora 22
    - Windows 7
    - Windows 8.1
    - Windows 10
    - Windows Server 2008 R2
    - Windows Server 2012 R2
    - Windows Server 2016 Tech Preview 4
  - X64, I440FX, no SMM
    - Fedora 24
    - RHEL-6.7
    - RHEL-7.2-ish
- ArmVirtQemu: boot test with virtio-gpu
  - AARCH64
    - Fedora 24
    - RHELSA-7.3
    - openSUSE Tumbleweed (4.8.4-based)

This change is connected to ASWG ticket
<https://mantis.uefi.org/mantis/view.php?id=1757>, which is now
closed/fixed.

Cc: Al Stone <ahs3@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Leo Duran <leo.duran@amd.com>
Cc: Michael Tsirkin <mtsirkin@redhat.com>
Cc: Phil Dennis-Jordan <phil@philjordan.eu>
Cc: Star Zeng <star.zeng@intel.com>
Reported-by: Phil Dennis-Jordan <phil@philjordan.eu>
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Phil Dennis-Jordan <phil@philjordan.eu>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/AcpiTableDxe: condense whitespace around FADT.{DSDT,X_DSDT}
Laszlo Ersek [Wed, 1 Feb 2017 14:10:02 +0000 (15:10 +0100)]
MdeModulePkg/AcpiTableDxe: condense whitespace around FADT.{DSDT,X_DSDT}

This patch incurs no functional changes, it just removes some whitespace,
and also makes sure we always assign

  AcpiTableInstance->Fadt3->Dsdt

first, and

  AcpiTableInstance->Fadt3->XDsdt

second. The goal is to separate the syntactic changes from the functional
changes implemented by the next patch.

Cc: Al Stone <ahs3@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Feng Tian <feng.tian@intel.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Leo Duran <leo.duran@amd.com>
Cc: Michael Tsirkin <mtsirkin@redhat.com>
Cc: Phil Dennis-Jordan <phil@philjordan.eu>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Feng Tian <feng.tian@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoBaseTools/UPT: Fix an issue of adding Event twice
Hess Chen [Tue, 7 Mar 2017 08:01:25 +0000 (16:01 +0800)]
BaseTools/UPT: Fix an issue of adding Event twice

Fix the issue of after installing a package the Event information is duplicated. The tool checks if the EVENT information existing in UserExtension or not. If already existing in UserExtension the tool will not add additional information.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hess Chen <hesheng.chen@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools/UPT: Man Page Update
Hess Chen [Tue, 7 Mar 2017 06:15:13 +0000 (14:15 +0800)]
BaseTools/UPT: Man Page Update

Update UPT man page to complete the description of options.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hess Chen <hesheng.chen@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools/UPT: Add a checkpoint for missing '"'
Hess Chen [Tue, 7 Mar 2017 06:01:34 +0000 (14:01 +0800)]
BaseTools/UPT: Add a checkpoint for missing '"'

Add a checkpoint for UNI file which is missing '"' at the end of a line.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hess Chen <hesheng.chen@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools/ECC: Fix an issue of parameter parser
Hess Chen [Tue, 7 Mar 2017 06:01:33 +0000 (14:01 +0800)]
BaseTools/ECC: Fix an issue of parameter parser

The original solution of getting parameter name is to skip "_" but
now it doesn't ignore "_" anymore.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hess Chen <hesheng.chen@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoArmPkg/CpuDxe: handle implied attributes in EfiAttributeToArmAttribute
Ard Biesheuvel [Wed, 8 Mar 2017 11:46:38 +0000 (12:46 +0100)]
ArmPkg/CpuDxe: handle implied attributes in EfiAttributeToArmAttribute

Some memory attributes are implied by the memory type, e.g., device memory
is always mapped non-executable and cached memory should have the inner
shareable attribute.

In order to prevent unnecessary memory attribute updates of mappings
created early on, make EfiAttributeToArmAttribute() return these implied
attributes in the same way as ArmMmuLib does already. This avoids false
positives when looking for differences between current and desired mapping
attributes.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmVirtPkg: apply PE/COFF memory protection to DxeCore as well
Ard Biesheuvel [Tue, 7 Mar 2017 11:24:57 +0000 (12:24 +0100)]
ArmVirtPkg: apply PE/COFF memory protection to DxeCore as well

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>
4 years agoArmPkg/UncachedMemoryAllocationLib: map uncached allocations non-executable
Ard Biesheuvel [Tue, 28 Feb 2017 12:13:12 +0000 (12:13 +0000)]
ArmPkg/UncachedMemoryAllocationLib: map uncached allocations non-executable

The primary use case for UncachedMemoryAllocationLib is non-coherent DMA,
which implies that such regions are not used to fetch instructions from.

So let's map them as non-executable, to avoid creating a security hole
when the rest of the platform may be enforcing strict memory permissions
on ordinary allocations.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPkg/UncachedMemoryAllocationLib: use CWG value to align pool allocations
Ard Biesheuvel [Tue, 28 Feb 2017 12:13:11 +0000 (12:13 +0000)]
ArmPkg/UncachedMemoryAllocationLib: use CWG value to align pool allocations

Uncached pool allocations are aligned to the data cache line length under
the assumption that this is sufficient to prevent cache maintenance from
corrupting adjacent allocations. However, the value to use in such cases
is architecturally called the Cache Writeback Granule (CWG), which is
essentially the maximum Dcache line length rather than the minimum.

Note that this is mostly a cosmetical fix, given that the pool allocation
is turned into a page allocation later, and rounded up accordingly.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPkg/UncachedMemoryAllocationLib: restore mapping attributes after free
Ard Biesheuvel [Tue, 28 Feb 2017 12:13:10 +0000 (12:13 +0000)]
ArmPkg/UncachedMemoryAllocationLib: restore mapping attributes after free

In order to play nice with platforms that use strict memory permission
policies, restore the original mapping attributes when freeing uncached
allocations.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoEmulatorPkg: Add the UefiBootManagerLib library
Dandan Bi [Tue, 7 Mar 2017 02:13:58 +0000 (10:13 +0800)]
EmulatorPkg: Add the UefiBootManagerLib library

Since shell has consumed the APIs in UefiBootManagerLib.

Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Andrew Fish <afish@apple.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: Refine casting result to bigger size
Hao Wu [Tue, 7 Mar 2017 01:00:48 +0000 (09:00 +0800)]
UefiCpuPkg/PiSmmCpuDxeSmm: Refine casting result to bigger size

The commit is a follow-up of commit 8491e30.

In file MpService.c line 786:
Pte[Index] = (UINT64)((UINTN)PageTable + EFI_PAGE_SIZE * (Index + 1)) |
mAddressEncMask ...

(Where PageTable is of type VOID*, Index is of type UINTN, mAddressEncMask
is of type UINT64 and Pte[Index] is of type UINT64.)

Since in this case, the code logic ensures that the expression will not
exceed the range of UINTN, the commit will remove the explicit type cast
'(UINT64)'.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
4 years agoMdeModulePkg/DxeCore: Fix coding style issues
Bi, Dandan [Fri, 3 Mar 2017 04:44:46 +0000 (12:44 +0800)]
MdeModulePkg/DxeCore: Fix coding style issues

Add comments for functions.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg: Add PROMPT&HELP string of pcd to UNI file
Bi, Dandan [Fri, 3 Mar 2017 04:44:45 +0000 (12:44 +0800)]
MdeModulePkg: Add PROMPT&HELP string of pcd to UNI file

Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Leo Duran <leo.duran@amd.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoArmVirtPkg: enable non-executable DXE stack for all platforms
Ard Biesheuvel [Wed, 1 Mar 2017 16:31:43 +0000 (16:31 +0000)]
ArmVirtPkg: enable non-executable DXE stack for all platforms

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>