Alex James [Mon, 26 Mar 2018 16:15:10 +0000 (00:15 +0800)]
CorebootPayloadPkg: Conditionally add DebugAgentLib for DXE drivers
To fix building with SOURCE_DEBUG_ENABLE, add DebugAgentLib for
LibraryClasses.common.DXE_DRIVER, as is done with Vlv2TbltDevicePkg.
Cc: Maurice Ma <maurice.ma@intel.com> Cc: Prince Agyeman <prince.agyeman@intel.com> Cc: Benjamin You <benjamin.you@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Alex James <theracermaster@gmail.com> Reviewed-by: Benjamin You <benjamin.you@intel.com>
Bin Wang [Mon, 26 Feb 2018 08:19:30 +0000 (16:19 +0800)]
BaseTools: compare GUID value should not case-sensitive
build report error when the same Guid value in FDF file use lowercase,
in tools_def.txt file use uppercase.
The guid value's compare should not case-sensitive.
Carsey, Jaben [Wed, 28 Mar 2018 16:04:57 +0000 (00:04 +0800)]
BaseTools: FdfParser and FdfParserLite share reg exp
FdfParser can share regular expression from FdfParserLite.
reduce overlap and reduce recompile of the same expression.
v2: fix missed replacement of Pattern with shared variable
* Remove references to drivers that are not present
* Enable the Tianocore boot logo and BGRT table
* Enable full UEFI shell
* Enable more debug messages in DXE phase
Cc: David Wei <david.wei@intel.com> Cc: Mang Guo <mang.guo@intel.com> Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com> Reviewed-by: David Wei <david.wei@intel.com>
There are use cases where a VOID * PCD needs to be generated from multiple
binary input files. This can be in the form of an array of fixed size
elements or a set of variable sized elements.
Update BinToPcd to support multiple one or more -i INPUTFILE arguments.
By default, the contents of each binary input file are concatenated in
the order provided. This supports generating a PCD that is an array of
fixed size elements
Add -x, --xdr flags to BinToPcd to encodes the PCD using the
Variable-Length Opaque Data of RFC 4506 External Data Representation
Standard (XDR).
The data format from RFC 4506 meets the requirements for a PCD that is a
set of variable sized elements in the Variable-Length Opaque Data format.
The overhead of this format is a 32-bit length and 0 to 3 bytes of padding
to align the next element at a 32-bit boundary.
Cc: Sean Brogan <sean.brogan@microsoft.com> Cc: Yonghong Zhu <yonghong.zhu@intel.com> Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com> Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
Laszlo Ersek [Wed, 28 Mar 2018 11:26:35 +0000 (13:26 +0200)]
OvmfPkg/PlatformPei: debug log "etc/reserved-memory-end" from fw_cfg
QEMU calculates the UINT64 value in "etc/reserved-memory-end" in a quite
complex way, in the pc_memory_init() function. Log the value as a
DEBUG_VERBOSE message to support debugging.
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org> Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.redhat.com/show_bug.cgi?id=1353591
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Laszlo Ersek <lersek@redhat.com> Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
If the platform creates the "TlsCaCertificate" variable as volatile, then
EnrollX509toVariable() shouldn't fail to extend it just because
TLS_AUTH_CONFIG_VAR_BASE_ATTR contains the EFI_VARIABLE_NON_VOLATILE
attribute.
Thus, if the variable exists, add the EFI_VARIABLE_APPEND_WRITE attribute
to the variable's current attributes. This is what DeleteCert() does
already.
Laszlo Ersek [Thu, 22 Mar 2018 11:00:55 +0000 (12:00 +0100)]
NetworkPkg/HttpDxe: sanity-check the TlsCaCertificate variable before use
In TlsConfigCertificate(), make sure that the set of EFI_SIGNATURE_LIST
objects that the platform stored to "TlsCaCertificate" is well-formed.
In addition, because HttpInstance->TlsConfiguration->SetData() expects
X509 certificates only, ensure that the EFI_SIGNATURE_LIST objects only
report X509 certificates, as described under EFI_CERT_X509_GUID in the
UEFI-2.7 spec.
Laszlo Ersek [Thu, 22 Mar 2018 09:59:59 +0000 (10:59 +0100)]
NetworkPkg/HttpDxe: drop misleading comment / status code in cert config
For TlsConfigureSession(), it makes sense to exempt EFI_NOT_FOUND from
TlsConfigCipherList() / gRT->GetVariable(), because there is a default
cipher list (SSL_DEFAULT_CIPHER_LIST) we can fall back to.
The same is not true of TlsConfigCertificate(), because there is no
default CA cert list. The platform (or the user of the Setup utility) is
required to configure a CA cert list first.
Remove the misleading comment and status code mapping in
TlsConfigCertificate().
Laszlo Ersek [Thu, 22 Mar 2018 13:22:40 +0000 (14:22 +0100)]
NetworkPkg/HttpBootDxe: fix typo in DHCPv4 packet parsing
The string "HTTPClient" has 10 non-NUL characters; the value 9 is a
copy-paste leftover from "PXEClient". Check for all 10 characters in the
vendor-class-identifier option when determining whether the DHCP offer is
an HTTP offer.
Dandan Bi [Fri, 23 Mar 2018 03:21:56 +0000 (11:21 +0800)]
OvmfPkg:Fix VS2012 build failure
Initialize local variable to suppress warning C4701/C4703:
potentially uninitialized local variable/pointer variable.
1.In VirtualMemory.c:
Read of "PageMapLevel4Entry" in SetMemoryEncDe() is only
reached when "PageMapLevel4Entry" is got correctly.
2.In VirtioBlk.c:
Reads (dereferences) of "BufferMapping" and "BufferDeviceAddress"
in SynchronousRequest() are only reached if "BufferSize > 0" *and*
we map the data buffer successfully.
3.In VirtioScsi.c:
Reads (dereferences) of "InDataMapping" and "InDataDeviceAddress",
in VirtioScsiPassThru() are only reached if
"Packet->InTransferLength > 0" on input, *and* we map the
input buffer successfully. The similar reason for "OutDataMapping"
and "OutDataDeviceAddress".
Liming Gao [Tue, 27 Mar 2018 02:29:48 +0000 (10:29 +0800)]
BaseTools: Update Rsa2048Sha256Sign to use openssl standard options
sha256 is not the standard option. It should be replaced by sha -sha256.
Otherwise, it doesn't work in MAC OS.
In V2, update the option to sha1 -sha256.
In late openssl version >= 1.1, there is no sha option, but has sha1,sha256.
In previous openssl version < 1.1, there is no sha256, but has sha,sha1.
To work with all openssl version, use sha1 -sha256 for it.
Booting from USB may fail while the macro USB_BOOT_IO_BLOCKS
set to 128 because the block size of some USB devices are exceeded
512, like some virtual CD-ROM from BMC, the block size is 2048.
So,the count blocks to transfer should be calculated by block
size of the USB devices.
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Ming Huang <ming.huang@linaro.org> Reviewed-by: Star Zeng <star.zeng@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Commit 5563281fa2b31093a1cbd415553b9264c5136e89
* ShellPkg/[hex]edit: use SimpleTextInEx to read console
changes shell edit and hexedit to read input through SimpleTextInEx.
It exposes a issue in UsbKeyboard driver:
Per UEFI Spec,
When interpreting the data from this function (ReadKeyStrokeEx), it
should be noted that if a class of printable characters that are
normally adjusted by shift modifiers (e.g. Shift Key + "f" key) would
be presented solely as a KeyData.Key.UnicodeChar without the
associated shift state. So in the previous example of a Shift Key +
"f" key being pressed, the only pertinent data returned would be
KeyData.Key.UnicodeChar with the value of "F".
UsbKeyboard driver does convert Shift Key + "f" to "F" without the
shift state. But it doesn't do the conversion for all printable
characters, e.g.: Shift Key + "1" --> "!".
The root cause is today's logic to check whether a character is
printable or not is as below:
if ((KeyDescriptor->AffectedAttribute & EFI_AFFECTED_BY_CAPS_LOCK)
!= 0) {
So it only converts Shift + "a"-"z", but doesn't for Shift + "0"-"9",
and Shift + "["...
The patch updates the check logic as below to fix the issue:
if ((KeyDescriptor->Unicode != CHAR_NULL) &&
(KeyDescriptor->ShiftedUnicode != CHAR_NULL) &&
(KeyDescriptor->Unicode != KeyDescriptor->ShiftedUnicode)) {
The above check is TRUE when the character is printable and
it's *really* affected by Shift key.
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com> Reviewed-by: Star Zeng <star.zeng@intel.com>
When -G option is used to convert a BMP file to a UX capsule,
the bitmap is centered horizontally and placed in the lower
half of the screen below the boot logo.
This matches examples shown in the following pages:
Checks are also made to make sure the bitmap provided
fits in the current GOP mode.
Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Fix pointer math when more than one capsule is passed
to the CapsuleApp. Use the ContinuationPointer from
the last array entry instead of the first array entry.
Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Jiaxin Wu [Tue, 13 Mar 2018 08:53:18 +0000 (16:53 +0800)]
NetworkPkg/UefiPxeBcDxe: Configure the ARP Instance/RouteTable with new address
After completed a DHCP D.O.R.A process and got the new address, the ARP Instance
and RouteTable should be configured so as to avoid the later Pxe.Arp failure.
Cc: Fu Siyuan <siyuan.fu@intel.com> Cc: Ye Ting <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0 Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com> Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Laszlo Ersek [Wed, 14 Mar 2018 09:58:34 +0000 (10:58 +0100)]
BaseTools/PosixLike: honor pre-set PYTHONPATH
Utilities written in Python may depend on external (preinstalled) Python
packages; for example, Ecc depends on "antlr_python_runtime-3.0.1". Such
packages need not be installed system-wide, as long as they are reachable
through PYTHONPATH. Therefore we shouldn't overwrite the user's PYTHONPATH
with "BaseTools/Source/Python"; instead, we should prepend the latter to
the former.
TPM2.0 command lib always assumes TPM device and transmission channel can
respond correctly. But it is not true when communication channel is exploited
and wrong data is spoofed. Add more logic to prohibit memory overflow attack.
Cc: Long Qin <qin.long@intel.com> Cc: Yao Jiewen <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Chao Zhang <chao.b.zhang@intel.com> Reviewed-by: Long Qin <qin.long@intel.com> Reviewed-by: Yao Jiewen <jiewen.yao@intel.com>
Yonghong Zhu [Wed, 21 Mar 2018 02:36:59 +0000 (10:36 +0800)]
BaseTools: Override Max size by build Option Pcd for HII type
Current code will generate maxsize for HII type PCD when parser DSC
file, while this HII type PCD value maybe override in build command
per --pcd option, so the max size need re-calculate.
Achin Gupta [Thu, 15 Feb 2018 13:31:50 +0000 (13:31 +0000)]
Maintainers.txt: Add StandaloneMmPkg and maintainers
This patch adds maintainers, reviewer and directory for the
StandaloneMmPkg. This package will host an implementation of Standalone
Management Mode as specified in the Platform Initialization (PI)
Specification, Volume 4: Management Mode Core Interface.
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Achin Gupta <achin.gupta@arm.com> Reviewed-by: Jiewen.yao@intel.com Reviewed-by: Michael D Kinney <michael.d.kinney@intel.com> Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Star Zeng [Mon, 19 Mar 2018 12:01:58 +0000 (20:01 +0800)]
SourceLevelDebugPkg DebugCommUsb3: Return error when debug cap is reset
When source level debug is enabled, but debug cable is not connected,
XhcResetHC() in XhciReg.c will reset the host controller, the debug
capability registers will be also reset. After the code in
InitializeUsbDebugHardware() sets DCE bit and LSE bit to "1" in DCCTRL,
there will be DMA on 0 (the value of some debug capability registers
for data transfer is 0) address buffer, fault info like below will
appear when IOMMU based on VTd is enabled.
Yonghong Zhu [Tue, 20 Mar 2018 05:16:12 +0000 (13:16 +0800)]
BaseTools: Fix bug for --pcd VOID* type when no max size is specified
when VOID* type non-structure pcd used in --pcd, and its max size is not
specified in DSC or its value is hex value, build break due to the code
int(Pcd.MaxDatumSize,10).
Now this patch remove this code, because tool will calculate the size
info in later phase.
Hao Wu [Mon, 19 Mar 2018 07:02:59 +0000 (15:02 +0800)]
MdeModulePkg/DxeMain: Fix BSP interrupts reenabled in ExitBootServices
Within function CoreExitBootServices(), this commit will move the call
of:
MemoryProtectionExitBootServicesCallback();
before:
SaveAndSetDebugTimerInterrupt (FALSE);
and
gCpu->DisableInterrupt (gCpu);
The reason is that, within MemoryProtectionExitBootServicesCallback(),
APIs like RaiseTpl and RestoreTpl maybe called. An example will be:
DebugLib (using PeiDxeDebugLibReportStatusCode instance)
|
v
ReportStatusCodeLib (using DxeReportStatusCodeLib instance)
|
v
Raise/RestoreTpl
The call of Raise/RestoreTpl APIs will re-enable BSP interrupts. Hence,
this commit refine the calling sequence to ensure BSP interrupts before
leaving CoreExitBootServices().
Cc: Jian J Wang <jian.j.wang@intel.com> Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Hao Wu <hao.a.wu@intel.com> Reviewed-by: Jiewen Yao <jiewen.yao@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Hao Wu [Mon, 19 Mar 2018 05:00:13 +0000 (13:00 +0800)]
UefiCpuPkg/MpInitLib: Disable interrupt at ExitBootServices AP Mwait
Within function ApWakeupFunction():
When source level debugger is enabled, AP interrupts will be enabled by
EnableDebugAgent(). Then the AP function will be executed by:
Procedure (Parameter);
After the AP function returns, AP interrupts will be disabled when the
APs are placed in loop mode (both HltLoop and MwaiLoop).
However, at ExitBootServices, ApWakeupFunction() is called with
'Procedure' equals to RelocateApLoop().
(ExitBootServices callback registered within InitMpGlobalData())
RelocateApLoop() never returns, so it has to disable the AP interrupts by
itself. However, we find that interrupts are only disabled for the
HltLoop case, but not for the MwaitLoop case (within file MpFuncs.nasm).
This commit adds the missing disabling of AP interrupts for MwaitLoop.
Also, for X64, this commit will disable the interrupts before switching to
32-bit mode.
Cc: Laszlo Ersek <lersek@redhat.com> Cc: Jian J Wang <jian.j.wang@intel.com> Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Hao Wu <hao.a.wu@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com> Reviewed-by: Jiewen Yao <jiewen.yao@intel.com> Reviewed-by: Jeff Fan <vanjeff_919@hotmail.com>
Pete Batard [Fri, 23 Feb 2018 09:50:03 +0000 (17:50 +0800)]
BaseTools/Conf: Add VS2017/ARM64 support
Build options for ARM64 are the same as for ARM, except for /BASE:0
which is removed from DLINK flags to avoid LNK1355 error:
invalid base address 0x0; ARM64 image cannot have base address below 4GB
Pete Batard [Fri, 23 Feb 2018 09:50:02 +0000 (17:50 +0800)]
MdePkg/Include: Add VA list support for VS2017/ARM64
We need to explicitly call the built-in __va_start() for ARM64, otherwise
the variable parameters are not properly enqueued for the next function
calls.
Also do the same for ARM, as as it doesn't harm us.
Yonghong Zhu [Thu, 15 Mar 2018 06:40:57 +0000 (14:40 +0800)]
BaseTools: Add PackageDocumentTools into Scripts folder
This tool is used to generate the document for edk2 packages. The
generated document will be in UDK release. For example, UDK2017
document can be found in:
https://github.com/tianocore/tianocore.github.io/wiki/UDK2017#documentation
Yonghong Zhu [Thu, 15 Mar 2018 11:40:18 +0000 (19:40 +0800)]
BaseTools: Fix bug for VOID* DynamicDefault Pcd use Flexible format
define a flexible pcd format in Dyanmic/DynamicExDefault section,
it cause build error.
[PcdsDynamicExDefault.common.DEFAULT]
pcdToken.Name|{GUID("11111111-2222-42eb-b5eb-fef31d207cb4")}
Option --hash --binary-destination generate Binaries section in
the inf file, but the path of ASL file is begin with
Output directory, so need replace Output directory with '',
will get the file name RamDisk.aml
Incorrect AML file path in inf file on linux:
[Binaries.X64]
PE32|RamDiskDxe.efi
Star Zeng [Thu, 15 Mar 2018 05:50:31 +0000 (13:50 +0800)]
SecurityPkg OpalPasswordDxe:Fix wrong BufferSize input to UnicodeSPrint
Current code uses string length as BufferSize input to UnicodeSPrint,
it is wrong and makes the pop up string trimmed. The BufferSize input
to UnicodeSPrint should be the size, in bytes, of the output buffer.
This is to use sizeof (mPopUpString) as the BufferSize input to
UnicodeSPrint, it also updates array size of mPopUpString from 256 to
100 that is enough, otherwise the pop up string may be too long.
Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Eric Dong <eric.dong@intel.com> Cc: Chao Zhang <chao.b.zhang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Star Zeng <star.zeng@intel.com> Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Laszlo Ersek [Thu, 15 Mar 2018 11:49:26 +0000 (12:49 +0100)]
OvmfPkg/PlatformBootManagerLib: process "-kernel" before boot devices
This improves the UEFI boot time for VMs that have "-kernel", many disks
or NICs, and no "bootindex" properties.
(Unlike in ArmVirt commit 23d04b58e27b, in OvmfPkg commit 52fba28994e9 we
introduced TryRunningQemuKernel() right from the start *after*
BdsLibConnectAll(). Therefore, unlike in patch
'ArmVirtPkg/PlatformBootManagerLib: return to "-kernel before boot
devices"', we adopt the logic as new in this patch.)
Laszlo Ersek [Thu, 15 Mar 2018 13:57:39 +0000 (14:57 +0100)]
OvmfPkg/PlatformBootManagerLib: rejuvenate old-style function comments
The old-style "Routine Description: ..." comments use the leftmost column
and are placed between the parameter list and the function body. Therefore
they cause git-diff to produce bogus hunk headers that fail to name the
function being patched.
Convert these comment blocks to the current edk2 style. While at it, clean
them up too.
For PlatformBootManagerBeforeConsole() and
PlatformBootManagerAfterConsole(), copy the descriptions from the call
sites in "MdeModulePkg/Universal/BdsDxe/BdsEntry.c". They are more
detailed than the comments in the lib class header
"MdeModulePkg/Include/Library/PlatformBootManagerLib.h"; ArmVirtPkg
already uses these comments.
Laszlo Ersek [Thu, 15 Mar 2018 11:49:26 +0000 (12:49 +0100)]
ArmVirtPkg/PlatformBootManagerLib: return to "-kernel before boot devices"
Move the TryRunningQemuKernel() call back to its original place. This
improves the UEFI boot time for VMs that have "-kernel", many disks or
NICs, and no "bootindex" properties. A well-known example is
guestfish/libguestfs.
Jian J Wang [Thu, 15 Mar 2018 06:19:00 +0000 (14:19 +0800)]
MdeModulePkg/PiSmmCore: fix #PF caused by freeing read-only memory
SMM core will add a HEADER before each allocated pool memory and clean
up this header once it's freed. If a block of allocated pool is marked
as read-only after allocation (EfiRuntimeServicesCode type of pool in
SMM will always be marked as read-only), #PF exception will be triggered
during memory pool freeing.
Normally EfiRuntimeServicesCode type of pool should not be freed in the
real world. But some test suites will actually do memory free for all
types of memory for the purpose of functionality and conformance test.
So this issue should be fixed anyway.
Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Jian J Wang <jian.j.wang@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Jian J Wang [Thu, 15 Mar 2018 04:45:43 +0000 (12:45 +0800)]
MdeModulePkg/Core: fix bits operation error on a boundary condition
If given address is on 64K boundary and the requested bit number is 64,
all SetBits(), ClearBits() and GetBits() will encounter ASSERT problem
in trying to do a 64 bits of shift, which is not allowed by LShift() and
RShift(). This patch tries to fix this issue by turning bits operation
into whole integer operation in such situation.
Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Jian J Wang <jian.j.wang@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Jian J Wang [Thu, 15 Mar 2018 04:43:20 +0000 (12:43 +0800)]
MdeModulePkg/PiSmmCore: fix bits operation error on a boundary condition
If given address is on 64K boundary and the requested bit number is 64,
all SetBits(), ClearBits() and GetBits() will encounter ASSERT problem
in trying to do a 64 bits of shift, which is not allowed by LShift() and
RShift(). This patch tries to fix this issue by turning bits operation
into whole integer operation in such situation.
Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Cc: Jiewen Yao <jiewen.yao@intel.com> Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Jian J Wang <jian.j.wang@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Jian J Wang [Wed, 14 Mar 2018 08:28:34 +0000 (16:28 +0800)]
MdeModulePkg/Core: allow HeapGuard even before CpuArchProtocol installed
Due to the fact that HeapGuard needs CpuArchProtocol to update page
attributes, the feature is normally enabled after CpuArchProtocol is
installed. Since there're some drivers are loaded before CpuArchProtocl,
they cannot make use HeapGuard feature to detect potential issues.
This patch fixes above situation by updating the DXE core to skip the
NULL check against global gCpu in the IsMemoryTypeToGuard(), and adding
NULL check against gCpu in SetGuardPage() and UnsetGuardPage() to make
sure that they can be called but do nothing. This will allow HeapGuard to
record all guarded memory without setting the related Guard pages to not-
present.
Once the CpuArchProtocol is installed, a protocol notify will be called
to complete the work of setting Guard pages to not-present.
Please note that above changes will cause a #PF in GCD code during cleanup
of map entries, which is initiated by CpuDxe driver to update real mtrr
and paging attributes back to GCD. During that time, CpuDxe doesn't allow
GCD to update memory attributes and then any Guard page cannot be unset.
As a result, this will prevent Guarded memory from freeing during memory
map cleanup.
The solution is to avoid allocating guarded memory as memory map entries
in GCD code. It's done by setting global mOnGuarding to TRUE before memory
allocation and setting it back to FALSE afterwards in GCD function
CoreAllocateGcdMapEntry().
Cc: Star Zeng <star.zeng@intel.com> Cc: Eric Dong <eric.dong@intel.com> Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Jian J Wang <jian.j.wang@intel.com> Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
projects / mirror_edk2.git / log